Efficiency in Learning: Evidence-Based Guidelines to Manage Cognitive Load

ffirs.qxd 9/29/05 19:09 Page v Efﬁciency in Learning Evidence-Based Guidelines to Manage Cognitive Load Ruth Colvin C...

Author: Ruth C. Clark | Frank Nguyen | John Sweller

82 downloads 2006 Views 6MB Size Report

This content was uploaded by our users and we assume good faith they have the permission to share this book. If you own the copyright to this book and it is wrongfully on our website, we offer a simple DMCA procedure to remove your content from our site. Start by pressing the button below!

Report copyright / DMCA form

DOWNLOAD PDF

ffirs.qxd 9/29/05 19:09 Page v

Efﬁciency in

Learning Evidence-Based Guidelines to Manage Cognitive Load

Ruth Colvin Clark Frank Nguyen John Sweller

ffirs.qxd 9/29/05 19:09 Page vi

Copyright © 2006 by John Wiley & Sons, Inc. Published by Pfeiffer An Imprint of Wiley. 989 Market Street, San Francisco, CA 94103-1741 www.pfeiffer.com Except as specifically noted below, no part of this publication may be reproduced, stored in a retrieval system, or transmitted in any form or by any means, electronic, mechanical, photocopying, recording, scanning, or otherwise, except as permitted under Section 107 or 108 of the 1976 United States Copyright Act, without either the prior written permission of the Publisher, or authorization through payment of the appropriate per-copy fee to the Copyright Clearance Center, Inc., 222 Rosewood Drive, Danvers, MA 01923, 978-750-8400, fax 978-646-8600, or on the web at www.copyright.com. Requests to the Publisher for permission should be addressed to the Permissions Department, John Wiley & Sons, Inc., 111 River Street, Hoboken, NJ 07030, phone 201-748-6011, fax 201-748-6008, or online at http://www.wiley.com/go/permission. Limit of Liability/Disclaimer of Warranty: While the publisher and author have used their best efforts in preparing this book, they make no representations or warranties with respect to the accuracy or completeness of the contents of this book and specifically disclaim any implied warranties of merchantability or fitness for a particular purpose. No warranty may be created or extended by sales representatives or written sales materials. The advice and strategies contained herein may not be suitable for your situation. You should consult with a professional where appropriate. Neither the publisher nor author shall be liable for any loss of profit or any other commercial damages, including but not limited to special, incidental, consequential, or other damages. Readers should be aware that Internet websites offered as citations and/or sources for further information may have changed or disappeared between the time this was written and when it is read. Certain pages from this book and all the materials on the accompanying CD-ROM are designed for use in a group setting and may be customized and reproduced for educational/training purposes. The reproducible pages are designated by the appearance of the following copyright notice at the foot of each page: Efficiency in Learning. Copyright © 2006 by John Wiley & Sons, Inc. Reproduced by permission of Pfeiffer, an Imprint of Wiley. www.pfeiffer.com This notice may not be changed or deleted and it must appear on all reproductions as printed. This free permission is restricted to limited customization of the CD-ROM materials for your organization and the paper reproduction of the materials for educational/training events. It does not allow for systematic or large-scale reproduction, distribution (more than 100 copies per page, per year), transmission, electronic reproduction or inclusion in any publications offered for sale or used for commercial purposes—none of which may be done without prior written permission of the Publisher. For additional copies/bulk purchases of this book in the U.S. please contact 800-274-4434. Pfeiffer books and products are available through most bookstores. To contact Pfeiffer directly call our Customer Care Department within the U.S. at 800-274-4434, outside the U.S. at 317-572-3985, fax 317-572-4002, or visit www.pfeiffer.com. Pfeiffer also publishes its books in a variety of electronic formats. Some content that appears in print may not be available in electronic books. Library of Congress Cataloging-in-Publication Data Clark, Ruth Colvin. Efficiency in learning: evidence-based guidelines to manage cognitive load / Ruth Colvin Clark, Frank Nguyen, John Sweller. p. cm. Includes bibliographical references and index. ISBN-13 978-0-7879-7728-3 (alk. paper) ISBN-10 0-7879-7728-4 (alk. paper) 1. Organizational learning. 2. Cognitive learning. I. Nguyen, Frank, 1975- II. Sweller, John, 1946- III. Title. HD58.82.C53 2006 658.3124—dc22

2005010921

Acquiring Editor: Matthew Davis Director of Development: Kathleen Dolan Davies Production Editor: Dawn Kilgore Printed in the United States of America Printing

10

9

8

7

6

5

4

3

2

1

Editor: Rebecca Taff Manufacturing Supervisor: Becky Carreño

ftoc.qxd 9/29/05 19:20 Page vii

CONTENTS

Contents of the CD-ROM Acknowledgments Introduction

ix

xi

xiii

PART ONE. AN INTRODUCTION TO EFFICIENCY IN LEARNING 1 1. Cognitive Load and Efﬁciency in Learning 2. The Psychology of Efﬁciency

5

27

PART TWO. BASIC GUIDELINES FOR MANAGING (IRRELEVANT) COGNITIVE LOAD 43 3. Use Visuals and Audio Narration to Exploit Working Memory Resources 47 4. Focus Attention and Avoid Split Attention

77

5. Weed Your Training to Manage Limited Working Memory Capacity 107 6. Provide External Memory Support to Reduce Working Memory Load 139 vii

ftoc.qxd 9/29/05 19:20 Page viii

viii

Contents

7. Use Segmenting, Sequencing, and Learner Pacing to Impose Content Gradually 161 8. Transition from Worked Examples to Practice to Impose Mental Work Gradually 189 PART THREE. INSTRUCTIONAL GUIDELINES FOR IMPOSING RELEVANT COGNITIVE LOAD 213 9. Put Working Memory to Work with Germane Load 217 PART FOUR. TAILORING INSTRUCTION TO LEARNER EXPERTISE 243 10. Accommodate Differences in Learner Expertise 247 11. Use Rapid Testing to Adapt e-Learning to Learner Expertise 275 PART FIVE. COGNITIVE LOAD THEORY IN PERSPECTIVE 289 12. Applying Cognitive Load Theory 293 13. The Evolution of Cognitive Load Theory: A Personal Perspective by John Sweller 313 Appendix: All About the Numbers Glossary 341 References 353 About the Authors Index

365

367

List of Figures and Tables

379

How to Use the CD-ROM

389

331

ftoc.qxd 9/29/05 19:20 Page ix

CONTENTS OF THE CD-ROM

• Asynchronous e-learning demonstration lesson that applies our guidelines: After Load Managed Excel Web-Based Lesson • Commentary on this lesson by Dr. John Sweller • Asynchronous e-learning demonstration lesson that violates our guidelines: Before Overloaded Excel Web-Based Lesson • Commentary on this lesson by Dr. John Sweller • Synchronous e-learning demonstration lesson that applies our guidelines: Virtual Classroom Example • Commentary on this lesson by Dr. John Sweller • A video discussion by Dr. John Sweller on the research and guidelines in this book, organized by chapter • Working Aid: Applying Cognitive Load Theory to Your Training

ix

flast.qxd 9/29/05 19:10 Page x

Ruth Colvin Clark offers her contribution to this book in memory of her father:

Almer Paul Colvin 1915–2005

flast.qxd 9/29/05 19:10 Page xi

ACKNOWLEDGMENTS

T

THE AUTHORS ARE grateful to the following individuals who read through drafts of the manuscript and offered invaluable advice for improvement: Charlie Adamchik Deidre Breitfeller Vincent Iglesias-Cardinale Nancy Curtis Matthew Hanzel Gail Bloom Jeff Krebs Timothy W. Spannaus Leslie Stephen Jon Stephenson Benhong Rosaline Tsai

xi

flast.qxd 9/29/05 19:10 Page xii

xii

Acknowledgments

We also thank two individuals who have greatly contributed to this book. Matt Davis, our acquisition editor, understood the importance of cognitive load theory to recommend acceptance of our book proposal. Leslie Stephen, our developmental editor, gave us invaluable suggestions for improving the organization and readability of the first draft. We are also grateful to Barry Galloway, who provided the narration on our CD demonstrations, and Global Knowledge, who provided examples of their electronic performance support systems.

flast.qxd 9/29/05 19:10 Page xiii

INTRODUCTION Getting the Most from This Resource

Purpose In this era of information overload and greater work complexity, instructional professionals are increasingly asked to accelerate the speed and quality of the learning process. This book provides you with evidence-based guidelines on how to create efficient instructional environments which result in faster learning, better learning, or both. At the core of Efficiency in Learning are instructional guidelines that work in harmony with human learning processes.

Audience This book was written for all instructional professionals in organizational training settings who are seeking proven methods to create efficient learning environments. This includes classroom and synchronous e-learning instructors, as well as designers and developers of classroom and multimedia instructional materials.

xiii

flast.qxd 9/29/05 19:10 Page xiv

xiv

Introduction

Package Components In addition to the book, we include a CD with the following elements: • Asynchronous e-learning demonstration lesson that violates our guidelines: Before Overloaded Excel Web-Based Lesson • Commentary on this lesson by Dr. John Sweller • Asynchronous e-learning demonstration lesson that applies our guidelines: After Load Managed Excel Web-Based Lesson • Commentary on this lesson by Dr. John Sweller • Synchronous e-learning demonstration lesson that applies our guidelines: Virtual Classroom Example • Commentary on this lesson by Dr. John Sweller • A video discussion by Dr. John Sweller on the research and guidelines in this book organized by chapter • Working Aid: Applying Cognitive Load Theory to your Training The CD is used to illustrate instructional strategies that cannot be captured in a book, such as audio narration, color, animation, and adaptive testing. We provide specific references to various elements of the CD throughout the text and at the end of each chapter point you to related contents on the CD. We also recommend that instructors using this book as a class text make use of these resources as an integral part of their instructional plan. For example, a class introductory exercise could assign students to review the Before Overloaded asynchronous demonstration lesson that violates our guidelines and make a list/discuss what they think is wrong with the example.

Product Description The book is organized into five parts as follows:

Part I. An Introduction to Efﬁciency in Learning (Chapters 1 and 2) This part provides an introduction to efficient learning that introduces cognitive load theory and overviews the scientific definition and metric for

flast.qxd 9/29/05 19:10 Page xv

Introduction

xv

efficiency, as well as the human learning processes from which cognitive load theory is derived.

Part II. Basic Guidelines for Managing Irrelevant Cognitive Load (Chapters 3 through 8) This part provides guidelines and supporting research for the use of graphics, text, and audio proven to make learning more efficient. It also discusses ways to reduce memory load by (1) eliminating extraneous content or themes, (2) imposing mental work gradually through appropriate lesson organization, and (3) replacing some practice exercises with worked examples.

Part III. Instructional Guidelines for Imposing Relevant Cognitive Load (Chapter 9) This part focuses on proven ways to help learners build new knowledge and skills through the use of diverse examples, better learner processing of examples, and rehearsal of new content.

Part IV. Tailoring Instruction to Learner Expertise (Chapters 10 and 11) This part offers guidelines and examples of ways to tailor your training for novice and experienced learners. It summarizes a model for adaptation of instruction in asynchronous e-learning environments based on a new technique of rapid testing.

Part V. Cognitive Load Theory in Perspective (Chapters 12 and 13) This part provides a retrospective look at the previous chapters. In Chapter 12 we integrate all of our previous guidelines into a context familiar to instructional professionals, and in Chapter 13 John Sweller discusses the past, present, and future of cognitive load theory from a personal perspective.

Appendix The appendix provides mathematical details regarding calculation of the efficiency metric as well as a review of basic statistical concepts, including statistical significance and effect size.

flast.qxd 9/29/05 19:10 Page xvi

xvi

Introduction

Glossary The glossary provides definitions of the cognitive load terminology used throughout the book. Each chapter in the book includes: • Several guidelines for instructional professionals along with examples, the research, and psychology supporting those guidelines • Summaries and application checklists to aid instructional professionals to apply the guidelines • A guide to the resources on the CD relevant to that chapter • Recommended readings for individuals desiring more information on the chapter theme

Explanation of Cognitive Load Theory The guidelines and examples in this book are based on over twenty-five years of valid experimental research conducted by John Sweller and his associates throughout the world. Many training professionals will recall the recommendation to shape their instruction around the “magical number of 7 plus or minus 2” in order to avoid overloading their learners. Cognitive load theory is the 21st Century update to that maxim. Cognitive load theory is a comprehensive and proven instructional theory that illustrates ways to reduce unproductive forms of cognitive load and at the same time maximize productive sources of cognitive load that lead to efficient learning environments. Chapters 1 and 2 provide an overview of cognitive load theory, which is the foundation for this book.

c01.qxd 9/29/05 17:53 Page 1

PA R T O N E

An Introduction to Efﬁciency in Learning

N PART I WE INCLUDE two chapters that summarize the basics of

I

cognitive load theory and that describe its psychological basis. In an age in which workers must learn more in less time, more likely than not, you have taken, produced, or taught training classes that resulted in mental overload. However, you may not realize that cognitive load has a scientific basis— and even better—that research offers many proven techniques that you can use to manage cognitive load either as a consumer or a producer of training. By applying these techniques you can produce efficient instructional environments—environments that result in faster learning, better learning or both. Cognitive load theory has its modern origins in experiments conducted by Dr. John Sweller at the University of New South Wales, Australia, in the early 1980s. Today cognitive load theory has grown into one of the most widely recognized sets of proven principles governing learning and instruction in the training profession. You can read a personal account by

1

c01.qxd 9/29/05 17:53 Page 2

2

Efficiency in Learning

John Sweller of how cognitive load theory started and evolved in Chapter 13. Because most of the research on cognitive load theory is found in technical journals not written for training practitioners, we wrote this book for instructional professionals to summarize the principles and evidence behind it. Part I lays the foundation for these principles.

Read

To Learn About

Chapter 1. Cognitive What cognitive load theory is Load and Efficiency Three Types of Cognitive Load: Intrinsic, in Learning Extraneous (Irrelevant), and Germane (Relevant) How instructional environments interact with learner backgrounds and content complexity to result in cognitive load The psychological basis and research evidence for cognitive load theory A formal definition of efficiency in learning Chapter 2. The Psychology of Efficiency

Human memories involved in learning: working memory and long-term memory Cognitive load and working memory limits Long-term memory and expertise The main psychological events of learning Instructional methods and psychological events of learning

c01.qxd 9/29/05 17:53 Page 3

An Introduction to Efficiency in Learning

On the CD Video Interview with John Sweller: Chapter Preview/Review Chapter 1. Cognitive Load and Efficiency in Learning. The basics of cognitive load theory, including a discussion of extraneous, intrinsic, and germane cognitive load. The scientific basis for cognitive load theory. Chapter 2. The Psychology of Efficiency. The features of working memory, the consequences of limited working memory on human learning, the features of long-term memory and the importance of long-term memory to human cognition.

3

c01.qxd 9/29/05 17:53 Page 4

CHAPTER OUTLINE The Costs of Inefﬁcient Instruction What Is Cognitive Load Theory? A Deﬁnition of Cognitive Load Theory

Types of Cognitive Load Intrinsic Load Germane (Relevant) Load Extraneous (Irrelevant) Load Balancing Mental Load in Your Training

No Yellow Brick Road: The Relativity of Cognitive Load Cognitive Load Theory and Human Learning Evidence-Based Practice Evidence for Cognitive Load Theory About the Numbers Limits of Research

Quantifying Efﬁciency The Efﬁciency Graph

The Bottom Line On the CD John Sweller Video Interview Sample Excel e-Lessons

c01.qxd 9/29/05 17:53 Page 5

1

+

+

Reduce Increase Manage irrelevant relevant intrinsic load load load

Cognitive Load and Efﬁciency in Learning

= Efficient learning

I

NFORMATION OVERLOAD AND LARGE FINANCIAL INVESTMENTS in worker learning and training demand efﬁcient instructional envi-

ronments. Efﬁcient instructional environments lead to better learning, faster learning, or both because they make the best use of limited human cognitive capacity. This book offers practical proven guidelines to make your instruction efﬁcient. In this chapter we set the stage by introducing cognitive load theory, which is the scientiﬁc basis for efﬁciency in learning. We will look at three types of cognitive load you must consider in your training, as well as the variations in cognitive load resulting from the interaction among instructional environments, learner prior knowledge, and the complexity of the learning task. Unlike many books offering training tips and techniques, our guidelines are based on recent valid scientiﬁc evidence. We will introduce the type of evidence that we present throughout the book and, since cognitive load theory is fundamentally about efﬁciency, we will deﬁne efﬁciency and show how it is measured in research studies.

5

c01.qxd 9/29/05 17:53 Page 6

6

Efficiency in Learning

The Costs of Inefﬁcient Instruction This is a book about how to create efficient learning environments. The guidelines in this book apply to all types of instructional delivery media, including computers, workbooks, and instructors. Instructional settings that are efficient result in learning that is faster and/or better than settings that are inefficient. Many popular books on learning and training techniques are based on little more than personal opinion. In contrast, our guidelines are based on scientific evidence—evidence accumulated over the past twenty-five years by an international team of instructional scientists. This evidence has important economic implications. As a consequence of high investments made in training programs coupled with rampant information overload, inefficient instructional environments exert a high toll in wasted economic and human resources. How high is our training investment? From customer service to manufacturing—from sales to supervision—50 to 60 billion dollars are spent each year on organizational training programs in the United States alone (Dolezalek, 2004). And this is a low estimate because it does not factor in the hidden costs that make up the most expensive element of any training program—the salary time of participants being trained. While staff are attending a week of training, they are earning their salaries and they are not producing. Even if we disregard lost opportunity costs, just adding the salary costs alone would bring the annual investment in training into the $300 billion range! More efficient learning environments increase training cost effectiveness by reducing instructional time, improving training outcomes, or both. Psychological work demands are growing in the 21st Century. Whether you call it info glut or data delirium, information overload has gotten so bad that it’s led to a new form of psychological stress called Information Fatigue Syndrome. A study from the University of California at Berkeley reports that the amount of new information created in the year 2002 disseminated in print, film, magnetic, and optical storage media equaled five exabytes (Lyman & Varian, 2003). Five exabytes is equivalent to the information contained in half a million libraries the size of the U.S. Library of Congress print collection, which exceeds nineteen million books!

c01.qxd 9/29/05 17:53 Page 7

Cognitive Load and Efficiency in Learning

7

Information overload erodes the quality of work. For example, primary care physicians cited information overload as a major cause of difficulties practitioners experience in diagnosing and managing heart failure (Fuat, Hungin, & Murphy, 2003). Not only is the sheer amount of information growing, but so also are the complexity and number of tasks many workers must juggle. More information and more complex tasks demand greater skills, which require more training. At the same time, organizations want to save costs by reducing time spend in training programs. These economic and psychological pressures call for efficient training environments—environments that are proven to work in harmony with the strengths and limitations of human learning processes.

What Is Cognitive Load Theory? As instructional professionals, many of you have probably heard of the “magical number 7 2” items of information, first published by George Miller in 1956. According to this guideline, our cognitive system can only process 7 2 items at one time. Once we exceed those limits, our thinking and learning processes bog down. Based on research conducted over the past twenty-five years, a growing international contingent of instructional scientists has expanded and refined the rule of 7 2 into a comprehensive set of instructional principles called cognitive load theory.

A Deﬁnition of Cognitive Load Theory Cognitive load theory is a universal set of learning principles that are proven to result in efficient instructional environments as a consequence of leveraging human cognitive learning processes. 1. Cognitive Load Theory Is Universal. Cognitive load theory applies to all types of

content, all delivery media, and all learners. Because cognitive load theory addresses how to use fundamental tools of training—text, visuals, and audio—it applies to everything from technical content to soft skills as well as to all delivery platforms from print to e-learning. Because of its universality, whether you are a classroom instructor or developer of

c01.qxd 9/29/05 17:53 Page 8

8

Efficiency in Learning

training materials for workbooks or computers, cognitive load theory applies to you. 2. Cognitive Load Theory Offers Principles and Related Instructional Guidelines.

Unlike many general educational theories, cognitive load theory offers principles that lead to very specific guidelines that all instructional professionals can implement. Throughout the chapters in this book we offer more than twenty-five specific guidelines for best ways to design, develop, and present training. Some of these guidelines are likely to be familiar methods that you may have used for years. Other guidelines, however, will be new—some even counter to prevailing instructional practice. 3. Cognitive Load Theory Is Evidence-Based. Cognitive load theory is based on

dozens of controlled experimental research studies. Throughout the chapters we summarize some of the experiments and show you the results. Because so much training advice is not based on evidence, we feel it is important for you to have the opportunity to review at least some of the research that supports cognitive load theory. For more details, we offer recommended readings, many of which are original research reports. In Chapter 13, John Sweller, originator of cognitive load theory, writes a personal perspective of how cognitive load theory started and has evolved during the last twenty-five years. 4. Cognitive Load Theory Leads to Efﬁcient Learning. Efficient instructional

environments lead to faster learning, better learning, or both. The scientists who have worked on cognitive load theory have created a metric for quantifying efficiency as well as an efficiency graph for display and visual comparison of lesson efficiencies. Since you will see research data displayed on the efficiency graph throughout the book, we define and illustrate this metric and graph in this chapter. 5. Cognitive Load Theory Leverages Human Cognitive Learning Processes. Learning

environments based on cognitive load theory minimize wasted mental resources and instead put those limited mental resources to work in ways proven to maximize learning. Because cognitive load theory is grounded in human learning processes, you will not only gain a set of proven instructional guidelines, but you will also understand why those guidelines work. Based on

c01.qxd 9/29/05 17:53 Page 9

9

Cognitive Load and Efficiency in Learning

this understanding, you can readily adapt them to your own instructional settings. You can also explain the basis for your instructional recommendations to your colleagues and clients. As an incidental benefit, you should also gain insights into your own cognitive processes!

Types of Cognitive Load Some forms of cognitive load are useful, while others waste mental resources. Your goal during training is to minimize wasteful forms of cognitive load and maximize the useful forms. The three main types of cognitive load you must consider in your training program are intrinsic load, germane load, and extraneous load. Since total mental capacity is limited, you will need to balance these three forms of load to maximize learning efficiency.

Intrinsic Load Intrinsic load is the mental work imposed by the complexity of the content in your lessons and is primarily determined by your instructional goals. For example, in Figure 1.1 we show a practice assignment from an e-lesson on

Figure 1.1. An Assignment in an Excel Lesson That Imposes Moderate Intrinsic Cognitive Load. From the CD Virtual Classroom Example.

c01.qxd 9/29/05 17:53 Page 10

10

Efficiency in Learning

Excel® formulas drawn from our demonstration lesson on the CD. To perform this task, the learner must coordinate at least seven steps, including locating the correct spreadsheet row, locating the correct spreadsheet column, combining these to locate the correct spreadsheet cell in which to input a formula, selecting that cell with the mouse, constructing the correct formula by applying Excel format rules (which, depending on the formula, may involve many steps), typing the formula in the cell, and pressing the enter key. For someone new to Excel, this is a complex task because it requires the coordination of multiple mental and physical components. In cognitive load terminology, we would say that this assignment imposes a moderately high intrinsic load because it involves a high amount of element interactivity. Element interactivity simply means that several knowledge elements must be coordinated in memory to accomplish the task. Some learning tasks are low in element interactivity because they can be accomplished in a serial rather than coordinated fashion. For example, when studying a foreign language, learning some types of vocabulary is relatively low in element interactivity because each word can be memorized independently of other words. However, when you start to construct sentences, element interactivity jumps dramatically. When composing sentences you need to consider not only the meaning of several words but also the grammar and syntax rules that must be applied to sequence and parse the words correctly. All of these elements must be coordinated simultaneously to produce a correct sentence. If your task is to respond verbally to a question posed in a new foreign language, the mental load is even greater. Ask any new foreign language student about the amount of mental load he or she experiences during early conversational practice! To respond verbally, the student must first interpret the question, then compose an answer by selecting the correct words and applying grammar rules, and finally pronounce the words correctly—all within a relatively short amount of time. Intrinsic cognitive load is determined primarily by the knowledge and skills associated with your instructional objective. Although you cannot directly alter the inherent intrinsic load of your instructional content, you can manage the intrinsic load of any given lesson by decomposing complex

c01.qxd 9/29/05 17:53 Page 11

Cognitive Load and Efficiency in Learning

11

tasks into a series of prerequisite tasks and supporting knowledge distributed over a series of topics or lessons. This is what instructional professionals do as they create outlines of their courses and lessons. As a byproduct of segmenting and sequencing content into a series of instructional events, instructors manage intrinsic cognitive load. In Chapter 7 we summarize guidelines and evidence for best ways to manage intrinsic cognitive load through course and lesson design decisions.

Germane (Relevant) Load Germane cognitive load is mental work imposed by instructional activities that benefit the instructional goal. For example, learners in an Excel spreadsheet class will have different work requirements for using spreadsheets. Some students will need to construct spreadsheets as the basis for regular income and expense reports. Other students will use spreadsheets to calculate compensation that factors in taxes, commissions, bonuses, and deductions. To accomplish such diverse goals, during training, the learners will need to build a robust set of skills that they can apply to various types of spreadsheets with different data sets when they return to their work assignments. To build this flexible skill set, instructional examples should incorporate different calculation goals and data values. For example, in Figure 1.1 the learner practices a compensation calculation. Other examples in the same lesson involve profit, inventory, and sales scenarios. Of course, learning would be easier if all of the examples used a single type of spreadsheet with similar data. However, the skills that emerge from a more homogenous set of examples have been proven to be much more limited than skills built from a diverse set of examples. By studying diverse context examples and assignments, learners end up with a much broader repertoire of spreadsheet skills applicable to many work situations. The extra mental load imposed by this diversity is an example of germane cognitive load. Diversity in examples adds cognitive load in the service of the instructional goal. Think of germane load as relevant load imposed by instructional methods that lead to a better learning outcome. Chapter 9 is devoted entirely to instructional guidelines that add germane load.

c01.qxd 9/29/05 17:53 Page 12

12

Efficiency in Learning

Extraneous (Irrelevant) Load Extraneous cognitive load is the main form of load discussed in this book because it is always under your control as the instructor or course developer. Extraneous load imposes mental work that is irrelevant to the learning goal and consequently wastes limited mental resources. Those wasted resources drain mental capacity that could be used for germane load. As an example, take a look at Figure 1.2. It’s a screen taken from our overloaded Excel CD demonstration lesson on how to construct formulas. A number of features in this lesson waste limited mental capacity. For example, note that the words in the example are narrated and are also visible in text in the box located in the lower right corner of the screen. This design taxes mental resources in two unproductive ways. First, the learner must expend mental effort integrating the text in the lower right-hand corner

Figure 1.2. A Screen from a Lesson on Excel with Many Sources of Extraneous Cognitive Load. From the Overloaded Web-Based Lesson on the CD.

Audio: Barb has entered her sales revenue and her overhead for last year into the spreadsheet. The first thing that Barb would like to know is how much profit she made for each month last year.

c01.qxd 9/29/05 17:53 Page 13

Cognitive Load and Efficiency in Learning

13

with the visual portion of the spreadsheet referenced by the text. Second, the learner must expend mental effort to coordinate the words presented in two modes: visually in the text and aurally in the narration. The information in the lower left hand “Did you Know” box is another source of extraneous cognitive load, since it distracts the learner from the lesson objective. There are many other cognitive load violations in this lesson that we will discuss throughout the book. The poor design of this instructional product imposes extraneous cognitive load that drains cognitive resources needed to achieve the learning objective. The result of inefficient training programs with many extraneous sources of cognitive load is longer times to learn, poorer learning outcomes, or both. Think of extraneous cognitive load as irrelevant load.

Balancing Mental Load in Your Training Intrinsic, germane, and extraneous forms of cognitive load are additive. If your training program includes content that is complex, it is high in intrinsic load. If your program includes design elements that add extraneous load as well, there may be very little capacity left for germane load. Your training program will be inefficient. Consequently the learners will take longer to acquire the intended skills and/or they will not achieve the learning objective to the desired standard. To create efficient instruction, you must maximize germane load and minimize extraneous sources of load. While you usually cannot control the intrinsic load associated with the learning goals, you can manage it by segmenting and sequencing content in ways that optimize the amount of element interactivity required at any one time. The chapters in Part II focus on ways to reduce extraneous cognitive load by: (1) optimizing the use of visual and auditory presentation modes; (2) supporting learner attention; and (3) reducing the amount of information that must be processed in memory. By minimizing extraneous load, you free limited cognitive capacity for relevant or germane load imposed by instructional techniques that serve the learning objectives. In Part III we focus on techniques that add germane load to your training.

c01.qxd 9/29/05 17:53 Page 14

14

Efficiency in Learning

No Yellow Brick Road: The Relativity of Cognitive Load Dorothy was lucky because there was a single well-defined path that led to the Emerald City. However, we will see that the path to efficiency in training is not always so straightforward. Cognitive load depends on the interaction of three components: the learning goal and its associated content, the learner’s prior knowledge, and the instructional environment. As we discussed in the previous section, intrinsic cognitive load can be high or low, depending on the amount of element interactivity required to accomplish a task. Learning outcomes that require coordination among multiple content elements will result in greater cognitive load than less complex tasks. Research shows that many of the cognitive load techniques that reduce extraneous load improve efficiency in the learning of complex tasks only. Low complexity content will not demand a great deal of mental resources. Therefore, learning of low complexity tasks is not impeded by extraneous cognitive load. In contrast, when tasks are complex, using techniques that minimize extraneous load improves learning efficiency. Therefore, a general guideline for achieving efficiency in learning is to minimize extraneous cognitive load in your instructional materials when learning tasks are complex. But what is complexity? Complexity is of course relative to the performer. Indeed, we really can only define complexity in conjunction with expertise. Landing an airplane does not impose much load on an experienced pilot. However, it’s an overwhelming task the first few times a novice tries it. For an experienced pilot, nothing associated with routine flying is complex. For a learner, almost everything is complex. Answering a simple question in Italian requires minimal effort by a Milanese but imposes heavy demands on mental resources from the recent learner of Italian visiting Milan for the first time. Experts have a large skill repertoire in memory based on years of practice that allows them to effortlessly perform tasks that are overwhelming to a novice. As a result, we need to expand our general guideline for achieving efficiency in learning as follows: Avoid extraneous cognitive load when lessons involve complex content and the learners are novices. As we will see in Chapter 10, the techniques used to minimize extraneous load are not

c01.qxd 9/29/05 17:53 Page 15

Cognitive Load and Efficiency in Learning

15

needed by learners with greater prior knowledge. In fact, many of them actually impede their learning! You will need to change your instructional strategies as your learners develop expertise during training. In Part IV we show you how.

Cognitive Load Theory and Human Learning The guidelines of cognitive load theory result in more efficient learning because they exploit the limits and strengths of human learning processes. Our psychological architecture includes two main memory systems. One, called working memory, has a very limited capacity but is the active processing center of our brain. The rule that we cannot remember more than 7 2 items applies to the limited capacity of working memory. Although its capacity is limited, working memory is the site of our thinking and learning processes. Another memory system, called long-term memory, has a huge capacity but is primarily a storage repository. Long-term memory cannot engage in thinking or learning processes, although, as we will see in Chapter 2, it can have a large effect on thinking and problem solving. These two memory systems work together. As learning takes place in working memory, the new knowledge and skills are stored in long-term memory. As we gain expertise in a domain, our knowledge repository in long-term memory expands. That knowledge repository in turn allows working memory to function more efficiently in ways we will discuss in Chapter 2. As a result of the knowledge stored in long-term memory, working memory can deal with much more information, and the risks of cognitive load during learning are much lower. That’s why learners with greater prior knowledge are not subject to the negative effects of instructional methods that impose extraneous load on novice learners.

Evidence-Based Practice The training profession has been shaped by fad and folk wisdom more than by scientific evidence of what actually works (Clark, in press). Whether it be discovery learning, edutainment, or learning styles, our training programs

c01.qxd 9/29/05 17:53 Page 16

16

Efficiency in Learning

are often the victims of various fads that at best waste time and resources and at worst are counterproductive to learning. Fortunately, we see some strong signals that policy makers are looking for valid research to guide instructional decisions. For the first time in history, in 1998 the U.S. Department of Education made school funding contingent on the use of funds for programs based on “proven, comprehensive reform models.” The No Child Left Behind Act mentions scientifically based research over one hundred times. Paragraph A Section 9109 defines scientific research as: “the application of rigorous, systematic, and objective procedures to obtain reliable and valid knowledge relevant to education activities and programs which includes research that is evaluated using experimental or quasi-experimental designs preferably with random assignment.” Organizations that collectively invest billions in training programs are also seeking instructional methods that are proven to work. Evidence-based practice means grounding decisions about the development and deployment of learning programs on the basis of valid evidence—not fads, fables, or folk wisdom.

Evidence for Cognitive Load Theory Cognitive load theory is based on dozens of experiments conducted over the past twenty-five years by instructional scientists in Australia, Europe, and the United States. All of the research we summarize uses random assignment of participants to an experimental lesson and a comparison lesson. After a study period, the participants rate the amount of effort they invested while studying the lesson and take a test to measure learning outcomes. These two measures—invested mental effort and learning—are combined in an efficiency metric that we describe later in the chapter. As cognitive load theory evolved, experiments were designed to measure the effects of cognitive load management methods under different conditions. Researchers compared lessons with and without cognitive load management techniques that included both high and low complexity content. For example, a study reported by Leahy, Chandler, and Sweller (2003) compared audio and text explanations of the temperature line graph shown in Figure 1.3. The test included some easy questions such as “How can you

c01.qxd 9/29/05 17:53 Page 17

17

Cognitive Load and Efficiency in Learning

Figure 1.3. A Graph of Temperature Changes over Time. From Leahy, Chandler, and Sweller, 2003.

Monday Tuesday

Temperature (C) 36 34 32 30 28 26 24 22 20 18 16 9am

10am

11am

12pm

1pm

2pm

3pm

4pm

Time of Day

recognize zero average rate of change just by looking at the graph?” as well as some complex questions such as “What is the average rate of change between 11:00 A.M. and 1 P.M. on Tuesday?” As you can see in Figure 1.4, the lesson that explained the graph with audio narration resulted in better learning of complex questions only. For easier tasks, there was no difference between the audio and text versions. We conclude from this study that a textual rather than an audio explanation of a graphic can impose an extraneous cognitive load that leads to depressed learning of complex tasks. We discuss this study in greater detail in Chapter 4.

About the Numbers As you read, you will find summaries of research experiments like the one mentioned in the preceding paragraph that support our guidelines. For example, in Figure 1.4 you can see that, for complex tasks, the audio narrated version resulted in learning outcomes that were significantly different from the outcomes from the text version. Statistical significance means that the outcome differences are unlikely to have occurred by chance alone. But statistical significance does not necessarily mean that the results have practical implications. A statistically significant result may in fact represent only a very small

c01.qxd 9/29/05 17:53 Page 18

18

Efficiency in Learning

Figure 1.4. Audio Explanations Result in Better Achievement Than Textual Explanations on Complex Questions. Based on data from Leahy, Chandler, and Sweller (2003).

Audio Version Mean Percentage Correct

100 Text Version SD=Significant Difference

80 60 40

SD 20

Easy Problems

Complex Problems

outcome difference that is not especially relevant from a practical perspective. Practical significance, also known as clinical significance, can be better judged by a statistic called effect size. Effect size is a relatively recently reported statistic and you will see it often reported in research published after 2000. When available, we have included effect size data in our research summaries. As a general guideline, effect sizes less than or equal to .30 are considered small and are of negligible practical importance. Effect sizes around .50 are considered medium and are of moderate practical importance. Finally, effect sizes of .80 or higher are large and are of crucial practical importance. See the Appendix in the back of the book for more details on how effect sizes are calculated.

Limits of Research Any one experiment—even one with a high effect size—is likely to have limited applicability to your instructional environment because the context of the experimental conditions are different from your situation. Some

c01.qxd 9/29/05 17:53 Page 19

Cognitive Load and Efficiency in Learning

19

factors for you to consider with respect to any experiment include the age and prior knowledge of the learners, the content and length of the lessons, the type of assessment used to measure learning, as well as whether learning was measured immediately and/or sometime after the instructional event. The good news about cognitive load theory is that so many experiments have been done that many of the guidelines we offer have been demonstrated in diverse environments. For example, Table 1.1 summarizes over sixteen studies that replicate the result shown in Figure 1.4—that audio explanations of visuals lead to better learning than text explanations of visuals. This guideline is called the modality principle. As you can see in Table 1.1, the modality principle has been demonstrated in controlled experiments using fourth graders and college students in lessons on geometry, electrical testing, and botany that lasted from just a few seconds to approximately half an hour. Most of the experiments measured different forms of learning, including recall of lesson content as well as application of that content to perform a task or solve a problem. As a result of the many diverse conditions in which the modality effect has been demonstrated, you can feel confident to use audio to explain visuals in many instructional situations summarized in Chapter 4. In contrast to the modality effect, some of the guidelines we offer are more recent and therefore do not yet have a large number of experiments to support them. You will need to attend to the details of those experiments to infer to what extent the results are likely to apply to your setting and/or wait until more evidence accumulates.

Quantifying Efﬁciency Fundamentally, cognitive load theory is about efficiency. Cognitive load theory defines efficiency in terms of two variables: learner performance and learner mental effort. Instructional environments that result in higher learning outcomes with less mental effort are more efficient than environments that lead to lower outcomes with greater mental effort. Instructional scientists use an efficiency metric to quantify the efficiency of an instructional product.

c01.qxd 9/29/05 17:53 Page 20

20

Efficiency in Learning

Table 1.1. A Summary of Experiments Demonstrating a Modality Effect. Date/Research Team

Learner Population

Lesson Topic

Lesson Length

Outcome Measures

1995—Mousavi et al., Expt 1

8th graders/ Australian

Geometry examples

Learner determined; up to 5 minutes per example on text and time to listen to audio twice

Scores on similar and different geometry problems

1995—Mousavi et al., Expt 2

8th graders/ Australian

Geometry examples

151 and 157 seconds for each example

Scores on similar and different geometry problems

1995—Mousavi et al., Expt 3

8th graders/ Australian

Geometry examples

Varied by treatment

Learning time, testing time, scores on similar and different geometry problems

1995—Mousavi et al., Expt 5

8th graders/ Australian

Geometry examples

Varied by treatment

Test solution times

1995—Mousavi et al., Expt 6

4th graders/ Australian

Geometry examples

55 seconds up to 3 minutes per example

Learning time and test solution times

1997—TindallFord et al., Expt 1

Trade apprentices/ Australian

How to conduct electrical tests

5 minutes

Recognition and application

1997—Tindall Ford et al., Expt 2

Trade apprentices/ Australian

How to interpret an electrical table

100 seconds and 170 seconds

Recognition, application, and efﬁciency

1997—Tindall Ford et al., Expt 3

Trade apprentices/ Australian

Electrical symbol identiﬁcation and how to interpret electrical circuit diagram

Approximately 3 minutes

Recall and application, test solution times, efﬁciency

c01.qxd 9/29/05 17:53 Page 21

21

Cognitive Load and Efficiency in Learning

Table 1.1. (Continued) Date/Research Team

Learner Population

Lesson Topic

Lesson Length

Outcome Measures

1998—Mayer & Moreno, Expt 1

College students/U.S.

How lightning forms

140 seconds

Retention, recognition, application

1998—Mayer & Moreno, Expt 2

College students/U.S.

How car brakes work

45 seconds

Recall, recognition, application

2001—Moreno et al., Expt 1

College students/U.S.

Botany concepts/ game format with agent

25 minutes

Recall and application

2001—Moreno et al., Expt 2

7th graders/ U.S.

Botany concepts/ game format with agent

Self-paced, up to 40 minutes

Recall and application

2001—Moreno et al., Expt 4

College students/U.S.

Botany concepts/ game format with agent

Self-paced, 24–28 minutes

Recall and application

2001—Moreno et al., Expt 5

College students/U.S.

Botany concepts/ game format with agent

Self-paced, 24–28 minutes

Recall and application

2002—Craig et al., Expt 2

College students/U.S.

How lightning works explained by agent

180 seconds

Recall, recognition, application

2003—Leahy et al., Expt 1

5th graders/ Australian

Interpretation of a line graph

No time limit and 185 seconds

Recognition, application

2003—Mayer et al., Expt 1

College students/U.S.

How an electric motor works

Approximately 20 minutes

Application

c01.qxd 9/29/05 17:53 Page 22

22

Efficiency in Learning

Conceptually, the efficiency metric is calculated by subtracting mental load (ML) from performance (P) outcomes. We express this mathematically as E P – ML. When performance is greater than mental load, the efficiency value is positive. When performance is lower than mental load, the efficiency value is negative. Performance is most often measured by a test taken at the end of the lesson. Sometimes however, performance is measured by the time required to complete a lesson or a test. Mental load is most commonly measured by learner estimates of lesson difficulty. The difficulty (mental load) of a lesson is assessed using a 1 to 7 or 1 to 9 scale in which 1 indicates extremely low mental effort (very, very easy) and 7 or 9 indicates extremely high mental effort (very, very difficult). Although learner estimates of mental load are subjective, studies that have compared these ratings with other physiological or psychological measures of mental load show that they are effective and are the most pragmatic way to assess mental effort. For a detailed technical discussion of measurement of mental effort, see the paper by Paas, Tuovinen, Tabbers, and Van Gerven (2003).

The Efﬁciency Graph To visually represent the efficiency metric, instructional scientists use an efficiency graph like the one shown in Figure 1.5. Mental effort is plotted on the horizontal axis with higher values to the right of the vertical line and lower values to the left. Performance is plotted on the vertical axis with higher values above the horizontal line and lower values below it. As you can see in Figure 1.5, the efficiency value represented by point A is high on the performance line and low on the mental effort line. High performance with low mental effort means high efficiency. The upper left quadrant of the graph is considered the high efficiency area of the graph. In contrast, point B represents an efficiency value that is low on the performance scale and high on the mental effort scale. The lower right quadrant of the graph is called the low efficiency area of the graph. For more details on the mathematics behind the efficiency value and graph, see the Appendix.

c01.qxd 9/29/05 17:53 Page 23

23

Cognitive Load and Efficiency in Learning

Figure 1.5. Hypothetical Efﬁciency Plots on the Efﬁciency Graph. Performance High Efficiency

E=0

1.0 0.8

A

0.6 0.4 Mental Effort

0.2 -1.0 -0.8 -0.6 -0.4 -0.2

0.2 0.4 -0.2

0.6

0.8

1.0

-0.4 -0.6

B

-0.8 -1.0 Low Efficiency

The Bottom Line In this chapter we set the stage for the book as follows: • Cognitive load theory is an evidence-based set of universal principles and guidelines that result in more efficient learning environments. • Efficient learning environments lead to better learning, faster learning, or both. • Efficient learning environments balance intrinsic, germane, and extraneous sources of load. • Cognitive load depends on the interaction among the expertise of the learner, the complexity of the content, and the instructional methods used in the training environment.

c01.qxd 9/29/05 17:53 Page 24

24

Efficiency in Learning

• Efficient learning environments exploit the strengths and compensate for the limits of human learning processes. • Efficiency of an instructional product can be quantified by an efficiency metric and displayed on the efficiency graph.

On the CD John Sweller Video Interview Chapter 1: Cognitive Load Theory and Efficiency in Learning. John defines cognitive load theory and describes intrinsic, extraneous, and germane forms of cognitive load. He also discusses the evidence for cognitive load theory.

Sample Excel e-Lessons We have several sample lessons on the CD to illustrate applications and violations of cognitive load theory. You may want to preview them now as an introductory supplement to this book and review them as you read the various chapters in order to focus on specific techniques discussed in the chapter. The samples include: 1. An asynchronous web-based lesson that violates many cognitive load principles: Before Overloaded Excel Web-Based Lesson. 2. An asynchronous web-based lesson that applies many cognitive load principles: After Load Managed Excel Web-Based Lesson. 3. A virtual classroom (synchronous) web-based lesson that applies many cognitive load principles: Virtual Classroom Example. In addition to each sample, there is also a commentary on the sample by John Sweller.

COMING NEXT Cognitive load theory works in harmony with human memory processes involved in learning. In the next chapter, we review the features of and interactions between working memory and long-term memory. We also describe

c01.qxd 9/29/05 17:53 Page 25

Cognitive Load and Efficiency in Learning

25

the psychological processes involved in translating lesson content into new knowledge and skills in long-term memory.

Recommended Reading Sweller, J. (2005). Implications of cognitive load theory for multimedia learning. In R.E. Mayer (Ed.), Cambridge handbook of multimedia learning. Cambridge: UK: Cambridge University Press.

c02.qxd 9/29/05 17:54 Page 26

CHAPTER OUTLINE Harnessing Human Learning Processes Our Memory Partners Working Memory and the Signiﬁcance of the Number 7 2 Long-Term Memory and Expertise Efﬁcient Instruction Provides Schema Substitutes for Novices

Visual and Auditory Centers in Working Memory Dual Task Experiments

Cognitive Load and Learning How Learning Happens The Processes of Learning Instructional Methods to Promote Instructional Events

Automaticity: A Working Memory Bypass The Bottom Line On the CD John Sweller Video Interview Sample Excel eLesson

c02.qxd 9/29/05 17:54 Page 27

2 Lesson

The Psychology of Efﬁciency

LTM

WM

I

NSTRUCTION IS EFFICIENT

only when it works in harmony with human learning processes. Efﬁcient learning environments are compatible with the features of working memory and long-term memory. Knowledge in the long-term memory (LTM) of experts allows them to process information efﬁciently. However, for novices who lack this knowledge, the instructional environment must provide knowledge substitutes as supports to enable efﬁcient learning. Information in the instructional environment is learned when it is encoded into long-term memory in ways that allow it to be transferred later in the workplace. In this chapter we describe the major psychological events that mediate this transformation.

Harnessing Human Learning Processes In Chapter 1 we defined efficient instruction as training that leads to better learning outcomes with less mental effort. Efficient learning leads to better achievement of instructional goals and/or faster learning. When you apply our cognitive load guidelines to your instructional environment, you help your learners use their mental resources to their best advantage. Unlike many 27

c02.qxd 9/29/05 17:54 Page 28

28

Efficiency in Learning

popular training programs, cognitive load theory is based on human learning processes. As we discussed in Chapter 1, there are no straightforward applications of our guidelines because they depend on an interaction among the learner’s prior knowledge, the complexity of the content and tasks, and the instructional environment. Therefore, you will need to tailor these guidelines to your own unique situation. Understanding not only the what’s of cognitive load theory but also the why’s will help you design and deliver efficient training. This chapter focuses specifically on those human learning processes that make up the why’s of cognitive load theory. In this chapter we discuss working and long-term memory systems and the main psychological events involved in learning that must be accommodated in efficient instructional environments. In particular we will see how our guidelines work to compensate for lack of knowledge in the long-term memory of novices.

Our Memory Partners All human learning and work activities rely on two of our memory systems— working memory and long-term memory and the partnership they share. As its name implies, working memory is the active partner. As you read this book and think about how it applies to your work, it is your working memory that does the processing. While in a learning mode, new information from the environment is processed in working memory to form knowledge structures called schemas, which are stored in long-term memory. Schemas are memory structures that permit us to treat a large number of information elements as though they are a single element. For example, when you read the word “element” you no longer think of seven separate letters; you think of a single entity, the word “element.” We have schemas for all aspects of our cognitive lives. While all trees differ from all other trees and include a huge amount of sensory information (color, shape, number of leaves, twigs, branches, and so forth), we can instantly recognize an object as a tree because we have a schema for trees. We similarly have schemas for letters, words, and combinations of words that allow us to read easily and rapidly. Schemas for the solution to specific mathematics problems may

c02.qxd 9/29/05 17:54 Page 29

T h e Ps y c h o l o g y o f E f f i c i e n c y

29

make us competent at mathematics. In any area in which you have competence, that competence derives from schemas stored in your long-term memory. Long-term memory is the main knowledge repository because working memory has very little storage capacity. Working memory’s main role is conscious processing. When you ask working memory to store anything beyond very limited amounts of information, its processing capacity is inadequate. We’ve all had the experience of being overwhelmed when trying to learn something new. That feeling of failing to understand leading to frustration or anxiety is your working memory telling you that it is overloaded. Overload stresses psychological processes in ways that diminish learning efficiency. When overload gets large enough, the learning system shuts down altogether. Your job as an instructional professional is to design and deliver training that will leverage germane cognitive load to promote learning and at the same time to minimize extraneous load and manage intrinsic cognitive load to free working memory capacity for learning.

Working Memory and the Signiﬁcance of the Number 7 2 The limits of working memory capacity were first made explicit in George Miller’s 1956 paper: The Magical Number Seven Plus or Minus Two. The phrase 7 2, refers to limited working memory capacity. Not only is the storage capacity of working memory very limited, but so also is the duration of anything that enters working memory. Unless you actively keep new data alive in working memory through repetition or process it for storage into long-term memory, even small amounts of new information will disappear from consciousness in just a few seconds. As a result of these working memory limits, if you look at a mid-play chess board very briefly, say for five seconds, you would probably only be able to recall the location of a few chess pieces. Chase and Simon (1973) counted how many times novice chess players needed to refer back to a mid-play board like this one in order to accurately reconstruct it from memory. As you can see from the solid line in Figure 2.1, novice chess players needed to refer back to the board more than six times before they were able to reconstruct most of it. However, when faced with the same task,

c02.qxd 9/29/05 17:54 Page 30

30

Efficiency in Learning

Figure 2.1. Number of Referrals Needed to Reproduce a Mid-Play Chess Board.

Number of Pieces Accurately Recalled

Source: Chase and Simon (1973).

Master Chess Players

30 25 20 15 10

Novice Chess Players

5 0 T1

T2

T3

T4

T5

T6

Number of Referrals to Chess Board

expert chess players had a different experience. As you can see from the broken line in Figure 2.1, master level chess players were able to recall the chess pieces with only a few referrals. From this experiment, we see that we need to revise our conception of 7 2. While the 7 2 chunks capacity limit is the same for everyone, chunk size varies. People with experience in a domain can form chunks of much greater size and in this way expand the virtual capacity of their working memory. What are chunks? These days, most people equate them with schemas. A large chunk is simply an extensive, sophisticated schema and for that reason we will use the term schema from this point on. Where are schemas stored? In long-term memory that we discuss next.

Long-Term Memory and Expertise In contrast to working memory, long-term memory has a massive capacity for information storage. However, it is the inert member of the memory partnership. All conscious processing takes place in working memory. But working memory and long-term memory work closely

c02.qxd 9/29/05 17:54 Page 31

31

T h e Ps y c h o l o g y o f E f f i c i e n c y

together. The more knowledge and skills stored in long-term memory, the greater the virtual capacity of working memory as a result of larger, more complex schemas. For example, a chess board includes about twenty-four elements of information for a novice player. However, experts represent the chess board in play patterns involving clusters of several pieces. Each cluster translates into a schema. Therefore the chess board of twenty-four pieces contains approximately eight or nine schemas for a chess expert. The recall of chess pieces shown in Figure 2.1 was based on a mid-play chess board. How would a random placement of chess pieces on the board affect the recall results? We might expect that experts would lose their advantage and need about the same number of referrals as novices. Chase and Simon (1973) tested this hypothesis. As you can see in Figure 2.2, expert player performance was actually worse than novice performance when trying to reconstruct a random chess board! The experts were trying to apply their schemas for chess play patterns to a meaningless environment. The extra psychological work they expended trying to make sense of what they were viewing depressed their memory performance so they were worse off than the novice players!

Figure 2.2. Number of Referrals Needed to Reproduce a Random Chess Board. Number of Pieces Accurately Recalled

Source: Chase and Simon (1973).

25 20

Novice Chess Players

15 10 5 0

Master Chess Players T1

T2 T3 T4 Number of Referrals to Chess Board

T5

T6

c02.qxd 9/29/05 17:54 Page 32

32

Efficiency in Learning

Efﬁcient Instruction Provides Schema Substitutes for Novices As a result of their enhanced schemas, experts have significantly different psychological capabilities than novices. Experts are able to tackle complex tasks that overwhelm less experienced workers. When learning new skills in their domain, experts are enabled by their rich storehouse of schemas to process much larger amounts of information as well as to guide much of their own learning processes. Novices, in contrast, lack such schemas and therefore need learning environments that compensate for them. Well-designed learning environments for novices provide schema substitutes by optimizing the limited capacity of working memory in ways that free working memory for learning. For example, packed lessons loaded with new knowledge and skills will overwhelm novice learners. Good instruction will segment and sequence the content in ways that reduce the amount of new information novices must process at one time. In contrast, experts can handle many elements of information because the elements are incorporated in a schema that is treated as a single element in working memory. Therefore, experts do not need the same instructional support as novices. For example, learners with some domain experience learn a new skill best by practicing it right away. In contrast, novices are overloaded by the mental demands of extensive practice and learn more efficiently when they can first study worked examples to help them build new schemas. In short, effective training environments for novices are designed to substitute for the schemas that learners with relevant experience already have. As we will discuss in Chapter 10, many of the instructional techniques that serve as schema substitutes for novices don’t work for experienced learners, who possess relevant schemas. In fact, like the expert chess players facing a random chess board, instructional environments designed for novices actually depress learning outcomes of experienced learners—an effect known in cognitive load theory as expertise reversal. In Chapters 10 and 11 we will describe in greater detail how to transition your training as learners gain expertise. In summary we see that working memory has a limited capacity of 7 2 information elements. However, not all information elements are equal. The size of any given element depends on the prior knowledge of the learner based on preexisting schemas stored in long-term memory. This is why

c02.qxd 9/29/05 17:54 Page 33

33

T h e Ps y c h o l o g y o f E f f i c i e n c y

experienced workers can easily handle complex tasks that are overwhelming for new workers. Similarly, when experienced workers serve as instructors, they often overload their learners by failing to compensate for the much more limited schemas of the learners. To compensate for lack of schemas in LTM of novices, efficient instructional environments based on cognitive load theory offer the psychological support that such schemas provide more experienced learners.

Visual and Auditory Centers in Working Memory In the last twenty-five years, we have learned more about working memory. Most relevant to instructional efficiency is the finding that working memory incorporates separate components. One component is for visual information and another one is for auditory (phonetic) information as represented by the two folders in Figure 2.3. While each of these working memory components is quite limited, we will see in Chapter 3 that under some defined circumstances you can maximize working memory capacity by using both components to explain visuals with words delivered in audio narration rather than by text.

Figure 2.3. An Overview of Cognitive Learning Processes. Source: Clark Training & Consulting.

MEMORY Long-term

Working

Encoding Retrieval

Schemas

Phonetic Visual Rehearsal

ATTENTION

c02.qxd 9/29/05 17:54 Page 34

34

Efficiency in Learning

Dual Task Experiments The visual and auditory subcomponents of working memory were inferred by performance outcomes from dual task experiments. Dual task experiments require multitasking. In a dual task experiment, participants are assigned a primary task such as to recall a list of numbers that is read to them at the same time as they perform a secondary task. A secondary task might require the participant to strike a key when they hear a tone or track a moving dot on a computer screen using a joy stick. Psychologists conducting dual task experiments found that, when participants are asked to perform a secondary auditory task (respond to a tone) while performing a primary auditory task (recall a list of numbers read aloud), performance on one or both tasks suffers. However, performing a secondary task that is visual (tracing a moving dot) while performing a primary auditory task leads to minimal performance declines. The reason for these results is that two tasks that involve the same modality, for example, both visual or both auditory, drain the limited working memory capacity in that center. In contrast, two tasks that call on different modalities, for example, one visual and one auditory, make use of the separate subcomponents in working memory.

Cognitive Load and Learning The goal of your instructional program is to free limited working memory from irrelevant mental effort and harness it for the work required to integrate new knowledge and skills into the schemas in long-term memory. Productive mental effort needed to build effective new schemas is what we call germane cognitive load. Studying examples is one type of germane cognitive load that leads to schema development in long-term memory. Unproductive mental effort that does not lead to learning is what we call extraneous cognitive load. For example, requiring learners to study an example in which a visual is explained by text that is separated from the graphic, such as the one shown in Figure 2.4, imposes extraneous cognitive load. That is because working memory must invest extra effort to search for relations between the text and graphic. Instead, the improved visual in Figure 2.5 integrates the text explanations close to the visual elements being explained, reducing the need

c02.qxd 9/29/05 17:54 Page 35

T h e Ps y c h o l o g y o f E f f i c i e n c y

Figure 2.4. An Instructional Display That Imposes Extraneous Load.

Step Action 1. Select the cell where you would like the quantity to be displayed. In this example, click cell B6. 2. Locate cell references for hourly wage and hours. In this example, the data is stored in cells B4 and B5, respectively. 3. Enter a formula to multiply cell references. The symbol for multiply is the asterisk (*). In this example, enter =B4*B5. Press Enter.

Figure 2.5. An Instructional Display That Minimizes Extraneous Load.

35

c02.qxd 9/29/05 17:54 Page 36

36

Efficiency in Learning

to search for relevant relations. In this way, the instructional display saves the mental work that learners have to invest when studying separated displays. That extra capacity can be devoted to processing the example and building a schema from it. Programs with considerable extraneous cognitive load lead to inefficient learning. We also preserve the limited resources of working memory by reducing intrinsic cognitive load that results from the complexity of the new knowledge and skills to be learned. As we will discuss in Chapter 7, you reduce intrinsic load by segmenting and sequencing content in ways that control the amount of new information the learner must process at one time.

How Learning Happens As illustrated in Figure 2.3, visual and auditory information that is attended from the instructional environment enters working memory. Unattended information does not enter the processing system. Once in working memory, mental effort is used to integrate the visual and auditory information in order to form a coherent message. Next, working memory must integrate the new information into preexisting related schemas that are brought into working memory from long-term memory. If successful, the result is an expanded schema in long-term memory that has integrated the new knowledge and skills from the instructional environment.

The Processes of Learning The following psychological processes summarized in Figure 2.3 work in concert to achieve learning goals: attention, activation of prior knowledge, elaboration-rehearsal, encoding and retrieval. First, learners must focus their attention on the instructional information that is relevant to the learning goal. Attention is critical in order to screen out irrelevant information that will clog working memory and to put relevant data into working memory. New information entering working memory must be integrated into preexisting schemas in long-term memory. Therefore those schemas must be transferred from long-term memory into working memory. This transfer is called activation of relevant preexisting knowledge. At

c02.qxd 9/29/05 17:54 Page 37

T h e Ps y c h o l o g y o f E f f i c i e n c y

37

this point, working memory must process new knowledge and skills in order to integrate them into the activated schemas from long-term memory. This processing is called elaboration of information. Elaboration of information results from rehearsal of new content in working memory. As a result of rehearsal and elaboration, incoming content from the instructional environment is transformed to result in expanded schemas stored in long-term memory. The result is encoding of new knowledge and skills in long-term memory. However, we can’t claim successful learning yet. It’s not sufficient to form new schemas in long-term memory. Since all conscious activity occurs in working memory, those new schemas must be brought back into working memory when needed on the job. This process is called retrieval. Retrieval of new knowledge and skills is the psychological basis for transfer of learning. Successful transfer means that new knowledge and skills acquired during training are available for use later in the work environment. If you’ve ever struggled to apply something that was covered in a training class, you know that transfer of learning is not always easy. Learning transfer will depend on how new schemas are formed. Schemas that incorporate relevant cues that reflect the context of the job are more likely to transfer than schemas lacking such cues. For example, when learning a new computer application, if the training program uses screens similar to those you will see on the job, your new schema will incorporate those visual cues. When you see those screens in the workplace, you are more likely to retrieve the new procedure since it was learned in conjunction with those screens. All of these processes rely on mental effort invested by working memory. If training programs require working memory to expend additional effort not directly related to these processes, learning slows and becomes inefficient. To review these learning processes in greater detail, refer to our recommended readings at the end of this chapter.

Instructional Methods to Promote Instructional Events Your training programs should use instructional methods that promote the key learning events summarized in the previous paragraph and at the same time avoid disrupting them. For example, in the beginning of our demonstration

c02.qxd 9/29/05 17:54 Page 38

38

Efficiency in Learning

virtual classroom lesson on Excel from the CD, the instructor presents learning objectives to help participants focus their attention on content relevant to the expected outcomes of the training. During the lesson introduction, a quick review of a prior lesson on cell references helps activate relevant prior knowledge. This makes the related schemas available in working memory for addition of the new content. The instructor promotes elaboration and rehearsal by asking questions that require learners to study and draw conclusions from the examples, as well as to complete a demonstration that the instructor begins. By using job-realistic situations and work environments during the lesson, the instructor is helping learners build schemas that are likely to transfer later. For example in the Excel lesson, as shown in Figure 2.6, the instructor uses an actual spreadsheet with different job-realistic contexts for demonstrations and practice.

Figure 2.6. A Virtual Classroom Excel Lesson Incorporates Demonstrations of Excel Applications. From the Virtual Classroom Lesson on the CD.

c02.qxd 9/29/05 17:54 Page 39

39

T h e Ps y c h o l o g y o f E f f i c i e n c y

Automaticity: A Working Memory Bypass So far we have reviewed two vehicles for maximizing the limited capacities of working memory. First, as individuals gain expertise, their schemas in long-term memory grow, and consequently they are able to process larger content segments. In this way, the virtual capacity of working memory expands. Prior to this point, the instructional environment must substitute for limited schemas of novices by segmenting, sequencing, and presenting content in ways that will avoid overload. Second, we saw that judicious use of the auditory and visual subcomponents of working memory, allow us to maximize its limited capacity. Automaticity is a third psychological mechanism that can allow you to largely bypass working memory limits. Any task that is performed hundreds of times becomes established in long-term memory. Once automated, the skill can be performed with little or no resources from working memory. In effect, these skills are performed unconsciously. For example, as I write this paragraph, I can type the words automatically. Since I do not need to invest any mental resource in the typing, I can use limited working memory capacity to compose meaningful sentences and paragraphs. All of us have the benefit of automated skills in long-term memory that free up working memory resources. As another example, you are able to read this paragraph very quickly because as an expert reader you have automated the decoding skills of the English language. Unless you are bilingual, however, if you were to read this book in Italian, your reading speed and comprehension would be greatly slowed. The path to automaticity is a long one requiring many hundreds of practice repetitions. Most organizational training programs do not allocate sufficient time to reach skill automaticity. New tasks that are performed frequently become automated after training as a result of job experience. Other new tasks that are not performed frequently will require conscious working memory capacity to perform. Good reference materials will help learners perform infrequent tasks more effectively. In Chapter 6 we summarize guidelines for the design and production of effective reference aids.

•••

c02.qxd 9/29/05 17:54 Page 40

40

Efficiency in Learning

The Bottom Line In this chapter we laid the psychological foundation for cognitive load theory as follows: • The guidelines that emerge from cognitive load theory are based on maximizing limited working memory resources, which can then be allocated to more efficient building of schemas in long-term memory. • Experts in a domain have a rich repository of schemas in long-term memory that allow them to use their working memory resources more efficiently. • Novices need support from the instructional environment to substitute for their lack of schemas. • Five essential instructional events mediate the transformation of new knowledge and skills from the learning environment into transferable schemas in long-term memory: attention to relevant instructional content; activation of related prior knowledge stored in long-term memory; elaboration of new content during rehearsal; encoding of new schemas in long-term memory; and retrieval of these schemas later when needed in the workplace. • All of the instructional events rely on working memory capacity and should be supported by instructional methods that manage cognitive load.

On the CD John Sweller Video Interview Chapter 2: The Psychology of Efficiency. In his discussion of human cognition, John includes the features of working memory, the consequences of limited working memory on human thinking processes, features of long-term memory, including schemas, the importance of long-term memory to human cognition, and the process of automaticity.

c02.qxd 9/29/05 17:54 Page 41

T h e Ps y c h o l o g y o f E f f i c i e n c y

41

Sample Excel eLesson The Virtual Classroom Example illustrates the application of the psychological events of learning as follows: 1. Attention—supported by cueing with the highlighter of relevant portions of the spreadsheet as well as setting of ground rules to limit the potential for distraction from use of diverse features of the virtual classroom interface. 2. Activation of prior knowledge—supported by a review of spreadsheet row and column designations. 3. Elaboration, Rehearsal, Encoding—supported by frequent questions requiring learners to process and apply new knowledge and skills. 4. Retrieval—supported by use of Excel interface and realistic job contexts as the basis for examples and practice.

COMING NEXT In Part II of the book we present guidelines and research regarding ways that instructional professionals can minimize extraneous cognitive load. In Chapter 3 we look at ways to minimize extraneous cognitive load by leveraging the auditory and the visual subcomponents of working memory. We will review guidelines related to use of visuals alone and to explanations of visuals that result in more efficient instructional environments.

