Designing Inclusive Systems: Designing Inclusion for Real-world Applications

Designing Inclusive Systems Patrick Langdon John Clarkson Peter Robinson Jonathan Lazar Ann Heylighen • • Editors ...

Author: Patrick Langdon | John Clarkson | Peter Robinson | Jonathan Lazar | Ann Heylighen

37 downloads 2099 Views 4MB 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

Designing Inclusive Systems

Patrick Langdon John Clarkson Peter Robinson Jonathan Lazar Ann Heylighen •

•

Editors

Designing Inclusive Systems Designing Inclusion for Real-world Applications

123

Patrick Langdon Department of Engineering Cambridge Engineering Design Centre University of Cambridge Trumpington Street Cambridge CB2 1PZ, UK

Jonathan Lazar Universal Usability Laboratory Department of Computer and Information Sciences Towson University,York Road 8000 Towson, MD 21252, USA

John Clarkson Department of Engineering Cambridge Engineering Design Centre University of Cambridge Trumpington Street Cambridge CB2 1PZ, UK

Ann Heylighen Department of Architecture Urbanism and Planning Katholieke Universiteit Leuven Kasteelpark Arenberg 1/2431 3001 Leuven, Belgium

Peter Robinson Computer Laboratory University of Cambridge JJ Thomson Avenue, Madingley Road William Gates Building 15 Cambridge CB3 0FD, UK

ISBN 978-1-4471-2866-3 DOI 10.1007/978-1-4471-2867-0

e-ISBN 978-1-4471-2867-0

Springer London Heidelberg New York Dordrecht British Library Cataloguing in Publication Data A catalogue record for this book is available from the British Library Library of Congress Control Number: 2012933436 Ó Springer-Verlag London 2012 This work is subject to copyright. All rights are reserved by the Publisher, whether the whole or part of the material is concerned, specifically the rights of translation, reprinting, reuse of illustrations, recitation, broadcasting, reproduction on microfilms or in any other physical way, and transmission or information storage and retrieval, electronic adaptation, computer software, or by similar or dissimilar methodology now known or hereafter developed. Exempted from this legal reservation are brief excerpts in connection with reviews or scholarly analysis or material supplied specifically for the purpose of being entered and executed on a computer system, for exclusive use by the purchaser of the work. Duplication of this publication or parts thereof is permitted only under the provisions of the Copyright Law of the Publisher’s location, in its current version, and permission for use must always be obtained from Springer. Permissions for use may be obtained through RightsLink at the Copyright Clearance Center. Violations are liable to prosecution under the respective Copyright Law. The use of general descriptive names, registered names, trademarks, service marks, etc. in this publication does not imply, even in the absence of a specific statement, that such names are exempt from the relevant protective laws and regulations and therefore free for general use. While the advice and information in this book are believed to be true and accurate at the date of publication, neither the authors nor the editors nor the publisher can accept any legal responsibility for any errors or omissions that may be made. The publisher makes no warranty, express or implied, with respect to the material contained herein. Cover Photography: Matt Willox in Design Futures, part of Sheffield Hallam University Packaging Accessibility Analysis, Dr Alaster Yoxall: Lab4living, Sheffield Hallam Universit Museum M in Leuven, Belgium designed by Stéphane Beel Architecten: Peter-Willem Vermeersch Cover design: eStudio Calamar S.L. Printed on acid-free paper Springer is part of Springer Science+Business Media (www.springer.com)

Preface

The Cambridge Workshops on Universal Access and Assistive Technology (CWUAAT) are a series of workshops held at a Cambridge University College every two years. This volume: “DESIGNING INCLUSIVE SYSTEMS” comes from the 6th in this series of highly successful events. The workshops are characterised by a single session running over three days in pleasant surroundings with delegates from home and abroad staying on site. Feedback suggests that allowing speakers longer presentation times, carrying discussion on through sessions into plenaries and shared mealtimes generates a highly cooperative and creative academic environment that is both enjoyable and informative. The workshop theme: “Designing inclusion for real-world applications” refers to the emerging potential and relevance of the latest generations of inclusive design thinking, tools, techniques and data, to mainstream project applications such as healthcare and the design of working environments. Inclusive Design Research involves developing tools and guidance enabling designers to design for the widest possible population, for a given range of capabilities. In the context of demographic changes leading to an increasing number of older people, the general field of inclusive design research strives to relate the capabilities of the population to the design of artefacts, environments and technology by better characterising the user and the task demand. Inclusive populations of older people, for example, contain a greater variation in sensory, cognitive and physical capabilities. These variations may be co-occurring and rapidly changing leading to a demanding design environment. Previous research developments in inclusive design have addressed issues of matching product and task demand to users’ capabilities in the context of simple daily living activities or specific products. New research developments are now extending the scope of the inclusive design approach into real-world applications by forming interdisciplinary links with systems engineering, industrial product design, healthcare and medical device design as well as education, policy development and architecture. This is a necessary stage of research because once design techniques and materials are fully developed for knowledge transfer, a v

Preface

vi

proving ground is required in real-world application and industry. This proving ground then tests the impact made by the original research. As in the previous years, this book contains the best reviewed papers invited for oral presentation. The papers that have been included were selected by blind peer review carried out by an international panel of currently active researchers. The chapters forming the book represent an edited sample of current national and international research in the fields of inclusive design, universal access, and assistive and rehabilitative technology. In the 2012 workshop, as well as the typical Inclusive Design themes of measuring demand and capability; emergent technologies, and design for inclusion, there has also been more focus on new themes such as cognitive interaction with new technologies, architecture, and healthcare. This reflects the newly developing transdisciplinary perspectives and ongoing research agendas. For example, can medical and neuroscientific models of thinking impairment be harmonised with functional descriptions to assist more inclusive design? Is it possible to motivate older generations to use modern healthcare software by better understanding the psychology of human motivation? Can we identify and quantify the differences between designers’ and users’ mental models of a product? In addition, researchers are increasingly investigating how public policies; both from governments and international non-governmental organisations, influence inclusive and accessible design, as well as the usage and adoption of assistive technology by individuals. Healthcare is a forcing domain: how can we provide architects with sufficient evidence to enable them to design healthcare buildings that better anticipate the needs of patients lying in a hospital bed? For this CWUAAT we have extended the editorial panel to include two esteemed colleagues, Ann Heylighen from KU Leuven in Belgium, and Jonathan Lazar from Towson University in the US. This reflects the growing importance of particular interdisciplinary fields such as inclusive architecture, and public policy related to inclusive design, to the CWUAAT workshops. It also acknowledges the substantial international contributions that have been made over the series. There are five main themes: I.

Designing for the Real-world addresses the application of Inclusive Design techniques in healthcare, public facilities and services, and hazardous traditional industries; II. Measuring Demand and Capabilities looks at ways of measuring capability-demand relationships for actual tasks, software, devices and buildings; III. Designing Cognitive Interaction with Emerging Technologies draws together a number of threads related to cognition including the alignment of design and user mental models, motivating older users and unifying models of cognitive impairment; IV. Design for Inclusion is a space specifically for design issues in inclusive design, from sampling through to policy and novel new ways to inform the designer about inclusive design features;

Preface

vii

V. Designing Inclusive Architecture highlights specific cases, such as inclusive heritage, architecture for dementia and virtual environment tools for design. In the tradition of CWUAAT, we have solicited and accepted contributions over a wide range of topics, both within individual themes and also across the workshop’s scope. We ultimately hope to generate more interdisciplinary dialogues based on focused usage cases that can provide the discipline necessary to drive further novel research, leading to better designs. The aim is to impact industry and end-users as well governance and public design, thereby effectively reducing exclusion and difficulty in people’s daily lives and society. We would like to thank all those authors and contributors who have submitted to CWUAAT 2012 and to the preparation of this book. Many thanks are also due to the reviewing members of the Programme Committee who continue to support the workshop series. Finally, thanks are particularly due to Mari Huhtala and Suzanne Williams, who both play a key role in bringing the resulting publication to fruition between final submission and the Workshop itself. As in previous years, we are grateful to the staff at Fitzwilliam College for their patience and service. Pat Langdon, John Clarkson, Peter Robinson Jonathan Lazar and Ann Heylighen The CWUAAT Editorial Committee University of Cambridge March 2012

Contents

List of Contributors …………………………………………………………..xiii

Part I Designing for the Real-world 1.

Hospital Reality from a Lying Perspective: Exploring a Sensory Research Approach M. Annemans, C. Van Audenhove, H. Vermolen and A. Heylighen .................................................................................. 3

2.

Inclusive Bus Travel - A Psychosocial Approach F. Nickpour, P.W. Jordan and H. Dong.............................................. 13

3.

Safe and Inclusive Design of Equipment Used in the Minerals Industry T. Horberry and T. Cooke .................................................................. 23

4.

Deploying a Two-player System for Arm Rehabilitation in Schools R.J. Holt, A.P.H. Weightman, J.F. Gallagher, N. Preston, M.C. Levesley, M. Mon-Williams and B. Bhakta ................................ 33

5.

Evaluating the Accessibility and Usability of Blogging Platforms for Blind Users B. Wentz, M. Cirba, N. Kharal, J. Moran and M. Slate ...................... 43

ix

x

Contents

Part II Measuring Demand and Capabilities 6.

A Population Perspective on Mobile Phone Related Tasks M. Bradley, S. Waller, J. Goodman-Deane, I. Hosking, R. Tenneti, P.M. Langdon and P.J. Clarkson ....................................................... 55

7.

How to Use Virtual and Augmented Reality Techniques to Design Highly Usable Human-machine Interfaces S. Ceccacci, M. Germani and M. Mengoni ....................................... 65

8.

Development and Evaluation of Sonified Weather Maps for Blind Users R. Weir, B. Sizemore, H. Henderson, S. Chakraborty and J. Lazar........................................................................................ 75

9.

Achieving Inclusion in Public Spaces: A Shopping Mall Case Study Y. Afacan............................................................................................ 85

10. Visibility Prediction Software: Five Factors of Contrast Perception for People with Vision Impairment in the Real World H. Dalke, A. Corso, G. Conduit and A. Riaz ...................................... 93

Part III Designing Cognitive Interaction with Emerging Technologies 11. Intrinsic Motivation and Design of ICT for the Ageing Population T.S. Goldhaber, P.M. Langdon and P.J. Clarkson........................... 105 12. A Framework for Studying Cognitive Impairment to Inform Inclusive Design E. Jokisuu, P.M. Langdon and P.J. Clarkson................................... 115 13. Interactive Error Correction Using Statistical Language Models in a Client-server Interface for Editing Mathematical Text D. Attanayake, G. Hunter, E. Pfluegel and J. Denholm-Price.............................................................................. 125 14. Understandable by Design: How Can Products be Designed to Align with User Experience? A. Mieczakowski, P.M. Langdon, R.H. Bracewell, J.J. Patmore and P.J. Clarkson............................................................................. 133

Contents

xi

Part IV Design for Inclusion 15. Design Advisor: How to Supply Designers with Knowledge about Inclusion? E. Zitkus, P.M. Langdon and P.J. Clarkson ..................................... 145 16. From Guinea Pigs to Design Partners: Working with Older People in ICT Design R. Edlin-White, S. Cobb, A. Floyde, S. Lewthwaite, J. Wang and J. Riedel ........................................................................................... 155 17. When Users Cannot be Included in Inclusive Design R. Herriott ......................................................................................... 165 18. What is Good Design in the Eyes of Older Users? N. Goddard and C. Nicolle ............................................................... 175 19. Equal Access to Information? Evaluating the Accessibility of Public Library Web Sites in the State of Maryland J. Lazar, B. Wentz, C. Akeley, M. Almuhim, S. Barmoy, P. Beavan, C. Beck, A. Blair, A. Bortz, B. Bradley, M. Carter, D. Crouch, G. Dehmer, M. Gorman, C. Gregory, E. Lanier, A. McIntee, R. Nelson Jr., D. Ritgert, R. Rogers Jr., S. Rosenwald, S. Sullivan, J. Wells, C. Willis, K. Wingo-Jones and T. Yatto ............................. 185 20. Clustering User Data for User Modelling in the GUIDE Multi-modal Set-top Box P.M. Langdon and P. Biswas ........................................................... 195

Part V Designing Inclusive Architecture 21. Inclusive Built Heritage as a Matter of Concern: A Field Experiment A. Heylighen ..................................................................................... 207 22. Designing a Virtual Environment Framework for Improving Guidance for the Visually Impaired S. Kammoun, M.J-M. Macé, B. Oriola and C. Jouffrais ................... 217 23. Spatial Clues for Orientation: Architectural Design Meets People with Dementia I. Van Steenwinkel, C. Van Audenhove and A. Heylighen .............. 227 Index of Contributors …………………………………………………..…..237

List of Contributors

Afacan Y., Department of Interior Architecture and Environmental Design, Faculty of Art, Design and Architecture, Bilkent University, Ankara, Turkey Akeley C., Department of Computer and Information Sciences, Universal Usability Laboratory, Towson University, US Almuhim M., Department of Computer and Information Sciences, Universal Usability Laboratory, Towson University, US Annemans M., Department of Architecture, Urbanism and Planning, Katholieke Universiteit Leuven, Heverlee, Belgium Attanayake D., Faculty of Computing, Information Systems and Mathematics (CISM), Kingston University, London, UK Barmoy S., Department of Computer and Information Sciences, Universal Usability Laboratory, Towson University, US Beavan P., Department of Computer and Information Sciences, Universal Usability Laboratory, Towson University, US Beck C., Department of Computer and Information Sciences, Universal Usability Laboratory, Towson University, US Bhakta B., Academic Department of Rehabilitation Medicine, Faculty of Medicine and Health, University of Leeds, Leeds, UK Biswas P., Engineering Design Centre, Department of Engineering, University of Cambridge, Cambridge, UK Blair A., Department of Computer and Information Sciences, Universal Usability Laboratory, Towson University, US Bortz A., Department of Computer and Information Sciences, Universal Usability Laboratory, Towson University, US Bracewell R.H., Engineering Design Centre, Department of Engineering, University of Cambridge, Cambridge, UK xiii

xiv

List of Contributors

Bradley B., Department of Computer and Information Sciences, Universal Usability Laboratory, Towson University, US Bradley M., Engineering Design Centre, Department of Engineering, University of Cambridge, Cambridge, UK Carter M., Department of Computer and Information Sciences, Universal Usability Laboratory, Towson University, US Ceccacci S., Department of Management and Industrial Organisation, Marche Polytechnic University, Ancona, Italy Chakraborty S., Department of Computer and Information Sciences, Universal Usability Laboratory, Towson University, US Cirba M., Division of Business, Management and Technology, Keystone College, La Plume, PA, US Clarkson P.J., Engineering Design Centre, Department of Engineering, University of Cambridge, Cambridge, UK Cooke T., Minerals Industry Safety and Health Centre, Sustainable Minerals Institute, The University of Queensland, St Lucia, Australia Cobb S., Human Factors Research Group, Faculty of Engineering, The University of Nottingham, Nottingham, UK Conduit G.J., Department of Condensed Matter Physics, Weizmann Institute of Science, Rehovot, Israel Corso A., Design Research Centre, Faculty of Art, Design and Architecture, Kingston University, UK Crouch D., Department of Computer and Information Sciences, Universal Usability Laboratory, Towson University, US Dalke H., Design Research Centre, Faculty of Art, Design and Architecture, Kingston University, UK Dehmer G., Department of Computer and Information Sciences, Universal Usability Laboratory, Towson University, US Denholm-Price J., Faculty of Computing, Information Systems and Mathematics (CISM), Kingston University, London, UK Edlin-White R., Human Factors Research Group, Faculty of Engineering, The University of Nottingham, Nottingham, UK Elton E., Ergonomics and Safety Research Institute, Loughborough University, Loughborough, UK Floyde A., Human Factors Research Group, Faculty of Engineering, The University of Nottingham, Nottingham, UK Gallagher J.F., Institute of Engineering Systems and Design, School of Mechanical Engineering, University of Leeds, Leeds, UK Germani M., Department of Industrial Engineering and Mathematical Sciences, Marche Polytechnic University, Ancona, Italy Goddard N., Ergonomics and Safety Research Institute, Loughborough University, Loughborough, UK Goldhaber T.S., Engineering Design Centre, Department of Engineering, Cambridge University, Cambridge, UK Goodman-Deane J., Engineering Design Centre, Department of Engineering, Cambridge University, Cambridge, UK

List of Contributors

xv

Gorman C., Department of Computer and Information Sciences, Universal Usability Laboratory, Towson University, US Gregory C., Department of Computer and Information Sciences, Universal Usability Laboratory, Towson University, US Henderson H., Department of Computer and Information Sciences, Universal Usability Laboratory, Towson University, US Herriott R., Aarhus School of Architecture, Aarhus, Denmark Heylighen A., Department of Architecture, Urbanism and Planning, Katholieke Universiteit Leuven, Leuven, Belgium Holt R.J., Institute of Engineering Systems and Design, School of Mechanical Engineering, University of Leeds, Leeds, UK Horberry T., Minerals Industry Safety and Health Centre, Sustainable Minerals Institute, The University of Queensland, St Lucia, Australia Hosking I., Engineering Design Centre, Department of Engineering, University of Cambridge, Cambridge, UK Hunter G., Faculty of Computing, Information Systems and Mathematics (CISM), Kingston University, London, UK Jokisuu E., Engineering Design Centre, Department of Engineering, University of Cambridge, Cambridge, UK Jouffrais C., IRIT-CNRS, University of Toulouse, Toulouse, France Kammoun S., IRIT-CNRS, University of Toulouse, Toulouse, France Kharal N., Division of Business, Management and Technology, Keystone College, La Plume, PA, US Langdon P.M., Engineering Design Centre, Department of Engineering, University of Cambridge, Cambridge, UK Lanier E., Department of Computer and Information Sciences, Universal Usability Laboratory, Towson University, US Lazar J., Department of Computer and Information Sciences, Universal Usability Laboratory, Towson University, US Levesley M.C., Institute of Engineering Systems and Design, School of Mechanical Engineering, University of Leeds, Leeds, UK Lewthwaite S., Department of American and Canadian Studies, University of Nottingham, Nottingham, UK Macé M.J-M., IRIT-CNRS, University of Toulouse, Toulouse, France Mengoni M., Design Tools and Methods Group, Marche Polytechnic University, Ancona, Italy Mieczakowski A., Engineering Design Centre, Department of Engineering, University of Cambridge, Cambridge, UK McIntee A., Department of Computer and Information Sciences, Universal Usability Laboratory, Towson University, US Mon-Williams M., Institute of Psychological Sciences, University of Leeds, Leeds, UK Moran J., Division of Business, Management and Technology, Keystone College, La Plume, PA, US Nelson Jr.D., Department of Computer and Information Sciences, Universal Usability Laboratory, Towson University, US

xvi

List of Contributors

Nickpour F., Inclusive Design Research Group, School of Engineering and Design, Brunel University, West London, UK Nicolle C., Ergonomics and Safety Research Institute, Loughborough University, Loughborough, UK Oriola B., IRIT-CNRS, University of Toulouse, Toulouse, France Patmore J.J., Member of Pembroke College, University of Cambridge, Cambridge, UK Pfluegel E., Faculty of Computing, Information Systems and Mathematics (CISM), Kingston University, London, UK Preston N., Academic Department of Rehabilitation Medicine, University of Leeds, Leeds, UK Riaz A., Design Research Centre, Faculty of Art, Design and Architecture, Kingston University, UK Riedel J., University of Nottingham Business School, University of Nottingham, Nottingham, UK Ritgert D., Department of Computer and Information Sciences, Universal Usability Laboratory, Towson University, US Rogers Jr.R., Department of Computer and Information Sciences, Universal Usability Laboratory, Towson University, US Rosenwald S., Department of Computer and Information Sciences, Universal Usability Laboratory, Towson University, US Sizemore B., Department of Computer and Information Sciences, Universal Usability Laboratory, Towson University, US Slate M., Division of Business, Management and Technology, Keystone College, La Plume, PA, US Sullivan S., Department of Computer and Information Sciences, Universal Usability Laboratory, Towson University, US Tenneti R., School of Primary, Aboriginal and Rural Health Care, Faculty of Medicine, Dentistry and Health Sciences, The University of Western Australia, Crawley, Western Australia Van Audenhove C., Faculty of Medicine, Katholieke Universiteit Leuven, Leuven, Belgium Van Steenwinkel I., Department of Architecture, Urbanism and Planning, Katholieke Universiteit Leuven, Leuven, Belgium Vermolen H., Osararchitects nv, Antwerpen, Belgium Waller S.D., Engineering Design Centre, Department of Engineering, Cambridge University, Cambridge, UK Wang J., University of Nottingham Business School, University of Nottingham, Nottingham, UK Weightman A.P.H., School of Mechanical Engineering, University of Leeds, Leeds, UK Weir R., Department of Computer and Information Sciences, Universal Usability Laboratory, Towson University, US Wells J., Department of Computer and Information Sciences, Universal Usability Laboratory, Towson University, US Wentz B., Department of Computer Science and Information Technology, Frostburg State University, Frostburg, MD, US

List of Contributors

xvii

Willis C., Department of Computer and Information Sciences, Universal Usability Laboratory, Towson University, US Wingo-Jones K., Department of Computer and Information Sciences, Universal Usability Laboratory, Towson University, US Yatto T., Department of Computer and Information Sciences, Universal Usability Laboratory, Towson University, US Zitkus E., Engineering Design Centre, Department of Engineering, Cambridge University, Cambridge, UK

Part I

Designing for the Real-world

Chapter 1 Hospital Reality from a Lying Perspective: Exploring a Sensory Research Approach M. Annemans, C. Van Audenhove, H. Vermolen and A. Heylighen

1.1 Real Buildings, Real Experiences Despite many efforts by healthcare providers, for most people a hospital stay is rarely a pleasant experience. The hospital building as such is part of this perception. Moreover, the specific situation of a hospital stay is largely determined by the material reality of the organisation. Studies on hospital environments tend to single out one particular aspect, e.g. the view through the window, or presence of green (Ulrich 1984a, 1984b) and try to prove its clinical outcome. Yet they fail to translate their results to the design of real-life settings (Rubin et al., 1998, Cbz 2008). Moreover, the influence of patients’ peculiar perspective, i.e. lying in a hospital bed, on the way they experience the reality of the hospital is largely under researched. The overall objective of our research is therefore to investigate what spatial aspects influence patients’ well-being in a hospital setting through an improved understanding of people’s spatial experience from a lying perspective. By developing a better understanding of the relationship between the patient, the objects that feature in his or her hospital life, especially the bed and the building, we hope to enable architects to design buildings that add to the healing character of the hospital environment. Ultimately, we aim to provide architects with sufficient evidence to design healthcare buildings that can better anticipate the needs of patients and other users. Since our research aims to gain insight in how patients experience a hospital from a lying perspective, we need a manner to make their spatial experience more explicit. Therefore, we explored different ways for patients to document their spatial experience. In this paper, we report on a pilot study which explores how three patients with a very different profile each deal with this task in their own way. The empirical material collected is not only very rich in itself, but also inspires the participating patients to talk about those aspects of the building that

P. Langdon et al. (eds.), Designing Inclusive Systems, DOI: 10.1007/978-1-4471-2867-0_1, © Springer-Verlag London 2012

3

4

Annemans et al.

affect them most. Certain themes frequently return in the conversations, yet the goal at this stage in our research is not so much to obtain valid information, but rather to explore the possibilities of using participant collected material to facilitate the interviews.

1.2 Sensory Reality Research about healing or wholesome environments in health care settings is mostly conducted in the field of evidence based design (EBD). This notion found its origin in the analogy with other evidence-based approaches to research and practice. When it comes to buildings for health care especially, evidence based medicine was a source of inspiration. In EBD studies evidence for the healing outcome of building aspects is being collected (Ulrich et al., 2004). Mostly individual aspects are investigated and clinical output is highly valued (Rubin et al., 1998). However, almost all reports that collect and evaluate the results of the conducted studies have to conclude that the evidence is not ready to be transposed to the real world. To start with, the settings in which the trials are carried out single out individual aspects and, as such, are not representative of real-world hospital settings. Secondly the studies do not provide insight in how the different aspects relate to each other (Rubin et al., 1998; Van den Berg, 2005; Cbz, 2008). However, a wholesome environment involves more than the measurable aspects that withstand the evaluation of EBD studies. After all, architecture is experienced through the senses (Rasmussen, 1964; Pallasmaa, 2005), so how a place feels, smells, sounds, and looks defines our impression of it. In this multisensory experience of the environment, time, movement and activities play a major role. The senses are indeed a key factor in the experience of everyday reality (Ingold, 2000). If we are to develop an articulate understanding of patients’ actual experience from a lying perspective, as is the aim of our research, we thus need to collect firsthand information about what they feel, sense, and think during their hospital stay, not only while in their room, but also while travelling through the hospital building.

1.3 Insight/Inside Spatial Experience In studying patients’ spatial experience from a lying perspective, we explore and develop a research methodology that respects the interrelatedness of how they experience the reality of things in the hospital, addressing different aspects of the built environment as well as their complex interaction. To obtain a broad spectrum of empirical material, participants should be challenged to interrogate all their senses. The method of collecting material should also appeal to a very diverse group of patients within the hospital, making them feel at ease with their participation. Only when these criteria are met, will the material be rich enough to gain a nuanced understanding of the complex reality of everyday hospital experience.

Hospital Reality from a Lying Perspective

5

In the first months of our research we conducted interviews with various actors in the field (medical staff, patients, technical directors of hospitals, hospital architects) to obtain a profound understanding of the meaning of the bed in the hospital (Annemans et al., 2011). Each interviewee shed light on the topic from his/her specific perspective. This resulted in a good understanding of the complexity of hospital design. Although our research focuses on the experience of (lying) patients, this background makes us aware of the importance of reflecting on and taking into account certain undeniable technical and organisational realities. During these interviews it became clear that people find it hard to talk about their spatial experience, especially those unfamiliar with the design and construction process. This lack of conversation skills about the research topic makes it hard to obtain the rich empirical material we are looking for. Part of the difficulty can be explained by a lack of vocabulary when talking about space (Franck and von Sommaruga Howard, 2009). Additionally, due to the complexity of experience, it is not easy to ensure that it is explored throughout the entire range of its various articulations (Throop, 2003). These obstacles force us to look for a suitable research approach that can give us access to patients’ personal ways of experiencing a hospital environment. For this reason we decided to explore alternative interviewing techniques. Aware of the difficulty for patients of expressing their spatial experience and given the explorative phase of the research, we opted to ask the participating patients only one, very broad question: “Can you document the hospital experience from a lying perspective?” After a short introduction patients are then invited to document their spatial experience in any way that pleases them. They are provided with pens, pencils, note block, drawing paper and a camera (with the possibility of recording sound and movies), or they can use their own camera. Afterwards the material produced during this process is used to facilitate a semi-structured interview with the participating patient about the spatial qualities of the building. As already pointed out, we experience the built environment through all our senses. For this reason we looked into the use of sensory methods. Visual techniques like photo-elicitation where people are interviewed using photographs are not new (Collier, 1967). Over the years the technique has gained popularity in various fields such as visual ethnography, visual anthropology, visual sociology, and visual cultural studies (Pink, 2007). A sensory method does not guarantee access to sensory realms; for example, the visual is not necessarily best accessed by a visual method (Mason and Davies, 2009). Nevertheless, introducing photography can be seen as a first contribution to a more ‘sensually complete’ methodology (Warren, 2002). In order to capture the real-life experience instead of a fictive reality constructed by the researcher, however, we shifted our focus beyond photo-elicitation to photo-production (Radley, 2010). In line with the work of other researchers (Herssens and Heylighen, 2009; Radley, 2010), our interest lies not so much in the meaning of pictures, we want to gain a good understanding not only of what has been made visible, but also why, and how. Indeed, how images are recorded may also tell us something about the hospital experience of the participating patients. As taking pictures in hospitals is not always appropriate or allowed (Radley and Taylor, 2003), we did not provide the patients with a camera only, and limit their way of expression to photographs, but

6

Annemans et al.

also gave them the opportunity to take notes or make a drawing. This approach yields richer empirical material, which initiates different topics in the interviews. In the elicitation process, there is no reason why the drawings and plans should be treated any differently to the photographs (Harper, 2002).

1.4 “Talking” Patients, “Speaking” Hospitals 1.4.1 Who is Talking To explore the methodology’s possibilities and limitations we try it out with three very different, pragmatically chosen patients. Because of the explorative stage of the research the validity and generalisability of the obtained results is subordinate to the test of the methodology and the participants’ engagement. The three of them are in the hospital for different reasons while documenting their stay. Although all are provided with the same equipment, they choose very diverse ways to document their experience. There is Mrs A, who stays a few days in the hospital for rather serious surgery. Through photographs she elaborately documents her experiences in two rooms, a double room before the operation, a single room afterwards. Despite her commitment to the task, she admits she does not feel comfortable or able to take the camera into the hallway or to treatment. Mrs B has a long history of hospital visits, and participates in this study when going for a check-up in a hospital where she has been a regular patient. Since she is not lying herself while travelling through the hospital this time, she interprets the question as illustrating what she remembers from previous visits and uses it to signal points of improvement in the treatment of patients. Depending on what she feels is appropriate, she switches between taking pictures and drawing sketches of what she observes. Even a brief story emerges. Mr C is on dialysis, making him visit the same part of the hospital several times a week. Taking pictures he considers redundant, as the researcher can see everything herself when conducting the interview. However, he does have some ideas about how the department should be rearranged, improving most of the disadvantages he experiences today. He sketches a plan of the current situation to explain to the researcher how it is and re-sketches it to illustrate how he would alter it.

1.4.2 How Patients “Talk” How patients choose to document their experience tells us a lot about how they experience certain situations. The pencils and papers, for instance, were initially introduced as a pragmatic solution to the problem that taking pictures in hospitals is not always appropriate or allowed. However, they turn out to provide insights that would have stayed undiscovered had only a camera been used.

Hospital Reality from a Lying Perspective

7

As one patient points out, sketching an awkward situation afterwards is much less confronting for the people involved than taking a picture as it unfolds. Another patient does not see the point of taking pictures of a building that the researcher can observe herself, however he does have some suggestions on how a future hospital setting should be designed. Whereas, for him, a photograph just shows an existing situation, drawing provides the opportunity to manipulate reality. Since we use the recorded images as a base for interviews, it does not come as a surprise that also this kind of material serves as an interesting source of information about how the hospital building is perceived, or which aspects are appreciated or considered problematic. Often just starting the conversation is enough to divert to aspects not, or only slightly, related to the collected material. As if enabling participants to start the conversation from their viewpoint is enough to open the door to unexpected insights.

1.4.3 What Hospitals “Say” Through the material collected by the participants, the hospitals tell us both about their organisation and about the building and its interior. The collected images and accompanying narratives give us a good overview of what kind of material can be expected. Obviously there are the images that literally represent visual elements in the environment, such as the ceiling. However mostly, here too, the explanation why the image is recorded reveals supplementary useful insights. In addition some images are not taken for what they show but have an iconographic meaning, telling us about other sensory experiences than the visual. Finally, the produced images can be a representation of what could be, giving the participating patients a voice to express how they would change the spatial situation they are in. Sometimes the same picture can be placed in different categories because of the multi-layered explanation that comes with it. Pictures of the windows or the ceiling are fairly straightforward in what they indicate. Fresh air and a view of the outside world are generally considered important aspects of a healing environment (Rubin et al., 1998; Devlin and Arneill, 2003; Dijkstra et al., 2006; Cbz, 2008). It is thus no surprise that one who wants to discuss these topics pictures windows. When lying in a hospital bed your perspective changes. Lying or sitting in bed has a great impact on what, or whether, you see through the window.

Figure 1.1. Difference between what you see through the window when lying down versus sitting in bed

8

Annemans et al.

The iconographic meaning can only be discovered when both the image and the narrative are combined. Indeed, the same object can be pictured for a completely different reason. One of the pictures made by Mrs A shows a window as well. However, this window does not represent fresh air or daylight. The window in the picture is located in an inner wall around the atrium accommodating the cafeteria. Through the window the sound of chattering people and laughing children enters the room. Although Mrs A describes this as a pleasant sound when in a different setting, she finds it quite disturbing while being in the hospital and trying to recover from surgery. Fairly banal objects can be icons of less tangible aspects that have a strong impact on the experience of the hospital. Mrs A photographs the little table and the chairs in her room while having visitors. Because of the table’s presence, the people in the chairs do not seem to be visiting in a hospital so much as just being at home. As she puts it: ‘I think it has a more homelike atmosphere when people can be sitting on a chair, at a table, where you can put something on, than when you would just have a row of chairs, then it would feel like they were watching me.’ (Mrs A, interview transcript)

Figure 1.2. Iconographic images: left: Window photographed to illustrate the sound that enters through it; right: The presence of a table changes the perception of visitors on the chairs.

Mrs B made a set of two photos and a drawing, picturing the transportation of a(n unknown) patient in bed. In this trilogy she combines a reflection on her own experiences while transported through the hospital in a bed and the observations from the bed.

Hospital Reality from a Lying Perspective

9

Figure 1.3. Nurse connecting a patient’s bed to a wagon, a view of the ceiling while travelling through the hospital, setup of how the bed is transported through the hospital

The first picture shows the hallway in the basement of the hospital. Mrs B mainly wants to point out how she felt when she was transported through it during an earlier stay. It is a very functional hallway, used as a storage space for carts, bicycles, obsolete beds and so on. It seems as if no one ever thought about the fact that patients who have to go to the nuclear department are transported through it. As Mrs B explains, “the lighting is not pleasant, and all the stuff that is stored there makes you feel uncomfortable”. The nurse in the picture is attaching a bed to a cart, used to pull the beds when large distances have to be covered. Mrs B reflects on how the patient in the bed must feel, being handled like that. Since she did not feel comfortable taking a picture while the patient could see her, she drew the setup, explaining how the bed is attached to the cart, how the patient does not have any contact with the nurse involved, and what s/he must perceive and feel while being towed like that. To illustrate that, Mrs B also took a picture of the ceiling a little further down the hall and concluded that seeing all those pipes, and the rags in between, is not how a disconcerted patient is comforted. The dust between the technical equipment on the ceiling makes her wonder about hospital hygiene. Although these pictures might seem a little banal at first, how they come together and are used as a backbone for the narrative provides new opportunities to talk about spatial experience. We do not just learn about Mrs B’s experience while visiting the hospital this time, but are also provided with a reflection on previous visits and when and why you feel most vulnerable as a patient, which she uses to explain her reluctance to take a picture. Even movement and time are touched upon during the conversation. The length of the travel through the hall, how the patient must undergo the transportation not knowing where he/she is going, it all adds to

10

Annemans et al.

the experience of the building without being necessarily related to visual perception. In response to the simple question we asked the patients to also express their view on how the building could be improved. For Mr C drawing a plan of the ward gives him the opportunity to point out the shortcomings of the current setup, while by sketching how he would organise it, he brings up spatial aspects that he thinks are important. More acoustic barriers, more visual transparency, and a reduction of the passage to create a quiet atmosphere are only some of the aspects mentioned.

Figure 1.4. Left: existing situation, right: how Mr C would organise the ward

Suggesting alterations to the hospital layout is not restricted to drawing plans. Mrs A documents how she improved her privacy by opening a second door to block the view from the hallway. When both her door and the one at the other side of the hall were fully opened, she could see right into the other room. Since she did not want to close her door completely, preferring some connection with the rest of the ward, she opened the door of the bathroom to block the view. The conversation about these pictures first focuses on the privacy but soon broadens to interactions with other patients and staff, feelings of safety and even how bed transport could be improved so there would be fewer obstacles on the way.

Figure 1.5. Left: open doors, giving a view into the room at the other side of the hall, right: opening the bathroom door can create a second barrier without closing the door of the room

Hospital Reality from a Lying Perspective

11

1.5 Conclusions and Future Work Developing a better understanding of how patients experience the complex reality of a hospital building from a lying perspective, asks for a sensitive methodological approach that addresses all senses and facilitates reflection on experience. Although we do not wish to push forward a single method, the technique explored so far seems to fulfil the requirements. By asking patients to document their spatial experience and providing them with multiple media to do so, we gain insights from the inside out, not only from the material they collect, but also from why and how they collect it. Using the illustrations generated by the participants themselves clearly adds an extra dimension to the interviews. Both the interviews and the findings are deepened. Even though the material collected is mostly visual, it can also illustrate auditory or haptic qualities, like the window through which a lot of noise enters. The images, complemented by the accompanying narratives, serve as a source of information about all senses, movement and time, illustrating the intangibility of sensory experience. Overall, this alternative way of questioning people definitely provides additional value given the aim of our research. Given the promising results of this exploratory phase, the approach will be further elaborated in future phases into a more generally applicable methodology that can, at least partially, bypass the difficulty of articulating spatial experience. The motivation of the participating patients is of crucial importance to the success of the approach. As such, finding patients willing to participate is a key concern. Although part of the strength of the approach lies in its flexibility, a balance should be found between freedom and guidance. In future research, when the key concern is the outcome of the interviews and not the methodology followed, the selection of the settings and participants will need to be more representative to obtain more valid and transferable results. When thinking about the translation of the research results to inform architects’ design process, we might even consider using the empirical material in its original form. This would imply that real data, directly from the patients, are used by those designing environments for them. Especially in health care buildings, the highly complex reality forms an important step for designers to acknowledge the needs of the primary users. Since the evidence provided by EBD studies conducted so far is not ready to be converted to real-life applications yet, it is definitely worth investigating whether different empirical material can be used to inform the design process.

1.6 Acknowledgements Margo Annemans’ research is funded by a PhD grant from the Baekeland program from the Institute for the Promotion of Innovation through Science and Technology in Flanders (IWT-Vlaanderen). The Baekeland program gives researchers the opportunity to complete a PhD in close collaboration with the industry, in this case Osar Architects nv provided this opportunity. Ann Heylighen received support form the European Research Council under the European Community’s Seventh

12

Annemans et al.

Framework Programme (FP7/2007-2013)/ERC grant agreement No. 201673. The authors thank the participating patients for sharing their time and insights.

1.7 References Annemans M, Van Audenhove C, Vermolen H, Heylighen A (2011) Lying architecture: Experiencing space from a hospital bed. In: Proceedings of the 1st International Conference Exploring the Multi-dimensions of Well-being (Well-being 2011), Birmingham, UK Cbz (2008) Kwaliteit van de fysieke zorgomgeving, stand van zaken onderzoek omgevingsvariabelen en de effecten op de (zieke) mens, College bouw zorginstellingen, Utrecht, The Netherlands Collier J (1967) Visual anthropology: Photography as a research method, Holt Rinehart and Winston, NY, US Devlin AS, Arneill AB (2003) Health care environments and patient outcomes: A review of the literature. Environment and Behavior, 35(5): 665-694 Dijkstra K, Pieterse M, Pruyn A (2006) Physical environmental stimuli that turn healthcare facilities into healing environments through psychologically mediated effects: Systematic review. Journal of Advanced Nursing, 56(2): 166-181 Franck K, von Sommaruga Howard T (2010) Design through dialogue: A guide for clients and architects, Wiley, Chichester, UK Harper D (2002) Talking about pictures: A case for photo elicitation. Visual Studies, 17(1): 13-26 Herssens J, Heylighen A (2009) A lens into the haptic world. In: Proceedings of the International Conference on Inclusive Design and Communications (INCLUDE 2009), London, UK Ingold T (2000) The perception of the environment: essays on livelihood, dwelling and skill, Routledge, London, UK Mason J, Davies K (2009) Coming to our senses? A critical approach to sensory methodology. Qualitative Research, 9(5): 587-603 Pallasmaa J (2005) The eyes of the skin: Architecture and the senses, John Wiley & Sons, NY, US Pink S (2007) Doing visual ethnography: Images, media, and representation in research, 2nd edn. Sage Publications, London, UK Radley A (2010) What people do with pictures. Visual Studies, 25(3): 268-279 Radley A, Taylor D (2003) Images of recovery: A photo-elicitation study on the hospital ward. Qualitative Health Research, 13(1): 77-99 Rasmussen SE (1964) Experiencing architecture, MIT Press, Cambridge, MA, US Rubin HR, Owens AJ, Golden G (1998) An investigation to determine whether the built environment affects patients’ medical outcomes, Center for Health Design, Martinez, CA, US Throop CJ (2003) Articulating experience. Anthropological Theory, 3(2): 219-241 Ulrich RS, Zimring C, Joseph A, Quan X, Choudhary R (2004) The role of the physical environment in the hospital of the 21st century: A once-in-a-lifetime opportunity, Centre for Health Design, Concord, CA, US Van den Berg AE (2005) Health impacts of healing environments: A review of evidence for benefits of nature, daylight, fresh air, and quiet in healthcare settings. University Hospital Groningen, Groningen, The Netherlands Warren S (2002) ‘Show me how it feels to work here’: Using photography to research organisational aesthetics. Ephemera Critical Dialogues on Organisations, 2(3): 224-245

Chapter 2 Inclusive Bus Travel - A Psychosocial Approach F. Nickpour, P.W. Jordan and H. Dong

2.1 Introduction 2.1.1 Public Transport - The New Climate Public transport is facing major challenges in the current economic and social climate; a considerable rise in demand for public transport and an ageing population that is mainly dependant on public transport and is increasingly in need of specialised and door-to-door services. The above challenges double when one considers the raised public awareness and the pressure from user organisations to improve the equality and quality of public transport for all. Public transport providers need to respond to increasing demand for service provision, both in terms of volume and diversity of service users. Transport for London (TfL), a major public transport provider in UK, is currently facing oversubscribed door-to-door services and an increasing demand for accessible and usable public transport by conventionally marginalised groups such as older people and people with disabilities. Issues of accessibility, reliability and quality of service are key indicators that are sometimes in conflict and need to be revisited. There is a need to keep the quality of service consistent and at the same time redefine and prioritise the areas of focus and improvement.

2.1.2 Public Bus Services Buses will continue to be - probably for many years - the main and only form of public transport that can be accessible to almost all (London TravelWatch, 2010). There is also evidence that bus services are often more frequently used by disadvantaged or vulnerable sections of society, therefore poor performance is more likely to impact on these groups (London TravelWatch, 2009). Thus, the bus

P. Langdon et al. (eds.), Designing Inclusive Systems, DOI: 10.1007/978-1-4471-2867-0_2, © Springer-Verlag London 2012

13

14

Nickpour et al.

service proves to be the single most powerful transport tool in terms of inclusivity and equality potential and provision in a mega-city like London. There have been great improvements in terms of making buses fully accessible. In London, all buses are now low-floor vehicles and have a space for one wheelchair (Transport for London, 2011). However, an ‘accessible bus’ does not necessarily guarantee an ‘accessible bus service’. An accessible bus service requires not only an accessible bus and an accessible bus stop but also an empathic well-trained driver and a user-friendly environment. As well as improving inclusivity, making local bus services more accessible brings wider benefits including facilitating social inclusion in the local community, making bus travel easier and more pleasurable for every member of the local community and reducing the need for dedicated services (e.g. Dial a Ride) which are not costeffective.

2.1.3 The Project Commissioned by Transport for London and one local London borough, a research project was conducted in order to address issues associated with bus travel in London. The aim of the project was to produce recommendations for improving the accessibility of bus travel through investigating barriers to a diverse range of people using (or not using) public buses and what makes a journey either pleasant or unpleasant. A variety of approaches and techniques were used in order to understand the barriers to accessibility and inclusivity and how these could be overcome. The research project aimed to assess and improve the accessibility of public buses through a holistic and comprehensive service-oriented approach, focusing on an accessible bus service as a whole rather than focusing on segments of the whole service such as bus or bus-stop. 2.1.3.1 Bus Service - Key Stakeholders Broadly, with respect to bus services, three major stakeholders were defined: • • •

Service user - mobility challenged people Service provider - bus drivers Service operator - bus companies

Addressing accessibility and inclusivity issues, the project focused on mobility challenged people as the critical bus service users. For the purpose of this project, a mobility-challenged person was defined: ‘A mobility challenged person is someone whose mobility has been challenged due to age, physical or mental impairment, or an external physical condition; each of the above could have substantial and long-term adverse effect on the person’s ability to use public transport.’ (Nickpour and Jordan, 2011)

This definition includes, but is not limited to, wheelchair users and those with other impairments that affect mobility. Other major groups with other mobility

Inclusive Bus Travel - A Psychosocial Approach

15

restrictions that may make it more difficult to use public transport are: older people, blind or visually-impaired people, deaf people or people with hearing difficulties, those with learning difficulties or social phobias, and guardians with buggies. 2.1.3.2 Bus Service - Stakeholder Issues Key issues concerning each stakeholder included: • •

•

For bus passengers: Positive experience from start to finish - every stage of the journey should be efficient, enjoyable and smooth, and the user should be and feel safe at all times. For bus drivers: Pleasant working environment - drivers should be treated politely and respectfully by all passengers. They should be equipped with the skills needed to carry out all aspects of their duties competently and receive the full support of both bus users and their employers in doing so. For bus operators: Profitable business - operators should be encouraged and enabled to fulfil the service requirements against suitable performance targets in a manner which is commercially viable.