Recommended Readings Clark, R.C. (2003). Building expertise (2nd ed.) Silver Spring, MD: International Society for Performance Improvement. Sweller, J., van Merriënboer, J.J.G., & Paas, F.G.W.C. (1998). Cognitive architecture and instructional design. Educational Psychology Review, 10(3), 251–296.

c02.qxd 9/29/05 17:54 Page 42

c03.qxd 9/29/05 17:56 Page 43

PA R T T W O

Basic Guidelines for Managing Extraneous (Irrelevant) Cognitive Load

N PART 2 WE INCLUDE six chapters that summarize all of the proven

I

ways you can reduce extraneous (irrelevant) and intrinsic cognitive load. When learners are novice and the skills are complex, you should keep extraneous cognitive load low and manage intrinsic cognitive load in order to free up working memory for learning. Many of the proven techniques that reduce extraneous cognitive load involve best ways to use the basic communication tools of visuals, text, and audio to present your content. Effective use of graphics, audio, and text can make the best use of limited working memory resources by: • Dividing content between the visual and auditory centers in working memory • Focusing attention to important content elements and avoiding split attention • Minimizing unnecessary or redundant content or presentation modes • Providing external memory supplements with performance aids In addition to these methods, you can reduce extraneous cognitive load by replacing some practice exercises with demonstrations, also called worked examples. Studying worked examples leads to learning with less extraneous 43

c03.qxd 9/29/05 17:56 Page 44

44

Efficiency in Learning

(unproductive) mental effort than that required by solving practice problems. As a result, learning is more efficient. Intrinsic cognitive load originates from the complexity of your instructional content. You can reduce intrinsic cognitive load by segmenting the content so that learners are exposed to supporting knowledge separate from task steps or process stages. In that way learners are introduced to small increments of new content gradually. Likewise, when learners can control the rate at which they access instruction, such as in self-paced e-learning or when reading a text, intrinsic load can be managed by the learner. In contrast, instructionally paced learning environments, such as classroom or synchronous e-learning sessions, will impose higher cognitive load and load reduction techniques must be used to compensate. Read

To Find Out How to

Chapter 3. Use Visuals and Audio Narration

Use diagrams to represent spatial information in training materials and working aids Use diagrams to build deeper understanding Explain diagrams efficiently with audio narration

Chapter 4. Focus Attention and Avoid Split Attention

Focus attention with signals and cues Avoid split attention by integrating content on pages or screens Avoid split attention in computer training by integrating all content in a single delivery medium

Chapter 5. Weed Your Training

Pare content down to essentials Eliminate extraneous visuals, text, and audio Eliminate redundancy in content delivery modes

Chapter 6. Provide External Memory Support

Supplement memory with performance aids Design efficient performance aids

Segment lesson content so that learners receive content gradually Chapter 7. Use Segmenting, Sequenc- Identify cognitive load risks of whole-task learning environments ing, and Learner Pacing Identify the cognitive load associated with instructionally paced learning environments Chapter 8. Transition Replace some practice problems with worked examples and from Worked Examples completion problems Use backwards fading to transition from worked examples to practice problems Display worked examples and completion problems in ways that avoid split attention

c03.qxd 9/29/05 17:56 Page 45

Basic Guidelines for Managing Extraneous

45

On the CD Video Interview with John Sweller: Chapter Preview/Review Chapter 3. Use Visuals and Audio Narration. The best techniques for display and explanation of spatial information using graphics and audio Chapter 4. Optimize Attention. Methods to use to focus attention/avoid split attention; a discussion of the cognitive load imposed by note-taking and by split attention formats in computer training Chapter 5. Weed Your Training. A definition of the redundancy effect followed by specific discussion of repetition versus redundancy, redundancy in computer training, and redundancy in instructor presentations Chapter 7. Use Segmenting, Sequencing, and Learner Pacing. Techniques you can use to manage intrinsic cognitive load Chapter 8. Transition from Worked Examples. Research on worked examples; how to adjust worked examples as learners gain expertise as well as formatting guidelines for worked examples

Sample Excel e-Lessons Before: Overloaded Excel Web-Based Lesson. This asynchronous e-learning sample violates most of the guidelines for reducing extraneous cognitive load discussed in this part of the book. John Sweller’s commentary of the Overloaded Excel Web-Based Lesson specifies the cognitive load violations in this asynchronous e-learning sample. After: Load-Managed Excel Web-Based Lesson. This asynchronous e-learning sample applies most of the guidelines for reducing extraneous cognitive load discussed in this part of the book. John Sweller’s commentary of the Load-Managed Excel Web-Based Lesson specifies the cognitive load guidelines exemplified in this asynchronous e-learning sample. Virtual Classroom Example. This synchronous e-learning sample applies most of the guidelines for reducing extraneous cognitive load discussed in this part of the book.

c03.qxd 9/29/05 17:56 Page 46

CHAPTER OUTLINE Accelerate Expertise with Dual Modalities Guideline 1: Use Diagrams to Optimize Performance on Tasks Requiring Spatial Manipulations How Diagrams Promote Efﬁcient Performance

Guideline 2: Use Diagrams to Promote Learning of Rules Involving Spatial Relationships Applying the Research Guideline 3: Use Diagrams to Help Learners Build Deeper Understanding Are All Diagrams Equal? The Psychology of Diagrams for Deeper Learning

Guideline 4: Explain Diagrams with Words Presented by Audio Narration Applying the Research When to Use Audio to Explain Visuals The Delivery Medium Must Carry Audio Back-Up Audio with Text to Accommodate Learners with Hearing Impairments Use Audio to Explain High Complexity Content Use Audio for Low Prior Knowledge Learners Use Diagrams and Audio Only When Diagrams and/or Text Require Explanations Use Text Rather than Audio When Learners Need Reference to Content

The Bottom Line Using Audio to Describe Text Rather Than Diagrams On the CD John Sweller Video Interview Sample Excel e-Lesson

c03.qxd 9/29/05 17:56 Page 47

3 Use Visuals and Audio Narration to Exploit Working Memory Resources

T

Lesson

WO PREVALENT AND POWERFUL TOOLS instructional profes-

sionals use to promote learning are visuals and auditory descriptions of those visuals. Whether delivered in a classroom or on a computer, the effective use of visual and auditory modalities is directly linked to efﬁcient instruction. What evidence do we have about the best ways to use the visual and the auditory modes to support learning? In this chapter we present research, psychological rationale, and examples to support four guidelines related to the use of visuals alone and to the explanation of visuals.

Guidelines for Audio-Visual Instruction 1. Use diagrams to optimize performance on tasks requiring spatial manipulations. 2. Use diagrams to promote learning of rules involving spatial relationships. 3. Use diagrams to help learners build deeper understanding. 4. Explain diagrams with words presented in audio narration.

47

c03.qxd 9/29/05 17:56 Page 48

48

Efficiency in Learning

Accelerate Expertise with Dual Modalities Imagine that you are developing an e-lesson on how to use a computer application such as how to input a formula into a spreadsheet. You’ve designed an animated demonstration showing how to activate the correct cell in the spreadsheet, type the formula in it, and press enter to see the result. You could explain the animated demonstration with either on-screen text or with audio narration. Which would be more efficient? What research do we have about the best ways to use visuals and audio for performance and for learning? As we have discussed in Chapters 1 and 2, cognitive load refers to the amount of work imposed on working memory. One factor that determines cognitive load is the complexity of the content—what instructional scientists call element interactivity. Complex content or instructional goals require learners to devote working memory capacity to coordinate multiple elements. For example, when learning from instructional materials that teach a new computer application like Excel, you must coordinate several actions on the keyboard and mouse with the screen formats. One way to accelerate expertise when your instructional goals involve coordination and integration of several elements is to exploit two subcomponents of working memory: the auditory (phonetic) center and the visual center. This is the first of three chapters on how to best make use of auditory and visual modalities to extend the limited capacity of working memory.

WHAT THE RESEARCH MULTITA SKING

SAYS

ABOUT

Recall in Chapter 2 that we discussed dual task experiments in which subjects are required to perform a primary task and a simultaneous secondary task. For example, subjects are asked to remember a list of random numbers read to them at the same time that they trace an image on a computer screen with a mouse. In dual task experiments in which each task requires a different mode, for example, an auditory task and a visual task, performance on the primary task is not depressed signiﬁcantly.

c03.qxd 9/29/05 17:56 Page 49

49

Use Visuals and Audio Narration

In contrast, when each task uses the same mode, for example, two auditory tasks, performance plummets. Even though we have two ears, our capacity for simultaneously assimilating multiple different auditory messages is quite limited. Based on these dual-task experiments, psychologists have inferred that working memory has two subcomponents: one for auditory information called the phonological loop and a separate one for visual information called the visual-spatial sketch pad (Baddeley, 1992). The phonological loop processes auditory, mainly verbal, information while the visual-spatial component processes diagrams and pictures. In Figure 3.1 we represent the phonological loop as a phonetic folder and the visualspatial sketch pad as a visual folder in working memory.

Figure 3.1. Working Memory Includes a Phonetic (Auditory) and Visual Component. Source: Clark Training & Consulting.

MEMORY Long-term

Working

Encoding Retrieval

Schemas

Phonetic Visual Rehearsal

ATTENTION

In this chapter we will look at evidence-based ways to accelerate learning through judicious use of visual modalities including graphics and diagrams as well as auditory modalities such as narration. Our guidelines regarding diagrams apply to any instructional medium that can display visual information including computer screens, white boards in physical or virtual classrooms, and workbooks. Our guidelines regarding audio apply to any instructional

c03.qxd 9/29/05 17:56 Page 50

50

Efficiency in Learning

medium that can deliver sound including instructors in classrooms, computers used for e-learning, and video. We begin with evidence-based guidelines for the best use of visuals followed by a discussion of how to maximize learning by describing visual information with audio narration.

Guideline 1: Use Diagrams to Optimize Performance on Tasks Requiring Spatial Manipulations Whether assembling your child’s bicycle or clearing a jam in the copy machine, all of us have faced tasks requiring us to manipulate objects. Commonly we refer to working aids with instructions to help us complete these kinds of tasks. Under what circumstances do diagrams aid performance on such spatial tasks? Research shows that we perform spatial tasks faster when we refer to self-explanatory diagrams than when we refer to text descriptions. This guideline applies primarily to individuals who are new to the task and to tasks involving relatively complex spatial relationships.

WHAT THE RESEARCH SAYS ABOUT USING DIAGRAMS IN PERFORMANCE AIDS Marcus, Cooper, and Sweller (1996) asked sixth graders to perform one of three tasks requiring connection of resistors in three different conﬁgurations. As shown in Figure 3.2, these tasks range from quite simple (single series) to relatively complex (parallel) conﬁgurations. Half of the participants were provided performance aids with textual directions, while the other half used performance aids with diagrams. Therefore the experiment compared six conditions: three resistor connection patterns explained by text and three explained by diagrams. Figure 3.2 shows the text and diagram versions for both the simple single series connection and the more complex parallel connection. As you can see, either the diagram or the text was self-explanatory

c03.qxd 9/29/05 17:56 Page 51

51

Use Visuals and Audio Narration

Figure 3.2. A Simple and Complex Assembly Task Explained with Text and with Diagrams. Adapted From Marcus, Cooper, and Sweller (1996).

Text Versions

Diagram Versions

Single Series Connections Using the resistors supplied, make the following connections: • Connect one end of a 2 ohm resistor to one end of a 3 ohm resistor • Connect one end of a 7 ohm resistor to one end of a 5 ohm resistor Etc.

2 OHM 7 OHM

3 OHM 5 OHM

Parallel Connections Using the resistors supplied, make the following connections: • Connect one end of an 8 ohm resistor to one end of a 3 ohm resistor, and connect the other end of the 8 ohm resistor to the other end of the 3 ohm resistor. • Connect one end of the 3 ohm resistor to one end of a 5 ohm resistor, and connect the other end of the 3 ohm resistor to the other end of the 5 ohm resistor.

8 OHM 3 OHM 5 OHM

and could support completion of the task. The research team measured the time participants needed to complete the connections, as well as the participants’ ratings of task difﬁculty. Figure 3.3 shows the time used to complete each of three connections. As indicated in Figure 3.3, the diagrams helped participants complete the connections faster for simple and complex connections. However, notice that the time differences between text and diagram were most pronounced with the more complex parallel connections. In this experiment the participants using diagrams as a performance aid completed the more complex task over two and one-half times faster compared to those using text! When rating task difﬁculty, all groups rated the parallel connections as more difﬁcult than the others. However, those using the text work aid rated the task more difﬁcult than those using the diagram aid.

c03.qxd 9/29/05 17:56 Page 52

52

Efficiency in Learning

Figure 3.3. Diagrams on Performance Aids Lead to Faster Task Performance. Based on Experiment 1 (Marcus, Cooper, & Sweller, 1996).

180 Average Time in Seconds

160 140

Text Instructions Diagram instructions SD = significant difference

120 110

SD

90 70 50 Single Series Multiple Series Parallel Type of Resistor Connection

Recall from Chapter 1 that we can visually compare efﬁciencies of different instructional materials with an efﬁciency graph. Figure 3.4 shows a plot of the six different conditions on the efﬁciency graph. Note that the data point corresponding to the parallel connections explained by text is located in the low efﬁciency portion of the graph (extreme lower right hand corner). Note also that the efﬁciency ratings of all the aids with diagrams fall quite close together in the high efﬁciency portion of the graph (upper left quadrant). The efﬁciency graph clearly illustrates the superiority of performance aids that use diagrams and the inefﬁciency of aids that use text—especially for the more complex parallel connections.

c03.qxd 9/29/05 17:56 Page 53

53

Use Visuals and Audio Narration

Figure 3.4. Diagrams Are More Efﬁcient Than Text as Work Aids. Source: Marcus, Cooper, and Sweller (1996).

High Efficiency Parallel Diagram

E=0

1.0 0.8

Multiple series Diagram Single series Diagram

- 1.0 - 0.8 - 0.6 - 0.4 - 0.2 Single series Text

0.6 0.4 0.2

Mental Effort

0.2 0.4 - 0.2

0.6

- 0.4

0.8

1.0

Multiple series Text

- 0.6 Parallel Text

- 0.8 - 1.0

Low Efficiency

The research supporting Guideline 1 shows that performance aids using diagrams are more efficient than aids using text. In other words, diagrams can serve as psychological equalizers by allowing you to complete high complexity tasks in about the same amount of time as low complexity tasks with less mental effort.

How Diagrams Promote Efﬁcient Performance What is it about diagrams that make them easier to process psychologically? All elements in a visual can be viewed simultaneously, unlike sentences, which must be processed sequentially one at a time. This leads to a lower visual search for tasks that involve coordination of multiple spatial elements. Greater psychological processing efficiency is the result. Likewise, diagrams provide more explicit representations of spatial tasks. A diagram requires fewer inferences because it shows spatial relationships

c03.qxd 9/29/05 17:56 Page 54

54

Efficiency in Learning

that must be inferred from textual descriptions (Larkin & Simon, 1987). There is a closer correspondence between the diagram and the requirements of the task.

Guideline 2: Use Diagrams to Promote Learning of Rules Involving Spatial Relationships For Guideline 1, we reviewed evidence that working aids for assembly tasks that used diagrams resulted in faster performance than aids that used textual instructions. The focus in that section was performance, rather than learning. In other words, the goal was to accomplish a task while using the aid, rather than to store new knowledge or skills in memory. Guideline 2 rests on evidence that diagrams support learning of content involving spatial relationships better than text does.

WHAT THE RESEARCH SAYS ABOUT USING DIAGRAMS TO PROMOTE LEARNING Carlson, Chandler, and Sweller (2003) constructed alternative chemistry lesson versions to teach students how to name carbon compounds by applying rules regarding the preﬁx and sufﬁx of the compound. The rule for the preﬁx portion of the carbon compound is a simple association between a name and the number of carbon atoms in the molecule. For example, the preﬁx propan- is applied to any compound with three carbon atoms. To apply the preﬁx rule, you simply need to count the number of carbon atoms and associate it with the correct name. In contrast, the naming rule for the sufﬁx portion of the compound is more complex. It requires the student to consider the relative positions and types of bonds of various atoms. Figure 3.5 shows an example of one sufﬁx rule presented in text and by diagram. Therefore, learning and applying the preﬁx rule is likely to impose less cognitive load than learning to apply the sufﬁx rule. In this experiment, the students were provided lesson materials that used either text or diagrams to explain the preﬁx rule and the sufﬁx rule. After studying the preﬁx and the sufﬁx rules (using either text or diagrams), they rated the difﬁculty of the

c03.qxd 9/29/05 17:56 Page 55

55

Use Visuals and Audio Narration

Figure 3.5. A Text and Diagram Version of a Chemistry Compound Sufﬁx Rule. Source: Carlson, Chandler, and Sweller (2003).

Text Version The suffix used if the oxygen structure is a carbon atom and oxygen atom joined by a carbon-oxygen double bond and to the same carbon atom a single bonded nitrogen atom which has two hydrogen atoms attached to it by two single bonds is -amide

Diagram Version O C N

H

-Amide

H

materials using a seven-point mental effort scale described in Chapter 1. The study period was followed by two test phases. The ﬁrst test phase required participants to apply the preﬁx and sufﬁx rule to twelve compounds not seen during the lesson. Participants could use their study materials as reference guides. Therefore this test phase was similar to the resistor connection study discussed in the previous section. In both situations the participants had access to a reference guide during task performance. The second test phase included six problems to be completed without reference to the instructional materials. Test Phase 1, in which learners had access to their instructional materials, resulted in a pattern similar to the one described for the resistor connection task. Materials with diagrams resulted in better scores for both the simple rules (preﬁx names) and complex rules (sufﬁx names) with the greatest advantage seen for the more complex sufﬁx rules. Difﬁculty ratings were highest for the more complex sufﬁx rule and for textual descriptions—most notably textual descriptions of the more complex sufﬁx rule. The efﬁciency graph revealed a pattern similar to the one seen in the resistor study. Test Phase 2 required a similar performance, naming of carbon compounds. However, on this test, learners were not allowed to reference their instructional materials. Therefore performance depended on learned ability to apply the two rules. Learners studying from materials with diagrams were able to apply the more complex sufﬁx rule from memory more effectively than learners who studied from the text versions. However, for the simpler preﬁx rule, the text materials resulted in better achievement.

c03.qxd 9/29/05 17:56 Page 56

56

Efficiency in Learning

The efﬁciency graph shown in Figure 3.6 illustrates the low efﬁciency of the textual explanation of the more complex sufﬁx rule at about the same graph location as the Phase 1 test results. Unlike Phase 1, the text explanation of the simpler preﬁx rule was more efﬁcient than the diagram, although both text and diagram materials fell into the high efﬁciency upper left quadrant. The authors conclude: “When instructional material requires learners to compare, relate, or simultaneously process multiple elements of information, then a diagrammatic representation of the materials is likely to enhance both the learning and understanding. Under conditions where instructional information can be processed serially and there are few if any relations between learning elements, then a textbased instructional format is likely to be as effective as a diagrammatic representation of the instructional material” (p. 638).

Figure 3.6. Diagrams Are More Efﬁcient Than Text for Learning of More Complex Sufﬁx Problems. Source: Carlson, Chandler, and Sweller (2003).

High Efficiency

Text Prefix

Diagram Prefix

E=0

1.0 0.8 0.6 0.4 0.2

- 1.0 - 0.8 - 0.6 - 0.4 - 0.2

0.2 0.4 - 0.2

Mental Effort 0.6

0.8

1.0

Diagram Suffix

- 0.4 - 0.6 - 0.8

Text Suffix

- 1.0

Low Efficiency

c03.qxd 9/29/05 17:56 Page 57

Use Visuals and Audio Narration

57

Applying the Research Taken together, the research supporting Guidelines 1 and 2 tells us that: • Diagrams enhance performance and learning efficiency. • Diagrams lead to greatest efficiency for spatial tasks of medium to high complexity. • For tasks of lower complexity, text representations are usually less expensive to produce and are as efficient for performance aids and instructional materials. • We recommend diagrams for both performance aids and for training materials in most situations.

Guideline 3: Use Diagrams to Help Learners Build Deeper Understanding So far we have seen that materials with diagrams alone can serve as better performance or learning aids than materials with text alone, especially with more complex content. Guideline 3 is concerned with problemsolving instructional goals that require a deep understanding of the content. Workers are often faced with challenging situations that require some degree of problem solving. For example, when troubleshooting equipment or when writing employee performance appraisals, workers must generate somewhat novel solutions to each situation. In order to perform novel tasks effectively, the worker must be able to call on a schema of sufficient flexibility that it can be applied to a variety of situations. For example, in a troubleshooting task, to isolate the fault the worker must apply knowledge of how the system works, along with a systematic troubleshooting process. A number of studies have shown that adding a relevant visual to explanatory text helps learners build a deeper understanding of the content that will aid problem solving.

c03.qxd 9/29/05 17:56 Page 58

58

Efficiency in Learning

WHAT THE RESEARCH SAYS ABOUT ADDING DIAGRAMS FOR DEEPER U ND E R S TA N D I N G Mayer (2001, 2002) has conducted a number of experiments using lessons that taught how something worked. For example, he has developed lessons designed to teach how a bicycle pump works, how lightning works, and how hydraulic brakes work. In these studies two lesson versions were tested. One consisted of text alone that explained how the system worked. The other included the same text but added an illustration of the process. In some lessons the illustrations were presented as a series of still visuals with arrows to indicate motion and in other lessons the diagrams were presented as animations. Figure 3.7 shows a portion of the text and the text-plus diagram versions from the bicycle pump lesson.

Figure 3.7. Text and Text Plus Diagram Versions from the Bicycle Pump Lesson. Source: Mayer (2001).

Text Version

Text Plus Diagram Version HANDLE As the rod is pulled out,

As the rod is pulled out, air passes through the piston and fills the area between the piston and the outlet valve Air passes through the piston PISTON INLET VALVE OUTLET VALVE

HOSE And fills the area between The piston and the outlet valve

c03.qxd 9/29/05 17:56 Page 59

59

Use Visuals and Audio Narration

After a study period, learners were asked to apply this new knowledge to problems that were not discussed during the lesson. For example, learners were asked: “Suppose you push down and pull up the handle of a pump several times but no air comes out. What could have gone wrong? or What could be done to make a pump more effective, that is, to move air more rapidly?” To answer questions like these, learners must have built a cause-and-effect understanding of how the parts of the equipment work together to produce the desired result. In six different experiments, learners who studied from lessons that added a diagram to the text created a median of 79 percent more solutions to questions like these than learners who studied from text alone. The median effect size was 1.37, which tells us that these results not only have statistical signiﬁcance but are of high practical signiﬁcance as well.

Are All Diagrams Equal? Not all visuals are equally effective. Different types of visuals are best suited for various instructional goals (Clark & Lyons, 2004). For example, a representational visual is a diagram that illustrates the appearance of an object. In contrast, explanatory visuals show relationships among the content elements. Some examples of explanatory visuals include organizational maps and bar charts to show qualitative and quantitative relationships and interpretive diagrams to illustrate abstract relationships and principles. When the learning goal requires a deep understanding, explanatory visuals that show relationships work best.

WHAT THE RESEARCH SAYS EXPL ANATORY VISUALS

ABOUT

Gyselinck and Tardieu (1999) compared learning of principles governing gas pressure from text alone and from text accompanied by two different types of illustrations. The illustration shown on left side of Figure 3.8 is a representational

c03.qxd 9/29/05 17:56 Page 60

60

Efficiency in Learning

Figure 3.8. A Representational and an Explanatory Illustration of Gas Pressure. Source: Gyselinck and Tardieu (1999).

Representational Visual

Explanatory Visual

Pressure

Bag Bag As the bag goes up in altitude, it will inflate because of the pressure.

visual. It does not show relationships. In contrast, the diagram on the right side of Figure 3.8 is an explanatory visual because it illustrates changes in air pressure in relation to altitude. An experimental comparison of the two diagrams with a text-only version found that learning was better from text illustrated with either of these diagrams than from text alone. However, the explanatory visuals such as the one shown on the right side were the most beneﬁcial. The research team concluded: “The results reported indicate that pictures that highlight the relationships between the objects being described in the text are the most beneﬁcial for readers, allowing them to build connections in order to draw inferences” (pp. 214–215).

The Psychology of Diagrams for Deeper Learning As we discussed in Chapter 2, expertise is based on a rich bedrock of schemas resident in long-term memory. Schemas are formed when new information entering the eyes and ears is integrated into prior knowledge. This process is called encoding. According to a theory called dual encoding,

c03.qxd 9/29/05 17:56 Page 61

Use Visuals and Audio Narration

61

learning can be deeper when participants have multiple opportunities to encode information. When studying materials such as those on the right side of Figure 3.8, learners have the opportunity to encode information in two ways—through the words and through the visuals. Learners read the words and view the visual and then integrate those two sources of information into a coherent explanation. Learners have two opportunities to build a new schema: one from the text and one from the visual representation. However, the extra work imposed on working memory to integrate the two sources of information will increase the cognitive load of materials that include text and related visuals. Since the added load results in deeper learning, this is a form of useful load—the kind we called germane. Since these types of lessons are more demanding of limited working memory capacity, we need to offset this germane load by finding ways to present the visuals and the words in the most cognitively efficient manner possible. We will look at ways to do this under Guideline 4 and in Chapters 4 and 5.

Guideline 4: Explain Diagrams with Words Presented by Audio Narration One way to reduce load imposed by the need to coordinate visuals and words is to divide the new incoming content among the two storage areas in working memory—the visual and the auditory. In our CD asynchronous demonstration lesson on creating Excel formulas, our default version uses audio narration to explain animated demonstrations. However, if the delivery platform does not have sound capability or some learners have hearing impairments, the audio can be turned off and is replaced by onscreen text. Research has indicated that the default version using audio narration results in more efficient learning. Cognitive efficiency gained when visuals are explained by auditory narration is referred to as the modality effect.

c03.qxd 9/29/05 17:56 Page 62

62

WHAT THE RESEARCH SAYS T HE M O D A L I T Y EF F EC T

Efficiency in Learning

ABOUT

One of the most compelling ﬁndings from cognitive load theory research is the modality effect. A variety of experiments using different content, learners, and contexts have shown that learning is better when diagrams are explained with audio narration than when diagrams are explained with text. This guideline applies to situations in which the visual is not self-explanatory; in other words the visual requires an explanation to be understood. The superiority of explanations provided in narration has been shown for static diagrams as well as for animated visuals. As summarized in Table 1.1 in Chapter 1, the modality effect has been demonstrated in lessons designed to teach electrical troubleshooting to trade school apprentices, mechanical and scientiﬁc processes to college students, and botany concepts to seventh graders, among others. In the following paragraphs we summarize a few of these studies on the modality effect.

LEARNING PROCEDURES Tindall-Ford, Chandler, and Sweller (1997) trained trade apprentices how to conduct electrical tests of appliances using a volt meter. Figure 3.9 shows the visual-only version that explained the diagram with text. A second version (AV format) described this same visual using the same words presented in audio narration. Each lesson lasted about ﬁve minutes. The test included written problem-solving questions as well as a practical test that required the learners to perform the electrical test using actual equipment. One month later a second training session was conducted with the same learners using the same materials. Figure 3.10 shows the test results from the second training session for both groups. As you can see, the AV format resulted in higher scores.

LEARNING PROCESSES Previously in this chapter we described experiments conducted by Mayer (2001, 2002) and his colleagues that focused on teaching college students how bicycle pumps, brakes, and lightning work. In a comparison of these lessons, in which words were presented in text, with lessons in which the same words were presented in

c03.qxd 9/29/05 17:56 Page 63

63

Use Visuals and Audio Narration

Figure 3.9. A Visual-Only Version from an Electrical Test Lesson. Source: Tindall-Ford, Chandler, and Sweller (1997).

Kettle

OFF ON

Switch Frame of Appliance Other Lead

neutral Plug of Appliance

active Earth Lead

earth Earth

M

PRESS TO TEST

Line

METER

1000 V 500 V 240 V

TEST BUTTON

Test 2: The insulation resistance between the electrical element and the frame. 1. Set the meter to read 500 V. 2. Make sure the appliance’s switch is “on.” 3. Place the earth lead on the active pin of the appliance’s plug. 4. Place the other lead on the frame of the appliance. 5. Press the test button. 6. Read the resistance from the meter. The required result is a reading of at least one M . 7. Remove the earth lead from the active pin and place it on the neutral pin. 8. Press the test button again. 9. Read the resistance. A reading of at least one M is again required.

c03.qxd 9/29/05 17:56 Page 64

64

Efficiency in Learning

Figure 3.10. Learning Is Better from Audio-Visual Presentations. Based on data in Phase 2 (Tindall-Ford, Chandler, & Sweller, 1997).

Visual Explained by Text

20

Visual Explained by Audio SD = Significant difference

Test Score

15

Test Max = 20 10 SD 5

audio narration, the narrated versions always resulted in better problem-solving learning. In four of four comparisons in which the visuals were animated, the audio versions resulted in better learning, with a median gain of 80 percent (Mayer, 2001). The median effect size from all of these experiments was 1.17, indicating that the results are not only statistically signiﬁcant but have considerable practical signiﬁcance as well.

LEARNING FROM AGENTS Learning agents are on-screen characters presented in human, animal, or inanimate forms included to help learners achieve the instructional goal. For example, Figure 3.11 shows a learning agent from our CD Excel asynchronous e-lesson. Research has shown that learning is better in the presence of a learning agent (Clark & Mayer, 2003).

c03.qxd 9/29/05 17:56 Page 65

Use Visuals and Audio Narration

65

Figure 3.11. A Learning Agent from an Excel Lesson.

Given the beneﬁts of learning agents, a number of recent experiments have examined how best to present agents to get the most learning value from them. Agents with different representations (human versus cartoon animal versus no image) that presented information and advice delivered either by text or by audio have been tested in lessons designed for college and seventh grade students. The image does not seem to be very important for learning. Moreno, Mayer, Spires, and Lester (2001) found their agent was equally effective in the form of a cartoon bug, a human video image, or no image at all! However, the modality of the agent’s words makes a difference. Four studies that used different types of agents and focused on different instructional goals all found that learning is better when the agent communicates with the learner by audio rather than by text (Atkinson, 2002; Craig, Gholson, & Driscoll, 2002; Mayer, Dow, & Mayer, 2003; Moreno, Mayer, Spires, & Lester, 2001).

c03.qxd 9/29/05 17:56 Page 66

66

Efficiency in Learning

Applying the Research Taken together the research supporting the modality effect tells us that: • When a visual requires an explanation, learning is more efficient when the explanation is provided with audio narration rather than with text. • Both still and animated visuals benefit from an audio explanation. • Providing content in a complementary visual and auditory format reduces cognitive load by dividing content between the visual and auditory centers in working memory.

When to Use Audio to Explain Visuals Of course not all situations lend themselves to using audio, and research shows that, even when it is available, it is not always beneficial. We recommend that you describe diagrams using audio narration rather than text when: (1) the delivery medium can support audio; (2) learners are not hearing impaired; (3) the instructional goal requires higher levels of mental work; (4) learners are novice to the skills being trained; (5) the diagram and/or text require an explanation; and (6) learners will not need reference to the content. In this section we discuss each of these conditions that will influence your decision to use audio narration.

The Delivery Medium Must Carry Audio Naturally your delivery medium has to be able to carry audio. Probably the two most common settings where audio can be used to describe visuals are multimedia training delivered on computer and instructor-led training either in actual or virtual classroom sessions. For example, in a synchronous web cast, learning is most effective when the instructor explains a relevant visual, as shown in our Excel virtual classroom lesson on our CD. If the instruction is delivered in books, you will most likely want to use other proven methods to effectively reduce load. We describe these in the next chapter.

c03.qxd 9/29/05 17:56 Page 67

67

Use Visuals and Audio Narration

Back-Up Audio with Text to Accommodate Learners with Hearing Impairments Many organizations require that all training programs accommodate learners with visual and auditory impairments. For these situations, secondary text explanations should be available to replace the audio. As we will see in Chapter 5, you should not present the same explanations in both audio and text simultaneously. Identical text and audio are redundant and depress learning by overloading working memory with unnecessary information.

Use Audio to Explain High Complexity Content Audio narration will be most helpful when the cognitive load is highest. If the instructional goal and/or content are relatively simple, presenting words with text will be as effective for learning as presenting words with audio narration. For example, consider the graph shown in Figure 3.12. Suppose you were asked to find the average rate of temperature change on Tuesday between 12 P.M. and 2 P.M.? How much mental effort would you need to answer this question compared to finding the times on Monday that had a zero average

Figure 3.12. A Graph of Temperature Changes over Time for Two Days. Source: Leahy, Chandler, and Sweller (2003).

Monday Tuesday

Temperature (C) 36 34 32 30 28 26 24 22 20 18 16 9am

10am

11am

12pm

1pm

Time of Day

2pm

3pm

4pm

c03.qxd 9/29/05 17:56 Page 68

68

Efficiency in Learning

rate of change? In order to find a time that has a zero average rate of change for any given day, the student needs only to locate the correct day and look at the time periods for a flat line. In contrast, in order to find an average rate of temperature change for a given day and time the student must (1) locate the correct day, (2) locate the first time period, (3) identify the temperature for that time, (4) locate the second time period, and (5) identify the temperature for that time. Next the student must (6) subtract the lowest temperature from the highest temperature, (7) subtract the earlier time from the later time and divide the answer for Step 5 by the answer for Step 6. As you can see, the average rate of change question requires you to identify a number of different variables on the graph and to perform several mathematical operations. In contrast, the zero rate of change question only requires you to identify straight line segments and associate them with specific hours. Leahy, Chandler, and Sweller (2003) compared a text explanation with an audio explanation of this graph on high and low complex questions such as the ones we just described. Figure 3.13 shows the results. As you

Figure 3.13. Audio Explanations Result in Better Achievement on Complex Questions. Adapted from Leahy, Chandler, and Sweller (2003).

Audio Version Text Version

Mean Percentage Correct

100

SD=Significant Difference 80 60 40 SD 20

Easy Problems

Complex Problems

c03.qxd 9/29/05 17:56 Page 69

Use Visuals and Audio Narration

69

can see, there are no real differences for the easy questions. As we would expect, the audio version helped learning on the more complex operations and had little effect on questions that did not require much mental effort.

Use Audio for Low Prior Knowledge Learners A third factor to consider is the experience of the learner. Novices are most subject to cognitive overload. As learners gain experience, we would expect that the modality effect would taper off. In fact, Kalyuga, Chandler, and Sweller (2000) found that, as learners became more experienced as they progressed through a series of lessons, the effectiveness of the audio versions decreased. Once learners were quite experienced, the diagram alone was the most effective treatment. At this point, the experienced learners understood how to use the diagram and adding audio explanations was counterproductive. This is another demonstration of the expertise reversal effect that we describe in Chapter 10. Naturally, audio narration will reduce load only in situations in which the language used in the instruction corresponds to the native language of the learner. In cases in which there is a mismatch in delivery and native language, a text version is likely to impose less load.

Use Diagrams and Audio Only When Diagrams and/or Text Require Explanations Note that the superiority of audio-visual instruction over visual-only instruction (the modality effect) is only obtainable when neither the diagram nor the text can be understood without the other. For example, in our Excel demonstration lessons on the accompanying CD, neither the visual animated demonstration nor the descriptive words is self-explanatory. Learners need both for understanding. You will need to apply different guidelines when, for example, a text redescribes a diagram that is self-explanatory. In these situations the text and the diagram become redundant. We describe how to deal with this situation in Chapter 5.

c03.qxd 9/29/05 17:57 Page 70

70

Efficiency in Learning

Use Text Rather than Audio When Learners Need Reference to Content Although audio explanations of a relevant visual can reduce extraneous cognitive load, there are times that presenting information in audio might add extraneous load. Because audio is a transient modality, any content that must be referenced during the training should be available in a text format. Directions or data needed to complete an exercise are two common examples. For example, in our demonstration Excel virtual classroom lesson on the CD, the instructor writes the exercise directions on the spreadsheet for learners to reference, as shown in Figure 3.14.

Figure 3.14. Exercise Directions Displayed in Text Rather Than Audio. From the CD accompanying this text.

c03.qxd 9/29/05 17:57 Page 71

71

Use Visuals and Audio Narration

Using Audio to Describe Text Rather Than Diagrams All of the studies we’ve described on the modality effect so far focused on the use of audio to explain diagrams. However, the same benefits are realized when the material being explained is in the form of text. If you have written text that requires an explanation, it is better to provide the explanation in spoken rather than written form.

WHAT THE RESEARCH SAYS ABOUT U S I N G AU D I O T O DES C RIBE TEXT Mousavi, Low, and Sweller (1995) evaluated learning from worked examples of geometry problems. Figure 3.15 shows a geometry problem statement presented in text and in diagrams. The solution statements were presented in text

Figure 3.15. A Geometry Problem and Solution Example Presented in Combinations of Text, Diagram, and Audio. Source: Mousavi, Low, and Sweller (1995).

1. Problem Statement Text Version

The length of a rectangle is 3 cm. Its width is 1 cm. What is the perimeter of the rectangle?

Diagram Version 3 cm

1 cm

2. Solution Statements Text Version

Length + width = 3 cm + 1 cm = 4 cm The perimeter = 4 cm X 2 = 8 cm

Audio Version

“Length plus width equals three centimeters plus one centimeter equals four centimeters. The perimeter equals four centimeters times two equals eight centimeters”

c03.qxd 9/29/05 17:57 Page 72

72

Efficiency in Learning

or by an audio narration. The time required to solve test problems was compared among four groups: (1) text problem explained with text solution; (2) text problem explained with audio solution; (3) diagram problem explained with text solution; and (4) diagram problem explained with audio solution. Figure 3.16 shows the time each group needed to complete the test. Note in the two white bars that, when the problem was presented in text, the solution presented in audio was a bit more effective than when presented in text. The authors conclude: “The results of this experiment demonstrate the superiority of auditory solution statements over written solution statements regardless of whether problem information is presented in written or diagrammatic form” (p. 329).

Figure 3.16. Diagrams and Text Explained by Audio Lead to Faster Performance. Based on data from Mousavi, Low, and Sweller (1995).

Problem presented in text Mean Seconds to Complete Test

300

Problem presented in diagram 250

200

150

100

Text Solution

Audio Solution

•••

Text Solution

Audio Solution

c03.qxd 9/29/05 17:57 Page 73

Use Visuals and Audio Narration

73

The Bottom Line There are solid research and psychological reasons for recommending that you: • Use diagrams in place of text to support performance of tasks involving spatial relationships when the diagram is self-explanatory • Add explanatory diagrams to words when the goal is to help learners build a deep understanding • Explain visual content—either diagrams or text with words presented in audio narration when the tasks or content is complex All of the above recommendations apply to workers or learners who are novice to the task or content as well as to tasks that are more complex.

Tips for Design of AV Lessons • When planning your training, begin by considering best ways to visualize your content. • Use explanatory visuals that show relationships when your goal is to help learners build a deeper understanding. • In synchronous e-learning or classroom instruction, use the white board or projector to show relevant visuals (diagrams or text); explain the visuals verbally. • In asynchronous e-learning, allow learners to control the delivery of audio by providing stop and replay options. • Keep audio scripts concise, relevant, and brief. • Use text rather than audio for content that must be referenced during the lesson. • Synchronize audio explanations with presentation of the visuals. • Provide text as a backup for audio explanations for learners who may not have sound functionality and/or may have hearing impairments.

c03.qxd 9/29/05 17:57 Page 74

74

Efficiency in Learning

On the CD John Sweller Video Interview Chapter 3: Using Visuals and Audio Narration. John describes the best techniques for display and explanation of spatial information using graphics and audio.

Sample Excel e-Lesson The asynchronous lesson illustrates effective use of visuals, audio, and text. Note the use of: 1. Relevant visuals to illustrate procedures. 2. Audio narration as the default modality to explain visuals. 3. Text on screens that require learner responses. The synchronous lesson Virtual Classroom Example illustrates effective use of visuals, audio, and text, including instructor explanations of relevant visuals on the white board and use of text to display directions. COMING NEXT In this chapter we have summarized basic guidelines for use of visuals and explanations of visuals using audio narration in ways that best manage cognitive load. In the next two chapters we will expand on these guidelines. In some cases your delivery medium will not have audio capability and you will have to rely on text to explain visuals. In Chapter 4 we describe ways to manage cognitive load by focusing learner attention to relevant portions of visual or text displays and to avoid split attention by integrating text close to the visuals. In Chapter 5 we review research recommending a minimal display of content needed for understanding. The guidelines in Chapter 5 mitigate against describing a visual with both text and audio.

Recommended Readings Clark, R., & Lyons, C. (2004). Graphics for learning. San Francisco, CA: Pfeiffer. Marcus, N., Cooper, M., & Sweller, J. (1996). Understanding instructions. Journal of Educational Psychology, 88(1), 49–63.

c03.qxd 9/29/05 17:57 Page 75

Use Visuals and Audio Narration

75

Mayer, R.E., & Moreno, R. (1998). A split attention effect in multimedia learning: Evidence for dual processing systems in working memory. Journal of Educational Psychology, 90(2), 312–320. Moreno, R., Mayer, R.E., Spires, H.A., & Lester, J.C. (2001). The case for social agency in computer-based teaching: Do students learn more deeply when they interact with animated pedagogical agents? Cognition and Instruction, 19(2), 177–213. Tindall-Ford, S., Chandler, P., & Sweller, J. (1997). When two sensory modes are better than one. Journal of Experimental Psychology: Applied, 3(4), 257–287.

c04.qxd 9/29/05 17:58 Page 76

CHAPTER OUTLINE Attention and Working Memory Guideline 5: Use Cues and Signals to Focus Attention to Important Visual and Textual Content Use Signals to Draw Attention to Verbal Information When to Use Signals Applying the Research

Guideline 6: Integrate Explanatory Text Close to Related Visuals on Pages and Screens to Avoid Split Attention Applying the Split Attention Principle Applying the Research

Integrate Words and Visuals for Teaching Computer Applications in One Delivery Medium to Avoid Split Attention Faster Learning from Materials Integrated on the Computer Applying the Integrated Media Research to Your Training

The Bottom Line On the CD John Sweller Video Interview Sample e-Lessons

c04.qxd 9/29/05 17:58 Page 77

4 Focus Attention and Avoid Split Attention

W

HEN LEARNERS ARE FACED WITH content and delivery systems

that impose a relatively high cognitive load, you should use instructional methods both to focus attention and to avoid split attention. To focus attention, use methods that direct cognitive resources toward relevant content in the training materials. For example, you can use arrows or circles as cues to draw attention to visual displays. And you can use bolding or vocal emphasis as signals to draw attention to verbal material presented in text or in narration. By split attention we refer to the extraneous cognitive load imposed when the learner needs to integrate two or more dependent sources of visual information that are physically separated. In this chapter we present research, psychological rationale, and examples to support three guidelines that increase instructional efﬁciency by optimizing attention.

Guidelines for Supporting Attention 5. Use cues and signals to focus attention to important visual and textual content. 6. Integrate explanatory text close to related visuals on pages and screens. 7. Integrate words and visuals used to teach computer applications into one delivery medium 77

c04.qxd 9/29/05 17:58 Page 78

78

Efficiency in Learning

Attention and Working Memory In 1972 an Eastern Airlines flight crashed in the Florida Everglades as a result of cockpit distractions. The crew became so preoccupied with a malfunction that no one noticed the altimeter reading or warnings until it was too late to make a correction. While attention failures during learning will not have as severe consequences, the loss in instructional efficiency from instructional environments that fail to support attention can be costly. Sternberg (1996) defines attention as “the phenomenon by which we actively process a limited amount of information from the enormous amount of information available through our senses, our stored memories, and our other cognitive processes” (p. 69). Attention helps learners filter out irrelevant data and focus on the important elements in the instructional environment. There are two sides to the coin when it comes to supporting attention during training. On the one side, you can use instructional methods called cues and signals to direct attention to the important elements of the training. On the other side, you can use instructional methods to minimize extraneous cognitive load arising from split attention. In this chapter we will look at evidence-based ways to do both in order to free up as much working memory as possible for learning processes. Support for attention will be most important when the learners are novices, when the learning content is complex, and when the training materials are presented dynamically, requiring immediate processing. A narrated animated sequence presented in a multimedia lesson as well as a classroom lecture are two examples of dynamic delivery formats. Both deliver content at a rate outside of the control of the learner and demand extra attentional support. In contrast, content presented in a workbook is static and typically can be reviewed and revisited at the learner’s own pace.

Guideline 5: Use Cues and Signals to Focus Attention to Important Visual and Textual Content Guideline 5 recommends the use of cues such as arrows and lines to draw attention to critical portions of complex visual displays as well as signals such as italics or vocal emphasis to draw attention to important words or sections presented as text or as audio narration.

c04.qxd 9/29/05 17:58 Page 79

Fo c u s A t t e n t i o n a n d A v o i d S p l i t A t t e n t i o n

79

WHAT THE RESEARCH SAYS ABOUT C U ES T O D R AW ATTEN TION TO VISUAL DISPL AY S In Chapter 3 we recommended that you explain complex visual displays with words presented with audio narration. The learning beneﬁts gained from audio explanations of visuals is called the modality effect. However, a study by Jeung, Chandler, and Sweller (1997) reported that the modality effect is only realized when cues reduce visual search. The research team compared learning from three different multimedia explanations of geometry examples: (1) a diagram with no cuing explained by audio narration; (2) a diagram cued with ﬂashing explained by audio narration; and (3) a diagram with no cuing explained by text. For the complex diagrams, learning from narration was better than learning from text only when the relevant portion of the diagram was cued to draw attention to the portion being explained.

Because audio explanations are transient, extra cuing of related visuals is needed when the visuals are complex. Whether teaching in the classroom or online, the use of highlighters, red circles, or arrows are common methods used to direct learner attention to a relevant section of a complex visual display.

Use Signals to Draw Attention to Verbal Information From Ezines to email to traditional print publications, professionals are inundated with textual information. One way to help readers find and process textual information efficiently is to provide signals. Signals in text are analogous to arrows and circles used to focus attention in visual displays. Signals include paragraph headings, bolding and italicizing of important words, signal phrases, and topical overviews written as content previews. A signal phrase is a short sentence or two that explicitly points out the relevance of the words to follow such as: “The following four methods are the most significant among the many techniques researched.” To see some examples of signals, compare a section from an unsignaled and a signaled text on airplane lift shown in Figures 4.1 and 4.2. In the signaled version italics and bolding draw attention to individual words. Paragraph headers are used to alert the reader to the new topic. In addition, the writer added two sentences as organizers just prior to the main text content.

c04.qxd 9/29/05 17:58 Page 80

80

Efficiency in Learning

Figure 4.1. First Paragraphs from Unsignaled Version of Passage on Airplane Lift. From Mautone and Mayer (2001). Airplane Flight – The Principles of Lift What is needed to cause an aircraft which is heavier than air to climb into the air and stay there? An aerodynamic principle formulated by Daniel Bernouille in 1738 helps explain it. Bernouille’s Principle explains how upward forces, called lift, act upon the plane when it moves through the air. A cross section of a bird’s wing, a boomerang, and a Stealth bomber all share a shape similar to that of an airplane wing. The upper surface of the wing is curved more than the bottom surface. The surface on the top of the wing is longer than on the bottom. This is called an airfoil. In order to achieve lift, air must flow over the wing. The wingspan of a 747 is more than 200 feet; that’s taller than a 14 story building. When the airplane moves forward, its wings cut through the air. As the air moves across the wing, it will push against it in all directions, perpendicular to the surface of the wing.

Figure 4.2. First Paragraphs from Signaled Version of Passage on Airplane Lift. From Mautone and Mayer (2001). Airplane Flight – The Principles of Lift

New organizer

Section heads

Bold, Italics

What is needed to cause an aircraft which is heavier than air to climb into the air and stay there? An aerodynamic principle formulated by Daniel Bernouille in 1738 helps explain it. Bernouille’s Principle explains how upward forces, called lift, act upon the plane when it moves through the air. To understand how lift works, you need to focus on differences between The top and bottom of an airplane’s wing. First, how the top of the wing is shaped differently than the bottom; second, how quickly air flows across The top surface, compared to across the bottom surface; and third, how the air pressure on top of the wing compares to that on the bottom of the wing. Wing Shape: Curved Upper Surface Is Longer A cross section of a bird’s wing, a boomerang, and a Stealth bomber all share a shape similar to that of an airplane wing. The upper surface of the wing is curved more than the bottom surface. The surface on the top of the wing is longer than on the bottom. This is called an airfoil. Air Flow: Air Moves Faster Across Top of Wing In order to achieve lift, air must flow over the wing. The wingspan of a 747 is more than 200 feet; that’s taller than a 14 story building. When the airplane moves forward, its wings cut through the air.

c04.qxd 9/29/05 17:58 Page 81

81

Fo c u s A t t e n t i o n a n d A v o i d S p l i t A t t e n t i o n

WHAT THE SIGNALS

RESEARCH

SAYS

ABOUT

Mautone and Mayer (2001) compared learning from full versions of these texts as well as a multimedia version. They used a signaled version with 496 words and an unsignaled version with 487 words. The ﬁrst experiment evaluated the effect of signals when words are presented in text, as shown in Figure 4.2. The second experiment evaluated the effect of signals when the same words were presented using audio narration. A third experiment compared signaled and unsignaled versions of the same lesson presented with narrated animation in multimedia. To signal important words in the audio versions, words that were italicized or bolded in text were read with a slower, deeper intonation. As you can see in Figure 4.3, all three signaled versions resulted in signiﬁcantly better learning, whether words were presented as text or as audio narration. Learning was between 48 percent and 44 percent better in the signaled versions. The effect sizes were .69 for textual signaling, .74 for auditory signaling, and .74 for signaling in the multimedia version, indicating moderate practical signiﬁcance.

Figure 4.3. Signaled Versions Led to Better Learning. Based on data reported by Mautone and Mayer (2001). Signaled Version 10 Unsignaled Version SD = significant difference

Test Scores

8

6

4 SD

SD

SD

2

Text Version

Audio Version

Animation-Audio Version

c04.qxd 9/29/05 17:58 Page 82

82

Efficiency in Learning

When to Use Signals Research studies have demonstrated that signals are most useful for lengthy and complex texts. Likewise, signals in the form of structured abstracts have been shown to speed reader access to technical information. Complex Texts Need Signals. Signals will benefit learning when the text is long

and complex. Lorch and Lorch (1996) compared learning from signaled and unsignaled texts—half of which were simple and half complex. They found that signals did not influence learning from simple texts but were beneficial for more complex texts. Simple texts do not need extra attentional support to make sense, as working memory can readily process unsignaled simple texts. In contrast, more complex texts make additional demands on working memory and adding signals helps to offload some of those demands. Of course, we need to keep in mind that what constitutes simple or complex material will depend not just on the material but also on the expertise of the learners. With relatively high levels of expertise, signals that aid novices may be redundant for more knowledgeable learners. Chapter 10 discusses how to adjust training materials for learner expertise. Signals Aid Access to Technical Information. Signals in the form of structured

abstracts can also speed reader access to a series of texts written by different authors. Figure 4.4 shows a structured abstract from research reported by Hartley and Benjamin (1998) that compared the readability of structured with unstructured journal article abstracts. The structured abstract uses bolded section heads to focus attention to each section of the abstract. As a whole, the structured abstract provides an organizer for the body of the article. If all authors use the same structure, all abstracts will be more consistent. Hartley and Benjamin (1998) found that structured abstracts offer readers more information in a more accessible form than do unstructured abstracts. In an age of high information density, knowledge workers must routinely skim large quantities of text materials. A structured abstract preceding a long technical article serves as a gateway into the article by

c04.qxd 9/29/05 17:58 Page 83

83

Fo c u s A t t e n t i o n a n d A v o i d S p l i t A t t e n t i o n

Figure 4.4. A Structured Abstract for a Journal Article. From Hartley and Benjamin (1998).

Background. In 1997 four journals published by the British Psychological Society – the British Journal of Clinical Psychology, the British Journal of Educational Psychology, the British Journal of Heath Psychology, and Legal and Criminological Psychology – began publishing structured abstracts. Aims. The aim of the studies reported here was to assess the effectiveness of these structured abstracts by comparing them with original versions written in a traditional, unstructured format. Method. The authors of articles accepted for publication in the four journals were asked to supply copies of their original traditional abstracts (written when the paper was submitted) together with copies of their structured abstracts (when the paper was revised). 48 such requests were made and 30 pairs of abstracts were obtained. These abstracts were then compared on a number of measures. Results. Analysis showed that the structured abstracts were significantly more readable, significantly longer, and significantly more informative than the traditional ones. Judges assessed the contents of the structured abstracts more quickly and with significantly less difficulty than they did the traditional ones. Almost every respondent expressed positive attitudes to structured abstracts. Conclusions. The structured abstracts fared significantly better than the traditional ones on every measure used in this enquiry. We recommend, therefore, that the editors of other journals in the social sciences consider adopting structured abstracts.

helping the reader quickly determine whether the article is relevant to their needs and also by providing a preview of the article.

Applying the Research The research supporting Guideline 5 tells us that: • Cues and signals improve processing of complex content. • Cues and signals are most useful when the content is presented in a dynamic format such as in multimedia animation or classroom instruction. • Structured abstracts increase accessibility of technical information.

c04.qxd 9/29/05 17:58 Page 84

84

Efficiency in Learning

Guideline 6: Integrate Explanatory Text Close to Related Visuals on Pages and Screens to Avoid Split Attention Have you ever read a book in which the text describes a visual located on the back of the page? In order to make sense of the information you need to read the text, turn over the page to study the visual, and then turn the page back again to review the text. The annoyance you feel is your working memory complaining about extraneous cognitive load needed to search for and integrate related information that is physically separated. When dependent information is laid out in separate locations, additional mental energy must be devoted to integrating the separated information sources. This additional mental effort caused by split attention takes a toll on limited working memory capacity and increases extraneous cognitive load of the materials. Learning complex information is degraded when studying materials that lead to split attention. Split attention occurs when two mutually dependent visual sources of information are laid out in a format that requires the learner to mentally integrate them. Split attention frequently arises from training materials in which a step-by-step text procedure for equipment or computer operation is displayed at a distant location from a visual of the equipment or application screen. An example is shown in Figure 4.5. Here, the text under the diagram lists nine steps for conducting an electrical test on the kettle. Neither the text nor the diagram is self-explanatory. To make sense of the procedure, the learner must read the step and then search for the relevant section of the diagram. Searching for referents and then mentally integrating them increases working memory load but does not contribute to learning. In contrast, when either the text or the visual is self-explanatory, split attention will not occur. For example, the visual on the airline safety card shown in Figure 4.6 is self-explanatory. The reader can perform the procedure simply by referring to the visual. This example would not lead to split attention because understanding the visual is not dependent on reading the text. In fact, as we will see in Chapter 5, the visual alone minus the text would be the optimal way to display this content.

c04.qxd 9/29/05 17:58 Page 85

85

Fo c u s A t t e n t i o n a n d A v o i d S p l i t A t t e n t i o n

Figure 4.5. A Separated Text Version from an Electrical Test Lesson. From Tindall-Ford, Chandler, and Sweller (1997).

Kettle

OFF ON

Switch Frame of Appliance Other Lead

neutral Plug of Appliance

active Earth Lead

earth Earth

M

PRESS TO TEST

Line

METER

1000 V 500 V 240 V

TEST BUTTON

Test 2: The insulation resistance between the electrical element and the frame. 1. Set the meter to read 500 V. 2. Make sure the appliance’s switch is “on.” 3. Place the earth lead on the active pin of the appliance’s plug. 4. Place the other lead on the frame of the appliance. 5. Press the test button. 6. Read the resistance from the meter. The required result is a reading of at least one M . 7. Remove the earth lead from the active pin and place it on the neutral pin. 8. Press the test button again. 9. Read the resistance. A reading of at least one M is again required.

c04.qxd 9/29/05 17:58 Page 86

86

Efficiency in Learning

Figure 4.6. Two Self-Explanatory Information Sources That Would Not Lead to Split Attention.

To fasten your seat belt, insert the flat tab into the metal buckle.

To tighten, pull up on the strap.

To release the seat belt, lift the tab on the top of the metal buckle.

Please stow any carryon securely underneath the seat in front of you or in the overhead bin.

Detachable luggage wheels should not be stored in the overhead bin.

In the event of an emergency landing, brace yourself using the seat in front of you.

WHAT THE RESEARCH SPL I T AT T E N T I O N

SAYS

ABOUT

The negative effects of split attention on learning have been demonstrated in controlled research experiments using paper and multimedia materials. LEARNING FROM INTEGRATED AND TRADITIONAL PRINT MATERIALS Tindall-Ford, Chandler, and Sweller (1997) compared learning how to conduct electrical tests on appliances from three different print-based lesson versions. In the traditional version, descriptive text steps were displayed under the diagram, as shown in Figure 4.5. In the integrated version, steps were placed next to the relevant portion of the diagram, as shown in Figure 4.7. A third version used an audiovisual format in which the diagram was described by words presented in audio narration.

c04.qxd 9/29/05 17:58 Page 87

87

Fo c u s A t t e n t i o n a n d A v o i d S p l i t A t t e n t i o n

Figure 4.7. An Integrated Text Version from an Electrical Test Lesson. From Tindall-Ford, Chandler, and Sweller (1997).

2. Make sure the appliance’s switch is “on.” Kettle

OFF ON

Switch 4. Place the other lead on the frame of the appliance.

Frame of Appliance

Other Lead

7. Remove the earth lead from the active pin and place it on the neutral pin. neutral Plug of Appliance

active 6. Read the resistance from the meter. The required result is a reading of at least one M .

Earth Lead earth 3. Place the earth lead on the active pin of the appliance’s plug.

Earth

M

Line

9. Read the resistance. A reading of at least one M is again required. PRESS TO TEST

5. Press the test button. 8. Press the test button again.

METER

1000 V 500 V 240 V

TEST BUTTON

1. Set the meter to read 500 V.

c04.qxd 9/29/05 17:58 Page 88

88

Efficiency in Learning

Figure 4.8. Text Separated from the Visual Led to Split Attention and Less Learning. Based on data from Phase 1, Phase 2, and practical test in Tindall-Ford, Chandler, and Sweller (1997).

60 Audio Narration Text Integrated on Page

40

Text at Bottom of Page Percent Correct

40

SD = Significant Difference

30

20

SD

10

Trade school apprentices studied the materials and then were tested performing hands-on equipment tests without reference to the instructional materials. Learning results from the three versions are compared in Figure 4.8. As you can see, those using either the integrated or audio versions learned about twice as much as those using the traditional format. LEARNING FROM INTEGRATED AND TRADITIONAL MULTIMEDIA MATERIALS Moreno and Mayer (1999) studied the effects of split attention in multimedia. They developed three versions of an animated multimedia lesson on how lightning forms. As shown in Figure 4.9, in the separated version (B) the descriptive text is located at the bottom of the screen. The integrated text version (A) placed descriptive text near the relevant portion of the diagram. A third version (C) tested the modality effect

c04.qxd 9/29/05 17:58 Page 89

89

Fo c u s A t t e n t i o n a n d A v o i d S p l i t A t t e n t i o n

Figure 4.9. A Screen from Three e-Lesson Versions on How Lightning Forms. Adapted from Moreno and Mayer (1999).

Cool moist air moves over a warmer surface and becomes heated.

A. Integrated Text Version

“Cool moist air moves over a warmer surface and becomes heated. “

C. Audio Version Cool moist air moves over a warmer surface and becomes heated.

B. Separated Text Version

by presenting the words with audio narration. As you can see in Figure 4.10, the narrated version resulted in signiﬁcantly more learning than the integrated text, which in turn resulted in more learning than the separated text. These results show that applying the modality principle (effect size of 1.06) was more beneﬁcial for learning than avoiding split attention (effect size .48). These results are different from the outcomes summarized in the previous paragraph on electrical tests, in which the integrated text and narrated versions resulted in equivalent learning. One reason for the discrepancy could be that in the lightning lessons the visuals were presented dynamically, whereas in the electrical test lessons the visual was static. In the lightning lessons the animated content played over a series of screens independent of learner control, whereas the electrical test lessons were paper-based, allowing learners to review the diagram at their own

c04.qxd 9/29/05 17:58 Page 90

90

Efficiency in Learning

Figure 4.10. Audio Description of Visuals Led to Best Learning, Followed by Integrated Text, Which Was Better Than Separated Text. Source: Moreno and Mayer (1999).

50

Audio Narration Text Integrated on Screen Text at Bottom of Screen

Percent Correct

40

SD = Significant Difference 30

20

SD

SD

10

pace throughout the study period. No doubt, lessons that display visual content dynamically, as in an animated sequence, will impose a greater cognitive load than content displayed statically on a single page. Therefore a stronger modality effect will be seen in narrated animated instruction.

SPLIT ATTENTION FROM STUDENT NOTE-TAKING Rickards, Fajen, Sullivan, and Gillespie (1997) compared learning from signaled and unsignaled lectures among learners who did and did not take notes. The signaled lecture used emphasis phrases such as: “It should be noted that” as well as verbal emphasis on important words. As you can see in Figure 4.11, note-taking contributed to learning only when the lecture included signals. Note-taking during unsignaled lectures became a source of extraneous cognitive load. To process unsignaled lectures, all cognitive resources need to be invested in organizing the content. The additional task of taking notes becomes a distraction leading to

c04.qxd 9/29/05 17:58 Page 91

91

Fo c u s A t t e n t i o n a n d A v o i d S p l i t A t t e n t i o n

Figure 4.11. Taking Notes Leads to Split Attention Unless the Lecture Is Signaled. Source: Rickards, Fajen, Sullivan, and Gillespie (1997).

Took notes

40

Test Scores

Did not take notes SD = Significant Difference

40

30

SD

SD

20

Signaled Lecture

Unsignaled Lecture

cognitive overload. Similarly, Titsworth and Kiewra (2004) found that lectures that incorporated verbal signals resulted in notes that included almost 40 percent more content than notes taken from unsignaled lectures. Recent research empirically demonstrates the extreme mental resource demands of note-taking during lectures. Piolat, Olive, and Kellogg (2005), using dual task experiments, report that taking notes from a lecture requires more cognitive effort than playing chess, reading a text, or memorizing a list of nonsense syllables. The authors conclude that “note-taking is an activity that strongly depends on the central executive functions of working memory to manage comprehension, selection, and production processes concurrently. Indeed, the severe time pressure of note-taking requires that information is both quickly comprehended and recorded in written form” (p. 306). However, when words are presented in a written format, outcomes of taking notes from the text are different. Rickards, Fajen, Sullivan, and Gillespie (1997)

c04.qxd 9/29/05 17:58 Page 92

92

Efficiency in Learning

found that when studying a simple text, note-taking increased learning whether or not signals were present in the text. Likewise, Piolat, Olive, and Kellogg (2005) report ”Not surprisingly, taking notes from reading requires less cognitive effort than from a lecture. When listening, more operations are concurrently engaged and, thus, taking notes from a lecture places more demands on working memory resources” (p. 303). When studying from text sources, learners can review and organize the content at their own pace throughout the study period. Like the two studies we reviewed in the previous paragraphs, signals as cognitive load management methods are more beneﬁcial when content is delivered in a dynamic format than when content is presented in a static form that can be processed and reviewed at the learner’s pace.

Applying the Split Attention Principle Instructional layouts that lead to split attention are quite common. In multimedia for example, mutually referring dependent text and visuals are sometimes placed on upper and lower segments of a scrolling screen or on two different screens. Likewise, it’s common to place explanatory text in the corner or bottom of the screen physically distant from the visual display. When describing on-screen visuals with dependent text, place the text near the relevant portion of the visual and, when needed, use cues to link specific text to different elements in the visual display. For example, in our asynchronous Excel e-lesson on the CD, we inserted lines from specific sentences to the relevant portion of the spreadsheet, as shown in Figure 4.12. These types of integrated layouts are most important when explaining complex tasks. Although we have emphasized alignment of visuals and text in our discussion, the split attention principle also applies to two sources of mutually referring text. For example, text may be needed to present a scenario or factual data for an online case study or text may be used to present feedback to on-screen text questions. In all cases, the screens should be designed to place related information in an integrated format. Avoid layouts like the one in Figure 4.13 that require the learner to page back to it in order to access information essential to the exercise.

c04.qxd 9/29/05 17:58 Page 93

Fo c u s A t t e n t i o n a n d A v o i d S p l i t A t t e n t i o n

93

Figure 4.12. Placement of Text and Use of Pointers to Minimize Split Attention.

Split attention is also common in training developed for print media. For example, it’s common to present procedural steps under (or worse, on a separate page from) related visuals. When planning materials for print-based working aids, workbooks, or handouts, consider the real estate limits and options of your delivery medium. For example, you might want to lay out visuals horizontally across a two-page spread or vertically along the sides of the page. In classroom presentations, you can use wall charts that duplicate commonly used visuals to minimize paging back to those visuals in the training materials. See Clark and Lyons (2004) for additional guidelines on design and layout of visuals in instructional materials. Split attention caused by taking notes from a lecture may reduce learning. The cognitive effort required to take notes reduces mental capacity that could be devoted to processing the content in ways that lead to learning. Note-taking

c04.qxd 9/29/05 17:58 Page 94

94

Efficiency in Learning

Figure 4.13. Having to Refer Back to These Directions During Practice Will Lead to Split Attention.

does not benefit higher order learning (compared to recall learning) of older learners (Kobayashi, 2005). We recommend that limited working memory resources be utilized in more productive ways than taking notes from a lecture. For example, written content summaries can be provided to learners and the instructional time used for practice exercises or discussions of the content.

Applying the Research In summary, the research on Guideline 6 tells us that: • Text explaining a visual should be placed near the visual and additional cues such as lines or arrows used to integrate the text to the relevant portion of the visual. • Text and related visuals should not be separated on a page, on different pages or screens. • Note-taking during a lecture demands considerable cognitive resources and should be minimized.

c04.qxd 9/29/05 17:58 Page 95

Fo c u s A t t e n t i o n a n d A v o i d S p l i t A t t e n t i o n

95

• When note-taking during a lecture is unavoidable, use signals in the lecture to minimize split attention.

Integrate Words and Visuals for Teaching Computer Applications in One Delivery Medium to Avoid Split Attention Under Guideline 6 we saw that it’s important to place dependent text and visuals near each other on pages or screens to avoid split attention. A common violation of split attention that occurs in training on computer applications or equipment is to present related information in two different media. For example, procedure steps are written in text in a manual and the learner simultaneously applies those steps to a computer or equipment. For example, in Figure 4.14 you see a typical computer training manual that includes steps needed to create a formula in Excel. This layout requires the learner to read each step in the manual, find the relevant portion of the screen on the computer and integrate the two. For a simple step such as “Place the cursor in Cell B2,” the cognitive load would not be great and the split manual-computer

Figure 4.14. Displaying Training Content in Two Media Leads to Split Attention.

c04.qxd 9/29/05 17:58 Page 96

96

Efficiency in Learning

format would not cause a problem. However, consider a more complex step such as “To calculate projected profit per employee, enter the formula in cell B8 using parentheses to tell Excel to perform the addition first and then a slash to tell Excel to divide the sum by the number of employees.” To apply this step the learner must read the text in the manual, find the correct entry cell on the computer screen, locate the correct cell reference to use in the formula, and apply the order of operations rule. This is a demanding task for a new Excel user, and trying to integrate the text in the manual with the various onscreen elements will add extraneous cognitive load. Instead, the instructional materials should perform the integration for the learner by placing all of the content into one delivery medium. In an integrated format either all of the content is placed in the manual and no computer is used or all of the content is placed on the computer screen and no manual is used. Figure 4.15 shows an example of an integrated manual in which all relevant visual and textual information is contained in the manual. Figure 4.16 shows one screen from a lesson in which all of the content is integrated on the computer. Figure 4.15. A Computer Training Manual That Minimizes Split Attention by Integrating Text and Visuals.

c04.qxd 9/29/05 17:58 Page 97

Fo c u s A t t e n t i o n a n d A v o i d S p l i t A t t e n t i o n

97

Figure 4.16. A Computer-Based Training Lesson That Minimizes Split Attention by Integrating Text and Visuals on the Computer. Adapted from Cerpa, Chandler, and Sweller (1996).

WHAT THE RESEARCH SAYS ABOUT T R A I N I N G CO M P UTER P ROC EDURES Learning of high complexity computer tasks is both faster and better when presented in an integrated format in the same delivery medium (Cerpa, Chandler, & Sweller, 1996; Chandler & Sweller, 1992; Chandler & Sweller 1996; Sweller & Chandler, 1994). Here we summarize the details from two of these reports. FASTER LEARNING FROM INTEGRATED MANUALS AND NO COMPUTER Sweller and Chandler (1994) compared learning of three different computer applications from conventional lesson materials (a manual with text steps plus a computer) with learning from an integrated manual where the steps and visuals were all presented in the manual. Figure 4.15 shows an example of an integrated manual for a spreadsheet lesson. The group studying the integrated manual did not have access

c04.qxd 9/29/05 17:58 Page 98

98

Efficiency in Learning

to a computer during the training phases. The group using the traditional manual with computer read each step and then tried it on the computer. The ﬁrst experiment focused on training trade apprentices several tasks involved in a CAD/CAM computer application. Participants were allowed to study the experimental materials until they had learned the tasks and completion times were recorded. After the study period, learners were given a written and practical test during which they were not allowed to refer to the training materials. The practical test was a hands-on test requiring the learners to draw a number of lines. The second experiment was similar to the ﬁrst, but it focused on training high school students spreadsheet tasks. The third experiment focused on teaching seventh graders simple word processing tasks such as how to move the cursor around on the screen. As you can see in Figure 4.17, the instructional time was faster among learners using the integrated manual in all three experiments. This is not surprising since these learners were able to study from a single source and did not spend time interacting with the computer, as did the learners using a traditional manual plus computer. On more complex tasks such as drawing a line in the CAD/CAM application or using a function in the spreadsheet application, learning was signiﬁcantly better among those who studied from the integrated manual, even though they had no access to a computer during the study period! In Experiment 3 dealing with simple word processing tasks, there was no signiﬁcant difference in learning among the instructional formats.

Figure 4.17. Integrated Materials Led to Faster Learning of Complex Software Skills. Source: Chandler and Sweller (1994). Integrated Manual No Computer

400

Traditional Manual With Computer

Time in Seconds

400

SD = significant difference 300 SD 200 SD 100 CAD/CAM

Spreadsheet

Word Processing

c04.qxd 9/29/05 17:58 Page 99

Fo c u s A t t e n t i o n a n d A v o i d S p l i t A t t e n t i o n

99

According to Guideline 7, when training computer tasks that involve interactions among several components such as the keyboard and on-screen elements, learning is faster and better when learners study from an integrated manual that is self-contained. That is, the manual includes all the necessary visual and textual information and the learner does not need to divide mental resources between the manual and the computer. Naturally, any training situation in which acquiring a new motor skill is an essential part of the task will require hands-on practice. Motor skills such as typing or performing surgery cannot be learned by studying a manual. However, in situations for which learners already have the requisite motor skills, that is, they are familiar with manipulations involving keyboards and the mouse, learning computer procedures can be more effective in the absence of the computer!

Faster Learning from Materials Integrated on the Computer In our discussion of Guideline 7 to this point, we saw that learning was faster when participants studied computer procedures from an integrated manual with no computer than when they used a traditional manual with hands-on computer practice. Alternatively, you can get efficient learning by integrating all of the instruction on the computer screen.

Cerpa, Chandler, and Sweller (1996) compared learning spreadsheet tasks of low and high complexity from a traditional manual plus computer with learning from a computer-based training format in which the instruction was contained within the computer. Similar to the integrated manual format, the CBT format incorporated all required text and visuals in a single medium—the computer in this case. Figure 4.16 shows a sample screen similar to one used in their study. The traditional manual group used the same text you see on the screen written in a manual in conjunction with a simulation of the software for computer practice. After the training period, learners took a practical test that included high- and low-complexity tasks, each of which had to be

c04.qxd 9/29/05 17:58 Page 100

100

Efficiency in Learning

completed within a given time limit. As you can see in Figure 4.18, the training formats had no effect on learning of low-complexity tasks. However, the integrated computerbased training led to signiﬁcantly better learning than the manual plus computer group on high-complexity tasks. The efﬁciency diagram shown in Figure 4.19 shows that the integrated CBT format was better for both high- and low-complexity materials, although the difference was only signiﬁcant for high-complexity tasks. In comparing the beneﬁts of a training format that integrates visuals and words in a single medium—either in an integrated manual as shown in Figure 4.15 or integrated CBT as shown in 4.16—Cerpa, Chandler, and Sweller (1996) conclude that: “Both sets of results examined together suggest that the instructional medium of delivery is largely irrelevant. The factor that seems to be more pertinent for learning is the extraneous load imposed by the instructional format. The computer-based training package used in these studies was an effective learning tool not because an electronic form of delivery was used but because it physically integrated disparate sources of information and reduced the extraneous load on working memory” (p. 364).

Figure 4.18. Integrated CBT Leads to Better Learning of HighComplexity Skills Than Manuals Plus Software. Source: Cerpa, Chandler, and Sweller (1996). 100

Integrated CBT Manual + software

Percent Correct

80

SD = Significant Difference

60

40 SD 20

Low Complexity

High Complexity

Instructional Skills

c04.qxd 9/29/05 17:58 Page 101

101

Fo c u s A t t e n t i o n a n d A v o i d S p l i t A t t e n t i o n

Figure 4.19. Integrated CBT Leads to More Efﬁcient Learning Than Manuals Plus Software. Source: Cerpa, Chandler, and Sweller (1996).

High Efficiency

E=0 0.4

CBT High Complexity

0.3

CBT Low Complexity

0.2 0.1 - 0.4

- 0.3

- 0.2

- 0.1

0.1 - 0.1 0.2

Mental Effort 0.3

0.4

- 0.2

Manual + Software Low Complexity

- 0.3

Manual + Software High Complexity

- 0.4

Low Efficiency

Applying the Integrated Media Research to Your Training While the integrated manual and integrated computer-based training both proved more effective for learning than dividing content between a manual and a computer, we recommend that when possible you integrate content on the computer. In our experience, learners are eager to work with the computer as much as possible and many learners will lose motivation if required to learn many computer tasks from an integrated manual without a computer. In addition, most computer training classes incorporate many tasks and insufficient time is available to commit all of the steps to memory by studying an integrated manual. Instead, a more active and motivational environment can be realized by presenting text and visuals on the computer.

c04.qxd 9/29/05 17:58 Page 102

102

Efficiency in Learning

Nevertheless, if you do provide a manual, make sure it uses a physically integrated, self-contained format that does not require the equipment to be present.

••• The Bottom Line There are solid research and psychological reasons to recommend that you: • Help learners focus their attention with cues for visuals and signals for text and auditory materials • Use cues and signals when the content is more complex and when the instruction is delivered dynamically, such as in classroom or synchronous e-learning • Minimize cognitive load from divided attention by physically integrating mutually referent content such as text and graphics on the page or screen • Minimize cognitive load imposed by note-taking during lectures by providing learners with content summaries and using class time for learning activities • Minimize cognitive load by integrating computer training on the computer so that attention is not split between a manual and the computer

Tips for Supporting Attention • Use arrows, lines, or circles to draw attention to relevant portions of complex visual displays, especially when the instruction is presented in a dynamic format. • Use bolding, italics, topic headers, organizers, and pointer phrases when writing lengthy and complex textual materials.

c04.qxd 9/29/05 17:58 Page 103

Fo c u s A t t e n t i o n a n d A v o i d S p l i t A t t e n t i o n

103

• Ask writers to follow a standard format when writing abstracts for chapters or articles. • Minimize note-taking during dynamic instruction such as lectures by providing learners with content summaries. • When taking notes during lectures is essential, be sure to signal lectures. • Keep related text and visuals closely aligned on the page or screen. • Avoid scrolling screens that separate text and visuals. • Avoid separating related text and visuals on different pages or screens. • Use lines to link text to its visual referent. • Consider the physical parameters of your display medium when planning layouts of visuals and text. • To maximize efﬁciency in computer learning, place instructional text on simulations of application screens so that the instruction is integrated on the computer.

On the CD John Sweller Video Interview Chapter 4: Focus Attention and Avoid Split Attention. John discusses the following issues related to attention: focusing attention, avoiding split attention, note-taking and attention, as well as split attention in computer training.

Sample e-Lessons The asynchronous Load-Managed Excel Web-Based Lesson illustrates effective use of cues and signals to focus attention and to minimize divided

c04.qxd 9/29/05 17:58 Page 104

104

Efficiency in Learning

attention. To view the text version, turn off the audio default option. Note the following: 1. Use of red circles to cue learners to relevant portion of the visual. 2. Use of underlining, bolding and italics in text. 3. Placement of text boxes and use of lines to link text to relevant portions of the spreadsheet. The synchronous Virtual Classroom Example illustrates effective uses of cues and signals to focus attention and to minimize divided attention during the presentation. Note: 1. Instructor use of highlighter to draw attention to relevant portions of spreadsheet. 2. Instructor ground rules to minimize split attention during presentations. 3. Placement of exercise directions in text on the same screen as the exercise. The asynchronous Overloaded Excel Web-Based Lesson fails to support attention effectively as a result of: 1. Lack of cues and signals in the graphics and text. 2. Placement of text at the bottom right-hand side of the screen with no lines or arrows to indicate links between text and visuals.

COMING NEXT In this chapter we looked at ways to help learners attend to the most important information in the lesson. In the next chapter we will look at additional strategies for reducing the amount of content presented to working memory by weeding out related but unnecessary lesson content, by concise writing, and by avoiding redundant content displays.

c04.qxd 9/29/05 17:58 Page 105

Fo c u s A t t e n t i o n a n d A v o i d S p l i t A t t e n t i o n

105

Recommended Readings Cerpa, N., Chandler, P., & Sweller, J. (1996). Some conditions under which integrated computer-based training software can facilitate learning. J. Educational Computing Research, 14(4), 344–367. Mautone, P.D., & Mayer, R.E. (2001). Signaling as a cognitive guide in multimedia learning. Journal of Educational Psychology, 93(2), 377–389. Moreno, R., & Mayer, R.E. (1999). Cognitive principles of multimedia learning: The role of modality and contiguity. Journal of Educational Psychology, 91(2), 348–368. Piolat, A., Olive, T., & Kellogg, R.T. (2005). Cognitive effort during note taking. Applied Cognitive Psychology, 19, 291-312. Tindall-Ford, S., Chandler, P., & Sweller, J. (1997) When two sensory modes are better than one. Journal of Experimental Psychology: Applied, 3(4), 247–287.

c05.qxd 9/29/05 18:19 Page 106

CHAPTER OUTLINE The Misconceptions of More Guideline 8: Pare Content Down to Essentials Write Concise Instructional Materials What Is an Effective Summary? Eliminate Related But Unnecessary Technical Content

Guideline 9: Eliminate Extraneous Visuals, Text, and Audio Omit Extraneous Words and Pictures Added for Interest Emotional vs. Cognitive Sources of Motivation Omit Extraneous Auditory Content Research Applications

Guideline 10: Eliminate Redundancy in Content Delivery Modes When Are Instructional Materials Redundant? Don’t Add Words to Self-Explanatory Visuals Research Applications Accommodating Audiences with Novices and Experts Don’t Describe Visuals with Words Presented in Both Text and Audio Narration Sequence On-Screen Text After Audio to Minimize Redundancy When to Narrate Text Research Applications Avoid Presenting Identical Words and Visuals in a Manual and on Computer in e-Learning

The Bottom Line On the CD John Sweller Video Interview Sample Excel e-Lessons

c05.qxd 9/29/05 18:19 Page 107

5 Weed Your Training to Manage Limited Working Memory Capacity

Redundant Redundant Redundant Redundant Redundant

HE GUIDELINES IN THIS CHAPTER run against common misconceptions that we help learners by providing instructional materials that include more information presented in multiple modes. By multiple modes we mean presenting words simultaneously in visual (text) and auditory (verbal) formats. In keeping with our goal of accommodating limited working memory capacity, it makes sense that instead of more instruction, we should offer less! We present research, psychological rationale and examples to support the following guidelines:

T

Guidelines for Less Is More 8. Pare content down to essentials. 9. Eliminate extraneous visuals, text, and audio. 10. Eliminate redundancy in content delivery modes.

107

c05.qxd 9/29/05 18:19 Page 108

108

Efficiency in Learning

The Misconceptions of More It’s a common misconception that the more you give learners the better. Figure 5.1 shows a typical example. On this e-learning screen from our overloaded Excel lesson on the CD, we have explained the spreadsheet example with text on the screen and also with audio narration of that same text. The rationale for delivering content in dual modes is that they offer twice the learning opportunity. In addition, many believe that two modes accommodate learners who have visual learning styles as well as learners who have auditory learning styles. Another related misconception is that we can improve motivation and learning by adding interesting information to our basic instructional materials. Unrelated themes and games are commonly

Figure 5.1. A Screen from Our Overloaded Excel e-Learning Lesson on the CD. Audio: Barb has entered her sales revenue and her overhead for last year into the spreadsheet. The first thing that Barb would like to know is how much profit she made for each month last year.

c05.qxd 9/29/05 18:19 Page 109

We e d Yo u r Tr a i n i n g

109

added to e-learning as a countermeasure for high dropout rates and to accommodate younger learners raised on computer games. An example of irrelevant information is found in Figure 5.1 in the “Did You Know” box in the lower left corner. In this chapter we review evidence that recommends a lean approach to your training materials by eliminating extraneous content in the form of pictures, words, or audio that may be related to the main topic but is not relevant to the instructional goal. We also recommend eliminating multiple modes of expression that communicate the same information, such as the text and audio shown in Figure 5.1. Mayer and Moreno (2003) recommend weeding as a solution to over-inflated instruction. Alternatively, you can avoid weeding by simply avoiding extraneous content or modes of expression in the original course design. By paring your content and delivery modes down to the essentials, you not only make learning more efficient, but you save development time as well. No need to spend extra resources on those entertaining but extraneous motivational stories or themes. No need to invest in redundant displays of content when one will do the job better. We know that working memory has a very limited capacity and we need to reserve those resources for learning. The most effective approach is to simply present the learner with the minimum required to achieve the instructional goals.

Guideline 8: Pare Content Down to Essentials Guideline 8 recommends two ways to limit your words to the minimum needed to present the content related to the learning goal: (1) write concise instructional materials and (2) eliminate or resequence technical details.

Write Concise Instructional Materials Vigorous writing is concise (Strunk & White, 2000). Technical writers have long known that succinct rendering of content is more effective than verbose versions. In the 1980s, John Carroll (1992), as a result of

c05.qxd 9/29/05 18:19 Page 110

110

Efficiency in Learning

many usability tests, designed and recommended computer manuals that he termed minimalist. In developing effective computer reference materials, he recommends that “One must deliberately not explain everything, only what must be explained. . . . Perhaps most important, when there is a problem one must not automatically add functions or build additional training modules; one must consider removing the function or documentation material associated with the problem” (p. 344). In concise writing every word contributes to the sentence and every sentence contributes to the learning goal. For example, Figures 5.2 and 5.3 show a flabby and a lean text version from our Excel load managed e-lesson on the CD. As you can see, our lean screen includes thirty fewer words than the inflated example. By ensuring that all your materials are as concise as possible, you can save instructional time and get better learning results!

Figure 5.2. A Screen with Overly Wordy Text.

c05.qxd 9/29/05 18:19 Page 111

111

We e d Yo u r Tr a i n i n g

Figure 5.3. A Concise Version of the Text in Figure 5.2 from Our Excel Load Managed Lesson on Our CD.

WHAT THE RESEARCH CONCISE MATERIALS

SAYS

ABOUT

Mayer, Bove, Bryman, Mars, and Tapangco (1995) conducted three experiments that compared learning about weather processes from summary lesson versions with learning from a complete passage. The summary version consisted of ﬁve captioned illustrations, two of which are shown in Figure 5.4. In Experiment 1, college students studied one of three lesson versions for ﬁve minutes. The versions were (1) a summary with forty-eight words and ﬁve illustrations; (2) a full passage with ﬁve hundred words and no illustrations; and (3) a combination of the full passage and the summary. The full passage alone (Version 2) resulted in signiﬁcantly less learning than either the summary alone (Version 1) or the full passage plus summary (Version 3), which were equivalent. As we discussed in

c05.qxd 9/29/05 18:19 Page 112

112

Efficiency in Learning

Figure 5.4. Two Captioned Illustrations from the Summary Lesson Version. Adapted from Mayer, Bove, Bryman, Mars, and Tapangco (1995).

Downdrafts Ice crystals Water droplets

Freezing level

Hailstones

Freezing level

Raindrops

Updrafts

Updrafts

Warm moist air 1. Warm moist air rises, water vapor 2. Raindrops and ice crystals drag condenses and forms cloud. air downward.

Chapter 3, learning is enhanced when words are illustrated with complementary explanatory visuals. In another experiment, the beneﬁts of brevity were tested by comparing summaries in which the same illustrations were described with varying numbers of words. The lean version (which was the same as that used in the previous experiment) described the illustrations with captions of ﬁfty words, the middle version added one hundred descriptive words, and the verbose version added 550 words. As you can see in Figure 5.5, the lean version led to the most learning. The research team concluded that (1) a summary that includes a combination of relevant words and pictures is better than a summary that includes either words or pictures alone and (2) a summary with words and pictures is more effective when it contains a small rather than a large amount of text. To understand the weather process, both words and pictures were needed since neither one could stand on its own. The research team cautioned, however, that “It would be incorrect to conclude that in all cases students can learn as much from studying a summary as from studying a full lesson along with a summary” (p. 72).

c05.qxd 9/29/05 18:19 Page 113

113

We e d Yo u r Tr a i n i n g

Figure 5.5. Learning Is Best from Concise Explanations of Visuals. Source: Mayer, Bove, Bryman, Mars, and Tapangco (1995).

SD = Significant Difference

5

5

Illustrations + 50 words Illustrations + 100 words

Test Score

4

Illustrations + 550 words

3

SD 2

1

The best approach is to provide just enough content, that is, words and pictures, needed to help the learner build the desired schemas. While in the weather lesson, only a few pictures and words were sufficient, other instructional goals might require more.

What Is an Effective Summary? An effective summary is one which is concise, coherent, and coordinated (Mayer, Bove, Bryman, Mars, & Tapangco, 1995). That is, it uses fewer words to present the content, the words and pictures relate to each other, and, in combination, the words and pictures are directly relevant to the instructional goal. In the above research, all learners were given the same amount of study time. However, if learners had been allowed to study until they understood the material, it is highly likely that the concise versions would have required less study time than the more verbose. Therefore more concise text not only leads to better learning but would save instructional

c05.qxd 9/29/05 18:19 Page 114

114

Efficiency in Learning

time as well. Since the most expensive element of any training program is the time learners spend away from the job, saving instructional time will directly contribute to cost benefit.

Eliminate Related But Unnecessary Technical Content It’s a common tendency of subject-matter experts to want to tell it all. As experts, they are not subject to the same cognitive load that novices will experience. Therefore, often they don’t even realize they have overloaded their instruction with way too much content for the audience and for the instructional goal. Good instructional design sorts out the need-to-know from the nice-to-know and tosses out the nice-to-know. Here we look at research that shows that omitting technical content that is not required to understand the lesson improves learning.

WHAT THE OMITTING DETAILS

RESEARCH SAYS ABOUT EXTRANEOUS TECHNIC AL

Mayer and Jackson (2004) prepared concise and expanded versions of a lesson on wave formation for delivery in both paper and multimedia formats. The concise materials used 553 words and ﬁve illustrations, whereas the expanded version included 980 words and eleven illustrations. The expanded version incorporated mathematical equations, symbols, and numerical computations relevant to three processes of ocean waves. College students were given seven minutes to study either the concise or expanded version. Following the study period, they took a test that required them to apply what they learned to answer conceptual questions such as: “Why do ocean waves move toward the shore?” In a second experiment using the paper-based materials, learners were allowed to study the materials as long as they wished beyond a minimum time period. As you can see in Figure 5.6, all concise lesson versions that omitted quantitative details, whether presented on paper or via computer animations, resulted in signiﬁcantly better learning. The authors suggested that, when learning a cause-and-effect process, learners ﬁrst form a qualitative mental model. Therefore they recommend that detailed quantitative information be deleted. Should the instructional goal require mathematical computations, sequence them after learners have built an initial qualitative understanding.

c05.qxd 9/29/05 18:19 Page 115

115

We e d Yo u r Tr a i n i n g

Figure 5.6. Learning Is Better from Concise Lessons That Omitted Quantitative Details, Regardless of Media. Source: Mayer and Jackson (2004). Concise

Max = 13

Expanded SD = Significant difference

8

7

SD SD

Test Score

5

SD

5

4

3

Paper Lessons 7 minute time limit

Paper Lessons No time limit

Multimedia Lessons

Guideline 9: Eliminate Extraneous Visuals, Text, and Audio Under Guideline 8, we recommended writing concisely. We also recommended omitting technical content not needed to achieve the instructional objective. Now we turn to extraneous content that is used primarily to add interest to lessons. Under Guideline 9 we review evidence recommending omission of words, pictures, and audio added primarily for the purpose of interest or appeal. While well intended, such adjuncts run the risk of depressing learning.

Omit Extraneous Words and Pictures Added for Interest e-Learning has been widely acknowledged for its high rate of dropout. Some suggest that younger learners weaned on computer games are especially prone

c05.qxd 9/29/05 18:19 Page 116

116

Efficiency in Learning

to drop out from multimedia instruction that is devoid of the motivational elements contained in such games. As a countermeasure, some instructional programs have adopted an approach called edutainment. In edutainment adjunct graphics, music, themes, vignettes, and games designed to make the instruction more appealing are added to the basic instructional materials. Some adjuncts are completely unrelated to the content, such as embedding technical computer training in a “Tomb Raiders” game theme augmented with high end graphics and music. Other adjuncts are related to the content but unnecessary to the instructional goal. For example, the box in the lower left corner of the screen in Figure 5.1 is an adjunct that we created to add interest to our Excel lesson. In keeping with our chapter theme of less is more, you won’t be surprised that evidence shows that adding motivational content—even content topically related to the lesson—depresses learning. The good news is that you can save the resources required to create these additions and the learner’s time invested in interacting with them. Instead of edutainment, you can invest your resources in promoting what instructional psychologists call cognitive motivation.

Emotional vs. Cognitive Sources of Motivation Related or unrelated themes, vignettes, or games added to technical materials in order to increase appeal are examples of emotional sources of motivation. These kinds of high-appeal additions are based on the theory that increasing the emotional impact of technical instruction by adding humor or interest leads to increased student engagement with the lesson and subsequently greater learning. In contrast, cognitive sources of motivation include instructional methods added to support the basic learning processes summarized in Chapter 2. Some examples include graphic organizers placed at the start of a lesson to preview the relationships among the content, an effective worked example to aid learners in building a mental model, as well as the many instructional methods we recommend throughout this book to manage cognitive load. Research we review next suggests that you omit additions designed to increase emotional motivation and invest your resources in cognitive motivational elements instead.

c05.qxd 9/29/05 18:19 Page 117

We e d Yo u r Tr a i n i n g

117

WHAT THE RESEARCH SAYS ABOUT ADDING HIGH EMOTIONAL INTEREST MATERIAL Harp and Mayer (1997) compared two versions of a short lesson on weather. The basic version contained 550 words and ﬁve captioned illustrations similar to the ones shown in Figure 5.4. Three high interest versions were created by adding 150 words, ﬁve pictures, or 150 words plus ﬁve pictures that described related but irrelevant facts about lightning. For example, one passage discussed the number of Americans killed by lightning and the fact that swimmers are highly vulnerable to lightning strikes. Another passage described what happens to airplanes that are struck by lightning during ﬂight. Yet another passage discussed a football player who was injured by lightning. Instructional scientists refer to content added to increase emotional interest as seductive details. Note that the seductive details were related to the topic but irrelevant to the instructional goal, which was to build an understanding of weather processes. College students randomly assigned to the different versions were allowed to study for as long as they wished and then completed an interest rating that asked questions about the emotional and the cognitive interest of the passages. For example, they were asked: “How interesting is this material?” to assess emotional interest and “How much does this material help you to understand the process of lightning?” to assess cognitive interest. Learners were tested with questions that required them to apply what they learned to problems such as: “What could you do to decrease the intensity of a lightning storm?” As shown in Figure 5.7, the base version lacking seductive details resulted in signiﬁcantly better learning. As you can see in Figure 5.8, learners rated the materials containing seductive details higher in emotional interest and the version that omitted seductive details as higher in cognitive interest. Harp and Mayer (1997) conclude that: “The current ﬁndings show that the best way to help students enjoy a passage is to help them understand it” (p. 100). Mayer, Heiser, and Lonn (2001) replicated these ﬁndings in a multimedia lesson in which seductive details similar to those described in the preceding paragraph were added in the form of video clips. Learners who studied lessons with seductive details learned signiﬁcantly less, with a moderate effect size of .55. Regardless of the delivery media, avoid adding related but extraneous information to spice up technical materials.

c05.qxd 9/29/05 18:19 Page 118

118

Efficiency in Learning

Figure 5.7. Learning Is Better from Lessons That Omit Seductive Details. Source: Harp and Mayer (1997).

ST = Seductive Text SI = Seductive Illustrations

4.0

Number of Solutions

SD = Significant Difference SD

3.0

2.0

1.0

Base

Base + ST

Base + Base + ST SI and SI

Figure 5.8. Ratings of Cognitive and Emotional Interest of Lessons with and Without Seductive Details. Source: Harp and Mayer (1997). Cognitive interest Emotional interest 10

Interest Ratings

SD = Significant Difference SD

8

SD

SD SD

5

4

2

Base Text

Base Illustrations

Seductive text

Seductive Illustrations

c05.qxd 9/29/05 18:19 Page 119

119

We e d Yo u r Tr a i n i n g

Omit Extraneous Auditory Content Many people prefer to study or work in the presence of background music. Likewise, music is often added to multimedia programs for the same reasons that seductive text and visuals are added—to increase lesson appeal. What is the effect of added audio either in the form of music or sounds related to the content? Here we review evidence regarding the effect of background music on learning, reading, and writing.

WHAT THE RESEARCH SAYS ABOUT QUIET LEARNING AND WORK EN V I R O N M E N T S Moreno and Mayer (2000) conducted two experiments that compared learning from multimedia lessons with and without extra audio. The base lessons used narrated animation to explain how lightning forms or how brakes work. The expanded lessons added either instrumental music, relevant environmental sounds, or both music and sounds to the base versions. None of the auditory additions obscured the narration. Following the lesson, learners were tested with questions that required them to apply their understanding of the content. The results from the lightning lesson versions are shown in Figure 5.9. In both lessons, the base version supported learning more effectively than versions with music and sound.

Figure 5.9. Learning Is Better Without Auditory Additions. Based on Data in Experiment 1 (Moreno & Mayer, 2000). No Significant Difference 3.5

Test Score

3

Narration + Narration Sounds only

No Significant Difference

Narration + Music 2

Narration + Sounds + Music 1

SD

c05.qxd 9/29/05 18:19 Page 120

120

Efficiency in Learning

Additional evidence for the negative effects of extraneous audio on learning from reading is reported by Knez and Hugge (2002). They compared learning from a seven-page text read in a quiet environment with reading of the same text in the presence of irrelevant conversational background speech. Recall of text ideas was signiﬁcantly better among those who read in a silent environment. The authors conclude that irrelevant background noise led to lower reading comprehension as a result of divided attention and/or from competition for limited working memory resources. Along similar lines, Ransdell and Gilroy (2001) found that the writing of essays in the presence of music (vocal and instrumental) took longer than when writing in a quiet environment. The researchers conclude that music can, like background speech, disrupt writing ﬂuency. To maintain quality of writing, writers slow down their production in the presence of background music. As for practical implications, the researchers suggest: “For all those college students who listen to music while they write on a computer, the advice from this study is clear. One’s writing ﬂuency is likely to be disrupted by both vocal and instrument music. A reliable reduction in ﬂuency, about ﬁfty words an hour, is likely to be caused by unattended background music. A student’s best bet would be to word process their work in silence, or at least try to reduce background sounds” (p. 147).

Research Applications Taken together, the research supporting Guidelines 8 and 9 suggest that you: • Write concisely using only the number of words needed to convey the content. • Omit nice-to-know information in technical lessons. • Start technical lessons by building a qualitative mental model before adding quantitative details. • Eliminate visual or audio materials included to add interest. • Study, read, or write in quiet environments.

c05.qxd 9/29/05 18:19 Page 121

121

We e d Yo u r Tr a i n i n g

Guideline 10: Eliminate Redundancy in Content Delivery Modes It is a common misconception that explaining content with multiple modes such as with text and audio narration of that text helps learning. One rationale for this misconception is that by using multiple modes of expression such as on-screen text and audio narration of that text you accommodate both visual and verbal learning styles. Counter to this folk wisdom, we have evidence that multiple content expressions actually overload working memory and depress learning!

When Are Instructional Materials Redundant? Redundancy in training refers to providing more expressions of content than needed for understanding. For example, in our airline safety card shown in Figure 5.10, the illustrations are clear and understandable in themselves. Adding explanatory text would be redundant. Or in a screen from our Excel overloaded example on the CD shown in Figure 5.1, the visual is explained by audio

Figure 5.10. An Airline Safety Card with Self-Explanatory Visuals.

To fasten your seat belt, insert the flat tab into the metal buckle.

To tighten, pull up on the strap.

To release the seat belt, lift the tab on the top of the metal buckle.

Please stow any carryon securely underneath the seat in front of you or in the overhead bin.

Detachable luggage wheels should not be stored in the overhead bin.

In the event of an emergency landing, brace yourself using the seat in front of you.

c05.qxd 9/29/05 18:19 Page 122

122

Efficiency in Learning

narration to leverage the modality effect. By adding on-screen text that replicates the audio narration, we have created a redundant expression of content. In Guideline 10 we review research related to redundancy as it applies to visuals alone, to narrated visuals and text, and to training of computer applications.

Don’t Add Words to Self-Explanatory Visuals A visual may be self-explanatory for one of several reasons. In some cases, such as the airplane safety card, the drawing itself is inherently self-explanatory to anyone on the aircraft. In other cases a drawing may include some brief embedded text that makes the drawing understandable. Finally, a graphic that is not inherently self-explanatory may be familiar to a specific audience with previous experience using that graphic. In all of these situations, adding words in the form of either text or audio narration is redundant and will depress learning because working memory is better off with just the minimum amount of information needed for understanding. As with all cognitive load management guidelines, redundancy is most detrimental when high demands are made on working memory, such as when the content is complex and the learners are novices.

WHAT THE RESEARCH SAYS ABOUT WORDS THAT ARE REDUNDANT TO ILLUSTRATIONS Chandler and Sweller (1991) compared learning and the instructional time needed to study three versions of a lesson on blood ﬂow through the heart. The diagram alone version consisted of a cross section of the heart with the chambers labeled and arrows illustrating the direction of blood ﬂow. The other two versions added textual descriptions such as “Blood from the upper and lower parts of the body ﬂows into the right atrium,” either embedded in the diagram or as a series of statements placed underneath the diagram. As you can see in Figure 5.11, the diagram alone version resulted in much better learning in one-half to one-third of the time of the other two versions. In this situation, the diagram with embedded

c05.qxd 9/29/05 18:19 Page 123

123

We e d Yo u r Tr a i n i n g

Figure 5.11. Learning Was Better and Faster with Self-Explanatory Diagrams Alone. Based on Experiment 5 (Chandler & Sweller, 1991).

Diagram Only Diagram + Separate Text

35 32

SD

SD = Significant Difference

150 140 130

28

Test Score

150

120 24

110 22

100

18

90

Instructional Time (Sec)

Diagram + Embedded Text

80

14

70 1

SD 50

arrows and labels was self-explanatory and adding text slowed instructional time and reduced learning. Leahy, Chandler, and Sweller (2003) contrasted learning from audio explanations of two versions of a temperature graph. One version was self-explanatory and the other was not. For the ﬁfth and sixth graders in the experiment, the basic graph shown in Figure 5.12 is not understandable and requires explanatory words. The second version, shown in Figure 5.13, rendered the graph self-explanatory by embedding short text captions. In the ﬁrst experiment the graph alone (which was not self-explanatory) was explained by words delivered either by text on the page or by audio narration. Learners were allowed to study the lesson versions for as long as they wished and were then tested with easy and with more complex graph problems. Consistent with the modality effect, the version that explained the graph with audio narration resulted in better learning than textual explanations on complex but not easy test questions.

c05.qxd 9/29/05 18:19 Page 124

124

Efficiency in Learning

Figure 5.12. A Graph of Temperature Changes Over Time for Two Days. Source: Leahy, Chandler, and Sweller (2003).

Monday Tuesday

Temperature (C) 36 34 32 30 28 26 24 22 20 18 16 9am

10am

11am

12pm

1pm

2pm

3pm

4pm

Time of Day

Figure 5.13. A Self-Explanatory Version of the Graph in Figure 5.12. Adapted from Leahy, Chandler, and Sweller (2003).

Temperature (C) 36 34 32 30 28 26 24 22 20 18 16

Monday Tuesday

3. Subtract the lower temperature from the higher temperature

2. Locate the two dots directly above the time

9am

10am

1. Select a time of day

11am

12pm

1pm

Time of Day

2pm

3pm

4pm

c05.qxd 9/29/05 18:19 Page 125

125

We e d Yo u r Tr a i n i n g

In the second experiment, learning from the self-explanatory version of the graph (Figure 5.13) alone was compared to learning from this graph with the addition of audio explanations. However, as shown in Figure 5.14, adding audio explanations to the self-explanatory version of the diagram depressed performance on more complex but not easy test questions. For the self-explanatory version of the graph, the words delivered by audio became redundant. The researchers conclude that: “If auditory explanations are used concurrently with, for example, a diagram, which contains sufﬁcient information to be understood alone, the dual-mode duplication of information is redundant and may hinder learning” (p. 414).

Figure 5.14. Adding Audio to Self-Explanatory Diagram Depresses Learning of Complex Tasks. Adapted from Leahy, Chandler, and Sweller (2003).

Diagram + Text Diagram + Text + Audio

Mean Percentage Correct

50

SD = Significant difference 40 30 20 SD 10

Easy Problems

Complex Problems

From this research, we see that audio explanations aided learning only when the tasks were more complex and only for visuals that were not selfexplanatory. Therefore, you cannot universally apply a rule such as “Explain visuals with audio narration.” Instead, you must consider the complexity of the task and the meaningfulness of the visual. Of course, the extent to which a visual is self-explanatory can depend on the experience of the learners. We discuss this issue next.

c05.qxd 9/29/05 18:19 Page 126

126

Efficiency in Learning

WHAT THE RESEARCH SAYS ABOUT NO VI CE A N D E X P E RIEN C ED L EA RN ERS Kalyuga, Chandler, & Sweller (2000) showed that a visual that was most effectively explained by words presented by audio narration in the beginning stages of learning was most effective when presented without any words during later stages of instruction. A complex diagram like the one shown in Figure 5.15 was presented to learners with explanations delivered with audio narration or without any words over three stages during which learners gained expertise with the diagram. As you can see in Figure 5.16, the version that led to best initial learning lost its effectiveness over time. By Stage 2 there was no difference in learning from the two versions, and by the ﬁnal stage, the version with the diagram alone was the better choice, both for learning and for instructional efﬁciency.

Figure 5.15. A Worked Example Using Audio to Explain How to Interpret the Diagram. Adapted from Kalyuga, Chandler, and Sweller (2000). Cutting speed for drills

Revolutions per minute (R.P.M.)

Cutting speed (m/min)

10 8

15 12

25 20

40 32

50

15 20 32 40 50 53 80 100 125 150 200 250 320 400 500 530 800 1000 1250 1500 2000 2500 3200

53 80 100 125

Material

Speed m/min

Aluminum Brass Bronze Cast Iron Copper

50-100 50-80 30-50 25-30 30-50

Assume you wish to determine the appropriate R.P.M. to drill a 25mm diameter hole in the bronze. Step 1. Select the cutting speed. From the table, select the cutting speed range for a given material, in this case, bronze. Step 2. Select the diagonal lines. At the right upper corner of the diagram, select the diagonal line that corresponds to the lowest available cutting speed without the suggested range for bronze.

8

5 5

12 10

20 15

32

50

80 125 200 320 500 800 1250

Step 3. Select the vertical line. At the bottom of the diagram, select the vertical line that corresponds to the required diameter of the hole, in this case, 25 mm.

Diameter in millimeters

Step 4. Find the intersection point. Follow the diagonal line until it intersects with the vertical line.

Help

Step 5. Select the horizontal line. Select the horizontal line that runs through the point.

25

40

53 100 150 250 400 530 1000

Done

Step 5. Read off the R.P.M. By following the horizontal line to the left, we can read off the appropriate R.P.M.

c05.qxd 9/29/05 18:19 Page 127

127

We e d Yo u r Tr a i n i n g

Figure 5.16. As Learners Gained Expertise, the Diagram Alone Led to Best Learning. Source: Kalyuga, Chandler, and Sweller (2000).

Diagram + Audio

7.0

Test Scores (max = 10)

Diagram Alone

5.0

5.0

4.0

Stage 1

Stage 2

Stage 3

Research Applications Taken together, the research supporting Guideline 10 suggests that: • Visuals may be self-explanatory because they visually incorporate all needed information; text is added to the visual; or the audience is familiar with the visual. • When visuals are self-explanatory, adding more explanations in the form of text or audio narration depresses learning. • When a visual does require an explanation, use audio (modality principle) or integrated text (to avoid split attention).

Accommodating Audiences with Novices and Experts If you have to design training for a mixed audience that includes both experienced and novice learners, we recommend violating the split-attention guideline and placing explanatory text underneath the diagram. Although it adds a split-attention processing burden to novice learners, this format will allow more experienced learners to easily bypass the text and avoid a redundancy effect. Alternatively, if using audio narration to explain an on-screen visual, provide options for the audio to be suppressed by more experienced learners.

c05.qxd 9/29/05 18:19 Page 128

128

Efficiency in Learning

Don’t Describe Visuals with Words Presented in Both Text and Audio Narration In e-learning, it’s common practice to include a combination of on-screen text and audio narration of that text to describe an on-screen visual. For example, in Figure 5.1 from our overloaded Excel lesson demonstration on the CD, a visual of a spreadsheet is explained by on-screen text and audio narration of that same text. This presentation is redundant. A similar redundancy occurs in the classroom or synchronous e-learning when the instructor projects a slide containing a visual and text and reads the text to the class. Incorporating redundant words in visual and auditory form is based on misconceptions that receiving the same message twice will enhance learning and/or that some learners benefit more from visual representations and some from auditory representations so it’s best to provide both. Both of these assumptions ignore working memory capacity limits. Evidence we summarize below consistently suggests that, when describing a graphic, a single representation of words—either in text or in audio—leads to better learning than a dual representation. In our Excel Load Managed asynchronous e-lesson on the CD, on some screens we use just audio narration to explain the content and on other screens we use just text. In contrast, in our overloaded lesson, we violate the redundancy principle by using both text and audio.

WHAT THE O MI T T I N G

RESEARCH SAYS ABOUT R ED U N DA N T TEXT

Figure 5.17 shows a complex technical graph called a Fusion Diagram used by Kalyuga, Chandler, and Sweller (1999) in an experiment that contrasted learning to interpret the diagram from explanations presented in audio, text, and audio plus text. Learners were randomly assigned to study the three different selfpaced lesson versions. Learners then rated the difﬁculty of the lesson and were tested. Reﬂecting the modality effect, the version that used audio alone to describe the diagram was the most efﬁcient. In contrast, the redundant version that combined text and audio was the least efﬁcient. Figure 5.18 illustrates the efﬁciency diagram for the three lesson versions. The research team concluded

c05.qxd 9/29/05 18:19 Page 129

129

We e d Yo u r Tr a i n i n g

Figure 5.17. Fusion Diagram Explained by On-Screen Text. Adapted from Kalyuga, Chandler, and Sweller (1999). 350 300 When cooling from the liquid state, 250 50/40 solder passes quickly 200 through the plastic state. 150

Liquid

Plastic

Plastic

Solid Solution Solid

100 58

Tin Content 8 Lead Content 100

18 98

28 88

38 78

48 58

58 58

58 48

78 38

88 28

98 100% 18 0%

Figure 5.18. Diagrams Explained by Audio Are More Efﬁcient Than Diagrams Explained by Text or by Text and Audio. Adapted from Kalyuga, Chandler, and Sweller (1999).

High Efficiency

E=0

1.0 0.8 0.5 Audio

0.4 0.2

- 1.0 - 0.8 - 0.5 - 0.4 - 0.2

0.2 0.4 - 0.2

Mental Effort 0.5

0.8

1.0

Text

- 0.4 - 0.5 - 0.8

Audio + Text

- 1.0

Low Efficiency

c05.qxd 9/29/05 18:19 Page 130

130

Efficiency in Learning

that: “It is widely believed (including among many multimedia instructional designers and educational software manufacturers) that duplicating informationally identical audio and visual material facilitates learning. The results of this experiment suggest that dual-mode presentation under such circumstances has negative rather than positive effects” (p. 352). Mayer, Heiser, and Lonn (2001) reported similar results when comparing learning from the weather lesson that described animated narrations with audio alone, or with audio plus either full or summary text. The versions that added on-screen text (in either full or summary forms) resulted in less learning. The research team calculated an effect size of .53, indicating moderate practical signiﬁcance.

Sequence On-Screen Text After Audio to Minimize Redundancy Some very recent experiments have evaluated the effects of first explaining a visual with audio narration followed by displaying the narrated words in text. By displaying the text after the narration, it is possible that a negative redundancy effect is reduced or even eliminated!

WHAT THE RESEARCH AUD I O F O L LO W E D BY

SAYS ABOUT TEXT

If on-screen text is displayed on the screen after completion of the audio narration, the negative effects of dual modes are minimized (Kalyuga, Chandler, & Sweller, 2004). A simultaneous audio and text explanation of a complex diagram like the one shown in Figure 5.17 resulted in lower efﬁciency than a nonconcurrent version in which the text was displayed after the audio had played. In both versions each step was presented one at a time and the learner could study the text for as long as he wished before moving to the next step. In a second experiment, the instructional presentation time was preset, eliminating differences in study time. As shown in Figure 5.19, under instructional pacing, the nonconcurrent version that presented text after audio resulted in better learning with a large effect size of 1.18 for instructional efﬁciency.

c05.qxd 9/29/05 18:19 Page 131

131

We e d Yo u r Tr a i n i n g

Figure 5.19. Audio Followed by Text Leads to Better Learning in Instructionally Paced Lessons. Source: Kalyuga, Chandler, and Sweller (2004). Diagram + Text Diagram + Text + Audio 7

SD = Significant difference

Test Score

5

SD

5 4 3 2 1

Learner Paced

Instructional Paced

When explaining a complex visual in asynchronous e-learning, you may want to first present the audio followed by on-screen text. When using this technique, the textual presentation provides a review opportunity for those wanting more study time but can be quickly bypassed by those not needing the additional support.

When to Narrate Text In the research discussed under Guideline 10, words were used to describe a visual that was not self-explanatory. In other words, the visual required a verbal description. The findings recommend that, when describing such visuals, you should stick with audio alone rather than audio plus identical text presented simultaneously. What do we know about the effects of narration of on-screen text when there is no other visual? For example, in the classroom a PowerPoint® slide presents a sentence of text and the instructor reads that text to the class. Alternatively, an e-learning screen displays three sentences and audio narration reads the same words.

c05.qxd 9/29/05 18:19 Page 132

132

Efficiency in Learning

WHAT THE RESEARCH NARRATING TEXT

SAYS

ABOUT

Two recent studies compared learning from content presented by text alone with content presented by text that was also narrated. As shown in Figure 5.20, the results from these two studies were opposite. Moreno and Mayer (2002) compared learning of college students from two versions of lessons on how lightning works. One version presented the words in audio alone, while the second version narrated the on-screen text. As shown in Figure 5.20 (left side), the version with text and audio resulted in greater learning on a transfer test. In contrast, Kalyuga, Chandler, and Sweller (2004) compared learning of adult trade apprentices from two lesson versions (audio alone and text plus audio) that included a discussion of around 300 to 400 words on four technical topics. As you can see in Figure 5.20 (right side), in this study the versions with audio alone resulted in better learning. What might account for these different results?

Figure 5.20. Two Experiments Comparing Learning from Audio Alone with Audio and On-Screen Text. Source: Moreno and Mayer (2002) and Kalyuga, Chandler, and Sweller (2004).

7 5 5 4 3 2

SD

Test Score: Kalyuga et al

Test Score: Mayer and Moreno

Audio + Text

25 25 24 23 22 21

SD

20

1

Short Text Pauses

Long Text no pauses

Audio alone

c05.qxd 9/29/05 18:19 Page 133

We e d Yo u r Tr a i n i n g

133

A close look shows that in the Moreno and Mayer experiments the words were presented in short segments with pauses between each segment. In contrast, in the Kalyuga, Chandler, and Sweller experiment, the words were presented in one long segment without any pauses. No doubt in the Kalyuga lessons, a lengthy text presented without pauses imposed greater load than did the shorter texts used in the Moreno and Mayer experiments. A long audio segment is difﬁcult to hold and process in working memory and even more difﬁcult to coordinate with visuals, resulting in a heavy load. We recommend that you can use audio and identical text when presenting short content segments. However, a negative effect may occur if you narrate lengthy text segments without pauses. We need additional research to guide decisions regarding narration of text in the absence of visuals.

Research Applications The research on text and audio suggests that you: • Avoid describing visuals with text and identical audio narration. • Follow audio narration with on-screen text in instructionally paced asynchronous e-learning programs. • Avoid audio narration of lengthy text passages when no visual is present.

Avoid Presenting Identical Words and Visuals in a Manual and on Computer in e-Learning In Chapter 4, we recommended that you integrate step-by-step computer instructions onto screen simulations of the application and present the instruction wholly on the computer. Often new computer application training programs will present the steps and visuals on the computer and replicate them in a manual for reference. Having exactly the same content duplicated in two media, that is, on the computer and in a manual, is a redundant expression of content and will lead to less efficient learning!

c05.qxd 9/29/05 18:19 Page 134

134

Efficiency in Learning

WHAT THE RESEARCH SAYS ABOUT REDUNDANT DELIVERY MEDIA Learning is better when words are placed on a simulation of spreadsheet screens than when words are presented in a manual (Cerpa, Chandler, & Sweller, 1996). The lesson contained partially in the manual and partially on the computer caused learners to split their attention between the manual and the computer screen. A third lesson version in the Cerpa, Chandler, and Sweller experiment displayed the words on a simulation of the application and, in addition, gave the learners a copy of all of the training screens in a manual. Thus the learners had two duplicate versions—one on the computer and a second in a manual. As you can see in Figure 5.21, for more complex tasks only, the version that incorporated everything on the computer resulted in better learning than the version that added a manual that duplicated the training screens. The self-contained computer version was also better than the split attention version in which the text was placed in the manual and the learner used the computer to apply the steps. In short, formats that led to either split attention or redundancy diminished learning in materials designed to teach computer applications. Similar results using materials to train different applications were reported by Sweller and Chandler (1994) and Chandler and Sweller (1996).

Figure 5.21. Integrated CBT Led to Better Learning of Complex Computer Skills Than Redundant or Split Attention Versions. Based on Data in Experiment 1 (Cerpa, Chandler, & Sweller, 1996). Integrated CBT

100

Manual + Software

Percent Correct

80

Integrated CBT + Manual

SD

SD = Significant Difference 50

SD

40

20

Low Complexity

High Complexity

Instructional Skills

c05.qxd 9/29/05 18:19 Page 135

135

We e d Yo u r Tr a i n i n g

An optimal format for learning computer applications is one that incorporates all instruction into the computer screens and omits either a redundant version that duplicates the screens in a manual or a split attention version that includes textual directions in the manual. In short, drop the manual.

••• The Bottom Line There are solid research and psychological reasons for recommending that you offer less rather than more in your training by: • Writing concise text to present your content • Omitting words, visuals, or audio that do not contribute to understanding • Building a qualitative mental model first in technical lessons and adding quantitative details later • Avoiding redundant words, modalities, and delivery media The benefits of a less-is-more approach can be realized in decreased time for development of instruction, reduced learning time, and better outcomes. In short, a lean approach is an efficient approach! Tips for Developing Lean Lessons • Limit words and visuals in instructional materials to just those relevant to the instructional goal. • Weed out extra words and visuals that are not essential to the learning objective. • Avoid adding related but unnecessary technical content as well as content added to stimulate emotional interest. • Present information in as few modes as needed to make it understandable. • Use diagrams alone when the diagram is self-explanatory or the learners are familiar with it.

c05.qxd 9/29/05 18:19 Page 136

136

Efficiency in Learning

• Diagrams that are not self-explanatory should be described by audio or by text, but not both. • Lengthy text segments on screens or slides should not be narrated. • Instructors should remain silent when presenting textual information to learners rather than reading that information. • Omit workbooks that duplicate on-screen information for training of computer applications. • In asynchronous e-learning, explain visuals with audio and then display text. Learners can skip the text or spend time reviewing it as needed.

On the CD John Sweller Video Interview Chapter 5: Weed Your Training to Manage Limited Working Memory Capacity. John begins by defining the redundancy effect followed by a discussion of repetition versus redundancy, redundancy in computer training, and redundancy in instructor presentations.

Sample Excel e-Lessons Compare our Before: Overloaded Excel Web-Based Lesson to the After: Load-Managed Excel Web-Based Lesson for the following elements: 1. The use of audio and text: Redundant in the Before version and not redundant in the After version. 2. Insertion of extraneous animation and content in the Before version removed from the After version. 3. All content is included in the computer lesson; no reference to an adjunct manual is made.

c05.qxd 9/29/05 18:19 Page 137

We e d Yo u r Tr a i n i n g

137

Notice that in the synchronous e-learning virtual classroom example: 1. Instructor does not read on-screen text to the class. 2. All content is included in the computer lesson. 3. The lesson content is succinct and relevant to the instructional goal. COMING NEXT In the past three chapters we have focused on ways to manage cognitive load by (1) making best use of the visual and phonological centers in working memory via the modality principle; (2) maximizing limited working memory capacity by focusing attention and avoiding split attention; and (3) streamlining lessons in ways that take a minimalist approach to instructional materials and interfaces. In the next chapter we suggest that in some situations you can manage load by providing well-designed working aids to support learning and performance.

Recommended Readings Cerpa, N., Chandler, P., & Sweller, J. (1996). Some conditions under which integrated computer-based training software can facilitate learning. Journal of Educational Computer Research, 15(4), 345–357. Harp, S.F., & Mayer, R.E. (1997). The role of interest in learning from scientific text and illustrations: On the distinction between emotional interest and cognitive interest. Journal of Educational Psychology, 89(1), 92–102. Leahy, W., Chandler, P., & Sweller, J. (2003). When auditory presentations should and should not be a component of multimedia instruction. Applied Cognitive Psychology, 17, 401–418. Sweller, J. (2005). The redundancy principle. In R.E. Mayer (Ed.), Handbook of multimedia research. Cambridge, UK: Cambridge University Press.

c06.qxd 9/29/05 18:21 Page 138

CHAPTER OUTLINE Bypassing Working Memory Guideline 11: Provide Performance Aids as External Memory Supplements What Are Performance Aids? When to Create Performance Aids Reference-Based Training When to Avoid Memory Support

Guideline 12: Design Performance Aids by Applying Cognitive Load Management Techniques For Spatial Content, Use Visuals as the Predominant Display Why Visuals Work Better When Possible, Design Self-Explanatory Visuals That Omit Text Applying the Research When Words Are Needed, Integrate Text into the Visual Integrate Performance Aids into the Performance Environment An Example of Integrated Performance Aids Applying the Research on Computer Performance Aids Use Memory Support in Training Environments Fading of Memory Support

The Bottom Line On the CD

c06.qxd 9/29/05 18:21 Page 139

6 Provide External Memory Support to Reduce Working Memory Load

A

NOTHER WAY TO REDUCE EXTRANEOUS COGNITIVE LOAD is to display content in external memory resources. External mem-

ory support is commonly implemented in the form of performance aids for use on the job, during training, or both. In this chapter we present research, psychological rationale, and examples to support the following guidelines:

Guidelines for External Memory Support 11. Provide performance aids as external memory supplements 12. Design performance aids by applying cognitive load management techniques

139

c06.qxd 9/29/05 18:21 Page 140

140

Efficiency in Learning

Bypassing Working Memory Throughout this book we have summarized instructional methods to avoid cognitive overload by accommodating working memory capacity limits. The methods include use of audio and visual modalities to maximize working memory resources, techniques to focus attention and to minimize split attention, and the elimination of redundant content and delivery modalities. Another strategy you can use to preserve limited mental resources is to reduce working memory load by providing external memory supports. Many job tasks do not require workers to perform or access knowledge from memory. Instead, they can use various reference resources to supplement their internally stored knowledge and skills. External memory supports can be used during training and after training on the job. These memory supports are commonly known as performance aids. In some cases a performance aid may replace training altogether; in other cases performance aids are used during and after the learning event. In this chapter we summarize guidelines on the design of performance aids useful for the work environment as well as the instructional environment. Compared to instructional materials, there is a paucity of research on the efficiencies of performance aids. However, many of the guidelines summarized in previous chapters apply to performance aids. Therefore this chapter offers an opportunity to review what you have read to this point in the context of performance aids.

Guideline 11: Provide Performance Aids as External Memory Supplements Under Guideline 11 we define performance aids and discuss situations that can benefit from performance support, either in the workplace or in an instructional setting.

What Are Performance Aids? Performance aids package content required for task completion in a format that is readily accessible when needed in the work and learning environment. Factual information and procedure guides are the most common types of

c06.qxd 9/29/05 18:21 Page 141

141

Pr o v i d e E x t e r n a l M e m o r y S u p p o r t

Figure 6.1. An Airline Performance Aid.

content included on performance aids. Some different types of performance aids include working aids, reference guides, wall charts, and “cheat sheets” in lesson materials. Figure 6.1 shows a working aid you have all most likely encountered—an airplane safety card. The goal of the aid is to illustrate critical procedures required of passengers. The wall chart shown in Figure 6.2 is a performance aid used in a seminar on instructional design. It summarizes five major types of content (concepts, facts, procedures, processes, and principles) and two levels of learning (remember and use) addressed in the seminar. During the class it serves as a presentation resource

Figure 6.2. A Wall Chart from an Instructional Design Class. Courtesy Clark Training & Consulting.

The Content ~ Performance Matrix Concepts

Facts

Procedures

Processes

Principles

Do It!

Solve Problems by Applying Process

Principles

USE (apply) 90% of all training needs

Classify

REMEMBER

R

e

KEYWORDS

Definition

m Facts

e Steps

m

b

Stages or Phases

e

r

Guidelines

c06.qxd 9/29/05 18:21 Page 142

142

Efficiency in Learning

when the instructor explains the content and as memory support when learners are completing exercises and tests based on the content. Figure 6.3 illustrates a memory support placed on an exercise worksheet used in our synchronous virtual classroom Excel lesson on the CD. The operators and generalized format for Excel formulas are displayed in the left-hand box as a resource for learners during demonstrations and initial practice exercises. Performance aids delivered electronically are also called electronic performance support systems (EPSS). A familiar example is the step-by-step assistance offered by “Clippy” in Microsoft applications such as Word® and PowerPoint®. In all of these examples, information that is essential to achieve a work or learning task is displayed in a format that is readily accessible to the performer while working on the task.

Figure 6.3. Memory Support Embedded in Instructional Materials. Source: Excel Virtual Classroom Lesson on the CD.

c06.qxd 9/29/05 18:21 Page 143

Pr o v i d e E x t e r n a l M e m o r y S u p p o r t

143

When to Create Performance Aids There are a number of occasions that benefit from external memory support. In some situations it is neither necessary nor practical to train skills needed to perform a task. All airlines provide safety briefings during which the safety card is pointed out. However, training of passengers on emergency procedures is not practical. First, the need for safety procedures is rare and consequently, any skills learned would be soon forgotten. Second, it’s not practical to give all passengers hands-on practice opening the emergency door. When training is neither needed nor practical and yet tasks must be performed, an effective working aid can offer a good solution. Even when training is provided, there are usually many tasks included in the training that are performed so rarely on the job that they will be forgotten. Working aids can offer a useful reminder for procedures once learned but rarely practiced after the class. Performance aids are also useful as memory support during and after training. Most organizational training courses incorporate much more content than human working memory can process in the time allotted. Very little time is available for building strong schemas in long-term memory. Commonly, many training courses offer little or no practice opportunities. At best in most situations, learners may have an opportunity to practice new skills one time. To help workers access the large amount of knowledge and skills typically transmitted but not learned, we recommend that you use performance aids to provide essential performance-related content. Facts and procedure steps are the most frequent content included on performance aids. For example, in new product training for sales staff, the product specifications can be summarized and packaged in easily accessible resources. Rather than devoting time to an extensive review of detailed

c06.qxd 9/29/05 18:21 Page 144

144

Efficiency in Learning

product specifications, limited training time can be spent helping sales staff make use of the reference resource in the context of the sales process. In computer classes, learners should have access to a performance aid with the steps needed to complete all major procedures. Give learners an opportunity to try out the main procedures during training by following the directions on the performance aids. As a result of gaining familiarity with and confidence in the support resources during training, there is a higher probability they will use them on the job.

Reference-Based Training The use of memory support in classes can lead to a confluence of reference and training. Reference materials should be designed as effective memory aids organized around the context of work tasks—not product features. Training classes should integrate reference materials by requiring learners to use them as performance aids in order to complete class assignments such as case studies. Design reference and training to complement each other and to be used together. The integration of training and performance aids is called reference-based training. Reference-based training is especially useful in courses that incorporate a large number of facts or procedures to complete tasks. Software training is a common instance. In reference-based training, lesson materials include supporting information, worked examples, and case studies. When it’s time to practice in-class exercises, learners use the reference as their guide. When reference-based training is successful, workers will rely on the reference guide in the workplace after training is completed.

When to Avoid Memory Support When effective performance aids are readily available, workers do not usually commit the content to memory during training. In that sense they will not “learn” the facts or steps in the aid. A failure to learn should have no adverse consequences for infrequently used content for which workers can rely on performance aids as reminders. However, in some situations relying on a performance aid can be counterproductive. Some tasks need to be completed quickly and accurately for safety or effectiveness.

c06.qxd 9/29/05 18:21 Page 145

Pr o v i d e E x t e r n a l M e m o r y S u p p o r t

145

Taking time to locate and study an aid could lead to serious adverse consequences. Landing an aircraft is one example. There is too much incoming data the pilot must attend in a short time frame during landing to also be referencing a working aid. The procedures required for landing must be stored in memory. Face-to-face sales situations may also suffer from too much reliance on performance aids. Better client response might result from an uninterrupted discussion of relevant product features. In work situations requiring immediate responses, the performance goal is better served by building knowledge and skill automaticity, as discussed in Chapter 9. Keep in mind the “dumbing down” potential of working aids. When workers use performance aids, in most cases they are following step-by-step directions in a rote fashion and may not build any understanding. In many cases, that is fine. It does not matter to me how my cell phone works as long as I can set it up to perform the functions I need. On the other hand, it might matter to a cell phone technician doing diagnostic and repair tasks. When the instructional goal involves tasks that require judgment, you will need to build knowledge and skills in memory based on understanding. At best, a performance aid could serve as a reminder of skills practiced in training and committed to memory. No matter when or where you plan to use external memory support resources, for best results you need to design them by applying the cognitive load management guidelines we have discussed in previous chapters. The majority of cognitive load research has been conducted in the context of learning—not performance support. In other words, in most experiments subjects are given time to study lesson materials and then are tested afterward without access to those materials. To measure performance, we need research that shows the influence of different types of performance aids on time and accuracy to complete a task while referencing the aid. Since there are few studies on performance aids, most of our guidelines related to design of performance aids will draw on research that measured learning. As more research on performance aids accumulates, we will adjust our guidelines accordingly. In the next

c06.qxd 9/29/05 18:21 Page 146

146

Efficiency in Learning

section we offer design guidelines for creating effective working and training aids.

Guideline 12: Design Performance Aids by Applying Cognitive Load Management Techniques Under Guideline 12 we show how to apply evidence-based guidelines described in previous chapters specifically to design performance aids.

For Spatial Content, Use Visuals as the Predominant Display Since text is so easy to construct, a common error is to design text-dominant working aids with few or no visuals to represent tasks involving spatial work interfaces. For example, in Figure 6.4 we show part of a working aid developed to support a computer procedure. Note that the aid is mostly text.

Figure 6.4. A Text-Dominant Working Aid.

c06.qxd 9/29/05 18:21 Page 147

147

Pr o v i d e E x t e r n a l M e m o r y S u p p o r t

Instead, when the work task involves visual interfaces such as computer screens or equipment, use a graphic representation of that interface as the predominant, and if possible, only display.

WHAT THE RESEARCH SAYS VISUALS IN PERFORMANCE

ABOUT AIDS

In Chapter 3, we summarized evidence showing that complex tasks involving spatial relationships are completed faster when tasks were illustrated with pictures rather than with words. Sixth graders were timed as they connected resistors together in simple combinations such as resistors joined end to end and in more complex combinations such as parallel connections in which three resistors must be joined in a speciﬁc sequence. As you can see in Figure 6.5, the visual aids resulted in faster performance for all tasks—but most dramatically on the more complex assembly tasks.

Figure 6.5. Visual Representations in Performance Aids Led to Faster Performance. Adapted from Experiment 1 (Marcus, Cooper, & Sweller, 1996). 180

Text instructions 160

Diagram instructions

SD

Average Time in Seconds

140

SD = Significant Difference 120 110 90 70 50

Base Text

Multiple Series Type of Resistor Connection

Parallel

c06.qxd 9/29/05 18:21 Page 148

148

Efficiency in Learning

Figure 6.6. An Inefﬁcient Performance Aid with Text Added to a Self-Explanatory Visual.

To fasten your seat belt, insert the flat tab into the metal buckle.

To tighten, pull up on the strap.

To release the seat belt, lift the tab on the top of the metal buckle.

Why Visuals Work Better Visuals are more efficient psychologically because they can be processed holistically, whereas words require serial processing. In addition, for tasks involving two- and three-dimensional interfaces, a visual provides a more direct and psychologically efficient representation. Words describing physical interfaces will require a mental transformation from text into a spatial representation (Clark & Lyons, 2004).

When Possible, Design Self-Explanatory Visuals That Omit Text In the resistor performance aids, learners were able to connect the resistors by just using visuals alone. Adding explanatory text would be redundant. For example, the airline safety card shown in Figure 6.6 is not as effective as the version shown in Figure 6.1 because it adds words to diagrams that are selfexplanatory. Adding more information than needed for understanding leads to redundancy and lower efficiency. WHAT THE RESEARCH SAYS ABOUT V I S UA L S A N D T E X T IN P ERF ORM A N C E AIDS Research by Chandler and Sweller (1991), summarized in Chapter 5, compared learning outcomes and learning time from lesson versions teaching human circulation. One version used a diagram alone that included arrows to illustrate the direction of blood ﬂow. The other version used the same diagram but added

c06.qxd 9/29/05 18:21 Page 149

149

Pr o v i d e E x t e r n a l M e m o r y S u p p o r t

explanatory text. As you can see in Figure 6.7, displaying diagrams alone led to better and faster learning.