2.2 Methodology and Methods 2.2.1 Methodology The research project followed a combined primary and secondary research methodology, with a heavy focus on primary research conducted through a diverse range of field research methods. A major focus for the project was consultation with people who had a wide range of mobility challenges. Many other stakeholders were also included in the consultation process. This included bus drivers and representatives from bus operating companies, TfL, police and advocacy groups representing mobility-challenged people. In addition to this consultation process, members of the project team gained first-hand experience of some of the issues faced by mobility-challenged people by taking bus trips while using wheelchairs. Information was also collected through observing mobility-challenged people travelling on buses and asking mobilitychallenged residents of London Borough of Hillingdon - where the study was conducted - to take bus journeys and report their experiences.

2.2.2 Methods A wide range of methods were used in order to collect first-hand information regarding the existing barriers and issues regarding accessibility and inclusiveness of bus services. All primary research was undertaken in the local London borough. In some cases, similar services were observed in other London boroughs as well. Due to space limitation, specific details in terms of participants’ process of

16

Nickpour et al.

selection, age, demographics, position, etc. are not included here. More detail on the above is provided in a technical report (Nickpour and Jordan, 2011). 2.2.2.1 Focus Groups Three focus group sessions with different focuses were run in order to provide a holistic understanding of the existing issues. Each session focused on one stakeholder group. Firstly, a focus group session was held with nine representatives of service providers and a cross-section of other stakeholders aiming to look at organisational and big-picture issues. The participants included representatives from TfL, the local Council, bus companies, Dial a Ride, Age UK, Metropolitan Police, Hillingdon Community Transport and Access and Mobility Forum. Then, one session was held with a diverse group of service users with a focus on mobility-challenged passengers. This included nine participants; one blind person, one person with learning difficulties, one wheelchair user and six older people. Finally, a session was held with service non-users including seven mobilitychallenged members of the public who did not use currently public buses for a variety of reasons. These included previous negative experience with using public buses and lack of trust and confidence in the service. 2.2.2.2 Access Audits Two sets of access audits were planned and carried out. The emphasis was on both immersion (Moore and Conn, 1985) and direct observation (Dray, 1997). The first series of audits included eight local bus journeys and were carried out by the project research team, role-playing by using a wheelchair, aiming to look at specific mobility issues. Each observation session was attended by two members of the research team. The second series of access audits were carried out by a diverse group including five local participants with mobility impairments. Participants included one male older person aged 72, and two wheelchair users, one with an electric wheelchair and one with a normal wheelchair. Also, one person with learning difficulty aged 21 and one blind person aged 42 carried out the access audits. All audit sessions were documented through various applicable audio, visual and textual formats. 2.2.2.3 Interviews and Meetings A number of meetings and interviews were held with individuals from various organisations and groups in order to look into a number of issues in more detail. Altogether, five interview sessions were held; these included interviews with three bus drivers, meetings with Hillingdon Community Transport general manager, the accessibility officer of Hillingdon Council, two officers from the Disablement Association of Hillingdon and six members of the local Youth Council. 2.2.2.4 Observations Two major observation sessions were held. One session focused on special services aimed at mobility-challenged passengers; the project team spent a day working with the Dial-a-Ride service that provided door to door transport for mobilitychallenged people. Another observation session took place at Bus Mentoring Day -

Inclusive Bus Travel - A Psychosocial Approach

17

a training day aimed at helping those who assist mobility challenged people with their travels. 2.2.2.5 Literature Review The literature review drew on a number of sources, reports and documents including reports by the Disabled Persons Transport Advisory Committee (DPTAC), Direct Gov, The Department of Transport and London TravelWatch. The main source for the literature review was the new report by the Greater London Authority (GLA), titled “Accessibility of Transport” (GLA, 2010) which looked at the accessibility of all public transport within the capital including buses. The report drew on inputs from a wide variety of advocacy groups representing mobility challenged people as well as on a wide array of statistics quantifying accessibility of buses and other modes of transport.

2.3 Findings Based on the access audits conducted, the journey was broken down into the stages shown in Figure 2.1.

Figure 2.1. Key stages of a bus journey

The findings are presented under three key categories; physical, psychosocial and operational issues. Due to the length available for this paper, only a summary of findings is presented here. Detailed breakdown and analysis of findings can be found in the ‘Inclusive Bus Travel in Hillingdon: Assessing Accessibility’ report (Nickpour and Jordan, 2011).

2.3.1 Physical Issues From a physical accessibility point of view, users tended to find the most problematic part of the journey was getting from home to the bus stop and getting from the bus to their final destination. Examples of problems here included: narrow pavements, loose paving stones, steep roads and difficult crossings. There were also accessibility difficulties at some bus stops - for example, the positioning of litter bins and other street furniture sometimes made deploying and using the ramp somewhat inconvenient.

18

Nickpour et al.

However, despite such difficulties, it was possible for mobility challenged people to board the bus at all of the stops examined in the audit. Improvements in the design of buses meant that, in general, once the user had reached the stop, the bus could be accessed OK and the on-board part of the journey completed.

2.3.2 Psychosocial Issues Various observational and immersive methods used also uncovered a number of other difficulties - mostly psychological and social - that users faced. These included: 2.3.2.1 Uncertainties There were many aspects to this including uncertainties as to whether users would be able to get on and off the bus OK, whether they would have a long wait at the stop and whether their interactions with others would be positive. 2.3.2.2 Overcrowding The start and end of the school day are times when the bus gets particularly crowded. This can sometimes mean that the bus is too crowded to let a wheelchair on. Even if it is possible to board, overcrowding can make it difficult for wheelchair users to get to the wheelchair bay and to move their chair into the proper position within it. Overcrowding is becoming an increasingly problematic issue as more and more people are using buses. This is due in part to the difficult economic conditions that we have had recently (bus travel tends to increase in times of financial hardship) and in part to the issuing of free bus passes to schoolchildren and older people. 2.3.2.3 Negative Experiences with Drivers Many users had also mentioned that they had had problems with the drivers. This could be because of inconsiderate driving - for example pulling away too quickly or because they were perceived as having an unfriendly or surly attitude towards the user. Indeed, during the access audits there were a number of incidents of drivers not stopping at bus stops when they saw a wheelchair user waiting to get on. Bus drivers mentioned that there were often problems with ramps failing to deploy and cited this as a reason why they could not always pick up wheelchair users. 2.3.2.4 Negative Behaviour of Other Passengers A number of participants reported being annoyed or intimidated by the behaviour of other passengers. In particular they mentioned teenagers who they said could be very loud and often used foul language. A number of participants also mentioned that they also found it annoying when people had loud conversations on mobile phones or played music so loudly that it could be heard through their headphones. The behaviour of other passengers when getting on and off the bus was also a source of annoyance and intimidation. In particular they mentioned pushing and

Inclusive Bus Travel - A Psychosocial Approach

19

shoving and people not waiting their turn in the queue. Other users had reported that they are wary of using buses in the evening or night because of the risk of encountering drunk or threatening people. 2.3.2.5 Off-putting Stories In some cases, participants were put off using the bus because of stories they heard about other people having bad experiences, in particular stories of violent or frightening incidents. These stories may have been told to them by friends or they may have read or heard about them in the media.

2.3.3 Operational Issues An issue that may be a contributory factor is the key performance indicators (KPIs) used to measure the performance of the bus operators. Currently, emphasis is mostly on reliability - that has to do with timeliness of the bus service. There are no measures in place to monitor either the number of mobility challenged people using buses or the quality of their experience as one performance indicator. It was observed that it can take some time for a mobility challenged person, such as a wheelchair user, to board the bus. This may lead to the bus running behind schedule with the consequence that it affects reliability. As reliability is the basis on which the bus companies are judged and the pressure is for them to run on time, drivers sometimes feel unenthusiastic about picking up mobility challenged passengers and hence may have a hostile attitude towards mobility challenged them or may try to avoid picking them up altogether.

2.4 Discussion 2.4.1 Physical Versus Psychosocial Issues Overall the research suggested that good progress had been made in terms of addressing the physical issues. There could be problems getting to and from the bus stop and sometimes there were problems with ramps and small wheelchair spaces. However, it was generally the case that it was physically possible to complete a journey without excessive difficulties. Perhaps the most striking issue to emerge from the research was the role that psycho-social factors played in affecting mobility-challenged people’s quality of experience of using public buses, in particular, the impact of the attitudes and behaviour of the driver and of other passengers. Bad experiences of this nature were the most frequently cited reasons for not enjoying a bus journey or for not using the bus at all. Previously, the emphasis of accessibility research and improvements schemes has been on the physical elements of accessibility. While these are certainly extremely important, the outcomes of our research suggest that psychosocial issues are equally, perhaps

20

Nickpour et al.

even more, so. This observation mirrors those within the field of design generally where there has been increasing attention in recent years on psychosocial issues and their emotional consequences (Norman, 2005).

2.4.2 Special Service Versus Public Service As part of this research we also looked at people’s experiences with door to door transportation schemes for mobility challenged people within London. These included Dial-a-Ride, a minibus-based service which picks up passengers at their home and takes them to a pre-requested destination. This service was very popular with users. In particular they enjoyed the friendly atmosphere on the minibus and the friendly, attentive and considerate behaviour of the driver. Mobility-challenged users praised the drivers for their empathy and understanding, for their cheerfulness and for making them feel valued and welcome whenever they used the service. They mentioned how much they looked forward to the social aspects of using the service and for the enjoyable conversations with other passengers. A challenge is to try and recreate some of these benefits on public buses and to put into place approaches and schemes that will help to foster a positive ambience.

2.4.3 Negative Interactions It should be emphasised that the picture is not entirely negative; Field research supported the fact that many of the drivers have an excellent approach to interacting with mobility-challenged people. They are friendly, welcoming, informative and help make the journey a great experience. Similarly, many teenagers are polite, well-behaved and kind towards other passengers. However, this was mainly the result of each individual’s intrinsic motivation and personal codes of conduct. Nevertheless, it is also important to recognise that there are genuine problems with some bus drivers’ and teenagers’ attitudes and behaviours. Negative drivers’ attitudes were observed and reported, such as being rude and uncommunicative towards mobility challenged people. Also, in some cases, some teenagers’ behaviour appeared inconsiderate and liable to make people feel uncomfortable. The effects of this negative behaviour tend to extend beyond the specific incidents that occur. When service users encounter a bad experience, they will remember this and will have a doubt in their minds about the quality of their experience next time. This uncertainty can have a very powerful and negative effect. Even if people subsequently have positive experiences, the memory of the previous bad experience can create a sense of doubt - will this happen again? This doubt can make people question whether they want to use the bus again and leave them with some negative feeling for the duration of their travel. Moving forward, the

Inclusive Bus Travel - A Psychosocial Approach

21

challenge is to find effective ways of improving the ambience on board and tackling some of the psychosocial issues that have been identified.

2.5 Conclusions and Recommendations There is need for a ‘Mentality Shift’ when addressing accessibility in public transport. This study suggests and highlights ‘psycho-social’ inclusion as the key area of focus. The findings suggest accessibility and inclusivity issues affecting public bus services fall into three broad categories: Physical, Psycho-social and Operational. Physical issues are to do with the design of the bus and the built environment and are the ‘typical’ issues considered when looking at accessibility. Findings suggest the key physical barriers identified include Getting to bus-stop, Space availability and priority on bus and Ramp technology & reliability. Psycho-Social issues are the ‘soft’ issues associated with the quality of people’s travel experience. Findings suggest the key psycho-social barriers identified are Ambience, Awareness and empathy and Communication. Operational issues concern the running of the service and cross-organisational strategies and regulations. The key identified operational barriers are Key Performance Indicators. Public bus service KPIs currently appear to focus only on efficiency rather than quality, inclusivity and pleasurability of service. The results indicate that it is the psycho-social issues that seem to be proving the biggest barrier to using public buses, in particular for mobility-challenged people. Addressing these issues requires a focus on people. It involves making them aware of the effect that their behaviour is having, convincing them to change it and giving them the skills and insights needed to do so. It also involves creating a desirable ambience throughout the bus journey, making the public transport experience not only efficient but also pleasurable. Overall - including both physical and psychosocial factors - the following nine recommendations are proposed as key principles for improving mobility challenged passengers’ experience of public bus travel. Create an inviting and friendly experience of the bus service. Perceptions about bus travel influence people’s decisions about whether to take the bus and the emotions associated with anticipating using it. Mobility challenged people should be confident that their bus journey will be a positive experience. Make bus stops reachable. Getting to and from the bus stop is, generally, the biggest physical barrier to bus travel for mobility challenged people. Making bus stops more reachable would significantly increase the numbers of people who could access public buses. Make all bus stops fully accessible. Once at the stop, mobility challenged people should be accurately informed about when the bus will arrive. The design of the stop should also facilitate quick and easy ingress for them. Promote and facilitate positive behaviour amongst passengers. Interactions with other passengers should be positive and friendly throughout the bus journey.

22

Nickpour et al.

Ensure that key aspects of the bus are fully operational. The aspects of the bus that affect accessibility should be fully operational at all times. Mobility challenged people should be confident that their journey will run smoothly and efficiently. Ensure that all users have a safe and comfortable space. All mobility challenged users should have a safe and comfortable space in which to complete their journey. They should be able to move into and out of this space easily. Welcome mobility challenged people aboard. Drivers should warmly welcome mobility challenged people aboard the bus. They should communicate clearly and cheerfully with them throughout the journey. Set off and drive smoothly. Ensure that mobility challenged people are settled before moving off. Make sure that this is done smoothly and that the drive is smooth and controlled throughout the journey. Provide information clearly through multiple channels throughout the journey. Mobility challenged people should be clear about when the bus is approaching their stop and have plenty of time to prepare to exit.

2.6 Acknowledgements This research project was commissioned by London Borough of Hillingdon and Transport for London. The authors would like to thank all local participants in the project and the user research team including Murtaza Abidi, Penelope Bamford, Thomas Wade and Jennifer McCormack.

2.7 References Dray SM (1997) Structured observation: Practical methods for understanding users and their work in context. In: Proceedings of CHI 97 Workshop on Human Factors in Computer Systems, SIGCHI, Atlanta, GA, US GLA (2010) Accessibility of the transport network. Greater London Authority, London, UK London TravelWatch (2009) TfL performance report. London TravelWatch, London, UK London TravelWatch (2010) Bus passengers’ priorities for improvement in London. London TravelWatch, London, UK Moore P, Conn CP (1985) Disguised: A true story. Word Books, Waco, TX, US Nickpour F, Jordan PW (2011) Inclusive bus travel in Hillingdon: Assessing accessibility. Technical Report, Brunel University, Brunel, UK Norman DA (2004) Emotional design: Why we love (or hate) everyday things. Basic Books, Basic Books, NY, US Transport for London (2011) TfL accessibility guide. Available at: http://www.tfl.gov.uk/gettingaround/transportaccessibility/1171.aspx (Accessed 13 August 2011)

Chapter 3 Safe and Inclusive Design of Equipment Used in the Minerals Industry T. Horberry and T. Cooke

3.1 Introduction The focus of this paper is upon the application of both safe and inclusive design to equipment used by operational and maintenance personnel in mining. It begins by introducing the minerals industry and outlining two important human-related issues that will greatly impact upon the design of future mining equipment. The paper then focuses on the importance of safe and inclusive design in this domain, and outlines a task-orientated risk assessment and design process called ‘OMAT’ that was developed by the authors. A series of semi-structured interviews with designers of mining equipment are then presented: they focus in particular on how designers currently obtain user-centred input and how the OMAT process might be integrated into their design practices. Finally, conclusions regarding the future safe and inclusive design of equipment (especially automated equipment) in this domain and why user-centred design processes should be of central importance to mining are presented.

3.1.1 An Overview of Mining and the Minerals Industry The minerals industry is a significant worldwide employer. For example, in Australia it employs approximately 136,000 personnel. Mining occurs across virtually the whole globe, with major areas in South Africa, North and South America, Australia, China and much of Europe. The worldwide injury, ill-health and fatality rates vary greatly, ranging from usually single figure deaths per annum in Australia through to many hundreds being killed in many third-world countries (Simpson et al., 2009). Whatever the precise figures, the minerals industry is a major global employer with many high hazards that can cause injuries and fatalities unless well managed (Komljenovic and Kecojevic, 2007).

P. Langdon et al. (eds.), Designing Inclusive Systems, DOI: 10.1007/978-1-4471-2867-0_3, © Springer-Verlag London 2012

23

24

Horberry and Cooke

Looking in more depth at the elements of the minerals industry system, there is no ‘typical’ arrangement that is used at the majority of mine sites (Sanders and Peay, 1988). Instead, it is a complex sociotechnical system where people, procedures, environments and equipment need to interact safely and efficiently. The main elements here include (Horberry et al., 2010): • • • • • • • •

an increasingly diverse group of people employed; a wide assortment of different jobs, tasks and roles; many different equipment manufacturers and suppliers; different worldwide mining companies; a wide array of national laws, regulations, and guidelines; different procedures, rules, practices and cultures at individual mine sites; differences in the built environment and precise mining method used; uncertainties in the natural environment.

3.1.2 Mining Equipment Focusing on the equipment element, there is a discernable trend in most advanced economies for mining equipment parameters to be improved in that the equipment needs to be bigger, stronger, quicker, safer and more reliable (Horberry et al., 2010). The images in Figure 3.1 show the size and complexity of two examples of mobile mining equipment.

Figure 3.1. Examples of mobile mining equipment

In tandem with this, there are continual equipment-related operational challenges, including the ever-present balance between safety and production. It is not within the scope of this paper to describe the range of equipment employed, but it certainly is within the scope to briefly describe two general human element challenges that impinge upon designing and deploying safe, inclusive and fit-forpurpose equipment. These are: the ageing workforce and the increased development and deployment of automation/new technologies in mining.

Safe and Inclusive Design of Equipment Used in the Minerals Industry

25

3.1.2.1 An Ageing Workforce Over twenty years ago it was noted that the average age of the mining workforce was getting younger (Sanders and Peay, 1988). In most industrialised countries this situation has now reversed: the minerals industry workforce is on average now getting older, and fatter. Of course, this mirrors many general trends in Western society; however, it does present a few specific issues for the design of mining equipment, especially to make such equipment inclusive and safe. Given the nature of the equipment in use then extreme ageing is not an issue in mining, but there are still many issues for safe and inclusive design that result from an older workforce. These include (adapted from Horberry et al., 2010): •

• •

•

Increased difficulty in learning new skills. Older workers do not automatise tasks as easily as their younger counterparts. This has implications for mining technology and automation use (e.g. remote control of a rock breaker) where skill requirements may change over time and require new automatic, over-learnt operating procedures. Reaction time increases, especially in reacting to unexpected stimuli. This is a particular concern for mobile mining equipment (e.g. haul trucks) operating in the complex transport environment in most mine sites. Loss of muscular strength, endurance and tone. Although many manual tasks in mining have been eliminated, they are still present in many equipment maintenance tasks. This problem is exacerbated if significant weight gain occurs in these older maintenance workers. Visual function changes including loss of precision, difficulty in focusing on near objects and declines in visual acuity and contrast sensitivity. Given the automated, round-the-clock nature of mining then this can present particular problems in bad lighting conditions or with a complex background.

3.1.2.2 Increasing Development and Deployment of New Mining Technologies and Automation Mining is already highly mechanised, but it seems certain that there will be more remote control and/or automation of mining equipment in the future. Indeed, automation in mining is now finally permeating into many cutting-edge sites. Although automation progress began initially during the 1960s, today many mining companies are investing heavily in automation initiatives (Lynas and Horberry, 2011). As in other industries, this may change the types of human element inputs required (Sheridan, 2002). For example, less manual operational tasks (at least when the equipment is working correctly) and many operators becoming more of a passive monitor of the system, rather than an active controller or driver of it. More broadly, the increasing uptake of mining automation presents many opportunities, but also many challenges related to safe and inclusive design and effective human systems integration (Lynas and Horberry, 2011). The role of the human in the overall mining system may change, but is still a central part rather than an optional extra. Thus, developing operator-centred approaches for the design and

26

Horberry and Cooke

integration of new/automated mining technologies is a key priority area for the technology to be successful: central to this is safe and inclusive design.

3.1.3 The Importance of Safe and Inclusive Design of Mining Equipment The design of mining equipment plays a critical part in the safety and efficiency of tasks that are conducted by operators using it. However, as much as in virtually any other occupational domain, the design of mining equipment is still heavily technology-centred rather than user-centred (Horberry et al., 2010). This has been for a variety of reasons in mining, including: •

•

•

Designers sometimes think they can use their knowledge/common sense/intuition, or they rely purely on designing to standards. This is particular acute in the minerals industry, where mine site access for designers can be a significant limitation (EMESRT, 2011). Adding a user-centred focus may be thought to alter an agreed design process. Similarly, an older version of a system may already be in place, and piecemeal alterations are subsequently made (for example, the design of haul trucks). The benefits and costs of using a user-centred approach for equipment design are not clear (Burgess-Limerick et al., 2007).

Because some mine equipment designers (and their customers) still view human element constraints to be less significant than technical challenges (such as equipment payload) there is a tendency not to adequately consider human factors in the equipment design life-cycle process, and it is common to see human factors concerns being passed from one phase to the next (Horberry et al., 2010). For example, during conceptual design, if the analysis does not adequately capture user requirements then subsequent inadequacies are hard to resolve in the detailed design phase. Human element problems (such as usability) that remain after the system has been designed cannot therefore be easily resolved during equipment build or implementation phases. This is a particular problem for mining equipment that still requires considerable human intervention in its maintenance and operation (e.g. mobile mining equipment maintenance such as changing air filters). Indeed, where such user-centred design does exist, it focuses more on safe rather than inclusive design. Safe design (sometimes known as safety in design, or prevention through design) has begun to receive an increasing amount of attention in both the scientific and occupational safety domains (including in mining, albeit slowly) and is generally applied to products and equipment. As the name implies, it involves occupational health and safety (OHS) through the original design, not safety by procedure or through retrofit trial and error.

Safe and Inclusive Design of Equipment Used in the Minerals Industry

27

“Safe Design is a design process that eliminates OHS hazards, or minimises potential OHS risk, by involving decision makers and considering the life cycle of the designedproduct.” (Safe Work Australia, 2011)

In contrast, inclusive design is still in its infancy in mining. As an occupational, high-hazard domain, it might be argued that a ‘conventional’ inclusive design process is in need of modification as the focus here is not on mainstream commercial products. However, the earlier-described changing workforce demographics (e.g. older operators) and the design possibilities raised by the uptake of automation should increase the focus on designing for an increasingly diverse user group that will engage in a wide range of new tasks.

3.1.4 The Operability and Maintainability Analysis Technique (OMAT) To help ensure safe and inclusive design, the authors of this paper recently developed and evaluated a process that could be used by mining Original Equipment Manufacturers (OEMs) (Cooke and Horberry, 2011). More specifically, OMAT is a task-oriented risk assessment process that focuses on human factors risks related to mobile mining equipment design. The starting point of the work was that poor equipment operability or maintainability can produce major safety and performance disbenefits. As such, many mining incidents and accidents are due to equipment design inadequacies, either in maintainability or operability, and are therefore theoretically preventable (Horberry et al., 2009). The behaviour of the operators and maintainers is largely shaped by their tasks which are, in turn, partly shaped by the equipment’s design. Therefore, in order to create the safest equipment possible designers must predict how their designs will shape the behaviours of miners in different sites around the world. With this in mind, a number of mining companies started discussing an approach to improve the human factors design of mobile mining equipment at an OEM level. This eventually resulted in a multi-company industry initiative known as EMESRT; the purpose was to establish a process of engagement between OEMs and mining customers to promote the development and adoption of leading practice user-centred designs (EMESRT, 2011). The OMAT tool was developed by support from EMESRT (Horberry et al., 2009). The OMAT process has been developed to align with existing OEM design processes (including major design milestones). OMAT has also been developed specifically to provide a user-engagement processes to identify and assess the risks in the design of mining equipment. Investigating such risks in operational or maintenance tasks involves six OMAT steps (after an initial step zero, to define the scope of the work) that are heavily dependent on OEM and mine site user engagement. A summary of these six stages is presented below.

28

Horberry and Cooke

1. Based on a comprehensive list of all operations and maintenance tasks performed using the equipment, the critical tasks are prioritised. 2. The constituent steps in these priority tasks are described and analysed. In effect this is a task analysis, whereby each task step and its order is uncovered, including deviations, short cuts or different methods of task completion for different user-groups. 3. Risks at each of the task steps are identified. Using the types of matrices commonly used in the mining industry the risks are identified, noting any current controls employed (e.g. guard rails for working at height). 4. Solutions are developed for the risks identified in stage 3. These should be primarily design solutions that eliminate risk. 5. Feedback received. The solutions developed in stage 4 are further developed by the OEMs. However, to continue the process of user-centred design, these solutions are then evaluated by mine site users. 6. A risk register is maintained to keep track of the whole process. This is of particular value for future design iterations of the equipment. The development and largely positive evaluation of the OMAT process to help promote safe and inclusive design of existing mobile mining equipment has resulted in important design changes being made (see Cooke and Horberry, 2010). But a comparatively neglected area to date has been an assessment of designers’ opinions of the process, and how it can be integrated with their existing safe and inclusive design processes.

3.2 Interviews with Mobile Mining Equipment Designers 3.2.1 Aim The aim of the study was to gauge the opinions and work practices of mobile equipment designers regarding user-centred design processes. In particular, it compared their current design methods to the OMAT task-based model and explored the designers’ current methods of gaining end-user feedback (i.e. from mine site operational and maintenance personnel).

3.2.2 Method A series of semi-structured interviews were held with personnel employed by a major manufacturer of mining equipment. Fourteen people who worked for the manufacturer were interviewed; all interviews took place in Finland. The participants worked in a variety of design-related roles including design engineers, project managers, safety experts and user interface specialists.

Safe and Inclusive Design of Equipment Used in the Minerals Industry

29

Rather than asking a rigid set of pre-defined questions, the interviews were topic based, allowing the interviewer to prompt, clarify and further probe the thoughts and practices of those being interviewed. Three topic areas were covered: Topic 1: The current design situation. Interviewees asked about the formal and informal human-related methods used to create and assess safety of the equipment. Specifically, they were asked to consider methods that considered end-user interaction. For each method noted, the interviewees were asked to describe the strengths and weaknesses of that technique. Topic 2: OMAT review. The OMAT process was then described to the participants (many already had a basic knowledge of OMAT). The participants were asked to judge and predict the strengths and weaknesses of this technique, and to state how it might be integrated within their overall design processes. Topic 3: End-user input. Of course, gaining end-user input is a common theme in safe, participatory and inclusive design. However, this has increasingly become difficult as the design, manufacture and use of equipment commonly occurs in different countries across varying cultural and language barriers. As an example of this, the interviews were conducted in Finland for equipment that would be used ultimately in Australia or elsewhere. This topic explored the currently-used methods of gaining end-user input into their design processes, especially as there was no homogeneous end-user group. It then explored two practices that could be used within the overall OMAT methodology to potentially gain more comprehensive end-user feedback: the inclusion of more (and varied) end-users in the OMAT workshops and the collection of video footage of tasks being performed with current equipment.

3.2.3 Results The results are described in terms of the three topics mentioned above. For reasons of space, only summary findings are reported here. Topic 1: The current design situation. The interviewees reported a large number of broadly human-centred methods to improve the safety of their equipment. They also reported that considering the end-user was a constant informal consideration. However, only four methods were noted that specifically included some consideration of the end-user. Two of these methods were only used by specific design specialists. The first was an “Ergonomic Checklist for the Operator’s Cabin”. This primarily involved ensuring that the adjustability of elements like reach distances to controls and chair design for weight of operator was accounted for using anthropometric data. The second method was in-field usability observation. However, this was limited to new digital controls used during operation (primarily of automated equipment). The interviewees thought that these methods were potentially useful; however, the scope and application of them was extremely limited.

30

Horberry and Cooke

The third method involved direct customer feedback from the use of current products. This was noted to be beneficial as many issues only emerged during equipment use. However, the major issue related to the ‘voice of the customer’ was that the input was only text-based and needed to pass through many hands before it reached the appropriate designer. By this stage it was commonly difficult to understand the precise issue and the person who raised it was not contactable. Therefore, only easily explainable and understandable issues fed through the current system with consistent success. The fourth, and most noted, method was that of a risk assessment workshop using qualitative risk matrices. The issues primarily came from equipment standards and were a mix of hazards and requirements. The primary positive issue noted about this method was that the workshop forced conversation amongst the stakeholders. The most significant negative issue noted was that the discussion focused only on the rather restrictive set of issues mentioned in existing standards. This commonly did not involve the end-user but, rather, ‘have we met the standard?’ This meant that significant issues related to operator interaction could be missed, and innovation might be stifled. Furthermore, the interviewees noted that extremely rarely did the method actually lead to design changes. As such, the interviewees mostly viewed this method as a legal cover rather than a useful design method. Topic 2: OMAT review. Despite the potential bias that might have been present (where one of the originators of the process was conducting the interviews), the interviewees’ opinions of OMAT were largely positive. In particular: • they welcomed the task-based, user-centred approach; • they found the whole OMAT process to be easy to understand; • it was viewed as comprehensive; • they predicted that it might encourage innovation in design solutions. However, some of the more negative features noted were: • • •

the OMAT process might take a long time (especially in workshops); it might be difficult to maintain/update this method; some concerned was expressed that it does not link sufficiently to standards. Although this was not fully articulated, there was more of a general feeling of unease about how much it became the workshop’s judgement (rather than an ‘objective’ standard) and having that judgement documented for a lawyer to subsequently investigate.

Topic 3: End-user input. The inclusion of a variety of end-users in design workshops (whether using OMAT or other methods) was reasonably positively received. This was because operational knowledge was seen to offer valuable design opportunities by knowing ‘what really happens’. However, there were also significant negative issues noted (that are also applicable to many other participatory, inclusive or safe design processes): • •

the end-user inputs received are limited to the knowledge of that person; legal issues (‘what happens if we don’t take their advice?’);

Safe and Inclusive Design of Equipment Used in the Minerals Industry

• • •

31

disclosure of information about design, so market value might be potentially reduced; they already have in-house operational ‘experts’ involved so current endusers might not add much extra information; pragmatically, and certainly when considering universal design, the potential users come from across the world and speak different languages. This was perceived to be a potentially costly exercise for little gain.

The technique of obtaining end-user inputs by means of video recording of current tasks was viewed by every interviewee as an extremely beneficial addition. In particular, it was opined that video records provided objective visual information that is easily shared, showed the designers how their equipment was actually used, acted as a memory aid for those designers who had actually been to a mine site and more generally could be an excellent reminder that real people will eventually have to use the equipment. Video records are, of course, not as an adequate replacement of comprehensive end-user feedback; the interviewees noted that they were unable to ask the person about the task, so had to interpret the actions on the video. Equally, other design options could not be explored, and if a design is changed significantly then the task itself changes. Finally, video records were only useful in the latter stages of the design process, so a designer would have to wait for the equipment to be out in the field before they can get this first iteration of the feedback.

3.3 Conclusions Including end-user input in some form in the design of equipment is a laudable goal in virtually any domain. As indicated in other domains that have successfully integrated automation (e.g. aviation), unless new technology in mining takes into account the human element that will ultimately operate or maintain the systems, then it is unlikely that such technology will flourish (c.f. Sheridan, 2002). This paper has hopefully shown some of the challenges and opportunities for those mining equipment designers who have a human-centred focus. Some of the challenges include designers actually getting access to mine sites, and obtaining appropriate end-user input, due to the widely diverse workforce involved. Not surprisingly, a disjointed group of user-centred methods currently exists, and these are employed to varying degrees by mining equipment designers. Despite this, safe design in some guise is certain to become more firmly entrenched in the design practices of larger OEMs. Consequently, it is anticipated that the task-based nature of OMAT will result in it being more widely accepted and deployed, especially for high frequency/high hazard tasks. A criticism of OMAT was that the method could be quite time-intensive, especially when involving extensive end-user workshops that needed to be formally documented. One way to address such a criticism would be to employ a streamlined version of the method that is still task-based and participatory, but only focuses on design issues of the highest priority tasks in a less formally documented workshop-style setting. Also, obtaining end-user inputs

32

Horberry and Cooke

by means of video recording of current tasks with the equipment is another way of streamlining the OMAT method whilst still retaining its task-based approach. Inclusive design of mining equipment is far less well-developed and wellaccepted than safe design. Indeed, philosophically, it might be maintained that specialised and hazardous mining equipment should not be designed for universal use unless reasonable controls (including training and following site procedures) are deployed. Still, designing accessible and usable mining equipment without excessive adaptations for the worldwide minerals industry user-group is certainly of importance for both safety and productivity. It is anticipated that user-centred design tools such as OMAT can be expanded and further employed to help achieve this goal.

3.4 References Burgess-Limerick R, Straker L, Pollock C, Dennis G, Leveritt S, Johnson S (2007) Participative ergonomics for manual tasks in coal mining. International Journal of Industrial Ergonomics, 37: 145-155 Cooke T, Horberry T (2011) The operability and maintainability analysis technique: Integrating task and risk analysis in the safe design of industrial equipment. In: Proceedings of the International Conference on Ergonomics and Human Factors 2011, Stoke Rochford, UK EMESRT (2011) Earth moving equipment safety round table. Available at: http://www.mirmgate.com/index.php?gate=emesrt (Accessed 8 August 2011) Horberry T, Burgess-Limerick R, Steiner L (2010) Human factors for the design, operation and maintenance of mining equipment. CRC Press, Boca Raton, FL, US Horberry T, Sarno S, Cooke T, Joy J (2009) Development of the operability and maintainability analysis technique for use with large surface haul trucks. Australian Coal Association Research Program report. Available at: http://www.acarp.com.au/ abstracts.aspx?repId=C17033 (Accessed 8 August 2011) Komljenovic D, Kecojevic V (2007) Risk management programme for occupational safety and health in surface mining operations. International Journal of Risk Assessment and Management, 7(5): 620-638 Lynas D, Horberry T (2011) Human factors issues with automated mining equipment. Ergonomics Open, 4(Suppl 2-M3): 74-80 Safe Work Australia (2011) What is safe design? Available at: http://www.safeworkaustralia.gov.au/SafetyInYourWorkplace/SafeDesign/Understanding /Pages/WhatIs.aspx (Accessed 8 August 2011) Sanders MS, Peay JM (1988) Human factors in mining (IC 9182). Department of the Interior, Bureau of Mines, Pittsburgh, PA, US Sheridan T (2002) Humans and automation. John Wiley, New York, US Simpson G, Horberry T, Joy J (2009) Understanding human error in mine safety. Ashgate Press, Farnham, UK

Chapter 4 Deploying a Two-player System for Arm Rehabilitation in Schools R.J. Holt, A.P.H. Weightman, J.F. Gallagher, N. Preston, M.C. Levesley, M. Mon-Williams and B. Bhakta

4.1 Introduction Cerebral Palsy (CP) is the commonest cause of disability among children in Europe (Johnson, 2002). Its effects and severity can be extremely varied, but a combination of arm impairments through weakness or spasticity and sensory deficits is common and can (i) significantly impair the ability of individuals with CP to carry out daily activities and (ii) create significant social barriers (Imms, 2008). Therapy is often used to aid the acquisition of motor skills, particularly in childhood, but a lack of physiotherapy resources means that this is often delivered through a self-managed home exercise programme. Exercises are frequently dull and repetitive, and children often lack the motivation to carry out these exercises, leading to poor compliance with the prescribed plan (Chappell and Williams, 2002). One solution to this is the use of Interactive Computer-Play (ICP)-based therapy (Sandlund et al., 2009), where therapy is delivered as a game through a computer-interface. Prior research at the University of Leeds has led to the development of a game-based system for home rehabilitation of upper limb impairment (Weightman et al., 2011). However, feedback from the participants in that project indicated that they preferred to play games with friends and suggested that this would improve their motivation further. Accordingly, a multiplayer ICPbased therapy system for upper limb rehabilitation intended for use in schools has been developed, and this paper describes both the system and the results of its initial deployment in a school environment. A brief overview of the background to ICP-based therapy and arm rehabilitation is given, followed by a description of the system itself, and finally the outcomes of deploying the system in a school environment for an eight-week period, including the amount of usage, feedback on the games used, and user comments received.

P. Langdon et al. (eds.), Designing Inclusive Systems, DOI: 10.1007/978-1-4471-2867-0_4, © Springer-Verlag London 2012

33

34

Holt et al.

4.2 Background ICP-based therapy has been growing in popularity in recent years, particularly with the advent of home computers and videogame consoles and the growing popularity of videogaming as a pastime in the last few decades. The development of consoles which use movement-based interaction in videogames, most notably the Nintendo WiiTM, has led to great interest in their use as a means of encouraging physical activity among children and making rehabilitation enjoyable (Deutsch et al., 2008; Lanningham-Foster et al., 2009). The use of off-the-shelf videogame consoles in rehabilitation has many benefits, as they enjoy the economies of scale of mass production, do not require specialist development, and games are already designed first and foremost to be enjoyable. However, they also have limitations: they are not necessarily accessible to players with more significant arm impairments; they do not provide the assistive force that a physiotherapist would provide (which is important in extending a patient’s capabilities); the games are not necessarily designed to provide therapeutic benefits; and these systems do not enforce compliance with a therapeutically desirable trajectory, which means that players can get away with making motions that are successful in the game, but do not provide therapeutic benefit. Research at the University of Leeds has promoted the use of Assisted Movement Devices as an adjunct to therapy, whereby a robotic system is used to provide both the assistive force and to encourage compliance with therapeutically desirable trajectories. Previous research has developed systems that use this approach with stroke patients in a clinical environment (Jackson et al, 2007) and for children with in a home environment (Weightman et al, 2011). In response to feedback gathered from the latter project, a two-player system has been developed, with the aim of deploying it in a school environment, where it would be easier to find multiple players. Social interaction such as co-operation and competition in games has long been identified as a motivator for playing (Malone and Lepper, 1987), and continues to be recognised as an important aspect of making games enjoyable (Sweetser and Wyeth, 2005). However, it does raise significant challenges in ICP-based therapy, as different players will have different levels of impairment (and in some cases, none at all), making it difficult to create a level playing field. The next section reviews the system developed, and how it addresses some of these issues.

4.3 The System The original home-based system had been based around a Microsoft SideWinderTM joystick, which had been adapted so that its force feedback system would provide the assistive force required to help players make the required movements (Weightman et al., 2011). This plugged via USB into a conventional PC, where the specially designed games would run. It was originally envisaged that the new system would take a similar form, with perhaps four to six such joysticks plugging

Deploying a Multiplayer System for Arm Rehabilitation in Schools

35

into a PC to allow multiplayer gaming. However, a number of factors meant that this approach was not feasible. Firstly, the Sidewinder was not able to provide as much force as desired, meaning a move towards a new design with larger motors and bespoke control software developed in LabVIEW and delivered via a National InstrumentsTM cRIO (Compact Reconfigurable Input-Output) controller. Furthermore, in working closely with teachers, children and parents to develop the system, it soon became apparent that class time was at an absolute premium for teachers, and they argued that they could not afford to spend take even a few minutes out of class time setting the system up and getting started without the children losing interest. Space was also at a premium, meaning that a four to six player system would be prohibitively large - teachers felt that this was redundant as they rarely had more than one or two children in a class requiring physiotherapy and were reluctant to take many unimpaired children out of lessons to participate in the games. Accordingly, the system was designed as a self-contained unit with two joysticks, two monitors, a PC and cRIO as shown in Figure 4.1. The system was wheeled so that it could be easily moved between classrooms or out of the way as needed, and required only a single button press to start up or turn off, booting directly into the games as needed.

Figure 4.1. The system

To address the need for a level playing field between different players, an Adaptation to Player Performance Algorithm (APPA) was developed. Four games were developed, representing different combinations of competitive and collaborative, sequential and simultaneous play. They all centred around the same premise, delivered through simple cut scenes: the players are monkeys trying to rescue their friends from a hungry crocodile in order to provide elements of fantasy

36

Holt et al.

and curiosity as recommended by Malone and Lepper (1987) and Sweetser and Wyeth (2005); and all were based around the simple back-and-forth movement of a conventional reach/retrieve exercise. The APPA first involved the players carrying out a simple single-player assessment task at the start of each session, in which each player guided their ‘monkey’ around the screen and tried to collect as many ‘bananas’ as possible in the time available. The system then adjusted the amount of assistance provided based on their performance in this task. The four actual games were developed with a user group of children with cerebral palsy who had participated in the previous home-based project (Weightman et al., 2011). These children gave feedback and made comments on early iterations, evaluating the initial concepts and gameplay proposals, giving feedback on early prototypes and testing the near final games for usability. The final four games selected were: 1. Van Game (Simultaneous, Cooperative): Players work together to collect bananas and destroy the crocodile’s van before s/he escapes with their monkey friends. 2. River Game (Simultaneous, Competitive): Players race against each other to collect the most bananas as they race along a winding river. 3. Chase Game (Sequential, Competitive): Players take it in turns to find their way through a maze, one playing as the crocodile, the other as a monkey - the player with the fastest time wins. 4. Maze Game (Sequential, Cooperative): Players work together to find a way through a maze before time runs out, as each player collects bananas so new paths open to the other player. Screenshots from these games are shown in Figure 4.2. Each game also had a single player variant, in which the player raced against the clock, rather than the other player.

4.4 Outcomes of Deployment To evaluate the feasibility of using the system in a real school environment, the system has been field-tested in seven schools with an overall total of eight children with CP aged between 8 and 12 years (to date). At this stage, the purpose of this testing is to assess whether the system can actually be delivered and used in a school without supervision by the research team, rather than assessing whether the system delivers therapeutic benefits.

4.4.1 Process A total of four systems were built. The systems were deployed into schools for two periods of four weeks, with a one-week “washout” period in between. In one period, the system was used in a single-player mode; in the other, it was used in a multiplayer mode. Half the schools were randomly assigned to use the single player mode first, while the other half used the system in multiplayer mode first.

Deploying a Multiplayer System for Arm Rehabilitation in Schools

37

As there were seven schools and just four systems, the systems were deployed in four schools first (Schools A, B, C and D) and in the remaining schools thereafter.

a) Van Game

b) River Game

c) Chase Game

d) Maze Game

Figure 4.2. Games for use with the system

The research team delivered the system to each participating school, explained to the member of staff who would be given responsibility for the system how to set it up and play the games, and then left the system under the supervision of the school for the four-week period. Teachers were introduced to the goals of the research project, made aware that the system was intended as a way of delivering therapy through gameplay and informed that the version of the system delivered was a prototype for evaluation with the aim of assessing the feasibility of its use in a school environment. The initial explanation was supplemented by an instruction manual, and a contact number to call in case of significant problems. The aim was to see whether and how each school would use the system in practice, without the research team’s input, so they were permitted to make whatever use of the system they saw fit. The systems gathered details of the amount that each child played, and a feedback questionnaire based on Read’s (2008) Smileyometer was used for each child to rate each of the games and the assessment task. Finally, a debriefing

38

Holt et al.

questionnaire was used to capture the views of adults and children about the system, any problems or benefits they encountered, and what might be done to improve it in the future. As a result of difficulties in getting time in teachers’ diaries for interviews in the early phase of the project, it had been agreed that the most effective way of gathering information was to provide paper feedback questionnaires with a stamped self-addressed envelope that the school could return to the researchers at their convenience, rather than attempting to arrange formal feedback interviews.