Figure 6.7. Learning Is Better (Left Bars) and Instructional Time Shorter (Right Bars) from Diagrams Alone. Adapted from Experiment 5 (Chandler & Sweller, 1991).

32

SD

Diagram Only

160

Diagram + Embedded Text

150 140

Test Scores

SD = Significant Difference 28

130

24

120 110

22 100 18

90 80

14

SD 10

Instructional Time (Sec)

36

70 60

Applying the Research The research we have on use of visuals and text in performance aids recommends that you: • Use graphics alone when the task can be effectively communicated visually. • Use arrows or other motion cues rather than text to depict motion. • Design the graphic to reflect the visual perspective of the performer to minimize cognitive work needed to translate from one perspective to another.

When Words Are Needed, Integrate Text into the Visual Of course in many cases, a visual alone is not sufficient to adequately communicate task performance. For example, most performance aids designed to support computer tasks require screen shots with actions explained by

c06.qxd 9/29/05 18:21 Page 150

150

Efficiency in Learning

text. In Chapter 4 we summarized research by Tindall-Ford, Chandler, and Sweller (1997) showing that apprentices learned electrical test procedures more effectively from instructional materials that integrated steps into the diagram than from materials that presented steps under the diagram. Similar results were reported by Moreno and Mayer (1999) when comparing e-learning versions in which explanatory text was integrated near the visuals or placed at the bottom of the screen. In these research studies, learners did not have access to the instructional materials when they were tested. Therefore, the conclusions apply most directly to learning from instructional materials rather than performing tasks using work aids. Nevertheless, we recommend that split attention be minimized in performance aids as well as in instructional materials. We predict that performance aids with integrated text will lead to faster and more accurate task completion. Figure 6.8 shows a poorly designed performance aid for a computer procedure in which the explanatory steps are located underneath the diagram at the bottom of the page. The performer will have to invest additional cognitive effort associating related text to the appropriate location in the screen. Figure 6.9 shows an improved version in which we integrated the text steps into the screen illustration.

Integrate Performance Aids into the Performance Environment Split attention is minimized when performance aids are co-located with the work interface. For example, some computer support aids are packaged in manuals or cards. Other computer support aids integrate help into the computer and/or into the application. Cognitive load theory recommends that you integrate procedural help as closely as you can into the application. First, design the performance aid to be delivered on the computer. Second, when possible, design the performance aid to be integrated into the application.

c06.qxd 9/29/05 18:21 Page 151

Figure 6.8. A Performance Aid with Separated Text and Diagram.

Step Action 1. 2. 3.

Select the cell where you would like the quantity tobe displayed. In this example, click cell B6. Locate cell references for hourly wage and hours. In this example, the data is stored in cells B4 and B5, respectively. Enter a formula to multiply cell references. The symbol for multiply is the asterisk (*). In this example, enter =B4*B5

Figure 6.9. A Performance Aid with Integrated Text and Diagram.

c06.qxd 9/29/05 18:21 Page 152

152

Efficiency in Learning

WHAT RESEARCH SAYS ABOUT INTEGRATING PERFORMANCE SUPPORT I NT O CO M P U T E R A P P L IC ATION S In Chapter 4 we reviewed research reported by Cerpa, Chandler, and Sweller (1996) that compared learning spreadsheet procedures from text directions integrated onto a simulation of the software to learning from text directions placed in a manual. When using the manual, learners had to divide their attention between the computer and the manual. A third lesson version reproduced both the text and computer screen visuals used in version one in the manual. Some of the lesson tasks, such as how to select a range of cells or how to enter data into cells, were easy. Others, such as how to use and enter a formula, were more difﬁcult. Learners were allowed to take as much time as they wished during the study period. As you can see in Figure 6.10, learning of difﬁcult tasks (but not easy tasks) was much better from the version in which the instructions were integrated into the screen. The integrated instructions avoided the split attention caused by having to reference a separate manual. The integrated instructions also avoided redundancy caused by having a duplicate of text and visuals in the manual.

Figure 6.10. Integrated Computer-Based Training Led to Better Learning of High Complexity Skills. Source: Cerpa, Chandler, and Sweller (1996). 100

80

SD = Significant Difference SD

Integrated CBT

Percent Correct

Manual + Software Integrated CBT + Manual 60

40

SD

20

Low Complexity Cerpa et al

High Complexity

c06.qxd 9/29/05 18:21 Page 153

153

Pr o v i d e E x t e r n a l M e m o r y S u p p o r t

This experiment focused on learning. We need research to verify that performance support directions that are integrated with the interface would lead to faster and more accurate performance. However, until we have more research on best design of performance aids, based on current research, we recommend that you use an integrated approach.

An Example of Integrated Performance Aids Global Knowledge designed the two online performance aids in Figures 6.11 and 6.12. Both aids are displayed in the computer application and thus avoid split attention between a manual and a computer screen. However, the aid shown in Figure 6.11 does a better job of integrating the steps presented in text to the relevant portion of the screen. The system tracks the user’s progress through the procedure and automatically checks off steps in the instructional text. It also paints a red box that acts as a visual cue to focus the user’s attention to the proper location on the screen.

Figure 6.11. An Online Performance Aid Integrates Text with Application. With Permission from Global Knowledge.

c06.qxd 9/29/05 18:21 Page 154

154

Efficiency in Learning

Figure 6.12. An Online Performance Aid Shows Steps Out of Context of Application. With Permission from Global Knowledge.

The version shown in Figure 6.12 does not integrate text with the software application. As a result, the users must manually advance the online performance aid as they complete the task, thereby splitting attention between the performance aid and software application. Unlike Figure 6.11, it does not provide an on-screen cue to indicate the location of the current task. Instead, it provides a small screen shot of the software application. This layout forces the user to manually match the content in the performance aid with the software application; it also offers a redundant visual display, thereby increasing cognitive load. Of course it may not be practical to create customized performance support. For example, PowerPoint® is a highly flexible software tool that offers numerous features. Microsoft online support presents text instructions for various tasks co-located next to a reduced version of the running application. Figure 6.13 shows an example. In addition to cognitive load issues, designers

c06.qxd 9/29/05 18:21 Page 155

155

Pr o v i d e E x t e r n a l M e m o r y S u p p o r t

Figure 6.13. A PowerPoint Performance Aid Aligned Next to the Running Application. With permission from Microsoft.

have to factor in time, cost, and technical constraints when designing performance aids.

Applying the Research on Computer Performance Aids The research we have on display of steps to guide performance in computer applications recommends that you: • Place performance support on the computer rather than in a manual to avoid inefficiencies arising from split attention. • Integrate computer support into the application as much as possible given budget and technological constraints.

Use Memory Support in Training Environments Your instructional environment—no matter what the topic or delivery media—can also benefit from memory support. By integrating memory

c06.qxd 9/29/05 18:21 Page 156

156

Efficiency in Learning

support into your instructional materials, you can increase efficiency during learning. For example, if you are teaching how to use formulas in Excel, place a small reminder of the correct format on the page or screen where learners are required to apply that knowledge in a practice exercise, such as the example from our virtual classroom lesson on the CD shown in Figure 6.3. In this lesson, we created a small text box showing the formatting guidelines and positioned it close to the cell where the formula should be entered. When delivering training with a manual, consider placing performance aids on the same page or on a facing page as pages that display a related exercise. Alternatively, if the content will be used in many places throughout a course, display it on a wall chart or reference card allowing learners flexible access throughout the course. Although memory support on a card or wall chart will lead to divided attention between the manual and the aid, it is better than not providing any memory support or providing it on a page in the manual in a way that prevents the learner from easily viewing the aid and the exercise simultaneously.

Fading of Memory Support If your final instructional goal requires learners to complete tasks from memory, you should fade the memory support as training progresses. For example, your first performance aid might show the entire format for an Excel formula. Later, you might reduce this to a reminder to start formulas with an equal sign. By the end of the training, you might remove all memory support. The decision regarding fading of support will depend on the extent to which effective job performance relies on learned skills and the extent to which those skills need to be built during training rather than later on the job. With the exception of safety-related training, most organizational training does not allow sufficient time for workers to practice enough to acquire new skills. Instead, they rely on work experience to complete the learning process. Of course, if memory supports are unavailable or unusable in the work environment, learning a task during training to the point where it can be performed without the memory support is essential.

•••

c06.qxd 9/29/05 18:21 Page 157

Pr o v i d e E x t e r n a l M e m o r y S u p p o r t

157

The Bottom Line Although there is a dearth of controlled research on performance aids, we apply cognitive load principles derived from research on instructional materials to recommend that you: • Provide workers and learners with external sources of content in the form of performance aids. • Use visuals in performance aids. • Avoid redundancy in performance aids by including the minimum needed to convey the content. • Integrate text into the visuals, if visuals require explanations. • Display performance aids for computer applications on the computer and into the application as much as is feasible. As more research accumulates, we will modify our guidelines accordingly. At this point, we consider it highly likely that exactly the same cognitive load principles will apply to supporting performance as to learning. Tips for Bypassing Working Memory • During your analysis, work with subject-matter experts to identify large bodies of facts or procedures needed to complete tasks; these are candidates for performance aids. • Determine whether facts or procedures can be accessed on a performance aid or must be learned during training to achieve the performance objective. • If you identify a large number of facts or procedures that can be delivered on a performance aid, consider a reference-based training approach. • Design reference and training resources to work in concert with each other. • Insert memory support into your instructional materials; fade the support as learners gain expertise. • Lay out and locate performance aids in ways that minimize split attention and redundancy. • Test early prototypes of performance aids with target users to validate their effectiveness.

c06.qxd 9/29/05 18:21 Page 158

158

Efficiency in Learning

On the CD Our sample synchronous Excel Virtual Classroom Lesson contains performance aids for use during instruction: 1. Onscreen reminders of formats for formulas placed close to formula location 2. Memory support faded as the lesson progresses

COMING NEXT In Chapter 5 we recommended weeding out extraneous content not directly related to the instructional objective. Even so, you are likely to have a great deal of relevant content to convey to your learners. In the next chapter we look at design strategies you can use to spread that content out in a lesson or course to avoid imposing it on working memory all at once.

c06.qxd 9/29/05 18:21 Page 159

c07.qxd 9/29/05 19:57 Page 160

CHAPTER OUTLINE Training Design and Cognitive Load Guideline 13: Teach System Components Before Teaching the Full Process What Is Process Knowledge? How to Segment and Sequence Process Content How to Design Process Lessons

Guideline 14: Teach Supporting Knowledge Separate from Teaching Procedure Steps What Are Procedures? How to Segment and Sequence Procedure Content Applying the Research

Design Alternatives at the Course Level Guideline 15: Consider the Risks of Cognitive Overload Before Designing Whole Task Learning Environments What Is Whole Task Learning? Tradeoffs for Whole Task Instruction

Guideline 16: Give Learners Control Over Pacing and Manage Cognitive Load When Pacing Must Be Instructionally Controlled Applying the Research

The Bottom Line On the CD John Sweller Video Interview Sample Excel e-Lessons

c07.qxd 9/29/05 19:57 Page 161

7 Use Segmenting, Sequencing, and Learner Pacing to Impose Content Gradually

1 5 9 13

1

2 6 10 14

3 7 11 15

2

4 8 12 16

3

O

NE OF THE GIFTS

that instructional professionals give to their students is content that is sliced, diced, and served in ways that help learners acquire needed knowledge and skills most efﬁciently. Managing cognitive load by reducing the amount of content presented at one time is an important outcome of your training design work. Content can be organized in different sequences. For example, directive course architectures group content into small lesson segments organized in a prerequisite sequence of knowledge and skills. In contrast, whole task architectures group content around realistic work assignments. Whole task course designs impose considerable cognitive load and should be used cautiously. Learning environments that are paced by the instructor, such as classroom training, require special attention to load management. When possible, design learning environments that allow learners to control their own rate of progress. In this chapter we present research, psychological rationale, and examples to support four guidelines related to content organization and learner access to content:

Guidelines for Segmenting, Sequencing, and Pacing Learning 13. Teach system components before teaching the full process. 14. Teach supporting knowledge separate from teaching procedure steps. 15. Consider the risks of cognitive overload before designing whole task learning environments. 16. Give learners control over pacing and manage cognitive load when pacing must be instructionally controlled. 161

c07.qxd 9/29/05 19:57 Page 162

162

Efficiency in Learning

Training Design and Cognitive Load For large formal instructional design projects, some effort is normally devoted to design of an instructional plan prior to development of training materials. In large-scale projects, the design phase typically includes defining the content of the instruction through job or skill analysis, drafting instructional objectives, and creating a blueprint that specifies the outline and architecture of the course. In smaller informal projects, the design might be a simple hand-written outline of topics and activities for a class. In either setting, a major activity underlying all design efforts is segmenting and sequencing the content. In Chapter 1 we described three types of cognitive load: extraneous, intrinsic, and germane. Intrinsic cognitive load arises from the complexity of the content and associated instructional objectives. Although the intrinsic cognitive load of your content cannot be controlled, you can manage it through segmenting and sequencing of your content. In most design projects, the content is segmented into progressively smaller course sections such as units, made up of lessons that in turn are made up of topics. In this chapter we focus on managing intrinsic load by adjusting the size and grouping of those various segments in order to accommodate working memory limits. Some course architectures segment and sequence content in ways that add considerable cognitive load. You should consider the cognitive load associated with alternative architectures during the design phase of your instructional project.

Guideline 13: Teach System Components Before Teaching the Full Process Under Guideline 13 we summarize guidelines for segmenting and sequencing content associated with process knowledge. We begin with a brief definition of processes followed by evidence-based suggestions on how best to sequence process information.

c07.qxd 9/29/05 19:57 Page 163

U s e S e g m e n t i n g , S e q u e n c i n g , a n d L e a r n i n g Pa c i n g

163

What Is Process Knowledge? One of the five major content types that make up the bulk of technical training is processes (Clark, 1999). Process content depicts a flow of events that summarize the operations of business, scientific, or mechanical systems. Some examples include how performance appraisals are administered, how the circulatory system functions, and how car brakes work. Understanding process knowledge is important for workers who are responsible for one or more stages within a process or who are responsible for troubleshooting and improving process functions. Process content is a major component of training for workers focused on systems, such as repair technicians, health professionals, and business managers. For more information on teaching of process content, see Clark (1999) and Clark and Lyons (2004).

How to Segment and Sequence Process Content We have evidence for how to best segment and sequence process content based on research described below that used lessons on mechanical systems. To avoid overloading learners with the entire process presented at once, you should first teach the system components and then show how those components interact in the entire process. For example, in a process such as how car brakes work that involves the interactions of various parts, you should first present the names, locations, and functions of each of the major parts by illustrating them in the context of the entire system or subsystem. Figure 7.1 shows the first part of a lesson on a braking system. As you can see, each major component of the system is labeled and its state changes are briefly described. Following the presentation of the individual components, teach the entire process flow showing how each component interacts with those around it to accomplish the final goal. For example, in the braking lesson, after studying the individual components shown in Figure 7.1, learners watched a narrated multimedia lesson that animated each stage in the braking process, starting with pressing the pedal and ending with slowing or stopping of the vehicle. If the lesson started with an explanation of the entire process, working memory would be overloaded—especially when the presentation is in the form of a multimedia animation. However, by first presenting each component individually, the learner is prepared to assimilate the stages of the entire process.

c07.qxd 9/29/05 19:57 Page 164

164

Efficiency in Learning

Figure 7.1. Pretraining Sheet Presenting Parts and Functions of Parts in a Car Braking Process. Source: Mayer, Mathias, and Wetzell (2002). This is the piston in the master cylinder. It can be back or forward. This is the brake fluid in the tube. It can be compressed or uncompressed.

This is the brake pedal. It can be up or down.

These are the smaller pistons in the wheel cylinder. They can be in or out.

This is the brake drum. It can be pressed against or not pressed against. This is the brake shoe. It can be pushing out or not pushing out.

WHAT THE RESEARCH SAYS ABOUT S E Q U E N C I N G P R O C E S S C O N T E N T Mayer, Mathias, and Wetzell (2002) compared learning from segmented and unsegmented versions of lessons on how a car braking system or a bicycle tire pump works. The segmented version included two stages (pretraining followed by full process), while the unsegmented version included only one stage showing the entire process. In the segmented lesson, the pretraining (ﬁrst stage of the lesson) presented the individual components of the system, including their names and their operating states. The pretraining was presented either on paper, as shown in Figure 7.1, or in multimedia, as illustrated in Figure 7.2. In the multimedia version, when the learner clicked on any one of the components, that component was cued with a circle and explained with short text that named the component and summarized its state changes. When the learner clicked on a “show me” button, she could see an animation of the component making its state changes. The second stage of the lesson showed an animation of the full braking process described with audio narration. College students were randomly assigned to study the segmented lesson version that

c07.qxd 9/29/05 19:57 Page 165

U s e S e g m e n t i n g , S e q u e n c i n g , a n d L e a r n i n g Pa c i n g

165

Figure 7.2. One Frame from Multimedia Pretraining Presenting Parts and Functions of Parts in a Car Braking Process. Adapted from Mayer, Mathias, and Wetzell (2002).

This is the piston in the master cylinder. It can be back or forward.

Back to Front Page

Show Me

included the pretraining topic followed by the full process or the unsegmented lesson that presented only the full process. Experiments 1 and 2 used lessons teaching the braking process. The segmented version presented pretraining on paper (shown in Figure 7.1) in Experiment 1 and in multimedia (shown in Figure 7.2) in Experiment 2. As you can see in Figure 7.3 (Experiment 1 and Experiment 2), both versions that segmented the content resulted in signiﬁcantly better understanding of the braking process. The effect sizes were at or greater than 1.0, indicating high practical signiﬁcance! You might note that in these two experiments, learners assigned to the segmented version had more instructional time than those using the unsegmented lessons. The segmented lessons ﬁrst presented the pretraining followed by the full lesson, whereas the unsegmented lessons included only the full lesson. To rule out the possible beneﬁts of greater content exposure, Mayer, Mathias, and Wetzell (2002) conducted a third experiment in which the pretraining was presented before the main lesson in one version or after the main lesson in a comparison version. In Experiment 3, the lesson topic

c07.qxd 9/29/05 19:57 Page 166

166

Efficiency in Learning

was how a bicycle pump works and the pre- or post-training was presented using a physical model of a pump explained by an instructor rather than paper or multimedia delivery used in Experiments 1 and 2. As you can see in Figure 7.3, Experiment 3, the group that received pretraining prior to the full lesson achieved much greater understanding than either of the other groups. The effect size was 2.16, which is very high. Even though the post-training group received the same amount of information, teaching the components after presenting the entire process did not help those learners. They needed familiarity with the components ﬁrst in order to understand the full process lesson. To help learners build a good understanding of processes, teach each component ﬁrst and then show how they interact in the entire process.

Figure 7.3. Learning from Process Lessons with and Without Pretraining in Three Experiments. Based on data reported by Mayer, Mathias, and Wetzell (2002).

Pretraining

Scores on Problem-Solving Test

No Pretraining 7.0

Post training

ES = 2.16 SD

ES = .91 5.0 ES = 1.54 SD 3.0

SD

1.0

Paper Pretraining Experiment 1

Multimedia Pretraining Hands on Pre/Post Training Experiment 2 Experiment 3

c07.qxd 9/29/05 19:57 Page 167

U s e S e g m e n t i n g , S e q u e n c i n g , a n d L e a r n i n g Pa c i n g

167

How to Design Process Lessons Mayer, Mathias, and Wetzell (2002) offer the following guidelines for organizing a process lesson: 1. Decompose the system into major components. As in all segmenting decisions, you need to consider the level of granularity for each component. If you segment too small, you would end up with so many components that working memory would be overwhelmed with the sheer volume of information. If you segment too large, you might end up with components that have no clear function. 2. Work with subject-matter experts to first divide the system into subsystems and then each subsystem into meaningful components. Keep in mind that the concept of meaningfulness is relative to your instructional goal, your target audience, and the level of detailed understanding that will be needed to perform job tasks successfully. If the target audience is novice, you can use your own judgment as one input to determine appropriate component size, assuming you are also novice to the skills. Otherwise, you will need to get input from representative members of the target audience since subject-matter experts typically overestimate appropriate segment size. 3. Visually segregate and present each component in context of the whole system using the capabilities of your delivery system. For example, in computer-delivered training you can use highlighting to cue the specific text to label each component as well as a simple animation to illustrate state changes for each component. Don’t feel, however, that you need to use animation to effectively teach processes. Research on other mechanical systems has shown that a series of still visuals showing state changes can be as effective as an animation (Hegarty, Narayanan, & Freitas, 2002).

c07.qxd 9/29/05 19:57 Page 168

168

Efficiency in Learning

Guideline 14: Teach Supporting Knowledge Separate from Teaching Procedure Steps Under Guideline 13, we provided guidelines for segmenting and sequencing process content. Here we will look at alternatives for organization of procedure lessons.

What Are Procedures? Almost all technical training includes a number of procedures. A procedure is knowledge underpinning performance of a task that is completed more or less the same way each time. For example, when you access your email you will generally follow the same sequence of steps. Procedures are also called recurrent skills or near transfer tasks. All of these terms emphasize the routine nature of procedures in which the same steps are applied in a similar way each time the procedure is performed. Procedures are made up of action steps as well as supporting information, which might include the goal of the procedure, the rationale for the procedure or for steps within the procedure, as well as concepts and facts associated with the steps. For example, consider the following step from an electrical test procedure: “The appliance is switched to ‘on’ to allow current generated from the megger meter to flow into each circuit. Make sure the appliance switch is ‘on.’ ” The second sentence, which we have italicized, communicates the step in this procedure. Depending on the prior knowledge of the learner, some knowledge associated with this step includes current flow and megger meter. In addition, the first sentence provides the rationale for the step. The step with the rationale and its supporting knowledge are all needed to ensure meaningful learning of the procedure. Imagine, however, that the procedure includes a large number of steps and/or steps that incorporate multiple concepts or components. Presenting the steps at the same time as the supporting knowledge will overload working memory. Therefore, you should teach the steps separately from some or all of the supporting knowledge.

c07.qxd 9/29/05 19:57 Page 169

U s e S e g m e n t i n g , S e q u e n c i n g , a n d L e a r n i n g Pa c i n g

169

How to Segment and Sequence Procedure Content There are two alternatives for segmenting lessons based on procedures. In one approach you begin by teaching only the steps. After an opportunity to practice the steps, present the steps again, this time incorporating all of the supporting knowledge. This approach has the advantage of separating the complex information into two major segments and actively engaging learners in hands-on activities right away. It has the disadvantage of risking a loss of understanding during Phase 1 when the learners are performing steps in a more or less rote fashion. Alternatively, you can first teach and practice the major supporting knowledge topics followed by the procedure steps (Clark, 1999). Like the first approach, this segmenting scheme has the advantage of separating complex information into at least two major instructional segments. It can also build understanding early in the lesson but has the disadvantage of deferring hands-on practice to later in the lesson. Although we don’t have evidence now supporting one approach over the other, our guess is that either sequence could be effective. The underlying goal of both alternatives is to segment and separate out instructional content in ways that minimize the amount of new information working memory must accommodate at one time.

WHAT THE RESEARCH SAYS ABOUT S EQ U E N CI N G P ROC EDURA L C ON TEN T Pollock, Chandler, and Sweller (2002) compared learning from segmented and unsegmented procedural lessons. The research team compared how ﬁrst-year industrial students learned electrical test procedures from segmented and unsegmented lesson versions. The segmented version trained steps only in Phase 1 followed by the steps plus supporting knowledge in Phase 2. The unsegmented version gave the learners two study opportunities with a lesson that integrated steps and supporting knowledge. Figure 7.4 shows a sample page from Phase 1 of a segmented lesson that displays just the steps. Figure 7.5 shows a version that includes supporting knowledge as well as the steps.

c07.qxd 9/29/05 19:57 Page 170

170

Efficiency in Learning

Figure 7.4. The Segmented Lesson Version Teaching an Insulation Resistance Test. Adapted from Pollock, Chandler, and Sweller (2002).

Kettle

OFF ON

Switch Frame of Appliance Other Lead

neutral Plug of Appliance

active Earth Lead

earth Earth

M

PRESS TO TEST

Line

METER

1000 V 500 V 240 V

TEST BUTTON

Test 2: The insulation resistance between the electrical element and the frame. 1. Set the meter to read 500 V. 2. Make sure the appliance’s switch is “on.” 3. Place the earth lead on the active pin of the appliance’s plug. 4. Place the other lead on the frame of the appliance. 5. Press the test button. 6. Read the resistance from the meter. The required result is a reading of at least one M . 7. Remove the earth lead from the active pin and place it on the neutral pin. 8. Press the test button again. 9. Read the resistance. A reading of at least one M is again required.

c07.qxd 9/29/05 19:57 Page 171

U s e S e g m e n t i n g , S e q u e n c i n g , a n d L e a r n i n g Pa c i n g

Figure 7.5. The Unsegmented Lesson Version Teaching an Insulation Resistance Test. Adapted from Pollock, Chandler, and Sweller (2002).

Kettle

OFF ON

Switch Frame of Appliance Other Lead

neutral Plug of Appliance

active Earth Lead

earth Earth

M

PRESS TO TEST

Line

METER

1000 V 500 V 240 V

TEST BUTTON

This test aims to check that there is no “leak” in the flow of current from either the active or neutral wires of the appliance to the frame of the appliance, which would cause the frame of the appliance to become electrically active. To achieve this, two circuits are set up: the first including the frame of the appliance and the active of the appliance and a second including the frame of the appliance and the neutral of the appliance. Very large measures of insulation resistance (ideally “infinite” resistance” indicates that there are no leaks in the flow of current to each circuit. A megger meter measures the resistance of the circuit. 1. By setting the meter to read 500 volts, a larger than normal voltage will be introduced into each circuit in order to test the appliance under a heavy load. Set the meter to read 500 volts. 2. The appliance is switched “on” to allow current generated from the megger meter to flow into each circuit. Make sure the switch is “on.” 3. To measure the resistance of the first circuit, the megger meter is connected into the circuit. Place the line lead on the frame of the appliance. 4. Place the earth lead on the active pin of the appliance’s plug. 5. The test button of the megger meter introduced 500 volts into the circuit. Press the test button.

171

c07.qxd 9/29/05 19:57 Page 172

172

Efficiency in Learning

After studying their assigned lesson (segmented or unsegmented) in Phase 1, all participants rated the difﬁculty of their instructional materials and were tested with a written and a practical test. The test included both easy items involving identiﬁcation of parts and more complex items that required understanding and application of the procedure. During the second phase conducted forty-eight hours later, all learners studied the unsegmented version. The efﬁciency diagram shown in Figure 7.6 shows that participants who studied the segmented lessons achieved much higher efﬁciency scores on complex test items than those who worked with unsegmented materials (full lesson). For low complexity questions, there was no difference between the two groups. In a second experiment that was similar in all respects except that the learners were more advanced, there were no differences in outcomes. From this study we can conclude that when novice learners must master complex procedural content, teaching the steps separately from the supporting knowledge leads to more efﬁcient learning. The research team concluded: “When dealing with very complex information, in order to allow novice students to process interacting elements, the intrinsic cognitive load of the material should initially be artiﬁcially reduced. This allows serial rather than simultaneous processing of information, thus reducing working memory load” (p. 83). Note that in this study, the segmented version presented steps ﬁrst followed by steps plus supporting knowledge. We need further research to verify whether a reverse sequence, that is, supporting knowledge ﬁrst followed by steps plus supporting knowledge would be as efﬁcient.

c07.qxd 9/29/05 19:57 Page 173

173

U s e S e g m e n t i n g , S e q u e n c i n g , a n d L e a r n i n g Pa c i n g

Figure 7.6. Segmented Lessons Are More Efﬁcient for Learning Complex Content. Based on data from written test results in Experiment 1 (Pollock, Chandler, & Sweller, 2002).

High Efficiency

Segmented Lesson Phase 2

Performance 1.0 .82 0.8

E=0

0.6 Segmented Lesson Phase 1

.24 0.4

Mental Effort

0.2 - 1.0 - 0.8 - 0.6 - 0.4 - 0.2

0.2 0.4 - 0.2 - 0.4 -.38 - 0.6 -.68 - 0.8

0.6

0.8

1.0

Full Lesson Phase 2 Full Lesson Phase 1

- 1.0

Low Efficiency

Applying the Research Taken together the research on segmenting content tells us that: • Learning is more efficient when supporting knowledge such as facts or concepts is taught separately from main lesson content. • Teaching of process stages should be preceded by teaching the names and functions of components in the process. • Teaching of task steps should be segmented from teaching of supporting knowledge such as the reasons for the steps and/or concepts associated with the steps. Next we consider design decisions at the whole course level with emphasis on the cognitive load implications of different course architectures.

c07.qxd 9/29/05 19:57 Page 174

174

Efficiency in Learning

Design Alternatives at the Course Level Consider the following four instructional architectures as alternative course design frameworks: receptive, directive, guided discovery, and exploratory (Clark, 2003). A brief summary of each of these architectures is shown in Table 7.1.

Table 7.1. Four Architectures of Instruction. Adapted from Clark (2003).

Architecture

Description

Examples

Receptive

Instructional environments that primarily transmit information to learners. The main goal is to promote awareness or provide a repository of information.

A lengthy classroom lecture An informational website A brieﬁng

Directive

Instructional environments that build new knowledge and skills in a bottom-up fashion primarily using instructional methods of rule, example, and practice. They primarily emphasize a deductive approach to learning.

Most math classes A step-by-step class to teach a new computer application

Guided Discovery (Whole Task)

Instructional environments that require the learner to build knowledge and skills through experience with a real-world task or set of examples. They primarily emphasize an inductive approach to learning.

A bank loan class that teaches guidelines as learners assess a loan applicant A presentation skills class that begins with a brieﬁng assignment

Exploratory

Instructional environments that emphasize a high level of learner control over content and content sequence.

An e-learning class in which learners can select various links to learn about topics The Internet

c07.qxd 9/29/05 19:57 Page 175

U s e S e g m e n t i n g , S e q u e n c i n g , a n d L e a r n i n g Pa c i n g

175

Most relevant to our discussion are directive and guided discovery architectures. The directive architecture is characterized by a number of short lessons that start with basic prerequisite knowledge and skills and build hierarchically to more complex skills. The directive architecture includes lessons that present small content segments followed by examples and practice with feedback. In directive learning, lessons explicitly present the knowledge and skills followed by practice that requires learners to rehearse the content. The menu of lessons shown in Figure 7.7 from our asynchronous Excel e-lesson on the CD reflects a typical directive course structure. This course includes a number of short lessons that build on each other. The early lessons such as Excel basics, entering and editing data, and using formulas lay the foundation for more advanced spreadsheet functions sequenced later in the course. Each lesson provides definitions, examples, and demonstrations, followed by skill practice that typically requires learners to apply new skills to realistic work tasks or scenarios.

Figure 7.7. A Typical Directive Course Architecture.

c07.qxd 9/29/05 19:57 Page 176

176

Efficiency in Learning

In contrast, guided discovery architectures emphasize inductive learning by providing learning environments designed to help learners derive new schemas through experience with concrete case studies and examples. Although directive architectures also use case studies and examples, these are sequenced after an explanation or demonstration of skills. In contrast, in guided discovery the learning of new skills is designed to occur in the context of solving a case study, which is usually the starting point of the lesson. As Jonassen (1999) states: “The fundamental difference is that the problem drives the learning, rather than acting as an example of the concepts and principles previously taught. Students learn domain content in order to solve the problem, rather than solving the problem as an application of learning” (p. 218). Since learners must invest much cognitive effort to build relevant schemas at the same time that they are engaged in completing learning task assignments, the guided discovery architectures can impose greater load than the directive architectures.

Guideline 15: Consider the Risks of Cognitive Overload Before Designing Whole Task Learning Environments Under Guideline 15 we focus on a type of guided discovery instruction called whole task learning. We will define whole task courses and discuss issues you should consider from a cognitive load perspective before selecting a whole task design.

What Is Whole Task Learning? Whole task learning is “a student-centered pedagogical strategy that poses significant, contextualized, real-world, ill-structured situations while providing resources, guidance, instruction, and opportunities for reflection to learners as they develop content knowledge and problem-solving skills” (Hoffman & Ritchie, 1997, p. 97). A central feature of whole task learning is that it uses a real-world task assignment as the starting point for learning. In contrast, directive architectures typically teach small skills out of context and often

c07.qxd 9/29/05 19:57 Page 177

U s e S e g m e n t i n g , S e q u e n c i n g , a n d L e a r n i n g Pa c i n g

177

Figure 7.8. A Virtual Ofﬁce Setting for Bank Loan Whole Task Course. Source: Moody’s Financial Services.

present realistic case studies at the end of a learning unit. Whole task learning appears in various forms and has been called problem-based learning, goal-based scenarios, competency-based learning, and cognitive apprenticeships (Collins, Brown, & Newman, 1989; Schank, Berman, & MacPerson, 1999; van Merriënboer, Kirschner, & Kester, 2003). Figure 7.8 shows the office setting for a whole task multimedia course on commercial loan analysis for bank loan agents. The learner plays the role of a bank loan agent and is assigned loan applicants to research and analyze in order to recommend funding to the loan committee. Some of the resources available to the learner through the office interface are a bookshelf with articles on the applicant’s industry, a FAX and telephone for obtaining credit reports and checking on references, as well as an opportunity to conduct a virtual client interview. As data relevant to the loan application are gathered, they are stored in the file cabinet in the lower left hand side of the

c07.qxd 9/29/05 19:57 Page 178

178

Efficiency in Learning

desk for reference at any point during the lesson. Guidance is offered with a process worksheet that leads the learner through various stages and poses questions to consider throughout the loan research and evaluation process. In addition, an on-screen agent offers assistance, including links to structured tutorials. You can find additional examples of whole task learning formats described in Clark (2003); Hummel and Nadolski (2001); and Jonassen, (2004). You do not need multimedia to deliver whole task learning courses. However, some reports suggest that in situations such as medical diagnostic training in which sounds and appearances offer important case information, multimedia offers a better delivery vehicle than a paper medium (Hoffman & Ritchie, 1997; Kamin, O’Sullivan, Deterding, & Younger, 2004). In many medical schools a form of whole task learning called problem-based learning (PBL) begins with a case presentation (in multimedia or in a handout). A small learning team discusses the case with emphasis on defining the main learning issues to be researched and resolved. A period of self-study research is followed by a continued team discussion of the case during which the new knowledge acquired during self-study is reviewed and applied to the case resolution. Figure 7.9 contrasts a traditional directive course architecture with a whole task plan for a course on use of synchronous software for classroom instructors migrating to a virtual classroom environment. As you can see, the directive architecture is organized around the various features of the software,

Figure 7.9. A Directive and Whole Task Outline for a Course on Use of Virtual Classroom Software. Directive Outline I. Using the white board II. Using polling III. Using direct messaging IV. Using application sharing V. Using breakout rooms VI. Teaching a lesson

Whole Task Outline 1. Problem 1 – Preparing a briefing 2. Problem 2 – Teaching an office skill 3. Problem 3 – Teaching computer procedures 4. Problem 4 – Teaching a sales process

c07.qxd 9/29/05 19:57 Page 179

U s e S e g m e n t i n g , S e q u e n c i n g , a n d L e a r n i n g Pa c i n g

179

whereas the whole task version organizes content around teaching tasks. In the whole task version, the components of the virtual classroom software will need to be integrated into each task-focused lesson and acquired while performing the assigned task.

Tradeoffs for Whole Task Instruction Although there is much enthusiasm for whole task learning, we have almost no controlled experiments demonstrating its effectiveness. The majority of research reports focus on medical education, where PBL has been widely adopted. Among these reports, however, are few controlled experimental studies involving random assignment of learners. Therefore, we have no definitive evidence comparing learning from PBL with learning from traditional directive instruction. In a review of PBL, Hmelo-Silver (2004) reports that PBL students scored slightly lower than medical students who took a traditional curriculum on standard multiple-choice tests, but performed slightly better on tasks related to clinical problem solving. Thus there is no clear evidence that PBL offers highly significant learning advantages. As important as learning, however, are measures of learner satisfaction. PBL medical students report substantially more positive attitudes than students engaged in traditional courses. However, Hmelo-Silver (2004) cautions that “It is important to note that in medical schools, the students are a fairly select group and the PBL curricula are well established. Moreover, PBL is used throughout the entire curriculum” (p. 259). It would be a mistake to assume that the positive reactions of medical students to whole task learning will apply to your learners, who may not adapt well to the psychological requirements of whole task instruction. Learning in a whole task course can be similar to the expression: “We’re trying to sail the boat at the same time that we are building it.” Very early in the course learners encounter a realistic task assignment. While they are completing that task assignment, they are expected to be acquiring a new set of knowledge and skills related to the assignment. For most novice learners the result will be cognitive overload. In many cases the boat is likely to sink before it is built.

c07.qxd 9/29/05 19:57 Page 180

180

Efficiency in Learning

Van Merriënboer (1997) and van Merriënboer, Kirschner, and Kester (2003) have proposed a detailed design model for whole task courses that includes strategies to manage cognitive load, including faded worked examples and guidelines for segmenting-sequencing content. While these procedures are effective, it should be recognized that whole task learning that does not divide a task into manageable components is likely to overwhelm working memory. For now, we suggest that you consider whole task course designs only for learners with considerable relevant prior experience because these learners are likely to have already learned many of the components. As we have discussed throughout this book, novices are most subject to cognitive overload. It is possible that whole task designs will be effective for moderate to high prior knowledge learners. Again, we will need more evidence to verify this assertion and to compare learning, satisfaction, and efficiency outcomes from whole task architectures with traditional directive courses.

Guideline 16: Give Learners Control Over Pacing and Manage Cognitive Load When Pacing Must Be Instructionally Controlled Throughout this chapter we have summarized the negative effects of overloading learners with too much content at one time. We have recommended that you consider ways to segment and sequence information so that learners can process the content presented at any point in the training. Another strategy to minimize the amount of content learners must absorb at once is to offer learners the opportunity to assimilate new content at their own pace through learner control. In some learning environments learner control over pacing is not feasible. For example, narrated multimedia animations, instructor-led classes, or video presentations typically are delivered at a pace controlled by the instruction. Learners have no way to control the rate at which they must assimilate the content. Such instructionally controlled environments most likely impose greater cognitive load than learner-controlled environments, such as during self-study of a manual where the text and visuals can be reviewed at the reader’s discretion.

c07.qxd 9/29/05 19:57 Page 181

U s e S e g m e n t i n g , S e q u e n c i n g , a n d L e a r n i n g Pa c i n g

181

The type of learner control used is critical. Learner control over what and when to study sounds seductively appealing, but for novices it can impose an impossibly high cognitive load. By definition, novices do not know the content of an instructional area. Asking them to determine, for example, the order in which units will be studied may be asking them to make decisions that they are not yet capable of making in any sensible fashion. Such decisions should be left to someone who does have knowledge of the content—you, the instructional professional. That is why you should generally avoid an exploratory architecture for novice learners. On the other hand, decisions on how rapidly an instructional sequence should proceed can best be made by the learner. There are two implications of learner control for efficient instruction. One is to design e-learning in a way that allows learners as much control as possible over the rate of content display. To see the effect of learner versus instructional control, compare our two Excel asynchronous demonstration lessons on the CD. The overloaded version includes several screen sequences in which content is presented under instructional control. In contrast, the load managed version provides navigation controls that allow the learners to adjust their pace through the instruction. In situations in which learner control is not feasible, pay special attention to the various cognitive load management techniques we have presented throughout the book. For example, apply the modality principle by explaining animations with words in audio rather than on-screen text or both audio and on-screen text, keep the content lean by omitting extraneous information, and provide options for learners to replay instructional sequences.

WHAT THE RESEARCH SAYS ABOUT L EA R N ER CO N TROL OF PA C IN G Three experimental studies give evidence for the beneﬁts of learner control over pacing.

EXPERIMENT 1: LESSON ON WAVE FORMATIONS In Chapter 5 we summarized a study by Mayer and Jackson (2004) showing that a more concise lesson on how waves are formed was better than a more

c07.qxd 9/29/05 19:57 Page 182

182

Efficiency in Learning

Figure 7.10. Learning from Concise Training Is Better in Paper Version That Is Learner Controlled. Adapted from Mayer and Jackson (2004).

Max = 13 8

Test Score

7 6

SD = Significant Difference Paper SD Multimedia

5 4 3

elaborated version that incorporated quantitative details. The concise version led to more learning, whether presented on paper or via narrated animations in multimedia. As a part of this study, Mayer and Jackson (2004) compared learning from the concise training delivered on paper with concise training delivered in multimedia using narrated animations. As you can see in Figure 7.10, the paper versions resulted in signiﬁcantly more learning. The researchers suggested that when studying materials delivered in paper media, learners have the freedom to interact with the instruction at their own pace and thus experience less cognitive load. In other words, less load is imposed in media such as print that is learner controlled. Therefore it may not be wise to expend resources to convert still diagrams into animations, since animations typically run under instructional control. There is additional evidence that, in many situations, static visuals teach as effectively (if not

c07.qxd 9/29/05 19:57 Page 183

U s e S e g m e n t i n g , S e q u e n c i n g , a n d L e a r n i n g Pa c i n g

183

better) than dynamic visuals (Clark & Lyons, 2004). Given the potential for high cognitive load imposed by animated visuals, we look forward to research guidelines as to how and when to most effectively use them.

EXPERIMENT 2: LESSON ON WEATHER PROCESSES Mayer and Chandler (2001) compared the same lesson on weather processes described in Chapter 5 presented either with or without learner control. The learner-controlled version presented a ten-second segment of narrated information and paused until the learner clicked the screen to move to the next segment. In contrast, the instructionally controlled version played the entire three-minute segment continuously from start to ﬁnish. As you can see in Figure 7.11, the learner-paced version resulted in more learning than the instructionally controlled version.

Figure 7.11. Learning Is Better from Learner-Paced Multimedia Training. Based on Experiment 2 (Mayer & Chandler, 2001).

7

Test Score

6 5

SD

SD = Significant Difference

4 3 2 1 Learner Paced

Instructional Paced

c07.qxd 9/29/05 19:57 Page 184

184

Efficiency in Learning

EXPERIMENT 3: LESSON ON INTERPRETING TECHNICAL DIAGRAMS In Chapter 5, we summarized evidence for increased cognitive load in instructionally controlled environments (Kalyuga, Chandler, & Sweller, 2004). They compared two versions of a lesson—a redundant version that explained a diagram with text and audio with a nonredundant version that ﬁrst delivered audio alone to explain the diagram and then added the text. When the lessons were learnerpaced, the redundant version did not result in any negative effects. However, when the lessons were instructionally paced, the redundant version signiﬁcantly reduced learning.

Applying the Research Taken together, these three studies suggest that cognitive load is greater in instructionally controlled learning environments and therefore you should: • Make learner control the default pacing option in multimedia instruction. • Consider whether a series of still diagrams might be as effective as and less costly than an animation. • Consider whether a text with still visuals might be as effective and less costly than a video version. • Invest special effort to manage cognitive load in environments that are inherently instructionally controlled, such as synchronous e-learning or classroom training.

••• The Bottom Line There are solid research and psychological reasons for recommending that you: • Segment information in ways that avoid presenting all lesson content at once.

c07.qxd 9/29/05 19:57 Page 185

U s e S e g m e n t i n g , S e q u e n c i n g , a n d L e a r n i n g Pa c i n g

185

• Consider the cognitive load overhead of whole task learning environments. • Allow learners to control their pacing in multimedia instruction. • Apply other cognitive load management techniques in instructionally paced environments.

Tips for Imposing Content Gradually • Work with target learners as well as subject-matter experts to deﬁne optimal content segment sizes and sequences. • Separate supporting knowledge from full content presentations when teaching process or procedure lessons. • Avoid using whole task course designs unless your target learner population has relevant skills and knowledge. • Allow learners to control their pace in asynchronous e-learning. • Manage cognitive load in instructionally controlled environments such as classrooms or synchronous e-learning by pausing frequently for practice or question opportunities; avoiding the redundancy caused by reading projected slides; and keeping lessons as streamlined as possible through concise explanations that avoid extraneous information. • Consider a series of still diagrams rather than an animated display. • Consider a lesson delivered in text with still visuals rather than a video version.

On the CD John Sweller Video Interview Chapter 7: Using Segmenting, Sequencing, and Learner Pacing to Impose Content Gradually. John discusses ways to manage high intrinsic cognitive load through segmenting and learner control.

c07.qxd 9/29/05 19:57 Page 186

186

Efficiency in Learning

Sample Excel e-Lessons See the following in our sample lessons: 1. Compare our Before: Overloaded Excel Web-Based Lesson to the After: Load-Managed Excel Web-Based Lesson for instructional control of pacing in the overloaded version changed to learner control in the load-managed version. 2. Notice in the After: Load-Managed Excel Web-Based Lesson the use of directive lesson design and segmenting of content into formulas for addition and subtraction, division and multiplication, and combination formulas using parenthesis. 3. Notice in the Virtual Classroom Example the use of directive lesson design; the format of the formula is explained prior to the procedure of how to enter the formula into the spreadsheet; and the instructor use of load management techniques such as highlighting, avoidance of split attention, use of on screen memory support, and instructor pauses to compensate for instructional control of pacing.

COMING NEXT In this chapter we have focused on ways to manage the intrinsic load of your training programs through design decisions involving segmenting and sequencing of your content. In the next chapter we will consider how best to transition from examples to practice in order to impose mental work gradually as a lesson or training program progresses and learners gain expertise.

Recommended Readings Hmelo-Silver, C.E. (2004). Problem-based learning: What and how do students learn? Educational Psychology Review, 16(3), 235–266.

c07.qxd 9/29/05 19:57 Page 187

U s e S e g m e n t i n g , S e q u e n c i n g , a n d L e a r n i n g Pa c i n g

187

Mayer, R.E., Mathias, A., & Wetzell, K. (2002). Fostering understanding of multimedia messages through pre-training: Evidence for a two-stage theory of mental model construction. Journal of Experimental Psychology: Applied, 8(3), 147–154. Pollock, E., Chandler, P., & Sweller, J. (2002). Assimilating complex information. Learning and Instruction, 12, 61–86. Van Merriënboer, J.J.G., & Sweller, J. (2005). Cognitive load theory and complex learning: Recent development and future directions. Educational Psychology Review, 17(2), 147–176. Van Merriënboer, J.J.G., Kirschner, P.A., & Kester, L. (2003). Taking the load off a learner’s mind: Instructional design for complex learning. Educational Psychologist, 38(1), 5–13.

c08.qxd 9/29/05 19:58 Page 188

CHAPTER OUTLINE Does Practice Make Perfect? What Is a Worked Example?

Guideline 17: Replace Some Practice Problems with Worked Examples Acceleration of Expertise in the Classroom The Psychology of Replacing Practice with Worked Examples

Guideline 18: Use Completion Examples to Promote Learning Processing What Are Completion Examples? The Psychology of Completion Examples

Guideline 19: Transition from Worked Examples to Problem Assignments with Backwards Fading What Is Backwards Fading? The Psychology of Fading Applying the Research

Guideline 20: Display Worked Examples and Completion Problems in Ways That Minimize Extraneous Cognitive Load Applying the Modality and Split Attention Principles to Worked Examples How to Display Worked Examples

The Bottom Line On the CD John Sweller Video Interview Sample Excel e-Lessons

c08.qxd 9/29/05 19:58 Page 189

8 10

Transition from Worked Examples to Practice to Impose Mental Work Gradually

5 1 Ex

pe

r t is

e

I

T’S CONVENTIONAL WISDOM

that the best way to build skills is practice—and lots of it! But practice is expensive. First, it’s costly to design effective practice exercises, and second, completing practice assignments absorbs a great deal of the learner’s instructional time. Psychologically, working many practice exercises imposes mental load on working memory that is counterproductive to initial learning. In this chapter we offer a proven method to make your training more effective and more efﬁcient by using a series of worked and completion examples that require learners to gradually ﬁll in more and more steps until they are assigned full practice problems. We present research, psychological rationale, and examples to support the following guidelines:

Guidelines for Transitioning from Worked Examples to Practice 17. Replace some practice problems with worked examples. 18. Use completion examples to promote learner processing of examples. 19. Transition from worked examples to problem assignments with backwards fading. 20. Display worked examples and completion examples in ways that minimize extraneous cognitive load.

189

c08.qxd 9/29/05 19:58 Page 190

190

Efficiency in Learning

Does Practice Make Perfect? Can you recall the hours you spent in math courses working practice problems both in class and afterward as homework assignments? Practice exercises matched to the learning objectives are one hallmark of effective courseware. For example, in an early draft of our demonstration lesson on Excel formulas, we assigned learners six different scenarios for practicing constructing formulas. Effective practice exercises that require learners to process content deeply also impose considerable load on working memory. For novice learners who need to devote working memory capacity to building new schemas, working many practice problems slows learning by overloading working memory. The good news is that recent research offers proven techniques to achieve the same learning in less time by starting with worked examples that transition gradually into practice exercises. You will see in our CD demonstration asynchronous load managed e-lesson that we replaced many of our practice assignments with worked examples!

What Is a Worked Example? A worked example is a step-by-step demonstration of how to perform a task or solve a problem. For example, Figure 8.1 shows one screen with the beginning of a worked example from our asynchronous Excel formula demonstration lesson on the CD. For procedures, worked examples typically take the form of a demonstration that illustrates the steps to complete the task. Worked examples may be presented in diverse formats and modalities, including text in books such as the algebra worked example shown in Figure 8.2 or in a narrated animation such as the Excel worked example shown in Figure 8.1. Examples are nothing new, since good instruction has always used them. However, what is new is the proven effectiveness of examples to replace practice and get equivalent learning results in less time and with less learner effort! In this chapter we help you apply four guidelines for design and display of worked examples in ways that will accelerate expertise. You will see how worked examples should be modified as a lesson progresses and your learners gain expertise.

c08.qxd 9/29/05 19:58 Page 191

Tr a n s i t i o n f r o m Wo r k e d E x a m p l e s t o P r a c t i c e

191

Figure 8.1. Part of a Worked Example from Asynchronous e-Lesson on Constructing Formulas in Excel from the CD.

Figure 8.2. An Algebra Worked Example Displayed in Text. (a+b)/c = d Solve for a 1. a + b = dc 2. a = dc - b

Guideline 17: Replace Some Practice Problems with Worked Examples Suppose your lesson ends with eight practice problems that require the learner to apply the lesson guidelines, such as analyze a loan application or construct a control chart. Your practice problems require learners to apply the lesson skills or guidelines to diverse work situations. For example, among the loan

c08.qxd 9/29/05 19:58 Page 192

192

Efficiency in Learning

applicant problems, some scenarios involve applicants with a previous bankruptcy, while others include diverse income levels. To apply Guideline 17, instead of assigning eight practice problems, you would instead create four worked examples and four practice exercises. We call these versions worked example-problem pair lessons. The lesson alternates worked examples with a similar practice problem. Thus a worked example involving a bankruptcy case would be followed by a practice problem with a different bankruptcy scenario.

WHAT THE RESEARCH SAYS ABOUT REPL ACING PRACTICE WITH WORKED EX AMPLES Sweller and Cooper (1985) were the ﬁrst to show that replacing some practice problems with worked examples could save time and still result in the same or even more learning as requiring learners to work all problems as practice exercises. They used algebra problems such as the one shown in Figure 8.2. Learners were randomly assigned to an all-problem lesson version that required them to work eight problems as a practice exercise. A second group was assigned a different lesson version in which a worked example was followed by a practice exercise four times. Therefore both groups were exposed to eight problems, with the worked example group only solving four of the eight. The test required learners to solve six new problems that were similar to those used in the lessons. Sweller and Cooper (1985) measured completion time as well as number of errors made in the instruction and the test. Table 8.1 includes data from one of their Table 8.1. Worked Example Problem Pairs Result in Faster Learning and Performance. Source: Sweller and Cooper (1985).

Outcome Training Time (Sec) Training Errors Test Time Test Errors

Worked Example-Practice

All Practice

32.0 0 43.6

185.5 2.73 78.1

.18

.36

c08.qxd 9/29/05 19:58 Page 193

Tr a n s i t i o n f r o m Wo r k e d E x a m p l e s t o P r a c t i c e

193

experiments. Note that in this experiment the all-practice group took almost six times longer to complete the lesson than the worked example group! Not only were learners in the worked example group faster completing the lessons, but they were also signiﬁcantly faster completing the test problems—taking about half the time of the all-problem group. A close look at the tests showed that the all-problem group made many more errors than the worked examples group.

Acceleration of Expertise in the Classroom While the replacement of some practice with worked examples looked promising, it was important to see whether these findings applied to actual classrooms. The experimental lessons described previously were quite brief— lasting approximately forty minutes. How would worked examples fare in an actual classroom with longer lessons and different topics? Zhu and Simon (1987) report one field trial conducted in Chinese middle schools in which a traditional three-year course consisting of two years of algebra and one year of geometry were successfully completed in two years by replacing some practice with worked examples! Here we see real-world evidence of acceleration of expertise through a cognitive load reduction technique. In summary, the laboratory research showed that learning time could be saved under controlled conditions by substituting some practice with worked examples, and the field trials showed that this technique could be adapted to actual classrooms.

The Psychology of Replacing Practice with Worked Examples It comes as no surprise that actively solving practice problems imposes much more mental work than reviewing worked examples that illustrate how to solve those problems. When studying worked examples, limited working memory capacity can be devoted to building a schema of how to perform the task. Having a worked example to study just prior to solving a similar problem provides the learner with an analogy available while solving the

c08.qxd 9/29/05 19:58 Page 194

194

Efficiency in Learning

problem. When having to actively solve a problem without the benefit of an analogous example, most working memory capacity is used up in figuring out the best solution approach, with little remaining for building a schema. Paas (1992) found that worked examples were superior to straight practice for learning of statistical concepts mean, median, and mode. As part of his experiment he asked the participants to rate the amount of invested effort on a scale ranging from very, very low to very, very high. Not surprisingly, participants rated the lessons with worked examples as significantly less difficult than the lessons with all practice, which suggests that the benefit of worked examples is due to reduced cognitive load.

Guideline 18: Use Completion Examples to Promote Learning Processing Although the use of worked examples to replace some practice proved effective in a number of experiments, there is a potential drawback to worked examples. To be effective, a worked example must be studied. An example ignored won’t promote learning. Some learners may be tempted to either skip the worked examples completely or to give them a cursory review and thus miss the benefits they offer. In contrast, practice problems demand deep processing for solution. In fact, many practice assignments ask learners to “show their work” as evidence of this deep processing. Deeper processing leads to better learning. One way to minimize learners ignoring worked examples is to replace worked examples with completion examples.

What Are Completion Examples? A completion example is a hybrid between a practice assignment and a worked example. In a completion example, some of the steps are demonstrated as in a worked example and the other steps are completed by the learner as in a practice problem. Figure 8.3 shows an Excel worked example from our CD converted to a completion example. The first step is done for the learner, and the learner must complete the next step himself. In order to complete the problem, the learner will have to actively process the worked out portion and then overtly respond to the open portion.

c08.qxd 9/29/05 19:58 Page 195

195

Tr a n s i t i o n f r o m Wo r k e d E x a m p l e s t o P r a c t i c e

Figure 8.3. Part of a Completion Example from an Asynchronous Load-Managed Lesson on Excel on the CD.

WHAT THE RESEARCH SAYS CO M P L E T I O N EX A M P L ES

ABOUT

Paas (1992) compared the efﬁciency and learning effectiveness of three lesson versions: (1) all practice problems; (2) worked examples and practice pairs; and (3) completion examples and practice pairs. The lessons were designed to teach statistical concepts such as mean, median, and mode. Table 8.2 shows the test scores, learning time, and effort ratings given for the test for the three lesson versions. The learners spent signiﬁcantly more time on the lesson versions with all problems than the lessons with worked examples, yet learned signiﬁcantly less. Learning from the worked examples and completion examples was the same. The mental effort ratings for the training were not different among groups. However, the effort invested during the test was rated signiﬁcantly higher among learners who studied the allproblems lessons. The author concludes that: “Training with partly or completely

c08.qxd 9/29/05 19:58 Page 196

196

Efficiency in Learning

Table 8.2. Worked Examples and Completion Examples Are More Efﬁcient Than All-Practice Lessons. Source: Paas (1992).

Outcome

Worked Examples

Completion Examples

All Problems

Test Score (0–24) Time in Training Effort During Test (0–9)

18.8 32.3 min. 3.8

16.1 39.8 min. 3.1

12.4 42.5 min. 4.9

worked-out problems leads to less effort-demanding and better transfer performance. Moreover, time on training was shortest in the worked condition” (p. 433). Additional experiments involving lessons with worked examples and completion examples using geometry and programming tasks conﬁrmed that both worked examples and completion examples were more efﬁcient and equally effective in terms of learning outcomes than lessons that required learners to work all problems (Paas & van Merrienboer, 1994; Trafton & Reiser, 1993).

The Psychology of Completion Examples Like worked examples, completion examples reduce cognitive load because schemas can be acquired by studying the worked-out portions. Requiring the learner to finish the worked example ensures that she will process the example deeply. Research has shown that learners who process examples more deeply learn more. Chi and her colleagues (1989) asked college students to talk aloud as they studied worked examples of physics problems. They compared the “self-talk” of students who scored higher with those who scored lower on a test. They found that the high scoring learners had not only more self-explanations but that the quality of their self-explanations was better. The more effective self-explanations: • Focused on when and why physics solution equations were used • Connected specific solution steps to the problem • Included more self-monitoring such as: “I don’t see why this example used this equation at this stage in the solution”

c08.qxd 9/29/05 19:58 Page 197

Tr a n s i t i o n f r o m Wo r k e d E x a m p l e s t o P r a c t i c e

197

In summary, a completion example offers psychological balance. It reduces cognitive load by incorporating some worked-out elements and it fosters deep processing by requiring completion of the remaining elements.

Guideline 19: Transition from Worked Examples to Problem Assignments with Backwards Fading So far we have seen that you can increase the efficiency of your lessons by replacing some practice with either worked examples or completion examples. This technique manages cognitive load of novice learners by freeing up working memory capacity to build new schemas by study of examples. However, as learners gain expertise during their training, eventually worked examples actually become detrimental and learners are better off with lessons in which they work all the problems. This is because once a learner has acquired a basic schema for the skill or concept, he learns best by applying the schema to problems, rather than investing redundant effort in studying more worked examples. The best way to accommodate learners as they build expertise is through a process called backwards fading of worked examples. In lessons that apply backwards fading, worked examples transition gradually into practice problems coincident with a gain in learner proficiency. Fading techniques allow you to accommodate a gradual learning process. Initially learners should devote as much working memory capacity as possible to building a schema that will underpin the new skills. As they gain expertise, understanding is promoted as they do more and more work. In more advanced stages of learning, learners need to build high competency levels by solving many practice problems themselves.

What Is Backwards Fading? Fading is a process in which completion examples evolve into full problem assignments by a gradual increase in the number of steps that must be completed by the learner. The lesson begins with a full worked example that provides a model for the learner. The next few worked examples are completion examples in which more and more of the work is done by the learner. By the end of the lesson, full problems are assigned.

c08.qxd 9/29/05 19:58 Page 198

198

Efficiency in Learning

Figure 8.4 shows a faded completion problem from our asynchronous Excel lesson on the CD. In this example the instruction fills in the first step and the learner completes the remaining three steps. A faded worked example sequenced just prior to this one would demonstrate how to complete the first two steps and then ask the learner to complete the last two steps. In backwards fading, the instruction fills in the first steps and the learner finishes the problem starting from the end. Thus in a five-step problem, the first completion problem shows Steps 1 through 4 worked out and the learner fills in Step 5. In the next completion example, Steps 1 through 3 are worked out and the learner fills in Steps 4 and 5. At the end of the sequence, the learner is solving a full problem. Figure 8.5 conceptually illustrates backwards fading. As you can see, a sequence of examples (circles) requires greater and greater learner input (white portion of circle) to complete. In this way, instructional support is faded gradually as the learner gains expertise.

Figure 8.4. A Faded Worked Example from the Asynchronous Excel Lesson on the CD.

c08.qxd 9/29/05 19:58 Page 199

199

Tr a n s i t i o n f r o m Wo r k e d E x a m p l e s t o P r a c t i c e

Figure 8.5. A Conceptual Model of Backwards Faded Completion Examples. Transitioning from Worked Examples to Problem Assignments = Worked in Lesson = Worked by the Learner

Step 1 Step 2 Step 3

Step 1 Step 2 Step 3

Step 1 Step 2 Step 3

Step 1 Step 2 Step 3

Worked Example

Completion Example 1

Completion Example 2

Assigned Problem

Lesson Start

Lesson End

WHAT THE RESEARCH SAYS ABOUT WORKED EX AMPLES AND LEARNER E X P E R I E N C E Under Guideline 19, we discuss two research ﬁndings related to the best use of worked examples for novice and for more experienced learners. First, we summarize studies showing that worked examples best serve the needs of novice learners. As the learners gain expertise during training, their learning is best promoted by actively working problems themselves. Second, we review experiments that recommend progressive backwards fading as a way to gradually transition from worked examples to full problem assignments.

WHEN PRACTICE IS BETTER THAN WORKED EXAMPLES Kalyuga and his colleagues (2001) demonstrated the beneﬁts of transitioning from worked examples to full problem assignments as learners gain expertise. They

c08.qxd 9/29/05 19:58 Page 200

200

Efficiency in Learning

constructed lessons to teach trade apprentices how to write equations for relay circuits of gradually increasing levels of complexity. Figure 8.6 illustrates the design of the experiment. All apprentices received two learning opportunities: experimental lessons and training lessons. Experimental lessons compared two versions—one with worked examples and paired problems and the other with all assigned problems. Training sessions used paired worked examples and problems for all participants. The study was conducted in two stages during which the participants gained expertise. At the start of the ﬁrst stage, learners studied one of the two experimental lessons and then made ratings of lesson difﬁculty and completed a performance test. After the test, all participants had a training session. One week later a second training session was followed by a second experimental session, another rating, and a test. Two experiments designed like this were conducted with the same group of learners for a total of four stages. By the second experiment, the learners had gained much more expertise. This allowed a comparison of the effectiveness of worked example problem pairs with all assigned problem lessons over time as learners gained expertise.

Figure 8.6. The Plan of an Experiment Comparing Worked ExamplesPractice Pairs to All Practice as Learners Gain Expertise. Training Session 1

Experimental Session 1

Training Session 2

Experimental Session 2

Worked Examples

Worked Examples Test

Worked Examples

Worked Examples

All Problems

Test All Problems

Stage One

1 WEEK

Stage Two

c08.qxd 9/29/05 19:58 Page 201

201

Tr a n s i t i o n f r o m Wo r k e d E x a m p l e s t o P r a c t i c e

As you can see in Figure 8.7, the worked example-problem pair lessons resulted in more efficient learning during initial stages of learning. This result replicates the beneficial effect of worked examples problem pairs that we summarized previously. However, as the learners gained experience, the allproblem lessons became more efficient. That is, they yielded equivalent or better learning with less effort. The authors conclude that: “Differing levels of learner experience should be taken into account when designing instruction using worked examples and problems. A series of worked examples should be presented first. After learners become more familiar with the domain, problem solving (as well as exploratory learning environments) can be used to further enhance and extend acquired skills” (p. 588). Instructional methods that manage load effectively for novices are no longer needed once learners gain more expertise. We discuss the changing needs of experts in greater detail in Chapter 10.

Figure 8.7. Worked Examples Are More Efﬁcient for Novices; All Problems Are More Efﬁcient for Experts. Based on data in Tables 8.1 and 8.2 (Kalyuga, Chandler, & Sweller, 2001).

1

Efficiency

Worked ExampleProblem Pairs

As learners gain expertise, worked examples depressed learning

0

All Problems -1 1

2

3 Stage Experience

4

c08.qxd 9/29/05 19:58 Page 202

202

Efficiency in Learning

TRANSITIONING FROM WORKED EXAMPLES TO PRACTICE VIA FADING Atkinson, Renkl, and Merrill (2003) compared the learning effectiveness of worked example-problem pairs with completion examples using progressive backwards fading. College students were randomly assigned to lessons on statistical probability that either used worked example—problem pairs or completion examples with backwards fading. Figure 8.8 shows a sample completion problem from this study in which the ﬁrst two steps are worked out and the learner is asked to complete the third step. The test scores showed that the lessons with progressive backwards fading completion problems resulted in better learning than the example-problem pair lessons, with a rather low effect size of .23 for near transfer test items and .27 for far transfer.

Figure 8.8. A Sample Completion Problem with Backwards Fading. A Completion Problem with Backwards Fading Problem Text: The bulb of Mrs. Dark’s dining room lamp is defective. Mrs. Dark had 6 spare bulbs on hand. However, 3 of them are also defective. What is the probability that Mrs. Dark first replaces the original defective dining room bulb with another defective bulb before then replacing it with a functioning one? First Solution Step: Total number of spare bulbs: 6 Number of defective spare bulbs: 3 Probability of a defective spare bulb first: 3/6 = 1/2 = .5 Second Solution Step: Total number of spare bulbs after a first replacement trial: 5 Probability of a functioning bulb second: 3/5 = .6 Third Solution Step: Total number of spare bulbs after a second replacement trial: Probability of a functioning bulb third:

c08.qxd 9/29/05 19:58 Page 203

Tr a n s i t i o n f r o m Wo r k e d E x a m p l e s t o P r a c t i c e

203

The Psychology of Fading The research reviewed in the preceding paragraphs suggests that novice learners who are most susceptible to cognitive overload will benefit the most from worked examples. As learners gain expertise, worked examples can actually depress learning because it requires more mental effort for an experienced learner to study a worked example than to simply work a problem herself. This is because a learner who has built a schema for working problems can best solidify his or her schema by exercising it. Having to study a worked example, once a schema has been established, becomes a redundant mental activity and learning is disrupted. We discuss the adjustments you need to make as learners gain expertise in Chapter 10.

Applying the Research Taken together, the experiments on worked examples recommend that you: • Begin your lessons with worked examples followed by a similar problem. • Follow the first worked example with a series of completion examples in which the first steps are worked out and the final steps are left for the learner to complete. • End the lesson with full practice assignments. This transition from worked example to practice by way of backwards faded completion examples provides learners with a smooth trajectory from novice to expert.

Guideline 20: Display Worked Examples and Completion Problems in Ways That Minimize Extraneous Cognitive Load We have seen that worked example-problem pairs and completion examples are more efficient vehicles for learning than assigning all problems. However, a poorly formatted worked example can add extraneous cognitive load and thus negate its potential benefits. To be effective, worked examples should be formatted in ways that minimize cognitive load.

c08.qxd 9/29/05 19:58 Page 204

204

Efficiency in Learning

In Chapters 3 and 4 we discussed the use of audio (modality effect) and use of integrated text formats that minimize split attention. The modality effect applies to multimedia or classroom instruction and recommends that you describe on-screen visuals with words presented in audio narration rather than with text. Presenting content in complementary visual and auditory modes distributes the load between the visual and auditory storage centers of working memory. The use of integrated text applies to either multimedia or paper-based materials and recommends that text that describes visuals be physically integrated nearby the visual. This reduces visual search by doing the integration work for the learner. These principles apply to worked examples that use visual illustrations.

Applying the Modality and Split Attention Principles to Worked Examples Figure 8.9 shows a screen capture from our demonstration Excel virtual classroom e-lesson on the CD. In this segment of the lesson, the instructor

Figure 8.9. Instructor Explains the Demonstration Verbally in Virtual Classroom Source: See complete lesson on CD.

c08.qxd 9/29/05 19:58 Page 205

Tr a n s i t i o n f r o m Wo r k e d E x a m p l e s t o P r a c t i c e

205

describes how to enter a correct formula while demonstrating the procedure in the application. The use of audio that describes the actions demonstrated in the application sharing window minimizes load for novice learners. If the same procedure is being presented in a paper medium, words must be presented in text. Figures 8.10 and 8.11 contrast an effective and ineffective way to explain such a worked example. The efficient example (Figure 8.10) integrates the steps into a visual that uses the two-page spread of the book to show the application screens. The inefficient example (Figure 8.11) explains screens with words in text placed under the visual that require the learner to expend extra mental effort to do the integration himself.

Figure 8.10. A Print-Based Worked Example That Integrates Text into Diagram to Minimize Split Attention. Source: Mark Palmer, from Clark and Lyons (2004).

c08.qxd 9/29/05 19:58 Page 206

206

Efficiency in Learning

Figure 8.11. A Print-Based Worked Example That Splits Attention Between Diagram and Related Text.

Step Action 1. 2. 3.

4. 5.

Select the cell where you would like the quantity to be displayed. In this example, click cell B6. Locate cell references for hourly wage and hours. In this example, the data is stored in cells B4 and B5, respectively. Enter a formula to multiply cell references. The symbol for multiply is the asterisk (*). In this example, enter =B4*B5 Press Enter to accept the formula. End of Procedure.

WHAT THE RESEARCH SAYS ABOUT FORMATTING OF WORKED EX AMPLES Tindall-Ford, Chandler, and Sweller (1997) compared learning from three versions of worked examples that illustrated how to conduct common electrical tests on appliances. Examples of the materials used are shown in Chapter 4 (Figures 4.5 and 4.7). One version placed step-by-step text under the kettle diagram while the second version integrated the same text into the kettle diagram. A third version described the kettle diagram with audio narration. As you can see in Figure 8.12, both lesson versions that minimized cognitive overload (integrated text and audio versions), improved learning outcomes.

c08.qxd 9/29/05 19:58 Page 207

207

Tr a n s i t i o n f r o m Wo r k e d E x a m p l e s t o P r a c t i c e

Figure 8.12. Examples Described with Integrated Text or Audio Improve Learning. Source: Tindall-Ford, Chandler, and Sweller (1997).

Score on Transfer Test Max = 12

12 Significantly better than separated text

9

6

3

Separate Text

Audio

Integrated text

How to Display Worked Examples Audio narration is more transient than text. When using audio to describe a visual, we recommend that some form of visual cueing be used to draw the eye to the portion of the visual being described by the audio. Jeung, Chandler, and Sweller (1997) found that when the visual component of the instruction is complex and it is explained by audio, it was only effective when a cueing device such as color, an arrow, or subtle animation was used. For example, in our asynchronous Excel lesson on the CD we used a red circle to draw attention to the portion of the spreadsheet being explained. In the virtual classroom lesson on the CD, the instructor used a highlighter to focus attention to the relevant portion of the spreadsheet. Second, when you are using faded completion examples, present the example with integrated text rather than audio so the learners can refer to it as they finish it. You will note in our asynchronous CD demonstration lesson

c08.qxd 9/29/05 19:58 Page 208

208

Efficiency in Learning

that our full worked examples are described with audio narration, whereas our completion examples are presented with on-screen text.

••• The Bottom Line There are solid research and psychological reasons for recommending that you: • Help novice learners build robust schemas by pairing worked examples with practice assignments. • Ensure that learners study worked examples by using completion examples. • Transition gradually from worked examples to full practice as learners gain expertise. • Use backwards faded completion examples to transition from worked examples to practice. • Format worked examples in ways that manage cognitive load in multimedia through audio narration of steps and cueing of related visuals and in print media through integration of text nearby the visual. • Format completion examples with text that is integrated into the visual to avoid split attention.

Tips on Using Worked Examples to Impose Mental Work Gradually • If workers need to practice a skill during training to build proﬁciency, use backwards faded worked examples. • When working with subject-matter experts (SMEs) during course design, ask them to show you examples of how to complete the task in a step-by-step

c08.qxd 9/29/05 19:58 Page 209

Tr a n s i t i o n f r o m Wo r k e d E x a m p l e s t o P r a c t i c e

209

manner. Prepare for your session ahead by asking them to identify and bring task samples with them to your interview session. • As the SME describes how to complete the task, use audio recorders and cameras to document the steps and the visual interface(s). • If the task involves spatial content, such as use of computer screens or equipment, minimize cognitive load in examples either by integrating words into the diagram or by using audio to present the words. • Use audio to explain full worked examples and integrated text to explain faded completion examples. • Because audio is transient, always provide a replay button so learners can repeat the example. • If your task involves spatial content, design your screen or page real estate to accommodate the graphics and integrated text. Research shows that multiple visuals involved in a procedure can be effectively displayed in paper formats using the entire page spread, as shown in Figure 8.10.

On the CD John Sweller Video Interview Chapter 8: Transition from Worked Examples to Practice to Impose Mental Work Gradually. John describes the best ways to use worked examples in training followed by a description of how examples should be altered as learners gain expertise as well as how to display worked examples effectively.

Sample Excel e-Lessons In the Load-Managed Excel Asynchronous Web-Based Lesson note the following: 1. Each topic begins with a worked example and progresses to completion examples and then to full practice assignments.

c08.qxd 9/29/05 19:58 Page 210

210

Efficiency in Learning

2. Full worked examples are described with audio narration; completion examples and practice are described with on-screen text. 3. Full worked examples described with audio narration are cued with red circles to help learners see the relevant portions of the spreadsheet as they are described. 4. On-screen text explaining examples is aligned close to the relevant portions of the spreadsheet. In the Virtual Classroom Example note the following: 1. Worked examples are explained by instructor narration. 2. The instructor engages all learners in completion examples by ask them to respond in the chat window. 3. Practice problem directions remain in text on the screen. COMING NEXT In Part II of the book, we have summarized a number of proven techniques for reducing extraneous cognitive load in your instructional and performance support materials. Reducing extraneous cognitive load increases instructional efficiency so learning can be faster. But the real reason to reduce extraneous load is to free working memory capacity for more effortful mental processing that results in better learning. Cognitive load that results from instructional strategies that augment learning is called germane load. Part III of the book includes Chapter 9, in which we continue our story about worked examples as we look at ways to use worked examples to increase germane load and thus improve learning. It also discusses germane load that arises from rehearsals that lead to automaticity of new schemas.

Recommended Readings Atkinson, R.K., Derry, S.H., Renkl, A., & Wrotham, D. (2000). Learning from examples: Instructional principles from the worked examples research. Review of Educational Research, 70(2) 181–214.

c08.qxd 9/29/05 19:58 Page 211

Tr a n s i t i o n f r o m Wo r k e d E x a m p l e s t o P r a c t i c e

211

Atkinson, R.K., Renkl, A., & Merrill, M.M. (2003). Transition from studying examples to solving problems: Effects of self-explanation prompts and fading worked out steps. Journal of Educational Psychology, 95(4), 774–783. Clark, R.C., & Lyons, C. (2004). Graphics for learning. San Francisco: CA: Pfeiffer. Tindall-Ford, S., Chandler, P., & Sweller, J. (1997). When two sensory modes are better than one. Journal of Experimental Psychology: Applied, 3(4), 257–287.

c08.qxd 9/29/05 19:58 Page 212

c09.qxd 9/29/05 18:27 Page 213

PA R T T H R E E

Instructional Guidelines for Imposing Relevant Cognitive Load

I

N PART III WE INCLUDE only one chapter on proven techniques you can use to increase germane cognitive load. The reason for reducing extraneous load and managing intrinsic load is to free up working memory capacity for processing that leads to learning. Instructional strategies that lead to learning require mental work, which imposes germane cognitive load. As research evolves, additional methods may be found that impose germane cognitive load. However, based on evidence currently available, in this part of the book we review four main methods that promote learning

213

c09.qxd 9/29/05 18:27 Page 214

214

Efficiency in Learning

and transfer of learning. These include: the use of diverse worked examples to build more flexible mental models that support transfer of learning; encouraging learner explanations of worked examples in order to process them deeply; rehearsals that lead to automation of new knowledge and skills; and mental rehearsal of complex content after initial mental models are formed.

Read Chapter 9. Put Working Memory to Work with Germane Load

To Find Out How to Support transfer of learning with diverse worked examples Promote learner self-explanations of worked examples Help learners automate new skills Promote mental rehearsal after initial formation of mental models

c09.qxd 9/29/05 18:27 Page 215

Imposing Relevant Cognitive Load

215

On the CD Video Interview with John Sweller: Chapter Preview/Review Chapter 9. Put Working Memory to Work with Germane Load. Overview of germane load, use of variable worked examples, selfexplanations, and mental rehearsals

Sample Excel e-Lessons After: Load-Managed Excel Web-Based Lesson. This asynchronous e-learning sample imposes germane cognitive load through varied context worked examples and practice exercises. Virtual Classroom Example. This synchronous e-learning sample imposes germane cognitive load through varied context completion examples and practice exercises.

c09.qxd 9/29/05 18:27 Page 216

CHAPTER OUTLINE Shifting from Extraneous to Relevant (Germane) Load Guideline 21: Use Diverse Worked Examples to Foster Transfer of Learning Near vs. Far Transfer of Learning Getting Beneath the Cover Story How to Build Flexible Schemas That Support Far Transfer What Are Varied Context Examples? The Psychology of Varied Context Worked Examples Applying the Research

Guideline 22: Help Learners Exploit Examples Through Self-Explanations What Are Self-Explanations? What Is a Good Self-Explanation? Applying the Research Self Versus Peer Explanations

Guideline 23: Help Learners Automate New Knowledge and Skills What Is Automaticity How Do Skills Become Automatic? When to Build Automatic Skills Applying the Research

Guideline 24: Promote Mental Rehearsal of Complex Content After Mental Models Are Formed What Is Mental Rehearsal? When to Use Mental Rehearsal Applying the Research

The Bottom Line On the CD John Sweller Video Interview Sample Excel e-Lessons

c09.qxd 9/29/05 18:27 Page 217

9 Put Working Memory to Work with Germane Load

I

N PART II WE FOCUSED ON WAYS

to minimize unproductive sources of cognitive load called extraneous load. In this chapter we shift our attention to a form of cognitive load that leads to learning called germane load. Your instruction should minimize extraneous cognitive load to free working memory for germane load. We present research, the psychological rationale, and examples to support the following guidelines for promoting germane cognitive load:

Guidelines for Imposing Germane Cognitive Load 21. Use diverse worked examples to foster transfer of learning. 22. Help learners exploit examples through self-explanations. 23. Help learners automate new knowledge and skills. 24. Promote mental rehearsal of complex content after mental models are formed.

217

c09.qxd 9/29/05 18:27 Page 218

218

Efficiency in Learning

Shifting from Extraneous to Relevant (Germane) Load In Part II we focused on ways you can keep cognitive load low. Why is elimination of extraneous mental load so important? Recall from Chapter 2 that learning involves the formation of new schemas that are stored in long-term memory. The difference between an expert and a novice in any domain is the expert’s larger number and better organization of schemas. Learners form new schemas in working memory by integrating incoming information with prior knowledge. After the learner has built a new schema, he still must invest a lot of mental effort when initially using it. All of these activities require working memory resources. By minimizing extraneous sources of cognitive load, you free up working memory for learning. Instruction has two main goals. The first goal is to help learners form new schemas. The second goal is to help them automate those new schemas in ways that lead to efficient and effective job performance. Both of these activities impose a load on working memory that we call germane load. By reducing extraneous load, we free up working memory capacity for germane load. In this chapter we review four guidelines that will add load in the service of fostering learning. The first two guidelines help learners build schemas, and the second two guidelines help them automate new schemas.

Guideline 21: Use Diverse Worked Examples to Foster Transfer of Learning At this point we pick up the story of worked examples from where we left off in the last chapter. In Chapter 8 we saw that, for novice learners, studying worked examples–practice pairs resulted in more efficient learning than working practice problems. We also saw that completion examples, partial worked examples that the learner finishes, encourage learners to actively study the worked examples. In fact, an ideal approach is to transition from full worked

c09.qxd 9/29/05 18:27 Page 219

P u t Wo r k i n g M e m o r y t o Wo r k w i t h G e r m a n e L o a d

219

examples through a series of increasingly faded completion examples to full practice problems. This gradual shift of mental work takes advantage of the growth of expertise as the learner progresses through the instructional program.

Near vs. Far Transfer of Learning How might you solve this problem?

Suppose you are a doctor faced with a patient who has a malignant tumor in his stomach. It is impossible to operate on the patient, but unless the tumor is destroyed, the patient will die. There is a kind of ray that at a sufficiently high intensity can destroy the tumor. Unfortunately, at this intensity, the healthy tissue that the rays pass through on the way to the tumor will also be destroyed. At lower intensities, the rays are harmless to healthy tissue but will not affect the tumor either. How can the rays be used to destroy the tumor without injuring the healthy tissue? Jot down a few ideas of how you might tackle the tumor problem. Then read the fortress story in Figure 9.1. Does the fortress story hold a clue you can use to solve the tumor problem? If you had read the fortress story first, would it have helped you solve the tumor problem? The solution to the tumor problem is to use weak rays from multiple locations around the body but all passing through the site of the tumor. The common principle linking these two stories is a convergence principle. Both fortress and tumor problems are solved by aligning many small components to converge onto a central target. Being able to apply a principle learned in one setting to a different setting is called far transfer of learning. Would you be surprised that in a research study only 30 percent of individuals who read the fortress story were able to apply the convergence principle to the tumor problem (Gick & Holyoak, 1980)? The context of the tumor problem is so different from that of the fortress problem that most

c09.qxd 9/29/05 18:27 Page 220

220

Efficiency in Learning

Figure 9.1. The Fortress Story. Source: Gick and Holyoak (1980).

The Fortress Invasion A small country was ruled from a strong fortress by a king. The fortress was situated in the middle of the country surrounded by farms and villages. Many roads radiated outward from the fortress like spokes on a wheel. A rebel general vowed to capture the fortress. He gathered his army at the head of one of the roads. However, the general learned that the king had planted mines on each of the roads. The mines were set so that small bodies of men could pass over them safely since the king needed to move his troops and workers to and from the fortress. But any large force would detonate the mines. It seemed impossible, therefore, to mount a full-scale attack on the fortress. The general divided his army into small groups and dispatched each group down a different road so that the entire army arrived at the fortress together.

people simply do not see the relationship. As a result, most people do not abstract the convergence principle from the fortress story. The transfer failure demonstrated in the tumor problem experiment is a common occurrence. Haskell (2001) begins his book on transfer with the discouraging observation: “Despite the importance of transfer of learning, research findings over the past nine decades clearly show that as individuals, and as educational institutions, we have failed to achieve transfer of learning on any significant level” (p. xiii). Transfer failure means you may have built some skills during the training. The learners may even score high on your test. But later, back on the job, they are unable to apply what they’ve learned. In the training world, we can distinguish between instructional goals that are relatively near transfer and those that are more far transfer. Near transfer tasks are activities that are completed more or less the same way each time. Accessing your email and processing a routine customer order are two common examples. Near transfer tasks are based on procedures that are done the same way each time they are performed. In contrast, far transfer tasks require the performer to adapt her skills to each new situation. Selling a new product or designing a web page are two examples. Because no two sales or websites are identical, the worker must adapt her skills. Of course, near and far transfer tasks are not really two separate categories but represent a

c09.qxd 9/29/05 18:27 Page 221

P u t Wo r k i n g M e m o r y t o Wo r k w i t h G e r m a n e L o a d

221

continuum. When you are developing training programs, you will have some near transfer outcomes, some far transfer outcomes, and some that are a combination of both.

Getting Beneath the Cover Story Recall the fortress and the tumor problems. One has a medical setting and the other a military context. Their surface features—what we call the “cover story” are quite different. That is why most people don’t apply the fortress solution to solve the tumor problem. The context of the tumor problem is too different. However, when you look beneath the cover story you find a common solution principle. The convergence principle makes up the structural feature the two stories have in common. The challenge to training far transfer tasks is to help learners see beyond the cover story of the problem or task they are facing and identify the relevant principles. In order to succeed at far transfer tasks, workers must be able to represent a problem in terms of its structural features. To do so they need a schema that is flexible enough to transfer to many situations. Instructional methods that help learners build flexible schemas will impose additional cognitive load. Unlike extraneous load discussed in earlier chapters, this type of load is productive. Therefore, we call it germane load.

How to Build Flexible Schemas That Support Far Transfer Gick and Holyoak (1980) looked for some effective ways to help learners solve the tumor problem. As we discussed, just reading the fortress problem alone was not sufficient. However, they found that asking learners to study two different problems with different cover stories but the same structure helped learners build a more transferable schema. Gick and Holyoak asked learners to read the fortress story and then a story about a fire on an oil rig that was put out by many small hoses aimed toward the middle of the fire. After reading, learners either drew a picture or wrote a summary of the common features of the solutions. This activity focused attention on the convergence principle underlying the two stories and helped learners build a

c09.qxd 9/29/05 18:27 Page 222

222

Efficiency in Learning

more flexible schema. These learners had much greater success solving the tumor problem because they had abstracted the convergence principle from two examples. Experts in all domains excel as a result of many schemas that incorporate knowledge of principles and knowledge of when to apply those principles. Most of these schemas have evolved from experience solving many problems over a long time period and, as a result, abstracting schemas for future use.

What Are Varied Context Examples? To accelerate this abstraction process, provide learners several examples or problems with similar structural features but different cover stories. We call these kinds of examples varied context examples. At the same time, encourage learners to focus on the structure, not on the cover story. For example, in our Excel demonstration lesson on the CD, our goal is to enable workers to construct formulas for a variety of calculations. We built examples using data from a hypothetical small business—Barb’s Bargain Basement. Learners saw examples involving calculation of profit per month, profit per employee, employee base pay, and employee net pay after commissions and taxes. All this diversity in examples and practice exercises will help learners apply formulas to whatever unique work situations they face. In other words, they will be able to successfully transfer what they have learned to their own applications.

WHAT THE RESEARCH SAYS ABOUT VARIED CONTEXT WORKED EX AMPLES In Chapter 8 we described experiments that used lessons designed to teach technical students basic statistics concepts such as mean, median, and mode. Paas (1992) evaluated learning from three lesson versions. Version 1 required the learners to work a series of practice problems. Version 2 required learners to study worked examples of those same problems. The third version required learners to complete partially worked examples (completion examples). Overall, lesson versions that contained either the worked examples or the completion examples resulted in better

c09.qxd 9/29/05 18:27 Page 223

223

P u t Wo r k i n g M e m o r y t o Wo r k w i t h G e r m a n e L o a d

learning outcomes than lesson versions that required learners to solve all of the problems. However, we did not mention in Chapter 8 that the test Paas used to evaluate learning included both near transfer problems and far transfer problems. Figure 9.2 shows an example of a problem presented in the training as well as a near and a far transfer test problem. The instructions and format of the near transfer test problem are almost identical to those of the training problem. In contrast, the far transfer test problem requires the learner to modify the procedure based on an understanding of a statistical mean. As you can see in Figure 9.3, the scores on the near transfer test items did not differ much among the various lesson versions. In contrast, the scores on the far transfer problems were double among the learners who studied from lessons that included the completion and worked example formats. This is because both the completion and the worked examples

Figure 9.2. Three Problems Used in Worked Examples Research. Adapted from Paas (1992).

Problem Used in the Training Calculate the mean temperature at 3:00 PM for the 5 days of data shown: Day

1

2

3

4

5

Temperature

18

16

20

21

19

Near Transfer Test Problem Calculate the mean for the following row of numbers: 13, 16, 16, 13, 14, 12, 15, 25

Far Transfer Test Problem For a known mean of 5, calculate the value of X in the frequency table: Number

Frequency

3 4 5 7

1 4 5 x

c09.qxd 9/29/05 18:27 Page 224

224

Efficiency in Learning

Figure 9.3. Lessons with Worked and Completion Examples Resulted in Better Learning of Far Transfer Test Problems Than All-Problem Lessons. Based on data from Paas (1992). Significant Difference

No Significant Difference

12

10

Near Transfer

Test Score

8

Far Transfer 6

4

2

All Problems Lessons

Worked Examples Lessons

Completion Examples Lessons

versions imposed less cognitive load, freeing working memory capacity to build a deeper understanding. As a result, these learners did better on far transfer test questions. Paas and Van Merriënboer (1994) measured the efﬁciency of learning from worked examples and all problem lessons that used either all similar or diverse instances. The instructional goals involved mathematical and geometric principles. Four experimental conditions were compared: worked examples of high and low variability; solving of problems of high and low variability. Learners who studied the worked example lessons learned faster, taking about 45 percent of the time required by those in the all problem lessons. Those studying the worked example lessons also rated the materials as requiring signiﬁcantly lower mental effort. As shown in Figure 9.4, the worked example lesson versions were more efﬁcient than the versions that required learners to solve all of the problems. Notice that the lessons with high variable worked examples resulted in better performance than the lessons with low variable worked examples.

c09.qxd 9/29/05 18:27 Page 225

225

P u t Wo r k i n g M e m o r y t o Wo r k w i t h G e r m a n e L o a d

Figure 9.4. Efﬁciency Is Greatest for High Variable Worked Examples. Source: Paas and Van Merriënboer (1994). High Efficiency

Performance 1.0

High Variable Worked Example

0.8

E=0

0.6 0.4 Low Variable Worked Example

−1.0 −0.8 −0.6 −0.4 −0.2

0.2

0.2

Mental Effort

0.4

0.6

0.8

1.0

−0.2 −0.4

Low Variable All Problems

−0.6 −0.8 −1.0

High Variable All Problems Low Efficiency

The Psychology of Varied Context Worked Examples The research we have reviewed supporting Guideline 21 suggests that when learners must adapt new knowledge and skills to problems or situations that differ from those presented in training, they need to build a flexible schema. Schemas that can adapt to far transfer outcomes are based on domainspecific principles such as the convergence principle illustrated with the tumor problem. To encourage learners to build principle-based schemas, provide a variety of instances in the form of worked examples and practice problems that vary the cover stories but incorporate the structural features. Through engagement with varied context examples and practice, learners formulate schemas that are principle-based. The result is learners who know not only the procedures but also the conditions under which those procedures apply. However, the deeper processing required to abstract principles from diverse examples imposes an additional load on working memory. Using

c09.qxd 9/29/05 18:27 Page 226

226

Efficiency in Learning

worked examples helps offset this load. Psychologically, a method that imposes germane cognitive load (diversity in examples and practice) is effective because extraneous cognitive load is minimized (through worked examples rather than all practice).

Applying the Research The research supporting Guideline 21 tells us that: • Examples and practice with diverse contexts lead to better far transfer learning than examples that are similar. • Examples and practice with diverse contexts add more cognitive load than examples that are similar. • Worked examples help reduce cognitive load. • When faced with far transfer instructional goals, you should provide a series of diverse context examples to add germane load and make them worked examples to reduce extraneous load.

Guideline 22: Help Learners Exploit Examples Through Self-Explanations We’ve seen in this chapter and in Chapter 8 that worked examples or completion examples are an efficient way to help learners build new mental models. However, often learners do not process examples very deeply. In the case of full worked examples, they may skip them altogether. In the case of completion examples, they may finish the example by filling in the steps but not pay much attention to the principles underlying the examples. As a result of shallow processing of the example, learning suffers. One way to promote better learning from worked examples is to encourage your learners to make effective self-explanations.

What Are Self-Explanations? A self-explanation is a mental dialog that learners have when studying a worked example that helps them understand the example and build

c09.qxd 9/29/05 18:27 Page 227

227

P u t Wo r k i n g M e m o r y t o Wo r k w i t h G e r m a n e L o a d

a schema from it. Chi and her colleagues (1989) reported that students who generated more self-explanations of examples learned more than students who generated fewer self-explanations. They found that, when studying physics examples, better learners generated on average 15.5 selfexplanations per example, compared to 2.75 from poorer learners. A good self-explanation leads to an understanding that builds a complete and accurate mental model. In the sample self-explanation of a physics problem shown in Figure 9.5, you see a physics worked example (diagram and bold text) and (in italics) the learner’s self-explanations. Note that in the first statement, the learner identifies his confusion between the mass (W) and the body (the knot). Misunderstandings often arise from a conflict between what is read in the instructional materials and one’s existing schema. In the process of the selfexplanation, this learner identifies the conflict and revises his conception of mass and body. Good self-explanations either identify and correct

Figure 9.5. A Student Self-Explanation of a Physics Problem. Source: Chi (2000).

Problem (with Self Explanations)

Diagram

The diagram shows an object of Weight W hung by strings. Consider the knot at the junction of the three strings to be the body #1 “Why should this (the knot) be the body? I thought W was the body?” The body remains at rest under the action of the three forces #2 “I see, so, the W will be the force and not the body. OK” ...of the three forces shown in the diagram #3 “Uh huh, so.. so they refer to the point as the body.....”

45˚ 30˚

W

c09.qxd 9/29/05 18:27 Page 228

228

Efficiency in Learning

misconceptions or they elaborate on the ideas through inferences. The positive effects of self-explanations have been seen in a variety of content domains and in ages ranging from five years to college students.

What Is a Good Self-Explanation? Many learners do not self-explain. Or when they do, their explanations are shallow. For example, they might just repeat the words in the example. Good self-explanations involve deep processing of the material. Researchers have asked learners to talk aloud while self-explaining and have analyzed the number and type of self-explanations given. They then correlated the patterns of self-explanations with learning. They found three main types of productive self-explanations: (1) monitor and correct; (2) try and check; and (3) make inferences by associating the examples with underlying principles or prior knowledge. In monitor and correct self-explanations, learners identify elements of the worked example they don’t understand and strive to resolve discrepancies. The self-explanations illustrated in Figure 9.5 are of the monitor and correct type. The try and check self-explanation involves reading a worked example, covering up the solution steps, and trying to work it. Then the learner checks her steps and answer against the example. Inferencing involves making new connections, either between separate sentences in the worked example or between the worked example and learner prior knowledge.

WHAT THE RESEARCH SAYS ABOUT HO W T O P R O M O T E P RODUC TIVE SELF-EXPL ANATIONS You can improve learning just by getting learners to engage in good selfexplanations while they are reviewing instructional materials. Here we offer three evidence-based strategies you can use either individually or in combination. First, we recommend training learners how to self-explain productively. Second, consider

c09.qxd 9/29/05 18:27 Page 229

P u t Wo r k i n g M e m o r y t o Wo r k w i t h G e r m a n e L o a d

229

using faded worked examples. Third, add questions to your worked examples that will stimulate self-explanations. All of these techniques will add germane cognitive load to your instruction.

EXPERIMENT ON TRAINING LEARNERS TO SELF-EXPLAIN Renkl and his colleagues (1998) compared learning from worked examples of calculating interest problems between two groups of bank apprentices. One group received a twenty-minute training program on how to effectively selfexplain. The training consisted of a good self-explanation example followed by a practice assignment during which the learners self-explained aloud and received feedback and coaching from an instructor. Both trained and untrained groups were asked to talk aloud while they studied nine worked-out examples of calculating interest problems. The research team found that the trained group produced twice as many self-explanations, more of which were of higher quality than those of the untrained group. The training resulted in better learning among low prior knowledge learners on near transfer test items and among both high and low prior knowledge learners on far transfer test items. In summary, a brief training program that included only one example and one practice resulted in more self-explanations of higher quality, which in turn led to better learning.

EXPERIMENT ON USING FADED WORKED EXAMPLES Renkl, Atkinson, and Grosse (2004) found that individuals learned most about steps that were faded regardless of their location in the worked example. In other words, whether a step was faded early or later in the sequence of steps, the principle associated with the faded step was learned best. An analysis of the learner’s study patterns showed that, when working with faded worked examples, learners addressed misunderstandings that they identiﬁed. In contrast, when studying from lessons with example-problem pairs, they ignored their misunderstandings. It seems that faded worked examples help learners incorporate more of the monitor and correct type of self-explanations that beneﬁt learning.

c09.qxd 9/29/05 18:27 Page 230

230

Efficiency in Learning

EXPERIMENT ON ADDING QUESTIONS REQUIRING SELF-EXPLANATIONS OF EXAMPLES Atkinson and his colleagues (2003) tested e-learning lessons that included backwards faded worked examples coupled with questions that required learners to identify the statistical principle that applied to the worked out solution steps. Figure 9.6 shows a sample problem from this study. Note that in this worked example the ﬁrst solution step has been worked out and the learner is required to select which probability rule or principle was applied to that step. As shown in Figure 9.7, the learners who responded to self-explanation questions scored higher on both near and far transfer test items. The effect sizes were .42 for near transfer questions and .37 on far transfer, indicating moderate practical signiﬁcance.

Figure 9.6. A Worked Example with First Worked Step Requiring a Self-Explanation. Source: Atkinson, Renkel, and Merrill (2003).

Problem: From a ballot box containing 3 red balls and 2 white balls, two balls are randomly drawn. The chosen balls are not put back into the box. What is the probability that a red ball is drawn first and a white ball is second?

First Solution Step Total number of balls: Number of red balls: Probability of red ball on first draw Please enter the letter or the rule/principle used in this step:

5 3 3/5 = .6

A. Probability of an event B. Principle of complementarity C. Multiplication principle D. Addition principle

c09.qxd 9/29/05 18:27 Page 231

P u t Wo r k i n g M e m o r y t o Wo r k w i t h G e r m a n e L o a d

231

Figure 9.7. Learning Is Better from Faded Worked Examples with Added Questions That Promote Self-Explanations. Based on data from Atkinson, Renkel, and Merrill (2003).

Percentage Correct on Test

60

Near Transfer

Significant Difference

Far Transfer 50

40

30

20

10

Without Questions

With Questions

Applying the Research The research reviewed in Guideline 21 recommends that you promote deeper processing of worked examples by: • Training learners to self-explain worked examples in small classroom sessions. – Demonstrate several effective self-explanations of examples. – Discuss the demonstrations. – Ask students to practice making self-explanations. – Use peer and instructor feedback to improve self-explanations. – Periodically remind learners to self-explain during training. • Using faded completion examples that will require learners to selfexplain the portions they fill in. • Adding questions to the worked-out portions of completion examples that require learners to understand principles behind the worked examples. (See Figure 9.8 for an example of a self-explanation question in an Excel e-lesson.)

c09.qxd 9/29/05 18:27 Page 232

232

Efficiency in Learning

Figure 9.8. A Question in Excel Lesson Requires Learner to Identify Rule Associated with Worked Step.

WHAT THE RESEARCH SAYS ABOUT SELF-EXPL ANATIONS OF DESIGN TA SKS Wetzstein and Hacker (2004) found that asking learners to answer questions posed by a partner resulted in a better end product from a design task. Sixty adult volunteers with no engineering background were asked to sketch their design for a garden grill of speciﬁed functionality. After their initial design, half of the participants worked individually with partners who asked them the following questions about their design: “How does this work?” “Why did you do it like this?” “Which advantages and disadvantages does this solution have?” “What could a better solution look like?” The other half of the participants completed an unrelated questionnaire. Following these activities, all participants were given an opportunity to revise their designs. The group that described, explained, and evaluated their design to another

c09.qxd 9/29/05 18:27 Page 233

P u t Wo r k i n g M e m o r y t o Wo r k w i t h G e r m a n e L o a d

233

person ended up with more signiﬁcant improvements to their design than individuals who did not participate in discussions. The research team found that the experimental group applied more new principles and improvements as well as added more written explanations to their design sketches than the comparison group.

Self Versus Peer Explanations Of course, since in the garden grill design experiment, the participants were explaining their ideas to another person, by definition this is not a selfexplanation. We wonder whether similar results would be achieved if participants simply had a self-dialog—like the self-explanations we have discussed above. Or would improved designs be achieved if students responded to a virtual partner who asked questions like the ones used in this study? We will look to future research to shed light on these questions.

Guideline 23: Help Learners Automate New Knowledge and Skills In the previous sections of this chapter, we have considered techniques to assist learners to use their cognitive resources in constructing schemas. Of course, learning is more than constructing schemas. Those schemas need to be automated as well so that working memory resources are freed to allow the construction of more complex schemas. In another example of germane cognitive load, the remainder of the chapter considers techniques that assist learners to use their cognitive resources to automate knowledge and skills. There are occasions when a performance aid is not practical, such as when work tasks require fast and accurate responses. Landing an airplane is one example. In addition, complex tasks rely on a number of subtasks. Attempting to perform these complex tasks while at the same time allocating mental capacity to performing the subtasks will overload working memory. In order to perform complex tasks, those prerequisite skills must be available to memory in a way that does not overload working memory. These situations require that knowledge and skills become automated. Under Guideline 23, we define automaticity and discuss when and how to build automaticity in your learners.

c09.qxd 9/29/05 18:27 Page 234

234

Efficiency in Learning

What Is Automaticity As you read this paragraph, you are using a number of automated perceptual and cognitive skills. When you drive a car, you are using automated perceptual and motor skills. In fact, all complex tasks can only be performed effectively because a myriad of underlying subtasks have been automated. When knowledge or skills become automated, they are coded into longterm memory and can be exercised with minimal or no resources from working memory. That is why automaticity offers an internal bypass of working memory. Since automated skills require no or minimal working memory resources, the cognitive system is freed up for multitasking. For example, as I type this sentence I concentrate on writing a series of words that make sense and that contribute to the point of the paragraph. I don’t need to spend any conscious effort locating and striking the computer keys, which would absorb working memory resources that would not be available for composing text. When learning new skills you begin by slowly building a schema in longterm memory. Initially your task performance is slow and inaccurate. Over time, with practice and feedback, performance becomes more accurate. With even more time and practice, the schemas eventually automate. At that point performance is both accurate and very fast.

How Do Skills Become Automatic? There is no magic path to automaticity. Skills become automatic only after many, many practice sessions. While the number of iterations needed for any specific skill may vary somewhat, Shiffrin and Schneider (1977) found that over two thousand practice sessions were required to automate a lettermatching task. Your current reading and math skills grew out of hundreds of drill and practice sessions you had in school. If you are a touch typist, you learned by devoting hours to typing drills. Automaticity can be achieved in training settings that allocate time for a great deal of drill and practice. More often, however, automaticity is actually achieved on the job as a result of repetitive task performance. While achieving automaticity can take a long time, the payoff can be dramatic. Kotovsky, Hayes, and Simon (1985) found that when a problem was presented in a form that permitted subjects to use automated information, on average, it was solved sixteen times faster than a structurally identical problem

c09.qxd 9/29/05 18:27 Page 235

P u t Wo r k i n g M e m o r y t o Wo r k w i t h G e r m a n e L o a d

235

that required the use of non-automated information. Without automation, many tasks become very difficult or even impossible.

When to Build Automatic Skills Workers will need to automate knowledge and skills when learning tasks that require a response that is both very fast and very accurate. For example, train engineers must immediately recognize a track signal and take appropriate action. There is no time to access a signal reference aid. Workers will also need to automate knowledge and skills when the job requires complex responses that rely on multiple subskills that must be coordinated. In other words, when the job requires multitasking, automation of underlying skills is essential.

Applying the Research Repetitive practice needed to build automaticity will impose cognitive load. However, this type of load is called germane load because it is relevant to the instructional goal. Cognitive load theory recommends the following: • Help learners build a new schema in long-term memory with explanations, worked examples, completion examples, and practice with feedback. Your learners have achieved an initial stage of learning when they can perform tasks accurately but not necessarily quickly. At this initial stage of learning, task performance is effortful and demands working memory resources. • Help learners automate these skills by giving learners many opportunities to rehearse new knowledge and skills. This type of repetitive rehearsal is often called drill and practice. Automaticity can be achieved through hands-on practice exercises or with mental rehearsal, as summarized under Guideline 24.

Guideline 24: Promote Mental Rehearsal of Complex Content After Mental Models Are Formed In this chapter we address two phases of learning. During the early stages of learning, help learners form robust mental models by encouraging them to engage in productive self-explanations of varied context faded worked examples. However, once the mental models are formed, automaticity is best

c09.qxd 9/29/05 18:27 Page 236

236

Efficiency in Learning

achieved through practice performing the task and/or solving problems. One form of practice leading to automaticity is mental rehearsal.

What Is Mental Rehearsal? Cooper, Tindall-Ford, Chandler, and Sweller (2001) define mental rehearsal as “the introspective or covert rehearsal that takes place within the individual who thinks through the performance of an activity” (p. 68). Mental rehearsal could involve imagining someone else performing while you observe or imagining doing a task yourself. For example, an Olympic diver is standing at the end of the high diving board with his eyes closed. He is mentally imagining the body moves he will make as he executes the dive. In fact, the interest in the use of imagery that emerged in sports psychology in the 1960s and 1970s continues to be a focus of research today (Etneir & Landers, 1996; Romero & Silvertri, 1990). For our purposes we will look at the effects of mental rehearsal on cognitive tasks such as constructing a spreadsheet formula, plotting points on a graph, or calculating basic statistics.

When to Use Mental Rehearsal Mentally rehearsing the steps to perform a task or solve a problem requires free capacity in working memory. Therefore learners should not rehearse while they are forming their schemas, since both activities would overload working memory. Instead, assign mental rehearsal to help build automaticity after learners have had a chance to form a mental model. In the next section we summarize research showing that asking learners to mentally rehearse a worked example is more effective than asking them to study that example only after they have built a basic mental model of the task.

WHAT THE RESEARCH MENTAL REHEARSAL

SAYS

ABOUT

A number of experiments have shown that engaging learners in mental rehearsal of a worked example is more effective than asking learners to study a worked example once they have already formed the basic schema required to perform the task (Cooper, Tindall-Ford, Chandler, & Sweller, 2001; Ginns, Chandler, & Sweller, 2003; Leahy & Sweller, in preparation; Tindall-Ford & Sweller, (in preparation). In

c09.qxd 9/29/05 18:27 Page 237

237

P u t Wo r k i n g M e m o r y t o Wo r k w i t h G e r m a n e L o a d

this section we summarize two experiments that illustrate the conditions under which mental rehearsal is more effective than studying a worked example. EXPERIMENT 1: COMPARING MENTAL REHEARSAL WITH STUDY DURING EARLY AND LATE LEARNING PHASES Cooper, Tindall-Ford, Chandler, and Sweller (2001) compared learning from mental rehearsal to learning by studying a worked example among two groups of eighth graders. The lesson focused on how to calculate the midpoint and gradient between two points on a number plane. Phase 1 of the instruction presented an explanation followed by two worked examples. The learners were randomly assigned to either study the examples or to mentally rehearse the steps in the examples. Phase 2 of the instruction presented a second pair of worked examples similar to those presented in Phase 1. However, in Phase 2 the student groups were reversed so that the group that had studied previously was now directed to mentally rehearse the steps and those who had mentally rehearsed were directed to study. In this way the effects of mental rehearsal compared to study assignments could be compared during initial and later learning of a new skill. As shown in Figure 9.9, the group that studied during the ﬁrst phase and engaged in mental rehearsal during the second phase scored higher and took less Figure 9.9. Better Learning of Complex Content from Study in Initial Sessions and Rehearsal in Later Sessions. Based on Data from Experiment 2 (Leahy & Sweller, in preparation). 100

Study First – Rehearse Second Rehearse First – Study Second 180 160

% Correct on Test

95

140

SD 120 90

SD

100 80

85

60

Test Scores

Test Time

Seconds to Complete Test

SD = Significant Difference

c09.qxd 9/29/05 18:27 Page 238

238

Efficiency in Learning

time on the test than the group that rehearsed during the ﬁrst phase and studied during the second phase. EXPERIMENT 2: COMPARING MENTAL REHEARSAL WITH STUDY ON LEARNING OF HIGH AND LOW COMPLEXITY CONTENT DURING EARLY AND LATE LEARNING PHASES Leahy and Sweller (in preparation) compared learning how to interpret a temperature graph among two groups of ﬁfth graders. In this experiment, they varied three conditions: rehearsal or study, early and later learning, and complexity of the content. Half of the learners were asked to study the material in the ﬁrst learning session and half were asked to mentally rehearse. They were then given a second learning session in which half were again asked to study and half asked to mentally rehearse, with learners randomly assigned to the two conditions in both phases. Leahy and Sweller anticipated that in the second session learners will have gained more expertise and the outcomes would differ as a result. They further compared learning outcomes on high and low complexity test questions. As we have discussed in previous chapters, cognitive load is most limiting when content is more complex.

Figure 9.10. Study Followed by Rehearsal Results in Better Learning of Complex Content. Source: Leahy & Sweller (in preparation) SD = Significant difference Session 2*

Study Rehearse

Session 1* 90

Session 2

85

Session 1

% Correct on Test

80 75 70 65 60 55

SD

50 SD

45 40 Low Complexity Content

High Complexity Content

* Test scores for low complexity content not significantly different

c09.qxd 9/29/05 18:27 Page 239

P u t Wo r k i n g M e m o r y t o Wo r k w i t h G e r m a n e L o a d

239

As shown in Figure 9.10, it’s not surprising that learning was greater for all learners after Session 2 than after Session 1. Likewise, overall learning of the low complexity content was greater. Of greatest relevance to our discussion is that, in Session 1, study was superior to rehearsal on high but not low complexity material, but in Session 2, rehearsal was superior to study, again only on high complexity material. This result was obtained because by Session 2 the learners had acquired a basic understanding of the high complexity material so they were able to proﬁt from mental rehearsal. Conversely, study was superior in the ﬁrst session because, when dealing with high complexity content, studying examples imposes less load than rehearsal. We know that, once a basic schema is formed, it can be automated through overt practice or through mental rehearsal. We will need additional research to help us determine when overt practice is more effective than mental rehearsal.

Applying the Research The experiments supporting Guideline 24 tell us that: • Mental rehearsal is an effective mechanism for automating new knowledge and skills. • For skills of high complexity, help learners build a basic schema first by studying worked and completion examples. • For skills of high complexity, once the basic schemas are formed, learners can automate them through mental rehearsal of the steps. • For skills of low complexity that do not demand as much working memory resources, the sequence of studying and mental rehearsal does not matter.

••• The Bottom Line There are solid research and psychological reasons to recommend that you should: • Provide learners with varied context examples to encourage building of more complex mental models that will support far transfer.

c09.qxd 9/29/05 18:27 Page 240

240

Efficiency in Learning

• Format diverse examples as faded completion examples to compensate for extra germane load. • Promote productive self-explanations of examples by training learners to self-explain or by requiring them to respond to questions that focus their attention on the underlying principles. • Ask learners to mentally rehearse steps involved in a task once a basic schema is developed. Mental rehearsal is most effective when the instructional goal involves content of moderate to high complexity. Tips for Design of Lessons That Impose Germane Load The experimental research reviewed in this chapter mostly used highly structured content such as mathematics for which it is easy to deﬁne underlying principles. Often, however, the instructional professional must derive the structure or principles underneath far transfer tasks, such as effective selling or coaching. To derive the structure you should:

• Identify several top performers in the tasks to be trained, for example, top sales professionals or best coaches. • Observe and record the activities of these performers and ask interview questions about the reasons they took the actions they did. • Distill the commonalities among these top performers into best practice principles. Once you have identiﬁed the key principles behind effective far transfer performance, you should then: • Use diverse contexts such as different clients and different coaching scenarios as the basis for varied context worked examples and practice. • Promote learner engagement with the varied context worked examples by converting them into completion examples and by requiring learners to identify the main underlying principles. • When training involves design tasks such as creating a website or an improved business process, encourage peer explanations of student products by responding to questions such as, What is the rationale for this design? How could it be improved?

c09.qxd 9/29/05 18:27 Page 241

P u t Wo r k i n g M e m o r y t o Wo r k w i t h G e r m a n e L o a d

241

• If skill automaticity is needed, ensure that learners rehearse (mentally or physically) until their responses are fast and error free. • Teach learners the beneﬁts of mental rehearsal so that they can use it themselves when appropriate.

On the CD John Sweller Video Interview Chapter 9. Put Working Memory to Work with Germane Load. Following a description of germane cognitive load, John discusses variable worked examples, self-explanations, and mental rehearsal techniques.

Sample Excel e-Lessons The asynchronous Load-Managed Excel Web-Based Lesson as well as the synchronous Virtual Classroom Example illustrate the use of: 1. Diverse contexts in examples and practice exercises for Excel formula construction. 2. Instructor promotion of processing of examples in the virtual classroom by asking learners to state the rules governing formulas as they review examples.

COMING NEXT Throughout this book we have provided a number of evidence-based techniques that reduce extraneous cognitive load and/or increase germane cognitive load. We have emphasized that these techniques are most relevant when the content is complex and when the learners are relatively novice. In Part IV, we will look at ways to adjust instruction as learners gain more expertise. Research we will review shows that, as learners gain expertise, many of the cognitive load techniques we have presented thus far not only do not help learners, but in fact they hinder their learning! Instructional scientists

c09.qxd 9/29/05 18:27 Page 242

242

Efficiency in Learning

call this phenomena the expertise reversal effect. In the next chapter you will learn how to accommodate individual differences among learners with greater and lesser amounts of prior experience.

Recommended Readings Chi, M.T.H. (2000). Self-explaining expository texts: The dual processes of generating inferences and repairing mental models. In R. Glaser (Ed.), Advances in instructional psychology educational design and cognitive science (Vol. 5). Mahwah, NJ: Lawrence Erlbaum. Leahy, W., & Sweller, J. (2004). Cognitive load and the imagination effect. Applied Cognitive Psychology, 18, 857–875. Paas, F.G.W.C., & Van Merriënboer, J.J.G. (1994). Variability of worked examples and transfer of geometrical problem-solving skills: A cognitive load approach. Journal of Educational Psychology, 86(1), 122–133. Renkl, A., Atkinson, R.K., & Grosse, C.S. (2004). How fading worked solution steps works: A cognitive load perspective. Instructional Science, 32, 59–82.

c10.qxd 9/29/05 18:29 Page 243

PA R T F O U R

Tailoring Instruction to Learner Expertise

N PART IV WE DESCRIBE some of the more recent research using

I

cognitive load theory. Specifically, we discuss how instructional methods must be adjusted as learners gain expertise. In cognitive load theory, the changes in cognitive load support required as learning progresses is called the expertise reversal principle. Since learning involves a gradual transition from novice to expert, ideal instruction should be dynamically adjusted to accommodate evolving expertise. Adapting instructional methods in e-learning by using rapid tests offers one recently discovered method you can use to assess learner expertise quickly and to tailor instruction appropriately.

243

c10.qxd 9/29/05 18:29 Page 244

244

Efficiency in Learning

Read

To Find Out How to

Chapter 10. Accommodate Differences in Learner Expertise

Write effective text for novice and expert learners Eliminate instructional elements that become redundant for experts Transition from worked examples to problem assignments as learners gain expertise Use directive course designs for novice learners

Chapter 11. Use Rapid Testing

Design tests that allow you to tailor instruction to changing expertise of learners

On the CD Video Interview with John Sweller: Preview/Review Chapter 10. Accommodate Differences in Learner Expertise. A discussion of which individual differences should influence instruction and a description of the expertise reversal effect. Chapter 11. Use Rapid Testing to Adapt e-Learning. A discussion of adaptive testing, including an example, potential future applications of rapid testing and learner control.

c10.qxd 9/29/05 18:29 Page 245

Ta i l o r i n g I n s t r u c t i o n t o L e a r n e r E x p e r t i s e

245

Sample Excel e-Lessons After: Load-Managed Excel Web-Based Lesson. This asynchronous e-learning sample starts with a rapid test that will branch the learner to the appropriate instructional methods based on their demonstrated expertise. If you view the entire program, you will notice a transition in examples so that initial worked examples gradually require more and more learner work to complete, eventually ending with full practice assignments. In this way learner input is increased and instructor input is decreased gradually as the learner gains expertise.

c10.qxd 9/29/05 18:29 Page 246

CHAPTER OUTLINE Beyond Learning Styles: Which Differences Matter? What Are Interactions?

How Cognitive Load Changes with Greater Expertise Cognitive Load Methods Are Schema Substitutes

Expertise Reversal Applied Evidence for Expertise Reversal Comparison Group Experiments Staged Experiments

Guideline 25: Write High Coherent Texts for Low Knowledge Readers Guideline 26: Avoid Interrupting Reading of Low Skilled Readers Applying the Research

Guideline 27: Eliminate Redundant Content for More Experienced Learners When to Drop the Diagram or Drop the Text

Guideline 28: Transition from Worked Examples to Problem Assignments as Learners Gain Expertise A Review of Research on Worked Examples as Learning Progresses Applying the Research

Guideline 29: Use Directive Rather Than Guided Discovery Learning Designs for Novice Learners Applying the Research

The Bottom Line On the CD John Sweller Video Interview Sample Excel e-Lesson

c10.qxd 9/29/05 18:29 Page 247

Accommodate Differences in Learner Expertise

Lesson A

Expert

Novice

Expert

Novice

Learning

10 Lesson B

I

NSTRUCTIONAL METHODS that promote efﬁcient learning serve as

substitutes for missing schemas in the long-term memory of novices. However, as a learner gains expertise, her new schemas compensate for limited working memory capacity. Many of the instructional methods that are effective for novices either have no effect or, in some cases, depress the learning of learners with more expertise. This outcome is called the expertise reversal effect. As a result, training designed for learners with greater prior knowledge requires different instructional methods than does training designed for novices. In this chapter we present research, psychological rationale, and examples to support the following ﬁve guidelines to adjust training to learner expertise:

Guidelines for Accommodating Differences in Expertise 25. Write highly coherent texts for low knowledge readers. 26. Avoid interrupting reading of low skilled readers. 27. Eliminate redundant content for more experienced learners. 28. Transition from worked examples to problem assignments as learners gain expertise. 29. Use directive rather than guided discovery learning designs for novice learners.

247

c10.qxd 9/29/05 18:29 Page 248

248

Efficiency in Learning

Beyond Learning Styles: Which Differences Matter? Learning styles are one type of unproductive instructional mythology pervasive in the training profession. At best, most learning style programs are a waste of resources, and at worst, they lead to instructional methods that actually retard learning. A good example is the popular idea that there are visual learning styles and verbal learning styles. According to this myth, some people learn more from their visual senses, while others learn more from their auditory senses. In a misguided attempt to accommodate visual and auditory styles, many e-learning courses present content using visuals, text to describe the visuals, and audio narration of that text. However, as we have seen in Chapter 5, presenting words simultaneously in text and in audio results in a psychological redundancy that overloads working memory and depresses learning. The concept of learning styles is based on an assumption of individual differences among learners that reliably interact with different types of instruction. For example, according to the visual-auditory learning style myth, an individual with a visual learning style will learn more from a visual presentation than from an auditory presentation, whereas an individual with an auditory learning style shows the reverse pattern. Any consistent response to a specific instructional method by a group of learners that share a common attribute results in what instructional scientists call an interaction. We’ve touched on interactions in previous chapters, but since this entire chapter is about interactions, it’s an opportune time to take a closer look.

What Are Interactions? An interaction means that individuals of Type A will have a better or worse learning outcome with a given instructional Method Z than will individuals of Type B. When two instructional methods have opposite effects on two different types of learners, the result is called a disordinal interaction. A disordinal interaction exists, for example, when Type A individuals learn more from Method Y than from Method Z and, at the same time, Type B individuals learn more from Method Z than from Method Y. Figure 10.1 shows what a disordinal interaction looks like

c10.qxd 9/29/05 18:29 Page 249

249

Accommodate Differences in Learner Expertise

Figure 10.1. A Disordinal Interaction Between Method Y and Z for Type A and Type B Learners. 70 Method Y

Percent Correct on Test

60

50

40

30

20

Method Z

Type A Learners

Type B Learners

when the outcome data are plotted. As you can see, for learners of Type A, instructional Method Y resulted in much better learning than instructional Method Z. In contrast, Type B individuals show the reverse pattern. To maximize learning when you have this type of interaction, you would first determine whether any given learner was a Type A or a Type B and then tailor the instruction to match. In order to adopt learning-style-based programs, we need replicated valid evidence (based on random assignment of learners to different instructional programs) showing that learners of Style A do better with one type of instructional approach, while learners of Style B do better with a different approach. We have not seen valid replicated evidence behind many of the popular learning styles, such as verbal and visual. However, prior knowledge is the one individual difference that has consistently been shown to interact with different instructional methods. Luckily, prior knowledge is an individual difference that is easier to assess than many other differences. In this chapter we will present guidelines, evidence, and psychology that show how you should adapt your instruction based on the amount of relevant prior knowledge possessed by your target learners.

c10.qxd 9/29/05 18:29 Page 250

250

Efficiency in Learning

How Cognitive Load Changes with Greater Expertise Recall the experiments we summarized in Chapter 2 that compared expert and novice chess player memory for chess boards. Not surprisingly novices needed to refer back to a mid-play chess board more times than did experts in order to reconstruct it. Most interesting, however, was the reversal in this pattern when the chess board was a random board—one on which pieces did not reflect any real-world play patterns. When the board was random, expert memory was actually poorer than that of novices. This result is analogous to a pattern that Kalyuga, Ayres, Chandler, and Sweller (2003) term the expertise reversal effect. As a novice chess player, your understanding of a mid-play and a random chess board are the same. Neither has any special meaning beyond a mix of game pieces placed on squares. However, as you gain expertise, your understanding of play patterns makes viewing a random board disruptive. Applied to training, expertise reversal predicts that a given instructional method that works well for novice learners is not only not useful for individuals with more expertise, but also results in depressed learning outcomes! In other words, in expertise reversal you will see an interaction between instructional method A that is effective for novices and instructional method B that is effective for experts. The reason that expert chess players don’t do well remembering random chess boards is that they view chess boards through eyes that are trained to look for meaningful patterns. These patterns are based on schemas that experts have built in their long-term memory over many chess games. When faced with a real mid-play chess board, these schemas allow experts to represent several pieces as one large item of information in working memory. A consequence of interpreting the world through schemas is a larger virtual working memory capacity for incoming information. However, when those familiar patterns disappear, as in a random chess board, expert performance is degraded due to the resources diverted to resolving the conflict between their schemas and the environment.

Cognitive Load Methods Are Schema Substitutes Throughout this book we have presented a number of evidence-based instructional methods that manage cognitive load. These methods are effective

c10.qxd 9/29/05 18:29 Page 251

Accommodate Differences in Learner Expertise

251

because they serve as schema substitutes for novice learners. Since novices don’t have relevant schemas, the instruction needs to serve the role that schemas in long-term memory would serve. The instructional methods that serve as schema replacements for novices are not needed by experts and consequently are redundant. Recall from Chapter 5 that a self-explanatory diagram should not be re-explained with words. Adding redundant explanatory information only burdens working memory with unneeded data and consequently depresses learning. As an individual gains expertise, he builds many relevant schemas applicable to further learning in his area of expertise. Therefore most of the instructional methods we have described throughout this book become redundant for learners with relevant knowledge in the instructional domain. In short, just about all of the guidelines we have presented in previous chapters apply to novice learners. Once learners have gained relevant prior knowledge, you need to readjust your instructional methods.

Expertise Reversal Applied As a consequence of expertise reversal, efficient instruction for more advanced learners will require different methods than training designed for entrylevel learners. If your training is extensive and supports a transition from low to high proficiency, you will need to adjust your more advanced lessons to accommodate the greater experience of your learners. A typical example is new hire training. Many new hire training programs extend over a number of weeks and combine formal training with on-the-job experience. During this process, trainees who started as complete novices build skills reaching relatively high levels of proficiency by the end of the training. In an ideal situation, the initial lessons that the new trainees study apply most of the load management guidelines we have suggested throughout this book. However, as learning progresses, the instructional materials transition to different methods that we summarize in this chapter. If you are delivering instruction by self-paced e-learning, you can take advantage of a newly discovered efficient testing method to quickly assess the expertise of each learner as she progresses in the training. Applying this

c10.qxd 9/29/05 18:29 Page 252

252

Efficiency in Learning

method results in an ongoing dynamic evaluation of learning accompanied by the assignment of the best instructional methods for a given level of expertise. We will provide more details of this method in the next chapter.

Evidence for Expertise Reversal Throughout this chapter we will summarize evidence for the expertise reversal effect. The experiments we describe use one of two approaches that we call comparison group and staged.

Comparison Group Experiments Most of the research we have discussed throughout the book uses a comparison group design in which learning from one lesson version is compared with learning from an alternative version. In a comparison group experiment designed to test prior knowledge differences, two or more instructional versions are tested on a group of low prior knowledge learners and on a different group of high prior knowledge learners. For example, two lesson versions on the same topic are developed: one with text alone and one with diagrams and text. Then a group of high and low prior knowledge learners is divided in half and randomly assigned to one of the two lesson versions. All learners are tested at the end of the study period. This experimental plan yields four combinations: 1. Low prior knowledge learners studying the version with text alone; 2. Low prior knowledge learners studying the version with text plus diagram; 3. High prior knowledge learners studying the version with text alone; and 4. High prior knowledge learners studying the version with text plus diagram. If the expertise reversal effect applies, you will see an interaction in which instructional methods that promote better learning among low prior knowledge learners are not as effective for high prior knowledge learners,

c10.qxd 9/29/05 18:29 Page 253

253

Accommodate Differences in Learner Expertise

and vice versa. When plotted on a graph, the results will look similar to Figure 10.1, in which Type A will be low prior knowledge and Type B will be high prior knowledge. This experimental design is frequently called a “cross-sectional” design because we use a cross section of lower and higher expertise learners, rather than observing the same learners as they gain expertise.

Staged Experiments The other experimental design frequently used to study differences in expertise is called a staged design and is illustrated in Figure 10.2. In these studies, learning outcomes are compared in the same group over time as learning progresses. The experiment starts with all novice learners, who are randomly assigned to study either the text alone version or the text plus diagram version. After the initial study period, all learners are tested. Then the entire group completes a well-designed lesson to ensure that everyone has an equally effective opportunity to build expertise. After completing the well-designed training, Stage 2 begins. Stage 2 may immediately follow Stage 1 or may be scheduled at a later time, such as the following week. During Stage 2, the learners are again assigned to either the text alone version or the text plus diagram version. After a study period, learners are tested again. By Stage 2, learners have gained more expertise, and the experimental lesson versions

Figure 10.2. A Generalized Plan of a Staged Experiment. Experimental Lesson Versions

Experimental Lesson Versions

Stage 2

Version 1 Test

Version 2

Training Session 2

Stage 1

Version 1 Test

Test

Version 2

Training Session 1

Version 1

Experimental Lesson Versions

Version 2

Stage 3

c10.qxd 9/29/05 18:29 Page 254

254

Efficiency in Learning

(text alone or text with diagrams) might have a different effect on learning. This plan may be followed for several stages. Each stage would include a period of study of the experimental lesson materials plus a training period with well-designed materials to ensure that everyone gains expertise. When the expertise reversal effect applies, the version that worked well for the learners in the early stages will not work as well for them as they gain expertise and eventually will result in less learning than the comparison versions. When the data is plotted on a graph, the pattern will look similar to that shown in Figure 10.1, with Type A learners replaced with Stage 1 and Type B learners replaced with Stage 2. This experimental design is frequently called a “longitudinal” design because we use the same learners whose expertise alters over time, rather than a cross section of learners with different levels of expertise. In either the comparison group or the staged approach, if the learners rate the difficulty of the experimental instructional materials at the end of their study period, an efficiency metric can be calculated. These metrics can be plotted on the efficiency graph, as we discussed in Chapter 1. The main difference you will see on these efficiency graphs is two data points for each test lesson: one set of data points representing efficiencies for low knowledge learners and one set representing efficiencies for high knowledge learners. In this chapter we provide evidence-based guidelines for instructional methods that work best with low knowledge learners and with high knowledge learners. We summarize what is known about individual differences regarding learning from text alone, text with and without diagrams, worked examples and problem assignments, as well as directive and guided discovery course designs.

Guideline 25: Write High Coherent Texts for Low Knowledge Readers This first guideline might sound odd, since this is a chapter about instructional methods for learners with high knowledge. However, we have evidence that you will get a return on investing resources in making texts very clear for novices—but not for readers with relevant experience.

c10.qxd 9/29/05 18:29 Page 255

255

Accommodate Differences in Learner Expertise

WHAT THE RESEARCH T E X T CO H ER E NC E

SAYS

ABOUT

McNamara, Kintsch, Songer, and Kintsch (1996) compared learning among high and low prior knowledge students after reading high and low coherent texts on heart disease. You can compare a segment from the low and high coherent versions in Figure 10.3. The underlined text in the revised version indicates edits made to add clarity to the less coherent version. For example, note that section headers have been added. Section headers are useful to signal paragraph topics to readers (Lorch, 1989). Second, pronouns such as “it” in the low coherence text were replaced with the appropriate reference, such as “heart” when there was a possibility of ambiguity in interpretation. Third, new sentences were added to explain unfamiliar concepts. For example, note the sentence added to

Figure 10.3. Excerpts from Low and High Coherent Texts (edits are underlined in the coherent version). Adapted from McNamara, Kintsch, Songer, and Kintsch (1996).

Minimally Coherent Version The heart is the hardest-working organ in the body. We rely on a regular blood supply every moment of every day. Any disorder that stops the blood supply is a threat to life. Heart disease is very common. More people are killed every year in the U.S. by heart disease than by any other disease. A congenital disease is one that a person is born with. Most babies are born with perfect hearts. In about one in every 200 cases something goes wrong. Sometimes a valve develops the wrong shape. It may be too tight or fail to close properly. Sometimes a gap is left in the septal wall between the two sides of the heart. This is called a septal defect. When a baby’s heart is badly formed, it cannot work efficiently. The blood does not receive enough oxygen. The baby becomes breathless. The blood cannot get rid of carbon dioxide through the lungs. It becomes

Coherent Version The heart is the hardest working organ in the body. We rely on it to supply blood regularly to the body every moment of every day. Any disorder that stops the heart from supplying blood to the body is a threat to life. Heart disease is such a disorder. It is very common. More people are killed every year in the U.S. by heart disease than by any other disease. There are many kinds of heart disease, some of which are present at birth and some of which are acquired later. 1. Congenital Heart Disease A congenital heart disease is a defect that a baby is born with. Most babies are born with perfect hearts. But one in every 200 babies is born with a bad heart. For example, hearts have flaps, called valves, that control the blood flow between its chambers.

c10.qxd 9/29/05 18:29 Page 256

256

Efficiency in Learning

the bottom of the coherent version that explains what heart valves are and how they function. Finally, new preview sentences were added. For example, at the end of the ﬁrst paragraph the authors added the following preview sentence: “There are many kinds of heart disease, some of which are present at birth and some of which are acquired later.” This preview sentence orients readers to the topics to follow. The low coherent text is understandable; however, the ideas have not been explicated to the same extent as the coherent version. For novices, comprehension of the low coherent text is likely to require greater mental effort. Prevalent wisdom suggests that the more coherent text version would be the best resource for everyone. After all, isn’t high coherence in writing always better? Surprisingly, the research team found that high prior knowledge readers who studied the low coherent version performed better on a problem-solving test than high prior knowledge learners who studied the high coherent version. As you can see in Figure 10.4, low prior knowledge readers showed the opposite pattern.