4.4.2 Usage Table 4.1 records the amount of usage made of the system for each child at each school (note that School C had two children), in terms of the number of days upon which the system was used, and the mean length of play on those days. In addition, it shows the mean number of days on which the system was used across all the children, and the mean length of all their sessions. Data were only recorded for children with cerebral palsy as they were the target users of the system: the project was not concerned with the amount of therapy delivered to children who did not require it. Table 4.1. Usage of system by school and child School

Child

Days Used Single Player (of possible 20)

Mean Session Length Single Player (mins)

Days Used Multiplayer (of possible 20)

Mean Session Length Multiplayer (mins)

A

1

10

6.68

19

9.38

B

2

5

5.26

13

10.2

C

3

15

21.5

19

17.2

C

4

15

17.7

16

19.0

D

5

12

8.71

10

9.51

E

6

15

3.38

12

6.20

F

7

13

15.5

10

9.70

G

8

10

13.3

16

20.5

11.9

12.7

14.4

13.4

Overall Mean

All the children made some use of the system and with just one exception (Child 2’s single player phase), it was used on at least half the available days. In some phases, days were lost owing to school closures in bad weather and bank holidays, but these were accepted as part of a realistic snapshot of how much use the system received in practice (as these were all factors that might limit its use they are not corrected for in the table above). On average, multiplayer mode was

Deploying a Multiplayer System for Arm Rehabilitation in Schools

39

played more often and for longer than the single player mode, despite the difficulty in having to take children without arm impairments out of lessons to accommodate this - though with such a small sample, the difference was not statistically significant.

4.4.3 Game Ratings Getting schools to complete and return questionnaires proved extremely difficult, and several rounds of chasing up were required. A total of fourteen smileyometers were returned, though we know that a total of thirty-two children across the seven schools (including the eight with CP) used the system altogether. Unfortunately, the questionnaires were not always correctly labelled, making it impossible at some schools to determine which feedback was from children with CP and which from children without. Accordingly, the results presented here represent the overall feedback from all the children that provided Smileyometers rather than trying to distinguish between the different groups of users. The results for each game, including the assessment task, are shown in Figure 4.3.

Figure 4.3. Smileyometer ratings for the four games and assessment tasks

40

Holt et al.

Overall, the two co-operative games - the Van and Maze Games - proved most consistently popular, with the Maze Game receiving no negative ratings (Not Very Good or Awful), and the Van Game receiving just one not very good. The River Game caused the greatest split in opinion, receiving the most Brilliants, but also receiving the most negative ratings of the four games. The Chase Game also received mixed feedback, but with a more even split across the ratings. The Assessment Task also split opinion, receiving more negative ratings than any of the games, but also surprisingly receiving a large number of positive ratings, with only the River Game receiving more positive ratings. It is worth noting that the ordering of the games varied from child to child quite significantly as well: each game was most popular for at least two children, and least popular for at least two others. On the whole, this suggests that the games were well-received, although there were substantial variations in individual preferences, and several children did suggest that more games would be welcome to provide greater variety.

4.4.4 Qualitative Comments Feedback comments from the school staff responsible for the system focused consistently on the size of the system, and ease of setup. One school indicated that manoeuvrability was not important, as the size of the system meant that it could not be moved anywhere else. The other schools indicated that being able to move the system around easily was very important, as it needed to be used in different classrooms, or moved out of the way when not in use. Every school was able to find space for the system, and none felt it was too big, but all agreed that a system any larger than this would be untenable. Ease of setup was also an issue, with all schools praising the ease with which the system could be started, but also reporting that the system sometimes failed to initialise properly the first time it was switched on. While they were always able to resolve this by restarting the system, School B in particular complained that this cut significantly into the short periods they were able to find for use. Finally, schools E, F and G all indicated that exam preparation cut into the amount of time available for therapy, particularly in the second phase of their deployment (the multiplayer phase for Schools E and F, single player for School G), and that this made it particularly difficult to find time to use the system. These comments all confirmed our initial finding from working with teachers when developing the system in the first place, that manoeuvrability, ease of setup and a small footprint are all important for a school environment. This demonstrates not only the merit of the revised concept of a self-contained unit over the initial notion of four to six independent joysticks, but also the importance of engaging with users to properly understand their needs.

Deploying a Multiplayer System for Arm Rehabilitation in Schools

41

4.5 Conclusions and Further Work This paper has reported the deployment of a novel multiplayer gaming system for delivering arm rehabilitation in a school environment. The system has been deployed in a real environment, and this has demonstrated that the system can be used in a variety of schools. The games presented have been well-received despite their simplicity although there were significant variations in individual preference and a wider range of games would be desirable if the system were to be used in the longer term. There are a further three schools still to test the system before final conclusions are drawn. In addition, this project has only assessed the feasibility of deploying and using the system in a school environment. Clinical trials with a larger number of children will be required to demonstrate the therapeutic efficacy of this approach.

4.6 References Chappell F, Williams B (2002) Rates and reasons for non-adherence to home physiotherapy in paediatrics. Physiotherapy, 88(3): 2-11 Deutsch JE, Borbely M, Filler J, Huhn K, Guarrera-Bowlby P (2008) Use of a low-cost, commercially available gaming console (Wii) for rehabilitation of an adolescent with cerebral palsy. Physical Therapy, 88(10): 1196-1207 Imms C (2008) Children with cerebral palsy participate: A review of the literature. Disability and Rehabilitation, 30(24): 1867-1884 Jackson AE, Holt RJ, Culmer PR, Makower SG, Levesley MC, Richardson RC et al. (2007) Dual robot system for upper limb rehabilitation after stroke: the design process. Proceedings of the Institution of Mechanical Engineers Part C: Journal of Mechanical Engineering Science, 221(7): 845-857 Johnson A (2002) Prevalence and characteristics of children with cerebral palsy in Europe. Developmental Medicine and Child Neurology, 44(9): 633-640 Lanningham-Foster L, Foster RC, McCrady SK, Jensen TB, Mitre N, Levine JA (2009) Activity-promoting video games and increased energy expenditure. Journal of Pediatrics, 154(6): 819-823 Malone TW, Lepper MR (1987) making learning fun: A taxonomy of intrinsic motivations for learning In: Snow RE, Farr MJ (eds.) Aptitude, learning and instructions, Vol. 3: Cognitive and affective process analyses. Laurence Erlbaum Associates, Hillsdale, NJ, US, pp 223-253 Read JC (2008) Validating the fun toolkit: An instrument for measuring children’s opinions of technology. Cognition, Technology and Work, 10(2): 119-128 Sandlund M, Mcdonough S, Hager-Ross C (2009) Interactive computer play in rehabilitation of children with sensorimotor disorders: A systematic review. Developmental Medicine and Child Neurology, 51(3): 173-179 Sweetser S, Wyeth P (2005) Gameflow: A model for evaluating player enjoyment in games. ACM Computers in Entertainment, 3(3): 3A Weightman APH, Preston N, Levesley MC, Holt RJ, Mon-Williams M, Clarke M et al. (2011) Home based computer-assisted upper limb exercise for young children with cerebral palsy: A feasibility study investigating impact on motor control and functional outcome. Journal of Rehabilitation Medicine, 43(4): 359-363

Chapter 5 Evaluating the Accessibility and Usability of Blogging Platforms for Blind Users B. Wentz, M. Cirba, N. Kharal, J. Moran and M. Slate

5.1 Introduction Web-based social media have become a dynamic way of allowing the inclusive communication of many perspectives from diverse backgrounds. In a recent survey from WebAIM, blogging was noted to be the dominant form of social media according to screen reader users (WebAIM, 2010). Social media and other forms of technology can provide a platform for users with disabilities to interact and communicate on a level playing field with anyone else throughout society. It is the responsibility of application designers to follow established standards of accessibility and usability in order to provide equal access to this dynamic technology.

5.2 Related Literature There are approximately 284 million individuals worldwide who are visually impaired (low vision), and this figure includes 39 million individuals who are completely blind with no residual vision (World Health Organization, 2011). A screen reader (such as JAWS, System Access, VoiceOver, or Window-Eyes) is a software application that reads the content of a computer screen out loud in a linear manner, using computer-synthesised speech. This is the primary way that blind users access computers and web sites. Braille devices are often too expensive, and the rate of Braille literacy among blind users very low. For example, in the United States, fewer than 10% of children who are blind are learning to read Braille (National Federation of the Blind, 2011). In 2005 a small study was conducted to evaluate the accessibility of several blogging platforms (American Foundation for the Blind, 2010). It was concluded that problems such as CAPTCHAs, which are required to register and create most blogs, create accessibility challenges. Also highlighted were accessibility

P. Langdon et al. (eds.), Designing Inclusive Systems, DOI: 10.1007/978-1-4471-2867-0_5, © Springer-Verlag London 2012

43

44

Wentz et al.

challenges related to navigating many of the blog platforms. Common problems discovered were related to basic principles of accessible design, such as properly labeling the fields on a web-based form. Recently, WordPress has made an attempt to provide some general accessibility guidelines for bloggers who use that platform (WordPress.org, 2011). Our project was intended to provide a more recent examination of a more narrow focus, namely to evaluate accessibility and usability by enabling users who rely on screen readers to access the platforms. Our usability testing examined the ability to read and post comments, and our accessibility evaluation evaluated the technical ability required to create a blog and access the main administrative console of a blog platform.

5.3 Research Methodology and Results 5.3.1 Usability Testing 5.3.1.1 Blog Platform Selection Blogger (a Google product) and WordPress were selected for usability testing due to suggestions that together they are the largest platforms, providing the background structure for approximately 85% of all blogs (Peltier, 2009). WordPress offers both WordPress.com (which provides free blog hosting and/or paid, ad-free blogging) as well as WordPress.org (which provides blogging software available as a download to be installed on a server for localised hosting). Since the default templates are the same for WordPress.org and WordPress.com, WordPress.com was selected since it is a hosted system through which any individual is permitted to establish a free blog. The focus of the following usability testing was on the default templates within both Blogger and WordPress, since both platforms offer numerous templates. 5.3.1.2 Participant Selection for Usability Testing The participants recruited for this study were self-labeled as blind, JAWS screen reader users, and at least 18 years of age. We sent recruitment emails to the Pennsylvania chapter of the National Federation of the Blind as well as the Pennsylvania Council of the Blind. Additionally, the recruitment email was posted on listservs that are commonly used by blind individuals within the state. Since there is no central directory of all blind individuals worldwide (or in the United States), a true random sampling would be technically impossible. A total of 15 participants were involved in the usability testing. Our goal with this evaluation was to conduct exploratory research with a group of blind users, in order to discover usability challenges that could be corrected to provide improved accessibility and usability of web-based blog platforms. 5.3.1.3 Data Collection for Usability Testing Usability testing of Blogger and WordPress was conducted in March and April 2011. The testing was conducted using an Acer Aspire One netbook with the Windows XP operating system and Internet Explorer 8. An external keyboard,

Evaluating the Accessibility and Usability of Dominant Blogging Platforms

45

external speakers, a wireless broadband card (for testing the web-based applications) and JAWS 12 (screen reader software) were used for the usability testing. A stopwatch was used to record the time spent on each task. We spent time creating test blogs and accounts that would make sure that no personal data of the participants was used. Each session was started by asking each participant questions about their background and blog experience, and then participants were asked to rank the importance of reading blog comments, posting blog comments and creating a blog. After the initial questionnaire, we read each task to the participants, and then they tried to complete each task without our assistance. Each task was timed and recorded for completion, and participants could choose to stop the task at any time. We noted problems and comments when they were mentioned by participants or observed by us. 5.3.1.4 Demographics of Usability Testing The participants in our study ranged in age from 34 to 63, with the mean age being 54. There was a close ratio of male (seven) to female participants (eight). Most of the participants were college graduates or had completed some college (12 out of 15). The most common operating system was Windows XP, and the most common Web browser was Internet Explorer 8. The versions of JAWS used ranged from version 8 to 12, with the most common version being version 12. The mean number of years of experience using JAWS was 10.8 years. Since the average blind individual might have less education and screen reader experience than the participants in this study, it is likely that they would have more difficulty with these blog interfaces than the study participants did. 5.3.1.5 Previous Experience and Value of Blogging When asked about their previous experience with blogs, 80% (12 out of 15) of the participants reported having previously read comments that were posted on a blog, and out of those participants, 47% (7 out of 15) recalled having no difficulty when reading blog comments. Participants were then asked about posting comments on a blog, and only 47% (7 out of 15) reported previously posting comments on blogs. Of those participants, 57% (8 out of 15) recalled having difficulty when posting blog comments. Only 27% (4 out of 15) of the participants reported previously creating or attempting to create a blog, and out of those, two of the four reported having difficulty when trying to create a blog. Participants were then asked three questions related to the value that they place on being able to read comments, post comments, and create blogs. On a scale of 15 (with 5 being the most important), the mean responses were very close for all three. Reading comments had a mean of 3.5, posting comments had a mean of 3.7, and creating blogs had a mean of 3.5. This indicates that for this sample, a moderate value is placed on all aspects of interacting with blogs. 5.3.1.6 Testing Results and Impact of Previous Experience After the preliminary questions about demographics and prior experience with blogs, users were then asked to complete two tasks in both Blogger and WordPress. For Task 1, participants were asked to read any comment posted on the

46

Wentz et al.

blog web page. For Task 2, participants were asked to post a comment about the article on the blog web page. The tasks with the results are illustrated in Table 5.1. Blogger had a higher rate of completion for Task 1, but the task took longer to complete than it did in WordPress. No participants were able to complete Task 2 in Blogger due to a usability problem which is described later. Table 5.1. Tasks and the completion rate, mean time, and SD for successful tasks Task 1: Read a comment 1: Read a comment 2: Post a comment 2: Post a comment

Blog Platform Blogger WordPress Blogger WordPress

Completion Percentage 80% (12 out of 15) 67% (10 out of 15) 0% 53% (8 out of 15)

Mean Time 144.3 sec. 126.8 sec. 129 sec.

Standard Deviation 75.2 sec. 36.1 sec. 54.3 sec.

The impact that previous experience with reading and posting comments had on the results of this usability testing is illustrated in Table 5.2. It is important to note that this does not reflect previous experience with a particular interface but rather previous experience with either reading a blog comment or posting a blog comment on any blog interface. For task 1 on the Blogger interface, it appears that there is an improvement in usability based on prior familiarity, but it is important to note that there were only a small number of users (three) who were inexperienced with task 1. Task 2 in WordPress does seem to indicate a usability curve based on experience, since the number of experienced to inexperienced users was much closer, yet the task completion rate was heavily weighted towards experienced users. In addition, the mean time to complete successful tasks was higher for inexperienced users for Blogger task 1, and for WordPress tasks 1 and 2. Table 5.2. Previous experience and mean completion rate/time for successful tasks Task/Interface Task 1, Blogger

Users with Experience 12 out of 15

Task 1, WordPress

12 out of 15

Task 2, Blogger

7 out of 15

Task 2, WordPress

7 out of 15

Experienced Users 92% (11/12) 135.6 sec. 67% (8/12) 119.6 sec. 0% 71% (5/7) 113 sec.

Inexperienced Users 33%. (1/3) 217.7 sec. 67% (2/3) 155.5 sec. 0% 38% (3/8) 155.7 sec.

5.3.1.7 Usability Problems Identified While many minor usability challenges were identified during the usability testing conducted for this study, there were several major usability challenges which, if corrected, could dramatically improve the usability of these blog interfaces. Usability problems with Blogger included a “Post Comment” button that did not work as expected and an audio CAPTCHA that is not embedded and requires an external plug-in to work. These problems effectively prevent users from successfully posting comments.

Evaluating the Accessibility and Usability of Dominant Blogging Platforms

47

Figure 5.1. Post comment problem in blogger

In Figure 5.1, there is a screen shot of the “Post a Comment” form for the Blogger interface. While this should be a simple, straightforward process, when a user (blind or sighted) selects the “Post Comment” button, the action is not performed, and instead the user is redirected back to the same page, and the comment is not posted. It takes several attempts before the user is directed to the login prompt for account credentials. This does not occur if a user is already logged into a Google Account. While this is frustrating for any user, the problem is not immediately evident to a blind user. The only indication that this is occurring is the small red image that is shown in the screenshot above. One user, who tried this process repeatedly and did reach the next step, was taken to a prompt for a CAPTCHA. Figure 5.2 illustrates the traditional CAPTCHA screenshot for the Blogger interface with a symbol indicating an audio option for accessibility. The alternate text for the audio option says “Listen and type the words you hear.” The problem is that the user was unable to proceed because the CAPTCHA did not rely on embedded audio but on a particular format of audio which required a plug-in for use.

Figure 5.2. Audio CAPTCHA plug-in problem in blogger

48

Wentz et al.

Usability problems with WordPress included a link that directs users to the wrong field in a form, and an error message that could be improved for usability. When a user selects the reply link to add a comment, the focus of the cursor goes to the “Comment” field on the form. For a sighted user, it would be evident that “Name” and “Email” are required fields, but for a user relying on a screen reader and keyboard navigation, the usability problem with the “Reply” link is confusing and could prevent a successful post unless the user realises what the problem is.

5.3.2 Accessibility Evaluations Accessibility evaluations often rely on automated software (such as A-Prompt, Deque WorldSpace, and WAVE). Automated evaluations can determine that a web page generally fails compliance with standards such as Section 508, but automated evaluations are often not as accurate as manual evaluations because accessibility problems such as alternate content descriptions that do not fit the context cannot be currently determined by automated means. Careful inspections using a screen reader (such as JAWS) are considered to be the most accurate form of accessibility evaluation (Mankoff et al., 2005), and that accuracy increases when multiple individuals evaluate the same interfaces (Lazar et al., 2010a). This process can be used to determine compliance with standards, such as the Section 508 standards for web sites. Accessibility evaluations complement usability evaluations in that usability evaluations can assess ease of use, whereas usability testing can often not completely evaluate compliance with standards since users may not be able to completely use all aspects of a system due to accessibility barriers. 5.3.2.1 Procedures for the Accessibility Evaluations The evaluations that we conducted utilised the web site accessibility guidelines of Section 508 (1194.22) of the US Rehabilitation Act, identified as paragraphs “a” through paragraph “p” (US Government, 2010). This inspection process involves an individual (previously trained on accessibility inspections) with vision inspecting a web site using a screen reader and a checking for compliance with each guideline from Section 508. Table 5.3 presents a list of the guidelines, along with a short description of each of the guidelines (note that the descriptions are summarised from Lazar et al., 2010a, and not a part of the Section 508 guidelines).

Evaluating the Accessibility and Usability of Dominant Blogging Platforms

49

Table 5.3. Description of each of the 16 paragraphs of the Section 508 web guidelines (a) Text Equivalent (have a text equivalent for any graphical elements) (b) Synchronised Equivalent Alternatives (have captioned video, transcripts of any audio, or other alternatives for multimedia) (c) Use of Color (color should not be used as the only method for identifying elements of the web page or any data) (d) Organisation (style sheets are encouraged, but users should still be able to utilise a web page when style sheets are turned off) (e) Redundant Text Links on Server-Side Image Map and (f) Client-Side Image Maps (redundant clickable links for server-side image maps, and accessible client-side image maps are preferred) (g) and (h) Row and Column Headers (use appropriate headers and markup to allow easy navigation of a table) (i) Frames (title all frames and label all frames for easy identification and navigation, e.g., use “navigation” “main content” and “search” rather than “top” or “bottom”) (j) Screen Flicker Frequency (limit or eliminate the use of flickering, which can provoke seizures) (k) Text-Only Page Default (if a web page cannot be made accessible, provide an equivalent text-only page, and make sure it is kept up to date) (l) Scripting Languages (make sure that equivalents for any non-accessible scripting are included, e.g., for those who are not using pointing devices) (m) Linked Plug-In or Applet (if any plug-ins are required, make sure to provide a link to an accessible version of the plug-in) (n) Online Electronic Forms (all forms must be properly labeled and accessible) (o) Method to Skip Repetitive Navigation Links (all web pages should have a link which allows a user to skip directly to the main content, bypassing any site navigation information) (p) Alerts on Timed Responses (if any page responses are timed, the user should be given the opportunity to indicate that more time is needed)

For this evaluation, we selected the pages used to create a blog as well as the main interface for managing the blog in both Blogger and WordPress (refer to Table 5.4). Table 5.4. Web pages on blogger and WordPress selected for evaluation Blogger.com Web Pages: Home page “Get started” page to create an account Page to name the blog Page to choose the template Blog management page

WordPress.com Web Pages: Home page Page to create an account Blog management page

For each web page, four individuals did a separate inspection using a screen reader, followed-up by a code inspection. After doing an individual inspection, the four individuals met, discussed the differences between their individual evaluations, interpreted, re-inspected the pages, and then agreed upon one common

Wentz et al.

50

evaluation for each web page. This is a common approach to provide a higher level of validity than a single individual review (Nielson and Mack, 1994). 5.3.2.2 Results of the Accessibility Evaluations The accessibility evaluation revealed some major accessibility barriers with the registration process and the management interface for both Blogger and WordPress. Each page that was evaluated had at least some minor accessibility violations. Some of the violations for Blogger included images with no alternate text (paragraph “a”), tables with headers that were missing mark-up (paragraph “h”), labels that were not properly associated with their controls, and poor error labeling (paragraph “n”). Blogger also violated paragraph “c” by using color alone to identify information on the “Get Started” page. During the blog template selection process, paragraph “l” was violated by the template selection feature being inaccessible with the keyboard alone. On the main page to manage the blog, Blogger violated paragraph “b” (with no caption for the video on how to use Blogger), paragraph “g” (with no headers in a table), and paragraph “l” (drop-down selections for text formatting were created using inaccessibly implemented JavaScript). On WordPress, the home page was missing alternate text for images (paragraph “a”), and a search form field was not properly labeled (paragraph “n”). On the sign-up page, color alone was used to convey meaning (paragraph “c”), there was a table without header information (paragraph “h”), there was a scripted drop-down list on a form that was not accessible (paragraph “l”), and there was another search form field that was not labeled (paragraph “n”). There were many violations on the main blog management page (7 out of 16 paragraphs), including missing alternate text (paragraph “a”), no captions for the tutorial video (paragraph “b”), missing headers for tables (paragraph “h”), untitled frames (paragraph “i”), a JavaScript “admin” bar that was not accessible (paragraph “l”), and no way to skip over navigational links to reach the main content (paragraph “o”). Figure 5.3 shows a screenshot of the JavaScript “admin” bar on the main blog management page which was only accessible by using a mouse.

Figure 5.3. Inaccessible “admin” bar in WordPress

Evaluating the Accessibility and Usability of Dominant Blogging Platforms

51

5.4 Discussion The major usability problems discovered during the usability testing process of the Blogger and WordPress interfaces are primarily simple issues that could easily be corrected by developers and could have been avoided through some basic usability evaluations of these interfaces. As far as the audio CAPTCHA plug-in problem with Blogger, it is ironic that Google owns both Blogger and reCAPTCHA, which is a company that develops CAPTCHAs for commercial use. The audio CAPTCHAs produced by reCAPTCHA use embedded audio, and while they may have usability problems related to audio CAPTCHAs in general (Lazar et al., 2010b), at least the external plug-in to play the CAPTCHA could be avoided. The accessibility problems identified during the accessibility evaluations of the Blogger and WordPress registration and management processes were also problems that could easily be corrected. Adding alternate text, adding labels to form fields, adding skip navigation links, and adding headers to tables are all very simple tasks for web designers. A basic accessibility evaluation would reveal these problems, and evaluations at regular intervals could prevent problems like these from occurring in the future. Companies who provide products such as Blogger and WordPress, and anyone who uses these products should be aware of these accessibility and usability problems. The impact of web-based blogs is far-reaching, and simply paying attention to the design of these common templates could significantly improve the online experience of millions of users.

5.5 Conclusions A few of the problems uncovered in this study could affect both blind and sighted users to some extent (such as the “Post Comment” problem in WordPress): however, navigating with only a keyboard and screen reader causes many of these problems to become significant. Designers should carefully follow national guidelines such as those set forth in US Section 508 and international guidelines such as the Web Content Accessibility Guidelines. Regular accessibility evaluations and usability testing would ensure that these popular blogging platforms are equally accessible to all users.

5.6 References American Foundation for the Blind (2010) Is blogging accessible to people with vision loss? Available at: http://www.afb.org/Section.asp?SectionID=57&DocumentID=2753 (Accessed 13 December 2010) Lazar J, Beavan P, Brown J, Coffey D, Nolf B, Poole R et al. (2010a). Investigating the accessibility of state government web sites in Maryland. In: Langdon PM, Clarkson PJ, Robinson P (eds.) Designing inclusive interactions. Springer, London, UK, pp 69-78

52

Wentz et al.

Lazar J, Feng J, Adelegan O, Giller A, Hardsock A, Horney R et al. (2010b) Assessing the usability of the new radio clip-based human interaction proofs. In: Poster Presentation at the 6th Symposium on Usable Privacy and Security (SOUPS 2010), Redmond, WA, US Mankoff J, Fait H, Tran T (2005) Is your web page accessible? A comparative study of methods for assessing web page accessibility for the blind. In: Proceedings of the 23rd ACM Conference on Human Factors in Computing Systems, Portland, OR, US, pp 41-50 National Federation of the Blind (2011) How many children in America are not taught to read? Available at: http://www.nfb.org/nfb/braille_initiative.asp (Accessed 5 May 2011) Nielson J, Mack R (1994) Usability inspection methods. John Wiley and Sons, NY, US Peltier J (2009) Which blogging platform do you use? Available at: http://peltiertech.com/WordPress/which-blogging-platform-do-you-use/ (Accessed 5 December 2010) US Government (2010) Section 508 standards guide. Available at: http://www.section508.gov/ index.cfm?fuseAction=stdsdoc#Web (Accessed 5 May 2011) WebAIM (2010) Screen reader user survey #3 results. Available at: http://webaim.org/projects/screenreadersurvey3/ (Accessed 5 January 2011) Accessibility (2011) WordPress.org. Available at: http://codex.WordPress.org/Accessibility (Accessed 15 January 2011) World Health Organization (2011) Visual impairment and blindness. Available at: http://www.who.int/mediacentre/factsheets/fs282/en/ (Accessed 5 May 2011)

Part II

Measuring Demand and Capabilities

Chapter 6 A Population Perspective on Mobile Phone Related Tasks M. Bradley, S. Waller, J. Goodman-Deane, I. Hosking, R. Tenneti, P.M. Langdon and P.J. Clarkson

6.1 Introduction For design to be truly inclusive, it needs to take into account the range of users’ capabilities. To do this appropriately, good data on those capabilities is needed. This paper reports on results from a postcode sampled survey of 362 people. The survey examined a wide range of user capabilities and characteristics, but the paper focuses on just a few of the survey measures. These measures examine some of the component activities involved in using mobile telephones: selection of a menu item via two different interaction patterns, use of differing sized pushbutton controls and insertion of two different types of electrical connector. These results can help to inform more inclusive design of mobile phones by examining how people’s capability to perform these activities varies across different activities and by age and gender. The survey from which this paper draws its data, aimed to test methods and materials in preparation for a follow-up survey with a bigger sample. Although the data in this paper was taken from a preparatory survey, there were 362 participants, and this sample size is big enough to achieve statistical significance for the conclusions drawn. Further details of how the survey was designed and conducted are now presented, followed by details of the specific tests that are reported here. Related research is also presented for each specific test.

6.2 Method The survey examined a wide range of human capabilities and characteristics related to product use. It was part of the i~design research programme (i~design, 2011). It was designed by the i~design research team, and conducted by the National Centre for

P. Langdon et al. (eds.), Designing Inclusive Systems, DOI: 10.1007/978-1-4471-2867-0_6, © Springer-Verlag London 2012

55

56

Bradley et al.

Social Research (NatCen). NatCen is a non-commercial professional survey organisation. 30 professional interviewers from NatCen conducted the actual surveys, and each interviewer received one day’s training supervised by the i~design team.

6.2.1 Sample The sample was recruited through random selection of private-household postcodes in England and Wales. Invitation letters were sent out to 990 postcode addresses, and each property was subsequently visited by one of the interviewers. At households that consented to take part, the interviewer selected a single occupant at random from those aged 16 and over, and the participants were not paid. 362 responses were obtained, and 53.6% of these were female. 23% of the sample were aged 16-34, 29% were aged 3549, 24% were aged 50-64 and 23% were aged over 65.

6.2.2 Survey Procedure The survey was conducted face-to-face using a computer assisted personal interviewing (CAPI) programme on a laptop. The survey examined a wide range of areas of human capability and product use: vision, hearing, dexterity, mobility, reach, cognitive function, technology/product use, psychological resources, and anthropometrics. The survey aimed to gather generic information that could be used to predict participants' ability to interact with products. Some of the tests were therefore designed to elicit information on participants’ basic capabilities, which are relevant to many aspects of product interaction. For example, vision tests measured participants’ visual acuity, which affects performance at a range of product tasks, such as reading text on a mobile phone screen and reading a manual. However, the survey also included some specific product tasks that require a combination of basic capabilities. For example, the ability to press particular buttons on a mobile phone was measured directly, because this typically requires hand-eye coordination and fingertip sensitivity. This paper focuses on the product tasks in the survey that relate specifically to mobile telephone use: selection of a menu item via two different interaction patterns, use of differing sized pushbutton controls and insertion of two different types of electrical connector. These measures are described in more detail below. Many of the other measures in the survey are also relevant to mobile phone use and design. The full results from the survey and its documentation are publicly available: please contact [email protected] for details.

6.2.3 Menu Selection As part of the survey, participants were presented with two interface styles for selecting an item from a list. We use the term ‘select and confirm’ to refer to an interface in which users navigate up and down the list using arrow keys and press a ‘select’ key when they reach their choice (Figure 6.1a). Similarly, ‘number navigation’ refers to an interface in which users choose an item by entering the number next to it (Figure 6.1b).

A Population Perspective on Mobile Phone Related Tasks

(a)

57

(b)

Figure 6.1. Mock-up paper prototype mobile phone interfaces used to assess (a) Select and confirm (b) Number navigation interaction styles

The ‘select and confirm’ style is more common in mobile telephones, but previous studies have indicated that some users experience problems with it. Lindholm and Keinonen (2003) describe how novice users can initially struggle with the concept of soft keys in interfaces of this type but this can be overcome through practice. The widespread adoption of this interface style for technology products means that many younger people are familiar with it, but many older users may still have limited or no experience with it. These older users are therefore more likely to struggle with this menu style. This difficulty was specifically observed for older users in digital set-boxes for television (Clarkson and Keates, 2003), which often have ‘select and confirm’ menus. Clarkson and Keates recommended the use of number lists as an alternative approach. Our study thus investigated whether the ‘number navigation’ approach could overcome the problems that some users experience with ‘select and confirm’. We expected that younger users would be experienced and proficient at using ‘select and confirm’ menus and that many older users (aged 65+) would not. The more direct interaction of ‘number navigation’, which requires only one key press to select a menu item, could be advantageous for such novices. These interaction styles were presented to users using a simplified paper prototyping method (Snyder, 2003). Respondents were shown images of mobile telephones with these interfaces (Figure 6.1), and were asked to indicate which buttons they would press to see what information is in the calendar. The order in which the interfaces were presented was counter-balanced. The interviewer scored whether the correct buttons were pressed in the correct order. The ‘number navigation’ interface requires fewer physical buttons to implement than the ‘select and confirm’ interface, and this reduction of complexity represents a confounding variable. However, the decision was made to present both types of user interface with the minimum number of buttons required to implement basic phone functionality. If this test revealed a useful performance difference, this could offer realworld benefit for the user. The test was limited by using a simplified paper prototyping method, where the participant was not given any feedback on button presses. This was necessitated by practical constraints; the survey was conducted by 30 different interviewers and with only one day of training to cover the whole survey, it was not feasible to train them to

58

Bradley et al.

reliably run any type of interactive prototyping. This limitation should be taken into account in interpreting the results, as failure rates on the tasks are likely to be higher without feedback to the user. However, the method was intended to measure whether, on first sight of an interface, the user’s first attempt to plan and execute the task would involve any incorrect or unnecessary actions. Even if a user could eventually succeed in a goal through trial, error and recovery, an interface that allows users to get it right first time is preferable. Another factor that should be taken into consideration is that ‘number navigation’ requires fewer key presses, and thus presents fewer opportunities for the user to make a mistake. This is a potential confounding variable which is hard to eliminate in a simple experiment of this type. The decision was made to present the user interfaces in a form that is likely to be used in practice. This has the advantage that, if there is a useful performance difference, then the improved interface can be incorporated in an actual mobile phone, even if there is potential uncertainty about the underlying reason for the benefit.

6.2.4 Pressing Buttons The survey also assessed participants’ ability to use differing sized pushbutton controls on a mobile telephone. The participants were presented with an LG KP170 mobile phone with a small five way keypad and four larger surrounding pushbuttons (Figure 6.2). They were shown the five way keypad and asked to press four of the keypad arrows, in the order of up, right, down and centre (‘OK’ button). They were then asked to press the four larger outer buttons in the order of bottom left, top left, top right and bottom right. After each task, the interviewer recorded whether the correct screen was shown. For the first task, this indicated whether the respondent had pressed the buttons correctly in the specified order. However, an experimental weakness meant that, for the second task, this only indicated whether the last of the buttons had been pressed correctly. Any participant who tried the task was also asked to rate how easy they found the task on a six point scale from very easy to impossible. They were not given any feedback on whether they had successfully completed the task before making this assessment. We expected participants (particularly older participants) to perform better with the large buttons, in accordance with Fitts’ Law (Fitts, 1954), because they require less accurate finger positioning and dexterity.

Five way keypad

Large outer buttons

Figure 6.2. LG KP170 mobile phone showing the buttons used in the study

A Population Perspective on Mobile Phone Related Tasks

59

6.2.5 Inserting Cables Participants were also tested on their ability to insert two different types of electrical connector. Participants inserted two connectors into an MP3 player (a Sansa Clip+): a 3.5mm headphone jack plug and a mini-USB plug (Figure 6.3). An MP3 player was used for purposes of experimental convenience, but it was felt that the task of insertion would not differ from that experienced on a mobile phone. Firstly, the MP3 player and headphone jack socket location were shown to the participants. They were then given the MP3 player and the headphone jack plug and were asked to insert the jack into the socket. The test was aborted if the participant took longer than 15 seconds. The participants were then shown the mini-USB socket location on the MP3 player, and the mini-USB jack (at a random but incorrect rotational orientation). They were handed the MP3 player the right way up with the mini-USB jack on top of it with the mini-USB symbol visible, and were asked to insert the jack in the indicated socket. After each task, the interviewer recorded whether the jack was inserted correctly. Participants who did the task successfully also rated how easy they found the task on a six point scale from very easy to very difficult.

Figure 6.3. MP3 player showing headphone jack and mini-USB sockets and plugs

We expected participants to perform better with the headphone jack than with the mini-USB cable because the jack on the mini-USB needs to be rotated to the correct orientation before insertion, but the headphone jack can be inserted in any orientation.

6.3 Results and Discussion 6.3.1 Menu Selection Results Overall, more people completed the ‘select and confirm’ task successfully than the ‘number navigation task’ (p<0.001, McNemar’s). For each of the tasks, successful task completion decreased with age (p<0.001, Pearson’s chi-squared) as shown in Figure 6.4. This was most evident in the two oldest groups, particularly for women in the 75+ group. For both tasks, there was no significant difference between men and women (p>0.05, Fisher’s exact).

60

Bradley et al.

As explained in Section 6.2.3, we expected that younger users would perform better with the ‘select and confirm’ interface than older users. This was supported by the results, with 72% of people under 65 using ‘select and confirm’ successfully, and 61% using ‘number navigation’ successfully (p<0.001, McNemar’s). We also expected that older users (65+) would perform better with the ‘number navigation’ style than with ‘select and confirm’. The results indicate that 28% of over 65s used ‘select and confirm’ successfully, and 33% used ‘number navigation’ successfully, but this was not statistically significant (p>0.05, McNemar’s), possibly due to only having 38 people in this age group. Further analysis and research is needed to better understand how different kinds of participants use different interface styles. 'Number navigation' menu task

80 60 40 20

age band Male

Female

75+

65-74

55-64

45-54

35-44

25-34

18-24

75+

65-74

55-64

45-54

35-44

25-34

0

18-24

% of age band correct

Select and confirm' menu task 100

age band Male

Female

Figure 6.4. Percentage of male and female participants succeeding in the menu selection tasks, by age category

As explained in Section 6.2.3, we expected that younger users would perform better with the ‘select and confirm’ interface than older users. This was supported by the results, with 72% of people under 65 using ‘select and confirm’ successfully, and 61% using ‘number navigation’ successfully (p<0.001, McNemar’s). We also expected that older users (65+) would perform better with the ‘number navigation’ style than with ‘select and confirm’. The results indicate that 28% of over 65s used ‘select and confirm’ successfully, and 33% used ‘number navigation’ successfully, but this was not statistically significant (p>0.05, McNemar’s), possibly due to only having 38 people in this age group. Further analysis and research is needed to better understand how different kinds of participants use different interface styles. In fact, the success rates for both menu selection styles were low overall, with only 61.8% of the sample using the ‘number navigation’ style successfully, and 73.4% succeeding with ‘select and confirm’. These rates are even lower for older age groups, with over 50% of participants aged over 75 failing on both tasks. It should be remembered that the partial paper prototyping method used is likely to give lower success rates than use in practice. Nevertheless, the success rates are sufficiently low to raise serious concerns about both menu selection styles. It seems that, although number navigation may perform slightly better with older people, it is not effective enough to work as an alternative. Other interface styles for selecting menu options are needed, and more work is required. Some possibilities are provided by touchscreen interfaces,

A Population Perspective on Mobile Phone Related Tasks

61

which allow menu options to be selected more directly. However, this technology may have other usability issues including those arising from the lack of tactile feedback.

6.3.2. Button Dexterity Results Overall, more people were successful with the larger buttons than with the fiveway keypad (p<0.001, McNemar’s), in agreement with our hypothesis. However, an experimental weakness emerged for this condition alone: participants only needed to press the last button correctly to be recorded as having completed the task correctly, thus potentially improving the success rate. For the fiveway keypad, participants needed to press all the buttons correctly to be recorded as completing the task successfully. Thus, it may be more useful to examine the results for each task separately. For the fiveway keypad (Figure 6.5a), there was no significant difference between men and women, but success on this task did decrease with age (p<0.001, Pearson’s chisquared), with 47% of those aged 75+ failing on this task. For the larger outer buttons (Figure 6.5b), there was no significant difference between men and women or between different age groups, with high levels of success across the sample. Participants also rated how easy they found each task (see Figure 6.6). Considering all ages and genders together, participants rated the larger buttons as easier to use than the small ones (p<0.001, Marginal Homogeneity). For the fiveway keypad, it seems that the ease of use ratings may decrease with age but increase again for the oldest age group, but further analysis and investigation is needed. Bigger buttons (success)

5 way keypad (success)

80 60 40

75+

65-74

55-64

45-54

age band

age band Male

35-44

75+

65-74

55-64

45-54

35-44

25-34

18-24

0

25-34

20

18-24

% of age band

100

Female

Male

Female

(a) (b) Figure 6.5. Percentage of male and female participants succeeding in the button pressing tasks, by age category

6.3.3 Cable Insertion Results Overall, people were more successful at inserting the mini USB cable than the headphone jack (p<0.001, McNemar’s). This contradicted our expectation that the headphone jack would perform better. Further examination of the data indicated that many people failed because they inserted the jack partially but not fully. It is likely that many of these partial insertions were because the required pushing force peaked at a

62

Bradley et al.

partial insertion point. Some participants may have believed that they had completed the task at that point and not thought to push the jack harder to insert it fully. This appeared to be more problematic for the female participants (p<0.001, Fisher’s exact), but it is unclear why. This may be due to muscle strength differences or a difference in prior experience such that participants did not know that the plug needed to be inserted further. Other age and gender differences are shown in Figure 6.7. Success on the headphone jack task decreased with age (p<0.01, Pearson’s chi-squared), as did success on the mini-USB task (p<0.05, Pearson’s chi-squared), although the latter may be affected by ceiling effects. There was no significant difference between genders for the mini-USB task, with almost all participants successfully completing this task. Bigger buttons (easy or v. easy)

5 way keypad (easy or v. easy)

80 60 40

75+

65-74

55-64

45-54

age band

age band Male

35-44

75+

65-74

55-64

45-54

35-44

25-34

18-24

0

25-34

20

18-24

% of age band

100

Female

Male

Female

Figure 6.6. Percentage of male and female participants rating the task as ‘very easy’ or ‘easy’ for the five way keypad (small) and outer large button tasks, by age category Mini USB (success)

Headphone jack (success)

% of age band

100 80 60 40 20

75+

65-74

55-64

45-54

35-44

25-34

age band

age band Male

18-24

75+

65-74

55-64

45-54

35-44

25-34

18-24

0

Female

Male

Female

Figure 6.7. Percentage of male and female participants succeeding in the cable insertion tasks, by age category

Participants also rated how easy they found each task (see Figure 6.8). Although almost all (over 99% of) participants completed the mini-USB task successfully, they considered it to be more difficult than the headphone jack task (p<0.001, Marginal Homogeneity). This agrees with our expectation that this task would be more difficult. It may be because participants needed to orient the plug with the socket in this task.

63

A Population Perspective on Mobile Phone Related Tasks

Mini USB (easy or v. easy)

Headphone jack (easy or v. easy) 100

60 40 20

75+

65-74

55-64

45-54

35-44

25-34

age band

age band Male

18-24

75+

65-74

55-64

45-54

35-44

25-34

0

18-24

% of age band

80

Female

Male

Female

Figure 6.8. Percentage of male and female participants rating the task as ‘very easy’ or ‘easy’ for the cable insertion tasks, by age category

6.4. Conclusions This study examined 362 participants’ performance in and ratings of ease of use for some component activities of mobile phone use. Across the activities, older users had markedly lower success rates and generally lower ratings for ease of use. In particular, the menu selection methods tested in the study had low success rates overall and particularly for older users. This raises concerns about the suitability of these common interface styles, and highlights the need for further research to develop new methods of selecting menu items on mobile devices. In addition, the button pressing tasks indicated that small buttons in the fiveway keypad were difficult for older users to press accurately. Yet buttons of this size and type are commonly used in mobile telephones. Somewhat, surprisingly, many participants also struggled with correctly inserting the headphone jack. Although most people did successfully insert the miniUSB cable, many people did not consider this task to be easy. Overall, these results indicate that current mobile phone designs are not meeting the needs and capabilities of older users. More detailed analysis and study are needed to illuminate some of the fundamental issues, and further work is needed to produce more inclusive mobile phone designs.

6.5 Acknowledgements The survey was conducted as part of the inclusive design research programme (i~design, 2011), funded by EPSRC. As well as the authors of this paper, several others made substantial contributions to the design and analysis of this survey, including Felicia Huppert, Kai Ruggeri, Eddy Elton, Jose Liht and John Ryan. Mike Bradley’s time was partly funded via the RCUK Digital Economy Programme’s BRIDGE project.

64

Bradley et al.