Figure 10.4. Opposite Learning Outcomes from High and Low Coherent Text by High and Low Prior Knowledge Readers. Source: McNamara, Kintsch, Songer, and Kintsch (1996). 70

Percent Correct on Test

60

50

High Coherent Text

40

30 Low Coherent Text

20

Low Prior Knowledge Readers

High Prior Knowledge Readers

c10.qxd 9/29/05 18:29 Page 257

Accommodate Differences in Learner Expertise

257

Here we see an example of the expertise reversal effect. A high coherence text that most effectively helped low prior knowledge learners understand its meaning exerted the opposite effect on high prior knowledge readers. High prior knowledge readers learned more from a low coherent text. An explanation for this result based on cognitive load theory is that high prior knowledge readers already have adequate schemas about heart functions. Adding schema support in the form of extra explanations and preview sentences was redundant for high prior knowledge readers. The load added to their working memory had a negative overall impact on their processing of the text. In contrast, because low prior knowledge readers lacked appropriate schemas, they experienced additional extraneous cognitive load when reading the less coherent text. We need more research on text coherence before we recommend writing low coherent texts for more experienced learners. We can say that adding explanations that are unnecessary for a given level of readership should, according to cognitive load theory, have negative effects, as found by McNamara, Kintsch, Songer, and Kintsch (1996). We do recommend that extra investment be made to prepare highly coherent instructional texts for novice learners. From a cognitive load perspective, reading difficulty was not rated, so we do not know whether the outcome observed in this study reflects a redundancy effect. We look forward to more research for evidence-based writing guidelines to guide our writing practices for high and low knowledge readers.

Guideline 26: Avoid Interrupting Reading of Low Skilled Readers While we are on the topic of learning from reading, we briefly mention research showing that it’s not a good idea to ask poor readers to answer questions about what they read while they are reading. That’s because individuals with weak reading skills need to invest all of their cognitive effort in processing the text. To ask them to respond to questions at the same time leads to cognitive overload and less learning.

c10.qxd 9/29/05 18:29 Page 258

258

Efficiency in Learning

WHAT THE RESEARCH SAYS ABOUT A SKING QUESTIONS DURING READING Van den Broek, Tzeng, Risden, Trabasso, and Basche (2001) compared recall of story facts among readers of different skill levels, including fourth graders, tenth graders, and college students. As you can see in Figure 10.5, the test scores show a signiﬁcant interaction. The readers with lowest skills (fourth graders), learned less when having to answer questions during reading than fourth graders who did not answer any questions. The reverse pattern is seen for the college students, who are skilled readers.

Figure 10.5. Answering Questions During Reading Had Opposite Learning Effects Among Readers of Different Expertise. Based on data from Van den Broek, Tzeng, Risden, Trabasso, and Basche (2001).

No Questions

Questions During Reading

4th graders

10th graders

college

While most organizational training professionals are not creating materials for fourth graders, in a global economy many learners may not be native speakers of English. In addition many trainees may have low reading proficiencies. Based on this evidence, we recommend that, when designing written materials for lower skilled readers, you should avoid inserting questions into the materials. Instead you can promote understanding by providing opportunities for questions and discussion after completing the reading assignment.

c10.qxd 9/29/05 18:29 Page 259

Accommodate Differences in Learner Expertise

259

Applying the Research The research supporting Guidelines 25 and 26 tells us that: • You should write considerate text for readers unfamiliar with the content. • Considerate text includes guidance such as – Organizing sentences or diagrams that preview or review the content – Definitions and examples of unfamiliar terms – Explicit statement that require minimal inferences – Headers to signal paragraph topics • You need not invest as much effort in writing considerate text for readers familiar with the topic for whom it may be redundant. • You are likely to decrease comprehension of low skilled readers by interrupting their reading with questions. • You can improve comprehension of high skilled readers by including questions with their reading assignments.

Guideline 27: Eliminate Redundant Content for More Experienced Learners In Chapter 3, we saw that relevant diagrams help learners to perform procedural tasks more effectively than text, as well as promote a deeper understanding of content. Further, we recommended that you explain visuals with audio narration. However, when a diagram is self-explanatory, it becomes redundant to add words in any form, as we discussed in Chapter 5. In this section we review guidelines showing that more experienced learners will profit as well as or better from either text or diagrams alone. Experienced learners will not need the redundant information contained in the combination of text and diagram. In fact, it is likely that such redundant information will depress their learning.

c10.qxd 9/29/05 18:29 Page 260

260

Efficiency in Learning

WHAT THE RESEARCH SAYS ABOUT PROV I D I N G E X P E R IEN C ED L EA RN ERS TE X T A LO N E Mayer and Gallini (1990) used a comparison group experimental design to evaluate the effectiveness of diagrams added to text on high and on low knowledge learners. They developed different lesson versions that described three different mechanical processes, including the operations of brakes, pumps, and generators. Among their lesson versions one included text alone and the other included the same text plus relevant diagrams. Figure 10.6 shows two sections from the text plus a diagram lesson on how brakes work. Prior to the experiment, participants answered a survey that assessed their prior knowledge of mechanical systems. On the basis of the survey, the researchers divided the participants into a low and a high knowledge group. These participants were then randomly assigned to study the two lesson versions. After the study period, the learners were tested with questions that asked them to apply what they learned. For example, they were asked, “Why do brakes get hot? What could be done to make brakes more reliable?” As you can see in Figure 10.7, the lesson versions with diagrams added to the text were very helpful for the low prior knowledge learners, but had no beneﬁcial effect on the learning of high prior knowledge learners. The high prior knowledge learners had a sufﬁcient schema to make sense of the text alone. In contrast, the diagram provided a schema substitute for low prior knowledge learners.

Figure 10.6. Text Plus Diagram Lesson Version of How a Brake Works. Adapted from Mayer and Gallini (1990). Parts Illustration

Steps Illustration When the driver steps on the car’s brake pedal... A piston moves forward inside the master cylinder

Tube

Brake drum

The piston forces brake fluid out of the master cylinder and through the tubes to the wheel cylinder

In the wheel cylinder, the increase in the fluid pressure makes a set of smaller pistons move

Wheel cylinder Smaller piston

Brake shoe

When the brake shoes press against the drum, both the drum and the wheel stop or slow down

c10.qxd 9/29/05 18:29 Page 261

Accommodate Differences in Learner Expertise

261

Figure 10.7. Lesson Versions with Diagrams Aid Understanding of Low but Not High Prior Knowledge Learners. Based on Data in Experiment 1 (Mayer & Gallini, 1990). .7

Proportion Correct

.6

Text + Diagram

.5 .4 .3

Text only

.2 .1

Low Prior Knowledge Learners

High Prior Knowledge Learners

In a summary of several experiments comparing effects of diagrams on high and low knowledge learners, Mayer (2001) reports that low prior knowledge learners studying texts with relevant diagrams will on average show performance gains 61 percent greater than performance gains of high prior knowledge readers. As with the reading studies summarized under Guideline 25, the data suggest that we invest more resources in design of effective materials for low prior knowledge than for high prior knowledge learners. First, we should pay attention to the coherence of the written text and, second, we should add relevant diagrams to illustrate the ideas in the text. Note, however, that the expertise reversal effect was not seen in the Mayer and Gallini (1990) study. Learning of high prior knowledge learners was the same from versions with text alone as from versions with text and graphics. Mayer and Gallini did not specifically seek out high and low knowledge individuals. Rather, they took a group of individuals and divided them into higher and lower knowledge groups. By increasing the knowledge base of the high knowledge group, the probability of a full expertise reversal effect is increased.

c10.qxd 9/29/05 18:29 Page 262

262

Efficiency in Learning

WHAT THE RESEARCH SAYS ABOUT PRO V I D I N G EX P ERIEN C ED L EA RN ERS DIAGRAMS ALONE Kalyuga, Chandler, and Sweller (1998) used a staged experiment to evaluate the effects of two lesson versions designed to teach how to interpret circuit diagrams of a motor on learning of novice and more experienced participants. One version included diagrams plus text, and the other included just the diagrams. The effects of the text alone versus text plus diagram were tracked among the same group of learners as they gained expertise over three learning stages similar to the design shown in Figure 10.2. As you can see in Figure 10.8, at Stage 1 the novice participants beneﬁted from lessons that included a combination of diagrams and explanatory text. The novice learners lacked relevant schemas to make sense of a circuit diagram alone. Notice, however, that by Stage 2, either lesson version was equally effective. By this point, the learners had sufﬁcient expertise to proﬁt from the diagram alone but insufﬁcient expertise to be disrupted by the combination of text and diagram. However, as their expertise

Figure 10.8. Diagrams Alone Resulted in Better Learning with More Expert Learners. Based on data from Experiments 2 and 3 (Kalyuga, Chandler, & Sweller, 1998). 11.0

Diagram Only 10.0

Test Scores

9.0

8.0

Diagram + Integrated Text 7.0

6.0

Stage 1

Stage 2

Stage 3 1 month later

c10.qxd 9/29/05 18:29 Page 263

263

Accommodate Differences in Learner Expertise

developed even more over the month between Stages 2 and 3, the text became redundant. By Stage 3, learning was better from a diagram alone. Here we see the expertise reversal effect. As novices, learners needed the schema support provided by text to explain the diagram. As they gradually developed expertise, the beneﬁts of the text disappeared and, once they had gained relatively high levels of expertise, the text became redundant and resulted in diminished learning. These results were replicated in a second experiment in which the diagrams were explained by words presented in audio rather than by text (Kayluga, Chandler, & Sweller, 2000). In Figure 10.9 we show a complex diagram for determining cutting speeds for drills that also includes a table of cutting speeds for various materials as well as the steps to follow to determine a cutting speed. Two lesson versions incorporated diagrams only or diagrams explained by words presented in audio narration.

Figure 10.9. A Worked Example Using Audio to Explain How to Use the Diagram to Determine Cutting Speeds for Drills. Adapted from Kalyuga, Chandler, and Sweller (2000). Cutting Speed for Drills

Revolutions per minute (R.P.M.)

Cutting speed (m/min)

10 8

15 12

25 20

40 32

50

15 20 32 40 50 63 80 100 125 160 200 250 320 400 500 630 800 1000 1250 1600 2000 2500 3200

63 80 100 125

Material

Speed m/min

Aluminum Brass Bronze Cast Iron Copper

60-100 50-80 30-60 25-30 30-60

Assume you wish to determine the appropriate R.P.M. to drill a 25mm diameter hole in the bronze. Step 1. Select the cutting speed. From the table, select the cutting speed range for a given material, in this case, bronze. Step 2. Select the diagonal lines. At the right upper corner of the diagram, select the diagonal line that corresponds to the lowest available cutting speed without the suggested range for bronze.

8

5 5

12 10

20 16

32

50

80 125 200 320 500 800 1250

Step 3. Select the vertical line. At the bottom of the diagram, select the vertical line that corresponds to the required diameter of the hole, in this case, 25 mm.

Diameter in millimeters

Step 4. Find the intersection point. Follow the diagonal line until it intersects with the vertical line.

Help

Step 5. Select the horizontal line. Select the horizontal line that runs through the point.

25

40

63 100 160 250 400 630 1000

Done

Step 5. Read off the R.P.M. By following the horizontal line to the left, we can read off the appropriate R.P.M.

c10.qxd 9/29/05 18:29 Page 264

264

Efficiency in Learning

Similar to the previous experiment, the same group of learners who started out as novices were trained and tested over three stages during which they gained expertise. The results were similar to those shown in Figure 10.8. The diagram explained by audio was most effective for novice learners. However, as their expertise increased, there was a point at which there was no signiﬁcant learning difference between the two versions, and eventually the diagram alone was best. As in the previous experiment (Kalyuga, Chandler, & Sweller, 1998) that used circuit diagrams, the task of determining a cutting speed could not be performed without the diagram. Therefore, once learners had developed a schema adequate to interpret the diagram, the explanatory words became redundant. The efﬁciency diagram shown in Figure 10.10 shows the reversal of efﬁciency of different lesson versions as the learners gained expertise. In both of the previous experiments, words presented as text or as audio narration became redundant to a diagram that was essential to accomplish the task.

Figure 10.10. Diagrams Plus Words Are More Efﬁcient for Novices; Diagrams Alone Are More Efﬁcient for Experts. Source: Kalyuga, Chandler, and Sweller (2000).

High Efficiency Diagram + Audio Stage 1

E=0

1.0 0.8 0.6

Diagram Only Stage 3

0.4 0.2

- 1.0 - 0.8 - 0.6 - 0.4 - 0.2 Diagram + Audio Stage 3

0.2 0.4 - 0.2

Mental Effort 0.6

0.8

1.0

- 0.4 - 0.6

Diagram Only Stage 1

- 0.8 - 1.0

Low Efficiency

c10.qxd 9/29/05 18:29 Page 265

Accommodate Differences in Learner Expertise

265

However, in the Mayer and Gallini experiment, we saw that more experienced learners proﬁted as much from a text-only version as from a version with text and diagram. So you might wonder whether you should drop diagrams or drop text for more experienced learners.

When to Drop the Diagram or Drop the Text The experiments of Mayer and Gallini (1990) and Kalyuga, Chandler, and Sweller (1998, 2000) evaluated different combinations of text and diagrams. None of them compared three conditions of text alone, text plus diagram, and diagram alone. Until we have data from such a comparison, we recommend that when the text alone is self-explanatory, it is more cost-effective to eliminate diagrams in lessons designed for more experienced learners. In contrast, if the diagram is essential to the task, drop the text. For example, interpreting cutting speed from a diagram like the one shown in Figure 10.9 cannot be done without the diagram. When the visual information is essential to performing the task, you will need to delete the text for more experienced learners familiar with the diagram. In contrast, when the visual information is not an integral part of the task, drop the diagrams in lieu of the text.

Guideline 28: Transition from Worked Examples to Problem Assignments as Learners Gain Expertise In Chapter 8, we presented a number of guidelines on how to best sequence and structure faded worked examples in order to gradually impose mental work on learners. Therefore here we will only briefly review how worked examples interact with learners of low and high prior knowledge.

A Review of Research on Worked Examples as Learning Progresses Kalyuga, Chandler, Tuovinen, and Sweller (2001) compared the learning benefits of lessons with worked example-problem pairs to lessons with all-problem

c10.qxd 9/29/05 18:29 Page 266

266

Efficiency in Learning

assignments for training how to write programs for relay circuits of various levels of complexity. Experiment 1 used a staged learning approach. They found that worked example-problem pairs led to greater learning than all-problem lessons for the first two stages. However, by Stage 3, the learning effects of the two lesson versions were equivalent. An expertise reversal effect had not yet occurred. The research team concluded that an expertise reversal would take place if the learners had even more experience with this type of programming. Because they had limited time to work with the participants, they conducted a second experiment using an easier task that could be learned faster. This experiment involved two stages. As you can see in Figure 10.11, sufficient expertise was gained by the learners so that by Stage 2 worked examples became redundant and learning was better with all practice problems.

Figure 10.11. Lessons with All Problems Led to Better Learning of Experienced Learners. Source: Kalyuga, Chandler, Tuovinen, and Sweller (2001). 11

All Problems

10 9

Test Score (out of 12)

8

Worked Examples

7 6 5 4 3 2 1

Stage 1

Stage 2 1 week

Applying the Research The research on worked examples and expertise recommends that you should: • Provide fully worked examples for novice learners.

c10.qxd 9/29/05 18:29 Page 267

Accommodate Differences in Learner Expertise

267

• Convert fully worked examples to completion examples that fade instructional support by requiring learners to fill in greater portions of the examples. • End the sequence with a full practice assignment. • Vary the cover story of the examples and practice when far transfer outcomes are desired. Based on cognitive load theory, once learners have formed their own schemas for performing a task, they are better off solving problems based on those schemas. Having to study worked examples of a task they already know adds an unnecessary and sometimes conflicting repetition, which adds load to working memory. The consequence is depressed learning.

Guideline 29: Use Directive Rather Than Guided Discovery Learning Designs for Novice Learners Recall from Chapter 7 that a directive lesson presents guidelines or steps, shows worked examples of how to apply the guidelines or steps, and then asks the learner to practice through faded worked examples and/or practice problems. For example, in our sample spreadsheet e-lessons on the CD, we initially explain rules for formatting Excel formulas followed by faded worked examples and then practice assignments. In contrast, lessons with a discovery or guided discovery design use a more inductive approach. Inductive lessons provide learners with stories, experiences, or exploratory opportunities from which learners are expected to derive the relevant guidelines or steps. For example, at one point in our Excel virtual classroom lesson, the instructor showed three sample formulas and asked learners to derive the major formatting rules. Various types of guided discovery course architectures are frequently touted as better because the learner is constantly engaged. However, engagement not directed toward schema acquisition and automation can impose an extraneous working memory load. Next, we review two experiments that

c10.qxd 9/29/05 18:29 Page 268

268

Efficiency in Learning

compared learning from a directive lesson in which worked examples were provided with learning from guided discovery lessons in which learners were asked to provide their own examples.

WHAT THE RESEARCH SAYS ABOUT DIRECTIVE AND GUIDED DISCOVERY L E S S O N S Kalyuga, Chandler, and Sweller (2001) prepared two lesson versions, both designed to teach how to write Boolean switching equations to design relay circuits of different levels of complexity. An initial lesson provided everyone with basic knowledge needed for both simple and complex circuits. Next, in Stage 1, participants were randomly assigned to study two experimental lesson versions. In the directive version, the lessons provided worked-example problem pairs for either simple or complex circuits. In the guided discovery version, learners were directed to design high and low complexity relay circuits. If their circuit design was appropriate, they were directed to type in the Boolean equation for their circuit. Therefore, in the guided discovery lesson, learners were basically constructing their own examples. Upon completion of their lessons, all learners rated lesson difﬁculty and were tested. At this point, all study participants were given a training session designed to build expertise in constructing circuits. Stage 2 started with a training session similar to the one that ended Stage 1 and proceeded to experimental sessions using directive or guided discovery approaches. As you can see in Figure 10.12, there was no difference in learning between the directive and guided discovery lessons on simple tasks. However, the worked example lessons were completed much faster and thus offer a more efﬁcient approach. However, for complex tasks the results are different. As shown in Figure 10.13, the directive lesson using worked examples resulted in better learning among novice learners than the guided discovery lesson. However, by the end of Stage 2 when the learners had more expertise, both versions resulted in equivalent learning. In summary, in this experiment learning of high and low complexity content was compared from a directive lesson that used worked-example problem pairs to a guided discovery lesson over time as learning increased. Working memory would be

c10.qxd 9/29/05 18:29 Page 269

269

Accommodate Differences in Learner Expertise

Figure 10.12. A Comparison of Directive with Guided Discovery Lesson Design on Learning and Training Time of Novice Learners on Simple Tasks. Source: Kalyuga, Chandler, and Sweller (2001). Directive Versions Guided Discovery Versions Simple Tasks 10

170

Test Scores

130 6 110

SD 4

90

Training Time (Secs)

150

8

70

2

50

Learning

Training Time

Figure 10.13. A Comparison of Directive and Guided Discovery Lesson on Learning of Complex Tasks by Novice and Experienced Participants. Source: Kalyuga, Chandler, and Sweller (2001). Complex Tasks 10 Directive Using Worked Examples

Test Scores

8

6

4 Guided Exploration

2

Stage 1

Stage 2

c10.qxd 9/29/05 18:29 Page 270

270

Efficiency in Learning

most vulnerable to overload when learners were novices and were working with more complex circuits. As expected, the directive lessons were most effective for novice learners working with high complexity content. In contrast, for low complexity circuits there was little difference in learning between the two lesson versions, although instructional time was faster with directive instruction. A similar result was reported by Tuovinen and Sweller (1999) using instruction on how to work with database programs. In this study a directive lesson that used worked example-problem pairs was more effective for learners unfamiliar with database programs than a guided discovery lesson in which learners were asked to create their own examples. However, for participants who had prior familiarity with databases, either version was equally effective.

Applying the Research The research on directive versus guided discovery lesson designs suggests that you: • Prepare directive lessons for novice learners that provide brief content segments that include explanations, examples, and practice. • Prepare directive lessons when it is important to save training time. • Prepare either directive or guided discovery lessons for more experienced learners.

••• The Bottom Line There are solid research and psychological reasons for recommending that you: • Adapt instructional methods based on individual differences in prior knowledge. – You may need to adjust methods within a lengthy training period as learners gain expertise. – In e-learning you can dynamically adapt training to differences in expertise, as described next in Chapter 11.

c10.qxd 9/29/05 18:29 Page 271

Accommodate Differences in Learner Expertise

271

• Invest effort in writing considerate text for readers unfamiliar with the topic. • Include questions during reading for skilled readers but not for unskilled readers such as English as a second language students. • Include relevant diagrams and text for novice learners. • Drop either the diagram or the text for experienced learners. – Drop the text if the diagram is required to complete the task. – Drop the diagram if the diagram is not essential to task completion. • Start with worked examples and transition to full practice exercises via completion examples as learners gain expertise. • Use directive architectures that provide steps, examples, and practice for novice learners.

Tips for Accommodating Differences in Learner Expertise • During the analysis phase, make a careful assessment of your target audience and instructional goals. – Are your learners novices, experts, or a mix of both? – Will your learners gain expertise through many lessons? – Will the instructional goals involve content that is complex for the target learners? • If you are dealing primarily with novice learners faced with moderate to high complexity goals, apply the many techniques we have described throughout the book to minimize extraneous cognitive load and to maximize germane cognitive load. • If your audience is mixed, consider ways to accommodate diversity in expertise. • If delivering via e-learning, see Chapter 11 on ways to adapt training to expertise. • If delivering via classroom, consider blended instruction in which self-study prework is used to make background knowledge more equivalent prior to class.

c10.qxd 9/29/05 18:29 Page 272

272

Efficiency in Learning

• Consider breaking the class into introductory and advanced sections. • If your audience is made up of learners with medium to high levels of expertise, consider less investment of resources in methods that minimize cognitive load by asking yourself: – Will these learners proﬁt from an investment made to write highly coherent textual materials? – Do these learners need both text and diagrams or would one representation sufﬁce? – Do these learners need worked examples or can they move directly to problem solving? – Should I create a guided discovery design to promote inductive learning?

On the CD John Sweller Video Interview Chapter 10: Accommodating Differences in Learner Expertise. John discusses the types of individual differences that are proven to make a difference to learning, followed by specific discussion of expertise reversal.

Sample Excel e-Lesson The Load Managed Excel Web-Based Lesson starts each topic with worked examples and progresses through a series of faded completion examples. All of the Excel demonstration lessons use a directive lesson design.

COMING NEXT In this chapter we have recommended that you adapt your instruction according to the expertise of your learner. In e-learning, individual measures of learner knowledge can be used as the basis for adapting instructional methods dynamically as each learner progresses through the instruction. Adaptive testing has not been used extensively, however, because it is time-consuming

c10.qxd 9/29/05 18:29 Page 273

Accommodate Differences in Learner Expertise

273

for learners to take tests throughout the training. However, Kalyuga and Sweller (2004) have reported a brand new method of rapid testing that makes tailored instruction in e-learning practical. We summarize their research in Chapter 11.

Recommended Reading Kalyuga, S., Ayres, P., Chandler, P., & Sweller, J. (2003). The expertise reversal effect. Educational Psychologist, 38(1), 23–31.

c11.qxd 9/29/05 18:54 Page 274

CHAPTER OUTLINE e-Learning and Adaptive Training A New Method for Rapid Testing Rapid Tests to Adapt Instruction to Learner Expertise Applying Rapid Assessment to Your e-Learning The Bottom Line On the CD John Sweller Video Interview Sample Excel e-Lessons

c11.qxd 9/29/05 18:54 Page 275

11 Use Rapid Testing to Adapt e-Learning to Learner Expertise

Pretest

1

2

2.1

3

4

2.2

NLIKE CLASSROOM INSTRUCTION, asynchronous e-learning

U

lets you tailor instruction to each learner. e-Lessons can adjust the content and the instructional methods based on learner needs. This match between lesson and learner is called adaptive instruction. Adaptive instruction matches content and methods based on learner expertise before and during the training. Pretests that assign learners to relevant lessons are a common example. A challenge for adaptive instruction is the amount of testing time needed to assess learner expertise. A traditional test that accurately assesses learner expertise must include several questions, and each question may take a number of minutes to complete. In this chapter we describe very recent research that shows how you can develop test questions that are both fast and accurate indicators of expertise. Because this is very recent research, we are not recommending guidelines in this chapter. However, at the end of the chapter we summarize some issues for you to consider as you think about how to apply adaptive testing to your instructional environment.

275

c11.qxd 9/29/05 18:54 Page 276

276

Efficiency in Learning

e-Learning and Adaptive Training One of the benefits of asynchronous e-learning is the potential to tailor training to just what the individual learner needs. If someone has already worked with a spreadsheet and knows the basics of cells and formulas, why should she spend time going through all of the introductory lessons on cells and formulas? Unlike the classroom, where learning mostly proceeds in a lock-step group pace and sequence, individualized instruction should in theory adjust instruction to each learner. As we mentioned in Chapter 1, in the United States we invest over $50 billion a year in organizational training. But that figure does not incorporate the most expensive element of any training program—the time of the learners. If we can save learning time by tailoring instruction to each person’s needs, there is a great potential for economic payoff in efficient instruction. The ability to individualize instruction is called adaptive learning. You are probably most familiar with adaptive learning in the form of pretests that assign learners to the courses or lessons that match their skills. Depending on the size and scope of the course, the learner will typically complete a pretest that includes fifteen to twenty test questions followed by an assignment to specific lessons. However, in most cases, once they start, all learners work through the same lesson, regardless of their rate of learning. Everyone sees the same sequence of examples and works the same number of practice exercises. We know that students learn at different rates and, if we could track learning dynamically in the lesson, we could adjust the amount of instruction to learner needs. However, taking frequent tests throughout a course or lesson is too time-consuming. Therefore, dynamic adaptive instruction has generally not been implemented. In Chapters 8, 9, and 10 we recommended that you use a series of gradually faded worked examples so that, as learners progress through a lesson, they receive completion examples that require them to fill in more and more steps. However, an implicit assumption behind this method is that all learners develop expertise at a linear rate correspondent with the progressively faded worked examples. A more accurate approach would adjust the sequence from worked examples through faded worked examples to full problems

c11.qxd 9/29/05 18:54 Page 277

U s e R a p i d Te s t i n g t o A d a p t e - L e a r n i n g t o L e a r n e r E x p e r t i s e

277

based on actual measures of learning. Some learners may need many fully worked examples to build a schema. Others will be ready for full practice exercises after only one or two worked examples. e-Learning offers a mechanism for tailoring instruction to each learner. However, for ongoing diagnosis of learning, we need a testing method that can quickly and accurately assess expertise as learners progress. Kalyuga and Sweller (2004, 2005) have recently reported a new rapid testing method that can serve this purpose.

A New Method for Rapid Testing What if, instead of requiring learners to provide complete problem solutions, as in traditional tests, the test only requires them to write down the first step they would take to solve a problem? Because experts have acquired and automated many schemas for solving problems in their domain, they can often jump from a problem statement directly to a solution with few or even no intermediate steps. In contrast, someone with a lower level of expertise might know the general approach to take but will need to invest considerable effort to write out several intermediate steps to solve the same problem. Of course, someone who is a novice may not even attempt the problem. Or he would write out an incorrect step. We illustrate this idea in Figure 11.1 with one

Figure 11.1. Alternative First Steps to Solve an Algebra Problem Among Learners of Diverse Experience. Adapted from Kalyuga and Sweller (2004).

Write in the first step that you would normally take when solving this equation: 2(3X-1) = 1 Possible Responses: Incorrect answer I don’t know

Novice Learner

2*3X – 2*1 = 1

Intermediate Learner

6X = 3 or X = .5

Expert

c11.qxd 9/29/05 18:54 Page 278

278

Efficiency in Learning

algebra question from a rapid diagnostic test and several possible responses. Test takers are instructed to quickly respond by writing down the first step they would normally take to solve the problem. Experts who are very familiar with these types of equations are likely to draw on their existing schemas and immediately jump to the solution or final stages of solution. Novices are more likely to acknowledge they don’t know or write out something incorrect. Learners with intermediate expertise will respond with a correct, intermediate step leading to the solution. Since learners are only writing out one step to a problem solution, to obtain a comprehensive picture of their range of skills, you would need to include problems at varying levels of complexity. For example, in Figure 11.2 we show four levels of algebraic problems along with their associated solution rules. If you work out the sample we label “advanced,” you will see that its solution incorporates all of the rules of the problems that precede it. If you were to construct a fast test to monitor which rules your learners had

Figure 11.2. Levels of Algebraic Equation Problems That Incorporate Increasing Numbers of Skills. Adapted from Kalyuga and Sweller (2004).

Levels of Algebraic Problems Level

Example

Skill

Basic

5x = 15

Dividing (or multiplying) both sides of the equation by the same number

Intermediate

4X – 5 = 3

Adding or subtracting the same number to both sides of the equation before dividing or multiplying

Intermediate

3(x+1) = 5

Expanding values in brackets

Advanced

2(2X-5)/3 = 5

Multiplying both sides of the equation by the same number

Note that each higher level incorporates the skills of the previous level

c11.qxd 9/29/05 18:54 Page 279

U s e R a p i d Te s t i n g t o A d a p t e - L e a r n i n g t o L e a r n e r E x p e r t i s e

279

acquired, you could develop a four-question test asking the learners to write out the first step they would take to solve each of these problems. Learners who miss the basic problem will most likely not be able to solve any of the more advanced problems, since they all incorporate the rule in Problem 1. Likewise, learners who solve Problem 1 but miss Problem 2 probably know Rule 1 but not Rule 2. By asking learners to write out the first step only, you can save a lot of testing time. By including several problems that incorporate the range of skills involved, you can infer your learners’ skill levels. For example in Figures 11.3 and 11.4 you can see Questions 1 and 3 from our Excel demonstration e-lesson pretest on the CD. The first question involves a simple addition operation, whereas the third question requires constructing a formula that involves four different variables. Depending on which questions were answered accurately, the learner is branched to the appropriate topic in the lesson. You can try this out on the load managed web-based lesson on our CD.

Figure 11.3. The First of a Three-Item Pretest in an Excel Lesson on the CD.

c11.qxd 9/29/05 18:54 Page 280

280

Efficiency in Learning

Figure 11.4. The Third of a Three-Item Pretest in an Excel Lesson on the CD.

In order for a rapid testing method like this one to be practical, we need to know that it does give us an accurate picture of the learner’s skills. One way to verify the value of such a rapid test is to compare it with test scores from a traditional test measuring the same skills.

WHAT THE RESEARCH SAYS ABOUT HOW RAPID TEST RESULTS COMPARE WITH TRADITIONAL TEST RESULTS Kalyuga and Sweller (2004) gave a group of ninth and tenth graders a traditional test that included a set of twelve algebraic equations with three questions at each of four levels of complexity, corresponding to the four levels shown in Figure 11.2. After completion of this test, the same learners took a second rapid test including twelve different algebraic equations at the same four levels of complexity. On the

c11.qxd 9/29/05 18:54 Page 281

U s e R a p i d Te s t i n g t o A d a p t e - L e a r n i n g t o L e a r n e r E x p e r t i s e

281

rapid test, the learners were asked to quickly write out their immediate ﬁrst solution step. You won’t be surprised to learn that the rapid test was completed much faster. In fact, the rapid test was completed in 118 seconds, compared to 574 seconds for the traditional test. This difference was both statistically and practically signiﬁcant, with an effect size of 2.23, which is very high. But the real issue is the extent to which the rapid test gave results similar to the traditional test. In correlating the scores from the rapid test with the traditional test, they found a very high correlation of .92. Recall that correlations range from –1 (very high negative association) to +1(very high positive association). This high degree of correspondence between the two tests tells us that the results from the rapid test are as valid a way of determining a learner’s skill as the traditional test that look much longer. We can conclude that a rapid test can give us virtually as good an assessment of learning as a traditional test, but in a much shorter time period. This kind of test offers us a practical way to dynamically assess skills prior to the lesson as well as progress during learning.

Rapid Tests to Adapt Instruction to Learner Expertise Adaptive online training can be used to tailor instructional content and methods to each learner’s level of expertise as they start a course and as they progress through the lessons. Such a program would use results from a rapid pretest to assign learners to the appropriate lesson or topic within a lesson. As the learner progresses through the training in a lesson or lesson topic, the instructional materials are tailored based on her learning. For example, suppose after taking a lesson pretest, Marcy is assigned to Topic B since she demonstrated an understanding of the prerequisite skill in Topic A. Marcy receives some instruction on Topic B followed by a rapid diagnostic test. If she passes the diagnostic test, she moves on to Topic C. Otherwise, she receives additional training on Topic B and repeats the testingtraining cycle until she demonstrates competency in Topic B. In this way, each learner moves through the lessons at his or her own pace and progresses only after demonstrating competency in each skill. To consider a more specific adaptive scenario, our demonstration asynchronous Excel lesson on formulas included three topics on how to construct

c11.qxd 9/29/05 18:54 Page 282

282

Efficiency in Learning

formulas for addition and subtraction, multiplication and division, and a combination of both. Three rapid pretest items like the ones shown in Figures 11.3 and 11.4 are used to assign the learner to the appropriate topic in the lesson. In that topic the learner views a worked example and one or two completion examples, ending with a full problem assignment. At that point the learners are tested again using an item like the one shown in Figure 11.5. If they respond correctly, they move to the next topic. Otherwise, they are assigned additional worked examples. You can try out our adaptive instruction when you view the asynchronous Excel lesson example on the CD. Dynamic rapid testing sounds good, but we need evidence that it works as intended. In the next section, we summarize very recent research reported by Kalyuga and Sweller (2005) that compared the learning from this type of adaptive training to learning of a comparison group that did not receive adaptive training.

Figure 11.5. A Diagnostic Test Given After Completion of Topic 1. From the CD accompanying this book.

c11.qxd 9/29/05 18:54 Page 283

U s e R a p i d Te s t i n g t o A d a p t e - L e a r n i n g t o L e a r n e r E x p e r t i s e

283

WHAT THE RESEARCH SAYS ABOUT A D A P T I V E E - L EA RN IN G BA S ED ON RAPID A SSESSMENT Figure 11.6 illustrates a high-level overview of the adaptive training plan tested by Kalyuga and Sweller (2005). The lessons were designed to teach the algebraic skills summarized in Figure 11.2. As you can see in Figure 11.6, the learners assigned to the adaptive training group start with a rapid pretest. Based on their pretest scores, they are assigned to lesson topics Level 1 through Level 4. Figure 11.7 summarizes the instructional methods used at each level. The basic instruction at Level 1 included two worked examples. However, in the basic instruction at Levels 2 and 3, the worked examples were faded and presented as completion examples. For example, if you started at Level 2 you would receive a faded worked example in which the ﬁrst step, which reﬂects the new skill in Level 2, is already worked and you would complete the second step based on your prerequisite knowledge. Level 4 included full problem assignments. A novice who started at Level 1 would

Figure 11.6. An Overview of an Adaptive Testing Learning Sequence.

Level 2 Rapid Pretest Level 3

Level 4

Rapid Post-Test

Rapid Diagnostic Test

Level 1

More Training More Training More Training

c11.qxd 9/29/05 18:54 Page 284

284

Efficiency in Learning

Figure 11.7. A Summary of the Instruction Methods at Each Level of an Adaptive Training Plan. Level

Basic Instruction

Additional Instruction

1.

2 fully worked examples (WE) of Level 1 equations, each followed by a problem to solve.

2 full WE

4 short WE

2.

2 faded WE of Level 2 equations where learners fill in the last step, each followed by a problem to solve

2 full WE

4 short WE

3.

2 faded WE of Level 3 equations where learners fill in the last two steps, each followed by a problem to solve

2 full WE 4 short WE

4.

Two problems at Level 4

begin with fully worked examples and move through progressively faded completion examples, ﬁnishing at Level 4 with full problem assignments. A learner only progresses to the next level after demonstrating mastery of her current stage on a rapid diagnostic test. If the learner fails to demonstrate learning on the rapid diagnostic test, she is branched to one of two levels of additional training that includes more worked examples. After completing the additional training, she retakes the rapid diagnostic test. When her solutions on the rapid test showed she had acquired the relevant rule, she was branched to the next level. Once participants completed Level 4, they took a rapid post-test with different items but a similar structure to the pretest. Kalyuga and Sweller (2005) compared learning from the adaptive training with learning from a comparison group. To ensure that learners in the comparison group received the same amount of instruction as individuals in the adaptive group, learners in the comparison group were randomly paired with an individual in the adapted group and received the same instruction as their partners. In that way, both groups of learners were exposed to the same amount of instruction. However,

c11.qxd 9/29/05 18:54 Page 285

U s e R a p i d Te s t i n g t o A d a p t e - L e a r n i n g t o L e a r n e r E x p e r t i s e

285

individuals in the comparison group received their instructional assignments independently of their actual learning. These individuals did not take any diagnostic tests and completed the pretest and post-test only. Kalyuga and Sweller (2005) found that the average pretest/post-test gain was 2.27 for the adaptive group and .67 for the comparison group. This difference was both statistically and practically signiﬁcant, with an effect size of .55, which is moderate.

In this research, the highly structured domain of algebraic equations allowed the researchers to cleanly develop rapid assessments and instructional treatments with faded worked examples matched to learner expertise. Our Excel formula lesson likewise involved sufficient structure in topics to permit us to easily apply the adaptive plan. We will need further research to determine how to best adapt this methodology to less structured learning goals. Since the rapid testing method is very new, we will need more research, including experiments and field trials, before recommending it for all types of instructional goals. In the next paragraphs, we discuss the benefits and drawbacks of adaptive learning.

Applying Rapid Assessment to Your e-Learning From a practical perspective, adaptive testing always requires a greater investment of resources than alternative designs that incorporate the same learning methods for everyone. First, the diagnostic tests must be built. Second, the additional instruction must be built for those learners who do not reach competency after completing the basic instruction. The potential benefits of pretests are efficiency gains for trainee populations with diverse prior knowledge. The pretest will allow learners to bypass training topics that they already know. If you have a sufficiently large population of learners with diverse skill sets, you can likely make a strong economic case for diagnostic pretesting. The potential benefits of dynamic in-lesson assessment are efficiency in learning for faster learners as well as confidence in equivalent skill achievement among learners requiring different levels of instructional support. The

c11.qxd 9/29/05 18:54 Page 286

286

Efficiency in Learning

dynamic in-lesson test will ensure that learners do not progress until they have mastered prerequisite skills. If you have a trainee population with sufficiently diverse instructional support requirements and it’s critical that all learners reach a similar performance level, you can likely make an economic case for dynamic assessment at the topic or lesson level. Naturally, the more fine-grained your adaptive learning, the more development work is involved. If you work at the lesson level, your diagnostic test items will be fewer than if you work at the topic level. Since e-learning lessons should be relatively short, you might gain most cost benefit at the lesson level.

••• The Bottom Line The research is quite recent on rapid testing. However, based on results to date we conclude that: • Rapid testing can give as valid an assessment of learning as full tests. • Greater efficiency in training can be promoted by tailoring instructional methods and content to diverse levels of learner expertise. • Rapid testing techniques may be applied to pretests and/or internal diagnostic tests. • You should consider adaptive instruction when: – The instruction will be delivered via asynchronous e-learning, – You have a large audience of widely varied skill levels for whom significant time will be saved by pretesting, and/or – You have a large audience with varied instructional support requirements, all of whom must reach a common level of a critical competency. Depending on your situation, you may want to use the rapid pretest only, the rapid diagnostic test only, or a combination of both. As you can see, the development of the test items and the extra instructional support will add considerably to the cost of the courseware and should be undertaken only

c11.qxd 9/29/05 18:54 Page 287

U s e R a p i d Te s t i n g t o A d a p t e - L e a r n i n g t o L e a r n e r E x p e r t i s e

287

when you anticipate sufficient return on investment. Whether using diagnostic assessment or not, you should consider the expertise reversal effect when designing more advanced lessons.

Tips for Applying Adaptive Instruction • If you are delivering training in instructor-led settings (classroom or virtual classroom), consider using rapid pretests to assign learners to relevant training sessions. However, dynamic adaptive instruction is not usually practical. • Is your audience relatively heterogeneous in relevant experience? If not, do not invest in rapid pretesting; if so, determine whether learning efﬁciency gained by adaptive assignment will warrant the cost to develop pretests. • Is your audience relatively heterogeneous regarding the amount of instructional support they are likely to need to achieve the instructional objective? If not, do not invest in dynamic adaptive learning; if so, is there sufﬁcient savings in time to justify the development costs? If so, then consider using rapid tests as part of a dynamic adaptive training program. • Is it critical that all learners achieve similar competency levels? If so, then consider using rapid tests as part of a dynamic adaptive training program. • Is your content relatively structured? If so, then it can likely be adapted to rapid testing techniques. If not, then you should wait for further research demonstrating the best ways to create rapid tests in less structured domains.

On the CD John Sweller Video Interview Chapter 11: Use Rapid Testing to Adapt e-Learning to Learner Expertise. Topics include an introduction to rapid testing, an example of a rapid test, a discussion of potential applications of rapid tests, as well as learner control.

Sample Excel e-Lessons 1. The asynchronous Load-Managed Excel Web-Based Lesson demonstrates our adaptation of rapid testing to a lesson on construction of formulas in Excel.

c11.qxd 9/29/05 18:54 Page 288

288

Efficiency in Learning

The lesson begins with a three-item rapid pretest. Each item corresponds to a topic on formula formatting in the lesson. Each lesson topic ends with a rapid diagnostic test. If the response is incorrect, the learner receives additional completion examples and another rapid test item. 2. The synchronous Virtual Classroom Example mentions (but does not show) a pretest used to assign learners to individual class sessions. COMING NEXT In Parts II and III, we summarized a number of proven techniques for reducing extraneous cognitive load and increasing germane load for novice and for experienced learners. In Part IV we have seen that, when it comes to experienced learners, many of the techniques that improve learning and learning efficiency in novice learners either have no effect or, in some cases, actually have a negative effect. In the last part of the book we include two chapters that summarize and integrate all of the guidelines we have generated throughout the book. Chapter 12 organizes cognitive load management guidelines into an instructional application context. Chapter 12 won’t present any new research or guidelines. It will integrate the guidelines we’ve discussed throughout the book in a context appropriate for individuals responsible for design, development, evaluation, and delivery of training products.

Recommended Readings Kalyuga, S., & Sweller, J. (2005). Rapid dynamic assessment of expertise to improve the efficiency of adaptive e-learning. Educational Technology Research and Development, 53, 83–93. Kalyuga, S., & Sweller, J. (2004). Measuring knowledge to optimize cognitive load factors during instruction. Journal of Educational Psychology, 96, 558–568.

c12.qxd 9/29/05 18:55 Page 289

PA R T F I V E

Cognitive Load Theory in Perspective

N PART V WE OFFER two quite different opportunities to inte-

I

grate all of the cognitive load theory principles that we have presented throughout the book. In Chapter 12 we discuss how and when you can apply cognitive load principles in the context of your instructional decisions as a training facilitator or designer. Specifically, we consider how cognitive load theory applies to your instructional planning, development of materials, and implementation and evaluation activities. In Chapter 13 you have a different opportunity to integrate the principles in the book through a personal account by John Sweller of how cognitive load theory originated, developed, and evolved over the past twenty-five years.

289

c12.qxd 9/29/05 18:55 Page 290

290

Efficiency in Learning

Read

To Find Out How

Chapter 12. Applying Cognitive Load Theory

What you discover during instructional planning regarding content, target audience, and delivery media will shape your application of cognitive load principles To best use visuals, audio, text, examples, and practice during development of your training To apply cognitive load guidelines during implementation of training by instructors and in training handouts To evaluate your instruction for efficiency

Chapter 13. The Evolution of Cognitive Load Theory

Cognitive load theory originated and evolved

On the CD Video Interview with John Sweller: Chapter Preview/Review Chapter 12: Applying CLT. An overview of the applications of cognitive load theory

Sample Excel e-Lessons You can see our application of cognitive load theory as a result of our analysis, design, and development decisions in our asynchronous and synchronous demonstration e-lessons:

c12.qxd 9/29/05 18:55 Page 291

C o g n i t i v e L o a d T h e o r y i n Pe r s p e c t i v e

291

Before: Overloaded Excel Web-Based Lesson. This asynchronous elearning sample violates most of the guidelines for reducing extraneous cognitive load discussed in this part of the book. After: Load-Managed Excel Web-Based Lesson. This asynchronous elearning sample applies most of the guidelines for reducing extraneous cognitive load discussed in this part of the book. Virtual Classroom Example. This synchronous e-learning sample applies most of the guidelines for reducing extraneous cognitive load discussed in this part of the book.

c12.qxd 9/29/05 18:55 Page 292

CHAPTER OUTLINE Applying Cognitive Load Theory to Instructional Planning Your Target Audience The Target Audience for the Excel Course Your Content Content for the Excel Course Your Delivery Media Delivery Media for the Excel Course

Training Development and Cognitive Load Theory Start with Visuals and Performance Aids Develop Explanations of Visuals and Performance Aids Develop Examples and Practice for Major Tasks Develop Supporting Knowledge Topics

Challenges Implementing Cognitive Load Theory with Many Authoring Software Packages Violations of Redundancy Limitations of Cueing Options Lack of Technology for Adaptive Learning

Applying Cognitive Load Theory to Training Implementation Instructor Implementation of Cognitive Load Techniques Learner Support in Asynchronous Training Training Handouts

Evaluating Courseware for Efﬁciency The Bottom Line On the CD John Sweller Video Interview Sample Excel e-Lessons

c12.qxd 9/29/05 18:55 Page 293

12

Design

Develop Cognitive Load Theory

Applying Cognitive Load Theory

Implement

Evaluate

T

HE GOAL OF THIS CHAPTER is to integrate the various guidelines

we have described for efﬁcient learning environments into a context that is relevant to your work as an instructional professional. We assume basic familiarity with course design processes and will emphasize primarily the speciﬁc aspects of planning, development, implementation, and evaluation of training that are most related to management of cognitive load. We will illustrate much of our discussion with the Excel demonstration lessons on the CD.

293

c12.qxd 9/29/05 18:55 Page 294

294

Efficiency in Learning

Applying Cognitive Load Theory to Instructional Planning Throughout the book we have summarized guidelines and supporting evidence for managing cognitive load in your instructional programs. In Part II we looked at ways to use instructional strategies to reduce extraneous load imposed on working memory and to manage intrinsic cognitive load resulting from your content complexity. In Part III, we summarized ways you should exploit available working memory capacity to help learners build relevant schemas with techniques that impose germane load. In Part IV we discussed the adjustments you need to make based on the expertise of your learners. In this summary chapter, we synthesize the guidelines presented previously to help you transfer our recommendations into your instructional activities. This chapter is relevant to readers who are responsible for training design and development or facilitation of classroom or synchronous e-learning, as well as for individuals who are reviewing off-the-shelf courseware. When you first receive your instructional assignment, take some time to evaluate these three factors: the relevant background of your target audience; the complexity of your content; and the media mix that will be used to deliver your training. To minimize split attention while reading this chapter, we recommend that you print a copy of Table 12.1, which is on the accompanying CD. We will illustrate our commentary on the three factors using our Excel demonstration courses shown on the CD.

Your Target Audience We have seen in Part IV of this book that instructional methods that are effective for novice learners don’t help—and sometimes even hinder experienced learners. So one important first step will be to research and define the backgrounds of the intended learners. Sometimes, such as when you are teaching a completely new computer application, everyone will be new to the content. In other situations, such as when you are teaching an advanced skill, just about everyone will have some relevant experience. Often however, your audience includes a mixture of novice and experienced learners.

c12.qxd 9/29/05 18:55 Page 295

295

Applying Cognitive Load Theor y

Table 12.1. Applying Cognitive Load Theory to Your Training. Factor

What to Do

Ways to Do It

Audience Prior Knowledge

If prior knowledge is low, minimize extraneous cognitive load by applying instructional methods that compensate for lack of schemas

• Provide performance aids • Use relevant visuals • Explain visuals with audio or integrated text—not both • Avoid unnecessary explanations o Don’t add words to self-explanatory visual o Don’t explain visuals with both audio and text o Write concisely o Avoid irrelevant themes • Use directive lesson designs • Use faded worked examples • Write considerate text

If prior knowledge is high, avoid instructional methods that are redundant with respect to learner schemas

• Do not add unnecessary explanations in the form of visuals or text • Assign practice exercises rather than worked examples or completion examples • Do not invest extra effort in high coherent text • Use directive or guided discovery lesson designs

If your audience is a mixture of low and high prior knowledge, accommodate differences when practical

• Use adaptive instruction in e-learning • Use prework to equalize understanding • Enforce prerequisites • Use pretests to assign learners to appropriate segments • Divide sessions into introductory and advanced (Continued)

c12.qxd 9/29/05 18:55 Page 296

296

Efficiency in Learning

Table 12.1. Applying Cognitive Load Theory to Your Training (Continued). Factor

What to Do

Ways to Do It

Content Complexity

If complexity is high, artiﬁcially reduce intrinsic cognitive load by segmenting supporting knowledge from major lesson content. Most importantly, ensure that extraneous cognitive load is maximally reduced

• Separate concepts and facts from process stages or task steps • Teach these concepts and facts in a separate section from the stages or steps • Consider each of the techniques designed to reduce extraneous cognitive load

Content Type

For near transfer content, build and automate schemas that will apply to the job

• Use worked examples and completion examples that incorporate the work environment • Use drill and practice in the form of mental or physical rehearsal if automatic responses are needed on the job • Provide performance aids if automaticity is not needed on the job

For far transfer content, build ﬂexible schema that will apply to diverse job environments and ensure those schemas are automated

• Use worked examples, completion examples, and practice that incorporate diverse work contexts • Encourage processing of diverse worked examples through self-explanations • Provide sufﬁcient learning time to automate schemas

For computer training, build instruc- • Incorporate all explanations into the computer tional environments that avoid split • Avoid displaying instruction in attention and redundancy a manual and on a computer, which leads to split attention or redundancy

c12.qxd 9/29/05 18:55 Page 297

297

Applying Cognitive Load Theor y

Table 12.1. (Continued). Factor

What to Do

Ways to Do It

Media

For instructionally paced delivery media such as classrooms, virtual classrooms, or video, minimize extraneous cognitive load

• Apply the guidelines listed above for reducing extraneous cognitive load

For learner-paced delivery media such as asynchronous e-learning or print media, less cognitive load is imposed and therefore you need not apply as many cognitive load guidelines

• For asynchronous e-learning, design the program to provide brief content sequences that are under navigational control of the learner to replay, pause, and move forward

When your learners are novice, they won’t have schemas relevant to the new knowledge and skills they need to acquire. Therefore, you will need to provide schema-substitutes as a support until learners have more of a psychological basis for standing on their own. These supports are all of the methods we described in Part II for reducing extraneous cognitive load. However, if your learners are more experienced or will become more experienced in a lengthy training program, then you should abandon most of these methods. With more experience, your audience has relevant schema and the instructional supports that aid novices will in fact slow down experienced learners. Refer to specific guidelines in Table 12.1 for methods appropriate for experienced learners. If you have a mixture of low and high experience learners, you have several alternatives for tailoring the training to individual needs, depending on your delivery media and organizational resources. If you are delivering via asynchronous e-learning, you can use pretests and adaptive diagnostic tests to guide learners to introductory and more advanced sections of the course. If you are delivering in the classroom, you can assign prework to help equilibrate prior knowledge before class starts. Alternatively, you can establish and enforce prerequisites. Along with prerequisites, you can organize courses

c12.qxd 9/29/05 18:55 Page 298

298

Efficiency in Learning

into introductory and advanced sections and in that way arrange for a more homogeneous group of participants.

The Target Audience for the Excel Course In our analysis of our Excel audience (financial analysts), we found a great deal of variation in learner background and work requirements. Some learners were experienced with Excel use for some basic functions but had diverse needs for specialized calculations and charts based on new work assignments. Other learners were completely new to Excel. Beyond a need for the basic functions of Excel, the requirement for different Excel applications varied based on the analysts’ work assignments. Therefore, we have a very heterogeneous audience regarding incoming skills and desired outcomes. Over the next three years the total population of analysts needing some Excel training will likely approximate one thousand. Many of the analysts who use Excel are highly paid and would prefer to invest time in acquiring only skills directly relevant to their work needs. The analysts work individually so that group collaboration opportunities available in face-to-face instruction were not a major satisfier or need. In response to these factors, an asynchronous e-learning program offered the best delivery option.

Your Content As you define the content and instructional objectives, consider content complexity, whether the objectives require near or far transfer learning, and whether computer application training is involved. We have solid cognitive load guidelines for these three content factors. For high complexity content, your learners will experience high intrinsic cognitive load. You can artificially reduce this load using the methods described in Chapter 5 and summarized in Table 12.1. The main strategy is to segment the content in ways that supporting knowledge is trained first, followed by major lesson content such as process stages or task steps. Keep in mind that content complexity is not the same as content difficulty. In

c12.qxd 9/29/05 18:55 Page 299

Applying Cognitive Load Theor y

299

cognitive load theory content complexity is a function of the number of elements that must be processed in working memory simultaneously. See Chapter 1 to review the differences between content complexity and difficulty. In some training, the learning objectives are primarily near transfer. That is, workers should apply new skills in a consistent way to a consistent environment. Many procedural tasks such as logging onto email or completing a customer order form are examples. The main challenge to achieving near transfer training success is to duplicate the work environment as closely as possible during the training. For example, if learning to use a new system, provide practice on the same system or a close simulation of it. In other cases your learning objectives will require far transfer performance. In far transfer, learners must build relevant schemas that must later adapt to a variety of different work contexts. Effective sales and coaching skills are two common examples. In far transfer training you will need to provide learners with diverse examples and scenarios. Because the diversity adds cognitive load, it will be very important to use worked and completion examples, as discussed in Chapter 9. In addition, far transfer almost invariably imposes a heavy cognitive load because learners must use working memory resources to work out how learned material applies to a new situation. That cognitive load can be reduced if the learned material is automated, as discussed in Chapter 9. Therefore, if far transfer is the goal, make sure the material is very well learned. Finally, a great many organizational training resources are directed toward teaching use of new technology. When training use of computer systems, embed all of the instruction in the computer in order to avoid split attention. If learners are reading steps in a manual and trying to apply them to an application screen on the computer, split attention results, as discussed in Chapter 4. Likewise, duplicating all of the computer screens and explanations in a manual that requires the use of the computer as well leads to redundancy, as discussed in Chapter 5. Your best bet is to get rid of the manual and incorporate all training onto the computer.

c12.qxd 9/29/05 18:55 Page 300

300

Efficiency in Learning

Content for the Excel Course We judged most of the content in our Excel lesson as moderate to high complexity, in that learners will need to devote memory capacity to several perceptual, physical, and cognitive factors involved in identifying relevant screen components, defining the mathematical goal, translating the goal into proper formula formats, and using the mouse and keyboard to implement the procedure. The Excel course is designed to support analysts who will need to produce various reports and displays to help decision makers evaluate data. It will be important that learners see how Excel procedures can be operationalized for diverse purposes and with diverse data sets. In other words, the outcomes are far transfer. Therefore, we included a variety of Excel scenarios as the basis for worked examples and practice exercises. Of course, Excel is a computer application. To minimize split attention and redundancy, we integrated all of the training onto the computer.

Your Delivery Media An important feature of delivery media relevant to cognitive load theory is the degree to which learners control the pace at which they are exposed to instruction. In a classroom—physical or virtual—the instructor controls the pace and learners must keep up. In contrast, in print or some forms of asynchronous e-learning, learners can control the rate at which they progress through training. They can usually reread and revisit various segments of the instruction as needed. Therefore, delivery media that permit learner control of pacing tend to impose much less cognitive load. If your delivery media will be instructionally paced, then it’s especially important that you apply as many of the guidelines as are practical to minimize extraneous and intrinsic cognitive load. Your delivery media, of course, will also affect how you implement other methods for managing cognitive load. Obviously, in print media you do not have the advantage of audio narration, and therefore you must concentrate on minimizing split attention by placing text close to the information to which it refers, using pointers to reduce the need to search for referents. In asynchronous e-learning you have capabilities to assess learner progress and

c12.qxd 9/29/05 18:55 Page 301

Applying Cognitive Load Theor y

301

tailor instruction to differences in their expertise through adaptive testing, as described in Chapter 11. Cognitive load theory applies to all media. Regardless of media, all lessons designed for novices that teach tasks or procedures should include worked examples, practice exercises, and relevant visuals to display main content. Likewise, all instruction should minimize divided attention, although it may do so in different ways. All instruction should also minimize extraneous cognitive load imposed by unneeded themes, vignettes, audio, technical information, or flabby writing.

Delivery Media for the Excel Course For our Excel demonstration course, a number of factors pointed to asynchronous e-learning as the major delivery medium. The learning audience is quite large and globally located. Likewise, the organization is most interested in maximizing efficiency in instruction of highly paid analysts by tailoring it to individual skill backgrounds and work requirements. However, design and development of asynchronous courseware tends to take longer than virtual classroom versions. This is because the asynchronous courseware must include all relevant feedback to exercises as well as audio narration, both of which are handled by the instructor in the virtual classroom version. In addition, the asynchronous version must be programmed. Since there is an immediate need for training, a virtual classroom training adapted from a classroom course will serve as an intermediate solution. The virtual class will be delivered first, followed by an asynchronous e-learning version to follow at a later date. You can see both the virtual classroom and asynchronous e-lesson samples on our CD. We exploited the capabilities of the asynchronous e-learning by including rapid diagnostic tests that will assign learners to relevant topics in the course and that will diagnose their learning progress and assign the optimal mix of completion examples and practice based on their level of expertise. Although dynamic assessment is not practical in the virtual classroom version, we did use a pretest (not shown in the example) to assign learners to the appropriate virtual sessions.

c12.qxd 9/29/05 18:55 Page 302

302

Efficiency in Learning

In both the e-learning versions, we had the opportunity to use audio to explain relevant visuals. However, in the case of completion examples and practice, audio is too transient and case data and directions are placed in text on the screen. Where text is used, it is placed close by relevant visuals with pointers to avoid split attention. Furthermore, although the default version in the asynchronous lesson is audio, learners can revert to text alone by turning off the audio controls. Both courses use a series of worked and completion examples ending in full practice.

Training Development and Cognitive Load Theory In this section of the chapter, we discuss some tips for developing your training materials in accordance with cognitive load theory.

Start with Visuals and Performance Aids Since many of the cognitive load guidelines at the development phase relate to treatment of visuals, we recommend defining the major visuals to be included in the program as an early step. In addition, since many organizational training programs do not allow sufficient time to practice many skills, we recommend planning your performance aids as another early step. In that way, you can build your instruction around the performance aids and the visuals. As we described in Chapter 3, appropriate visuals can greatly promote learning and increase engagement in the instructional materials. Conversely, inappropriate visuals can add extraneous cognitive load. Examine the major tasks associated with each lesson to define the dominant visuals associated with those tasks as well as ways to most effectively create a performance aid. In our Excel lesson on formulas, a spreadsheet Excel screen capture was the natural predominant visual interface. For more details on appropriate visuals to use for different instructional outcomes, refer to Chapter 3 and to Clark and Lyons (2004).

Develop Explanations of Visuals and Performance Aids As you define your visuals and performance aids, consider whether they will need explanations or whether they can be designed to be self-explanatory,

c12.qxd 9/29/05 18:55 Page 303

Applying Cognitive Load Theor y

303

similar to the airline emergency procedure visuals that use arrows to communicate motion. Remember that to add words to a self-explanatory visual results in redundancy that slows learning. In most cases, you will need to explain your visuals and performance aids, and we recommend using audio narration to do so in order to leverage the modality effect. As you draft the visuals and performance aids, you can also write the scripts to be used by instructors or by narrators in asynchronous e-learning. The format and wording of the script will vary depending on your delivery medium. For example, a classroom instructor with subject-matter expertise might need only bulleted text points to prompt her explanations. On the other hand, a professional narrator will need detailed scripts when creating audio for asynchronous e-learning. Regardless of media, make every word in your scripts count. The result should be just enough words to communicate the content, as discussed in Chapter 5. If you are using asynchronous e-learning delivery, you may not have software or hardware to deliver audio narration. Even if you are using audio narration, you may also need to provide a text option for a subset of learners lacking audio delivery capabilities or who have hearing loss. You will then need to write descriptions in lean text and place them close to the visual being explained. In addition, there are some places in the lesson where you should not use audio because it is too transient and learners need an opportunity to review the words over a period of time. For example, all directions for exercises or completion examples should be delivered in text only. Likewise, any data needed for case studies and feedback to exercises should remain visible for the learner to review.

Develop Examples and Practice for Major Tasks As you are working out the visuals, performance aids, and scripts to teach the major tasks, plan a series of worked examples and practice exercises. In general, we recommend starting with a fully worked example that is explained with instructor or audio narration. Next present faded worked examples in which the learners are required to fill in one or more steps. For the faded worked examples, present all words in on-screen text. End the sequence with a full practice exercise presented in text. If learners will

c12.qxd 9/29/05 18:55 Page 304

304

Efficiency in Learning

need to apply the skills to different scenarios later, on the job, be sure to incorporate different surface features into your examples and practice exercises, as discussed in Chapter 9. For example, when practicing constructing formulas in Excel, we provide examples and exercises that use profit, payroll, and sales projections. The different scenarios add germane cognitive load that is offset by using faded worked example progressions, as described in Chapters 8 and 9.

Develop Supporting Knowledge Topics Many tasks will have associated supporting knowledge, such as processes and concepts. The supporting knowledge should be sequenced in the lesson either before or after the major lesson task to manage the amount of new content presented to learners all at once. For example, the task of constructing a formula in Excel requires supporting knowledge of what is a valid format for an Excel formula. In the virtual classroom lesson, this concept was taught prior to the worked examples of how to input a formula in a spreadsheet. To make the lesson more interactive, the instructor displayed several examples of formulas and asked the learners to define some of the features associated with formulas. Learners input their responses using the chat facility. Using frequent questions that require learners to study examples and form conclusions is important to maintain a high level of interactivity in e-learning environments.

Challenges Implementing Cognitive Load Theory with Many Authoring Software Packages As you implement our research findings and guidelines, you will find that software packages that are currently used to create asynchronous courses— such as web development tools and even specialized courseware authoring tools—do not support cognitive load guidelines by default. Even worse, some cannot be configured to support principles to manage cognitive load at all. In general, most courseware authoring programs currently available on the market violate the redundancy principle, do not always provide appropriate visual cues to direct attention, and do not support diagnostic testing.

c12.qxd 9/29/05 18:55 Page 305

Applying Cognitive Load Theor y

305

Violations of Redundancy Training on how to perform procedures in software is quite commonly delivered as asynchronous e-learning. Software simulations are the most common method used for instruction and practice in such courses. Tools used to develop these simulations are quite robust, often offering lesson-user interactions, data tracking, and even audio. However, the vast majority of software simulation packages violate the redundancy principles by providing a visual of a software screen explained by on-screen text and audio narration. As a result, the student may experience cognitive overload with the redundant text and audio in the presence of a complex visual. To implement methods such as those used in our Excel asynchronous e-lesson demonstration, such software simulation tools should ideally offer facilities to explain software screen visuals with audio narration by default. In addition, in the event that the student does not have audio capabilities on her computer or the language in the course may not be her native language, the software simulation should provide the student with the ability to turn off the audio. With the sophistication of current learning management systems (LMS) and courseware, the course could even be launched with audio on or off, depending on the student’s personal preference configured in the LMS.

Limitations of Cueing Options In the event that the user turns off audio narration in his course and opts to read on-screen text instead, cognitive load theory suggests that the course developer should integrate the text with appropriate locations on the courseware’s screen to minimize split attention. As shown in Figure 12.1, we used red lines to connect on-screen text to the corresponding location in the spreadsheet. While most courseware development applications do provide various tools to draw the student’s attention to different locations on the screen, visual cues are often limited to bold lines or circles, on-screen arrows, or even animated lines that can lead to split attention.

Lack of Technology for Adaptive Learning Implementation of diagnostic assessment for adaptive e-learning is generally not supported by authoring software. In our Excel demonstration we

c12.qxd 9/29/05 18:55 Page 306

306

Efficiency in Learning

Figure 12.1. An Asynchronous Course with Lines Used to Integrate Text.

begin with a rapid assessment in order to assign learners to just the topics they need. Unfortunately, some practice and assessment software packages on the market do not provide this dynamic, prescriptive ability. Therefore, extensive technical coding ability is required to program it. In addition, judging of learner responses to questions and assessments is very inflexible. For instance, if the correct answer to a question is The Sky Is Blue, and the student enters the sky is blue or simply blue, the assessment software may judge the answer as incorrect unless it is explicitly told to ignore the case of the text and is also programmed with all the possible answers that the student may answer. Although the recent research on rapid assessment has used open-ended responses to math problems, we suggest that multiplechoice items may be a more practical format, given the constraints of the software packages. These limitations in authoring software present some difficulties in implementing cognitive load principles to your asynchronous courseware today.

c12.qxd 9/29/05 18:55 Page 307

Applying Cognitive Load Theor y

307

In time, however, the effort and cost to incorporate these guidelines in your content should decrease as vendors who provide software to create courseware incorporate the facilities to readily create training that reduces extraneous cognitive load.

Applying Cognitive Load Theory to Training Implementation Implementation involves the deployment of courseware to the learners. For instructor-led courses, implementation requires training and supporting instructors to deliver the instruction in the most efficient manner. Implementation of asynchronous e-learning from a cognitive load perspective requires ensuring that learners have structured opportunities and distractionfree environments for learning.

Instructor Implementation of Cognitive Load Techniques Since the use of visuals, text, and audio are fundamental to learning efficiency, the instructor plays a major role in applying cognitive load theory in classroom or synchronous e-learning environments. We recommend that all instructors be trained in basic cognitive load theory principles so they understand how to best utilize the limited capabilities of learners’ working memory during training. For example, instructors should provide succinct explanations of visuals displayed on slides. For content displayed in text, instructors should display slides and pause while learners read them. Instructors should avoid redundancy caused by reading onscreen text—especially when there is a visual on the slide. Instructors provide germane load through explanations of worked examples, by asking learners to complete examples, and by calling on one or more participants for their answers. They can also promote self-explanations of worked examples by asking learners to list the principles or rules illustrated by various aspects of the example and by calling on one or more participants to give their explanations. Instructors are also responsible for gaining and sustaining attention during learning. They do so by eliminating

c12.qxd 9/29/05 18:55 Page 308

308

Efficiency in Learning

distractions of extraneous events, stories, and audio during explanations, and by frequent questions and other interactive assignments relevant to the instructional goal.

Learner Support in Asynchronous Training In many situations learners complete asynchronous training at their work stations using their production computers during breaks or lulls in their normal work routine. Depending on the work environment, this arrangement can reduce efficiency due to (1) production platforms that do not support audio; (2) distractions and interruptions in the work setting; and (3) overreliance on workers to initiate and complete training in an environment of conflicting priorities. If any of this sounds familiar, consider an investment in regional learning centers that are located in a separate area from the work environment and that include platforms that support efficient courseware. Managers should work closely with their workers regarding expectations and progress in training no matter what the delivery mode. However, manager support in asynchronous learning environments is especially important to ensure that learners are engaged in courseware that is appropriate to their background and assignments and are making progress in learning and applying new skills.

Training Handouts As you plan your instructional handouts, you need to consider the risks of redundancy. It’s common practice to hand out a copy of the slides used in an instructional presentation on which learners take notes. This may not be an efficient approach for two reasons. First, the visuals are a redundant expression of what learners are seeing in projected slides. Second, taking notes during an instructional presentation becomes a divided attention task. However, learners will need working aids to apply new skills on the job after training. To best apply cognitive load theory, we recommend that handouts be provided after the instructional presentation. This frees working memory to focus solely on learning from the instruction. Following the lesson, reference materials should be provided. For example, in our synchronous Excel lesson, note that exercise directions and memory support are included on the instructional screens, as illustrated in Figure 12.2. Learners do not need to

c12.qxd 9/29/05 18:55 Page 309

Applying Cognitive Load Theor y

309

Figure 12.2. The Screen Used in a Synchronous Course Includes On-Screen Memory Support.

refer to any additional materials during the lesson. For later use, a reference document is provided that includes the definitions and examples of formulas and the steps to enter a formula. For task reference support, we recommend a visual of the interface with embedded steps, like the one in Figure 12.3.

Evaluating Courseware for Efﬁciency You may be more interested in evaluation of courseware than in the processes involved in producing it. Before beginning any evaluation, you will need to consider the three factors we summarized in Figure 12.1. Begin with a planning process (also called a training needs assessment) to (1) define the audience background experience; (2) classify the type and complexity of the content involved; and (3) determine the most efficient mix of delivery media. Table 12.1 summarizes guidelines to consider for each of these three criteria. You can create an expanded list of quality assurance factors by building on

c12.qxd 9/29/05 18:55 Page 310

310

Efficiency in Learning

Figure 12.3. A Handout That Summarizes Procedural Steps.

each of these guidelines. To make them more useful, you can illustrate each guideline with relevant examples and non-examples from your own training. Such a list can serve as an evaluation vehicle when reviewing off-the-shelf or custom-developed software, as well as a quality guide for internal developers.

••• The Bottom Line In this chapter we have integrated the many guidelines discussed throughout our book into a context relevant to instructional professionals. Overall you will need to apply appropriate guidelines based on: • The background knowledge of your target audience • The complexity, transfer requirements, and computer involvement of your content • The mix of delivery media used in the training

c12.qxd 9/29/05 18:55 Page 311

Applying Cognitive Load Theor y

311

On the CD John Sweller Video Interview Chapter 12: Applying Cognitive Load Theory. John discusses applications of cognitive load theory

Sample Excel e-Lessons The asynchronous Load Managed Excel Web Based Learning and the synchronous Virtual Classroom Example both illustrate most of the principles discussed in this chapter. The Overloaded Excel Web-Based Lesson violates most of the guidelines pertaining to extraneous cognitive load. You can see a critique of this lesson by selecting Listen to Dr. John Sweller’s Commentary.

COMING NEXT In this chapter we integrated many of our guidelines into a context in which instructional professionals will apply them. In the next and final chapter, written by John Sweller, you have the opportunity to review the principles and research behind cognitive load theory from his personal account of how it started and has evolved over the past twenty-five years.

c13.qxd 9/29/05 18:58 Page 312

CHAPTER OUTLINE Origins Early Years—Problems with Problem Solving Middle Years—The Importance of Failed Experiments Recent Years—The Internationalization of Cognitive Load Theory Current Work Measuring Expertise Imagination (Mental Rehearsal) Effect Evolution of Human Cognitive Architecture

Conclusions

c13.qxd 9/29/05 18:58 Page 313

13 al

Journ

The Evolution of Cognitive Load Theory

John

er

Swell

A PERSONAL PERSPECTIVE John Sweller

I

N THIS CHAPTER, we will look at cognitive load theory’s origins, evolu-

tion, and prospects. Today, the theory is a major governing framework for instructional professionals, but surprisingly, the origins of the theory lie not in instructional work, but rather in some of my early research into the psychology of problem solving. At the time, instruction was not on my or my research students’ radar. Instruction only became preeminent when we saw the results of some of our early experiments. Those results led to a twenty-ﬁve-year program of research on the consequences of human cognitive architecture for instruction. That program is now global in scale. It has not only led to the procedures discussed in this book, but it is now beginning to throw light on the origins and structures of human cognitive architecture itself.

313

c13.qxd 9/29/05 18:58 Page 314

314

Efficiency in Learning

Origins The origins of cognitive load theory can be found in the results of a failed experiment. In the late 1970s, my research students and I were studying how people learned while solving problems (see Sweller, Mawer, & Howe, 1982). We were giving UNSW educational psychology undergraduates number transformation problems in which they were given a start number that had to be transformed into a goal number by finding the right sequence of moves. The only moves permitted were multiplying by 3 or subtracting 69. Problem solvers could use each of these moves as many times as they needed and in any order until they reached the goal number. For example, problem solvers had to convert 31 into 3 by successively multiplying by 3 or subtracting 69. (The answer is 3, –69, 3, –69.) Each move could be made just by pressing a key on a computer keyboard, so no mental arithmetic was involved. There were many problems, but each problem could be solved only by alternating the two possible moves a certain number of times. The aim of the experiment was to see what factors would assist problem solvers in learning the rule, but we faced an immediate and inexplicable difficulty. As we sat there watching the participants in the experiment solving their series of problems, two things stood out. First, the problems were not very difficult, with most people solving them relatively quickly. Second, despite easily solving the problems, and despite the fact that successfully solving a problem meant that the alternation rule had been followed because alternating the two possible moves was the only way the goal could be reached, very few people were aware of the rule. They could solve up to sixteen problems, with most problem solvers remaining quite oblivious to the fact that they had alternated the two moves to solve every single problem. Looking at how people learned while solving problems was proving a quite futile exercise because all our evidence suggested that most of the participants in our experiment were learning very little. Either we had stumbled on a spectacularly dense set of experimental participants, a conclusion I was reluctant to accept because they had already demonstrated considerable sagacity by choosing to study educational psychology with me, or alternatively, everyone has difficulty learning while problem solving. If learning and problem solving are incompatible, that incompatibility suggested we needed a new view of problem solving, because the field was

c13.qxd 9/29/05 18:58 Page 315

The Evolution of Cognitive Load Theor y

315

embarking on a long excursion in which learning through problem solving was a basic assumption. The publication of Newell and Simon’s (1972) book on problem solving inaugurated a flowering of the field. By studying people’s cognitive processes while solving a problem and with the help of artificial intelligence programs, Newell and Simon were able to uncover some of the important mechanisms of problem solving. For about twenty years, problem solving became one of the central fields of cognitive psychology, with a large number of researchers engaged in studying the mental processes we engage in when solving a problem. Furthermore, since problem solving was central to education and training, descriptions by researchers of mental processes while students were solving a problem tended to be associated with advice from the researcher on how to design instruction. Most of that advice implicitly assumed that problem solving in one form or another was also the best form of learning. While it is now quite clear that this advice was misguided, it is instructive to consider why such erroneous views could have been so influential. The proponents of the learning through problem solving recommendation collected a vast amount of data on the processes used by problem solvers. They avoided running controlled experiments in which learning was compared under problem solving as opposed to alternative conditions. No instructional recommendation should ever be accepted without that recommendation having been tested using controlled experiments in which the recommended new procedure is compared to currently used alternatives. Cognitive load theory has proved successful not only because of its reliance on a particular view of human cognition but also because no instructional recommendation generated by the theory has been offered without first being extensively tested using controlled experiments. All of the recommendations found in this book fall into this category. We should never accept instructional procedures that have not been tested in this fashion. Without controlled experimental testing, we face an unending list of instructional fads.

Early Years—Problems with Problem Solving The failure of our problem solvers to learn anything useful about the structure of the problems we presented to them immediately led to the obvious question: How should we have taught them? In the case of the puzzle problems we were using, the answer was as obvious as the question. If you show people

c13.qxd 9/29/05 18:58 Page 316

316

Efficiency in Learning

the alternating rule (3, 69) rather than have them attempt to discover it, they will learn it immediately. As confirmed by both commonsense and controlled experiments, while the alternating rule may be hard to discover, it is trivially easy to learn. Why is the rule hard to discover? While our language has altered somewhat over the years, from a cognitive load theory perspective, the explanation is straightforward. Solving a problem by searching for a solution places very heavy demands on working memory capacity. Those demands may result in a successful problem solution, but they do not leave sufficient working memory capacity to note which moves are appropriate for particular problem situations. Specifically, in the case of the numerical problems requiring alternation of the problem-solving moves, problem solvers devoted all of their working memory resources to working out which move was best at each point. They had no working memory capacity remaining to attend to the relations between moves. Once they made a move, they ignored both that move and all preceding moves and devoted all of their resources to working out what to do next. Without attending to the relation between moves, it is not possible to discover the alternation rule. The result was a successful solution with no realization that an alternating sequence of moves had been used. While this result clearly had implications for the psychology of problem solving, in isolation, it could have no consequences for instruction. No findings using puzzles could have instructional implications. At this point, it was possible to either continue studying the psychology of problem solving as pure research or attempt to find applications for the finding by branching into instructional design. I decided that the finding had sufficient potential to risk attempting to replicate it using real educational materials rather than puzzle problems. The first step was to find instructional areas where our puzzle problem findings might be relevant. Mathematics and science provided obvious examples. These areas had fairly stereotyped instructional procedures. A new area is introduced, one or two worked examples demonstrating how problems are solved using the new information are presented, followed by a relatively large number of problems for learners to solve. Since students are expected to spend a considerable time solving these problems, it is assumed that considerable learning will occur during this problem-solving process. If our puzzle problem results generalized to mathematics and science, this assumption may not be valid.

c13.qxd 9/29/05 18:58 Page 317

The Evolution of Cognitive Load Theor y

317

We tested two techniques that hypothetically should have been superior to conventional problem solving. The first required people to practice solving goal-free rather than conventional problems. For example, a goal-free geometry problem requires problem solvers to “calculate the value of as many angles as you can,” rather than “calculate a value for Angle X.” When calculating a particular angle, problem solvers must consider what they have now, consider the angle for which they must find a value, and search for a route between them. Working memory load is high. In contrast, when calculating the value of as many angles as possible, problem solvers merely look at their current problem state and calculate a value for any angle they can. Working memory load is substantially decreased. Our results strongly supported this hypothesis, with learners given goal-free problems performing much better on subsequent tests consisting of conventional problems, providing an example of the goal-free effect (Sweller, Mawer, & Ward, 1983). The second technique we tested was the use of worked examples, as indicated in Chapter 8. We taught students how to solve simple algebra problems of the sort: (a b)c d, solve for a. Then we gave a problem group lots of problems to practice while a worked example group were given the same problems with a worked solution for every second problem. On subsequent test problems, the worked example group did better than the problemsolving group, demonstrating the worked example effect (Cooper & Sweller, 1987; Sweller & Cooper, 1985). We suggested the superiority of the worked example group was due to cognitive load because they spent less time studying worked examples than the problem group spent solving the equivalent problems. Paas (1992) and Paas and van Merriënboer (1994) conclusively demonstrated that the effect was due to a reduced cognitive load for the worked example group using subjective ratings of task difficulty.

Middle Years—The Importance of Failed Experiments The obvious next step was to try the effectiveness of worked examples in other areas, such as geometry and physics. We ran exactly the same experiments using geometry and physics problems. To our surprise, the results indicated an utter failure of worked examples, with no evidence of superiority

c13.qxd 9/29/05 18:58 Page 318

318

Efficiency in Learning

over problems. At that time, we had no idea why presenting learners with worked examples in algebra worked so well, while presenting geometry or physics worked examples was no better than solving the equivalent problems. It took considerable thought over several years to figure out why worked examples failed in some areas and, as often happens, the answer was staring us in the face. As instructional professionals, our own working memory load is reduced by simply focusing on a particular instructional procedure such as the use of worked examples, rather than considering the cognitive processes underlying the instructional procedure. Assuming that practice problems should be transformed into worked examples is not cognitively taxing for us. Unfortunately, it is too easy and we need to analyze our instructional materials much more deeply. No simple technique is going to be universally applicable and, according to cognitive load theory, an instructional technique intended to reduce extraneous cognitive load will be effective only insofar as it does just that—reduce extraneous cognitive load. Algebra worked examples were effective, not just because they were worked examples but because they were worked examples that happened to be structured in a way that reduced extraneous cognitive load. Geometry and physics worked examples were ineffective because they were structured in a manner that failed to reduce extraneous cognitive load. We had structured our worked examples in a way that was conventional for their respective areas and, in the case of algebra, that structure coincidentally imposed a minimal, extraneous cognitive load. In the case of geometry and physics, the conventional structure of worked examples imposed a load on working memory that was no less than that imposed by solving problems. As a consequence, it made no difference whether we got people to study our worked examples or solve our problems. How should we structure geometry or physics worked examples? The answer is—in a manner that reduces split attention. Chapters 4 and 8 provide examples. For example, when faced with a diagram and text that are unintelligible in isolation, learners should not have to search for referents. Searching for which parts of a diagram go with which parts of text requires working memory resources that consequently become unavailable for learning. We found that constructing worked examples that reduced or eliminated

c13.qxd 9/29/05 18:58 Page 319

The Evolution of Cognitive Load Theor y

319

that search by physically integrating text into diagrams reinstated the worked example effect in geometry and physics and, in the process, demonstrated the split-attention effect (Sweller, Chandler,Tierney, & Cooper, 1990;Tarmizi & Sweller, 1988; Ward & Sweller, 1990). Eliminating split attention is, of course, important in all instruction, not just worked examples. We had failed to generalize the worked example effect in a variety of areas until we considered cognitive load theory more deeply. That failure allowed us to use cognitive load theory to generate the split-attention effect. In many ways, that sequence of events provides a template for the history of cognitive load theory over the next fifteen years. We repeatedly found a new instructional effect or procedure, found the limits of that effect when attempting to generalize it to new conditions, and used cognitive load theory to generate a new effect. Over many years, we realized that a successful experiment almost never leads to either theoretical or practical advances. In our case, advances seemed to come from failed experiments. Successful experiments did little more than confirm what we already knew. The redundancy effect grew from the split-attention effect in exactly this manner. We had found that learning was facilitated by integrating text with diagrams. We automatically assumed that integrating disparate sources of information would always be successful but, of course, instructional design is never that easy. The cognitive load implications of integrating text and diagrams matters, not the act of integration. It took us a few months to realize that, if text merely re-described a diagram, little was gained by integrating the sources of information. Integration was very successful when both sources of information were unintelligible in isolation and so had to be integrated either mentally or physically to reduce cognitive load. When the same information was merely repeated in a different form, cognitive load was increased by having to unnecessarily coordinate both sets of information, irrespective of whether they were integrated or not. The best way to reduce extraneous cognitive load was to eliminate the redundant version. Chandler and Sweller (1991) demonstrated the redundancy effect by showing that two forms of the same information resulted in less learning than one form. It turned out that we were not the first to demonstrate the redundancy effect. That effect had been demonstrated, forgotten, and then demonstrated

c13.qxd 9/29/05 18:58 Page 320

320

Efficiency in Learning

again on several occasions over many decades. Why was it forgotten on each occasion? For many people, the redundancy effect is counter-intuitive. Most of us intuitively feel that presenting learners with the same information in several different ways cannot be harmful and could be beneficial. In fact it is harmful, and cognitive load theory explains why it is harmful. If we have to unnecessarily coordinate multiple sources of the same information, scarce working memory resources are being used for activities unrelated to schema acquisition and automation, depressing learning. Demonstrations of a counter-intuitive effect without a proper theoretical explanation tended to be ignored and forgotten, and that may be why early examples of the redundancy effect had no impact on the field. It is to be hoped that the current, cognitive load theory explanation can establish the redundancy effect as a major weapon in the armory of instructional professionals. The split-attention effect led indirectly to the redundancy effect as described above. It led directly to the modality effect. It has been known for some time that both auditory and visual working memory can be used simultaneously and that, in combination, the use of both processors increased the capacity of working memory to some extent (Penney, 1989). When faced with a diagram, which must be visual, and words, which can be either visual or auditory, it seemed to follow that audio/visual presentation with the words presented in auditory (that is, spoken) form should increase available working memory by transferring some of the visual working memory load to auditory working memory. Mousavi, Low, and Sweller (1995) tested the hypothesis that, under split-attention conditions (that is, where the two sources of information were unintelligible in isolation), presenting text in spoken rather than written form would be beneficial. Controlled experiments comparing geometry diagrams and written or spoken text demonstrated the superiority of the spoken text, providing an example of the modality effect. The split-attention, redundancy, and modality effects proved strong, robust effects that could be easily demonstrated and so provided clear principles for instruction. The only problem was that, mixed with some large experimental effects, we kept obtaining evidence that under some conditions there was no effect at all. Again, these failures forced us to return to basic cognitive load theory to provide reasons that could be used to generate new

c13.qxd 9/29/05 18:58 Page 321

The Evolution of Cognitive Load Theor y

321

instructional principles. In the process, cognitive load theory underwent a massive expansion. To this point, we had only been concerned with extraneous cognitive load, which is cognitive load caused by instructional procedures. Many of the cognitive load principles discussed in Part 2 of this book are examples of extraneous cognitive load because they are under the control of the instructional designer. There are other forms of cognitive load—with intrinsic cognitive load being an important example. Here’s how the concept of intrinsic cognitive load developed. In the early to mid-1990s, we realized that some effects, such as the split-attention, redundancy, and modality effects, could not be obtained with some materials. We needed an explanation. We eventually discovered that the effects invariably failed when the nature of the material was such that it could be processed in working memory one or two elements at a time (Sweller, 1994; Sweller & Chandler, 1994). Learning to translate some of the nouns of a foreign language provides an example. You can learn each noun without reference to any of the other nouns, and so working memory load is naturally low. In contrast, the natural structure of some material is such that working memory load is high during learning. In order to understand such material, you needed to process many elements simultaneously because they interacted (for example, learning to deal with an equation or formula). Using this high element interactivity material, split-attention, redundancy, and modality effects could be readily obtained. In other words, complex material with many interacting elements gave the effects, but simple material with very few or no interacting elements did not. The reason instructional effects could be obtained using high but not low element interactivity material was that if the material included many interacting elements, it imposed a high intrinsic cognitive load—intrinsic because it was not determined by what the instructor did, but by the nature of the material. If a high extraneous cognitive load due to instructor activity was added to a high intrinsic cognitive load due to high element interactivity, we got the various effects. If intrinsic cognitive load was low due to low element interactivity, it hardly mattered what the instructor did because working memory was not overloaded. In other words, with a low intrinsic cognitive

c13.qxd 9/29/05 18:58 Page 322

322

Efficiency in Learning

load, extraneous cognitive load didn’t matter. We called this effect the element interactivity effect. At that time, we assumed intrinsic cognitive load was immutable. It could not be varied because it was “intrinsic” to the material. Only extraneous cognitive load due to instructional design could be varied. With Pollock, Chandler, and Sweller (2002), we realized that there had to be ways of reducing intrinsic cognitive load; otherwise very complex material could never be learned. We had to modify the theory to say that you can reduce intrinsic cognitive load but you cannot simultaneously maintain full understanding. You can eliminate and reduce some of the interacting elements to allow working memory to handle the material. For example, in a computer application, you may omit explanations and just tell learners what steps to follow. Those steps may be easily processed in working memory. At that point, understanding won’t occur, but once the reduced material has been learned it can be put together with the omitted information to give understanding. In that sense, intrinsic cognitive load is to some extent under the control of the instructor. Learning can be facilitated by reducing the number of interacting elements and only reintroducing them later when the essential elements have been learned. For these reasons, the split-attention, redundancy, modality, and other effects are defined as being due to extraneous cognitive load because they are under the full control of the instructor. Intrinsic cognitive load is not. It is intrinsic to the material being taught. The instructor will need to devise instruction to take account of intrinsic cognitive load by omitting some of the interacting elements, but cannot alter it except in such artificial ways and with an initial loss of understanding. Nevertheless, by presenting material with some interacting elements initially omitted, learning can be facilitated. We called the effect the “isolated/interacting elements effect.”

Recent Years—The Internationalization of Cognitive Load Theory A third form of cognitive load, germane cognitive load, was discovered after cognitive load theory attracted international interest. That interest first arose in Europe. Until the early 1990s, cognitive load theory was exclusively studied

c13.qxd 9/29/05 18:58 Page 323

The Evolution of Cognitive Load Theor y

323

at the University of New South Wales in Sydney. While we were a relatively large group, we were the only group. That situation began to alter when Jeroen van Merriënboer and his then student, Fred Paas, working in Holland, began using the theory. Paas and van Merriënboer (1994) found that if they gave learners worked examples that differed considerably in variability, cognitive load was increased compared to worked examples that were all very similar. Nevertheless, despite the increase in cognitive load, high variability worked examples resulted in better learning than low variability examples, giving the variability effect. Clearly, this was a different type of cognitive load from the more commonly studied extraneous and intrinsic load. They labeled this form of cognitive load “germane” cognitive load because it was a load that was germane to schema acquisition and automation. In effect, the aim of reducing extraneous cognitive load is to free working memory capacity for germane load. If the only consequence of reducing extraneous cognitive load is to reduce the mental work that learners do, it will not result in improved learning. The introduction of germane cognitive load was not the only advance provided by Paas and van Merriënboer (1994). Along with Paas (1992), they provided a very timely technique for directly measuring cognitive load. In the early 1990s, cognitive load theory was beginning to be noticed for the first time, but it was also beginning to be criticized. Much of this criticism was ideological in nature. Instructional theory was going through one of its frequent fads based on ideology, rather than data or on any coherent conception of human cognitive architecture. This particular fad insisted that instruction should never directly provide students with information; rather, they should discover it themselves. The name of the movement kept altering as it failed to find credible data, but you may recognize discovery learning, problem-based learning, enquiry-based learning, or constructivism. By the early 1990s, this movement was at its height, and the last thing its adherents wanted was someone from the wrong side of the world telling them that learning was facilitated by directly instructing learners and ensuring that the instruction reduced unnecessary mental activities. Up to that point, there was a large hole in the empirical evidence for cognitive load theory, and that hole could be exploited by anyone wishing to

c13.qxd 9/29/05 18:58 Page 324

324

Efficiency in Learning

attack the theory. While we could and had used the theory to generate lots of novel instructional techniques, our evidence that the effects were caused by cognitive load rather than some other factors were indirect at best. A reduced learning time was the most frequently used evidence. Paas (1992) and Paas and van Merriënboer (1994) demonstrated the worked example effect, but in addition used subjective measures of task difficulty to directly measure cognitive load. They not only found that the use of worked examples facilitated learning but also found that learners experienced a reduced working memory load as a consequence of studying worked examples. We immediately began using their techniques at UNSW to demonstrate that cognitive load was indeed the reason for the cognitive load effects. One of the other findings provided by our Dutch colleagues at this time was the example completion effect. Rather than giving people full worked examples to study, they gave them partially completed problems that had to be completed by the learner. Paas (1992) and Paas and van Merriënboer (1994) found this technique equally as effective as worked examples and better than full problems, thus providing an example of the example completion effect. This effect was subsequently used by Renkl and Atkinson (2003) to demonstrate the guidance fading effect. They provided learners with full worked examples initially. As learner expertise increased, those full examples were replaced by partially completed examples. With additional expertise, the partially completed examples were replaced by full problems. Some new findings at UNSW indicated the need for this guidance fading procedure. All of the effects described so far were demonstrated using novices. In the late 1990s we discovered that, as expertise increases, the effects gradually disappear. With further increases in expertise, they reverse (see Kalyuga, Ayres, Chandler, & Sweller, 2003). Again, cognitive load theory can be used to explain why. Consider worked examples. For novices, they are needed to demonstrate how a problem or class of problems is solved. As expertise increases, learners may still need to practice, but worked examples may now be redundant. Unnecessarily studying a worked example may impose a greater extraneous cognitive load than simply solving the problem, resulting in a reverse worked example effect. At this point, problem-solving practice may be

c13.qxd 9/29/05 18:58 Page 325

The Evolution of Cognitive Load Theor y

325

more useful than studying worked examples (Kalyuga, Chandler, Tuovinen, & Sweller, 2001). In other words, worked examples need to be faded through completion problems to full problems for maximum learning efficiency. At this time, many other researchers began to use cognitive load theory. In the United States, Richard Mayer incorporated the theory into his cognitive theory of multi-media learning (Mayer, 2001). He, along with Roxana Moreno (Mayer & Moreno, 2003) focused the use of cognitive load theory principles on instruction that used words, pictures, and sound. Several researchers in Germany began to use the theory in their experiments. The work of Alexander Renkl, along with Robert Atkinson (the latter in the United States) in combining the worked example, completion, and expertise reversal effects to develop the guidance fading effect, has been described above. Roland Brunken, Detlev Leutner, and Jan Plass (the latter in the United States) conducted some very substantial work on the use of secondary tasks as an alternative to subjective rating measures of cognitive load (Brunken, Plass, & Leutner, 2003). Peter Gerjets along with Katharina Scheiter and Richard Catrambone (the latter in the United States) investigated instructional techniques for handling a large intrinsic cognitive load (Gerjets, Scheiter, & Catrambone, 2004). All of these researchers have done more than merely use cognitive load theory. As a consequence of their work, the theory has altered and developed.

Current Work Currently, many lines of work are being pursued within a cognitive load theory framework. Summaries of some of that work can be found in recent special issues of journals devoted to the theory (Educational Technology Research and Development, 2005, 53[3]; Educational Psychologist, 2003, 38[1]; Instructional Science, 2004, 32[1–2]; Learning and Instruction, 2002, 12[1]). At UNSW, we are pursuing many different lines of work. Three lines that I am closely involved with are (1) finding measurement devices that can be used to determine what sort of information should be provided to learners; (2) looking at the effect of asking learners to imagine concepts or procedures; and (3) using our instructional findings to further develop our knowledge of

c13.qxd 9/29/05 18:58 Page 326

326

Efficiency in Learning

human cognition, especially the reasons why human cognitive architecture evolved in its particular way.

Measuring Expertise The expertise reversal effect indicates that instructional techniques that are effective for novices are ineffective for more expert learners, and it follows that the instructional procedures we use should alter as expertise increases. Nevertheless, while expertise reversal provides an interesting scientific effect, it can have few practical instructional implications unless we can determine learners’ levels of expertise and the instructional procedures appropriate to different levels. Currently, levels of expertise can be determined using normal tests of achievement, but these tests take a relatively long time to administer, far too long if one wishes to determine how instruction should be presented at a given point during a learning episode. Rapid, accurate tests of achievement that take no more than a few minutes to administer are required. What might such tests look like? They are described in Chapter 11 and are illustrated in our demonstration lesson on the CD. Based on cognitive load theory, these tests assume that expertise is determined by the acquisition of automated schemas. If one has an automated schema in a particular area, one can immediately recognize a problem situation and the best move associated with it. Therefore, instead of presenting students with a full test, we simply present a problem situation and ask students what is the best move that follows. Students’ answers can be used to indicate the extent to which they have acquired automated schemas and, in turn, the next instructional episode is determined by the extent to which automated schemas have been acquired (Kalyuga & Sweller, 2005). Obviously, modern computer technology is essential for this procedure.

Imagination (Mental Rehearsal) Effect Assume that instead of asking students to “study” material that they need to learn, we ask them to “imagine” or “mentally rehearse” the material instead. What differences might we expect to see in their cognitive processes, and what consequences might those differences have on learning? Rehearsing by imagining involves processing procedures or concepts in working memory, while

c13.qxd 9/29/05 18:58 Page 327

The Evolution of Cognitive Load Theor y

327

studying is more likely to involve reading and thinking about material. Initially, for novices, reading and thinking about material is likely to be essential. Furthermore, until at least rudimentary schemas have been constructed, processing procedures through working memory is likely to be impossible. For novices, instructions to study are likely to be superior to instructions to imagine. In an example of the expertise reversal effect, with the development of expertise the advantages can be predicted to reverse. Continually reading over the same material is likely to yield diminishing returns. Imagining or rehearsing relevant procedures and concepts should now be possible and should assist in transferring material to long-term memory and automating that material. Thus, for more expert learners, instructions to imagine are likely to be superior. These finding have been obtained (Cooper, Tindall-Ford, Chandler, & Sweller, 2001). We might expect this imagination effect to interact with other cognitive load effects such as the split-attention and modality effects. If material in integrated or dual mode form reduces cognitive load, that should make it easier to imagine or rehearse, which in turn should facilitate learning. Our current research is beginning to provide precisely these results.

Evolution of Human Cognitive Architecture This issue may appear to be quite unrelated to instruction, and indeed it is doubtful that instructional effects will flow directly from it. Nevertheless, it needs to be remembered that, without its base in human cognitive architecture, there would be no cognitive load theory. Without knowledge of human cognitive architecture, meaningful, large-scale advances in instructional procedures are probably impossible. Cognitive load theory is firmly grounded in its view of human cognition, and any strengthening of that view strengthens the theory. Why do we have our particular cognitive architecture? We know that our cognitive architecture must have evolved according to the rules of natural selection, in the same way as every other biological structure and function evolved, but why, for example, do we have such a limited working memory associated with a very large long-term memory? Here are some suggestions (see Sweller, 2003 and 2004, for details).

c13.qxd 9/29/05 18:58 Page 328

328

Efficiency in Learning

Human cognition is a natural information processing system, and as such it shares common characteristics with other natural information processing systems and differs from artificial systems created by humans. Evolution by natural selection is also an example of a natural information processing system. Because natural information processing systems share common characteristics and because evolution by natural selection drove the evolution of human cognitive architecture, in effect, the processes of human cognition mimic the processes of evolution. That is useful because we know far more about evolution than we do about cognition. If they share a common underlying logic, that logic is likely to provide us with information about cognition. Here are some of the common features of both evolution and cognition. A massive base of information is central to both systems. In the case of evolution, that base is a genetic code, while in the case of cognition it is longterm memory. All initial changes to the information base involve a random generation followed by tests of effectiveness procedure. In the case of evolution, that process is known as random mutation, followed by differential ability to reproduce. In the case of cognition, it is an unavoidable feature of problem solving. When making a problem-solving move, if we do not have knowledge in long-term memory or knowledge available in someone else’s longterm memory, we have no choice other than to randomly make a move and test it for effectiveness. All human knowledge can be ultimately sourced to this procedure, just as all information in a genetic code can be sourced to random mutations. If changes to a large store of information are random, each change must be very small, because large changes are likely to destroy the functionality of the store. Accordingly, major genetic alterations are likely to occur over millennia (or longer!). A limited working memory ensures that alterations to long-term memory are small and incremental. Together, these points provide an underlying logic for human cognition, and cognitive load theory is based on them. While, as indicated above, they do not directly provide us with instructional procedures, they do explain why cognitive load theory has followed its particular direction of emphasizing

c13.qxd 9/29/05 18:58 Page 329

The Evolution of Cognitive Load Theor y

329

knowledge held in long-term memory, with working memory closely tied to the needs of long-term memory. This base explains why cognitive load theory rejects some alternative instructional movements such as discovery learning or constructivism. Teaching learners how to make random problemsolving moves and test them for effectiveness is likely to be futile.

Conclusions From its beginnings as an attempt to understand why problem solvers learn little while solving problems to its current concerns with how our cognitive architecture is structured, cognitive load theory has continually based its instructional recommendations on the outcomes of controlled experiments. That emphasis on research into human cognition to provide theory and controlled experiments to provide data has led to the development of the instructional principles described in this book. While most of those principles are concerned with learning, a limited working memory affects all of our activities when dealing with novel information, including perceiving information and understanding instructions. The limitations and strengths of our cognitive architecture affect all of our cognitive activities. Ultimately, beyond the science of cognition and instruction, the usefulness of a cognitive load theory based approach will be determined by practitioners in the field.

c13.qxd 9/29/05 18:58 Page 330

bapp01.qxd 9/29/05 19:03 Page 331

APPENDIX: ALL ABOUT THE NUMBERS

The Numbers The guidelines we present in this book are based on experimental research evidence of instructional methods that lead to more efﬁcient learning. In our research summaries, we present the main statistical data that support the conclusions, including outcome averages, statistical signiﬁcance, and when reported, effect sizes. In Chapter 1 we introduced the efﬁciency metrics and the efﬁciency graph that displays them. Some readers may want to know more about how the efﬁciency metric is calculated and displayed on the graph. In addition, some readers may want to review the meaning of the statistical data we present. To manage the cognitive load of previous chapters, we decided to put these details in an appendix that describes and gives examples of the mathematical and statistical concepts behind the numbers. We begin with the numbers behind the efﬁciency metric, followed by a review of statistical concepts.

331

bapp01.qxd 9/29/05 19:03 Page 332

332

Appendix

Calculating and Displaying the Efﬁciency Metric As we mentioned in Chapter 1, the efficiency metric is calculated by subtracting mental load from performance outcomes. We can express this mathematically as E P ML. When performance is greater than mental load, the efficiency value is positive. When performance is lower than mental load, the efficiency value is negative. Performance is most often measured by a test taken at the end of the lesson. Sometimes, however, performance is measured by the time required to complete a lesson or a test task. Mental load is most commonly measured by learner ratings of lesson difficulty on a 1 to 7 or 1 to 9 scale. In a typical experiment, learners are randomly assigned to take one of two variations of a lesson. For example, one version uses audio narration to explain a visual, and the comparison version uses the same words presented in text to explain the visual. After studying their assigned version, learners rate the difficulty of the lesson and then take a performance test. Because the test scores and the mental effort ratings use such different numeric values, the data must be converted to a standardized score called a Z-score to make them comparable for purposes of calculating the efficiency values and plotting them on an efficiency graph. Figure A.1 summarizes mathematical details about Z scores.

The Efﬁciency Graph To visually represent efficiency, instructional scientists use an efficiency graph, shown in Figure A.2, in which the horizontal axis represents the range of mental effort ratings in Z scores and the vertical axis represents the performance outcomes (test scores or time) in Z scores. The origin or point of intersection of the two axes represents the average of all test scores and all ratings, which when converted to Z scores equals 0. Therefore a performance Z score greater than 0 will fall above the axes, and a mental load score greater than 0 will fall to the right of the axes. The efficiency graphs are shown with a theoretical reference line for which efficiency 0. Any point falling along this line represents a situation in which the Performance Z score equals the Mental Effort Z score. Since

bapp01.qxd 9/29/05 19:03 Page 333

333

Appendix

Figure A.1. What Is a Z Score? Z Scores allow instructional scientists to convert metrics that use different scales to a standardized scale. As a result, we can calculate and plot an efficiency metric based on test scores and mental effort ratings. A Z score converts the average of a set of numbers to 0 and the standard deviation to 1. Thus, any individual Z score is expressed in terms of its distance in standard deviations from the average. For example a Z score of +.4 represents a score that is four tenths of a standard deviation above the average. In contrast, a Z score of -.8 tells us that the score is eight tenths of a standard deviation below average.

Figure A.2. The Efﬁciency Graph with Hypothetical Plots. High Efficiency

Performance

E=0

1.0 0.8 0.6 0.4 0.2 - 1.0 - 0.8 - 0.6 - 0.4 - 0.2

0.2 0.4 - 0.2

Mental Effort 0.6

0.8

1.0

- 0.4 - 0.6 - 0.8 - 1.0

Low Efficiency

the efficiency is calculated by subtracting mental effort from performance, when they are equal, the result is 0. Note that the line is labeled E 0. Any lesson that results in high performance (above 0 on the vertical line) and at the same time requires low mental effort (to the left of 0 on the

bapp01.qxd 9/29/05 19:03 Page 334

334

Appendix

horizontal line) will fall into the upper left quadrant of the graph. This quadrant is called the high efficiency quadrant, since all points that fall within it represent higher than average performance and lower than average mental effort. Lesson Version A in Figure A.2 represents a high efficiency lesson with an average performance value of around 0.5 and an average mental effort value of around 0.7. In contrast, lesson Version B falls into the low efficiency quadrant because it resulted in performance that fell below the average and mental effort that fell above the average.

Calculating the Efﬁciency Metric To calculate the efficiency value of a lesson version, you use the equation: Average Performance in Z Scores – Average Difficulty Rating in Z Scores √2 Based on this equation, the average Z score for mental effort ratings for a lesson is subtracted from the average Z score for performance outcomes for that lesson, and the difference is divided by the square root of 2 (a mathematical operation required for the calculation of distances between points). To see a sample calculation and efficiency graph that uses a small model data set, refer to Figures A.3 and A.4. In this step-by-step example we use three hypothetical data sets from two lesson versions. We convert the raw scores into Z scores, determine the efficiencies for each version, and plot them on the efficiency graph. In the next section we illustrate some efficiency data from an actual experiment plotted on the efficiency graph.

Example of Actual Efﬁciency Values Plotted on the Efﬁciency Graph The data below come from an actual experimental comparison of two versions of lessons designed to teach interpretation of a complex electrical table (Tindall-Ford, Chandler, & Sweller, 1997). In the text version, the table was explained by text printed below the table. In the audio version, the table

bapp01.qxd 9/29/05 19:03 Page 335

Figure A.3. A Worked Example of Efﬁciency Calculation. Test Difficulty Rating Version 1 Actual score Z Score Actual rating Student A 70 .36 2.3 Student B 80 1.3 3.4 Student C 75 .81 2.9 Average Z .82 Plot Values: Performance = .82 and Mental Effort = –.72

Z Score –1.25 –0.22 –0.68 –0.72

Version 2 Student D 64 –0.18 4.3 Student E 52 –1.26 3.6 Student F 55 –0.99 5.3 Average Z –1.03 Plot Values: Performance = –0.81 and Mental Effort = .70 Version 1 + 2 Average 66 Standard Deviation 11.08

.63 –0.02 1.5 .70

3.63 1.06

Version 1 E = 0.82 – –0.72 = 1.09 √2

Version 2 E =

–0.81 – 0.70 = –1.07 √2

Formula for a Z score: raw score – average of all scores or Z = X – X standard deviation of all scores S

Figure A.4. Efﬁciency Graph for Worked Example in Figure A.3. High Efficiency

Performance

E=0

1.0 0.8 Version 1 E = 1.09

0.6 0.4

Mental Effort

0.2 - 1.0 - 0.8 - 0.6 - 0.4 - 0.2

0.2 0.4 - 0.2 - 0.4

0.6

0.8

1.0 Version 22 E = - 1.07

- 0.6 - 0.8 - 1.0

Low Efficiency

bapp01.qxd 9/29/05 19:03 Page 336

336

Appendix

Figure A.5. Data from Experiment Comparing Text and Audio Explanations of a Visual (Tindall-Ford, Chandler, & Sweller, 1997).

Mental load (max = 7) Test score (Max = 12) Efficiency

Lesson Version Text Explanation Average St. Dev 3.9 0.5

7.3

2.3

-.53

Audio Explanation Average St. Dev 2.6 * 0.5

9.8*

1.4

1.06*

* Significant difference p<.05 Conclusions: 1. There is a significant main effect in favor of the audio-visual group 2. Learning with audio-visual instruction appears to require significantly less mental effort than learning with a visual only format. 3. The audio-visual group’s efficiency score was significantly higher than that of the visual-only group.

was explained by narration of the same words presented in the text version. After presentation of each lesson, learners rated the difficulty of the lesson and were tested with problems that required them to interpret the table. The data and conclusions from this study are displayed in Figure A.5. Table A.1 shows the Z scores from this data. As you can see, for the text version, the test Z scores were below average and mental effort was higher than average. In contrast, for the audio version, test performance was high, while mental effort was low. Figure A.6 shows these values plotted on the efficiency graph. This experiment is a demonstration of the modality effect, which states that a complex visual is understood more efficiently when explanations are presented in audio narration than when explanations are presented in text. (Note that Z and efficiency scores should add to 0. They do not in this example because there were other groups in the study that are not discussed here.)

bapp01.qxd 9/29/05 19:03 Page 337

337

Appendix

Table A.1. Data from Tindall-Ford, Chandler, and Sweller (1997) Study Converted to Z Scores. Average Z and Efficiency Scores Test Performance

Mental Effort

Efficiency

Text

-0.06

.21

-0.19

Audio

.78

-.85

1.16

Figure A.6. Efﬁciency Graph for Data in Figure A.5. High Efficiency

Performance

E=0

1.0 0.8 0.6 Audio Version E = 1.16

0.4 0.2

- 1.0 - 0.8 - 0.6 - 0.4 - 0.2

0.2 0.4 - 0.2

Mental Effort 0.6

0.8

1.0

- 0.4 - 0.6 - 0.8

Text Version E = -0.19

- 1.0

Low Efficiency

Statistics and Evidence-Based Practice The guidelines we offer in this book are based on evidence from many controlled research studies. In each chapter we describe several experiments that form the basis for our guidelines. All of the research results we present are

bapp01.qxd 9/29/05 19:03 Page 338

338

Appendix

based on experimental designs in which learners are randomly assigned to two or more lesson versions. Typically, one lesson version applies cognitive load management instructional methods, and the other version does not. For example, in the previous section we illustrated efficiency data from an experiment in which learning and mental effort were compared between a lesson that explained a complex electrical table with text positioned below the table with another lesson version that explained the same table with the same words presented in audio narration. The data from this experiment are shown in Figure A.5. As you can see, the average test scores are higher among learners who studied the lessons with audio explanations. In addition the standard deviations are smaller. What do these numbers really mean? How much should you rely on results like these?

Standard Deviations and Means You may have come across standard deviations before. If so, you may recall that a standard deviation tells us how spread out the scores are. It is an average deviation of each score from the mean in a set of scores. If we give ten students a test, we will end up with ten test scores. We can calculate the mean of those ten test scores. We can then find the difference between each score and the mean score. If we then find the mean of those differences, that will give us the standard deviation. It is the average difference of the ten scores from the mean score. If the standard deviation is close to 0, the scores are all very similar because they are close to the mean score. If the standard deviation is large, the scores are spread out a long way from the mean. In general, more successful lessons result in test score averages that are higher—indicating more learning. If standard deviations are lower, that means greater consistency among the individual learners. That is exactly what you can see in the results shown in Figure A.5. In these data you can see that the mean or average score in the audio explanation group is 9.8 with a standard deviation of 1.4. In contrast, the text explanation group averages 7.3 with a standard deviation of 2.3. The audio group has a higher mean, indicating that they had learned more than the text explanation group, and a lower standard deviation, indicating that more people had scores that were close to the group mean scores.

bapp01.qxd 9/29/05 19:03 Page 339

339

Appendix

Statistical Signiﬁcance Because the learners were randomly assigned to each of the lessons, individual differences among participants should be the same in the two experimental groups. Therefore any differences in performance or mental load ratings are due to either chance or to real differences in learning caused by the instructional materials. To rule out chance explanations, researchers submit the data to a statistical test. In the experiment shown in Figure A.5, the statistical tests all showed that there is less than a 5 percent probability that the differences between the two groups could have occurred by chance alone. On this basis, we say that the results are statistically significant.

Practical Signiﬁcance and Effect Sizes However, even very small differences in scores can be statistically significant— especially if the experiment includes a large number of participants. Therefore, statistical significance does not necessarily translate into practical significance—especially if production of the “better” version is more expensive or time-consuming. As a result, many recent experiments report an additional statistic called effect size. The effect size tells you how many standard deviations the test group is from the control group. For example, an effect size of .5 means that you could expect any individual score of someone studying the less effective lesson version to increase by half a standard deviation if he or she took the more effective version. For example, suppose Sam had scored 60 points after studying a lesson that did not apply cognitive load guidelines. Let’s assume that the more effective lesson had an effect size of .75 and the average standard deviation of both the more and less effective lesson groups was 8. We could expect Sam’s score to increase by 8 times .75, or 6 points. As a general guideline, effect sizes less than or equal to .20 are considered small and are of negligible practical importance. Effect sizes around .50 are considered medium and are of moderate practical importance. Finally, effect sizes of .80 or higher are large and are of crucial practical importance (Hojat & Xu, 2004). Let’s return to our sample study comparing a text with an audio explanation of a graphic. We saw that the data had statistical significance. But is there sufficient difference to warrant recommending audio explanations of

bapp01.qxd 9/29/05 19:03 Page 340

340

Appendix

visuals? To calculate effect size, we divide the difference between the average scores for the two groups by the average standard deviations of both groups as follows: 9.8 7.3 1.35 1.85 This means that a learner in the text group that scored 5.0 could expect to score 5.0 (1.35 times 1.85 2.5) or about 7.5 if they studied the audio version. As you can see, the effect size from this experiment falls into the high range and results in a considerable improvement in learning. If we replicated this study and generally obtained high effect sizes like this one, we could recommend the use of audio to explain visuals with a high degree of confidence in its practical payoffs.

bgloss.qxd 9/29/05 19:04 Page 341

GLOSSARY

Activation of prior knowledge

A psychological process of learning in which relevant schemas or mental models stored in long-term memory are brought into working memory for the purpose of integrating new instructional content to result in larger, more complex schema or mental models.

Adaptive training

Instruction that matches content or instructional methods based on individual differences in learners. For example, a more experienced learner receives different lesson versions than a novice learner. Typically adaptive training is implemented in e-learning in which learners are assessed and branched to the appropriate instructional versions based on the assessment results. Also known as prescriptive learning.

341

bgloss.qxd 9/29/05 19:04 Page 342

342

Glossar y

Asynchronous e-learning

Instructional programs delivered on a computer that are designed primarily for self-study. These programs can be taken at any time by any one at his or her own pace. May or may not include options for synchronous or asynchronous communication. Common examples include some web-based training and computer-based training.

Attention

A psychological process in which limited mental capacity is used more efficiently by focusing mental resources on relevant features in the environment.

Automaticity

The status of any knowledge or skill that has been used so many times that it can be activated from long-term memory and applied using minimal working memory resources. Some common examples of automated skills among educated adults include driving and reading.

Backwards fading

An instructional technique in which an expanding number of steps in a worked example are left for the learner to complete starting from the last steps and working backwards. For example, a five-step example would begin with the first four steps worked out and the last step left for the learner to complete. A second five-step example would work out the first three steps, leaving the last two steps for the learner to complete. See also faded worked examples and completion examples.

Chunking

A technique in which information in long-term memory is used to chunk or group together multiple elements of information into a single element that can be easily processed in working memory. Seeing the letters C.H.U.N.K.I.N.G. as a single word rather than eight letters provides an example.

Cognitive load theory

A universal set of instructional principles and evidencebased guidelines that offer the most efficient methods to design and deliver instructional environments in ways that best utilize the limited capacity of working memory.

bgloss.qxd 9/29/05 19:04 Page 343

343

Glossar y

Cognitive motivation

Instructional strategies that increase the probability of initiating or completing an instructional event as a result of making the instructional materials more relevant or comprehensible. Some examples include the use of an organizational visual to illustrate the relationships among topics or the inclusion of job-realistic case examples to illustrate application of new skills to the learner’s work assignments.

Comparison group experiments

Experiments in which participants are randomly assigned to two or more different lesson versions and outcomes are compared between the two groups. Also known as controlled experiments.

Completion examples

An instructional technique in which a step-by-step worked example is partially filled in by the instruction and finished by the learner. For example, an algebra problem example in a lesson demonstrates the first three steps and asks the learner to complete the last three steps. See also faded worked examples and backwards fading.

Cues

Instructional methods used to draw learner attention to elements in the visual instructional interface. Common examples include arrows, highlighting, and circles.

Directive course architectures

Instructional programs in which content is explicitly presented to learners in short lessons that typically include rules, examples, and practice with feedback.

Disordinal interaction

When two instructional methods have opposite effects on two different types of learners. For example, worked examples benefit low prior knowledge learners but hinder learning of learners with greater expertise.

Dual encoding

A theory that accounts for the beneficial effects of visuals on learning resulting from formation of two memory codes: one from words and a second from visual representation of content.

bgloss.qxd 9/29/05 19:04 Page 344

344

Glossar y

Dynamic delivery format

Any delivery medium in which the instruction is typically delivered outside of learner control of pacing. A central feature of dynamic delivery is that the content is presented in a transient manner. Some examples include video, classroom lectures, and multimedia animations.

Edutainment

An instructional approach in which themes, games, or vignettes are added to a technical lesson for the purpose of increasing emotional sources of motivation.

Effect size

A statistic that indicates the practical significance of experimental outcome data. Effect sizes indicate the proportion of a standard deviation difference that would be realized if the control group learned from the experimental lesson version. Effect sizes of .5 and above indicate practical significance.

Efficiency

A property of instructional products that results in faster learning (or performance), better learning (or performance), or both. Technically an efficiency metric is calculated by subtracting the standardized (Z) average mental difficulty rating of an instructional product from the standardized (Z) average performance score realized after studying that product.

Efficiency graph

A two-dimensional diagram in which average performance Z scores are plotted on the vertical axis and average mental effort or program difficulty Z scores are plotted on the horizontal axis. The most efficient programs are those that fall into the upper left quadrant of the graphic, indicating programs that result in high performance with low mental effort.

Element interactivity

A property of instructional content that reflects the extent to which multiple content components must be held and/or processed simultaneously in working memory in order to be learned or to achieve a performance objective. For example, learning some of the nouns of a foreign language

bgloss.qxd 9/29/05 19:04 Page 345

345

Glossar y

vocabulary is a low element interactivity task because each word can be learned separately. In contrast, composing a sentence in a foreign language is a high element interactivity task because all words must be considered in relationship to each other and to grammar and parsing rules. Emotional motivation

Instructional strategies that increase the probability of initiating or completing an instructional program as a result of making the materials more engaging or humorous through the use of themes, games, or stories that are not related to the instructional goal. See also edutainment.

Encoding

A psychological process of learning resulting in storage of new instructional content in schema in long-term memory.

Evidence-based practice

The incorporation of valid research evidence into decisions about selection or design of instructional programs. Such research is usually based on comparison or controlled group experiments.

Expertise reversal

The negative effect of instructional methods that aid the learning of novices on the learning of experts. Because experts have a relatively large schema relevant to the instructional goal, they are able to manage their own cognitive learning processes without external instructional support. In some cases instructional methods such as worked examples interfere with learning of experts because of conflict between the instruction and the existing schema of experts.

Explanatory visuals

A diagram that illustrates relationships among content and helps learners build deeper understanding. Some examples include graphic expressions of quantitative data, concept maps, and schematic illustrations.

Exploratory course architectures

Instructional programs that allow a high degree of learner control over content and/or instructional methods, allowing learners to select the portions of the instruction they want to study.

bgloss.qxd 9/29/05 19:04 Page 346

346

Glossar y

Extraneous cognitive load

Work imposed on working memory that uses mental capacity but does not contribute to learning. Extraneous sources of cognitive load should be minimized in order to free working memory for processes that lead to learning. Some examples of instructional methods that impose extraneous cognitive load include redundant expressions of content such as narrating text to learners or explaining a self-explanatory visual, as well as split attention caused by separating explanatory text from a related visual.

Faded worked examples

A technique in which a step-by-step example is partially filled in by the instruction and finished by the learner. For example, an algebra problem example in a lesson demonstrates the first three steps and asks the learner to complete the last three steps. See also completion examples and backwards fading of worked examples.

Far transfer

Knowledge and skills that must be applied in diverse contexts on the job. Some examples of far transfer tasks include making a sales presentation or designing training. To perform far transfer tasks, the worker must use judgment to adapt guidelines to diverse work situations. Also known as nonrecurrent skills or principle-based tasks.

Germane cognitive load

Work imposed on working memory that uses mental capacity in ways that contribute to learning. Germane sources of cognitive load should be used to build the mental models appropriate to the instructional goal. Examples include the use of varied context examples and practice exercises that result in more robust mental models than examples or practice exercises that use similar contexts.

Guided discovery course architectures

Instructional programs in which learners are encouraged to explore an instructional environment but are also provided with a degree of guidance. Guided discovery courses will often rely on a job-realistic problem or case assignment as a vehicle to promote learning.

bgloss.qxd 9/29/05 19:04 Page 347

347

Glossar y

Interaction

When two or more factors combine to provide a result different to either factor in isolation. Two instructional methods may have different or opposite effects on different learners. For example, instructional method X works well for novice learners but either has no effect or depresses learning of experienced learners.

Intrinsic cognitive load

Work imposed on working memory as a result of the amount of element interactivity of the content to be learned. Intrinsic cognitive load is high when multiple content elements must be processed simultaneously in working memory, such as when learning to compose a sentence in a foreign language or when learning a number of steps in using a spreadsheet. Intrinsic cognitive load is an inherent feature of the knowledge and skills to be trained, although it can be artificially reduced by instructional designers through chunking and sequencing of content.

Learner control

The extent to which learners can select their own instructional pacing and/or content. Learner control may apply to pacing of the instruction and/or selection of content or instructional methods. Instructional programs may be high or low in learner control, depending on the delivery medium and their design. For example, classroom instruction is typically low in learner control because the pace and content are determined by the instructor.

Long-term memory

A relatively permanent mental repository of knowledge and skills in the form of schema that provide the basis for expertise. The schemas in long-term memory interact directly with working memory to influence the virtual capacity of working memory.

Mental model

A memory structure located in long-term memory that incorporates our knowledge and skills. Also known as schemas. Mental models can be larger or small and grow over time as learning progresses. Mental models are the basis for expertise.

bgloss.qxd 9/29/05 19:04 Page 348

348

Glossar y

Mental rehearsal

Modality effect

Near transfer

Performance aids

A psychological process of learning in which new knowledge and skills are practiced in memory in ways that lead to learning. A cognitive load principle stating that complex visuals are understood more efficiently when explanatory words are presented in an audio modality than when presented in a written modality. Because working memory includes separate processing areas for visual and auditory information, using the auditory mode along with the visual makes most efficient use of limited working memory resources. Knowledge and skills that are applied in more or less the same way each time they are used. Refers to routine tasks such as logging into email or starting an automobile. In research studies, near transfer tests ask learners to demonstrate new skills in a similar context that they were presented in the instruction. Also called recurrent tasks or procedures. External sources of content that aid workers to complete job tasks. They include low-technology support such as fact cards and high-technology support such as online help and wizards.

Phonological loop

A component of working memory that is responsible for storage and processing primarily of auditory information.

Procedures

Content that involves a series of steps to complete a task. Also known as near transfer or recurrent tasks. Includes routine tasks such as logging onto a computer or testing an electrical appliance.

Processes

Content that describes how things work. Can focus on activity flows in mechanical, scientific, and business systems. Some examples include how workers are hired or how the circulatory system works.

Rapid testing

A testing technique based on cognitive load theory that allows rapid assessment of learner expertise followed by adaptive instruction based on that assessment. In rapid

bgloss.qxd 9/29/05 19:04 Page 349

349

Glossar y

testing the learner is asked to quickly produce the next step he or she would take to solve a problem or complete a task. Because experts have more complete schemas, their next steps approximate a full solution, in contrast to novices, who either do not know the next step or produce a next step that represents an early stage in problem solution. Receptive course architectures

Instructional designs that include minimal learner interaction. Some examples include typical briefings, lectures, or text readings that do not include practice opportunities.

Recurrent skills

Skills that are based on routine steps common to a variety of situations. Also called procedures or near transfer tasks.

Redundancy principle

A cognitive load principle stating that content or content expressions that are duplications either of each other or of knowledge already in memory impede learning. Some examples include a text or audio explanation of a selfexplanatory visual as well as an explanation of a visual presented with both text and a duplicate audio narration of that text. Frequently applied to any information presented that is irrelevant to schema acquisition.

Referenced-based training

An instructional program that is based on job reference materials in the form of manuals or working aids. Typically the training materials include learning objectives, explanations of supporting knowledge, examples, and exercises. Learners use the reference materials while completing exercises.

Rehearsal

A psychological learning process in which new knowledge and skills are processed in working memory in ways that lead to learning. May include rote processing known as maintenance rehearsal or deeper processing known as elaborative rehearsal.

Representational visual

A graphic used for the purpose of illustrating the actual appearance of objects. Some examples include photographs of equipment, screen captures, or line drawings of people.

bgloss.qxd 9/29/05 19:04 Page 350

350

Glossar y

Retrieval

A psychological learning process in which relevant knowledge and skills stored in long-term memory are brought into working memory when needed to complete a mental or physical activity. Retrieval is the psychological basis for transfer of learning.

Schema

A memory structure located in long-term memory that is the basis for expertise. Allows the chunking of many elements of information into a single element. Schemas are also called mental models. Schemas can be large or small and grow over time as learning progresses.

Seductive details

Text or visual content added to instructional materials that is unrelated to the instructional goal in order to add emotional interest to the instruction. For example, pictures and descriptions of people injured by lightning strikes are irrelevant to a lesson designed to teach how lightning works. See also edutainment.

Segmenting

An instructional technique in which content is divided into segments and distributed over a series of instructional events such as topics or lessons in order to artificially reduce intrinsic cognitive load.

Self-explanations

Mental processing of examples in which learners attempt to clarify or elaborate on an example presented in instruction. Effective self-explanations lead to better learning of knowledge or skills illustrated by the examples.

Seven plus or minus two A popular phrase coined by George Miller in 1956 to characterize the limited capacity of working memory for information. This concept of seven plus or minus two has been updated in cognitive load theory with the guideline that working memory has a limited capacity of two to four elements during high element interactivity processing events. Signals

Instructional techniques used to draw attention to important elements in the textual instructional interface. Common examples include bolding, highlighting, or italics.

bgloss.qxd 9/29/05 19:04 Page 351

351

Glossar y

Split attention

A source of extraneous cognitive load caused by separation of related instructional elements that must be processed together for understanding. For example, split attention results when a visual is explained by text placed in a distant location from the visual, requiring the viewer to expend mental resources to integrate the two sources of information.

Staged experiments

Research studies in which experimental lessons are presented to the same group of learners over time and outcome data collected several times as those learners gain expertise. Commonly used to evaluate the effects of instructional methods on differing levels of expertise. Commonly known as a repeated measures design.

Standard deviation

A statistic that measures the degree of dispersion of data around the average of the data. Higher standard deviations reflect greater dispersion of data around the average. In most educational research, programs that result in lower standard deviations are more desirable, since they reflect greater consistency among the scores.

Statistical significance

A statistical measure of the probability that the data outcomes did not occur by chance alone. Statistical significance is usually expressed as probabilities less than .05, meaning that there is less than a 5 percent likelihood the differences among experimental programs occurred by chance alone. Statistical significance is influenced by the size of the experimental population and does not necessarily indicate practical significance.

Synchronous e-learning

Instructional programs delivered on a computer that are designed for group participation in different locations at the same time. Typically, these programs are conducted by an instructor and are instructionally paced. Facilities of synchronous e-learning depend on the delivery software, but typically include a white board for visuals, chat, and polling options, as well as audio.

bgloss.qxd 9/29/05 19:04 Page 352

352

Glossar y

Varied context worked examples

A series of step-by-step demonstrations in which the context changes. For example, one illustration of a statistical calculation uses typing scores, while another uses temperatures. Varied context worked examples impose a form of germane cognitive load that is beneficial for fartransfer learning of content such as concepts and principles.

Visual-spatial sketch pad

A component of working memory responsible for processing of diagrams and pictures.

Weeding

An instructional design strategy in which unnecessary content or redundant content modalities are eliminated in order to minimize extraneous cognitive load.

Whole task course designs

Instructional programs in which learning is designed to take place in the context of solving job-realistic problems or tasks. Also called problem-based learning or goal-based scenarios.

Worked example

A step-by-step demonstration used to illustrate how to complete a task. Replacing some practice exercises with worked examples has been shown to increase learning efficiency.

Worked exampleproblem pairs

A technique that decreases extraneous cognitive load by replacing some practice exercises with a series of worked examples, each followed by a similar practice exercise. The worked example-problem pairs technique has been shown to result in more efficient learning than all-practice lessons.

Working memory

A central element of human cognition responsible for active processing of data during thinking, problem solving, and learning. Working memory has a limited capacity and storage duration for information. Cognitive load theory is a set of instructional principles designed to accommodate the limits and exploit the strengths of working memory.

bref.qxd 9/29/05 19:08 Page 353

REFERENCES

Atkinson, R.K. (2002). Optimizing learning from examples using animated pedagogical agents. Journal of Educational Psychology, 94(2), 416–427. Atkinson, R.K., Derry, S.H., Renkl, A., & Wrotham, D. (2000). Learning from examples: Instructional principles from the worked examples research. Review of Educational Research, 70(2), 181–214. Atkinson, R.K., Renkl, A., & Merrill, M.M. (2003). Transition from studying examples to solving problems: Effects of self-explanation prompts and fading worked out steps. Journal of Educational Psychology, 95(4), 774–783. Baddeley, A. (1992). Working memory. Science, 255, 556–559. Brunken, R., Plass, J., & Leutner, D. (2003). Direct measurement of cognitive load in multimedia learning. Educational Psychologist, 38, 53–61.