6.6 References Clarkson PJ, Keates S (2003) Digital television for all. A report on usability and accessible design. Appendix E - investigating the inclusivity of digital television set-top receivers. Department of Trade and Industry, UK Fitts PM (1954) The information capacity of the human motor system in controlling the amplitude of movement. Journal of Experimental Psychology, 47(6): 381-391 i~design (2011) Available at: www-edc.eng.cam.ac.uk/idesign3/ (Accessed 31 August 2011) Lindholm C, Keinonen T (2003) Mobile usability: How Nokia changed the face of the mobile phone. McGraw-Hill, New York, US Snyder C (2003) Paper prototyping: The fast and easy way to design and refine user interfaces. Morgan Kaufmann Publishers, San Francisco, US

Chapter 7 How to Use Virtual and Augmented Reality Techniques to Design Highly Usable Human-machine Interfaces S. Ceccacci, M. Germani and M. Mengoni

7.1 Introduction Modern technological products, working and living environments, are ever more rich in potential functionalities for end-users (e.g. communication support, facilitation of physical actions, etc.). The increasing functionalities make these products complex to develop. Research efforts are focused on the development of highly usable products and environments that should be easily and intuitively used by a wide range of people (children, elderly, disabled, etc.). User-centred design approaches have to be adopted. This is not easy to handle when complex environments populated by different products have to be conceived, analysed and evaluated. In this scenario, the aim of the present work is to study a user-centred design method in relation to the kitchen environment. It has been developed in the context of an Italian National Research Programme, called “e-kitchen: smart and highly usable kitchen”, started in April 2011. The project involves 16 important Italian and international companies (e.g. Indesit, Lube Industries, Telecom Italia, Faber-Franke), seven SMEs and five research centres, under the coordination of Università Politecnica delle Marche. The project aims to achieve a complete rethinking of the kitchen environment taking into account safety, comfort, ecosustainability and energy efficiency. One of the main goals is the improvement of its usability for aged and disabled people. In this context it is essential to manage and optimise the interaction between the user and the different devices so as to let the individual live in an environment offering well-being, safety and new intelligent functionalities. The interaction studied involves three elements: human, machines and environment. Multi-sensory Virtual Reality and Augmented Reality techniques are used to support the inclusive design of all target devices and of the kitchen environment as a whole. These technologies give the advantage of simulating user interaction from the first conceptual design stages and hence of evaluating different design

P. Langdon et al. (eds.), Designing Inclusive Systems, DOI: 10.1007/978-1-4471-2867-0_7, © Springer-Verlag London 2012

65

66

Ceccacci et al.

alternatives by minimising development cost and time. Moreover, their high usability allows the involvement of customers in the entire design process thus implementing a user-centred design approach. In this way it is possible to achieve customised products by directly analysing different design solutions with end-users and finding which one meets their needs better. AR/VR technologies allow the assessment of both physical and cognitive ergonomics. Specific protocol analysis methods are under development in order to examine the implicit and explicit user behaviour.

7.2 Which Supporting Prototyping Technique Fits the Requirements of Designing Interactive Devices? In order to design highly usable interactive systems, it is crucial to understand the needs and limitations of end users. This can be achieved by adopting a user-centred perspective at each stage of the design process. According to ISO 13407 (1999) standard, a proper User-Centred Design (UCD) process is structured in four iterative phases. A critical one involves the building of interactive prototypes to support rapid product usability evaluation (Mengoni and Peruzzini, 2010). Two main prototyping techniques, which differ in the level of fidelity and interaction they provide, can be classified. Low-fidelity prototypes (e.g. paper sketches, cardboard mock-up) are good to test aspects such as the layout of controls. However, they do not allow the evaluation of the effects of tactile, auditory and visual feedback (Hall, 2001). High-fidelity prototypes (e.g. softwarebased, physical mock-up) allow users to realistically appraise product aesthetic attributes and functionalities (Sonderegger and Sauer, 2010), but they are expensive and cannot be realised during the first conceptual design stage when the product design has not yet been completed. To reduce the gap between low and high fidelity prototypes, Virtual Reality (VR)-based technologies have been introduced to create, manipulate and explore virtual prototypes as well as to simulate product behaviour in different working conditions. They aim to replace physical mock-ups with virtual ones (Wilson and D’Cruz, 2006). Some studies demonstrate how they can be used to rapidly carry out usability testing, to reduce evaluation time and costs and to involve end-users from the earliest stages of the design process (Kuutti et al., 2001). However, VR environments often show multiple technological limitations, such as low sense of immersion, poor physical interaction, high complexity, low realism, unnatural manipulation, intrusiveness and non-intuitiveness. Mixed Reality (MR) environments represent a compromise solution in which real and virtual worlds are combined in various proportions and presented as a unified whole. MR fuses the two extremes of the Virtuality-Reality continuum by synchronising information from the digital space and the physical one in a more natural way (Milgram et al., 1994). Within the MR framework the Augmented Reality (AR) technique is one of the most adopted due to the low cost of the

How to Use Virtual and Augmented Reality Techniques

67

technologies and to its ability to enhance the real scene with computer graphics and emerging tactile and auditory displays (Zhou et al., 2004). Many different solutions have been proposed in literature with the intent of providing devices to interact with the AR environment in a more intuitive way. Some researchers integrate handheld haptic technologies to reproduce the real contact with objects during the exploration of virtual space (Bordegoni et al., 2009). Others try to eliminate the gap between the interaction with a natural environment and the interaction with a computer system by adopting Tangible Augmented Reality (TAR) techniques (Park et al., 2008). A lot of studies have been carried out in order to analyse how VR interfaces are able to support elderly people to improve rehabilitation involvement (Flores et al., 2008; Alankus et al., 2010). These studies can be useful for the development of products and services tailored to elderly people with cognitive issues, in order to support them in their daily lives (Pittarello and De Faveri, 2006; Kim and Dey, 2009). However, no studies have been conducted yet to assess what the optimum VR technology is for the involvement of elderly people in design solutions evaluation. Most of the projects which aim to address the relationship between older people and technology, including PERSONA (2007) and UTOPIA (2011), have been limited in their development of a new methodological approach only to design for elderly people, without taking into consideration design-oriented technologies issues. However, this aspect is becoming highly important to really implement a UCD approach whilst reducing development costs. This is demonstrated by two integrated projects of the 7th Framework Programme, which are VAALID (2011) and VERITAS (2011).

7.3 The Proposed UCD Approach Based on VR/AR Techniques User involvement is particularly important in inclusive design because designers have very different awareness of elderly and disabled users’ needs. However, it can be a complicated task. The general proposed design process described below (Figure 7.1) is based on a proper user-centred methodology where user characteristics are the main drivers of the design solution identification. The information which is required to feed into the design process is grouped into two areas. The first includes all the information about end-user characteristics and abilities (characteristics of elderly people) and environmental and social factors which can affect their activities in their daily life. The information collected into the second area is related to the characteristics of suitable technologies (e.g. cost, flexibility, extendibility, upgradeability, etc.) and to the respective potentially valuable applications (e.g. safety monitoring, health and assistive application, communication, etc.). User Information is collected and analysed in a proper User Analysis process, in order to extract user needs and preferences. Preference data can be gathered using interviews, workshops, surveys, site visits, artefact analysis, focus groups, observational studies, and contextual inquiry. The analysis of user needs should

68

Ceccacci et al.

take into consideration the fact that elderly people present a very diverse range of abilities. This diversity of abilities exists not only within groups, but also among individuals. Therefore, it is impossible to draw up a simple profile, or to identify a single stereotypical user. In this context, the International Classification of Functioning, Disability and Health (ICF) is a valuable tool for understanding the abilities of a user, depending on his/her characteristics.

Functional Modelling

Figure 7.1. The proposed UCD approach for inclusive design

The adoption of a Functional Modelling approach allows the system/environment functions and flows to be structured. In this way it is possible to represent the design context in terms of actions to be supported and mutually correlated. It is fundamental to analyse how a particular user can act in an environment according to their capabilities. For example in a typical kitchen it is possible to perform a particular set of macro-tasks, such as preparing meals, eating, drinking, doing the housework, etc. Each macro-task consists of a number of actions which require specific physical and mental skills to be performed. Once the functional model of the environment has been built, it is possible to deduce how the lack of special skills impacts on the performance of each task, and hence to define the characteristics of a clear set of appropriate facilitators to be provided for a barrier-free kitchen. Prototyping activity is fundamental to communicate and assess the design intent. As mentioned above, the creation of high fidelity prototypes represents a cost that increases if alternative design solutions need to be evaluated. Moreover, at the beginning of the design stages it is very difficult to build a prototype which is similar to the final product in terms of implemented functions, adopted materials, aesthetic features, interaction modes, etc. because all design choices have not been made yet. In parallel, the involvement of end-users during conceptual design is fundamental to address subsequent decision-making. As a consequence, thanks to the advantages of novel AR/VR technologies, high-fidelity prototypes can be replaced with virtual ones. The problem lies in choosing the technology according to particular user characteristics. In fact, the experienced quality of a VR environment, in terms of interaction, immersion and navigation, generally depends on the human mental and physical ability.

How to Use Virtual and Augmented Reality Techniques

69

The evaluation of product usability is fundamental for the creation of userappropriate products. It allows designers not only to pinpoint design errors, but also to measure the quality perceived by the user and the resulting product experience, so that the design changes can be better finalised. This is carried out through usability tests, which support the collection of detailed information on the product experience and on the way consumers use it. To fully evaluate product usability it is necessary to develop an experimental protocol, which allows the correlation of the user response with the specific product features by considering emotional, affective and cognitive aspects of the user-product interaction.

7.4 The e-Kitchen Case Study and Experimental Results A preliminary study was conducted to assess how much TAR technology can be useful in involving the elderly in the early usability evaluation of the centralised control panel of the e-kitchen model (called smart kitchen user interface). The structure of the e-kitchen system is represented in Figure 7.2. The control panel manages all household appliances, receives suggestions and feedback from them, monitors consumption, etc. TAR usability assessment is carried out through the following steps: 1. definition of an experimental protocol to assess the usability of different prototyping environments; 2. construction of a traditional high-fidelity prototype and development of TAR prototype operating as the final system and enabling the user to interact with the physical elements to trigger an emotional and affective response; 3. application of the experimental protocol in order to compare user performance in two different environments (real and virtual). The proposed experimental protocol aims to measure efficiency, effectiveness and satisfaction in use provided by logical and physical elements of the interface, as prescribed in ISO 9241-11 (1998). The protocol is general enough to be able to be applied in different interactive devices and different virtual reality environments. Efficiency was assessed as the completion time required to complete a task. Effectiveness was evaluated by error occurrence, meaning the number of mistakes during task completion such as incorrect action sequence, repeated actions or commands and wrong actions (ISO/IEC TR 9126-4, 2004). This makes it clear if the product interface correctly guides users toward task completion. Finally, some metrics have been defined to evaluate satisfaction in use: pleasure in use, absence of monotony, accessibility and ease of use (UNI EN ISO 10075-2, 2002; KuijtEvers et al., 2004; Park and Han, 2010). Pleasure in use measures perceived pleasure and satisfaction during use. It can be defined as a mental state depending on the quality of user-product interaction. Absence of monotony measures how well the system avoids excessive mental workload and repetitive actions.

70

Ceccacci et al.

Accessibility measures intelligibility and adequacy of the information provided, and so it is closely connected with the understanding of product behaviours, functionalities and modes of use. Ease of use measures the simplicity of using the product interface and the perceived sense of friendliness.

Figure 7.2. The e-Kitchen model

The adopted TAR prototype is characterised by a tangible interaction in an AR environment and by a functional simulation of the GUI behaviour. The tangible interaction is achieved by using a physical prototype realised by low cost Rapid Prototyping techniques. A real scale projection of the virtual prototype is overlapped with the physical interface to simulate the aesthetic appearance (i.e. surface finishing, shape) and the visualisation display that changes according to the user options. Reliable real time interaction is guaranteed by an optical tracking system (Optitrack by NaturalPoint) which tracks the user’s hand, the user’s point of view and the product prototype position in the same space. The modelling of the GUI functional states and the management of state changing in real time is achieved by 3DVIA Virtools (by Dassault Systems). A portable clip-USB camera is mounted on see-through glasses (iWear VR920 by Vuzix) to record the scene. Such a set-up allows users to handle the product prototype and simultaneously see the virtual interface projected on the physical object. The high fidelity prototype (HF) is produced by building all the electronic and mechanical components of the device. It is made of a moulded square containing a case (PMMA for the front side and TPU for the back), a lion battery, a Bluetooth module, an electronic board, a LCD 3.5” capacitive APR touch screen, eight piezoelectric actuators, and a polycarbonate closing ring. The SW system is developed in C++ on Linux adopting Nokia QT graphical libraries. This set-up provides a powerful product representation since the highest level of interaction and fidelity is achieved. However, it entails high costs and a long time to produce the whole prototype.

How to Use Virtual and Augmented Reality Techniques

71

Figure 7.3. On the left the TAR prototype (A) the RP model with AR and Optitrack IR marker and (B) the augmented image of interface. On the right the HF prototype.

Involved end-user composition was defined by adopting a subjective sampling method based on the field studies previously conducted by industrial partners on age and gender of the target market. A total of 16 users were involved in the testing sessions, eight users for each experimental set-up: this number is sufficient for a qualitative assessment, because it permits the detection of more than 80% of all usability problems in a product, as supported by different studies (Virzi, 1992; Nielsen and Landauer, 1993; Nielsen, 2000). The groups were matched in gender (for each group 75% were women and 25% were men), age distribution (75% 6065 years old, 25% were 66-70) and technological expertise with respect to the different set-ups. Tests were carried out in parallel groups, as participants were statistically matched. Each group had to evaluate a product prototype produced according to the respective test set-up methods. To evaluate metrics of satisfaction, users were asked to express their personal judgment for each protocol metric (5-point scale). To evaluate efficiency and effectiveness, users were asked to perform five different tasks on the product interface which were defined so as to evaluate the quality of the interaction with every product function according to the above metrics of measurement. The joint use of VIA and interviews assured an accurate analysis of user response. A One-Way ANOVA (alpha level of 0.1) has been calculated in order to compare experimental data (task completion time, number of errors and satisfaction in use) in order to analyse differences between TAR and HF set-ups. On analysing the results, the adopted protocol proved to be appropriate to evaluate the GUI and the TAR prototype has shown it can be useful to investigate both cognitive and physical ergonomics due to its dual nature (i.e. virtual interactive graphic interface and physical interaction with the product body). In particular it is possible to infer that the technological set-up does not affect task errors and satisfaction in use. However, time performance is significantly different (Table 7.1). Results point out the effectiveness of adopting a tangible augmented prototype to create interactive prototypes for the usability testing. To prove the effectiveness of TAR techniques to carry out usability assessment in place of high fidelity prototypes an additional evaluation is imperative. It consists of analysing the effort required to virtually prototype alternative solutions and to set the proper experimental environment (Table 7.2). Table 7.2 provides a time estimation for building the two tested prototypes. The assessment is carried

72

Ceccacci et al.

out considering five main activities in prototyping as listed in Table 7.2. A total time comparison shows that TAR sensibly reduces the time for prototype development and testing. While previous results show the reliability of usability assessment on virtual prototypes instead of high fidelity ones, Table 7.2 demonstrates that designers do not require extra work to build them. Both achievements are fundamental to increase the use of these technologies in industry. Table 7.1. One-way ANOVA results for satisfaction, task completion time and errors

However, although this result is crucial to demonstrate the importance of a UCD approach based on TAR for industrial applications, it is not enough to show that it is suited to test usability for elderly and disabled people. Although aged people were chosen to carry out the testing sessions, their response in terms of usability of the adopted technologies has not been structurally assessed. The adopted protocol actually aims to evaluate the usability of the control panel independently from the technique used to realise it. A proper post-hoc questionnaire should be designed to focus on the user experience of the deployed prototypes. A preliminary analysis of users’ observations outlines that: 1. they are quite at ease with the AR glasses, which are similar to those they wear in everyday lives, but they show problems in vision if they need their own glasses to focus on the displayed images; 2. they have problems in selecting the virtual icon by positioning their finger in the right place to allow the panel to work properly. This problem is particularly evident in the presence of slight hand tremors; 3. they have some difficulties in identifying the right distance to position the physical prototype in respect to their viewpoint filtered by the AR glasses.

73

How to Use Virtual and Augmented Reality Techniques

Table 7.2. Design process performance resulting from the different prototyping techniques (evaluation in time) ACTIVITIES 1. Modelling of the design alternatives CAD modelling GUI modelling

2. Prototyping File import/export Setting (machine, rendering scene, etc.) Prototype generation Finishing

3. Experimental design 4. Pre-testing (users’ training, etc.) 5. Testing TOTAL TIME

TAR prototype 4 days

High-fidelity Prototype 10 days

2 days 2 days

4 days 6 days

5 days

20 days

1 day 3 days 1 days -

1 day 18 days 1 day

0.5 hour 1 hour 3 hours 9 days 4.5 hours

2.5 hours 30 days 2.5 hours

7.5 Conclusions The use of VR/AR technologies can successfully support the design and evaluation of Ambient Assisted Living environments. They can facilitate the assessment of user interfaces from both a cognitive and a physical point of view. In this context an Inclusive design approach has been defined in the scenario of a funded National project called e-kitchen. A part of this approach addressed to the VR/AR technology choice to virtually test different software user interfaces. In this case the technology assessment has to be properly correlated to a specific class of users (aged people). Experimentation highlights that Tangible AR is the best system to allow subjects to experience the interaction. Future work will be focused on a deep analysis of the usability of the deployed TAR technologies with specific attention elderly people and people with disabilities.

7.6 References VAALID (2011) Accessibility and usability validation framework for AAL interaction design process. Available at: http://www.vaalid-project.org/ (Accessed 4 November 2011) Alankus G, Lazar A, May M, Kelleher C (2010) Towards customizable games for stroke rehabilitation. In: Proceedings of the 28th ACM Conference on Human Factors in Computing Systems (CHI 2010), Atlanta, GA, US Bordegoni M, Cugini U, Caruso G, Polistina S (2009) Mixed prototyping for product assessment: A reference framework. International Journal on Interactive Design and Manufacturing, 3(3): 177-187 Flores E, Tobon G, Cavallaro E, Cavallaro FI, Perry JC, Keller T (2008) Improving patient motivation in game development for motor deficit rehabilitation. In: Proceedings of the International Conference on Advances in Computer Entertainment Technology (ACE 2008), Yokohama, Japan Hall R (2001) Prototyping for usability of new technology. International Journal of HumanComputer Studies, 55(4): 485-501

74

Ceccacci et al.

ISO 13407 (1999) Human-centred design processes for interactive systems. International Organization for Standardization. Available at: http://www.iso.org/iso/catalogue_detail. htm?csnumber=21197 (Accessed 4 November 2011) ISO 9241-11 (1998) Ergonomic requirements for office work with visual display terminals (VDTs) - Part 11: Guidance on usability. International Organization for Standardization. Available at: http://www.iso.org/iso/iso_catalogue/catalogue_tc/catalogue_detail.htm?csnum ber=16883 (Accessed 4 November 2011) ISO/IEC TR 9126-4 (2004) Software engineering - product quality- Part 4: Quality in use metrics. Available at: http://www.iso.org/iso/iso_catalogue/catalogue_tc/catalogue_detail.htm?csnumber =39752 (Accessed 4 November 2011) Kim S, Dey AK (2009) Simulated augmented reality windshield display as a cognitive mapping aid for elder driver navigation. In: Proceedings of the 27th ACM Conference on Human Factors in Computing Systems (CHI 2009), Boston, MA, US Kuijt-Evers LFM, Groenesteijn L, De Looze MP, Vink P (2004) Identifying factors of comfort in using hand tools. Applied Ergonomics, 35(5): 453-458 Kuutti K, Battarbee K, Sade S, Mattelmaki T, Keinonen T, Teirikko T et al. (2001) Virtual prototypes in usability testing. In: Proceedings of the 34th Hawaii International Conference on System Sciences (HICSS 2001), Maui, Hawaii, US Mengoni M, Peruzzini, M, (2010) Usability assessment method to improve interaction design: How to get it. In: Fisher X, Coutellier D (eds.) Research in Interactive Design, Springer-Verlag, Paris, France Milgram P, Takemura H, Utsumi A, Kishino F (1994) Augmented reality: A class of displays on the reality-virtuality continuum. In: Proceedings of SPIE, Telemanipulator and Telepresence Technologies. Nielsen J (2000) Why you only need to test with 5 users. Alertbox. Available at: http://www.useit.com/alertbox/20000319.html (Accessed 4 November 2011) Nielsen J, Landauer TK (1993) A mathematical model of the finding of usability problems. In: Proceedings of the ACM INTERCHI’93 Conference, Amsterdam, The Netherlands Park H, Moon HC, Lee JY (2008) Tangible augmented prototyping of digital handheld products. Computers in Industry, 60(2): 114-125 Park YS, Han HS (2010) Touch key design for one-handed thumb interaction with a mobile phone: Effects of touch key size and touch key location. International Journal of Industrial Ergonomics, 40(1): 68-76 PERSONA (2007) Perceptive spaces promoting independent aging. Available at: http://www.aalpersona.org/index.html (Accessed 4 November 2011) Pittarello F, De Faveri A (2006) Improving access of elderly people to real environments: a semantic based approach. In: Proceedings of the Working Conference on Advanced Visual Interfaces (AVI 2006), Venezia, Italy Sonderegger A, Sauer J (2010) The influence of design aesthetics in usability testing: Effects on user performance and perceived usability. Applied Ergonomics, 41(3): 403-410 UNI EN ISO 10075-2 (2002) Ergonomic principles related to mental workload - design principles. Italian Organization for Standardization. Available at: http://www.uni.com/en/ (Accessed 4 November 2011) UTOPIA (2011) Usable technologies for older people: Inclusive and appropriate. Available at: http://www.dcs.gla.ac.uk/utopia/default.html (Accessed 4 November 2011) VERITAS (2011) Virtual and augmented environments and realistic user interactions to achieve embedded accessibility designs. Available at: http://veritas-project.eu/ (Accessed 4 November 2011) Virzi RA (1992) Refining the test phase of usability evaluation: how many subjects is enough? Human Factors, 34(4): 457-468 Wilson JR, D’Cruz M (2006) Virtual and interactive environments for work of the future. International Journal of Human-Computer Studies, 64(3): 158-169 Zhou Z, Cheok AD, Yang X, Qiu Y (2004) An experimental study on the role of 3D sound in augmented reality environment. Interacting with Computers, 16(6): 1043-1068

Chapter 8 Development and Evaluation of Sonified Weather Maps for Blind Users R. Weir, B. Sizemore, H. Henderson, S. Chakraborty and J. Lazar

8.1 Introduction In recent years there have been significant advances in developing websites that are accessible for individuals with disabilities. In particular, the Web Content Accessibility Guidelines (WCAG) from the Web Accessibility Initiative (http://www.w3.org/wai) provide clear standards for developing accessible web based content, and have been the foundation for legal guidelines developed in many countries (Meiselwitz et al., 2010). However, a challenging aspect of accessible web design is developing equivalents of data visualisations for blind users, since data visualisations are often used to allow quicker comprehension of large and complex data sets (Fritz and Barner, 1999). In this research we investigate the potential use of sonification for presenting weather data to blind users. At this juncture we would like to emphasise that that the term “blind users” has different meanings depending on context and country. For instance, in the US, “blind users” refers to anyone with any type of visual loss, whereas in the UK, “visually impaired” is often used to describe people with low vision, and “blind” is used to describe someone with no useful residual vision. Maps represent a very common visualisation approach on web pages. Figure 8.1 (adapted from www.recovery.gov) shows a typical example of map based data visualisation. This map of demographic data allows the user to quickly reach a conclusion that southwestern states in the USA have a higher percentage of residents under 18 through an interpretation of the colour coding in the map. Equivalents of visualisations for blind users tend to be simple tabular representations of the same data (Plaisant, 2004). While this is technically equivalent, it does not provide an “overview of data” that is the key first step of usage of information visualisations.

P. Langdon et al. (eds.), Designing Inclusive Systems, DOI: 10.1007/978-1-4471-2867-0_8, © Springer-Verlag London 2012

75

Weir et al.

76

Figure 8.1. Map based visualisation of population data

Previous research has addressed this problem through the development of sonified maps, which used non-textual audio output to allow users to comprehend detailed as well as trend data rendered within maps (Zhao et al., 2008; Walker and Mauney, 2010). In this investigation, we build on existing research to investigate the potential use of sonification to represent weather map data for blind users. The rest of the article is organised as follows. First we describe existing research and applications related to sonified maps. Next we describe our methodological approach in developing and evaluating a sonified solution for providing accessible weather map based data for blind users. Finally, we discuss the implications and future research directions.

8.2 Literature Review An important implication of developing websites following accessibility guidelines such as WCAG is that they will work for most users with perceptual and/or motor impairments, including blind users. However, information visualisations on web pages continue to be inaccessible. Visualisations are inherently directed at sighted users and developing an equivalent rendering for blind users is not a trivial problem. One option is to use tactile printed media (such as a raised bar chart printout). However research (e.g. Wall and Brewster, 2006) suggests that these have limitations both in terms of flexibility of option and extent of usage. Effective visualisations need to incorporate a capability to drill-down to specific data values in addition to providing an overview of data (Shneiderman and Plaisant, 2010). Previous research has examined the potential use of sonification of maps for blind users, utilising non-textual sound (Zhao et al., 2008; Pauletto and Hunt, 2009). Two existing sonification tools are iSonic (http://www.cs.umd.edu/hcil/audiomap/) and Earth+ (http://prime.jsc.nasa.gov/earthplus/). The iSonic tool was originally created as a graduate project at the University of Maryland and allows blind users to hear population trends and patterns on a map of the United States. At a very basic level the application sound pitch to provide the user overview or trend level information about population within a geographical region. iSonic also provides alternative

Development and Evaluation of Sonified Weather Maps for Blind Users

77

map views of data that include a broader regional view and also a view down to the state level and associated counties. Earth+ (developed by NASA in 2005) is another tool that was evaluated. Earth+ provides accessible map interpretation for blind users. Unfortunately, the Earth + application is limited in its functionality. The application allows a user to load an image and provides a piano note with a pitch unique to the colour of the image based on the position of the cursor. Therefore, we chose to focus our work on using iSonic, and received permission from the University of Maryland to continue building onto the iSonic application.

8.3 Requirements Gathering To modify an existing tool for blind users to access new types of data content (weather data), we needed to understand how blind users access weather information. We were therefore motivated in our requirements gathering approach toward answering questions about the different kinds of weather information considered important by blind users and how it needed to be rendered, resources available to obtain the requisite weather related data feeds, and the most appropriate interface technology that could be used. Our data collection was greatly facilitated by our access to a diverse set of users within the blindness community. There were four interviews and six survey responses. Our four interviews took place with individuals who had 1) expertise about the domain of interest (meteorology), 2) a high level of motivation in helping develop and use sonified weather maps, 3) expertise on contextually appropriate assistive technology, and also with one individual who was a meteorologist but was not blind. In addition, there were also 6 responses to a survey, from typical users who had interest in gaining weather related information but moderate to low technological and domain related expertise. Space constraints prevent us from providing specific details about the interviews and the survey. Users were either interviewed or surveyed based on their availability, as is typical for humancomputer interaction research involving people with disabilities (Lazar et al., 2010). We employed interviews when face time was possible, and surveys when face time was not an option. The interviews with the expert blind users and the survey responses of the nonexpert blind users provided valuable input for the development of the prototype weather map. The data from the above user interaction provided requirements for the application related to the nature of weather related information expected by blind users, the most useful representations of such information, as well as modes of interaction that would be considered helpful. In addition to the interactions with blind users, the data collection also involved an interview with a meteorologist from the National Oceanic and Atmospheric Administration (NOAA) who is not blind. This interview provided valuable directions with regards to obtaining a continuous live weather data feed for the application from NOAA web sites. After the collection and analysis of interview and survey data, a series of user scenarios were created to obtain a clearer understanding of what the proposed application could be expected to do. These scenarios were created to explicitly map

78

Weir et al.

out how a user would go through the program for related tasks, what these tasks would show, and what types of outputs they would have. Special care was taken to explicitly include features and capabilities identified in the user interviews and surveys while developing the user scenarios. The scenarios proved to be helpful in narrowing down and creating a focus for what the application should do and how. Based on the requirements gathering and the user scenarios, it was possible to identify an initial set of specifications for the sonified weather map application. These specifications are described as follows: • provide weather information related to a) current and future temperature b) precipitation and c) wind speed; • allow the users to obtain discrete weather related information for a geographical location (e.g. temperature in a particular city); • allow the users to obtain trends with regards to weather related information for a geographical region (e.g. change of temperature within a state); • provide the users with a sense of the physical geographical reference within the context of the map; • allow the user to choose a) the nature of weather information and b) the level of detail for the weather information; • provide the user with means for multimodal interaction to enable perception of information through multiple senses; • be easy to use for users without access to sophisticated assistive technology; • be easy to use for users with minimal experience with assistive technology. The above list of specifications provided the basis for the design of the sonified weather map application. In addition a decision was taken to develop the initial prototype only for the state of Maryland and the 24 counties within Maryland (Baltimore City is not in a separate county, and demographically is usually counted as the 24thcounty of Maryland). The specifics of the design are described in the next section.

8.4 Design of Sonified Weather Map 8.4.1 System Architecture for Retrieving Weather Data The accessible weather map was built by modifying the iSonic application. The iSonic application in its current version works with static population data and requires the source data to be in a CSV file format. However, weather data is more dynamic, changing often at an hourly rate. Design modifications were therefore required to accommodate the dynamic data. The source for the weather data was the National Oceanic and Atmospheric Administration (NOAA) website. This website provides real-time weather data formatted for easy retrieval. The data querying was developed using Microsoft Excel macros that allow retrieval of external data from tables located on HTML

79

Development and Evaluation of Sonified Weather Maps for Blind Users

pages. The weather data for the different counties within Maryland was retrieved through the manipulation of the longitude and latitude values in the data retrieval queries. This allowed the importing of weather data for multiple points in Maryland into the excel sheet. The particular data points were developed through the compilation of a list of cities that were completely within the boundaries of each county, usually the county seat. The assumption was that these cities would be representative of the weather within the county in most instances, since most of these counties are not geographically large. A second Excel sheet was created and formatted appropriately to make it readable by the iSonic application. The cells of the two excel sheets were linked so that the second excel sheet could automatically pull data from the first excel sheet containing imported weather data from NOAA’s web page. A small java program was developed that that allowed automatic storing of the data from the second excel sheet as a CSV file. In addition VBA macros were created for both excel sheets that initiate a refresh, save, close window sequence. Finally a batch file was created so that, the weather data is automatically updated and converted into a file readable by the application. A system architecture schematic for the data retrieval process is shown in Figure 8.2. Supplies NOAA Web Server

Supplies Excel Sheet

Master Data

Converts to Sonic Map Interface

Is read by CSV File

Figure 8.2. System architecture schematic

8.4.2 User Interaction with the iSonic Weather Map The prototype iSonic weather map application was developed by overlaying weather data on the basic iSonic Application. The application design was therefore constrained to use the existing mapping between data and auditory feedback in this platform. An associated risk of this design decision is the possibility that the mapping would not effectively translate to a different domain (weather data). While this represented a design dilemma, a decision was taken to accept this as a constraint for this exploratory research, and also to use this as an opportunity to test the usefulness of the existing mapping. The prototype weather map was developed to provide the user with three different types of weather information – temperature, wind speed and percentage chance of precipitation (see Figure 8.3). These types of weather related information are presented to the user separately to reduce information clutter. The user has the ability to switch between three sonified maps.

80

Weir et al.

The primary interaction mechanism in the iSonic interface is keyboard-based. For the study, an additional interaction strategy was adopted, involving a touch screen with a tactile map overlay to traverse through the iSonic weather maps. A KeyTec Magictouch touch screen was used to evaluate the potential for touchscreen based interaction with the application. The initial plan was to use a tactile map of Maryland obtained using the Braille/tactile printing facilities at the National Federation of the Blind. However, this tactile map did not calibrate well to the scale expected by the touch screen. A work around was achieved, by tracing the map of Maryland from the iSonic interface and then creating tactile contours of the map boundaries using needles, on a sheet of paper that was thick enough not to rip, but thin enough that the touchscreen could still sense human touch. This tactile map overlaid on the touch screen calibrated well to the iSonic application. The iSonic interface provides two information presentation choices (or data views) to the user. The first is the default map-based representation. The second is a tabular data view where the first column represents the geographical elements of the data (counties of Maryland in this instance) and the other columns represent the specific data domains being presented (e.g. temperature, wind speed and precipitation, in this case). The user can switch between the two data views using the TAB key. The tabular data view allows the user to toggle between the three different weather maps (for temperature, wind speed, and precipitation). iSonic provides the user with two different choices of navigation through the map - absolute and relative. The absolute navigation allows the user to sweep through (from left to right and up to down) and provides an overview of map values. The interface provides a percussion sound at the end of each row and bell sound at the end of the sweep. For relative navigation the user can press the four arrow keys to navigate up, down, left and right in the map commencing from the left top corner of the map. A chirping sound alerts the user when the navigation takes him/her outside the boundaries of the map. The map is always traversed in units defined by the geographical boundaries of the Maryland counties. As the user traverses through the state, they are given auditory feedback about the weather related data for that county. The user has the choice of receiving this feedback with different degrees of detail or information level. The lowest level information is provided by a tonal sound. The pitch of this sound is used to represent the value of the data, a higher pitch being associated with higher value of data. The higher level of information augments this tonal sound with spoken feedback, such as the name of the county and the temperature in the county. For this study the users were also provided with the means of navigating using the touch screen and the tactile map overlay. With this alternative, the authors were able to select a particular county by tracing its contours in the tactile map and then tapping on it to get weather related auditory feedback for that county.

Development and Evaluation of Sonified Weather Maps for Blind Users

81

Figure 8.3. Prototype map interface

8.5 Usability Evaluation An important aspect of developing any application for users with disabilities is to do a usability evaluation. Furthermore, the previous research (Zhao et al., 2008) on iSonic only evaluated the software using the keyboard, so we felt that evaluating using the touchscreen functionality would provide useful feedback for both developers and researchers. We also wanted to learn more about the effectiveness of just a touchscreen as compared to a touchscreen with tactile overlay. As this research project had no funding, and the volunteer participants did not have much time available, it was not possible to conduct the usability testing in stages over months and determine the effect over time of getting experience with the application. So this was a shorter, more formative usability evaluation, rather than a summative evaluation, which also focused on qualitative, rather than quantitative evaluation. There were a large list of potential tasks, such as naming the five counties with the lowest temperature, comparing the temperatures in two counties that bordered a third county, assessing trends in temperature (e.g. “does the temperature rise or fall as you go from western Maryland to central Maryland?”), where in the state are the windiest counties located, and what portion of the state has the highest precipitation. Due to the large number of potential tasks and the limited time for usability evaluation, the participants were mostly asked to perform different tasks. We planned for testing to take place with five blind individuals, at the International Braille and Technology Centre in Baltimore, Maryland. These individuals were recruited through the National Federation of the Blind, and had specifically expressed interest in weather maps, sonification, or science education. Two of the participants in the usability evaluation were the same people who took part in the interviews discussed earlier (the Blind user with expertise in meteorology, and the Blind user interested in developing accessible weather maps). Our 5 blind individuals had an average age of 45.6 (range 28-68), and there were 4 males, 1 female, all of whom were screen reader users who are unable to use screen magnification. The participants had used computers an average of 30.8 years, and used screen reading software an average of 22 years. Clearly, the participants had a high level of computing experience: however, it is expected that any potential users of sonification software for weather maps would have a high

82

Weir et al.

level of experience. Of the five participants, only one had previously used any type of sonification software (for computer gaming). As the software was being set up before the usability evaluation, there turned out to be technical problems. Due to some security patches recently installed at the International Braille and Technology Centre computers, the sonification tones would not work on any of the computers there. The researchers made a last minute decision to use the Macbook laptop (booted in Windows mode) that they brought with them to take notes, to have the participants evaluate instead the keyboard version (including the sonification tones), and then used a PC in the International Braille and Technology Centre (without the sonification tones) to evaluate the effectiveness of the touch screen and the touch screen with tactile overlay. This was not ideal, but since the usability evaluation was formative and qualitative, it allowed for a lot of feedback from users and discussion. The researchers explained to users how the software worked and demonstrated it using keyboard and touchscreen, gave the participants a few minutes to play around with it, and then asked the participants to attempt some tasks. In general, all five participants liked the application, and were able to figure out how to use it, to successfully complete a few tasks, within a few minutes of first using the application. User 1 spent several minutes trying to get a feel for the program, and initially said that they liked the keyboard application but they wanted to be able to feel the edges of a tactile map. When using the touchscreen, the user perceived the touchscreen as being a little bit jumpy and stated that they didn’t like a touchscreen without a tactile overlay, preferably on the touchscreen, as compared to the keyboard. User 1 also noted that they listened more to the speech than to the tones. One interesting challenge is that user 1 assumed that the touchscreen was a multitouch screen, which it was not. User 2 did not like using the keyboard to navigate around the Maryland state map, but quickly got comfortable with the application and noted points like “central Maryland is definitely hotter” and “there’s a weather front somewhere here.” User 2 said that they didn’t like the tones, because “I’m not musically inclined, so I like numbers, not sounds, it’s my learning style.” User 2 also expressed a strong preference for the tactile overlay on the touchscreen. User 2 also felt there could be more clarity in the feedback provided for navigating off the map and suggested, “You have entered Virginia” or something similar. User 2 felt that the irregular shapes of the counties made it difficult to navigate on the keyboard but suggested that the software would be good for geography lessons. User 3 also stated that they listened more to the speech data than the tones, but that they could understand there was a difference in the tones, and they wondered if headsets would be more useful to perceive stereophonic sounds. When using the touchscreen, the user stated that they preferred this method, as it enabled nonsequential navigation. User 3 didn’t seem to have a preference as to the touchscreen with or without the tactile overlay. User 4 had a good sense of the different tones and their correlations with the data and understood the trends. Unlike the other users, user 4 seemed to find the tones to be very useful. Like user 2, they noted that this application would really be useful for learning the geography of a new area. User 4 wondered why, when you cross the Chesapeake Bay, a large body of water, it didn’t make a “splash” sound,

Development and Evaluation of Sonified Weather Maps for Blind Users

83

instead of a “chirp” sound (which is the current sound for crossing a body of water). Using the tones, user 4 could immediately determine that the west side of Maryland had the highest chance of precipitation, and wondered if the researchers could add “elevation” to the software application, to help them learn more about the geography of Maryland. When user 4 started using the touchscreen, the application crashed, and while it rebooted, user 4 had to leave for a work-related appointment. Therefore, user 4 was the only one who was not able to evaluate the touchscreen interaction with the application. User 5 liked being able to hear the trends using the sonified tones, and immediately picked up important trends, for instance, the chance of rain was higher in the northern and western parts of the state. User 5 really enjoyed using the application, and wondered how much data you could present to a user before they became overwhelmed: over-time comparisons (e.g. checking the data at 7AM and again at 11AM) might be most useful. User 5 was equally enthusiastic about the tactile map over the touchscreen, and was able to easily complete tasks using both approaches, commenting “now it starts to mean something, because now I’m touching it on the map.” User 5 further noted “Now, I get the information that I don’t normally get. This is a very different sense than I get from using the Braille note. This is exactly what I have been looking for!” User 5 further noted “I’ve always had to calculate the weather trends in my mind, until today!” In summary, while the usability evaluation was formative and qualitative, there were some clear trends. The users clearly preferred the tactile map over the touchscreen, as compared to either the touchscreen alone, or keyboard alone. While some users found the sonification tones useful, other users did not. Two users who had recently moved to Maryland thought that this software application would be very useful for learning state geography, which was not a stated scenario or development goal for the project, but could be a potential goal. Suggestions for improvement included a textual notification when you left a state border (such as text saying, “You are now in Virginia”), a splash sound instead of a bird chirp to notify you when you are crossing a body of water, and headsets to get a better spatial sense of where the sounds are coming from.

8.6 Implications and Discussion While the current investigation into developing sonified weather maps is at an early stage, there are some interesting implications for future research. Feedback from users indicates that visualisation maps present important and commonly used data representations that remain mostly inaccessible to blind users. This inaccessibility becomes more pertinent in information contexts with an inherent spatial component (such as weather data) where data values change rapidly and map sweeps can lead to important trend analyses. The enthusiasm of our usability evaluators (during interactions with the touchscreen and the tactile map) at being able to, for the first time, get an idea of the spatial orientation of the state of Maryland on the computer, indicates that there are interesting implications of research into designing and employing accessible maps as a pedagogical tool for

84

Weir et al.

learning not only weather, but also geography. Furthermore, because mobile phones and tablet devices are increasingly using multi-touch screens, and weather information is often needed on-the-go, it is important to investigate in the future how these applications could be used in portable devices such as Droid Phones and the Apple iPad. Touchscreens can be fully accessible to blind users (by using speech output, touchscreen gesturing, and Braille overlays to indicate where the visual keyboard appears), and it seems that a next step for research might be to make the application more robust, and determine how it might be implemented on a mobile or tablet device.

8.7 References Fritz JP, Barner KE (1999) Design of a haptic data visualization system for people with visual impairments. IEEE Transactions on Rehabilitation Engineering, 7(3): 372-384 Lazar J, Feng J, Hochheiser H (2010) Research methods in human-computer interaction. John Wiley and Sons, Chichester, UK Meiselwitz G, Wentz B, Lazar J (2010) Universal usability past, present and future. Foundations and Trends in Human Computer Interactions, 3(4): 213-333 Pauletto S, Hunt A (2009) Interactive sonification of complex data sonic interaction design. International Journal of Human-Computer Studies, 67(11): 923-933 Plaisant C (2004) The challenge of information visualization evaluation. In: the Proceedings of the Working Conference on Advanced visual interfaces, Gallipoli, Italy Shneiderman B, Plaisant C (2010) Designing the user interface: Strategies for effective human-computer interaction, 5th edn. Addison-Wesley, Boston, MA, US Walker BN, Mauney LM (2010) Universal design of auditory graphs: A comparison of sonification mappings for visually impaired and sighted listeners. ACM Transactions on Accessible Computing, 2(3): 12 Wall S, Brewster S (2006) Feeling what you hear: Tactile feedback for navigation of audio graphs. In: Proceedings of CHI 96 Workshop on Human Factors in Computer Systems, SIGCHI, Montreal, Quebec, Canada www.recvery.gov. (Accessed on 28 November 2011) Zhao H, Shneiderman B, Plaisant C, Lazar J (2008) Data sonification for users with visual impairments: A case study with geo-referenced data. ACM Transactions on Computer Human Interactions, 15(1): 4

Chapter 9 Achieving Inclusion in Public Spaces: A Shopping Mall Case Study Y. Afacan

9.1 Introduction Designing inclusive spaces can be seen as a response to accommodate diverse people within the built environment as efficiently, effectively, and satisfactorily as possible, regardless of health, body size, strength, experience, mobility and/or age. Although technological innovation has brought many benefits into architecture and planning, it is still difficult to embed inclusive design into real-world applications. Reviewing the literature on inclusion in the architectural design context indicates that an in-depth understanding to the diverse user of matching marketing purposes is lacking. Consequently, defining the user in the built environment as an ‘average person’ creates user-unfriendly public spaces. “Design exclusion does not come about by chance: it comes about through neglect, ignorance and a lack of adequate data and information” (Cassim et al., 2007). One of the main reasons for that is: most design practitioners are unable to take inclusive design into account during the initial phases of the design process, which leads to wrong decisions that can have a large impact on the overall design success and cost. The second reason is related to theory-practice inconsistency (Afacan and Erbug, 2009). Although there are guidelines and accessibility standards, designers have difficulty in utilising this academic source of information (Gregor et al., 2005). However, Nicolle et al. (2003) added, “Designers are under a great deal of time pressure: if knowledge is not presented in a usable format, it will be either discarded or ignored.” Therefore, although most designers are aware of universal design, problems appear in the integration of theories and guidelines into design practice (Demirkan, 2007). Despite the extensive literature on inclusive design, it is not easy to navigate the mass of data and interpret it into the cultural context. In Turkey, in the last decade, there has been a rise in the number of elderly and disabled people. It is traditional in Turkey to give a place of respect to the aging population in public spaces. However, compared to Europe and the US, there are still problems of integration of elderly and disabled people into social life because of environmental barriers, such as lack of ramps, disabled toilets, and inaccessible entrances to

P. Langdon et al. (eds.), Designing Inclusive Systems, DOI: 10.1007/978-1-4471-2867-0_9, © Springer-Verlag London 2012

85

86

Afacan

buildings. According to the Turkey Disability Survey (2002), most disabled people are still excluded in public spaces because designs do not provide the same opportunities of use for all users. Although the Turkish government realised the importance of inclusion within built environments and is developing policies (Republic of Prime Ministry Administration for Disabled People, 2011), still there is a need of studies to promote a positive attitude to inclusive design in the public, to encourage designers to design inclusively, and to make society sensitive and informed about diverse user needs, capabilities and expectations. This study is a further step of the previous study by Afacan and Erbug (2009), which promoted an interdisciplinary heuristic evaluation process for the universal usability of shopping malls. Afacan and Erbug (2009) highlighted the importance of working together with various design professions to lead to inclusion in real applications. According to this study, the lack of empathy for the requirements of diverse users is one of the three critical issues that make it difficult to integrate the inclusive design into current design practice. So, this study now delves deeper into how to include shopping malls from the user perspective. The aim of the study is not to evaluate the building performance of the case building, rather to focus on what makes a shopping mall more inclusive and how important each of the universal design criteria is in defining a mall as a user-friendly public space.