353

bref.qxd 9/29/05 19:08 Page 354

354

References

Carlson, R., Chandler, P., & Sweller, J. (2003). Journal of Educational Psychology, 95(3), 629–640. Carroll, J.M. (1992). Minimalist documentation. In H.D. Stolovitch & E.J. Keeps (Eds.), Handbook of human performance technology. San Francisco, CA: Jossey-Bass. Cerpa, N., Chandler, P., & Sweller, J. (1996). Some conditions under which integrated computer-based training software can facilitate learning. J. Educational Computing Research, 14(4), 344–367. Chandler, P., & Sweller, J. (1991). Cognitive load theory and the format of instruction. Cognition and Instruction, 8(4), 293–332. Chandler, P., & Sweller, J. (1992). The split attention effect as a factor in the design of instruction. British Journal of Educational Psychology, 62, 233–246. Chandler, P., & Sweller, J. (1996). Cognitive load while learning to use a computer program. Applied Cognitive Psychology, 10, 141–170. Chase, W.G., & Simon, H.A. (1973). The mind’s eye in chess. In W.G. Chase (Ed.), Visual information processing. London: Academic Press. Chi, M.T.H. (2000). Self-explaining expository texts: The dual processes of generating inferences and repairing mental models. In R. Glaser (Ed.), Advances in instructional psychology educational design and cognitive science (Vol. 5). Mahwah, NJ: Lawrence Erlbaum. Chi, M.T.H., Bassok, M., Lewis, M.W., Reimann, P., & Glaser, R. (1989). Self-explanations: How students study and use examples in learning to solve problems. Cognitive Science, 13, 145–182. Clark, R.C. (1999). Developing technical training (2nd ed.). Silver Spring: MD: International Society for Performance Improvement. Clark, R.C. (2003). Building expertise (2nd ed.). Silver Spring, MD: International Society for Performance Improvement. Clark, R.C. (in press). Evidence-based practice and professionalization of HPT. In J. Pershing (Ed.), Handbook of performance technology (3rd ed.). Silver Spring: MD: International Society of Performance Improvement.

bref.qxd 9/29/05 19:08 Page 355

References

355

Clark, R.C., & Lyons, C. (2004). Graphics for learning. San Francisco, CA: Pfeiffer. Clark, R.C., & Mayer, R.E. (2003). e-Learning and the science of instruction. San Francisco, CA: Pfeiffer. Collins, A., Brown, J.S., & Newman, S.E. (1989). Cognitive apprenticeship: Teaching the craft of reading, writing, and mathematics. In L.B. Resnick (Ed.), Knowing, learning, and instruction: Essays in honor of Robert Glaser (pp. 453–493). Mahwah, NJ: Lawrence Erlbaum. Cooper, G., & Sweller, J. (1987). The effects of schema acquisition and rule automation on mathematical problem-solving transfer. Journal of Educational Psychology, 79, 347–362. Cooper, G., Tindall-Ford, S., Chandler, P., & Sweller, J. (2001). Learning by imagining. Journal of Experimental Psychology: Applied, 7(1), 68–82. Craig, S.D., Gholson, B., & Driscoll, D.M. (2002). Animated pedagogical agents in multimedia educational environments: Effects of agent properties, picture features, and redundancy. Journal of Educational Psychology, 94(2), 428–434. Dolezalek, H. (2004). Industry report. Training, 41(9), 20–36. Driskell, J.E., Copper, C., & Moran, A. (1994). Does mental practice enhance performance? Journal of Applied Psychology, 7, 481–492. Etneir, J.L., & Landers, D.M. (1996). The influence of procedural variables on the efficacy of mental practice. The Sport Psychologist, 9, 48–57. Fuat, A., Hungin, A.P.S., & Murphy, J.J. (2003). Barriers to accurate diagnosis and effective management of heart failure in primary care: Qualitative study. British Medical Journal, 326, p. 196. Gerjets, P., Scheiter, K., & Catrambone, R. (2004). Designing instructional examples to reduce intrinsic cognitive load: Molar versus modular presentation of solution procedures. Instructional Science, 32, 33–58. Gick, M.L., & Holyoak, K.J. (1980). Analogical problem solving. Cognitive Psychology, 12, 306–355.

bref.qxd 9/29/05 19:08 Page 356

356

References

Gick, M.L., & Holyaok, K.J. (1983). Schema induction and analogical transfer. Cognitive Psychology, 15, 1–38. Ginns, P., Chandler, P., & Sweller, J. (2003). When imagining information is effective. Contemporary Educational Psychology, 28, 229–251 Gyselinck, V., & Tardieu, H. (1999). The role of illustrations in text comprehension: What, when, for whom, and why? In H.V. Van Oostendrop & S.R. Goldman (Eds.), The construction of mental representations during reading. Mahwah, NJ: Lawrence Erlbaum. Harp, S.F., & Mayer, R.E. (1997). The role of interest in learning from scientific text and illustrations: On the distinction between emotional interest and cognitive interest. Journal of Educational Psychology, 89(1), 92–102. Haskell, R.E. (2001). Transfer of learning. New York: Academic Press. Hartley, J., & Benjamin, M. (1998). An evaluation of structured abstracts in journals published by the British Psychological Society. British Journal of Educational Psychology, 68, 443–446. Hegarty, M., Narayanan, N.H., & Freitas, P. (2002). Understanding machines from multimedia and hypermedia presentations. In J. Otero, J.A. Leson, & A.C. Graesser (Eds.), The psychology of science text comprehension. Mahwah, NJ: Lawrence Erlbaum. Hmelo-Silver, C.E. (2004). Problem-based learning: What and how do students learn? Educational Psychology Review, 16(3), 235–266. Hoffman, B., & Ritchie. D. (1997). Using multimedia to overcome the problems with problem based learning. Instructional Science, 25, 97–115. Hojat, M., & Xu, G. (2004). A visitor’s guide to effect sizes. Advances in Health Sciences Education, 9, 241–249. Hummel, H.G.K., & Nadolski, R.J. (2002). Cueing for schema construction: Designing problem-solving multimedia practicals. Contemporary Educational Psychology, 27, 229–249. Jeung, H., Chandler, P., & Sweller, J. (1997). The role of visual indicators in dual sensory mode instruction. Educational Psychology, 17(3).

bref.qxd 9/29/05 19:08 Page 357

References

357

Jonassen, D.H. (1999). Designing constructivist learning environments. In C.M. Reigeluth (Ed.), Instructional-design theories and models (Volume II). Mahwah, NJ: Lawrence Erlbaum. Jonassen, D.H. (2004). Learning to solve problems. San Francisco, CA: Pfeiffer. Kalyuga, S., Ayres, P., Chandler, P., & Sweller, J. (2003). The expertise reversal effect. Educational Psychologist, 38(1), 23–31. Kalyuga, S., Chandler, P., & Sweller, J. (1998). Levels of expertise and instructional design. Human Factors, 40(1), 1–17. Kalyuga, S., Chandler, P., & Sweller, J. (1999). Managing split-attention and redundancy in multimedia instruction. Applied Cognitive Psychology, 13, 351–371. Kalyuga, S., Chandler, P., & Sweller, J. (2000). Incorporating learner experience into the design of multimedia instruction. Journal of Educational Psychology, 92, 126–136. Kalyuga, S., Chandler, P., & Sweller, J. (2001). Learner experience and efficiency of instructional guidance. Educational Psychology, 21(1), 5–23. Kalyuga, S., Chandler, P., & Sweller, J. (2004). When redundant on-screen text in multimedia technical instruction can interfere with learning. Human Factors, 46, 567–581. Kalyuga, S., Chandler, P., Tuovinen, J., & Sweller, J. (2001). When problem solving is superior to studying worked examples. Journal of Educational Psychology, 93(3), 579–588. Kalyuga, S., & Sweller, J. (2004). Measuring knowledge to optimize cognitive load factors during instruction. Journal of Educational Psychology, 96, 558–568. Kalyuga, S., & Sweller, J. (2005). Rapid dynamic assessment of expertise to improve the efficiency of adaptive e-learning. Educational Technology Research and Development, 53, 83–93. Kamin, C., O’Sullivan, P., Deterding, R., & Younger, M. (2003). A comparison of critical thinking in groups of third-year medical students in text, video, and virtual PBL case modalities. Academic Medicine, 78(2) 204–211.

bref.qxd 9/29/05 19:08 Page 358

358

References

Knez, I., & Hugge, S. (2002). Irrelevant speech and indoor lighting: Effects on cognitive performance and self-reported affect. Applied Cognitive Psychology, 15, 709–718. Kobayashi, K. (2005). What limits the encoding effect of note-taking? A meta-analytic examination. Contemporary Educational Psychology, 30, 242–262. Kotovsky, K., Hayes, J.R., & Simon, H.A. (1985). Why are some problems hard? Evidence from Tower of Hanoi. Cognitive Psychology, 17, 248–294. Larkin, J.S., & Simon, H.A. (1987). Why a diagram is (sometimes) worth a thousand words. Cognitive Science, 11, 65–69. Leahy, W., Chandler, P., & Sweller, J. (2003). When auditory presentations should and should not be a component of multimedia instruction. Applied Cognitive Psychology, 17, 401–418. Leahy, W., & Sweller, J. (2004). Cognitive load and the imagination effect. Applied Cognitive Psychology, 18, 857–875. Leahy, W., & Sweller, J. (in preparation). Cognitive load theory: Interactions between the imagination, expertise reversal and element interactivity effects. Lorch, R.F. (1989). Text-signaling devices and their effects on reading and memory processes. Educational Psychology Review, 1, 209–234. Lorch, R.F., & Lorch, E.P. (1996). Effects of organizational signals on free recall of expository text. Journal of Educational Psychology, 88(1), 38–48. Lyman, P., & Varian, H. (2003). How much information? 2003 Report from U.C. Berkeley accessible at www.sims.berkeley.edu/research/projects/ how-much-info-2003/ Marcus, N., Cooper, M., & Sweller, J. (1996). Understanding instructions. Journal of Educational Psychology, 88(1), 49–63. Mautone, P.D., & Mayer, R.E. (2001). Signaling as a cognitive guide in multimedia learning. Journal of Educational Psychology, 93(2), 377–389. Mayer, R.E. (2001). Multimedia learning. New York: Cambridge University Press.

bref.qxd 9/29/05 19:08 Page 359

References

359

Mayer, R.E. (2002). Using illustrations to promote constructivist learning from science text. In J. Otero, J.A. Leon, & A.C. Graesser (Eds.), The psychology of science text comprehension (pp. 333–356). Mahwah, NJ: Lawrence Erlbaum. Mayer, R.E., Bove, W., Bryman, A., Mars, R., & Tapangco, L. (1995). When less is more: Meaningful learning from visual and verbal summaries of science textbook lessons. Journal of Educational Psychology, 88(1), 54–73. Mayer, R.E., & Chandler, P. (2001). When learning is just a click away: Does simple user interaction foster deeper understanding of multimedia messages? Journal of Educational Psychology, 93(2), 390–397. Mayer, R.E., Dow, G.T., & Mayer, S. (2003). Multimedia learning in an interactive self-explaining environment: What works in the design of agent-based microwords? Journal of Educational Psychology, 95(4), 806–813. Mayer, R.E., & Gallini, J.K. (1990). When is an illustration worth ten thousand words? Journal of Educational Psychology, 82(4), 715–726. Mayer, R.E., Heiser, J., & Lonn, S. (2001). Cognitive constraints on multimedia learning: When presenting more material results in less understanding. Journal of Educational Psychology, 93(1), 187–198. Mayer, R.E., & Jackson, J. (2004, April). Concise science instruction: When quantitative details hurt qualitative understanding. Paper presented at the American Educational Research Association Conference, San Diego, California. Mayer, R.E., Mathias, A., & Wetzell, K. (2002). Fostering understanding of multimedia messages through pre-training: Evidence for a two-stage theory of mental model construction. Journal of Experimental Psychology: Applied, 8(3) 147–154. Mayer, R.E., & Moreno, R. (1998). A split attention effect in multimedia learning: Evidence for dual processing systems in working memory. Journal of Educational Psychology, 90(2), 312–320. Mayer, R.E., & Moreno, R. (2003). Nine ways to reduce cognitive load in multimedia learning. Educational Psychologist, 38(1), 43–52.

bref.qxd 9/29/05 19:08 Page 360

360

References

McNamara, D.S., Kintsch, E., Songer, N.B., & Kintsch, W. (1996). Are good texts always better? Interactions of text coherence, background knowledge, and levels of understanding in learning from text. Cognition and Instruction, 14(1), 1–43. Miller, G.A. (1956). The magical number seven, plus or minus two: Some limits on our capacity for processing information. Psychological Review, 63, 81–97. Moreno, R., & Mayer, R.E. (1999). Cognitive principles of multimedia learning: The role of modality and contiguity. Journal of Educational Psychology, 91(2), 348–368. Moreno, R., & Mayer, R.E. (2000). A coherence effect in multimedia learning: The case for minimizing irrelevant sounds in the design of multimedia instructional messages. Journal of Educational Psychology, 92(1), 117–125. Moreno, R., & Mayer, R.E. (2002). Verbal redundancy in multimedia learning: When reading helps listening. Journal of Educational Psychology, 94(1), 155–153. Moreno, R., Mayer, R.E., Spires, H.A., & Lester, J.C. (2001). The case for social agency in computer-based teaching: Do students learn more deeply when they interact with animated pedagogical agents? Cognition and Instruction, 19(2), 177–213. Mousavi, S.Y.L., Low, R., & Sweller, J. (1995). Reducing cognitive load by mixing auditory and visual presentation modes. Journal of Educational Psychology, 87(2), 319–334. Newell, A., & Simon, H. (1972). Human problem solving. Englewood Cliffs, NJ: Prentice Hall. Paas, F.G.W.C. (1992). Training strategies for attaining transfer of problemsolving skill in statistics: A cognitive load approach. Journal of Educational Psychology, 84(4), 429–434. Paas, F.G.W.C., Tuovinen, J.E., Tabbers, H., & Van Gerven, P.W.M. (2003). Cognitive load measurement as a means to advance cognitive load theory. Educational Psychologist, 38(1), 63–71.

bref.qxd 9/29/05 19:08 Page 361

References

361

Paas, F.G.W.C., & Van Merriënboer, J.J.G. (1994). Variability of worked examples and transfer of geometrical problem-solving skills: A cognitive load approach. Journal of Educational Psychology, 86(1), 122–133. Penney, C.G. (1989). Modality effects and the structure of short term verbal memory. Memory and Cognition, 17, 398–422. Piolat, A., Olive, T., & Kellogg, R.T. (2005). Cognitive effort during note taking. Applied Cognitive Psychology, 19, 291–312. Pollock, E., Chandler, P., & Sweller, J. (2002). Assimilating complex information. Learning and Instruction, 12, 61–86. Quilici, J.L., & Mayer, R.E. (1996). Role of examples in how students learn to categorize statistics word problems. Journal of Educational Psychology, 88(1), 144–161. Ransdell, S.E., & Gilroy, L. (2001). The effects of background music on word processed writing. Computers in Human Behavior, 17, 141–148. Renkl, A., & Atkinson, R.K. (2003). Structuring the transition from example study to problem solving in cognitive skill acquisition: A cognitive load perspective. Educational Psychologist, 38(1), 15–22. Renkl, A., Atkinson, R.K., & Grosse, C.S. (2004). How fading worked solution steps works: A cognitive load perspective. Instructional Science, 32, 59–82. Renkl, A., Stark, R., Gruber, H., & Mandl, H. (1998). Learning from worked out examples: The effects of example variability and elicited self-explanations. Contemporary Educational Psychology, 23, 90–98. Rickards, J.P., Fajen, B.R., Sullivan, J.F., & Gillespie, G. (1997). Signaling, note taking, and field independence-dependence in text comprehension and recall. Journal of Educational Psychology, 89(3), 408–417. Romero, K., & Silvertri, L. (1990). The role of mental practice in the acquisition and performance of motor skills. Journal of Instructional Psychology, 17, 218–221. Schank, R.C., Berman, T.R., & MacPerson, K.A. (1999). Learning by doing. In C.M. Reigeluth (Ed.), Instructional design theories and models (pp. 213–246). Mahwah, NJ: Lawrence Erlbaum.

bref.qxd 9/29/05 19:08 Page 362

362

References

Shiffrin, R., & Schneider, W. (1977). Controlled and automatic human information processing: I. Detection, search and attention. Psychological Review, 84, 1–66. Stark, R., Mandl, H., Gruber, H., & Renkl, A. (2002). Conditions and effects of example elaboration. Learning and Instruction, 12, 39–60. Sternberg, R.J. (1996). Attention and consciousness. Cognitive psychology (pp. 68–107). New York: Harcourt Brace. Strunk, W., & White, E.B. (1999). The elements of style (4th ed.). Boston: Allyn and Bacon. Sweller, J. (1994). Cognitive load theory, learning difficulty and instructional design. Learning and Instruction, 4, 295–312. Sweller, J. (2003). Evolution of human cognitive architecture. In B. Ross (Ed.), The psychology of learning and motivation (Vol. 43, pp. 215–266). San Diego: Academic Press. Sweller, J. (2004). Instructional design consequences of an analogy between evolution by natural selection and human cognitive architecture. Instructional Science, 32, 9–31. Sweller, J. (2005). Implications of cognitive load theory for multimedia learning. In R.E. Mayer (Ed.), Cambridge handbook of multimedia learning. Cambridge: UK: Cambridge University Press. Sweller, J. (2005). The redundancy principles. In R.E. Mayer (Ed.), Cambridge handbook of multimedia learning. Cambridge: UK: Cambridge University Press. Sweller, J., & Chandler, P. (1994). Why some material is difficult to learn. Cognition and Instruction, 12(3), 184–233. Sweller, J., Chandler, P., Tierney, P., & Cooper, M. (1990). Cognitive load and selective attention as factors in the structuring of technical material. Journal of Experimental Psychology: General, 119, 176–192. Sweller, J., & Cooper, G.A. (1985). The use of worked examples as a substitute for problem solving in learning algebra. Cognition and Instruction, 2, 59–89.

bref.qxd 9/29/05 19:08 Page 363

References

363

Sweller, J., Mawer, R., & Howe, W. (1982). The consequences of historycued and means-ends strategies in problem solving. American Journal of Psychology, 95, 455–484. Sweller, J., Mawer, R., & Ward, M. (1983). Development of expertise in mathematical problem solving. Journal of Experimental Psychology: General, 112, 634–656. Sweller, J., van Merriënboer, J.J.G., & Paas, F.G.W.C. (1998). Cognitive architecture and instructional design. Educational Psychology Review, 10(3), 251–296. Tarmizi, R., & Sweller, J. (1988). Guidance during mathematical problem solving. Journal of Educational Psychology, 80, 424–436. Tindall-Ford, S., Chandler, P., & Sweller, J. (1997). When two sensory modes are better than one. Journal of Experimental Psychology: Applied, 3(4), 257–287. Tindall-Ford, S., & Sweller, J. (in preparation). Altering the modality of instructions to facilitate imagination: Interactions between the modality and imagination effects. Titsworth, B.S., & Kiewra, K.A. (2004). Spoken organizational lecture cues and student notetaking as facilitators of student learning. Contemporary Educational Psychology, 29, 447–461. Trafton, J.G., & Reiser, B.J. (1993). The contribution of studying examples and solving problems to skill acquisition. Proceedings of the 15th Annual Conference of the Cognitive Science Society (pp. 1017–1022). Mahwah, NJ: Lawrence Erlbaum. Tuovinen, J.E., & Sweller, J. (1999). A comparison of cognitive load associated with discovery learning and worked examples. Journal of Educational Psychology, 91(2), 334–341. Van den Broek, P., Tzeng, Y., Risden, K., Trabasso, T., & Basche, P. (2001). Inferential questioning: Effects on comprehension of narrative texts as a function of grade and timing. Journal of Educational Psychology, 93(3), 521–529. van Merriënboer, J.J.G. (1997). Training complex cognitive skills: A fourcomponent instructional design model for technical training. Englewood Cliffs, NJ: Educational Technology Publications.

bref.qxd 9/29/05 19:08 Page 364

364

References

van Merriënboer, J.J.G., Kirschner, P.A., & Kester, L. (2003). Taking the load off a learner’s mind: Instructional design for complex learning. Educational Psychologist, 38(1), 5–13. van Merriënboer, J.J.G., & Sweller, J. (2005). Cognitive load theory and complex learning: Recent developments and future directions. Educational Psychology Review, 17(2), 147–176. Ward, M., & Sweller, J. (1990). Structuring effective worked examples. Cognition and Instruction, 7, 1–39. Wetzstein, A., & Hacker, W. (2004). Reflective verbalization improves solutions: The effects of question-based reflection in design problem solving. Applied Cognitive Psychology, 18, 145–156. Zhu, X., & Simon, H. (1987). Learning mathematics from examples and by doing. Cognition and Instruction, 4, 137–166.

babout.qxd 9/29/05 19:00 Page 365

ABOUT THE AUTHORS

Ruth Colvin Clark has focused her professional efforts on bridging the gap between academic research in instructional methods and practitioner application of that research. To that end she has developed a number of seminars and written five books, including e-Learning and the Science of Instruction, that translate and illustrate important research programs for organizational training specialists. A science undergraduate, Ruth completed her doctorate in instructional psychology/educational technology in 1988 at the University of Southern California. Ruth is a past president of the International Society of Performance Improvement and a member of the American Educational Research Association. Ruth is currently a dual resident of Southwest Colorado and Phoenix, Arizona, and divides her professional time among consulting, teaching, and writing. For more information, consult her website at www.clarktraining.com.

365

babout.qxd 9/29/05 19:00 Page 366

366

About the Authors

John Sweller is a psychologist who holds the position of professor of education at the University of New South Wales, Australia. His research interests are in the areas of cognitive processes and instructional procedures. With many others from around the globe, he developed cognitive load theory to provide a vehicle that generates instructional procedures based on our knowledge of human cognitive architecture. That ongoing, international research effort has led to the development of a large number of instructional principles that are discussed in this book. Furthermore, those instructional principles are beginning to throw further light on the very cognitive structures that generated the instructional principles in the first instance, and in the process, are potentially changing our conceptions of human cognitive architecture and its origins. Frank Nguyen is an e-learning technology manager at Intel Corporation. He has managed the design, development, and deployment of corporate e-learning and electronic performance support systems. He holds a master’s and is currently pursuing a doctorate in educational technology from Arizona State University. His research interests are in the areas of electronic performance support and human performance technology.

bindex.qxd 9/29/05 19:06 Page 367

INDEX

Note: Subentries under the entry for cognitive load theory evolution are listed chronologically.

A Abstracts, 82–83 Activation of prior knowledge, 37, 41 Adaptive learning, 276, 282–287; benefits and drawbacks of, 285–287; lack of technology for, 305–307 Adaptive training, 275, 276–277, 281; tips for applying, 287 Animations, 167, 171, 180–181, 184, 207–208 Architecture, human cognitive, 327–329 Architectures of instruction, 162, 174–176; directive, 174–175, 178–179, 186, 267–270; exploratory, 171, 174; guided discovery/whole task, 174, 176–180, 185, 267–270; receptive, 174 Asynchronous e-learning, 133–135, 275; audiovisuals and, 73, 303; e-Lessons for, 45, 311; and instructional planning, 297, 301–302; learner support in, 308

Atkinson, R. K., 65, 202, 211, 229, 230–231, 242, 324, 325 Attention, 37, 77–105; cues and signals and, 78–83; definition of, 78; delivery format and integrated presentation for, 95–102; e-Lessons on, 41, 103–104; tips for supporting, 102–103; and working memory, 78. See also Split attention Audio narration, 61–75; backups for hearingimpaired learners, 75; design tips for, 73; e-Lesson on, 74; for experienced learners, 126–127; in explaining diagrams, 61–70; extraneous, 119–120; followed by text, 130–131, 133; and high complexity content, 75–77; integrated materials and, 86–90, 127; for novices, 69, 126–127; and selfexplanatory visuals, 125, 127; sequencing on-screen text after, 130–131, 133; when to avoid, 133–135; when to use, 66–73, 131–133 Audiovisuals. See Audio narration; Visuals 367

bindex.qxd 9/29/05 19:06 Page 368

368

Index

Auditory centers, 33–34, 43, 48–49

Cognitive learning process. See Learning

Auditory content, extraneous, 119–120

Cognitive load, 9–10, 34–36; balancing of, 13; and expertise, 250–254; extraneous (irrelevant) load, 12–13, 34–35, 95–102; germane (relevant) load, 11, 34, 216–242; intrinsic load, 9–11, 44, 162, 321–322; management tips for, 185; pacing and, 180–185; performance aids and, 146–157; relativity of, 14–15; segmenting and sequencing and, 161–173; training design and, 162; whole task learning and overload, 176–180, 185; worked examples and, 203–208. See also Extraneous cognitive load; Germane cognitive load

Automaticity: automating new knowledge and skills, 233–235; definition of, 234; mental rehearsal and, 235–236, 239; repetitive practice and, 235; when to build, 235; as working memory bypass, 39 Ayres, P., 250, 273, 324

B Backwards fading, 197–203, 208 Baddeley, A., 49 Basche, P., 258 Bassock, M., 196, 227 Benjamin, M., 82–83 Berman, T. R., 177 Bove, W., 111–112, 113 Brown, J. S., 177 Brunken, R., 325 Bryman, A., 111–112, 113

C Carlson, R., 54–56 Carroll, J. M., 109–110 Catrambone, R., 325 Cerpa, N., 97, 99–100, 101, 105, 134, 152 Chandler, P., 16, 17, 18, 21, 54–56, 62, 63, 64, 67, 68, 69, 75, 79, 85, 86–88, 97, 98, 99–100, 101, 105, 122–125, 126–127, 128–129, 130–131, 132–133, 134, 148–149, 150, 152, 169–173, 183, 184, 187, 199, 201, 206, 207, 211, 236–237, 250, 262–266, 268–269, 273, 319, 321, 322, 324, 325, 327 Chase, W. G., 29, 30, 31 Cheat sheets, 141 Chi, M.T.H., 196, 227, 242 Clark, R. C., 41, 59, 64, 74, 93, 148, 163, 169, 174, 178, 183, 211 Classroom lectures. See Lectures Coding. See Encoding Cognitive architecture, human, 327–329

Cognitive load theory, 293–311; application to instructional planning, 294–302; application to training implementation, 307–309; definition of, 7–9; delivery media and, 300–302; e-Lessons on application of, 290–291, 311; and efficiency in learning, 5–14, 23–24, 40–41; evidence for, 16–19; human learning and, 15, 28, 327–329; instructional content and, 298–300; instructional guidelines and, 8; internationalization of, 322–325; software challenges in implementing, 304–307; target audience considerations for, 294–298; training development and, 302–304 Cognitive load theory evolution, 313–329; origins, 314–315; early years: problems with problem solving, 315–317; middle years: importance of failed experiments, 317–322; recent years: internationalization, 322–325; current work, 325–329 Cognitive motivation, 116–118 Coherence of text, 254–257, 261 Collins, A., 177 Comparison group experiments, 252–253 Completion examples, 194–197, 208; backwards fading and, 197–199; definition of, 194; psychology of, 196–197; research on, 195–196 Complexity: audio narration and, 67–69; cognitive load and, 44, 162; in computer learning, 99–101; of content, 13, 78, 102, 162; cues and, 102; diagrams and, 56; instructional planning and, 298–300; learning and, 14, 16; redundancy and, 122. See also Content; Text

bindex.qxd 9/29/05 19:06 Page 369

369

Index

Computer applications and teaching: delivery formats and, 95–102, 133–135; guidelines for, 296–297, 299; integrated performance aids and, 150–155; integrated presentation for, 95–102, 103; manuals for, 97–98, 110, 133–135, 180, 299; performance aids and, 142, 147, 150–155; redundancy and, 133–135; research on, 97–101; speed of integrated approach to, 99–101 Computer software challenges, 304–307 Concise materials, 109–114, 120, 135 Constructivism, 323 Content: complexity and attention support, 78, 82; complexity and cognitive load/ applications, 162, 296, 298–299; complexity and signal use, 82, 83; instructional planning and, 298–300; management tips for, 185; motivational/emotional, 116–118; paring down to essentials in, 108–114; redundancy in, 121–135; segmentation and sequencing of, 162–167, 184, 185; tips for developing lean lessons in, 135–136; tips for imposing gradually, 185; type, in applications, 296; use of text format and, 70

displays, 79; when to use, 82, 102; for worked examples displays, 207–208

D Deeper learning, 57–61, 94 Delivery format: attentional support for, 78, 83; dynamic, 83; and instructional planning, 297; paring down, 109; redundancy in, 121, 133–135; and sequencing/segmenting, 167; single, for integrated presentations, 95–102; and use of audio, 66 Derry, S. H., 211 Deterding, R., 178 Diagrams: audio narration and, 61–70; and deeper understanding, 57–61; diagrams-alone for experienced learners, 262–265; effectiveness of, 59–60; for novices, 57–61; as performance aids, 50–54; psychology of, 60–61; representative versus explanatory visuals, 59–60; for spatial manipulation tasks, 50–54; and spatial relationship rules, 54–57; when to eliminate, 265, 271; when to use, 69–73

Cooper, G., 236–237, 317, 327, 192

Directive course architectures, 174–175; e-Lesson on, 186; software for the classroom, 178–179; use for novices, 267–270; versus guided discovery, 268–270

Cooper, M., 52, 74, 147, 319

Discovery learning, 323

Copper, C., 355

Disordinal interaction, 248–249

Course architectures. See Architectures of instruction

Dolezalek, H., 6

Course design: content, 298–300; delivery media, 300–302; evaluation of, 309–310; instructional planning, 294–302; software challenges, 304–307; target audience, 294–298; training development materials, 302–304; training implementation, 307–309

Driscoll, D. M., 21, 65

Craig, S. D., 21, 65

E

Convergence principle, 219

Cross-sectional design. See Comparison group experiments Cues and signals, 78–83; as attentional support, 41, 102; e-Lessons on, 103–104; limitations of cueing options, 305; research on, 79–82; for verbal information, 79–83; for visual

Dow, G. T., 21 Driskell, J. E., 355 Dual encoding, 60–61 Dual modalities, acceleration of expertise, 48–50 Dual task experiments, 34

E-learning: adaptive training and, 275, 276–277, 281, 287; asynchronous, 73, 133–135, 297, 301–302, 308; cognitive load theory applied to, 7, 301–302; in exploratory course architecture, 174; individual differences in expertise and, 270, 281–285; instructional

bindex.qxd 9/29/05 19:06 Page 370

370

Index

planning for, 297, 301–302; and learner control, 181, 184; rapid testing/assessment and, 281–285; tips for applying adaptive instruction in, 287

reading interruptions and, 257–259; redundancy and, 122, 259–265; redundant content elimination recommended for, 259–265; schema substitutes and, 250–251; text-alone presentation and, 260–261; worked examples and, 193, 197, 199–201, 203, 265–267. See also Expertise reversal; Experts; Novices

Edutainment, 116 Effect size, 18, 339–340 Efficiency: cognitive load theory and, 5–14; diagrams and, 56; evaluation of, 309–310; quantification of, 19–23 Efficiency graph, 22–23, 332–337 Efficiency metric, 22, 332–337 Elaboration of information, 37 Elaboration-rehearsal. See Rehearsal

Expertise reversal, 32, 243, 247, 250–254, 261 Experts: applications of cognitive load theory to, 295; audio narration and, 126–127; in mixed audiences, 127, 297–298; preexisting schemas and, 32, 41; subject-matter experts, 209; visuals without narration and, 126–127. See also Expertise

Electronic performance support systems (EPSS), 142

Explanatory visuals, 59–60, 73

Element interactivity, 10, 48

Exploratory course architectures, 174, 181

Emotional motivation, 116–118

Extraneous cognitive load, 12–13, 34; cues and signals to minimize, 78–83, 102; e-Lessons on, 45, 103–104, 136–137, 186, 210, 311; guidelines for minimizing, 43–44, 311; inefficient learning and, 36; integrated presentation and, 95–102; split attention and, 78, 84–93. See also Working memory

Encoding, 36, 60–61; dual, 60–61; e-Lesson on, 41 Enquiry-based learning, 323 EPSS (electronic performance support systems), 142 Etneir, J. L., 236 Evaluation of courseware for efficiency, 309–310 Evidence-based practice, 15–19

Extraneous content and text, 109–115, 116–119 Extraneous visuals and audio, 115–120

Examples. See Completion examples; Worked examples

F

Excel Course: content for, 300; delivery media for, 301–302; target audience for, 298

Faded worked examples, 197–203, 208, 303–304

Experienced learners, 126–127, 297. See also Expertise; Experts

Fading: backwards, 197–203; of performance aids, 156

Experimental designs, 252

Fajen, B. R., 90, 91

Expertise, 243, 247–273; acceleration of, with dual modalities, 48–50; acceleration of, with worked examples, 193; accommodating differences in, 271–272; cognitive load changes with, 250–254; coherent text and, 254–257; diagrams-alone presentation and, 262–265; directive designs and, 267–270; e-Lessons on, 245, 272; experimental design and, 252–254; learning styles and, 248; measuring, 326; problem assignments and, 265–267; rapid testing and, 281–285;

Far transfer, 219–221, 296, 299 Freitas, P., 167 Fuat, A., 7 Funding for schools, 16 Fusion Diagram, 128–129

G Gallini, J. K., 260–261, 265 Gerjets, P., 325

bindex.qxd 9/29/05 19:06 Page 371

371

Index

Germane cognitive load, 11, 34, 213–242; automating new knowledge/skills and, 233–235; beneath the cover story on, 221; definition of, 218; e-Lessons on, 215, 241; far transfer and flexible schemas in, 221–222; mental rehearsal and, 235–239; self-explanations and, 226–233; shifting from extraneous load to, 218; tips for design of lessons on, 240–241; transfer of learning and, 218–226; varied context examples, 222–226; worked examples and, 218–226 Gholson, B., 21, 65 Gick, M. L., 219–220, 221 Gillespie, G., 90, 91

Hummel, H.G.K., 178 Hungin, A.P.S., 7

I Imagination effect. See Mental rehearsal Inefficient learning, costs of, 6–7 Information Fatigue Syndrome, 6 Instruction goals, 218 Instructional architectures. See Architectures of instruction Instructional guidelines: cognitive load theory and, 8; five essential learning processes and, 36–37

Grosse, C. S., 229, 242

Instructional materials: concise, 109–115; effective summary of, 112, 113–114; lean approach to, 109; memory support embedded in, 142; redundancy in, 121–122; tips for developing, 135–136; for training, 143–144, 155–156, 308–309. See also Performance aids

Gruber, H., 229

Instructional planning, 294–302

Guided discovery course architectures, 174, 176–180, 267–270

Integrated methods, 297, 299; integrated media and, 95–102, 103; split attention and, 84–93

Gilroy, L., 120 Ginns, P., 236 Glaser, R., 196, 227 Global Knowledge, 153–155

Gyselinck, V., 59–60

H Hacker, W., 232 Harp, S. F., 117 Hartley, J., 82–83 Haskell, R. E., 220 Hayes, J. R., 234 Hearing-impaired learners, 66, 67 Hegarty, M., 167 Heiser, J., 117, 130 Hmelo-Silver, C. E., 179, 186 Hoffman, B., 176, 178

Integrated text, 84–93, 204 Interaction, 248–250; disordinal, 248–249 Interactivity, element, 10, 48 Intrinsic cognitive load, 9–11, 44, 162, 321–322 Irrelevant load. See Extraneous cognitive load

J Jackson, J., 114–115, 181–182 Jeung, H., 79, 207 Jonassen, D. H., 176, 178

K

Howe, W., 314

Kalyuga, S., 69, 126–127, 128–129, 130–131, 132–133, 184, 199, 201, 250, 262–266, 268–269, 273, 277–278, 280, 282, 283–285, 288, 324, 325, 326

Hugge, S., 120

Kamin, C., 178

Human cognitive architecture, 327–329

Kellogg, R. T., 91, 92, 105

Hojat, M., 339 Holyoak, K. J., 219–220, 221

bindex.qxd 9/29/05 19:06 Page 372

372

Knez, I., 120 Kester, L., 177, 180, 187 Kiewra, K. A., 91 Kintsch, E., 255–256, 257 Kintsch, W., 255–256, 257 Kirchner, J.J.G., 177, 180, 187 Knez, I., 120 Kobayashi, K., 94 Kotovsky, K., 234

Index

Long-term memory, 15, 27, 29; overview of function, 34–39; schemas stored in, 31, 33, 235 Longitudinal design. See Staged experiments Lonn, S., 117, 130 Lorch, E. P., 82 Lorch, R. F., 82, 255 Low, R., 20, 71, 72 Lyman, P., 6 Lyons, C., 59, 74, 93, 148, 163, 183, 211

L Landers, D. M., 236

M

Larkin, J. S., 54

MacPerson, K. A., 177

Leahy, W., 16, 17, 18, 21, 67, 68, 123–125, 236–238, 242

Mandl, H., 229

Learner control of pacing, 161, 180–186

Manuals, computer, 97–98, 110, 133–135, 180, 299 Marcus, N., 52, 74, 147

Learner expertise. See Expertise

Mars, R., 111–112, 113

Learning, 7–9, 34–38; adaptive, 276; deeper, 57–61, 94; five elements of, 36–37; methods, 213–214; mixed audiences and, 127, 297–298; overload, 29; overview of, 33–34, 36–38; procedures, 62; styles, 248. See also Cognitive load; Memory; Modality effect

Mathias, A., 164–166, 167, 187

Learning agents, 64–65 Learning environments, 50, 119–120 Learning management systems (LMS), 305 Learning objectives, 299 Learning processes, five essential, 36–38 Learning styles, 248 Learning transfer, 38, 214, 218–226, 299 Lectures: attention support for, 78; note-taking in, 90–92, 93–95, 102, 103; in receptive course architecture, 174; signaled versus unsignaled, 90–91

Mautone, P. D., 80, 81, 105 Mawer, R., 314, 317 Mayer, R. E., 21, 58, 64, 65, 75, 80, 81, 88–89, 90, 105, 109, 111–112, 113, 114–115, 117, 119, 130, 132, 133, 150, 164–166, 167, 181–182, 183, 187, 260–261, 265, 325 Mayer, S., 21, 65 McNamara, D. S., 255–256, 257 Mean, statistical, 338–340 Media. See Delivery format; Multimedia Medical schools, problem-based learning in, 178, 179

Leutner, D., 325

Memory: human, 327–329; long-term, 15, 27, 29, 34–39, 235; support, external, 139–158; support, when to avoid, 144–146; support, when to create, 143–144; working, 15, 28–41, 48–49. See also Performance aids; Working memory

Lewis, M. W., 196, 227

Mental capacity, limits on, 9, 29–31

LMS. See Learning management systems (LMS)

Mental effort: in learning, 36; unproductive, 34

Load managed web-based lesson, 279

Mental model, 114, 120

Lester, J. C., 21, 65, 75

bindex.qxd 9/29/05 19:06 Page 373

373

Index

Mental rehearsal, 37, 235–239, 326–327; definition of, 236; e-Lesson on, 42; research about, 236–239; when to use, 236

directive architecture for, 267–270; exploratory architecture to be avoided with, 181, 267–270; learner control of pacing and, 181; in mixed audiences, 127; redundancy detrimental to, 122; schema substitutes for, 32–33, 41; use of audio for, 69; use of diagrams for, 57–61; whole task courses and, 179. See also Expertise

Mental work, 189, 213; tips for imposing gradually, 209 Merrill, M. M., 202, 211, 230–231 Miller, G. A., 7, 29 Mixed audiences, 127, 297–298

Numbers and statistics, 17–18, 331–340

Modality effect, 20–21, 61–66; cues and, 79; integrated materials and, 88–90; redundancy and, 128–130; worked examples and, 204–206

O

Modality principle, 19, 181

Olive, T., 91, 92, 105

Monitor and correct self-explanations, 228

O’Sullivan, P., 178

Moran, A., 355

Overload: e-Lesson on, 45; and whole task learning, 176–180, 185

Moreno, R., 21, 65, 75, 88–89, 90, 105, 109, 119, 132, 133, 150, 325

Older learners, notetaking and, 94

Motivation, 116–118

P

Mousavi, S.Y.L., 20, 71, 72

Paas, F.G.W.C., 22, 41, 194, 195–196, 222–225, 242, 317, 323, 324

Multimedia training: attention support for, 78, 83, 92–94; audiovisuals and, 66, 180; integrated materials and, 88–89; leanercontrolled pacing and, 184; modality effect and, 204

Pacing, learner control of, 162, 180–186 Paragraph headers, 79 PBL (problem-based learning), 178

Multitasking, 48

Penney, C. G., 361

Murphy, J. J., 7

Performance: audiovisuals and, 72; diagrams and, 53; performance related content, 143–144. See also Performance aids

N

Performance aids, 50–54, 139–158; cheat sheets, 141; for computers, 142, 147, 150–155; definition of, 140–141; design of, 146–157; e-Lesson on, 158; electronically delivered, 142; fading of, 156; integrated, 150–155; redundancy and, 157; in reference-based training, 144; reference guides as, 141; for spatial content, 146–150; testing of, 157; text in, 146–150; tips for, 157, 302–303; visuals in, 50–52, 146–150; wall charts as, 141; when to avoid, 144–146; when to create, 143–144; working aids, 141

Nadolski, R. J., 178 Narayanan, N. H., 167 Narrated animated instruction, 88–90 Narration. See Audio narration Near transfer, 219–221, 296, 299 Near transfer tasks. See Procedures Newell, A., 315 Newman, S. E., 177 No Child Left Behind Act, 16 Note-taking, 90–92, 93–95, 103; older learners and, 94; signaling and, 91–92, 95, 102 Novices: applications of cognitive load theory to, 295, 297; attentional support for, 78;

Phonological loop, 49 Piolat, A., 91, 92, 105 Plass, J., 325 Pollock, E., 169–173, 187, 322

bindex.qxd 9/29/05 19:06 Page 374

374

Power-Point®, 153–155 Practical significance, 18, 339–340 Practice, 190, 303–304; practice problems replaced with worked examples, 191–194; repetitive, 235 Prescriptive learning. See Adaptive training Pretests, 276, 279–280, 288 Prior knowledge, 249, 251–253; activation of, 36; and expertise reversal, 257; in instructional planning, 295 Problem assignments, 197–203 Problem-based learning (PBL), 178, 179, 323 Procedures: alternatives for organizing procedural lessons, 168–173; definition of, 168; segmenting and sequencing and, 169–173; supporting knowledge taught separately from, 168–173 Process content: design of lessons on, 167; preceded by teaching components of processes, 173; research on, 164–166; segmenting and sequencing and, 163–167

Index

Redundancy, 121–135, 259–265, 319–320; audio explanations and, 125; in delivery format, 121, 133–135; e-Lesson on, 136–137; in instructional materials, 121–122; and learner-controlled pacing, 184; mixed audiences and, 127; novice and experienced learners and, 126–127, 259–265; omitting redundant test, 128–130; in performance aids, 157; research applications for, 123; sequencing on-screen text after audio recommendation for, 130–131, 133; tips for developing lean lessons, 135–136; violations in software packages, 305; words redundant to illustrations, 122–125, 128 Reference-based training, 144, 157 Reference guides, 141 Rehearsal, 36–37, 41, 235–239, 326–327; definition of, 236; e-Lesson on, 41; research about, 236–239; when to use, 236 Reimann, P., 196, 227 Reiser, B. J., 196

Process knowledge, 163

Relevent load. See Germane cognitive load

Q

Renkl, A., 202, 211, 229, 230–231, 242, 324, 325

Qualitative mental model, 114, 120, 135

Repetitive practice, 235

Quiet learning environments, 119–120

Representative visuals, 59–60

Quiet work environments, 119–120

Retrieval, 36–37, 41

Quilici, J. L., 361

Rickards, J. P., 90, 91 Risden, K., 258

R

Ritchie, D., 176, 178

Ransdell, S. E., 120

Romero, K., 236

Rapid testing, 277–281, 301; benefits and drawbacks of, 285–287; e-learning based on, 283–285; e-Lessons on, 270, 288; and learner expertise, 281–285; versus traditional test results, 280–281 Reading: asking questions during, 258; avoiding interruption during, 257–259; using coherent text for low knowledge readers, 254–257 Receptive course architectures, 174 Recurrent skills. See Procedures

S Schank, R. C., 177 Scheiter, K., 325 Schemas: activation of, 37, 38; avoid rehearsing while forming, 236, 239; for far transfer, 221–222; in guided discovery course architecture, 176; pairing problem assignments and worked examples to build, 197, 208; preexisting, stored in long-term memory, 33; and self-explanations, 226–228;

bindex.qxd 9/29/05 19:06 Page 375

375

Index

storage of, 31, 32–33, 235; substitutes for, 32–33, 250–251; in working and long-term memory, 30–33, 218, 235 Schneider, W., 234 School funding, 16 Seductive details, 117–118 Segmenting, 161–173, 184, 185; design of process lessons and, 167; e-Lessons on, 186; of procedure content, 169–173; of process content, 163–167; segmented lesson examples, 171, 173; unsegmented lesson example, 172 Self-explanations, 196, 226–233; definition of, 226–227; of design tasks, 232–233; good, 228; research on, 229–231; schemas and, 226–228; self versus peer explanations, 233 Self-explanatory text, 84, 86, 265 Self-explanatory visuals, 84–86, 121–125, 127, 148–149, 259 Sequencing, 161–173, 184, 185; design of process lessons and, 167; e-Lessons on, 186; of procedure content, 169–173; of process content, 163–167; research on, 164–166, 169–173; of system components before full process, 162–167 Seven plus or minus two (7 2), 29–30 Shiffrin, R., 234 Signals, 78–83, 102; e-Lessons on, 103–104; for notetaking, 91–92, 95; paragraph headers as, 79; research on, 79–82; signal phrase, 79; signaled versus unsignaled lectures, 90–91; for verbal information, 79–83; when to use, 82, 102 Significance: practical, 18, 339–340; statistical, 17–18, 339 Silvertri, L., 236 Simon, H., 193 Simon, H. A., 29, 30, 31, 54, 234, 315 SMEs (subject-matter experts), 209

Spatial manipulation rules, 54–57 Spatial manipulation tasks, 50–54 Spires, H. A., 21, 65, 75 Split attention, 77–105, 320; avoiding, 84–93, 95–102; cues and signals to minimize, 78–83; definition of, 77; instructional planning and, 301; integrated methods and, 84–93, 95–102; note-taking and, 90–92, 93–95, 102, 103; performance aids and, 150; research on, 86–92; split attention principle, 92–94; worked examples and, 204–206 Staged experiments, 253–254 Standard deviation, 338–340 Stark, R., 229 Statistical significance, 17–18, 339 Statistics, 17–18, 331–340; effect size, 18, 339–340; practical significance, 18, 339–340; standard deviations and means, 338–340; statistical significance, 17–18, 339 Sternberg, R. J., 362 Structured abstracts, 82–83 Strunk, W., 109 Subject-matter experts (SMEs), 209 Sullivan, J. F., 90, 91 Summary, 112, 113–114 Supporting knowledge: taught separately from procedures, 168–173; topic development and, 304 Sweller, John, 1–2, 16, 17, 18, 20, 21, 25, 41, 52, 54–56, 62, 63, 64, 67, 68, 69, 71, 72, 74, 75, 79, 85, 86–88, 97, 98, 99–100, 101, 105, 122–125, 126–127, 128–129, 130–131, 132–133, 134, 147, 148–149, 150, 152, 169–173, 184, 187, 192, 199, 201, 206, 207, 211, 236–238, 242, 250, 262–266, 268–269, 273, 277–278, 280, 282, 314, 317, 319, 321, 322, 324, 325, 326, 327 Sweller, John, personal perspective from, 313–329

Songer, N. B., 255–256, 257

Sweller, John, video interviews with, 3, 24, 41, 45, 74, 103, 136, 185, 209, 215, 241, 244, 272, 287, 290, 311

Spatial content, 146–147, 209

System components, 162–167

Software challenges, 304–307

bindex.qxd 9/29/05 19:06 Page 376

376

Index

Tapangco, L., 111–112, 113

Transfer of learning, 37–38, 214; in applying cognitive load theory, 296, 299; near versus far, 219–221; worked examples and, 218–226

Tardieu, H., 59–60

Tuovinen, J. E., 22, 265–266, 269, 325

Target audience, 294–298

Tzeng, Y., 258

T Tabbers, H., 22

Tarmizi, R., 319 Technical information: qualitative mental model and, 114, 120, 135; signals and, 82–83; unnecessary, 114–115

U Usability tests, 110

Technical training, 163 Testing, 275–291; adaptive, 275; benefits and drawbacks of rapid testing in e-learning, 285–287; e-learning based on rapid assessment, 283–287; e-Lessons on, 270, 288; learner expertise and rapid testing, 281–285; rapid testing, 277–291, 301; traditional versus rapid testing, 280–281 Text: audio use to describe, 71–72; coherence of, 254–257, 261; conciseness in, 109–114, 120, 135; extraneous, 115–118, 119; integration of, 84–93; omitting redundant text, 128–130; on-screen, sequencing after audio, 130–131, 133; in performance aids, 146–150; redundant content of, 259–265; self-explanatory, 84, 86, 265; summary of, 112, 113–114; text-alone for experienced learners, 260–261, 271; unnecessary technical content in, 114–115; use to describe content, 70; when to eliminate, 265, 271; when to narrate, 66–73, 131–133 Tierney, P., 319 Tindall-Ford, S., 20, 62, 63, 64, 75, 85, 86–88, 105, 150, 206, 211, 236–237, 327 Titsworth, B. S., 91 Trabasso, T., 258 Trafton, J. G., 196 Training: design, 167, 174–176; development, 302–304; implementation of cognitive load theory techniques in, 307–309; memory support in, 155–156; reference-based, 144, 157. See also Architectures of instruction; Course design Training materials. See Instructional materials

V Van den Broek, P., 258 Van Gerven, P.W.M., 22 van Merriënboer, J.J.G, 41, 177, 180, 187, 196, 224–225, 242, 317, 323, 324 Varian, H., 6 Varied context examples, 222–226 Visual centers, 33–34, 43, 48–49 Visual-spatial sketch pad, 49 Visuals, 47–75; animations, 167, 180, 181, 184, 207–208; audio narration and, 61–70; for complex problems, 56; concise use of, 113; cues for, 79, 207–208; for deeper understanding, 57–61; delivery medium and, 66; design tips for, 73; diagrams-alone use for experienced learners, 262–265; effectiveness of, 59–60; explanatory, 59–60, 73; extraneous, 115–118, 119; integrated materials/text for, 86–90, 127; novices and, 57–61; in performance aids, 50–52, 146–150; placement in relation to text, 94, 102, 103; psychology of diagrams in, 60–61; redundancy in, 121–125, 259–265; representative, 59–60; self-explanatory, 84–86, 121–125, 127, 148–149, 259; for spatial manipulation tasks, 50–54; and spatial relationship rules, 54–57; split attention and, 84; tips for development of, 302–303; when to eliminate, 265, 271; when to use, 69–70

W Wall charts, 141 Ward, M., 317, 319

bindex.qxd 9/29/05 19:06 Page 377

377

Index

transfer of learning, 219–226; transition to problem assignments from, 197–203, 265–267; varied context examples, 222–226

Web casts, 66 Wetzell, K., 164–166, 167, 187 Wetzstein, A., 232

Whole task learning, 176–180, 267–270; avoid for novices, 267–270; and cognitive load management, 180; cognitive overload risks of, 176–180, 185; definition of, 176–177; examples of, 177–178; in medical schools, 178, 179; problem-based learning in, 178, 179, 323; software for the classroom, 178–179; tradeoffs for, 179–180

Working memory, 15, 28–31, 33–40, 48–49; attention and, 78; auditory centers and, 33–34, 43, 48–49; bypassing, 39, 140, 157; distraction to, 78; limits on, 29–31, 140, 162; mental effort in learning and, 37; overview of, 33–38; paring down content for, 108–114; schemas in, 31–32, 218, 235; Seven plus or minus two significance for, 29–31; visual and auditory centers in, 33–34, 48–49; worked examples and, 194. See also Germane cognitive load; Performance aids

Work. See Mental work

Writing, conciseness in, 109–114, 120, 135

Worked example-problem pair lessons, 192

Wrotham, D., 211

White, E. B., 109 Whole task course architectures, 174, 176–180, 185, 267–270

Worked examples, 189–211, 265–266; acceleration of expertise by, 193; backwards fading and, 197–203, 208, 303–304; completion examples versus, 194–197; definition of, 190–191; display of, 203–208; e-Lessons on, 210, 241; experience/expertise and, 197, 199–201, 203; failed experiments with, 317–319; formatting of, 203–208; in instructional planning, 296; psychology of, 193–194, 225–226; replacing practice problems with, 191–194; self-explanations and, 226–233; tips on using, 209; and

X Xu, G., 339

Y Younger, M., 178

Z Z scores, 332–337 Zhu, X., 193

bindex.qxd 9/29/05 19:06 Page 378

blof.qxd 9/29/05 19:07 Page 379

L I S T O F F I G U R E S A N D TA B L E S

Figures 1.1 An Assignment in an Excel Lesson That Imposes Moderate Intrinsic Cognitive Load 9 1.2 A Screen from a Lesson on Excel with Many Sources of Extraneous Cognitive Load 12 1.3 A Graph of Temperature Changes over Time

17

1.4 Audio Explanations Result in Better Achievement Than Textual Explanations on Complex Questions 18 1.5 Hypothetical Efficiency Plots on the Efficiency Graph 23 2.1 Number of Referrals Needed to Reproduce a Mid-Play Chess Board 30 2.2 Number of Referrals Needed to Reproduce a Random Chess Board 31 2.3 An Overview of Cognitive Learning Processes

33

2.4 An Instructional Display That Imposes Extraneous Load 2.5 An Instructional Display That Minimizes Extraneous Load

35 35

379

blof.qxd 9/29/05 19:07 Page 380

380

L i s t o f F i g u r e s a n d Ta b l e s

2.6 A Virtual Classroom Excel Lesson Incorporates Demonstrations of Excel Applications 38 3.1 Working Memory Includes a Phonetic (Auditory) and Visual Component 49 3.2 A Simple and Complex Assembly Task Explained with Text and with Diagrams 51 3.3 Diagrams on Performance Aids Lead to Faster Task Performance 52 3.4 Diagrams Are More Efficient Than Text as Work Aids 53 3.5 A Text and Diagram Version of a Chemistry Compound Suffix Rule 55 3.6 Diagrams Are More Efficient Than Text for Learning of More Complex Suffix Problems 56 3.7 Text and Text Plus Diagram Versions from the Bicycle Pump Lesson 58 3.8 A Representational and an Explanatory Illustration of Gas Pressure 60 3.9 A Visual-Only Version from an Electrical Test Lesson 3.10 Learning Is Better from Audio-Visual Presentations 3.11 A Learning Agent from an Excel Lesson

63 64

65

3.12 A Graph of Temperature Changes over Time for Two Days

67

3.13 Audio Explanations Result in Better Achievement on Complex Questions 68 3.14 Exercise Directions Displayed in Text Rather Than Audio 3.15 A Geometry Problem and Solution Example Presented in Combinations of Text, Diagram, and Audio 71 3.16 Diagrams and Text Explained by Audio Lead to Faster Performance 72 4.1 First Paragraphs from Unsignaled Version of Passage on Airplane Lift 80 4.2 First Paragraphs from Signaled Version of Passage on Airplane Lift 80

70

blof.qxd 9/29/05 19:07 Page 381

381

L i s t o f F i g u r e s a n d Ta b l e s

4.3 Signaled Versions Led to Better Learning 4.4 A Structured Abstract for a Journal Article

81 83

4.5 A Separated Text Version from an Electrical Test Lesson

85

4.6 Two Self-Explanatory Information Sources That Would Not Lead to Split Attention 86 4.7 An Integrated Text Version from an Electrical Test Lesson

87

4.8 Text Separated from the Visual Led to Split Attention and Less Learning 88 4.9 A Screen from Three e-Lesson Versions on How Lightning Forms 89 4.10 Audio Description of Visuals Led to Best Learning, Followed by Integrated Text, Which Was Better Than Separated Text 90 4.11 Taking Notes Leads to Split Attention Unless the Lecture Is Signaled 91 4.12 Placement of Text and Use of Pointers to Minimize Split Attention 93 4.13 Having to Refer Back to These Directions During Practice Will Lead to Split Attention 94 4.14 Displaying Training Content in Two Media Leads to Split Attention 95 4.15 A Computer Training Manual That Minimizes Split Attention by Integrating Text and Visuals 96 4.16 A Computer-Based Training Lesson That Minimizes Split Attention by Integrating Text and Visuals on the Computer 97 4.17 Integrated Materials Led to Faster Learning of Complex Software Skills 98 4.18 Integrated CBT Leads to Better Learning of High-Complexity Skills Than Manuals Plus Software 100 4.19 Integrated CBT Leads to More Efficient Learning Than Manuals Plus Software 101

blof.qxd 9/29/05 19:07 Page 382

382

L i s t o f F i g u r e s a n d Ta b l e s

5.1 A Screen from Our Overloaded Excel e-Learning Lesson on the CD 108 5.2 A Screen with Overly Wordy Text

110

5.3 A Concise Version of the Text in Figure 5.2 from Our Excel Load Managed Lesson on Our CD 111 5.4 Two Captioned Illustrations from the Summary Lesson Version 112 5.5 Learning Is Best from Concise Explanations of Visuals

113

5.6 Learning Is Better from Concise Lessons That Omitted Quantitative Details, Regardless of Media 115 5.7 Learning Is Better from Lessons That Omit Seductive Details 118 5.8 Ratings of Cognitive and Emotional Interest of Lessons with and Without Seductive Details 118 5.9 Learning Is Better Without Auditory Additions

119

5.10 An Airline Safety Card with Self-Explanatory Visuals

121

5.11 Learning Was Better and Faster with Self-Explanatory Diagrams Alone 123 5.12 A Graph of Temperature Changes Over Time for Two Days 124 5.13 A Self-Explanatory Version of the Graph in Figure 5.12

124

5.14 Adding Audio to Self-Explanatory Diagram Depresses Learning of Complex Tasks 125 5.15 A Worked Example Using Audio to Explain How to Interpret the Diagram 126 5.16 As Learners Gained Expertise, the Diagram Alone Led to Best Learning 127 5.17 Fusion Diagram Explained by On-Screen Text

129

5.18 Diagrams Explained by Audio Are More Efficient Than Diagrams Explained by Text or by Text and Audio 129 5.19 Audio Followed by Text Leads to Better Learning in Instructionally Paced Lessons 131

blof.qxd 9/29/05 19:07 Page 383

383

L i s t o f F i g u r e s a n d Ta b l e s

5.20 Two Experiments Comparing Learning from Audio Alone with Audio and On-Screen Text 132 5.21 Integrated CBT Led to Better Learning of Complex Computer Skills Than Redundant or Split Attention Versions 134 6.1 An Airline Performance Aid

141

6.2 A Wall Chart from an Instructional Design Class

141

6.3 Memory Support Embedded in Instructional Materials 6.4 A Text-Dominant Working Aid

142

146

6.5 Visual Representations in Performance Aids Led to Faster Performance 147 6.6 An Inefficient Performance Aid with Text Added to a SelfExplanatory Visual 148 6.7 Learning Is Better (Left Bars) and Instructional Time Shorter (Right Bars) from Diagrams Alone 149 6.8 A Performance Aid with Separated Text and Diagram

151

6.9 A Performance Aid with Integrated Text and Diagram

151

6.10 Integrated Computer-Based Training Led to Better Learning of High Complexity Skills 152 6.11 An Online Performance Aid Integrates Text with Application 153 6.12 An Online Performance Aid Shows Steps Out of Context of Application 154 6.13 A PowerPoint Performance Aid Aligned Next to the Running Application 155 7.1 Pretraining Sheet Presenting Parts and Functions of Parts in a Car Braking Process 164 7.2 One Frame from Multimedia Pretraining Presenting Parts and Functions of Parts in a Car Braking Process 165 7.3 Learning from Process Lessons with and Without Pretraining in Three Experiments 166 7.4 The Segmented Lesson Version Teaching an Insulation Resistance Test 170

blof.qxd 9/29/05 19:07 Page 384

384

L i s t o f F i g u r e s a n d Ta b l e s

7.5 The Unsegmented Lesson Version Teaching an Insulation Resistance Test 171 7.6 Segmented Lessons Are More Efficient for Learning Complex Content 173 7.7 A Typical Directive Course Architecture

175

7.8 A Virtual Office Setting for Bank Loan Whole Task Course 177 7.9 A Directive and Whole Task Outline for a Course on Use of Virtual Classroom Software 178 7.10 Learning from Concise Training Is Better in Paper Version That Is Learner Controlled 182 7.11 Learning Is Better from Learner-Paced Multimedia Training 183 8.1 Part of a Worked Example from Asynchronous e-Lesson on Constructing Formulas in Excel from the CD 191 8.2 An Algebra Worked Example Displayed in Text

191

8.3 Part of a Completion Example from an Asynchronous Load-Managed Lesson on Excel on the CD 195 8.4 A Faded Worked Example from the Asynchronous Excel Lesson on the CD 198 8.5 A Conceptual Model of Backwards Faded Completion Examples 199 8.6 The Plan of an Experiment Comparing Worked ExamplesPractice Pairs to All Practice as Learners Gain Expertise 200 8.7 Worked Examples Are More Efficient for Novices; All Problems Are More Efficient for Experts 201 8.8 A Sample Completion Problem with Backwards Fading 202 8.9 Instructor Explains the Demonstration Verbally in Virtual Classroom 204 8.10 A Print-Based Worked Example That Integrates Text into Diagram to Minimize Split Attention. 205

blof.qxd 9/29/05 19:07 Page 385

385

L i s t o f F i g u r e s a n d Ta b l e s

8.11 A Print-Based Worked Example That Splits Attention Between Diagram and Related Text 206 8.12 Examples Described with Integrated Text or Audio Improve Learning 207 9.1 The Fortress Story

220

9.2 Three Problems Used in Worked Examples Research

223

9.3 Lessons with Worked and Completion Examples Resulted in Better Learning of Far Transfer Test Problems Than All-Problem Lessons 224 9.4 Efficiency Is Greatest for High Variable Worked Examples 225 9.5 A Student Self-Explanation of a Physics Problem

227

9.6 A Worked Example with First Worked Step Requiring a Self-Explanation 230 9.7 Learning Is Better from Faded Worked Examples with Added Questions That Promote Self-Explanations 231 9.8 A Question in Excel Lesson Requires Learner to Identify Rule Associated with Worked Step 232 9.9 Better Learning of Complex Content from Study in Initial Sessions and Rehearsal in Later Sessions 237 9.10 Study Followed by Rehearsal Results in Better Learning of Complex Content 238 10.1 A Disordinal Interaction Between Method Y and Z for Type A and Type B Learners 249 10.2 A Generalized Plan of a Staged Experiment

253

10.3 Excerpts from Low and High Coherent Texts (edits are underlined in the coherent version). 255 10.4 Opposite Learning Outcomes from High and Low Coherent Text by High and Low Prior Knowledge Readers 256 10.5 Answering Questions During Reading Had Opposite Learning Effects Among Readers of Different Expertise 258 10.6 Text Plus Diagram Lesson Version of How a Brake Works

260

blof.qxd 9/29/05 19:07 Page 386

386

L i s t o f F i g u r e s a n d Ta b l e s

10.7 Lesson Versions with Diagrams Aid Understanding of Low but Not High Prior Knowledge Learners 261 10.8 Diagrams Alone Resulted in Better Learning with More Expert Learners 262 10.9 A Worked Example Using Audio to Explain How to Use the Diagram to Determine Cutting Speeds for Drills 263 10.10 Diagrams Plus Words Are More Efficient for Novices; Diagrams Alone Are More Efficient for Experts 264 10.11 Lessons with All Problems Led to Better Learning of Experienced Learners 266 10.12 A Comparison of Directive with Guided Discovery Lesson Design on Learning and Training Time of Novice Learners on Simple Tasks 269 10.13 A Comparison of Directive and Guided Discovery Lesson on Learning of Complex Tasks by Novice and Experienced Participants 269 11.1 Alternative First Steps to Solve an Algebra Problem Among Learners of Diverse Experience 277 11.2 Levels of Algebraic Equation Problems That Incorporate Increasing Numbers of Skills 278 11.3 The First of a Three-Item Pretest in an Excel Lesson on the CD 279 11.4 The Third of a Three-Item Pretest in an Excel Lesson on the CD 280 11.5 A Diagnostic Test Given After Completion of Topic 1 11.6 An Overview of an Adaptive Testing Learning Sequence

282 283

11.7 A Summary of the Instruction Methods at Each Level of an Adaptive Training Plan 284 12.1 An Asynchronous Course with Lines Used to Integrate Text 306 12.2 The Screen Used in a Synchronous Course Includes On-Screen Memory Support 309

blof.qxd 9/29/05 19:07 Page 387

387

L i s t o f F i g u r e s a n d Ta b l e s

12.3 A Handout That Summarizes Procedural Steps A.1 What Is a Z Score?

310

333

A.2 The Efficiency Graph with Hypothetical Plots A.3 A Worked Example of Efficiency Calculation

333 335

A.4 Efficiency Graph for Worked Example in Figure A.3

335

A.5 Data from Experiment Comparing Text and Audio Explanations of a Visual (Tindall-Ford, Chandler, & Sweller, 1997) 336 A.6 Efficiency Graph for Data in Figure A.5

337

Tables 1.1 A Summary of Experiments Demonstrating a Modality Effect 20 7.1 Four Architectures of Instruction

174

8.1 Worked Example Problem Pairs Result in Faster Learning and Performance 192 8.2 Worked Examples and Completion Examples Are More Efficient Than All-Practice Lessons 196 12.1 Applying Cognitive Load Theory to Your Training

295

A.1 Data from Tindall-Ford, Chandler, and Sweller (1997) Study Converted to Z Scores 337

blof.qxd 9/29/05 19:07 Page 388

bhowto.qxd 9/29/05 19:05 Page 389

H O W

T O

U S E

T H E

C D - R O M

System Requirements PC with Microsoft Windows 98SE or later Mac with Apple OS version 8.6 or later

Using the CD With Windows To view the items located on the CD, follow these steps: 1. Insert the CD into your computer’s CD-ROM drive. 2. A window appears with the following options: Contents: Allows you to view the files included on the CD-ROM. Software: Allows you to install useful software from the CD-ROM. Links: Displays a hyperlinked page of websites. Author: Displays a page with information about the Author(s).

389

bhowto.qxd 9/29/05 19:05 Page 390

390

How to Use the Accompanying CD-ROM Contact Us: Displays a page with information on contacting the publisher or author. Help: Displays a page with information on using the CD. Exit: Closes the interface window. If you do not have autorun enabled, or if the autorun window does not appear, follow these steps to access the CD: 1. Click Start-Run. 2. In the dialog box that appears, type d: start.exe, where d is the letter of your CD-ROM drive. This brings up the autorun window described in the preceding set of steps. 3. Choose the desired option from the menu. (See Step 2 in the preceding list for a description of these options.)

In Case of Trouble If you experience difficulty using the CD-ROM, please follow these steps: 1. Make sure your hardware and systems configurations conform to the systems requirements noted under “System Requirements” above. 2. Review the installation procedure for your type of hardware and operating system. It is possible to reinstall the software if necessary. To speak with someone in Product Technical Support, call 800–762–2974 or 317–572–3994 M–F 8:30 A.M.–5:00 P.M. EST. You can also get support and contact Product Technical Support through our website at www.wiley.com/techsupport. Before calling or writing, please have the following information available: •

Type of computer and operating system

•

Any error messages displayed

•

Complete description of the problem.

It is best if you are sitting at your computer when making the call.