9.2 Inclusion in Public Spaces: Shopping Malls in the Turkish Context Achieving inclusion means embracing difference and celebrating human diversity. Inclusiveness is crucial for design success, business power and socially responsive societies. Since a public space is defined as a building that must accommodate everybody regardless of age, ability and size, the architectural features of those spaces need to adapt to these differences (Grosbois, 2001). Everyone needs to be part of society through the use of public buildings (Build for all manual, 2006). “To achieve this, the built environment, everyday objects, services, culture and information - in short, everything that is designed and made by people to be used by people - must be accessible, convenient for everyone in society to use and responsive to evolving human diversity” (EIDD, 2004). In that respect, focusing on enabling environments, which means featuring physical and intellectual accessibility and the sustainability of built structures, together with their impact on work, mobility and leisure within the community (Build for all Manual, 2006), will help to make the world more inclusive, bringing the government on board and engaging with users, in terms of expectations and cost implications. Public space in the study is exemplified by a shopping mall. Shopping malls are particularly important for leisure activities in large urban centres, which should ensure that all people are equally welcome and that all visitors can participate in facilities that have no design stigmatisation and that enrich their lives and enhance autonomy and flexibility (Resolution ResAP 3, 2001). Inaccessible public buildings for leisure activities are obviously holding disabled people back from

Achieving Inclusion in Public Spaces: A Shopping Mall Case Study

87

productive spheres of society (Haque, 2005). Moreover, the changing leisure and consumption patterns of Turkish people have made shopping malls among the most important additions to urban life in Turkey (Erkip, 2003). Within ten years (from 2000 to 2010), the number of shopping malls in Turkey has increased tremendously due to structural reforms and the introduction of foreign capital. “Crowding, traffic problems, and lack of pedestrian safety in the city centre served to create demand for these new areas” (Erkip, 2003). Although these malls currently provide a modern well-maintained atmosphere, their spatial and social characteristics still leave much to improve in terms of common activities, social participation, independence and well-being. As Lebbon and Hewer (2007) say, researching inclusion in those spaces is divided between three broad communities: business, design professionals and design education. However, the user is always at the heart of these three paths. Thus, seeing user issues from a wider perspective will not only encourage designers to design for inclusion, but also help society to increase awareness about enabling environments and develop empathy with others. Therefore, this study does not limit the impact of a shopping mall to the field of consumption only, but also highlights its importance in supporting inclusion as part of social and environmental considerations.

9.3 Methodology 9.3.1 Data Collection This exploratory study on investigating diverse users’ needs, capabilities and expectations was carried out in a shopping mall in Ankara, Turkey. The selected building is one of the biggest of six shopping centres in Ankara. It has an indoor area of 12.70m2 over nine storeys (five storeys for leisure facilities and four storeys for car parking) and was built in 2007. It also includes a hypermarket, 195 shops, nine cinemas, cafes, food court, an entertainment centre and offices. A survey instrument with a comprehensive list of 110 items was developed to gather data. It is based on a structured questionnaire format with close-ended questions. As in Afacan and Erbug’s study (2009), the questions in the survey instrument were grouped under five categories with reference to the seven universal design principles. Based on Danford and Tauke’s (2001) definitions these five essential design elements of a universal city, which should be considered when applying the seven principles of universal design in built environments, are as follows: 1. circulation systems: ramps, elevators, escalators, hallways, and corridors; 2. entering and exiting: identifying and approaching the entrance and exit and manoeuvring through them; 3. wayfinding: paths/circulation, markers, nodes, edges, and zones/districts; and graphical wayfinding: text, pictogram, map, photograph, and diagrams; 4. obtaining product/services: service desks, waiting areas, and shops; 5. public amenities: public telephones, restrooms (toilets), and seating units.

88

Afacan

9.3.2 Procedure A total of 120 randomly selected users participated in the survey, including 40 adults (between ages 25-55), 40 elderly (between ages 56-85) and 40 adult with impairments including 20 physically impaired adults (between ages 28 and 51) using wheelchairs (n = 13), prostheses (n = 7) as mobility aids and 20 visually impaired adults (between ages 30 and 59) having total loss of sight (n = 7) and mild loss of sight (n = 13). The data were collected during face-to-face surveys with all the participants in a café of the mall. At the beginning, a brief summary of the procedure and the aim of the study was explained. In the survey, participants were asked to rate their importance level for each item on a scale of 1-5, (1 being the least important and 5 the most important) and to mark the appropriate boxes to identify how important each features is in spending time satisfactorily and comfortably in a inclusive public space. The items that may not have been clear to participants were explained as part of the questionnaire. Further information was obtained through an unstructured interview. Further, to avoid any biases, participants were not allowed to listen to others while they were being surveyed.

9.4 Results and Discussion The ratings of the participants on 110 items were analysed with the Statistical Package for the Social Sciences (SPSS). First, exploratory factor analysis was used to carry out data analysis. Through the Varimax method, a frequently used rotation option (Argyrous, 2005), a rotated component matrix was constructed to identify the number of potentially interpretable factors among the set of correlations within the data. The matrix indicated the extracted factors with their factor loadings.

9.4.1 Development of Inclusive Public Space Factors Before carrying out factor analysis, firstly the survey instrument is checked whether there are any items creating floor and/or ceiling effects. It means that items at the extreme ends should be eliminated. Since the scale used in the study is 5, items below 1.5 and above 4.5 are regarded as extreme ends. There were no items at the extreme ends. Secondly, the strength of the correlations among the survey items was calculated through exploratory factor analysis, which helps identify common issues and exclude unrelated ones. Pearson product-moment correlations of the response scores were calculated and a correlation matrix was constructed, in which all the response items were illustrated in rows and columns of statistical relationships with a correlation score. Items having a correlation score lower than 0.30 are avoided for the study, because for a useful statistical approach a correlation coefficient of 1.00 indicates a perfect association between two variables (Argyrous, 2005). However, in the study all correlations between item response scores were greater than 0.30.

89

Achieving Inclusion in Public Spaces: A Shopping Mall Case Study

The study defined factor loadings in excess of 0.55 as suitable and excluded factors with factor loading values below 0.55. Total variance of factors was calculated. “Total variance shows all the factors extracted from the analysis along with their eigenvalues, the percentage of variance attributable to each factor, and the cumulative variance of the factor” (Mieczakowski et al., 2010). So, factor analysis resulted in a five-factor solution that accounted for 54.646% of the total variance, 110 items had 54.646% variances in common, so they correlated highly with five common themes; each theme was considered to be a factor (Table 9.1.). Table 9.1. Total variance explained Factor

Scale

Eigenvalue

Variance

Cumulative

1

Circulation consistency with user expectations Ease of access to services Multiple path of travel Comfortable use of services Appropriate use of tactile and audio-visual design features

8.041

21.385

14.078

3.698 3.054 1.990 1.537

11.239 7.361 6.123 5.038

22.340 32.124 42.125 54.646

2 3 4 5

The inclusive meanings assigned to the five factors are explained below: 1. Factor 1, ‘circulation consistency with user expectations’ is defined by equitable and simple use of the stairs, moving ramps, elevators and escalators. The appropriate uses of the tactile, aural, visual design features to maximise their legibility are as important as ease of use of circulation elements. Variables on this factor also include provision of clear surfaces for effective manoeuvring, which is an essential design consideration of public spaces for physically disabled people. 2. Factor 2, ‘ease of access to services’, deals with using shops, waiting desks and other public services with low physical effort and equitably. Walking along unimpeded should be a consideration for all people. Any level changes can create barriers for all disabilities and should be avoided or replaced by gentle slopes. 3. Factor 3, ‘multiple path of travel’, is defined by flexibility and simplicity of circulation, entering/exiting and way finding. Diverse choices of these elements help create inclusion in public spaces. Variables on this factor also include entering and exiting with low physical effort. 4. Factor 4, ‘comfortable use of services’, is defined as being use with low physical effort. Comfort in the public spaces can be achieved with lighting, public seating and sheltering structures. A calm, welcoming, user-friendly atmosphere of the shops and urban facilities is required by everyone (Burton and Mitchell, 2006). All components of the services should be designed to be comfortable and safe to reach. 5. Factor 5, ‘Appropriate use of tactile and audio-visual design features’, is defined by the provision of perceptible information. Public space should help all people regardless their ability to understand where they are and guide them the way they need to go. Legible spaces with clear signs and tactile surfaces are easy to navigate.

Afacan

90

Most of the participants had lots of interesting and useful ideas and comments on how to improve a shopping mall and public space. Table 9.2. lists commonly made suggestions. Table 9.2. Reported suggestions for improvement of a public space Design Elements

Suggestion for improvement

Circulation

More elevators

Entering and exiting

Wayfinding

Obtaining product/services

Public amenities

General

Smooth ramps Wider pedestrian walks Adequate clear space in circulation elements Ease of access to entrances/exits Adequate dimensions Clear path to/from the site Close to car park Understandable signs Ease of navigation Simple layouts More and comfortable waiting areas Wide passages in shops Ease of reach to all products Knee spaces at desks More disabled toilets Audio-visual features in ATMs Ease of reach to public telephones Audio-visual design features Textured floor material No obstructions on pavements Enough car parking Knee clearances

9.4.2 Correlation Differences in Diverse User Groups The study also utilised ANOVA analyses with Bonferroni post-hoc comparisons on each factor scale score and calculated the F-ratio in order to analyse whether the scale means of the user groups were significantly different from each other. The study found statistically significant differences between the user groups in factors 1, 3 and 5. Both the physical and visual impairments of users can affect the design and usage of a public space. For factor 1 and 3, the mean values indicated that most of the adults found simple and intuitive use of means of circulation moderately important, whereas physically impaired and elderly participants consider using circulation with low physical effort more important. Most of the physically impaired participants want to enter/exit easily and see multiple choices for entering

Achieving Inclusion in Public Spaces: A Shopping Mall Case Study

91

and exiting. For visually impaired, it was difficult to use elevators without audiovisual systems, so they found legibility more important. During the unstructured interviews regarding circulation, the elderly participants stated that they have a fear of falling and getting lost while they are spending time in a public space. For factor 5, most of the visually impaired participants emphasised the importance of having perceptible information for the use of public space elements (toilets, public phones, and doors). Since most of the services do not make use of a variety of techniques, such as colour-contrasts, Braille markings, large-print readouts, 16 of 20 participants with visual limitations had difficulties in knowing where and how to use what. However, the others (adults, elderly and physically impaired) considered safety features and warning of hazards more important. Furthermore, all participants regardless of their ability or disability found equitable use of public amenities very important.

9.5 Conclusions This study shed light on the needs, capabilities and expectations of diverse user groups in a shopping mall. The majority of the participants (regardless of their ability) stated that current real-world applications do not consider diverse user expectations and public spaces are designed for an average person which leads to exclusion. The most commonly offered improvements are understandable signs and ease of navigation. The graphics in signs are small to read for the elderly and difficult to understand for visually impaired people. Regarding navigation, all people experience problems because of obstructions and level changes. The results of the study relate highly to the design principles and recommendations that have been explained by Burton and Mitchell (2006) for inclusive urban design. According to Burton and Mitchell, there are six key design principles; (1) familiarity, (2) legibility, (3) distinctiveness (4) accessibility, (5) comfort, (6) safety, which make urban life more inclusive, easy and enjoyable for all members of society. Although these six principles are suggested for ‘streets for life’, both they and the factors developed in the study emphasise the urgent necessity of allowing equal access and opportunity regardless of ability and size. The results of this study also provided an understanding of the importance levels and attitudes of users towards inclusive environments that maximise quality of life. The developed factors highlight the significance of a user-friendly public space, which provides many ways of contact for elderly and disabled people. Since high quality in public space design is also a key consideration for sustainable communities both in Turkey and all over the world, equality of access and opportunity should be achieved to meet inclusion targets and to eliminate the disabling effects of built environments. However, more analysis should be conducted in other public spaces and outdoor areas, such as restaurants, cafes, museums, theatres, libraries and parks. Future research will continue to develop methods and tools to help designers achieve inclusion in public spaces.

92

Afacan

9.6 References Afacan Y, Erbug C (2009) Application of heuristic evaluation method by universal design experts. Applied Ergonomics, 40(4): 731-744 Argyrous G ( 2005) Statistics for research. Sage Publications, London, UK Build for all Manual (2006) Promoting accessibility for all to the built environment and public infrastructure. Available at: http://www.build-for-all.net/en/reference/ (Accessed on 28 October 2011) Burton E, Mitchell L (2006) Inclusive urban design: Streets for life. Elsevier, Oxford, UK Cassim J, Coleman R, Clarkson PJ and Dong H (2007) Why inclusive design? In: Coleman R, Clarkson PJ, Dong H and Cassim J (eds.) Design for inclusivity: A practical guide to accessible, innovative and user-centred design. Gower Publishing Ltd, Hampshire Danford GS, Tauke, B (eds.) (2001) Universal design: New York. Mayor's Office for People with Disabilities, NY, US Demirkan H (2007) Housing for the aging population. European Review of Aging and Physical Activity, 4(1): 33-38 Erkip F (2003) The shopping mall as an emergent public space in Turkey. Environment and Planning, A 35(6): 1073-1093 EIDD (2004) European Institute for Design and Disability. The EIDD Stockholm Declaration. Stockholm, Sweden Gregor P, Sloan D, Newell A (2005) Disability and technology: Building barriers or creating opportunities. In: Zelkowitz M (ed.) Advances in computers. Elsevier, Amsterdam Grosbois LP (2001) The evolution of design for all in public buildings and transportation in France. In: Preiser WFE, Ostroff E (eds.) Universal Design Handbook. McGraw-Hill, MA, US Haque S (2005) Accessibility for all: Role of architects to make a barrier free environment. In: Proceedings of UIA Region IV Work Programme ‘Architecture for All’ UIA/ARCASIA Workshop, Istanbul, Turkey Lebbon C, Hewer S (2007) Where do we find out? In: Coleman R, Clarkson PJ, Dong H and Cassim J (eds.) Design for inclusivity: A practical guide to accessible, innovative and user-centred design. Gower Publishing Ltd, Hampshire Mieczakowski A, Langdon PM, Clarkson PJ (2010) Investigating designers’ cognitive representations for inclusive interaction between products and users. In: Langdon PM, Clarkson PJ and Robinson P (eds.) Designing inclusive interactions, Springer-Verlag, London Nicolle C, Rundle C, Graupp H (2003) Towards curricula in design for all for information and communication products, systems and services. In: Proceedings of the International Conference on Inclusive Design and Communications (INCLUDE 2003), London, UK Republic of Prime Ministry Administration for Disabled People (2011) Available at: http://www.ozida.gov.tr/ENG/ (Accessed on 27 October 2011) Resolution ResAP 3 (2001) Towards full citizenship of persons with disabilities through inclusive new technologies. Available at: http://www.coe.int/t/e/social_cohesion/socsp/ResAP(2001)3E.pdf (Accessed on 14 July 2011) Turkey Disability Survey (2002) Available at: http://www.ozida.gov.tr/ENG/ (Accessed on 28 June 2011)

Chapter 10 Visibility Prediction Software: Five Factors of Contrast Perception for People with Vision Impairment in the Real World H. Dalke, A. Corso, G. Conduit and A. Riaz

10.1 Contrast and Vision Impairment The use of colour contrast in the built environment for people with low vision has been largely unsupported for architects, access consultants or designers, with little information available and no easy-to-use tools. Accessible environments assist everyone including vision impaired people (VIP); yet often people can be disabled by buildings, not directly by their impairment (Pullin, 2009); in 2002 a total of 0.6% of the world’s population were listed as blind (Harle and McLannahan, 2008). A recent critique of accessibility recommendations showed a lack of understanding of the five key factors we identified for predicting an object’s visibility (Dalke, 2011) namely, visual ability (VA) of the observer, contrast, lux level, dimension of the object and distance away from the observer. These were established as fundamental for the perception of objects, texts or elements for VIPs, contrast being one of those five interdependent variables (Dalke et al., 2010); they are fundamental to the software that has now been developed to predict object visibility. The research carried out revealed gaps in how to achieve contrast practically for the professionals who should be more familiar with the process. In the USA, the ADA Standard for Accessible Design makes reference to contrast but it is ambiguous and open to interpretation. In the 1991 standard, 70 points of contrast difference are prescribed for marking warnings on walkways (ADA, 1991). But how to gauge contrast, by calculating the difference between the light reflectance values (LRV) of two surfaces, is always missing. In the UK there is no advice on how to check and deliver on contrast for accessible buildings and products (DDA, 2004; EHRC, 2010). Previous studies conducted in a laboratory, in simulated real-world scenarios and on real-world sites (Dalke et al., 2010), indicated that contrast difference is critical for perception of the world for people, especially those with low vision (Rogers-Ramachandran and Ramachandran, 1998). Luminance is vital for visual

P. Langdon et al. (eds.), Designing Inclusive Systems, DOI: 10.1007/978-1-4471-2867-0_10, © Springer-Verlag London 2012

93

94

Dalke et al.

accessibility (ANSI/IESNA, 2007); it is not the absolute difference in luminance that is important, but the relative difference, expressed as contrast (Barten, 1999). Colour is not that significant for accessibility. It can increase perception if people have good colour vision, but 8% of the total male population may be colour vision impaired; people may also be coping with multiple disabilities (Goldsmith, 1967). There is an infinite variation in the visual capacity, acuity, and fields of vision of the partially sighted community so perception prediction is a challenge. In real-world investigations into what visually impaired people actually see, and how they use contrast and lighting to navigate, we identified key factors for perception and defined V4+ mathematically as a boundary for the software by averaging participants’ results during the tests. Findings from research highlighted: the disabling effects of glare from white surfaces with black text, rendering them painful or even impossible to read; the beneficial impact of raising lux levels further by even just 50 lux; the importance of graduating lighting levels to assist adaptation to changes from exterior to interior illumination. Not surprisingly the performance of VIP participants was better in lab tests, without the compounding variables of visual noise in real-world settings. The final iteration of the software was derived from experimentation in everyday environments, with VIPs and released as an app on the iPhone using the five interdependent variable factors; two of the five are fixed - the VA at 4 and currently the lux at 400. Contrast, dimension and distance are selectable using sliders on the screen which deliver a result of VISIBLE or NOT VISIBLE. The app has been tested in real-world scenarios and in this paper we describe the sequence of these Phases 1, 2, 3 and 4, which informed the algorithms for software ‘A’, ‘B’, ‘C’ and now ‘D’; the software allows architects to address the needs of the visually impaired population.

10.1.1 Phases of Testing There have been four phases of testing a series of software iterations ‘A’, ‘B’, ‘C’ and ‘D’; these tests were used for fine tuning the software for prediction of visibility for vision impaired people. 2001 to 2003 Phase 1 briefly extensive testing with 35 people with low vision, in transport hubs, recording the angle, height and distance from 380 objects which informed Software ‘A’ and identified five factors of interdependent variables. It was a PC based Windows program, published by Dalke et al. (2010) and predicted a visibility determining distance line for each VA level. Participants were vision tested for Visual Acuity, Visual Field and colour vision impairment. 2008 Phase 2 test recruited ten volunteers through eSight, with categories of vision impairment (V1 - V10) who self-defined their vision using a visual scale (Grundy et al., 1999; Douglas et al., 2006; Dalke et al., 2010). Participants were 18+ years with a variety of vision impairment conditions such as Retinitis Pigmentosa, 97% blind with very low field of vision some colour vision, participants with Nystagmus, night blindness, and others having problems with bright light, tunnel vision, cone dystrophy, some sight, light sensitive but relying

Visibility Prediction Software: Five Factors of Contrast Perception

95

on rods since birth, peripheral vision only, long sight and reasonable visual acuity 24/60, and another person with wet macular disease. There were no participants from V1, V7 and V9 (V1 has no possible perception of light). Phase 2 was conducted in a controlled laboratory testing Software ‘A’ predictions (Dalke et al., 2010). Each participant was logged as lowest VA group self-defined that ensured a margin of error for the algorithm. This phase tested distances at which VIPs were able to see the contrast difference between 2187 greyscale patches, presented in three different sized patches of ten grey LRVs on ten grey LRV backgrounds of premixed NCS colours (see Figure 10.1); 150mm2, 300mm2 and 750mm2 patches on a 1800mm by 2400mm background, from up to 10 metres distance away were tested (Figure 10.2). The LRVs of backgrounds and patches were 5, 10, 21, 27, 40, 53, 62, 71, 82 and 93% LRV positioned randomised at eye level. All greyscales were measured with a spectrophotometer (xyY) and each test conducted in a daylit room; the distance on the grid at which the participant observed the patch was recorded.

Figure 10.1. Three different grey patches on a background

Figure 10.2. Diagram of 10m distance testing environment Phase 2

Testing found unacceptable margins of error for small and large objects and the data was used to assess the accuracy of predictions from Software ‘A’. Software ‘B’ was developed - in the form of spreadsheet 2D ‘lookup’ charts with data lines for all contrasts that predicted visibility. 2010 Phase 3 was conducted in a lab and explored the boundaries of Software ‘B’. It extended the data range of the Phase 2, with two participants of V4 and V8 and a new test distance was added of 20m. Two extra size patches were tested, 1000mm2 and 50mm2 of identical LRVs to Phase 2 which were presented on backgrounds of grey with an increased test course of 20m. Participants observed the patches placed on three 1500mm2 backgrounds of 5%, 53% and 93% LRV, in randomised sequences. Each test was in a controlled environment of a lux level range of 200-400 lux (see Figure 10.3). The distance the patch was observed by the participant on the background was recorded (see Figure 10.4).

Dalke et al.

96

Figure 10.3. 1000mm2 patch on 1500mm2 background (left). 50mm2 patch on 1500mm2 background (right).

Figure 10.4. Diagram of testing environment and course Phase 3

The results for this phase can be seen (see Figure 10.5a) where a 50mm2 object of 90 points of contrast did not achieve better than 5m distance perception for a V4 participant, and a V8 perception improved significantly after 20 points of contrast difference (see Figure 10.5b).This phase of testing informed Software ‘C’, a PC based DOS program that encompassed the lookup charts developed in Software ‘B’.

Contrast Points

Contrast Points

(a)

(b) (b)

(a)

Figure 10.5. (a) 50mm and 1000mm Object dimension line and distances seen by V4 participant using results Phase 3. (b) 50mm and 1000mm Object dimension line and distances seen by V8 participant using results Phase 3.

2010 Phase 4 used a real-world environment for testing which included variables such as visual noise, and assessed the accuracy of predictions from

Visibility Prediction Software: Five Factors of Contrast Perception

97

Software ‘C’ with six participants V2 to V8. Previous lab tests Phase 2 and Phase 3 had all variables carefully controlled and measured - that is lux levels, dimension, distance and contrast. Phase 4 test was set in an environment with a mix of visual noise and stimuli - The Food Store at Kingston University. The six chosen participants’ visual abilities were logged as self-defined on the visual acuity scale (Grundy et al., 1999) and were V2, V4, V4, V4, V5 and V8 with a V10 control. Five locations were selected for testing. Participants were asked to stand at marked and measured predetermined locations, established by using the Software ‘C’. Objects at each location were listed on a record sheet with their data, e.g. distance and lux. The following were noted: smallest dimension of the object, LRV difference between object and background, lux level. At each location, the participants were invited to begin by observing the environment in front of them and move forward and ‘describe their view’. The target object was not singled out by the researcher; the researcher recorded the distance at which an object in the environment was clearly perceived by a participant (Figure 10.6). Software ‘D’ predictions were validated and established the V4+ boundary (V4 to V9 is 93% of VIP) as the software’s minimum default. Final tuning of the charts from a comparison of Phase 3 and 4 test results, for the algorithm used in Software ‘C’, then ‘D’ was used for an iPhone App, released in November 2010 (www.cromocon.com).

Figure 10.6. Walking in the corridor testing distance perception of objects

10.2 Analysis of Results and Algorithm Development Previous sections detailed the tests in Phases 2 and 3, obtaining of empirical data, including Phase 4, used to test the predictions of Software ‘C’ and ‘D’. The analysis of the results constructed and evaluated a novel algorithm to predict visibility. A comparison between Phase 2 and 3 results with the predictions from Software ‘A’ showed significant innacuracies to warrant development of further software described in the following sections. A new algorithm was developed using an empirical approach. Results from Phase 2 tests developed Software ‘B’ to overcome the two problems described below.

Dalke et al.

98

10.2.1 Extending the Range of the Test Data to 20 Metres Phase 2 test course of 10 metres was increased to 20 metres for Phase 3 with two participants (V4 and V8). Phase 2 data was extended by ‘typical differences’ where unconstrained data was analysed to identify the differences between object dimension ranges, which were then applied to constrained data. The figure below (10.7) shows results for a V4 subject before (a) and after the adjustments (b).

(a)

(b)

Figure 10.7. Test results from Phase 2 and 3, participant A05 (V4), for all contrast points. Adjusted test data for participant A05 (V4) extended to range of 20m, for all contrast points.

10.2.2 Adjusting the Data to Normalise for a Constant Lux of 400 A problem with Phase 2 results was optimising the varying lux values between each participant’s tests (from ~200 to ~1000 lux) for a lux of 400, that would be effective for a brightly lit task based working environment (Williams, 1999); results from Phases 2 and 3 were rescaled. Lux-distance relationships under different conditions of contrast, object dimension, and the VA level were found using Software ‘A’. Linear approximations to these curves were used to rescale data points. Figure 10.8 shows this applied to a sample data point (a).

Visibility Prediction Software: Five Factors of Contrast Perception

99

Figure 10.8. Lux-distance lines plotted using data points from Software ‘A’, V4 contrast of 50 and object dimensions of 150, 300, and 750mm. (a) - Phase 2 data point from participant A05 (V4) at contrast 50, dimension 300mm, and lux of 240 (b) - Data point (a) adjusted for lux, from 240 to 400 using linear trend-line (distance, contrast 50, and dimension 300mm).

The linear approximation depended on conditions of the data point requiring adjustment, including the VA level. Figure 10.9 shows results adjusted from 240 to 400 lux can be compared to Figure 10.7b for before/after lux adjustment.

Figure 10.9. Results for participant A05(V4) at 150mm, 300mm, and 750mm dimensions, adjusted from 240 to 400 lux

In averaging results from Phase 2 and 3 to find a Typical Data Set for Software ‘B’ we focused on the VA 4 level. Results from Phase 2 and 3 were averaged across all VA levels. A large number of results were used without having to rely on V4 data sets and allowed us to attenuate the visual anomalies to give a better typical visual impairment prediction. The visual range was guaranteed by averaging across VA levels (2+3+6+4+6+3+4+5+8) / 9 = 4.55. Measures of central tendency were investigated with results from Phase 2, a geometric mean was used; extreme

Dalke et al.

100

anomalous points did not skew these. Central tendency measures were found to have a geometric mean that gave the best representation for each contrast data set.

10.2.3 Interpolation between Data Points and Final Adjustments Linear interpolation is un-representative of a relationship between sight and distance; techniques were investigated such as logarithmic, exponential, and polynomial interpolation. Non-Uniform Rational B-splines avoided the oscillatory nature of previous interpolations as the degree of the curve is fixed independently of the number of points fitting the curve. The interpolation was also extended in the contrast direction creating a NURB surface (Piegl et al., 1997) in a 3D visual space consisting of distance, dimension and contrast axes. The surface describes a visual threshold - above the surface signifies no perception, and below, signifies perception. The numbers of data points in the contrast and dimension directions were used as the basis of knot vectors that gave a control net of 5 x 9 points. The Cox-de Boor recursion formula (Piegl et al., 1997) was used to define the blending functions in each parametric direction. Figure 10.10a shows a sample of interpolated curves superimposed on the same 2D axis.

(a)

(b)

Figure 10.10. NURB interpolated curves showing all contrasts for a typical V4 subject (a) and (b) adjusted NURB interpolated contrast curves for a typical V4 subject

Software ‘B’ was composed of nine lookup charts in Excel of increasing contrasts from 10 to 90 points. An algorithm was formulated and used these curves developed in DOS, Software ‘C’. These curves were found to contain anomalies (seen in Figure 10.10a) where lower contrast lines had higher distances at certain dimensions; these were resolved by lowering original test data points to fit the same behaviour as all other data lines, and by readjusting the control points defining the NURB interpolation between data points. A sample of the revised curves is shown above (Figure 10.10b). Software ‘C’ was adjusted before porting

Visibility Prediction Software: Five Factors of Contrast Perception

101

the revised algorithm into the app for the iPhone with a graphical user interface, labelled Software ‘D’.

10.2.4 Testing the Software Tests in Phase 4 (see Figure 10.6) and results provided data for comparisons of the predictions of software with participants A02, A06, and A10; A07 and A05 are not compared, as visibility lines were not established with the participants viewing all objects at all distances. All the distances in the Phase 4 tests were normalised for 400 lux. Contrast 60 and 30 comparisons had the most data points at each contrast stage and represented the majority of the results (See Figure 10.11). The V4 subject (A02) in the charts above includes the percentage deviation from predicted, with each data point; apart from points labelled (1) which are considered anomalous. Deviation in both contrasts begins very high with 87% and 90% but these are deceptive if distance differences between A02 and prediction are considered. They are similar at all dimensions except for 1000mm in Figure 10.11a. Considering the differences that can exist in visual impairment within the same VA level, subject A02 compares well with the predictions for a typical V4. The two V5 subjects perform lower than the predicted V4, in most of their data points, when they should be higher. Participant A06 (V5) is especially low at contrast 30 points.

(a)

(b)

Figure 10.11. Comparison of Software ‘D’, typical V4 predictions with Phase 4 test results at contrast 60 points (a) and comparison of Software ‘D’, typical V4 predictions with Phase 4 test results at contrast 30 points (b)

102

Dalke et al.

10.3 Conclusions The studies highlighted the inconsistency of participants’ self defined VA level, the range of particular vision impairments (no two people with identical impairments), and the unpredictable impact of each individual eye condition on the observation of targets tested in the ‘real-world’. Not surprisingly participants performed much worse in busy real-world test locations (Figure 10.6) proving the danger of relying solely on lab testing for the development of assistive models of visual perception. However, a secure model has now been synthesised and developed through the investigation of the broad range of variables for perception. Five factors were explored and integrated into a practical tool, the app (www.cromocon.com), which is proving to be a robust and valuable tool for the design of inclusive environments. Further work is being undertaken with a large cohort of visually impaired people to extend the empirical work of the study.

10.4 References ADA (1991) Americans with Disabilities Act. ADA Standards for accessible design. Accessibility guidelines for buildings and facilities. U.S. Department of Justice ANSI/IESNA (2007) Lighting and the visual environment for senior living, IES RP-28-07. New York, US Barten PGJ (1999) Contrast sensitivity of the human eye and its effects on image quality. SPIE Press, Bellingham, WA, US Dalke H, Conduit GJ, Conduit B, Cooper R, Corso A, Wyatt DF (2010) A colour contrast assessment system: Design for people with visual impairment. In: Langdon PM, Clarkson PJ, Robinson P (eds.) Designing inclusive interactions. Springer, London, UK Dalke H (2011) The contrast guide. Cromocon, London, UK DDA (2004) Disability Discrimination Act. Her Majesty’s Stationery Office, London, UK Douglas G, Corcoran C, Pavey S (2006) Network 1000: Opinions and circumstances of VIP people in Great Britain. Visual Impairment Centre for Teaching and Research (VICTAR), University of Birmingham, Birmingham, UK EHRC (2010) European Human Rights Commission Equality Act. HMSO London, UK Goldsmith S (1967) Designing for the disabled. RIBA Publications Ltd, London, UK Grundy E, Ahlburg D, Ali M, Breeze E, Sloggett A (1999) Disability in Great Britain: Results from the 1996/97 disability follow-up to the family resources survey. Technical Report 94. Department of Social Security, Leeds, UK Harle J, McLannahan H (2008) Visual impairment: A global view. Oxford University Press, UK Piegl LA, Tiller W (1997) The NURBS book, 2nd edn. Springer-Verlag, New York, NY, US Pullin G (2009) Design meets disability. MIT Press, Cambridge, MA, US Rogers-Ramachandran DC, Ramachandran VS (1998) Psychophysical evidence for boundary and surface systems in human vision. Vision Research, 38(1): 71-77 Williams W (1999) Footcandles and lux for architectural lighting. An introduction to illuminance, 2nd edn. Available at: http://www.mts.net/~william5/library/illum.htm (Accessed on 28 November 2011)

Part III

Designing Cognitive Interaction with Emerging Technologies

Chapter 11 Intrinsic Motivation and Design of ICT for the Ageing Population T.S. Goldhaber, P.M. Langdon and P.J. Clarkson

11.1 Introduction In many countries around the world, the average of the population is rapidly increasing (OECD, 2006). On average, people require more healthcare as they age (Blaschke et al., 2009), meaning that in a population in which more people are of an advanced age, there will need to be a higher level of healthcare provision (Gray, 2005). However, existing medical systems have neither the finances nor capacity to insure a high standard of care for older individuals (Rechel et al., 2009). Information and Communication Technology (ICT) systems provide an increasingly promising platform with which to improve the efficiency and effectiveness of healthcare, particularly in a preventative context (Blaschke et al., 2009; Gupta et al., 2009; Rechel et al., 2009). For example, ICT can enable better communication with carers and medical professionals, allow individuals to better manage their own care, increase independence for older individuals, and reduce some of the problems of social isolation associated with old age (Chaffin and Harlow, 2005; Blaschke et al., 2009). Unfortunately, older people have persistently been slow to adopt ICT and consequently miss out on many of the benefits that could drastically improve their well-being (Mynatt et al., 2004; Czaja and Hiltz, 2005; Czaja et al., 2006; Wagner et al., 2010). Poor user interface (UI) design is partially responsible for some of the problems with ICT adoption in the ageing population (Hawthorn, 2000, 2007; Dickinson et al., 2005; Czaja et al., 2006). The problems with interface adoption and use in the ageing population are partially problems of motivation: older adults are not necessarily motivated to learn to use existing or new ICT interfaces (OFCOM, 2006). However, while both motivation in older adults (Lawton et al., 2002; Melenhorst et al., 2006) and the application of motivation theory to UI design (Williams et al., 2007; Jung et al., 2010) have been studied to some extent, the direct application of motivation theory to UI design for the ageing population has not. Because ICT will continue to present opportunities for improved healthcare and independence for the ageing population, motivating them to use it is essential. The

P. Langdon et al. (eds.), Designing Inclusive Systems, DOI: 10.1007/978-1-4471-2867-0_11, © Springer-Verlag London 2012

105

Goldhaber et al.

106

purpose of this paper is to begin to investigate how existing theories of intrinsic motivation might be applied to motivate the adoption and learning of ICT among older individuals.

11.2 ICT Adoption, Learning, and Use in the Ageing Population Because of factors such as declining fluid intelligence, spatial ability, memory, and attention span, coupled with increased computer anxiety, older users have a much harder time learning to use new ICT (Hawthorn, 2000). In addition, older people are less likely to engage in activities that they do not find enjoyable. As explained by Melenhorst et al. (2001): “Even if they are relatively healthy, older adults realise their place in the life cycle…they tend to be present-oriented and are reluctant to spend their time in an unpleasant way” (Melenhorst et al., 2001). According to Socioemotional Selectivity Theory (SST), imposed time constraints, including ageing, can lead to a change in overall goals (Carstensen et al., 1999), and many older users do not see modern ICT as related to their interests or goals (Hawthorn, 2007). In addition, they may not find the use of an initially difficult and confusing interface enjoyable and may therefore be reluctant to learn to use it even if they are aware of the potential benefits (Melenhorst et al., 2006). Learning problems, anxiety, and lack of familiarity increase the probability that ICT use will not initially be either enjoyable or useful, leading to reduced adoption and use rates (Hawthorn, 2000). Earlier research shows two things: • •

There are significant motivational barriers to technology adoption in the ageing population. Motivation theory can be applied to user interface design to increase motivation in users (e.g. Williams et al., 2007; Jung et al., 2010).

The area that needs to be addressed, therefore, is how motivation theory can be applied to interface design to increase motivation to adopt and fully learn to use new ICT. Critically, interfaces aimed at older users must motivate use of the interface itself, not just the accomplishment of an external goal, so the enjoyment users get from using the interface must be maintained as well. This makes intrinsic motivation a crucial area of study.

Intrinsic Motivation and Design of ICT for the Ageing Population

107

11.3 Intrinsic Motivation and Design The idea of intrinsic motivation is most easily traced back to White (1959), who summarised a great deal of research supporting the conclusion that not all motivation can be explained by extrinsic factors. For example, exploratory behaviour in young animals is rarely extrinsically motivated. Clearly, as White explained, there is some underlying framework for intrinsic motivation that needs to be understood. White identified competence, or the ability to affect and cope with one’s surroundings, as the primary driver of intrinsic motivation (White, 1959). Bandura’s concept of self-efficacy, influenced by White’s work, also relies heavily on the concept of competence, which is referred to as “performance accomplishment.” Bandura emphasised that a sense of efficacy early on in an activity correlates with motivation and, later, persevering through challenges and setbacks (Bandura, 1977). Intrinsic motivation is critical in design for the ageing population because they are much less likely to undertake a task if they do not find it enjoyable (Carstensen et al., 1999; Melenhorst et al., 2001). To expand on these ideas, the following sections summarise some prominent theories that deal either in whole or in part with intrinsic motivation.

11.3.1 Self-determination Theory Self-Determination Theory (SDT) is based on “the hypothesis that there is a set of universal psychological needs that must be satisfied for effective functioning and psychological health” (Deci and Ryan, 2008b). Importantly, the fulfilment of these needs predicts not only overall well-being but also the extent to which motivation becomes internalised - intrinsic motivation being the mostly highly internalised form of motivation (Ryan and Deci, 2000b) - and the consequent potential for learning (Deci et al., 1991). These core needs are defined by Deci and Ryan as autonomy, relatedness, and competence (Deci and Ryan, 2000). Autonomy lies at the heart of SDT. Autonomous actions are defined as those “for which people feel a full sense of choice and endorsement of an activity” (Deci and Ryan, 2008a). If someone does not feel a sense of volition and control over their decisions or actions, they cannot be intrinsically motivated because they are by definition working to fulfil externally set goals or avoid potential imposed consequences. It is the nature of those goals and consequences that leads to a lack of autonomy, which in turn can seriously degrade intrinsic motivation. Relatedness, although not as immediately applicable to interface design, is one of the key psychological needs specified in SDT. It is defined as “the need to feel belongingness and connectedness with others,” and is “centrally important for internalisation” (Ryan and Deci, 2000b). Competence is “a sense of efficacy” (Przybylski et al., 2010): the feeling acquired when someone feels that they have the skills necessary to complete a task or influence their environment. Both rewards and feedback, among other factors

108

Goldhaber et al.

such as an optimal level of challenge, can enhance feelings of competence, but only if there are sufficient feelings of autonomy - an “internally perceived locus of causality” (deCharms, 1968; Ryan and Deci, 2000a). The idea of competence as an important factor in intrinsic motivation was first introduced by White (1959). 11.3.1.1 Cognitive Evaluation Theory Cognitive Evaluation Theory (CET) is a sub-theory of SDT that focuses specifically on the causes of intrinsic motivation, particularly the SDT needs of competence and autonomy (Ryan and Deci, 2000b). Feelings of self-determination are obviously most affected by the perceived locus of causality and the resulting feelings of autonomy, but feelings of competence also strongly affect intrinsic motivation. CET has been tested in over 100 studies and has been found to have effects in many activities, such as sports and educational environments (Deci et al., 1999; Ryan et al., 2006). One of the more interesting and consistent experimental results is the finding that extrinsic rewards actually undermine intrinsic motivation, meaning that what might promote extrinsic motivation can have the opposite effect intrinsically (Ryan et al., 1983; Rummel and Feinberg, 1988; Deci et al., 1999; Ryan and Deci, 2000b). The concept of competence is complemented and expanded on by the theory of Flow (Csíkszentmihályi, 1990; Sheldon and Filak, 2008), which is explained in the next section. 11.3.1.2 Research Implications of SDT and CET OFCOM (2006) surveyed a large number of older users about their use of computers and the internet. Those who were interested in learning to use this technology, but either had not tried or had experienced difficulty, mentioned lack of skills, benefit perception, and motivation as well as social/environmental factors as the main barriers to adoption and use. While the fear of becoming socially isolated due to ICT use points more to the relatedness need of SDT, the problems with skills clearly point to competence as a key problem. Since ICT use is rarely mandated or externally rewarded, it is also likely that the need for autonomy may already be filled for many older users. Therefore, it makes sense that future research focus first and foremost on filling the competence need. This approach is supported in the literature. Both White (1959) and Bandura (1977) cite competence/sense of efficacy as critical to intrinsic motivation, explaining how acquiring feelings of competence early in an experience can prevent later negative reactions to errors and promote perseverance in the face of challenge, both noted issues for older people learning to use ICT. Bandura in particular emphasises that individuals must have many early successes, which tend to lessen the negative impact of future failures. UI design in general, therefore, should focus on creating as many successful experiences as possible, particularly with initial tasks.

Intrinsic Motivation and Design of ICT for the Ageing Population

109

11.3.2 The Theory of Flow Flow is “optimal experience”: a state of total engagement with or immersion in an activity (Csíkszentmihályi, 1990). The experience of Flow, by definition, requires an activity or experience to be enjoyable and is therefore related to intrinsic motivation (Privette, 1983). Flow in system use has also been shown to lead to more exploratory behaviour in users (Ghani and Deshpande, 1994). Therefore, understanding what leads to a Flow experience has the potential to unlock system properties that could encourage intrinsic motivation in users. Csíkszentmihályi (1990) identifies three main components of an activity that lead to Flow. The first is a challenge level that is matched to the skill set of the user. Second, an activity must have clear and bounded goals such that the person involved has a direction for their activity. Third, there must be immediate feedback so that the user or actor knows how they are doing. It is essential in order to experience Flow that someone knows that they are accomplishing the goals that have been laid out. (Csíkszentmihályi, 1990). A separate but nonetheless important feature of the Flow experience is a sense of control (Fisher, 1978), which although not elaborated in as much detail in Csíkszentmihályi’s writings is a major component, autonomy, in SDT/CET. Flow also expands nicely upon the competence and autonomy aspects of SDT, and competence in particular, (Sheldon and Filak, 2008). This overlap creates both confidence in the potential of these theories to improve interface design and a space for experiments and studies testing how these psychological needs and system properties can be best incorporated into interfaces. In addition, Csíkszentmihályi reveals that “the more a job inherently resembles a game - with variety, appropriate and flexible challenges, clear goals, and immediate feedback the more enjoyable it will be regardless of the worker’s level of development” (Csíkszentmihályi, 1990). Indeed, games provide an interesting platform for the study of intrinsic motivation in addition to indicating that interfaces can provide flow and intrinsic motivation if they are structured properly.

11.3.3 Other Theories of Intrinsic Motivation and Design While SDT/CET and Flow are broad theories that deal specifically with intrinsic motivation, there are other theories that touch on the precursors for intrinsic motivation and/or design principles for better motivation. This section provides brief overviews of those theories and their implications for UI design. 11.3.3.1 Theory of Intrinsically Motivating Instruction (TIMI) In the early 1980s, Thomas Malone studied different educational games, both interviewing students about their reactions, likes, and dislikes and also modifying the interfaces in various ways in order to try to isolate motivating properties (Malone, 1981). As the result of this research, he identified three factors that appear to facilitate intrinsically motivating instruction within educational

110

Goldhaber et al.

interfaces: challenge, fantasy, and curiosity. Malone’s experimental methods also provide a good framework for future experimental work in this area. 11.3.3.2 The 2 x 2 Achievement Goal Framework The 2 x 2 Achievement Goal Framework deals with two dimensions of motivation: performance/mastery and approach/avoidance. This framework comprises four achievement goals: “mastery-approach (focused on attaining task-based or intrapersonal competence), performance-approach (focused on attaining normative competence), mastery-avoidance (focused on avoiding task-based or intrapersonal incompetence), and performance-avoidance (focused on avoiding normative incompetence” (Cury et al., 2006). In general, mastery goals prove more intrinsically motivating than performance goals, and approach goals more so than avoidance goals. 11.3.3.3 Positive Design Theory Ping Zhang (2008a, 2008b) identifies the need to take a motivational approach to ICT design, explaining that “understand[ing] technology use behaviour and eventually feed[ing] this understanding into technology design…naturally calls for a motivational approach because motivation explains what gives behaviour its energy and direction” (Zhang, 2008b). Zhang’s papers provide good overviews of how various theories of motivation relate to different design problems. 11.3.3.4 Funology and Gamification The idea that non-game interfaces should be fun is often overlooked in the field of Human-Computer Interface (HCI) design (Monk et al., 2002). While it is obviously not always a desirable feature to have interfaces be fun, particularly when those interfaces are a means to an end, Draper (1999) identifies three scenarios in which fun is an important consideration in design: (1) when fun is the main purpose of the interface, such as a game; (2) when learning is the main function of the interface; and (3) when “learnability is considered as an important secondary requirement of software with some other main function” (Draper, 1999). Monk et al. (2002) proposes the term “Funology” to refer to the study of making interfaces fun for purposes including but not limited to games, the importance of which is recognised in the literature, if not pervasively (Rieber et al., 1998; Draper, 1999).

11.3.4 Summary In addition to Zhang’s design recommendations and the concepts of Funology and Gamification, four theories relating to intrinsic motivation were presented above: Self-Determination Theory, Flow, the Theory of Intrinsically Motivating Interfaces, and the 2 x 2 Achievement Motivation Theory. Interestingly, these theories all relate to and inform one another to some extent. Figure 11.1 shows the relationships between the different aspects of each theory.

Intrinsic Motivation and Design of ICT for the Ageing Population

111

Figure 11.1. How the intrinsic motivation theories relate to each other

There are many features of these theories that prove incredibly useful for design, particularly as variables that can be manipulated within experiments. Most importantly, the idea of competence and challenge is present in all four theories, although it is expressed and broken down differently in all of them. This is useful because it is the most relatable to interface design and the varying of challenge level is one of the most salient features of computer games, which have been studied extensively. Because of this, because the absence of feelings of competence has been shown to be one of the primary barriers to the adoption of ICT, and because feelings of competence have been shown to have effects such as increasing exploratory behaviour in users (Chaffin and Harlow, 2005) competence generation must be a primary focus of future UI design research.

11.4 Conclusions and Recommendations Because of lack of prior experience, reduced fluid intelligence, memory capacity and spatial ability, among other factors, some older individuals can have a harder time learning new skills than younger users Older users also have low levels of computer self-efficacy, knowledge, experience, and benefit perception, all of which reduce the desire to adopt the technology. In addition, older people are much less likely to engage in activities that they do not find enjoyable. This means that increasing the intrinsic motivation of users through the application of motivation theory to design could encourage older individuals to adopt new ICT. Many theories in social psychology identify elements that lead to intrinsically motivating experiences. Cognitive Evaluation Theory, a sub-theory of SelfDetermination Theory, identifies competence and autonomy as critical factors for intrinsic motivation. Flow theory unpacks the experience of competence into the critical components of appropriate challenge, clear goals, and useful feedback. The Theory of Intrinsically Motivating Instruction captures how better design can

112

Goldhaber et al.

promote learner engagement. Finally, the 2 x 2 Achievement Goal Framework shows how different kinds of goals affect intrinsic motivation. Feelings of lacking competence and self-efficacy have been identified in the literature as one of the single biggest barriers to adoption, learning and use of ICT among older individuals. Therefore, design research based on intrinsic motivation must focus on trying to generate feelings of competence in users and measuring subsequent changes in intrinsic motivation and learning. Moreover, as competence is the area of greatest overlap among the theories of intrinsic motivation presented, this is likely the most important feature on which to focus in future research.

11.5 Acknowledgements The authors are grateful for the help of Dr Andrew Przybylski in providing and discussing many valuable sources of information and ideas regarding intrinsic motivation.

11.6 References Bandura A (1977) Self-efficacy: Toward a unifying theory of behavioral change. Psychological Review, 84(2): 191-215 Blaschke C, Freddolino P, Mullen E (2009) Ageing and technology: A review of the research literature. British Journal of Social Work, 39(4): 641-656 Carstensen LL, Isaacowitz DM, Charles ST (1999) Taking time seriously - a theory of socioemotional selectivity. American Psychologist, 54(3): 165-181 Chaffin AJ, Harlow SD (2005) Cognitive learning applied to older adult learners and technology. Educational Gerontology, 31(4): 301-329 Csíkszentmihályi M (1990) Flow: the psychology of optimal experience. Harper & Row, New York, US Cury F, Elliot AJ, Da Fonseca D, Moller AC (2006) The social-cognitive model of achievement motivation and the 2x2 achievement goal framework. Journal of Personality and Social Psychology, 90(4): 666-679 Czaja SJ, Charness N, Fisk AD, Hertzog C, Nair SN, Rogers WA et al. (2006) Factors predicting the use of technology: Findings from the center for research and education on aging and technology enhancement (CREATE). Psychology and Aging, 21(2): 333-352 Czaja SJ, Hiltz SR (2005) Digital aids for an aging society. Communications of the ACM, 48(10): 43-44 deCharms R (1968) Personal causation. Academic Press, New York, US Deci EL, Koestner R, Ryan RM (1999) A meta-analytic review of experiments examining the effects of extrinsic rewards on intrinsic motivation. Psychological Bulletin, 125(6): 627-668 Deci EL, Ryan RM (2000). The “what” and “why” of goal pursuits: Human needs and the self-determination of behavior. Psychological Inquiry, 11(4): 227-268 Deci EL, Ryan RM (2008a) Facilitating optimal motivation and psychological well-being across life’s domains. Canadian Psychology-Psychologie Canadienne, 49(1): 14-23 Deci EL, Ryan RM (2008b) Self-determination theory: A macrotheory of human motivation, development, and health. Canadian Psychology-Psychologie Canadienne, 49(3): 182-185

Intrinsic Motivation and Design of ICT for the Ageing Population

113

Deci EL, Vallerand RJ, Pelletier LG, Ryan RM (1991) Motivation and education: The selfdetermination perspective. Educational Psychologist, 26(3/4): 325- 346 Dickinson A, Eisma R, Gregor P, Syme A, Milne S (2005) Strategies for teaching older people to use the World Wide Web. Universal Access in the Information Society, 4(1): 3-15 Draper SW (1999) Analysing fun as a candidate software requirement. Personal and Ubiquitous Computing, 3(3): 117-122 Fisher CD (1978) Effects of personal control, competence, and extrinsic reward systems on intrinsic motivation. Organizational Behavior and Human Performance, 21(3): 273-288 Ghani JA, Deshpande SP (1994) Task characteristics and the experience of optimal flow in human-computer interaction. Journal of Psychology, 128(4): 381-391 Gray A (2005) Population ageing and health care expenditure. Ageing Horizons, 2: 15-20 Gupta M, Chotard L, Ingporsson O, Bastos J, Borges I (2009) Health@home - an e-service model for disease prevention and healthcare in the home. Electronic Healthcare, 1: 1724 Hawthorn D (2000) Possible implications of aging for interface designers. Interacting with Computers, 12(5): 507-528 Hawthorn D (2007) Interface design and engagement with older people. Behaviour and Information Technology, 26(4): 333-341 Jung JH, Schneider C, Valacich J (2010) Enhancing the motivational affordance of information systems: The effects of real-time performance feedback and goal setting in group collaboration environments. Management Science, 56(4): 724-742 Lawton MP, Moss MS, Winter L, Hoffman C (2002) Motivation in later life: Personal projects and well-being. Psychology and Aging, 17(4): 539-547 Malone TW (1981) Toward a theory of intrinsically motivating instruction. Cognitive Science, 5(4): 333-369 Melenhorst AS, Rogers WA, Bouwhuis DG (2006) Older adults’ motivated choice for technological innovation: Evidence for benefit-driven selectivity. Psychology and Aging, 21(1): 190-195 Melenhorst AS, Rogers WA, Caylor EC (2001) The use of communication technologies by older adults: Exploring the benefits from the user’s perspective. Human Factors and Ergonomics Society Annual Meeting Proceedings, 45(3): 221-225 Monk A, Hassenzahl M, Blythe M, Reed D (2002) Funology: Designing enjoyment Mynatt ED, Melenhorst AS, Fisk AD, Rogers WA (2004) Aware technologies for aging in place: Understanding user needs and attitudes. IEEE Pervasive Computing, 3(2): 36-41 OECD (2006) Live longer, work longer: A synthesis report of the Ageing and Employment. Policies Project. OECD, Paris, France OFCOM (2006) Older people and communications technology. OFCOM Consumer Panel. Privette G (1983) Peak experience, peak performance, and flow: A comparative analysis of positive human experiences. Journal of Personality and Social Psychology, 45(6): 13611368 Przybylski AK, Rigby CS, Ryan RM (2010) A motivational model of video game engagement. Review of General Psychology, 14(2): 154-166 Rechel B, Doyle Y, Grundy E, McKee M (2009) How can health systems respond to population ageing? World Health Organization, Copenhagen, Denmark Rieber LP, Smith L, Noah D (1998) The value of serious play. Educational Technology, 38(6): 29-37 Rummel A, Feinberg R (1988) Cognitive evaluation theory - a meta-analytic review of the literature. Social Behavior and Personality, 16(2): 147-164 Ryan RM, Deci EL (2000a) Intrinsic and extrinsic motivations: Classic definitions and new directions. Contemporary Educational Psychology, 25(1): 54-67

114

Goldhaber et al.

Ryan RM, Deci EL (2000b) Self-determination theory and the facilitation of intrinsic motivation, social development, and well-being. American Psychologist, 55(1): 68-78 Ryan RM, Mims V, Koestner R (1983) Relation of reward contingency and interpersonal context to intrinsic motivation - a review and test using cognitive evaluation theory. Journal of Personality and Social Psychology, 45(4): 736-750 Ryan RM, Rigby CS, Przybylski A (2006) The motivational pull of video games: A selfdetermination theory approach. Motivation and Emotion, 30(4): 347-363 Sheldon KM, Filak V (2008) Manipulating autonomy, competence, and relatedness support in a game-learning context: New evidence that all three needs matter. British Journal of Social Psychology, 47: 267-283 Wagner N, Hassanein K, Head M (2010) Computer use by older adults: A multi-disciplinary review. Computers in Human Behavior, 26(5): 870-882 White RW (1959) Motivation reconsidered - the concept of competence. Psychological Review, 66(5): 297-333 Williams GC, Lynch M, Glasgow RE (2007) Computer-assisted intervention improves patient-centered diabetes care by increasing autonomy support. Health Psychology, 26(6): 728-734 Zhang P (2008a) Motivational affordances: Reasons for ICT design and use. Communications of the ACM, 51(11): 145-147 Zhang P (2008b) Toward a positive design theory: Principles for designing motivating information and communication technology. In: Avital M, Boland R, Cooperrider D (eds.) Designing Information and Organizations with a Positive Lens. Elsevier Science, Oxford, UK

Chapter 12 A Framework for Studying Cognitive Impairment to Inform Inclusive Design E. Jokisuu, P.M. Langdon and P.J. Clarkson

12.1 Introduction Cognitive impairment is an exceptionally complicated phenomenon, for which no simple solutions exist. There is no unified definition of what constitutes cognitive impairment and the terminology is varied. There is no uniform pattern of impairment occurring with any one medical condition and individual differences are significant: the same disease can affect people in different ways and the same condition can cause diverse patterns of impairment in different individuals. Some medical conditions causing cognitive impairment are chronic, others progressive, which adds to the difficulty of assessing their impact. In addition, the co-morbidity of certain medical conditions causes unexpected patterns of impairment. With certain conditions, such as brain injury, the impact of the impairment depends entirely on the location and extent of brain damage. (Groome et al., 2006) Furthermore, the number of people living with cognitive disability is growing due to population ageing and the improved survival rates for trauma patients (Ghajar, 2000; WHO, 2008). Neuropsychiatric illnesses and traumatic brain injury are among the leading causes of disability (WHO, 2008). It is perhaps due to these complicating factors that cognitive impairment has not been adequately addressed in current research. Related research conducted in the field of assistive technology has given valuable insight into the design of devices for specific medical conditions and impairment types (e.g. Gartland, 2004; LoPresti et al., 2004; Cohene et al., 2007). Design guidelines also exist to outline some general aspects of product usability, partly aiming to minimise the cognitive load to users (e.g. Centre for Universal Design, 1997; W3C, 2008). However, there is no systematic and comprehensive review of design issues relating to cognitive impairment. Assistive technology products tend to be developed for a very specific purpose and tailored to each individual, whereas design guidelines are often too general and fail to provide justification or deeper understanding of the needs of the people with cognitive impairment. In addition, it is not easy to estimate the number of people excluded from using a product, as

P. Langdon et al. (eds.), Designing Inclusive Systems, DOI: 10.1007/978-1-4471-2867-0_12, © Springer-Verlag London 2012

115

116

Jokisuu et al.

there are no comprehensive statistical data available about the prevalence and incidence of cognitive impairment. Understanding users and knowing their needs and requirements is vitally important to the success of inclusive design (Carlsson et al., 2002). Still, as important as user involvement is in the design process, there are particular challenges when involving users with cognitive impairment. Designers need support if they are to make design inclusive of such people. One way of supporting them is to provide information that is accurate, relevant, offers them insight and inspiration and is presented in a way that makes it easy to apply. This paper reports the first stage of a study that seeks to develop a model of cognitive impairment providing such design guidance. In this first stage, a framework for studying cognitive impairment is developed. This framework aims to identify those key aspects of cognitive impairment that designers need to acknowledge whilst also capturing the complexity of the phenomenon. The purpose of this study is a) to define the cognitive functions which are relevant to design; b) to categorise the medical conditions which can cause impairment in these functions; and c) to develop a framework which can be used to systematically collect information about cognitive impairment as it applies to design. This is the first stage of a larger study which aims to model cognitive impairment in a way that translates medical and psychological information into design guidance. This paper is organised as follows: Firstly, the research method is described briefly. Then the development of categories for both cognitive impairment and the medical conditions causing it is presented, followed by a description of the framework created based on these categories. Next an example of using the framework is given. The paper concludes with a discussion of the work so far and future steps of the research.

12.2 Research Method Cognitive impairment can be diagnosed through the medical condition causing it. In order to identify and define the different types of cognitive impairment and the medical conditions behind them, interviews with experts on neuropsychology and occupational therapy were conducted. The purpose of these interviews was to create two categorisations: one of cognitive impairment and another of medical conditions causing cognitive impairment. In addition, the purpose was to discuss ways to assess the links between the medical conditions and the types of impairment.

12.2.1 Interview Technique Semi-structured interview was chosen as the method for collecting information from the subject matter experts. As the area of interest was not very well defined, instead of pre-defined specific questions, a list of general themes for discussion

A Framework for Studying Cognitive Impairment to Inform Inclusive Design

117

was covered during the interview to explore the issues of interest. The interview outline is presented in Table 12.1. Each interview lasted for about one hour. The interviews were recorded using a voice recorder and then transcribed. A consent form was signed by all interviewees. Table 12.1. Interview themes What do you think of these categories of medical conditions? What do you think of these categories of cognitive functions? How would you assess the strength of the link between the medical condition and the cognitive functions? You can think of the strength in terms of how often do these cooccur or how typical it is for a person who has this disorder to have this cognitive problem. What is the most representative/most characteristic disease/disorder of each disease category? What are the most debilitating functions/conditions? Other diseases typical of the category? Is it possible to categorise disorders based on impaired cognitive functions? Is it possible to describe the strength of the link between a disease and a function? Is it possible to link brain areas to specific functions?

12.2.2 Participants Altogether four people were interviewed: two neuropsychiatrists with both clinical and academic experience and two occupational therapists specialising in cognitive impairment. The interviewees were chosen based on their relevant expertise in both academic and clinical settings. Each of them had over 10 years of experience of working with people with cognitive impairment. Participant 1 was a specialist in developmental neuroscience. His area of expertise covered different forms of learning disabilities and included Alzheimer’s disease, autism and depressive disorders as well as birth defects and congenital, developmental intellectual disabilities. He also had an interest in the citizenship of people with learning disabilities and their role in the society. Participant 2 was a neuropsychiatrist specialising in epilepsy and autism. His research interests included neuropsychiatric features and symptoms of neurological diseases as well as the biological processes and brain areas associated with certain neuropsychiatric diseases. Participants 3 and 4 were occupational therapists. They were interviewed simultaneously. One of them was an expert in traumatic brain injury, the other in learning disability.

Jokisuu et al.

118

12.3 Categorisation of Cognitive Impairment The International Classification of Functioning, Disability and Health (ICF) by the World Health Organisation was used as the basis for the identification and grouping of cognitive functions. There are many ways of assessing the functional capability of an individual, with differing definitions of functions and capability, which is why an international well-recognised classification - ICF - was chosen. It is a classification of human functioning and disability, and it covers body functions and structures and their impairments as well as the potential activity limitations and participation restrictions following from them. (WHO, 2001) The classification of cognitive functions presented in ICF did not require many changes to be applicable to this study. Based on the interviews, the categorisation of cognitive functions seems relatively standard, although ‘the challenge is to find categories which cover everything and are recognised by people’ (Participant 3). However, the occupational therapists emphasised that there is no standard way of categorising diseases and functions and there are many different ways of labelling them. In addition, people from different professions interpret the terms in different ways. For example, occupational therapists and doctors understand the terms very differently. The categorisation based on ICF and the categorisation based on the interviews are presented in Table 12.2.

12.4 Categorisation of Medical Conditions In order to get a comprehensive view of the medical conditions with potential impact on cognitive capability, existing and established classifications were used. Two of the most widely used classifications are the International Classification of Diseases (ICD-10) by the World Health Organization and the Diagnostic and Statistical Manual of Mental Disorders (DSM-IV) by the American Psychiatric Association (Mezzich, 2002). These were used together as the basis for developing the categorisation of medical conditions appropriate for this study. While both classifications have been developed as diagnostic aids, there are significant differences between them. ICD-10 has a basic grouping of epidemic diseases, constitutional diseases, local diseases arranged by site, developmental diseases and injuries, whereas DSM-IV focuses on diagnosing mental health issues with some correspondence to the mental disorders of ICD-10 (WHO, 2007; APA, 1994). The categorisation of medical conditions changed considerably as a result of the interviews. As Participant 1 mentioned, the medical classifications are ‘not the most logical way’ of categorising diseases and disorders: ‘Medical conditions can be categorised in a number of different ways, and the way of classifying them does not matter as long as there is some logic to it.’

His suggestion was to have broader categories and present examples of diseases in each category. Table 12.3. describes the changes from the categorisation in ICD10/DSM-IV to the adapted categorisation based on the experts’ opinion.

A Framework for Studying Cognitive Impairment to Inform Inclusive Design

119

Table 12.2. Developing a categorisation of cognitive functions based on ICF (WHO, 2001) Categorisation of Cognitive Impairment in ICF

Categorisation Adapted for the Framework

Perception

Visual perception Auditory perception Tactile perception Watching Listening

Perception

Visual perception Auditory perception Visuospatial perception Olfactory, gustatory and tactile perception

Attention

Sustaining attention Shifting attention Dividing attention Sharing attention

Attention

Focusing attention Shifting attention Dividing attention Sharing attention

Memory

Sensory memory Short-term memory Implicit memory Episodic memory Semantic memory Retrieval of memory

Memory

Short-term memory Long-term memory Retrieval of memory

Thinking

Abstraction Organization and planning Time management Cognitive flexibility Insight Judgment Problem solving Making decisions

Thinking

Pace, form, content and control of thinking Abstraction Organization and planning Time management Cognitive flexibility Insight Judgement Problem-solving Decision-making

Language

Comprehending messages Producing messages Conversation Calculation

Language

Reception of language Expression of language Integrative language functions Calculation

Learning

Copying Rehearsing Learning to read, write Acquiring skills

Learning

Copying Rehearsing Learning to read, write Acquiring skills

Psychomotor functions

Psychomotor control Quality of psychomotor functions Sequencing complex movements

Psychomotor functions

Psychomotor control Quality of psychomotor functions Sequencing complex movements Vestibular functions Proprioceptive function

120

Jokisuu et al.

Table 12.3 Developing a categorisation of medical conditions based on ICD-10 (WHO, 2007) and DSM-IV (APA, 1994) Categorisation of Medical Conditions in ICD-10/DSM-IV

Categorisation Adapted for the Framework

• Infectious diseases, e.g. Tuberculosis • Neoplasms • Diabetes • Dementia • Delirium • Disorders due to psychoactive substance use • Schizotypal disorders • Mood disorders • Neurotic and stress-related disorders • Disorders of adult personality and behaviour • Mental retardation • Disorders of psychological development • Behavioural and emotional disorders • Inflammatory diseases of the central nervous system, e.g. Encephalitis • Systemic atrophies primarily affecting the central nervous system, e.g. Huntington’s disease • Extrapyramidal and movement disorders, e.g. Parkinson’s disease • Degenerative disorders of the nervous system, e.g. Alzheimer’s disease • Demyelinating diseases of the central nervous system, e.g. Multiple sclerosis • Episodic and paroxysmal disorders, e.g. Epilepsy • Polyneuropathies and other disorders of the peripheral nervous system • Cerebral palsy and other paralytic syndromes • Cerebrovascular diseases • Congenital malformations of the nervous system • Head injuries

Disorders of development affecting cognition • Down’s syndrome • Autistic spectrum disorders, e.g. autism, Asperger’s syndrome • Cerebral palsy • Specific learning difficulties, e.g. dyslexia Neurodegenerative disorders • Alzheimer’s disease • Parkinson’s disease • Multiple Sclerosis Acquired brain injuries • Neoplasms • Traumatic brain injury • Infections, e.g. Meningitis, Encephalitis Vascular disorders • Stroke • Cardiovascular accidents Episodic disorders affecting the nervous system • Epilepsy • Delirium • Migraine Illnesses of other systems affecting the brain • Diabetes and other metabolic and endocrine disorders Psychiatric disorders affecting cognition • Schizophrenia • Depression • Bipolar disorder

121

A Framework for Studying Cognitive Impairment to Inform Inclusive Design

12.5 Categories as a Matrix

Psychomotor

Learning

Language

Thinking

Memory

Attention

Perception

Cognitive impairment can be diagnosed through the medical condition causing it. Therefore, building the link between the medical conditions and the cognitive functions they impair is essential. For this purpose, the categorisations described previously were used to create a matrix with medical conditions on rows and cognitive functions in columns (see example in Figure 12.1). The idea was to use the matrix to reveal patterns of impairment characteristic of certain medical conditions by mapping the links between medical conditions and cognitive functions. The mapping was done by assessing the strength of the connections between any given medical condition and the cognitive functions it typically impairs. The strength assessment had a scale of 1-3, with one denoting a weak link and three a strong link. Zero was used to indicate that there is no known link between the medical condition and the cognitive function. This matrix forms the basis of the framework. In this stage of the study, the matrix was used as a tool in the interviews to get a better understanding of the potential usefulness and applicability of the categorisations, to elicit opinions and experiences and to guide the experts in thinking in the field of cognitive impairment as a whole. In addition to getting valuable feedback on the categories, interviews also offered a chance to start learning about the possible patterns linking medical conditions and cognitive impairment. All the participants thought that the matrix was an interesting tool and a new way of approaching the area. In theory there should be a pattern of signs and symptoms evident that identifies each disease; this pattern is used to diagnose the disease. With accurate enough data, patterns could be seen but the data would probably be ‘noisy’ (Participant 2).

Down’s syndrome Dyslexia Autism Brain injury Schizophrenia Figure 12.1. Example of using the framework (from the study data). The darker the shade, the more significant the impairment.

All of the participants highlighted potential challenges of the approach. The main problem would be that the data collected with the matrix would not

122

Jokisuu et al.

necessarily be meaningful due to individual variations and the huge range of severity in conditions: the severity of the impairment depends on the progression of the disease and varies with the stage of the illness. Another challenge is those medical conditions, for example stroke and brain injury, which could cause problems in all of the cognitive functions. As Participant 2 emphasised: ‘It must be remembered that the matrix would only provide an average, a good approximation perhaps but very variable. […] The biggest challenge is probably going to be the immense amount of information that needs to be extracted, synthesised and summarised.’

12.6 Using the Framework After the initial series of interviews, a brief validation study was conducted to ensure the categorisation was as accurate as possible. Of the four interviewees, three reviewed the adapted categorisations to make sure they correctly reflected the participants’ views. To get a better idea of how the framework would work in practice, Participant 2 filled in the matrix based on his professional experience; meaning, he assessed the significance of each medical condition in terms of causing cognitive impairment (Figure 12.1). His experience was that it takes a considerable amount of time and effort to complete the matrix. The main advantage of the framework as an information collection tool lies, in his opinion, in its flexibility: it is possible to look at cognitive impairment as a whole or drill down to details of specific medical conditions. ‘It would help designers to understand where the individual’s strengths are by understanding the weaknesses and where tailoring of a product would be advantageous. The key is to find a useful way of conveying the information and not get impeded by the complexities.’

After this initial study, the framework has been used to collect information regarding various types of cognitive impairment and the medical conditions which cause them. This work is still ongoing but the framework seems to work quite well in that it allows a multifaceted view into the phenomenon of cognitive impairment: a general overview as well as more detailed information, as needed. The framework can also be used to identify the most important areas of cognitive impairment; that is, the most common and severe types of impairment which designers should focus on. An example of the framework is presented in Figure 12.1 in which the darker shades indicate areas where impairment is particularly significant. As a result, these would be the areas that designers should pay attention to. The example is discussed in more detail in Jokisuu et al. (2011). At this stage, however, the data is incomplete, as analysis is still ongoing. Incorporating the expert opinion of more participants into the framework will improve the reliability and validity of the results. Similarly, differences in participants’ assessments will highlight areas requiring further examination.

A Framework for Studying Cognitive Impairment to Inform Inclusive Design

123

12.7 Conclusions The framework presented in this paper can be used to systematically analyse the various types of cognitive impairment and the medical conditions that cause them. It is based on medical and disability classifications which were modified according to experts’ opinions. Through interviewing the experts, it became evident that there is a need for an information tool which would translate the sometimes complicated medical material into a more accessible format that would be relevant to designers. However, there are several potential problems with the framework which were highlighted during the study. The individual differences and varying levels of impairment, particularly in progressive diseases, are extremely difficult to capture with the framework. Terminology is not clear and the definitions of functions and disorders are not understood the same way, even when a well-established international classification is used. It should also be noted that impairment in a function does not necessarily result in an inability to perform an action; people develop compensatory strategies to cope in everyday life. These are not captured by a somewhat simplistic theoretical model. In addition, at this first stage of the study, only four people participated in the development of the framework, which might limit the validity of the work; however, this issue will be addressed in later stages when the framework is further developed and applied in a larger-scale study. The strength of the framework is that it offers designers a multifaceted view of the complexity of cognitive impairment: an overview of the general issues as well as the details and background. In the next stages of the study, the framework will be used to collect information that will provide a more complete picture of cognitive impairment, including examples of assistive technology and existing design guidance, stories of people living with cognitive impairment as well as a general idea of the number of people affected by various types of impairment. The framework will then be further developed with designers to see which elements are most useful and applicable to them.

12.8 References APA (1994) Diagnostic and statistical manual of mental disorders, 4th edn. American Psychiatric Association, Washington DC, WA, US Carlsson G, Iwarsson S, Stahl A (2002) The personal component of accessibility at group level: Exploring the complexity of functional capacity. Scandinavian Journal of Occupational Therapy, 9(3): 100-108 Centre for Universal Design (1997) Universal Design Principles. Available at: http://www.design.ncsu.edu/cud/about_ud/udprinciples.htm (Accessed 21 October 2011) Cohene T, Baecker R, Marziali E, Mindy S (2007) Memories of a life: A design case study for Alzheimer’s Disease. In: Lazar J (ed.) Universal usability: Designing computer interfaces for diverse users. John Wiley & Sons Ltd, Chichester, UK Gartland D (2004) Considerations in the selection and use of technology with people who have cognitive deficits following acquired brain injury. Neuropsychological Rehabilitation, 14(1/2): 61-75 Ghajar J (2000) Traumatic brain injury. The Lancet, 356(9233): 923-929

124

Jokisuu et al.

Groome D, Brace N, Dewart H, Edgar G, Edgar H, Esgate A et al. (2006) An introduction to cognitive psychology. Processes and disorders, 2nd edn. Psychology Press, Hove, UK Jokisuu E, Langdon PM, Clarkson PJ (2011) Modelling cognitive impairment to improve universal access. In: Stephanidis C (ed.) 6th International Conference on Universal Access in Human-Computer Interaction (UAHCI 2011), Orlando, FL, US LoPresti E, Mihailidis A, Kirsch N (2004) Assistive technology for cognitive rehabilitation: State of the art. Neuropsychological Rehabilitation, 14(1/2): 5-39 Mezzich JE (2002) International surveys on the use of ICD-10 and related diagnostic systems. Psychopathology, 35(2-3): 72-75 W3C (2008) Web Content Accessibility Guidelines WCAG 2.0. Available at: http://www.w3.org/TR/WCAG20/ (Accessed 21 October 2011) WHO (2001) International classification of functioning, disability and health. World Health Organization, Geneva, Switzerland WHO (2007) International statistical classification of diseases and related health problems, 10th edn. Available at: http://www.who.int/classifications/apps/icd/icd10 (Accessed 21 October 2011) WHO (2008) The global burden of disease: 2004 update. Available at: http://www.who.int/entity/healthinfo/global_burden_disease/GBD_report_2004update_ful l.pdf (Accessed 21 October 2011)

Chapter 13 Interactive Error Correction Using Statistical Language Models in a Clientserver Interface for Editing Mathematical Text D. Attanayake, G. Hunter, E. Pfluegel and J. Denholm-Price

13.1 Introduction Learning and using mathematical notation poses particular difficulties for people with various disabilities, partly due to its wide range of symbols and rather complicated layout. These pose great challenges, often affecting the educational and career opportunities of people who are visually impaired or have limited (or no) use of their hands or arms. Assistive systems to alleviate these difficulties would be of considerable benefit to such groups of people. In this paper, we provide an overview of the development of the new clientserver architecture of TalkMaths, a web-based speech interface system for dictating and editing mathematical text in electronic documents. First, we describe the current web-based architecture of TalkMaths and its new multi-modal features. We then discuss the process of building a data corpus by “crawling” through source codes of carefully selected web sites containing mathematical equations at the appropriate level. Each equation in this data set is then converted to the corresponding sequence of mathematical ‘keywords’ in order to use the data to build language models based on unigram, bigram and trigram statistics (Young, 1996). Finally we explain how these language models will be used within the new version of TalkMaths to give it automatic predictive or corrective abilities to improve its performance, usability and user satisfaction levels.

P. Langdon et al. (eds.), Designing Inclusive Systems, DOI: 10.1007/978-1-4471-2867-0_13, © Springer-Verlag London 2012

125

126

Attanayake et al.

13.2 Related Systems So and Watt (2006) used the predictive power of statistical language models in their system for recognising hand-written mathematical symbols using optical character recognition. They extracted mathematical expressions from a public ePrint server, arXiv.org, which covers a broad range of mathematically based subjects at research level, and analysed statistics on the most commonly used expressions. The key difference between our work and that of So and Watt is that even though the domain (mathematics) is essentially the same in both, the input methods and related problems are quite different as we are dealing with spoken input as opposed to hand-written characters. Also, the level of mathematics in question differs significantly between the two pieces of work, as we focus on relatively elementary levels of “high school” mathematics whereas So and Watt consider research level mathematics . There are various other existing systems that enable the dictation of mathematics such as MathTalk (Stevens et al., 1997), Math Speak & Write (Guy et al., 2004) and CamMath (Elliott and Bilmes, 2007). The first two use as their underlying system commercially available mathematical editors, which are not freely available. Also they offer somewhat limited methods for navigating and editing each mathematical expression. The Math Speak & Write system is freely available, although it only allows a rather small set of mathematical vocabulary and seems to require use of somewhat non-intuitive editing paradigms. Bernareggi and Brigatti (2008) developed a system for speech input of mathematics targeted at blind users, but it is currently not widely available and only supports spoken input in Italian. Hanakovič and Nagy (2006) also propose a speech input mathematical editor for the visually impaired. However, their system requires use of XHTML voice (X+V) technology which is restricted to a limited range of web browsers, such as the Opera web browser.

13.3 Current TalkMaths Architecture The TalkMaths system is primarily aimed at helping people with various disabilities (Wigmore et al., 2010a), particularly those with limited or no use of their hands or arms. The current TalkMaths prototype system uses an “Application Proxy architecture” (Attanayake et al., 2011), based on a client-server model, and is a web-based editor for mathematical text. Our client application is a Mozilla Firefox browser and requires a speech recognition engine to be installed on the client machine. A web server and a Google apps service are used to realise web logic (controlling the behaviour of the application independent from any speech input) and speech logic (parts of the logic that deal with interpreting input in the form of speech commands) respectively. In contrast to the earlier desktop version of TalkMaths, the new version now benefits from several advantages offered by web-standards (well-established technologies for creation and interpretation of web-based content). First, and perhaps most importantly, it offers easy access to a large number of users and, secondly, computationally expensive processes like parsing a sequence of commands can be carried out on a fast web server

Interactive Error Correction Using Statistical Language Models

127

(Gruenstein et al., 2008). This web-based architecture also allows TalkMaths to store an archive of a user’s interactions with the system, especially the speech commands used. This can be used for a variety of purposes. Users can retrieve work they have previously done using the system, without being tied to a particular machine. The archived data can also be used for analytical purposes, to build a speech data corpus that can be used both to improve the general centralised statistical language model (see Section 13.4 below) and/or to develop specialised language models specific to particular topics or individual users. TalkMaths is also now a multi-modal system that can accept input by speech and/or keyboard and mouse.

13.4 Building Statistical Language Models from Web Data Sources Our approach to building a substantial mathematical database has been evolving continuously over the last few years. Our first attempt at obtaining data on the ways in which mathematical equations are spoken was based on part of the British National Corpus (BNC) which consisted of transcriptions of conversations from school and college mathematics classes (Wigmore et al., 2009; Wigmore, 2011). The vocabulary (number of different words) found in this dataset was 4,355 (the total number of words was 123,821) and the perplexity of the statistical model obtained was rather high, due to both the relatively large vocabulary and the high proportion of non-mathematical words and “chat” in the conversations, making the data somewhat unpredictable. Our second attempt at modelling mathematical text was to manually populate a dataset using trigonometric equations from some mathematical text books (Wigmore et al., 2010b; Wigmore, 2011). Not surprisingly, both the vocabulary (102 distinct words) and the dataset overall (the total number of words was 7,857) were smaller than for the BNC-based data. The statistical language models built from the second (trigonometry) dataset had much lower perplexity scores, indicating that this data was more predictable. Interestingly, consistent with earlier studies (Hunter and Huckvale, 2006), the results of these first two attempts at creating statistical language models from mathematical material confirmed that, for data of a given type, if the training dataset is increased in size, the perplexity of the resulting language model decreases and hence its predictive power is increased. With this in mind, in the current study we have attempted to create a much larger high quality dataset of mathematical expressions on which to base new statistical language models. We identified and “crawled” a handful of public “tutorial” web sites containing mathematical equations at a similar level of complexity to that which TalkMaths is currently capable of processing. For this work, we developed a “web-crawler” that can identify LaTeX or MathML content within the source code of the web site. This mathematical content is then extracted into a database. A “filtering script” was then applied to remove display instructions from LaTeX code and illegal characters from the equations. Finally, a LaTeX/MathML to “spoken mathematical expressions” converter was

128

Attanayake et al.

designed and developed using the yapps2 (Patel, 2009) parser generator. The converter introduced additional keywords “begin” and “end” for denoting subsections (such as fractions or square roots) within the “linearised spoken descriptions” of each mathematical expression, to create a word string which would be identical to the “correct” way in which a TalkMaths user would dictate that expression. We then used the Carnegie Mellon Statistical Language Modeling (CMU SLM) Toolkit (Clarkson and Rosenfeld, 1997) to build statistical language models based on an initial sample of our corpus (3,194 mathematical expressions containing a total of 61,479 words with a vocabulary of 100 words). The process of building these models was analogous to that used in our earlier studies, which have been described elsewhere (Wigmore et al., 2010b; Wigmore, 2011). Table 13.1. Cross-validation perplexity calculations on statistical language models of 3,194 spoken mathematical expressions using CMU Language Modelling Toolkit

subsets [1-9]

subset10

54907

6572

7.07

Vocabulary (words) 100

subsets[1-8 & 10]

subset9

54968

6511

7.17

100

subsets[1-7 & 9-10]

subset8

55294

6185

7.11

100

subsets[1-6 & 8-10]

subset7

55688

5791

7.31

100

subsets[1-5 & 7-10]

subset6

55172

6307

7.25

100

subsets[1-4 & 6-10]

subset5

55597

5882

7.74

100

subsets[1-3 & 5-10] subsets[1-2 & 4-10]

subset4 subset3

55340 55805

6139 5674

7.01 7.65

100 98

subsets[1 & 3-10]

subset2

55177

6302

7.53

100

subsets[2-10]

subset1

55363

6116

7.02

100

Training set

Test Set Training words Test words Perplexity

Initial experiments to investigate the quality of this data confirmed the trends noted by previous studies (Hunter and Huckvale, 2006; Wigmore et al., 2010b; Wigmore, 2011). The results from this new study are summarised in Table 13.1. Interestingly, the predictive power of the models based on the sample tested was better than those of both our earlier studies (Wigmore et al., 2010b; Wigmore, 2011). There are several possible reasons for this improvement. The vocabulary is still relatively small and the training datasets used in our latest study are considerably larger than the trigonometric dataset used in Wigmore et al. (2010b). Perhaps the most important reasons could be the higher quality and increased amount of training data. We will describe below how we will use this statistical model in the context of our new web-based TalkMaths architecture.

Interactive Error Correction Using Statistical Language Models

129

13.5 Speech Error Correction In our usability studies (Wigmore et al., 2010a), we noted that commercial speech recognition engines are not primarily designed for the dictation of mathematics. The speech recogniser’s error rates are normally higher for dictating mathematics in contrast to “everyday” natural language. However, together with extensive user training and restricting the vocabulary, this problem can be substantially solved. While some researchers try to correct recognition errors at the parser level using error recovery strategies, others rely solely on the language models associated with the speech recognition engine to solve the problem. Since the users should themselves know exactly what they meant to dictate in the first place, we believe a collaborative effort between the user and the system, in the form of an interaction that is natural to the user, is required to carry out a repair on an erroneous input. Studies in Suhm, et al. (1996) also discuss interactive error recovery strategies. In addition, They deduce “user preference factors” in interactive error recovery such as the “naturalness of” the interaction, interpretation accuracy and the time it takes for a user to provide an input and for the system to interpret it. We will address these issues in our error correction environment to minimise the time and effort a user has to spend to complete a correction with ease.

13.6 Design of Interactive Prediction Mechanism Since speech recognition is carried out at the client side, the statistical language model can be used to predict what will be said next, preferably also at the client side. To realise this, the statistical language model is transmitted to the client as a one-off transaction when the client first connects to the web server. This model is then used by our companion application, residing in the client machine, which captures and passes recognition results from the speech recognition engine on to the browser. If the statistical language model produces some alternative predictions such as a ranked list of most likely alternatives to or continuations of the recognised word sequence, these will be sent to the remote server for parsing, and the results sent back to the client in the form of MathML expressions, which will be displayed as “alternative” mathematical equations in two-dimensional format on the screen. Note that the original utterance from the speech recognition engine will always be transmitted to the server for parsing and consequently displayed to the user as the “primary” mathematical expression on the screen. The user can subsequently check the appearance of the primary expression and, if this is not what he originally intended, he can correct it, either by editing the primary expression or by selecting the appropriate choice from the set of displayed alternatives, by speech or using keyboard and mouse. The process of our “interactive predictive alternative” system is illustrated in Figure 13.1 overleaf.

130

Attanayake et al.

SLM at server

Send SLM to client

Wait for next utterance or instruction

Display maths (including alternatives if present)

Maths?

Correction or navigation

No

Yes

Parse main utterance or instruction

No

Alternative(s)?

Yes Parse alternative(s)

Figure 13.1. Flowchart of TalkMaths “predictive alternative” system

13.7 Conclusions and Future Work We believe our “predictive alternative” system (described above) driven by a statistical language model, together with natural editing strategies, can significantly improve user satisfaction levels for TalkMaths in the future. Furthermore, by reducing speech recognition and parser errors in TalkMaths, disabled people such as repetitive strain injury (RSI) sufferers should find TalkMaths to be an intuitive

Interactive Error Correction Using Statistical Language Models

131

and useful tool. We will carry out an empirical evaluation of our multi-modal system with both disabled and non-disabled users in order to answer several key scientific questions relating to its design and usability. We are also investigating more sophisticated parsing methodologies and editing paradigms with a view to making further improvements to TalkMaths.

13.8 References Attanayake D, Hunter G, Pfluegel E, Denholm-Price J (2011) Architectures for speechbased web applications. In: Proceedings of the International Conference on Semantic EBusiness and Enterprise Computing (SEEC 2011), Kingston upon Thames, UK Bernareggi C, Brigatti V (2008) Writing mathematics by speech: A case study for visually impaired. In: Proceedings of the 11th International Conference on Computers Helping People with Special Needs (ICCHP 2008), Linz, Austria Clarkson P, Rosenfeld R (1997) Statistical language modeling using the CMU-Cambridge Toolkit. In: Proceedings of the 5th European Conference on Speech Communication and Technology (Eurospeech 1997) Elliott C, Bilmes JA (2007) Computer based mathematics using continuous speech recognition. In: Proceedings of the CHI 2007 Workshop on Striking a C[h]ord: Vocal Interaction in Assistive Technologies, Games, and More, San Jose, CA, US Guy C, Jurka M, Stanek S, Fateman R (2004) Math speak & write, a Computer program to read and hear mathematical input. Electrical Engineering and Computer Sciences Department Technical Report, University of California, Berkeley, CA, US Gruenstein A, McGraw I, Badr J (2008) The wami toolkit for developing, deploying, and evaluating web-accessible multimodal interfaces. In: Proceedings of the 10th International Conference on Multimodal Interfaces (ICMI 2008), Chania, Greece Hanakovič T, Nagy M (2006) Speech recognition helps visually impaired people writing mathematical formulas. In: Proceedings of the 10th International Conference on Computers Helping People with Special Needs (ICCHP 2006), Linz, Austria Hunter G, Huckvale M (2006) Is it appropriate to model dialogue in the same way as text? A comparative study using the British National Corpus. In: Proceedings of the 2006 European Modelling Symposium, London, UK Patel A (2009) Parsing with yapps. Available at: http://theory.stanford.edu/~amitp/yapps/ (Accessed 25 August 2011) So CM, Watt SM (2006) Determining empirical properties of mathematical expression use. In: Proceedings of the 4th International Conference on Mathematical Knowledge Management (MKM 2005), Bremen, Germany Stevens RD, Edwards, ADN, Harling, PA (1997) Access to mathematics for visually disabled students through multimodal interaction. Human-Computer Interaction 12(1): 47-92 Suhm B, Myers B, Waibel A (1996) Designing interactive error recovery methods for speech interfaces. In: Proceedings of the CHI 96 Workshop on Designing the User Interface for Speech Recognition applications, SIGCHI, Vancouver, Canada Wigmore A (2011) Speech-based creation and editing of mathematical content, PhD, Kingston University, UK Wigmore AM, Hunter G, Pfluegel E, Denholm-Price J, Binelli V (2009) “Let them TalkMaths!” Developing an intelligent system to assist disabled people to learn and use mathematics on computers through a speech interface: the TalkMaths and VoiceCalc

132

Attanayake et al.

systems. In: Proceedings of the 5th IEEE/AAAI International Conference on Intelligent Environments, Barcelona, Spain Wigmore AM, Hunter G, Pfluegel E, Denholm-Price J, Colbert M (2010a) “TalkMaths Better!” Evaluating and improving an intelligent interface for creating and editing mathematical text. In: Proceedings of the 6th International Conference on Intelligent Environments, Kuala Lumpur, Malaysia Wigmore AM, Pfluegel E, Hunter G, Denholm-Price J, Colbert M, Attanayake D (2010b) Evaluating and improving the TalkMaths speech interface for dictating and editing mathematical text. In: Proceedings of 5th European Workshop on Mathematical and Scientific E-Contents, Salamanca, Spain Wigmore AM (2011) Speech-based creation and editing of mathematical content. PhD Thesis, Kingston University, London, UK Young S (1996) Large vocabulary speech recognition: A review. IEEE Signal Processing Magazine, 13(5): 45-57

Chapter 14 Understandable by Design: How Can Products be Designed to Align with User Experience? A. Mieczakowski, P.M. Langdon, R.H. Bracewell, J.J. Patmore and P.J. Clarkson

14.1 Introduction Understanding users increases the likelihood that the final designed product will meet the needs of heterogeneous people (Kouprie and Sleeswijk Visser, 2009). However, the process of learning to understand users and their experiences requires qualitative research (Mattelmäki and Battarbee, 2002) and a structured investment of time that can be achieved through the development of a manageable design plan in the early stages of design (Yang and Epstein, 2005). Although the fields of accessibility and usability have made significant advances in the last decade in facilitating the design of products and services that satisfy the needs of different users and are easy to use, the majority of design companies still fail to acknowledge users’ needs early on in the design process (Gulliksen et al., 2003). This is because of, among other things, tight schedules and a limited budget (Dong, 2005). Studies by Ricability (2001) show that products designed for people with reduced capability are also easier for everyone else to use. Many products, however, are designed as if every user were an agile 25-30 year old professional (Benktzon, 1993) and, therefore, are largely inaccessible and unusable for less capable users (Keates and Clarkson, 2003). User-centred design approaches such as Inclusive Design have been developed in order to help designers expand the boundaries of product usage for as many people as possible by iteratively attuning product design to the needs of heterogeneous users from the beginning of the design process (Keates and Clarkson, 2003). There are moral and financial benefits associated with application of the Inclusive Design philosophy, as products which are designed in line with users’ capabilities are believed to improve customer satisfaction (Hewer and

P. Langdon et al. (eds.), Designing Inclusive Systems, DOI: 10.1007/978-1-4471-2867-0_14, © Springer-Verlag London 2012

133

134

Mieczakowski et al.

James, 1998), which in turn allows companies which value good design to achieve high growth (Keates and Clarkson, 2003). The work presented in this paper discusses the ways in which design companies can effectively adopt the inclusive design ethos and addresses the development of a modelling approach for supporting designers in determining early stages in the design process, whether specific product features evoke similar understanding and responses among the users and the designers of those features.

14.2 Inclusive Product Design To design more inclusively, designers must “understand the users, their cognitive behaviour, attitudes and the characteristics of their work tasks” (Gulliksen et al., 2003). Likewise, Preece et al. (2002) argue that a product should be based on users’ needs, requirements and the actions that they will perform when performing tasks and that it should be tested iteratively. Chamberlain et al. (2011) cite the case of a large global communications company which serves customers in 170 countries, BT plc, to illustrate how usercentred design can be effectively adopted in the development of products and services. The study by Chamberlain and colleagues is based on five recommendations: (1) it is vital for senior management to support the principles of user-centred design and identify a champion or champions who will promote this change within an organisation; (2) this design approach will be adopted more successfully if it is positioned as a design that improves the lives of everyone and advocates that small changes can make a big difference; (3) rethinking the design and accommodating it to the needs of heterogeneous users makes for good social responsibility and lucrative commercial decisions; (4) the adoption of inclusive design requires guidelines and tools for supporting the act of designing, as well as providing training on how to use different support materials; and (5) the most effective way of designing intuitive interactions in products and services is by having direct contact with a range of different users. The subsequent sections of this paper focus on the fourth element of the above recommendations for adopting the principles of inclusive design, i.e. they describe the development of a modelling approach for helping designers align their understanding and use of products with that of users.

14.3 Human Understanding and Experience with Designed Products Human understanding is often perceived as a model and the theoretical construct of a mental model has, for over sixty years, been studied at length by researchers from a wide range of disciplines (Craik, 1943; Johnson-Laird, 1983; Norman, 2002; Payne, 2008). Mental models are believed to be heavily influenced by previous knowledge and experience with similar products (Langdon et al., 2007).

Understandable by Design

135

One of the biggest advocates of mental models, Norman (2002) posits that accessible and usable products can be designed by means of matching the designer’s conceptual model of a product with the users’ mental models of that product through the use of the product interface. However, so far there is a significant lack of a simple-to-use support method which would guide designers in representing and comparing their understanding and intended use of a given product with the users’ understanding and actual use of that product. The aim of such a support method would ideally be to help designers “reduce the number of actions and decisions that users have to make in order to reach their goals” (Langdon and Thimbleby, 2010).

14.4 Modelling Approach for Aligning Design Intent with User Experience Since designers are in general comfortable and familiar with interpreting abstract visual representations such as models and diagrams because they are often involved in their daily work (Andreasen, 1994), the new inclusive design support tool proposed in this paper is a data model containing information gained from user observations and engineering and design insights. The key requirements for the proposed modelling approach are: (1) to be used early in the design process to identify potential problems and increase the chances of producing a more inclusive product; (2) to assist designers in gathering information about users’ understanding and behaviour in relation to products and comparing it with their own design intent; (3) to be easy to understand, implement and use; and (4) to be a semantic modelling representation capturing key pieces of information about users’ understanding and use of a given product. This paper proposes the Goal-Action-Belief-Object (GABO) modelling approach for designers to enable the assessment and comparison of designers’ and users’ understanding and usage of everyday products. This approach works on the premise that the contents of people’s minds (i.e. their knowledge, theories and beliefs) should be explored by designers in order to better understand users’ behaviour in relation to products (Payne, 2008). However, it should be noted that this approach does not focus on the structure of the human mind and does not represent information about the actively changing parts of mental models. Instead, it represents the current and static contents of users’ mental models during interaction with a given product. The GABO approach encompasses both elicitation and representation (Klimoski and Mohammed, 1994). Fundamentally, the GABO approach aims to encourage designers to pay greater attention to users’ understanding and use of products through representing: (1) goals that users want to achieve during interaction with products; (2) actions that they exert on product interfaces; (3) beliefs about product features that they bring from previous interactions with other products to interactions with new products; and (4) understanding of the impact of their actions on a product’s functional objects.

136

Mieczakowski et al.

During the use of the GABO approach, designers are recommended to create three types of model: (1) an engineering model of a given product indicating how its different parts interact with one another; (2) a designer model of a given product (in order to be compared with the engineering model to see what features should be mounted on the top of the underlying functional parts); and (3) a number of individual user models encompassing how users understand and use different product features (in order to be compared with the designer model). The designer model and the user models are of the highest importance as they are to be compared for compatibility, while the engineering model acts as a reference to the designer model. During the trial of the GABO approach, the designer, engineering and user models constructed using the approach have been tested on an example of a coffee maker with an elaborate interface shown in Figure 14.1.

Figure 14.1. A coffee maker with an elaborate interface

This coffee maker has a dual functionality: its left side makes an espresso type of coffee and its right side is for making a café type of coffee. The espresso side has such main features as: a start/stop button, a 1 and 2 cup espresso selection button, a button for steaming hot water and milk, a blue LED display (shared with the café side), a scroll button for setting time, an ‘OK’ button (shared with the café side). The café side has the following main parts: a start/stop button, a button for selecting the programmable mode of the coffee maker, a toggle for selecting strong or normal café strength, a blue LED display (shared with the espresso side), a scroll button for setting time and choosing the automatic mode of the coffee maker (shared with the espresso side) and an ‘OK’ button. Thirty users, two designers and two engineers were consulted during the trial with this coffee maker. Information from the users was obtained by observing them using the coffee maker. Information from designers and engineers was gathered from discussions based around a semi-structured interview technique. The engineers and designers consulted worked for the same design company, were

Understandable by Design

137

highly educated and had over two years of experience in product design. The thirty users were of different ages, capabilities and levels of education. Based on the information they provided, the engineering, designer and user models were created by the use of the Design Rationale Editor (DRed) software (Aurisicchio and Bracewell, 2009). The semantic coding language used to create the designer and user models was composed of different verb and noun combinations depending on which type of element—goal (verb + noun), action (verb+ing + noun), belief (noun + to+verb), object (noun)—it was describing. This is in line with the work of Andreasen (1994) who argues that “the modelling needs to be made semantic to function like a language”. Subsequently, the constructed GABO models were compared for similarities and differences to the designer model and each individual user model using an appropriate algorithm (discussed as follows). The GABO approach stipulates that any two models (a designer model and an individual user model) can be compared based on: (1) presence of the same nodes in the two models and (2) connectivity between two given nodes in the two models. The comparison procedure can be carried out using an algorithm from set theory that measures similarity between graphs with common node and edge sets (Goldsmith and Davenport, 1990). This algorithm is used for measuring both the presence of nodes and the connectivity between nodes in the designer model and the individual user models, with the designer model acting as the standard model against which each user model is checked for compatibility. The assumption is that by using the GABO approach’s algorithm, designers will be able to make close estimates of the compatibility of their intended goals, actions, beliefs and objects regarding product usage with the goals, actions, beliefs and objects of heterogeneous users.

14.4.1 The GABO Algorithms for Checking Presence and Connectivity of Nodes The algorithms for checking the presence and connectivity of nodes in the designer model and each individual user model are both inspired by the correlation coefficient, which assesses the shared number of nodes/edges between two sets of nodes/edges standardised by a measure of the total number of nodes /edges from both sets. The presence algorithm and the connectivity algorithm consist of the following steps: (1) examine the designer model (D) and enumerate all nodes/edges sequentially; (2) examine an individual user model ( U n ) for each node/edge and identify, by checking the type of each node and keyword, the type and keywords of the two nodes it connects to, whether it has previously been enumerated in the designer model (D): if not, then allocate a new number to it following the numerical sequence from the designer model (D); (3) count the nodes/edges shared between the designer and user sets of nodes/edges; (4) count the number of nodes/edges in the designer model; (5) divide the number of shared nodes/edges between the two sets of nodes/edges by the total number of nodes/edges from the designer model; and (6) add any nodes/edges that are unique to an individual user model.

Mieczakowski et al.

138

The

presence

algorithm

is: P ( D, U ) = (| N D ∩ N U | / | N D |) + UN U

where P = value between 0 and 1, where • 0 = total incompatibility between nodes in the designer and user models • 1 = 100% compatibility between nodes in the designer and user models • Any value between 0 and 1 = the degree of compatibility between nodes in the designer and user models (e.g. if P ( D,U ) = 28/88, then the compatibility level = 0.31 or 31%) D = designer model U = user model N D = the set of nodes in the designer model

N U = the set of nodes in the user model N D ∩ N U = the set of all nodes that are members of both N D and N U UN U = the set of unique nodes in the user model. The presence algorithm assumes that two nodes are the same (one from the designer model and one from the user model) if they belong to the same element type, for instance the belief element, and contain the same semantics. The implementation of the presence algorithm in DRed uses the transclusion function (Kolbitsch and Maurer, 2006). It should be noted that no distinction is made in the weightings of certain types of nodes and edges as the approach only checks for the presence or lack of similar concepts in the understanding of designers and users and does not estimate the significance of different aspects of their understanding. In other words, each node or edge present is allocated the same number. In a similar fashion, each missing node or edge has the same number deducted. The connectivity algorithm is: C ( D, U ) = (| E D ∩ EU | / | E D |) + UEU where C = value between 0 and 1, where • 0 = total incompatibility between edges in the designer and user models • 1 = 100% compatibility between edges in the designer and user models • Any value between 0 and 1 = the degree of compatibility between edges in the designer and user models (e.g. if C ( D,U ) = 30/114, then the compatibility level = 0.26 or 26%) D = designer model U = user model E D = the set of edges in the designer model

EU = the set of edges in the user model E D ∩ EU = the set of all edges that are members of both E D and EU UEU = the set of unique edges in the user model. The connectivity algorithm assumes that two edges are equal if they join two nodes that are exactly the same element-wise and semantically in the designer model and the user model.

Understandable by Design

139

The assumption with regard to the comparison procedure of the GABO approach is that it requires both computer assistance and human input. Computer assistance is needed for comparing the count of the GABO approach’s elements (nodes and edges) and generating a value between 0 and 1, which indicates the degree of presence and connectivity compatibility of each individual user model with the designer model. Human input is required for creating data structures for each model, comparing the engineering model with the designer model, as these two types of models are structured in a different way, and verifying at the end of the automatic comparison whether the elements in the designer model and the user models have been compared correctly by the software system. Once the GABO comparison algorithm was developed, it was programmed into DRed software using the TCL scripting language by DRed’s proprietor, Rob Bracewell, based on the research requirements. This code enabled a comparison of compatibility between the designer model and thirty individual user models of any product in the DRed software (discussed as follows).

14.4.2 Usage Trial of the GABO Modelling Approach During the trial of a coffee maker with an elaborate interface, distinct differences were found between designers’ and users’ understanding and usage of the function for activating coffee making. In particular, to activate the coffee maker users first had to set a clock and all the participants failed to envisage that a clock would be the first feature to be used on the coffee maker interface. Because of this cumbersome design, two participants were not able to activate the coffee making function, and the rest spent a significant amount of time (an hour was the longest) trying to figure it out by pressing all buttons on the interface including buttons on the espresso side. Another difficulty that most of the people had was with finding the position of the café water chamber. Since the coffee maker is of dual functionality, designers thought that a logical solution would be to design a water chamber for each of the functional sides. However, most users assumed that there was only one water chamber. Four users could not locate the water chamber on the café side when making filter coffee and poured water in the water chamber on the espresso side. Moreover, one participant poured water over the filter holding the coffee grounds, which resulted in all the water dripping out. Figure 14.2 shows a comparison screen produced at the end of the comparison procedure of the designer and individual models of thirty users of a complex-to-use coffee maker in the DRed software. This figure shows scores of compatibility for presence of nodes in the designer and thirty user models and scores of compatibility for connectivity between nodes in the designer and thirty user models. In addition, it shows a count of nodes and edges that are unique to individual user models. The instances of unique nodes are: new goal (NG), new action (NA), new belief (NB), new object (NO), repeated node (REP), rejected node (REJ) and misunderstood function (MF). The instances of unique edges are: new edge (NE) and repeated edge (RE). Overall, Figure 14.2 indicates that the average index of compatibility for the presence of nodes is approximately 86% and most users had a substantial number of unique nodes as they had difficulty with figuring out that a clock function had to be set before the coffee making function could be activated.

Mieczakowski et al.

140

NODES

Unique Nodes

EDGES

Unique Edges

Figure 14.2. GABO evaluation of a coffee maker with an elaborate interface

In addition, since this coffee maker was of dual functionality, when most users were confused about the functionality of the café side they moved over and used the functionality of the espresso side thinking that it would help them activate the function on the café side. Hence an increased number of new actions, new beliefs, repeated nodes and rejected nodes. The overall score for the presence of nodes is about 85% and there are large numbers of new edges and repeated edges for linking the increased number of nodes. For example, the model of user 7 scored 68% on the compatibility scale for the presence of nodes and it contained eight new actions, three new beliefs, twenty-five repeated nodes and ten rejected nodes. It needs noting that this user was not able to activate the coffee making function and gave up after having pressed all interface buttons with no expected feedback from the coffee machine. Although this user was not able to make coffee using a complex coffee maker, the overall count of compatibility for this user is relatively high as this person, while employing a ‘trial and error’ method, managed to use a number of relevant interface features. A high count for new actions, new beliefs, repeated nodes and rejected nodes is due to this user pressing a large number of unrelated buttons to activate the coffee making function. Regarding the score of compatibility for the presence of edges, this user scored 61% and her/his model contained forty new edges and twelve repeated edges for connecting eight new actions, three new beliefs, twenty-five repeated nodes and ten rejected nodes.

14.4.3 Evaluation of the GABO Modelling Approach The usefulness and effectiveness of the GABO approach was evaluated with eight designers from a range of small and large organisations based in the UK during two five hour workshop sessions.

Understandable by Design

141

Designers, aged between 29 and 52, were asked to work on two redesign tasks, one of which required them to redesign a complex-to-use interface of a household product (either a toaster or a coffee maker) using a method of choice and the other task required them to redesign an interface of one of the aforementioned two products using the GABO approach. When the tasks were completed, each designer was asked individually to fill out an evaluation questionnaire composed of a number of quantitative and qualitative questions. Overall, the designers marked, on average, point 5 on a 7-point scale indicating how useful the GABO approach was in identifying and capturing users’ understanding and the problems users encounter during product use. This procedure was mirrored when investigating designers’ opinion regarding their understanding of product functionality compared with the understanding of users, the result being point 5.5 on the scale. Likewise, indicating ease-of-use, the designers on average gave the GABO approach a score of 4.3. In addition, five designers believed that the GABO approach helped them to produce a better design than their alternative method of choice, while three designers said that they would need more time to use the GABO approach to determine as to whether it was better or worse than the alternative method. Furthermore, it should be noted that there was general fear among most evaluators that the GABO approach would require a lot of time and effort in training and actual usage to prove useful and they would not always be able to use it to its full potential as most projects are time and cost restricted.

14.5 Conclusions This paper discussed the ways in which companies can effectively adopt the principles of inclusive design and the development of a modelling approach for helping designers align their understanding and use of products with those of users. Since there seems to be a gap in the availability of an easy-to-use and pragmatic technique for representing and comparing designers’ and users’ understanding and usage of everyday products, this paper proposes the GABO approach that aims to bridge that gap for designers. The GABO approach consists of four stages in which designers need to: (1) analyse the engineering model of a product to better understand how different product parts interact with one another; (2) create a designer model of that product using appropriately annotated GABO elements and compare it with the engineering model to see what features should be mounted on the top of the underlying functional parts; (3) investigate how different individuals understand and use product features, create several individual user models using the GABO elements annotated in the same semantic style as their counterpart elements in the designer model; and (4) compare the designer model with individual user models using a simple algorithm from set theory, check the degree of compatibility between the designer model and the user models and make appropriate design decisions relating to the inclusivity of future product features. The results of the evaluation study with eight designers show that the use of the GABO modelling approach is feasible in real design situations as designers found it beneficial in representing their own and users’ understanding and use of products and subsequently comparing matches and mismatches in the understanding of the two groups. The research on the GABO approach continues and the next stage is to evaluate its usefulness and effectiveness in creating novel designs.

142

Mieczakowski et al.

14.6 References Andreasen MM (1994) Modelling: the language of the designer. Journal of Engineering Design, 5(2): 103-115 Aurisicchio M, Bracewell RH (2009) Engineering design by integrated diagrams. In: Proceedings of the International Conference on Engineering Design, Stanford, CA, US Benktzon M (1993) Designing for our future selves: The Swedish experience. Applied Ergonomics, 24(1): 19-27 Chamberlain M, Esquivel, J, Miller, F, Patmore, J (2011) BT’s adoption of customer centric design. Applied Ergonomics Special Issue, in press Craik KJW (1943) The nature of explanation. Cambridge University Press, Cambridge, UK Dong H (2005) Barriers to inclusive design in the UK. PhD Thesis, Cambridge Engineering Design Centre, University of Cambridge, Cambridge, UK Goldsmith TE, Davenport DM (1990) Assessing structural similarity of graphs. In: Schvaneveldt RW (ed.) Pathfinder associative networks, Ablex Publishing Corporation, Norwood, NJ, US Gulliksen J, Goransson B, Boivie I, Blomkvist S, Persson J, Cajander A (2003) Key principles for user-centered systems design. Behaviour and Information Technology, 22(6): 397-409 Hewer S, James L (1998) Realising potential: Two complementary views from the RSA, London. In: Placencia-Porrero I, Ballabio E (eds.) Improving the quality of life for the European Citizen: Technology for inclusive design and equality. IOS Press, London, UK Johnson-Laird PN (1983) Mental models. Harvard University Press, Cambridge, MA, US Keates S, Clarkson PJ (2003) Countering design exclusion: An introduction to inclusive design. Springer, London, UK Klimoski R, Mohammed S (1994) Team mental model: Construct or metaphor? Journal of Management, 20(2): 403-437 Kolbitsch J, Maurer H (2006) Transclusions in an HTML-based environment. Journal of Computing and Information Technology, 14(2): 161-173 Kouprie M, Sleeswijk Visser F (2009) A framework for empathy in design: Stepping into and out of the user’s life. Journal of Engineering Design, 20(5): 437-448 Langdon PM, Lewis T, Clarkson PJ (2007) The effects of prior experience on the use of consumer products. Universal Access in the Information Society, 9: 209-225 Langdon P, Thimbleby H (2010) Inclusion and interaction: Designing interaction for inclusive populations. Interacting with Computers, 22(6): 439-448 Mattelmäki T, Battarbee K (2002) Empathy probes. In: Proceedings of the Participatory Design Conference, Palo Alto, CA, US Norman DA (2002) The design of everyday things. Basic Books, London, UK Payne SJ (2008) Mental models in human-computer interaction. In: Jacko JA, Sears A (eds.) The human-computer interaction handbook: Fundamentals, evolving technologies and emerging applications, Taylor & Francis, NY, US Preece J, Rogers Y, Sharp H (2002) Interaction design: Beyond human-computer interaction. John Wiley & Sons, NY, US Ricability (2001) Inclusive design: Products that are easy for everyone to use. Disability Rights Commission, London, UK Yang MC, Epstein, DJ (2005) Study of prototypes, design activity and design outcome. Design Studies, 26(6): 649-669

Part IV

Design for Inclusion

Chapter 15 Design Advisor: How to Supply Designers with Knowledge about Inclusion? E. Zitkus, P.M. Langdon and P.J. Clarkson

15.1 Introduction In an ideal scenario accessibility issues such as legibility, usability and associated cognitive load, with respect to users of different age groups and impairments, would be considered whenever a new product was created. This means that designers would be challenged by the relation between users’ capability loss and the features of the new product. In this context, understanding what would cause exclusion from use of the products could support designers to make decisions resulting in more accessible products. Consequently, different approaches to supply designers with knowledge about inclusion have been studied throughout the last decades. This chapter analyses how interactive design advisors could provide knowledge about inclusion according to the design activity.

15.1.1 The Need for Integrating User Requirements A way of dealing with accessibility problems is by supplying designers with techniques to evaluate the physical interactions that will occur between users and new concept designs. Currently, there is a range of techniques to evaluate accessibility in new product development, some of which are listed below: •

•

user-centred techniques: well known techniques that enable designers to understand user’s needs and develop empathy to them through users’ participation in the design process (Norman, 2002). Examples of codesigning methods are seen in the works of Sanders (2000) and Rode et al. (2004). The user-centred method is also applied in user trials, user observations and more recently, user theatre (Newell et al., 2006); Designer trial techniques: these techniques enable designers to experience physical restrictions that they are unfamiliar with. Although there are different types of simulation apparatus such as Third-Age Suit, Age

P. Langdon et al. (eds.), Designing Inclusive Systems, DOI: 10.1007/978-1-4471-2867-0_15, © Springer-Verlag London 2012

145

Zitkus et al.

146

•

Explorer and Simulation Toolkit, all of them follow the same principle of restricting the motion and sensorial capabilities (Hitchcock et al., 2001; Cardoso and Clarkson, 2007). Virtual techniques are computer-based techniques in which assessments occur through software applications. They can be integrated to CAD models, through task performance simulation like HADRIAN, INCLUSIVE CAD, and in the future VERITAS and VICON. There are also tools that simulate vision and hearing capability loss by loading sounds or images, like the Impairment Simulator or tools that calculate the exclusion caused by capability demand, like the Exclusion Calculator (Marshall et al., 2004; Macdonald et al., 2006; VERITAS D4.1.3_v2, 2010; Clarkson et al., 2007).

The advantages and disadvantages of adopting these techniques in the industrial context were discussed in a previous study and a summary of each approach is shown on Table 15.1 (Zitkus et al., 2011) Table 15.1. Different techniques and their integration into design in the industrial context Process integration User trials / User observation Self observation Third-Age Suit / Age Explorer

Early in the conceptual phase, through similar products, or later through rapid prototypes.

Simulation Toolkit INCLUSIVE CAD HADRIAN

During the conceptual phase through CAD models.

VERITAS project VICON project Impairment Simulator Exclusion Calculator

Interface Observation of real users and/or getting their feedback after the trial. Observation of themselves. Designers observe themselves with physical restrictions or different levels of restrictions. Simulation of functional demand on lower limb muscles, hip and knee joints. Virtual interaction with user avatars. Virtual simulation.

During the conceptual phase through new concept images. Early in the concept phase, through task analyses.

Simulation of some of vision and hearing capability loss Virtual interaction with a range of applicable tasks.

Results Inspiring. Exclusion is not quantifiable. Re-assessing the product is an issue due to the sample selection. Inspiration is limited as the design teams do not represent a wide range of people. Inspiring. Exclusion is not quantifiable. Re-assessing the product means to wear the suit again. Exclusion is limitedly quantifiable due the range of tasks and the focus on physical capabilities. Exclusion is limitedly quantifiable due the range of tasks & users’ database. Exclusion could be quantifiable due to a broad anthropometric and capabilities database. Exclusion is limitedly quantifiable due the focus on sensorial capabilities Dependent on the knowledge of the designer.

Although the range of tools cited above seems to be a good support for assessing the accessibility of industrial products, they are barely used in the industry as part of the design process. This indicates that either the techniques to assess the product accessibility do not work in tandem with the design process or that there may be inadequacies on the application interface. Consequently, techniques that integrate the user’s needs to the design activity and its process in the industrial context are still needed.

Design Advisor: How to Supply Designers with Knowledge about Inclusion?

147

15.1.2 Proposing an Interactive Technique to the Design This chapter analyses and discusses the study carried out with experienced industrial designers (average of 10 years of experience) in two well recognised design consultancies in the United Kingdom. The design consultancies are specialised in product design, research and innovation within a broad range of industrial sectors and clients. The study was based on observations of five designers at work and recorded interviews with six other designers. At the end of each interview there was a demonstration of an interactive technique that informs the designers about inclusion. The study is part of a research project that aims to consider the design activity and its process while developing new interactive methods. The principle of the interactive technique proposed is to help designers to obtain information about the accessibility afforded by the new concept design during the time the designers are creating it. The demonstration enabled the designers to talk about interactive ways to evaluate accessibility that are informed by the design activity. The designers were asked to comment on the technique. The interactive technique was built into Google SketchUp software through Ruby programme language, both open free-sources available on internet. The demonstration emulated the design of a simple medicine pack and proposed an interactive way to check the legibility of the letters on the pack. The demonstration (illustrated in Figure 15.1) followed the sequence below: • • • • • •

designing the box (with colour and material); adding the text (with font size and style); setting the ambient light; setting the reading distance; selecting visibility test in the inclusive design tool in the tools window; receiving advice from a window that pops up on the screen, which describes the range of population excluded from reading the text in the pack and advice regarding font size, style and background/foreground colour contrast.

The findings are presented and discussed in the following sessions.

15.2 Interactive Interface for Inclusive Design The idea of having an interactive adviser built into a computer graphic tool had a positive response. The proposed interactive interface built into a computer graphic system (in this case Google SketchUp) was associated with tools to evaluate the mechanical or structural aspects of three dimensional (3D) models. For instance, mould flow analysis, stress-strain and FEM analysis were mentioned. Many of the comments, however, highlighted that the use of this kind of tool would be guided mainly by the needs of the client, instead of the designer’s work routine.

Zitkus et al.

148

Figure 15.1. The interactive interface demonstrated to the designers in the study

“So, if you can turn around and say “we’ve got stuff built into this programme could you check it (accessibility) later and tell if the design is suitable” then they (the client) would probably love it…. It could be useful if you haven’t got access to users and if the company need this kind of data… as we know, not every company can get access to, then it is probably better if they’ve got some kind of evaluation ….” Some designers mentioned other options that they use to evaluate the accessibility of new concept designs. Among them are information brought to the project by the clients themselves, guidelines, user opinions and self- evaluations.

15.2.1 Briefing, Research and Guidelines All the interviewees mentioned a similar process that happens at the initial stage of the design activity. Usually, designers are guided by a brief that is a source of information of the new product’s functionality, the components and their respective size, manufacturing, environmental considerations and by information from the potential user. It was highlighted that user data in the brief is limited to market information, and thus designers have to manage their time and budget to get user information

Design Advisor: How to Supply Designers with Knowledge about Inclusion?

149

from other sources. The designers mentioned that often some research starts taking place at this stage. This is described in detail in another study (Zitkus et al., 2012). The accessibility, however, “is not something that always got designated time within the process...”. Nevertheless, according to the designers, if it is part of the project requirement to consider accessibility, then they would normally look for ergonomic data from books, tables and internet, or they would look for specifications in guidelines. In fact, the responses indicate that designers mainly rely on guidelines, though their comments also highlighted that they find the information on these sources deficient and sometimes incompatible with their needs. They mentioned that they balance the deficiency of the guidelines by including some live assessments, such as self-evaluations and user-trials: “a lot of it I would say is based on common sense, we tend to tell to ourselves what is legible or not… I think lots of it comes with experience, the way our minds work, it becomes obvious if something is small and illegible… There are standards which drive how large a piece of text should be, you can print things out in various sizes and get feedback from the user group.” “I think lots of time that happen, that stuff (accessibility considerations) comes from experience… you are making subconscious decisions of what is good and bad accessibility. So, I think most of that coming from trying and testing ways of doing things.” “My approach would be to print out or to create different variants of the design and then just test that with people, just quickly get people there, or taking it to people, just going to Tesco, just finding people being demographically, just find some people of specific age group or if you got the profile of the user you will need.” Although the possibility of incorporating users in accessibility tests was mentioned, all interviews stressed that user observation or user trials only take place if the research allowance covers that or if the client defines this as a method of assessment. The user’s involvement in the process therefore occurs very rarely. “…even when I worked in companies that project things specifically for the elderly, it was rare in the extreme anybody who was elderly would be involved in the process… the users were not part of the process…” It is important to underline that the designers highlighted that the user’s needs, such as those related to accessibility and usability are only one part of the requirements that they have to deal with. They emphasised that design is a compromise activity, where decisions are made all the time and costs are involved in every option taken. “… the product is not only the users themselves, we have to consider who gonna assemble it by making assembly easier, how it is built, how it, if it gonna need maintenance, let's say, filters have to be changed, something like that, we have to consider how can we make that appropriate, easier or not, and at these days, go through the product going out and ending his life and needing to be recycled. So, to assemble, the materials, all that side of things…” “… to be honest, in many stages there are simple costs and practicality costs but all the primary driver before we get things like accessibility.”

150

Zitkus et al.

The designers also had some opinions about the interface of the interactive technique demonstrated. Their opinions bring to the light suitable places for a possible application and also the effectiveness of the information supplied to them.

15.2.2 Cognisant Applications for the Design Practice According to some interviewees an interactive adviser for graphic designers regarding the legibility of text, icons or any other graphic interface should be built for graphic designers into their respective tools: “I think most of the designers agencies use Illustrator or Photoshop so if it is incorporated on that it would be part of the tool you are used to using rather having to learn something… you don’t have to spend ages learning a new programme;” For other designers the idea of having all the analysis in 3D systems would facilitate their work. The interface could be two dimensional, but it should be integrated to a CAD software (such as SolidWorks and Pro-Engineer). “So, if there was a tool which can plug into CAD which allow us to set up different ergonomic parameters, certain aspects of dexterity, or visual acuity, or things like that, and then to be able to see and to get some kind of feedback with that, so that would be useful, I can’t see why it wouldn’t be… If the principle is to design in CAD, I think you should have as much information as possible in that environment.” “Two dimensional could be useful in 3D software not in terms of graphic stuff, but I guess a 2D ergonomic database, like knee height. That easily gives you that information, it would be quite useful.”

15.2.3 The Relevance of the Inclusion Information Another aspect of the inclusive interactive interface that caught the designers’ attention was the percentile of the population excluded. For them there is always a target market that guides the design activity. Advices based on the percentile of the entire population are generally not significant in such cases where a portion of the population is not part of the target market. The exclusion value that appeared on the pop-up window is therefore ineffective. “In nearly all of the products we work on it, the portion of the population might be not applicable to that, because they might be irrelevant, the usability. We would assume that some people will be excluded when you see that product; and that is acceptable for that product be successful, still it is from client perspective… I’m not sure, when you get a result like 7.3% of the population will be excluded the question is “which part of the population is it talking about?” Because if it is excluding 80% of people over 75 years old females from North East, so that is really important if the product is aiming that; those people in the North East in England. The detail of that is useful, that is what we need to know I guess.”

Design Advisor: How to Supply Designers with Knowledge about Inclusion?

151

The next section discusses the study and its relation with previous work in the design activity, as well as the relationship of the design activity with the accessibility evaluation techniques.

15.3 Deficiencies of User Information The interactive technique proposed and demonstrated to the designers supported the study in understanding the way accessibility evaluation techniques are selected and the role of the client in considering the user’s requirements. The designers highlighted ways in which the interface could be applied in computer graphic systems as well as the relevance of the information provided. In the following sections the outcomes are compared with past research and discussed in detail.

15.3.1 Sources of User information Past studies had already mentioned that time and budget constrain the adoption of user trials in the industrial context (Gill, 2009; Goodman-Deane et al., 2010). The findings described in the previous section confirm that direct involvement with users rarely happens in commercial projects. Users are consumers and their profiles are brought to designers in the form of target markets, generally consisting of a demographic view of potential consumers. User requirements, such as those related to the diverse range of capabilities among the potential consumers, are not necessarily part of the market information provided by the client. Any extra information regarding user data should come from the designer’s own research. This research, however, is limited due to the project budget and the client’s procedures, as a result of which user requirements are generally restricted to ergonomic tables available in books or on the internet. However, these tables generally do not consider a wide range of people, including the elderly and the disabled. In situations where accessibility is part of the design requirements, then designers make use of guidelines to comply with the users’ needs. In this case, the findings of the study confirm the deficiency implicit in the guidelines that were already mentioned in past studies. Cardoso and others (2003) noted that the new possibilities and the range of features of a new concept design are not covered by guidelines. Past studies also indicated that the broader the scope of the guideline the less it supports the design activity (Burns et al., 1997; Law et al., 2008). The deficiencies of guidelines are confirmed in the pilot study by the fact that the information given is not sufficient to the designers’ needs and they have to supplement it with tests and evaluations.

152

Zitkus et al.

15.3.2 User Participation in the Design Process Although some of the interviewees mentioned that they could test the features of a new concept design with potential users, the majority of the comments indicate that users are not part of the design routine. Normally, as described in the interviews, the design activity does not include elderly or disabled people in user trials, even in projects where they are the target market. Designers rely on their own assessments, limited as they are by not being able to feel other individuals’ capabilities. In addition, accessibility and usability are described as only some among the numerous requirements that designers have to comply with. User requirement is not a special requirement above the others, which means that designers have to balance the user’s needs with functionality, manufacturing, maintenance, environmental aspects, aesthetics and other requests. Moreover, it was also highlighted that accessibility does not always have a place in the product design requirements. The deficiencies in the mechanisms that provide user information to designers, added to the fact that user-trials are rarely adopted as part of the design process, result in the need already mentioned at the beginning of the chapter: the need to integrate user requirements into the design activity.

15.3.3 How Interactive Interfaces Provide Information Two aspects of interactive techniques were highlighted in the interviews after the demonstration; one is the interface, and the other is the information provided. 15.3.3.1 The Usability of the Interface The outcomes of the demonstration of the interactive technique proposed in this study confirmed an aspect already highlighted by Macdonald et al. (2006). The authors highlighted that the integration of supportive techniques with the designer’s tool is a way to enhance the acceptability of the technique. In the study carried out by Macdonald et al, the integration was with CAD models, whereas the designers’ responses included other graphic tools as well. Although some designers stressed that an interactive tool should be built into (or within) CAD, others underlined that the technique should provide the information according to the design practice and its related tool. There are reasons for both opinions. On the one hand, industrial designers tend to design in CAD and at some point all the information about the new concept, including graphic information, will be there for photorealistic renderings, which explains the easy integration with CAD. On the other hand, there are some variations across different design domains. For instance, according to this study, packaging design does not necessarily use CAD software. Although the design generally goes to computer graphics (Corel Draw for instance) before going to the laser cut, there is no need of a 3D digital model as the drawing has to be the opened/flat cut-out of the cardboard. Moreover, all the graphic design (the visual communication) to feature on the package does not need to get into CAD either: it can be done in Photoshop or Illustrator, both 2D graphic software.

Design Advisor: How to Supply Designers with Knowledge about Inclusion?

153

The above findings stress the need for generating additional interactive interfaces to be integrated and tested in a range of computer graphic tools. This would test the usability of the techniques by getting more feedback from the intended users, in this case the designers. 15.3.3.2 The Information Provided in the Interface It seems that normally designers think about a target market from which a reasonable amount of people from the entire population are excluded. Therefore, unless clients request or designers understand the benefits of changing the design, the objective of providing information about inclusion will not succeed. Information on inclusion, however, is not only a matter for designers; as already mentioned by Gill (2009) small to medium-sized design consultancies tend to face the pressure of costs and tight deadlines from the client, which constrain the designer’s decisions. Among all the interviewees the role of the client was always underlined as the main supplier of information, which includes information on potential consumers and their needs. Moreover, if the client does not act as an information supplier, they guide the designers to the kind of information they have to get through research. Therefore, further investigation is needed to understand which type of inclusive information that supports the design activity could also be practical for the client.

15.4 Conclusions Computer graphic systems seem to be well integrated in the design activity, and thus they are an important means of supplying designers with knowledge about inclusion. This knowledge about inclusion is necessary as designing is a trade-off activity that commonly does not prioritise accessibility among other requirements. However, to implement this scenario the techniques have to consider not only the design activity, but also the client’s view about the inclusive information supplied. Therefore, further research has to be carried out to understand how the benefits of knowledge about inclusion could be explicit to both designers and clients.

15.5 Acknowledgements The authors would like to thank Dr Alaster Yoxall and Prof Steve Gill for their support of this study, and also all the designers involved.

15.6 References Burns CM, Vicente KJ, Christoffersen K, Pawlak WS (1997) Towards viable, useful and usable human factors design guidance. Applied Ergonomics, 28(5-6): 311-322

154

Zitkus et al.

Cardoso C, Keates S, Clarkson PJ (2003) Assessment for inclusive design. In: Clarkson PJ, Coleman R, Keates S, Lebbon C (eds.) Inclusive design: Design for the whole population. Springer-Verlag, London, UK Cardoso C, Clarkson PJ (2007) User simulation in product evaluation. In: Coleman R, Clarkson PJ, Dong H and Cassim J (eds.) Design for inclusivity: A practical guide to accessible, innovative and user-centred design. Gower Publishing Ltd, Hampshire, UK Clarkson PJ, Coleman R, Hosking I, Waller SD (2007) Inclusive Design Toolkit. Engineering Design Centre, University of Cambridge, Cambridge, UK Gill S (2009) Six challenges facing user-oriented industrial design. The Design Journal, 12(1): pp 41-67 Goodman-Deane J, Langdon PM, Clarkson PJ (2010) Key influences on the user-centred design process. Journal of Engineering Design, 21(2-3): pp 345-373 Hitchcock D, Taylor A (2003) Simulation for inclusion - true user centred design. In: Proceedings of the International Conference on Inclusive Design and Communications (INCLUDE 2003), London, UK Law CM, Yi JS, Choi YS, Jacko JA (2008) A systematic examination of universal design resources: Part 1, heuristic evaluation. Universal Access in the Information Society, 7(1-2): 3154 Macdonald AS, Loudon D, Rowe PJ, Samuel D, Hood V, Nicol AC, et al. (2006) InclusiveCAD: A software resource for designers. In: Clarkson PJ, Langdon PM, Robinson P (eds.) Designing accessible technology. Springer, London, UK Marshall R, Case K, Porter JM, Sims R, Gyi DE (2004) Using HADRIAN for eliciting virtual user feedback in ‘design for all’. In: Proceedings of the Institution of Mechanical Engineers, Part B: Journal of Engineering Manufacture, 218(9): 1203-1210 Newell AF, Carmichael A, Morgan M, Dickinson A (2006) The use of theatre in requirements gathering and usability studies. Interacting with Computers, 18(5): 996-1011 Norman DA (2002) The design of everyday things. Basic Books, London, UK Rode JA, Toye EF, Blackwell AF (2004) The fuzzy felt ethnography: Understanding the programming patterns of domestic appliances. Personal and Ubiquitous Computing, 8(3-4): 161-176 Sanders EB-N (2000) Generative tools for co-designing. In: Scrivener SAR, Ball LJ, Woodcock A (eds.) Collaborative Design, Springer-Verlag, London, UK VERITAS D4.1.3_v2 (2010) Project presentation and project description leaflet. Fraunhofer IAO, Stuttgart, Germany Zitkus E, Langdon PM, Clarkson PJ (2011) Accessibility evaluation: Assistive tools for design activity in product development. In: Proceedings of the 1st International Conference on Sustainable Intelligent Manufacturing, Leiria, Portugal Zitkus E, Langdon PM, Clarkson PJ (2012) How are user requirements dealt with in the design process. In: Proceedings of the 12th International Design Conference (DESIGN 2012), Dubrovnik, Croatia, in press

Chapter 16 From Guinea Pigs to Design Partners: Working with Older People in ICT Design R. Edlin-White, S. Cobb, A. Floyde, S. Lewthwaite, J. Wang and J. Riedel

16.1 Introduction 16.1.1 Ageing Societies and the Implications for Research Governments worldwide are recognising the reality and challenges of aging societies. Aging is often accompanied by increased incidence of impairments in the physio-motor, sensory and cognitive domains, as well as health problems, reduced socialisation, poorer finances, reduced sense of purpose and sometimes marginalisation from society. Technology can be very beneficial for older people but is too often problematical, sometimes creating digital exclusion. Researchers and funding bodies are beginning to focus specifically on the ICT needs of older people, and excellent work is being conducted, but as yet research methods and models of engagement for working with older people are fragmentary or at an early stage of development (Rice and Carmichael, 2007). This paper explores methods and approaches currently used in HCI research with older users, and those used by the authors in the MyUI European research project, and presents a list of themes which the authors believe will be relevant to other work in the area and could inform a more consistent methodology.

16.1.2 User-centred Methods and Approaches in HCI A critical problem in the design of usable, accessible technology is that most designers are highly motivated by technology per se and have an inadequate understanding of the world of end users. Their effectiveness at user-centred design may be impaired by positivist world views, assumptions that user experiences are similar to their own, and a propensity for confirmation bias when determining user

P. Langdon et al. (eds.), Designing Inclusive Systems, DOI: 10.1007/978-1-4471-2867-0_16, © Springer-Verlag London 2012

155

156

Edlin-White et al.

needs (Bader and Nyce, 1998; Norman, 1998; Simonsen and Kensing, 1998). Design teams and processes should be more user-centred. Muller et al. (1997) noted the chasm between the worlds of designers and users, each world having its own culture, language, space and assumptions. The HCI community, they say, may need to develop “anthropological and sociological methods for working across differences in language and abilities” to improve “the field’s competence in intercultural communications”. This is even more necessary when users are older, retired people, and designers are young and employed. Cultivating a developing rapport with user communities, through field study, or better still by including users in the design process, can help designers bridge the chasm, understand the world, work and requirements of users, and by combining User Needs Analysis and Requirements Engineering can lead to more accessible and effective products (Muller, 2002; Lindgaard et al., 2006). Muller presents a thorough, persuasive account of Participatory Design. User engagement, he finds, is too often one-directional, creating applications (of technology) rather than solutions (to user problems). He appeals for mutual learning and openness to surprise (Muller, 2002; see also Farrel et al., 2006). Crabtree (1998) believes ethnography can usefully complement Participatory Design in user studies. People cannot always describe their own work effectively, but an observer can see and record the activity. An approach based on reviews of prototypes, he says, rarely empowers the users to challenge underlying premises such as “are we solving the right problem?”

16.1.3 HCI Methods Currently Used with Older Participants It has been argued that standard HCI approaches to user engagement are not appropriate when working with older people (Eisma et al., 2003, 2004; Rice and Alm, 2008). Much academic HCI research has used (mainly young) students as the main source of participants, age-biasing not only research findings but also the evolution of research approaches and methods. Older people, and those with disabilities, are currently under-represented as participants in academic research, especially HCI research (Lewis, 2006; Newell et al., 2007; Walker, 2007). Some HCI researchers working with older users have published their methods, usually in a descriptive way, not formulated as a standard recommended approach. Newell and Gregor (2000) discuss “User Sensitive Inclusive Design” - a paradigm for including users with disabilities, including older users, within a development framework. They appeal for respectful user involvement, but also note that - for a variety of reasons - potential end-users are not always the sole or best source of information about how a novel system may be of benefit to them or their diverse peers. They and other researchers (Muller, 2002; Hansson et al., 2006; Rice and Carmichael, 2007) recognise that other parties - partners, carers, clinicians or other professionals, teachers, family members and union representatives - can provide valuable supplementary insights. Eisma et al. (2003) appeal for a process involving “mutual inspiration” when working with older people; this can happen most readily in informal studies and interactions with the participants; in enjoyable, sociable group activities with

From Guinea Pigs to Design Partners

157

refreshments, and engaging in hands-on activities using research-related stimuli. This can provide rich, grounded and often serendipitous findings. A questionnaire conducted in a group context can stimulate conversations which generate more findings than the written responses. Observation by researchers can reveal more information than user reports alone (Eisma et al., 2003). Working with older and disabled people as users, Picking et al. (2009) developed the EDUCATID framework, with ethical considerations and user engagement at the heart of the process. The ethical approach is based on Abascal and Nicolle’s (2005) detailed and thoughtful discussion of ethics in HCI work. A genuine and open concern for the user can highlight impacts and consequences of a technological intervention which might otherwise be overlooked. In EDUCATID, user narratives are elicited through workshops, focus groups etc. and these are used to generate requirements and later concept designs. The method continues with formative and summative evaluations of prototypes and systems. Focus Groups are noted as useful in this field by many researchers (Barrett and Kirk, 2000; Eisma et al., 2003; Goodman et al., 2004). The OASIS project (Lindsay et al., 2008) developed the Oasis Participatory Analysis Framework (OPAF), for involving older users throughout a design process, including some very practical guidance in areas like recruitment. They advocate group discussions, focus groups and design workshops in field settings as methods of engagement, which they believe are less time consuming than pure ethnographic methods. They also commend extensive use of PICTIVE techniques (Muller, 2002) and low fidelity prototyping, in workshop and group sessions. Similar advice to all the above, most relevant for plentifully resourced work, is provided by the respected CREATE centre for research on aging and technology in chapters 3, 12 and 13 of Designing for Older Adults (Fisk et al., 2009). Few quantitative methods appear to be used. Pernice and Neilsen (2002) describe a number of qualitative and quantitative studies with older participants, concluding (p122) that quantitative studies need a large number of participants to achieve statistical significance, are time consuming to design and set up, and generate good numeric data but fewer insights than qualitative methods. Qualitative methods, while generating less objective data, and with the generalisability of the findings sometimes open to question, are better at creating deep and rich understanding, and are therefore particularly well suited to generating ideas and hypotheses, in understanding fields characterised by ambiguity and complexity, and in requirements elicitation and early evaluations in a process of iterative refinement (Creswell, 2007; Lamsweerde, 2009). The above and other related literature informed the approaches adopted in the MyUI project. It provided valuable insights but leaves many possibilities open.

158

Edlin-White et al.

16.2 Involvement of Older People in MyUI Project 16.2.1 The Project MyUI is a European Union funded research project which sets out to facilitate the development of highly accessible life-enhancing technology-mediated services for older people, including those who are experiencing age-related changes and impairments in the physio-motor, sensory or cognitive domains. Web-based services (in areas such as social communication, simple exercise programmes monitored and modified by experts, entertainment, education and reminders) will be delivered through the attractive medium of interactive TV using traditional or novel hand-held remote controls. Accessibility will be enhanced by the use of adaptive systems which detect a user’s changing accessibility needs during normal usage and adapt the user interface in real time to better match those needs. The University of Nottingham Human Factors Research Group was responsible for establishing the methodological approach, coordinating all end-user studies across Europe and conducting most of them.

16.2.2 Methodology for Participant Involvement It is clear that most of the main objectives of MyUI - to build accessible, lifeenhancing, valued services for older people - are based on subjective social constructs. Notions such as “beneficial”, “accessible” and “usable” will vary from person to person. A quality of life measure must include the viewpoint of the person living that life. People who make an informed choice to opt out of a technology are not necessarily “lost sheep” (Klecun, 2008). This understanding profoundly affects our style of engagement with communities of older people. We aim to listen to their articulations of these realities rather than imposing our own. Based on experience of previous projects and on the literature reviewed above, a hybridised and evolving methodology has been adopted, influenced by Participatory Design, ethnography and Participatory Research, and shaped by social constructivist models. Participants are treated as co-creators, the foremost experts in their own needs, whose contribution is recognised and valued as a vital part of the work. In Radermacher’s (2006) classification of participatory methods our main approach is “Researcher initiated/shared decisions with participants”. The project has also involved some empirical studies with older people, an approach based on a more positivist paradigm, in which the older person is - at least temporarily - a subject (or “guinea pig”) rather than an equal participant. In order to avoid Hawthorne effects, the purpose of the study and the nature of researcher observations were sometimes not fully disclosed until the end of the study, and the researchers temporarily adopted a slightly more detached stance. This was prompted by demands for “scientific” studies, and has provided convincing, formative findings, but as an approach sits somewhat uneasily with our ethical stance and quality of relationships with our user communities.

From Guinea Pigs to Design Partners

159

16.2.3 Ethical Approach Our ethical approach on this project goes well beyond meeting bureaucratic requirements. An ethical stance should permeate every aspect of the work, treating all stakeholders - but especially any who are vulnerable - with honesty, respect, courtesy and care. We believe the research should benefit the participants or future generations of people in similar circumstances, and not just our own interests as researchers. Informed consent is “gently invited”, recognising that hearing, reading, remembering etc. may be impaired. If later a participant asks us what we’re doing or why, we willingly reiterate and remind them they can withdraw. Part of our stance is the belief that accessibility problems generally arise from poor design, not inadequate users, so we avoid the use of pejorative terms such as “computer literacy” and are more concerned to ensure our designs should be informed by “user literacy”. Where possible we avoid medical questions and instead ask accessibility questions; e.g. we ask “would this display be adequately readable for you”, and not “is your eyesight adequate for this display”. Our questionnaires and consent forms are designed so they can also be administered verbally, in case the participant has any problems with reading or writing, e.g. due to literacy, eyesight, or digital dexterity limitations. We attempt to use age-appropriate measures of accessibility, usability or benefit of a technology, as older and retired people tend to have different priorities. These measures are mainly subjective, e.g. enjoyment, satisfaction, feeling of autonomy, mastery, ease of use and ease of understanding, and not feeling patronised or frustrated by the technology. We recognise that placing a technological artefact into someone’s domestic environment and expecting them to use it can have far-reaching effects in many domains. For example, enabling someone to do online shopping might affect the frequency of going out to shop, and hence their amount of socialisation, exercise, sunlight and fresh air, perhaps therefore affecting their physical and mental health. It might reduce the amount of help with shopping from family or neighbours again reducing social contact. Different grocery buying patterns may affect the diet and finances for better or for worse. Provision of technology may lead to a need for training, support and maintenance activities which may or may not be satisfied. Some ICT projects for older users have concluded with brief and superficial usability studies, optimistically interpreted, failing to take into account the benefit of support and training from researchers which could never be afforded in a commercial product (Dickinson and Gregor, 2006; Blaschke et al., 2009). We believe it is ethically appropriate to also try to estimate the broader Quality of Life (QoL) so that these many factors are not ignored.

16.2.4 Study Settings We established regular contact with three centres where older people routinely congregate: two Day Centres (DC1, DC2) and a self-contained retirement village

160

Edlin-White et al.

(RV1), between them representing over 600 older people and providing a good range of contrasting demographics. Through our Spanish consortium partners we organised research in an older people’s complex in Getafe, near Madrid. There were clear contrasts in lifestyle and aging and therefore in requirements. Compared to our UK groups, the Getafe cohort’s lifestyles seemed to include more swimming, dancing, and sociable outdoor activities (perhaps due to the climate) but less literacy. A popular activity among the Getafe women was making costumes for fiestas and carnivals (EdlinWhite et al., 2010). Older people with such physically, socially and cognitively active lifestyles are less susceptible to age-related impairments. All research activities have taken place in field settings - DC1, DC2 and RV1. Lab-based studies with older people can create artificial responses due to changes to their routine, as well as creating practical problems (e.g. transport, diets, carers etc.) and sometimes duty of care issues (Edlin-White et al., 2011). However studies carried out in “open” field settings, e.g. a day room, can be susceptible to all manner of distractions. During one of our early focus groups, staff brought round trays of sherry to mark a 40th wedding anniversary. It seems uncharitable to describe this as a confounding factor and not ecological validity. To improve research focus, we created unobtrusive “pop-up” labs in these locations, by arranging to use a side room and bringing in sufficient equipment and materials to create a mini usability lab. This allowed fewer distractions, more experimental controls, a more formal and comparable participant experience, and in individual studies eliminated the tendency for participants to influence each other’s responses. For some studies, within the implanted lab we created an inner environment (furniture, décor, lighting and equipment) to emulate a domestic TV watching experience for the participant.

16.2.5 Study Methods - Progressive Engagement Process The research involved significant journeys of increasing trust and collaboration for researchers, staff and participants. Initial contact was through senior staff (“gatekeepers”) at each venue. Early communication included presentations at a non-academic level, to explain the project and our research aspirations, stressing aspects likely to be of interest to this audience. In all cases the result has been positive; staff believed that older people are not consulted sufficiently in technology design. Initially study design was heavily influenced by the need to build trust and to ensure staff and participants believed the visit was enjoyable and beneficial. This meant working within the unique established culture of each venue and user community. In DC2, for instance, this meant working with a group of around 18 older people with a variety of abilities and limitations, sitting on seats arranged in a horseshoe shape along three walls of a large room. The research opportunity might be a 45 minute slot following a raffle and preceding a seated exercise session, and there would be the sound of a radio playing light classical music from the adjacent kitchen as well as the smells of food being prepared. Within this context it was possible for us to address the whole group, to obtain consent, and to conduct an

From Guinea Pigs to Design Partners

161

informal focus group, stimulated by handing round a variety of hand-held devices (iPod, phone, TV remote, DVD remote, calculators etc.). In DC1 people were often sitting “cabaret style” around small tables in groups of 4 to 6, perhaps knitting, chatting, doing puzzles, etc.. There would be periodic activities led from the front using a microphone, e.g. quizzes. We therefore used this “pub quiz” format to conduct early research. As trust and rapport grew, it became possible to increase the research rigour and shape the culture accordingly; e.g. quiz or puzzle formats but including questions about technology attitudes and aptitudes. With staff support it became possible to create the pop-up lab implants, and invite participants to leave the communal area for short “activities” (a better word than “studies” for this audience). In early visits we avoided asking personal questions, e.g. concerning disabilities; latterly it became acceptable to include such questions. The responses help to interpret study results and to recruit more purposefully. Consent for photography, video and audio recording was forthcoming too, as trust grew. Research methods have included focus groups, collaborative participatory design workshops, questionnaires, interviews, usability studies and empirical studies. Participant selection within the venues was initially opportunistic, based on availability and willingness of participants and the possibilities afforded by communal space or pop-up labs. Later, more purposeful sampling was used. Recruitment at RV1 was initially by advertising an informational presentation event. Further recruitment has been through networking, involvement and announcement at their well supported Monday night pub quiz, and with the assistance of two enthusiastic residents who have acted as recruiting champions. We have not offered financial incentives to participants; staff advised these would be unnecessary, but refreshments (hot drinks and cakes) have been provided, especially for longer (>45 minute) sessions, or to facilitate group interaction. We have instead explored whether we can support the venue, e.g. by provision of equipment or expertise. Most participants seem glad to be involved, embracing a volunteering ethos, hoping that it may lead to better technology in the future, even if not in their lifetimes. Significant effort has been dedicated to keeping the staff and participants at all three locations informed, agreeable and supportive. We strive to make the research activities enjoyable for participants. Afterwards we seek feedback from staff to ensure they are happy with our visit and with the well-being and enjoyment of the participants during the study. Participants have taken on a number of roles. In our few empirical studies, they have to some extent been subjects (or “guinea pigs”) rather than participants. However, in most activities the participant is engaged as an expert in their own accessibility needs and technology requirements, and indeed a valued design partner. Some of the more able participants have also acted as proxies for others not in the group, perhaps with greater accessibility needs. Some of our questions target this proxy role, with phrases such as “you or other older people”. More able participants sometimes act as carer or support for a less able person; e.g. helping them with understanding or reading or writing. Some of our more enthusiastic and supportive older people have become champions for our cause, helping with recruitment, room booking and even in aspects of facilitation of sessions.

162

Edlin-White et al.

16.3 Conclusions It appears that there is significant expertise but limited methodological guidance in this field. Methods and approaches are evolving. The MyUI project drew on the literature but also learned much through experience and continuous refinement of methods. We have identified the following common themes from both literature and project experience; themes which would have been beneficial to us at the start of our project, may be relevant for future projects, and may contribute towards more formal methodological recommendations. • • • • • • • • • • • •

Recognise that HCI work with older people is social research, based on subjective social constructs such as accessibility, usability and quality of life. Adopt an ethical and user-centred approach, perhaps utilising elements of Participatory Design, Accessible Design and Ethnography. User perspective to be taken seriously and respectfully, but self-knowledge can be incomplete or imperfect. Include supplementary perspectives. Allow time in the project plan to build trust with participant groups and gatekeepers, and cultivate good working relationships on an ongoing basis. Study settings - mainly field settings for ecological validity. Pop-up labs implanted in field settings can be effective, though time consuming. Recognise the challenge and difficulty of recruiting a representative sample. Study methods - Focus Groups and Design Workshops are effective; also questionnaires and interviews and possibly retrospective verbal protocols. Design all aspects of studies to accommodate participants with very varied abilities. This includes the Informed Consent process. Allow for different learning speeds and varied learning styles of older people. Quantitative methods with inferential statistics are unlikely to be effective unless there is a lot of time and resources and access to many participants. Study measures need to be appropriate to older users, and will probably include more subjective measures than mainstream HCI studies. Be flexible and open to change while conducting a study. Allow time for “social niceties” and off-topic digressions - some of which prove to be useful. Quality of life impacts of technology are important. QoL impact measurement is difficult, costly and most properly conducted over a long time period.

It is hoped that our experiences documented here, together with those of others, will contribute towards the development of more fully formed and finely honed methodological guidance for involving older people in technology design.

From Guinea Pigs to Design Partners

163

16.4 References Abascal J, Nicolle C (2005) Moving towards inclusive design guidelines for socially and ethically aware HCI. Interacting with Computers, 17(5): 484-505 Bader G, Nyce JM (1998) When only the self is real: Theory and practice in the development community. Journal of Computer Documentation, 22(1): 5-10 Barrett J, Kirk S (2000) Running focus groups with elderly and disabled elderly participants. Applied Ergonomics, 31: 621-629 Blaschke C, Freddolino P, Mullen E (2009) Ageing and technology: A review of the research literature. British Journal of Social Work, 39(4): 641-656 Crabtree A (1998) Ethnography in participatory design. In: Proceedings of the 1998 Participatory Design Conference, Seattle, WA, US Creswell J (2007) Qualitative inquiry and research design: Choosing among five approaches. Sage, Newbury, CA, US Dickinson A, Gregor P (2006) Computer use has no demonstrated impact on the well-being of older adults. International Journal of Human-Computer Studies, 64(8): 744-753 Edlin-White RW, Floyde A, D’Cruz M, Cobb S, Riedel JCKH (2010) User interface adaptation for accessibility by older users and the challenges of cultural diversity. In: Proceedings of the 14th International Conference on Human-computer Interaction: Towards Mobile and Intelligent Interaction Environments (HCII 2011), Orlando, FL, US Edlin-White RW, Cobb S, D’Cruz M, Floyde A, Lewthwaite S, Riedel JCKH (2011) Accessibility for older users through adaptive interfaces: Opportunities, challenges and achievements. In: Proceedings of the 14th International Conference on Human-computer Interaction: Towards Mobile and Intelligent Interaction Environments (HCII 2011), Orlando, FL, US Eisma R, Dickinson A, Goodman J, Mival O, Syme A, Tiwari L (2003) Mutual inspiration in the development of new technology for older people. In: Proceedings of the International Conference on Inclusive Design and Communications (INCLUDE 2003), London, UK Eisma R, Dickinson A, Goodman J, Syme A, Tiwari L, Newell AF (2004) Early user involvement in the development of information technology related products for older people. Universal Access in the Information Society, 3(2): 131-140 Farrell V, Farrell G, Mouzakis K, Pilgrim C, Byrt P (2006). PICTIOL: A case study in participatory design. In: Proceedings of the 18th Australia conference on ComputerHuman-Interaction (OZCHI ’06), Sydney, Australia Fisk AD, Rogers WA, Charness N, Czaja SJ, Sharit J (2009) Designing for older adults: Principles and creative human factors approaches. CRC Press, London, UK Goodman J, Dickinson A, Syme A (2004) Gathering requirements for mobile devices using focus groups with older people. In: Keates S, Clarkson PJ, Langdon P, Robinson P (eds.) Designing a more inclusive world. Springer, London, UK Hansson C, Dittrich Y, Randall D (2006) How to include users in the development of offthe-shelf software: A case for complementing participatory design with Agile development. In: Proceedings of the 39th Annual International Conference on System Sciences, Kauai, Hawaii, US Klecun E (2008) Bringing lost sheep into the fold: Questioning the discourse of the digital divide. Information, Technology and People, 21(3): 267-282 Lewis C (2006) HCI and cognitive disabilities. Interactions, 13(3): 14-15 Lindgaard G, Dillon R, Trbovich P, White R, Fernandes G, Lundahl S et al. (2006) User needs analysis and requirements engineering: Theory and practice. Interacting with Computers, 18(1): 47-70

164

Edlin-White et al.

Lindsay S, Jackson D, Olivier P (2008) Guidelines on UCD Requirements extraction. Available at: http://irgen.ncl.ac.uk/oasis/D5.5.2.doc (Accessed 4 September 2011) Muller MJ (2002) Participatory design: The third space in HCI. In: Jacko JA, Sears A (eds.) The human computer interaction handbook: Fundamentals, evolving technologies and emerging applications. Lawrence Erlbaum, Mahwah, NJ, USA Muller MJ, Wharton C, McIver WJ, Laux L (1997) Toward an HCI research and practice agenda based on human needs and social responsibility. In: Proceedings of CHI 97 Workshop on Human Factors in Computer Systems, SIGCHI, Atlanta, GA, US Newell AF, Gregor P (2000) User sensitive inclusive design - in search of a new paradigm. In: Proceedings of the ACM Conference on Universal Usability (CUU’00), Washington, DC, US Newell A, Carmichael J, Morgan M (2007) Methodologies for involving older adults in the design process. In: Proceedings of the 4th International Conference on Universal Access in HCI, Beijing, China Norman DA (1998) The invisible computer. MIT Press, Cambridge, MA, US Pernice K, Nielsen J (2002) Web usability for senior citizens. Design guidelines based on usability studies with people age 65 and older. Nielsen Norman Group, Fremont, CA, US Picking R, Robinet A, Grout V, McGinn J, Roy A, Ellis S et al. (2009) A case study using a methodological approach to developing user interfaces for elderly and disabled people. The Computer Journal, 53(6): 842-859 Radermacher HL (2006) Participatory action research with people with disabilities: Exploring experiences of participation. PhD Thesis, Victoria University, Victoria, Australia. Available at: http://wallaby.vu.edu.au/adt-VVUT/uploads/approved/adtVVUT20060904.160428/public/02whole.pdf (Accessed 10 March 2008) Rice M, Carmichael A (2007) Effective requirements gathering for older adults. Computing and Accessibility Special Issue on Accessible Europe, SIGACCESS Accessible Computing Newsletter, 88: 15-18 Rice M, Alm N (2008) Designing new interfaces for digital interactive television usable by older adults. ACM Computers in Entertainment, 6(1) Simonsen J, Kensing F (1998) Make room for ethnography in design! The Journal of Computer Documentation, 22(1): 20-30 Van Lamsweerde A (2009) Requirements engineering: From system goals to UML models to software specifications. John Wiley & Sons, London, UK Walker A (2007) Why involve older people in research? Age and Ageing, 36(5): 481-483

Chapter 17 When Users Cannot be Included in Inclusive Design R. Herriott

17.1 Introduction Inclusive Design (ID) methods place a strong emphasis on user participation in designing mainstream products. In recent years researchers in the field of assistive technology (AT) have drawn on and contributed to the ID approach. There are good grounds for this association. However, the linkage elides the differences in methods that are available and appropriate to designers in the respective fields. The demands made by strategies such as co-creation, focus groups, cultural probes and even simple interviewing can be above the capacities of the users of AT. Yet the impairments of ill and disabled users make the need for usability and pleasurability (Green and Jordan, 2002) even more important since alternative products are comparatively few. This paper examines the workarounds two teams of designers have used to reduce the demands placed on emphysema patients and elderly users during inclusive design processes. In the case of a student design project it was necessary to focus on a super-user, use prototyping as a creative tool and to use improvised ergonomic simulation. In a second case a consultancy was required to place more emphasis on ethnographic, observational methods and personas where co-creation and co-design proved to be beyond the capacities of the user-group.

17.1.1 Structure The paper begins with a discussion of ID and its relation to AT design. There is an overview of approaches to AT. This is followed by a description of two cases. The first concerns a breathing apparatus for patients with chronic obstructive pulmonary disorder. The second is a municipally funded research project into innovations to help the elderly and handicapped. Both projects were based in Denmark. To conclude there is a discussion of the findings and their implications.

P. Langdon et al. (eds.), Designing Inclusive Systems, DOI: 10.1007/978-1-4471-2867-0_17, © Springer-Verlag London 2012

165

166

Herriott

17.2 Inclusive Design and AT The literature on inclusive design shows a close linkage between it and AT. The CWUAAT series (2004 onwards), and the Include Conference series (2005 onwards), show a strong association between the methodology of inclusive design and the goals of AT products (e.g. Orpwood et al., 2004, Dhiensa et al., 2005, Mayagoitia et al., 2006, Mountain et al., 2006, Orpwood et al., 2008, Linnott, 2011). There is an overlap of interests driving this association. Inclusive design is a set of methods to optimise product designs so as to accommodate a much broader range of capabilities (BSI, 2005). AT in contrast is not a design method but a set of solutions to the problems caused by chronic ill-health or long-term disability. There are two ways to look at the distinction. One is that AT products have an element of compulsion lacking in mainstream consumer goods. If one wants to walk despite having weakened leg muscles, one must use a walking stick or rolling-frame. In contrast, it is only in a broad sense that one is compelled to own many consumer goods. Such goods are additions to our range of standard capabilities rather than replacements. AT products stand in where patients’ abilities have declined severely or were not there originally: the wheelchair, hearing aid and speech synthesiser are not products the average user will seek to purchase. Furthermore, potential AT users may even prefer to avoid using such products even if they could help; this is sometimes the case because the stigmatising appearance of the product is considered worse than the problems caused by not using it (e.g. Bichard et al., 2007). Another way to consider the cleavage between mainstream ID and AT is that the latter represents the point at which there is a step-change in the needs and capabilities of the user groups. A person who can read a newspaper with glasses in average light is qualitatively different from the person who effectively cannot read standard text. A person who walks comfortably though slowly is qualitatively different from one for whom walking demands 100% of their respiratory capability. It is worth pointing out that the binary distinction between impaired and mainstream users is a somewhat artificial construct since ability loss can vary in many cases. However, in the cases discussed in this paper there were many indicators which could objectively point towards the users being clearly on that side of the continuum where AT devices were the appropriate design solution. Inclusive design is thus a generally applicable range of tools and strategies originally devised to improve mainstream product design. These tools are aimed at making users a bigger part of the design process resulting in more valid design solutions (e.g. Coleman et al., 2007). While assistive technology can take inspiration from the ideas of better design and higher quality aesthetics embodied by inclusive design, it must be recognised that the user groups have qualitatively different characters. The design processes for each therefore can not be the same. I am not here suggesting that AT designers think that their users´ profiles are identical with the profile of the broader average. I wish instead to draw attention to the need for a different emphasis in using the tools devised from one area, mainstream design, in another area, assistive technology. The paper proceeds from this starting point to examine how designers manage the different capabilities of users who are unable to be as large a part of the design process as designers might wish.

When Users Cannot be Included in Inclusive Design

167

17.2.1 Approaches to AT Design One approach taken is to view AT as a branch of medical technology which uses design methods allied to those used for medical and rehabilitation devices (Cook and Hussey, 2002). This has tended to use a hard systems model of development: “Within design for disability, where teams tend to come from clinical and engineering backgrounds, the dominant culture is still one of solving problems” (Pullin, 2009). A second approach is to borrow from the more user-centred, socialscience based methods of ID. As Reed and Monk (2006) put it: “The key to understanding user-centred design is the aphorism ‘Know your user.’” This is a more recent development. The research presented at CWUAAT and at RCA’s INCLUDE conferences shows that this second approach is still a developing field. Torrens (2010) classifies the two approaches as being respectively quantitative and qualitative. Torrens suggests a third, synthetic approach using mixed methods which “offers the optimum opportunity for triangulation: validation of data through multivariate methods of collection” (ibid.). This synthetic approach still assumes not unreasonably - a significant role for the user in the design process. The three approaches to AT differ in the extent to which social-science methodology is deployed. Whichever approach is chosen, the range of tools to bring the user into the design process is the same: questionnaires, video ethnography, interviews, focus groups, critical user forums, cultural probes, and workshops. Ideally, the user is also consulted when the information initially gathered is turned into a specification and again after concept designs are translated into alpha and beta prototypes (Clarkson et al., 2007). Ideally, ID methods should be as comprehensive as possible. This builds redundancy into the information gathering process so that what one method may miss another may pick up. But such redundancy can be extremely demanding for patients with debilitating conditions such as emphysema, severe arthritis or muscular dystrophy. This situation points to an asymmetry in the inclusive design process. While the aim of the process is to design a more inclusive product, the input process of information gathering, processing and validation may itself be violating the sixth principle of universal design: low physical demand (Story et al., 1998).

17.3 Cases There now follow two cases of how designers responded to the diminished possibility for user-involvement in the design of AT products. The information was derived from interviews carried out by the author. The interviews were recorded and transcribed. For clarity the quotes have been edited.

Herriott

168

17.3.1 Breathe Easy Tonirie Sembach Lauridsen and Marco Barbiani, students at the Aarhus School of Architecture, identified problems with the existing oxygen supply systems for patients with reduced lung-capacity. The project was carried out in spring 2011. Their user-group consisted of patients with chronic obstructive pulmonary disease (COPD). This is “a major cause of morbidity and mortality. In Europe COPD, together with asthma and pneumonia, is the third most common cause of death. It is a disorder characterised by reduced maximum expiratory flow and slow forced emptying of the lungs, features which do not change markedly over several months” (Jeffery, 1998). This condition results in a sharply reduced capacity for physical activity and a much reduced will for and interest in non-essential tasks. COPD patients supplement their air intake using wearable inhalers. The limited physical capacity of these patients requires that the equipment is as a light as possible and that it requires the least physical effort to operate. As a product for people with severe lung disease, Breathe Easy falls into the category of assistive technology. It demonstrates the added difficulties of working with ill users and keeping them involved as much as is required by the ideal inclusive design process.

Figure 17.1. Breathe Easy portable unit

Initially, the designers formulated a design strategy approximately conforming to that recommended by the Cambridge Engineering Design Centre (EDC). However, they discovered that their users were unable to participate very fully in the design process due to their medical condition. The planned design process was one where the users played a key role as per ID best practice. When asked if the design process followed the planned design process the answer was “No. Definitely not” (Barbiani, 2011). The format of Barbiani and Lauridsen’s design was that of the concentrator, which extracts oxygen from the atmosphere. This class of device has a concentrator which is a large stationary element (principally a tank) and a smaller portable reserve carried by the patient. Technical problems related to the reloading of the portable unit, making the portable part comfortable to wear and managing the inconvenience of the nasal catheter. The existing designs required more force to connect the portable unit to the concentrator than the users were comfortable

When Users Cannot be Included in Inclusive Design

169

with so a simpler connection was needed. This design also had to accommodate the difficulty of treating the super-cooled connecting elements. In existing products these are exposed thus presenting the risk of frost-burn. Finally, the concept had to look more like a regular consumer good than something medico-industrial. Users resented having scuffed, crude-looking equipment in their homes. Ideally, inclusive design involves user-centred design methods in order to bring users actively into the design process. The students discovered that a combination of sick users and a technically demanding product rendered these objectives difficult to attain. The result was a project which aimed for user-centred design but where the users were few in number and often too sick to interact with. When asked whether the designers had a process in mind, the answer was that “in our mind actually it was very structured, from A to Z…” (ibid.). The actual process was described as “like a spaghetti and meatballs sauce thing because it was very difficult to divide the few things, the techniques from the aesthetics, from the problems the users have...” (ibid.). Finding users was the initial difficulty for the students: “I would say to find users took a very long time......quite some weeks actually. I would say the first one and a half months went on research...” (ibid.). The project duration was 12 weeks. Seven users were identified but they were not able to contribute as expected. The designers tried to use cultural probes but the resultant material was sparse and unsatisfactory. Workshops and brainstorming sessions were also planned but found unworkable. The actual strategy adopted was to video record interviews. This proved to be time consuming and created ethical and welfare problems: “[The patients] get very tired and just speaking for one hour is a project that takes more or less one day for them [...] For them to go from the kitchen to a table two metres away would take a lot of energy. And the interviews were very calm, relaxed [but] it was very difficult for them. You could see the effort they were making. We did not want to take a lot out of them, we did not want to kill them in our interviews....” (ibid.). To work around these problems the designers interviewed medical personnel and they used themselves as subjects in testing parts of the proposed designs. The problem of finding an optimum location for the nasal catheter (a thin tube connected at one end to the portable unit) was one which side-stepped difficulties of working with their users directly. This was a matter of prototyping the range of positions which eliminated obstruction caused by the tube connecting from the nose to the transportable device. This necessitated a vertically mounted tube in contrast to the side-mounted tubes of standard designs. Inclusive design and standard ergonomic practice warn against using the designer as substitute for the user. However, in the case of frail and weak users this goal was difficult practically and ethically. The designers used a breath restriction device - essentially breathing through a narrow tube - to simulate the reduction in oxygen intake. This allowed them to understand to some degree their users´ difficulties. A physiotherapist provided insight concerning the practical ergonomics of the design. This focused on minimising interference with thoracic muscle movement due to the weight of the portable unit. Having settled on a design offering the best compromises, the design was shown to what the designers called their “super-user”. This person was the most committed individual in the user-group who, coincidentally, had a better

Herriott

170

than average understanding of the design process. They were able to spend more time and to more fully explain their needs and experiences. In this case the superuser was atypical because of their engagement rather than because they were a more extreme case of COPD. After this consultation a full-size mock-up was prepared and the project thus completed.

17.3.2 The Four Municipalities Project Four municipalities (Aarhus, Esbjerg, Næestved and Roskilde) in Denmark formed a consortium to develop devices to increase the independence of elderly people with a variety of medical conditions. The intention was that the concepts should also not require additional manpower or, at best, reduce the need for it. The themes of the project were 1) the toilet, 2) outdoor and indoor mobility, 3) home training and personal care and 4) mobility for obese citizens. The project was managed by CAT in Roskilde, Denmark. CAT, formerly known as the Centre for Applied Technology, is a privately owned company which helps entrepreneurs and innovators to find the capital and competences required for converting ideas into enterprises. The design was carried out by CPH Design from Copenhagen (the initials “CPH” refer only to the official code for Kastrup Airport). The design process consisted of the following steps: a start-up process, research, analysis, conceptual development and evaluation. The start-up process involved assembling expertise from the Copenhagen Business School, anthropologists from Aarhus University, the Danish Centre for Assistive Technology, the Danish Technological Institute and interest groups representing the elderly and handicapped. The stakeholders thus had an extensive role in a product’s development as per ID’s tenets.

Figure 17.2. Body drier concept drawing

When Users Cannot be Included in Inclusive Design

171

The research phase was intended to involve a process approximating to a form of co-design where users themselves wielded pens and cameras to document their activities. This turned out not to be possible. The project manager Bente Rugaard Thorsen described the problems in this way: “I thought I could use a lot more creative methods. I wanted to bring the elderly in for interviews where I could ask them to draw their daily life, to ask them to make pictures of what they saw as being difficult for them...some of them could not hold a pen” (Rugaard Thorsen, 2011). Another device, the user diary, also had to be dropped for the same reason. An extended explanation of the problem was provided: “Having studied learning and innovative processes in the bank sector, I used drawing and sketching, but these people [in the study process] were not able to write their own name; many of them suffer from spasms or suffered from paralysis which often affected both sides of their bodies....or had lots of psychological problems…” (ibid.). Verbal communication was sometimes hindered by the effects of stroke where “you have difficulty structuring your speech... or they could have difficulties remembering and explaining…” (ibid.). The participants’ physical weakness limited the time for discussions and taking part in processes such as workshops which were viewed as disruptions to daily routines. Other methods had to be used to “get a glimpse of the challenges of their daily life…” (ibid.). An anthropological approach was taken in response to these problems. This required interviewing and observing approximately 50 participants. The very particular problem of how to convey very personal, intimate problems related to hygiene was encountered. “It would have been too intimate for these people to sit in a workshop, talking about their own problems and to be analysed openly in front of care-givers and also the companies [who would produce the products]” (ibid.). Thus, the strategy of using personas was adopted. While this is less than ideal it allowed for the anthropologist to communicate the needs of the users to the designers. Interview data was generalised and combined into figures that represented typical problems of each sub-group. The information was presented verbally and through posters. Idea cards were used to supplement this information. The investigation phase generated 140 ideas. From these the designers selected 12 core concepts to be refined: a modified shower-head, a virtual rehabilitation training programme, a Scandinavian-style treadmill with a virtual street view, a powered zipper, a harness for rising and sitting, a body dryer, a heated bathrobe, mark-up stickers for training with wheelchairs, a uridom and a folding bath. Two of the concepts - not listed - have been taken on for further development and for reasons of commercial sensitivity have not been presented publicly. There was blending of the research, analysis and conceptual development stages. To quote the project manager: “It’s never a linear process, but in a broad sense, we followed our planned design process” (ibid.). This project showed how both psychological inhibitions and physical frailty were barriers to gaining clear user insights. Specifically, when dealing with frail individuals there is an extra complication when the physical problems involved relate to matters of bodily privacy. The process highlighted a paradox. Subjects reported that it was humiliating to ask for help but in order to reduce their need for help, they must first agree to be helped. And not only helped in the sense of momentary assistance, but helped in the form of discussing their problems.

172

Herriott

17.4 Discussion It has been taken as given that when designers wish to include users then those users will be willing and able to be included. These cases show that the ideal of maximising user input into ID processes is met by the difficulties those users themselves have. Both the Four Municipalities project and Breathe Easy show that the conditions which make it necessary to think inclusively can make inclusivity difficult to implement. The AT user group is not merely slightly or moderately less physically capable than the “standard user group.” They are sometimes much less able to be active contributors to the design process aimed at helping them and which depend on their input. This does not mean that inclusive design is not possible. Rather a change in approaches is needed, what could be termed an economical rationing of user input and an emphasis on less intrusive methods. The Breathe Easy project used physiotherapists as surrogates: they had the means to communicate observations that a designer would not have time to make. But they are not the users we think of when thinking user-centred design. These projects also demand that we re-evaluate certain assumptions. For example, Porter and Porter (2002) list eight fallacies regarding ergonomics. The first fallacy (and the only one relevant here) is: “the design is satisfactory for me - it will therefore be satisfactory for everyone else.” This is generally a good assumption but it is not an absolute. The Breathe Easy team were faced with either self-testing their designs or not testing at all. Their cases indicate that there are some instances where designers can use their experience to make decisions which are (perhaps) less than ideal but which can still produce work which, by any objective standard, is an appreciable improvement on the alternative. To be abstracted from this is that we may need to counterbalance inclusive design’s centripetal force that drags users into the process. Designers in the end have to make design decisions. They may have to use their understanding and judgement, distilled from their other work, to provide substitutes and surrogates for the absent, unwilling or disabled user. In the case of the Four Municipalities project, the intention to use a co-design process was unfeasible. Again, designers had to take on work that ideally users would participate in. The problems involved recall those described by BoydGraber et al. (2006) where a “modified participatory design approach was used in which proxies, that is, speech-language pathologists who work with aphasic individuals, assumed the role normally filled by users.” Underwriting the validity of their work was the use of extensive interviewing and the use of anthropological professionals to guide this. This extensive interviewing provided a rich seam of information with which to cross-reference the findings. Such work is expensive and time consuming and it was made possible only by a favourable context for inclusive design (patient and well-funded clients). Other designers might not be so fortunate. ID and AT have much in common but are not the same. This paper has shown how common linkage of the two elides the differences in methods that are appropriate to designers in the respective fields. Those who are too ill to do the things we take for granted in the “broader mainstream” are by definition a distinctly different group. In metaphorical terms, the toolbox for ID and AT is the

When Users Cannot be Included in Inclusive Design

173

same. However, AT requires certain tools to be used in different ways. Designers must adjust their expectations of what can be achieved with users according to the nature of the capability loss in question. Regarding future work, it might be fruitful to consider re-engineering the inclusive design model to account for the differing capabilities of the target user group. In this case we might want to find out how to focus user-involvement more on the key functionalities and less on areas where surrogates can be found to provide information. One temptation in response to an inclusive design problem is that of devising more guidelines. Such guidelines as designers need are better created on a case by case basis. There is tremendous diversity in capability loss and a general guideline would, in all likelihood, not be able to accommodate this. If anything, this work shows the difficulty inherent in a priori formulations concerning problem-solving.

17.5 References Barbiani M (2011) Interview with author, conducted April 4, 2011 Bichard J-A, Langdon PM, Coleman R (2007) Does my stigma look big in this? Considering acceptability and desirability in the inclusive design of technology products. In: Stephanidis C (ed.) Universal access in human-computer interaction: Coping with diversity, 4th International Conference on Universal Access in Human-Computer Interaction (UAHCI 2007), Springer, Berlin, Germany Boyd-Graber JL, Niklova SS, Moffat KA, Kin KC, Lee JY, Mackey LW et al. (2006) Participatory design with proxies: Developing a desktop PDA system to support people with aphasia. In: Proceedings of CHI 96 Workshop on Human Factors in Computer Systems, SIGCHI, New York, US BSI (2005) Design management systems - part 6: Managing inclusive design - guide. British Standards Institution, London, UK Center for Universal Design (1997) The principles of universal design. North Carolina State University, Raleigh, NC, US Cook A, Hussey SM (2002) Assistive technologies - principles and practices. Mosby Inc, St Louis, MO, US Coleman R, Clarkson PJ, Dong H, Cassim J (2007) Design for inclusivity: A practical guide to accessible, innovative and user-centred design. Gower, Aldershot, UK Clarkson PJ, Coleman R, Hosking I, Waller SD (2007) Inclusive design toolkit. Engineering Design Centre, University of Cambridge, Cambridge, UK Dhiensa J, Machin C, Stone R (2005) Assistive technology: Going beyond the disability. In: Proceedings of the International Conference on Inclusive Design and Communications (INCLUDE 2005), London, UK Green WS, Jordan PW (2002) Pleasure with products. Taylor & Francis, London, UK Jeffery PK (1998) Structural and inflammatory changes in COPD: A comparison with asthma. Thorax, 53(2): 129-136 Linnott A (2011) Influencing the assistive technology marketplace. In: Proceedings of the International Conference on Inclusive Design and Communications (INCLUDE 2011), London, UK Mayagoitia RE, Kitchen S, Harding R, King A, Turner-Smith A (2006) User-centred approach to the design and evaluation of a stair climbing aid. In: Clarkson PJ, Langdon PM, Robinson P (eds.) Designing accessible technology. Springer, London, UK

174

Herriott

Mountain GA, Ware PM, Hammerton J, Mawson SJ, Zheng H, Davies R et al. (2006) The Smart project: A user-led approach to developing applications for domiciliary stroke rehabilitation. In: Clarkson PJ, Langdon PM, Robinson P (eds.) Designing accessible technology. Springer, London, UK Orpwood R, Gibbs C, Adlam T, Faulkner R, Meegahawatte D (2004) The Gloucester smart house for people with dementia - user interface aspects. In: Keates S, Clarkson PJ, Langdon P, Robinson P (eds.) Designing a more inclusive world. Springer, London, UK Orpwood R, Chadd J, Howcroft D, Sixsmith A, Torrington J, Gibson G et al. (2008) Userled design of technology to improve quality of life for people with dementia. In: Langdon PM, Clarkson PJ, Robinson P (eds.) Designing Inclusive Futures. Springer, London, UK Pullin G (2009) When design meets disability. MIT Press, Cambridge, MA, US Porter M, Porter CS (2002) Occupant accommodation: An ergonomics approach. In: Happian-Smith J (ed.) An introduction to modern vehicle design. Butterworth Heinemann, Oxford, UK Reed DJ, Monk A (2006) Design for inclusion. In: Clarkson PJ, Langdon PM, Robinson P (eds.) Designing accessible technology. Springer, London, UK Rugaard Thorsen B (2011) Interview with author, conducted May 31, 2011 Story M, Mueller JL, Mace R (1998) The universal design file: Designing for people of all ages and abilities. Center for Universal Design, North Carolina State University, Raleigh, NC, US Torrens G (2010) Assistive technology design: A mixed methods approach. Loughborough Design School, Loughborough University, Loughborough, UK

Chapter 18 What is Good Design in the Eyes of Older Users? N. Goddard and C. Nicolle

18.1 Introduction With the population of older consumers increasing and with the recent changes in legislation and attitudes towards this group, there have been corresponding changes in product design practice and a growing attempt to adopt an inclusive design approach. This recognises that people can become excluded from using products, services or environments if the needs and capabilities of all potential users are not taken into account. The inclusive design approach has developed from collaborations between industry, designers and researchers. One major influence in this area is the i~design project, whose definition is simply that “inclusive design is better design” (EDC, 2011). The Inclusive Design Toolkit website, a key output from the i~design project, states that a successful product must be “functional, usable, desirable and ultimately profitable” and that a key to good design is to reduce the demand on the user when capabilities decline with age or disability (EDC, 2011). It is also important to consider more emotional aspects, such as social acceptability and whether the potential user would actually want to use or be seen using the product (Keates and Clarkson, 2003). Other authors also emphasise that whilst inclusive design research and practice to date have focused primarily on the physical accessibility and usability of products, a better understanding is required of people’s emotional needs, such as social acceptability and desirability of products (Coleman et al., 2007; Lee, 2010). Similar views regarding the required shift in design focus are reflected in a number of other sources: the need to consider the less tangible human factors such as identity, emotion, delight and selfexpression (Cassim et al., 2007); simplicity, aesthetics, pleasure, personality, conspicuousness and fashion (Pullin, 2009); the product’s visual appearance (Crilly et al., 2004); creating pleasurable experiences (Jordan, 2000; Demirbilek and Sener, 2003); and the importance of the emotional aspects of design for a successful product (Norman, 2004), as well as needs related to specific cognitive conditions (e.g. Baumers and Heylighen, 2010).

P. Langdon et al. (eds.), Designing Inclusive Systems, DOI: 10.1007/978-1-4471-2867-0_18, © Springer-Verlag London 2012

175

176

Goddard and Nicolle

However, knowledge about what is good design can only arise by cooperation between designers and the end users; designers themselves cannot always judge what is good without understanding the point of view of the users (Heylighen and Bianchin, 2010). Therefore, the objective of the current study was to take a usercentred approach to investigate what makes good product design in the eyes of older people themselves and how their criteria for good design compare with those considered by designers and researchers.

18.2 Methods The first stage involved an ethnographic approach, combining semi-structured interviews and observation in people’s homes, enabling an understanding of older users’ individual opinions and use of everyday products in their natural environment. All procedures were approved by the University’s Ethical Advisory Committee. Participants were either emailed or handed (as per their preference) a preinterview questionnaire to complete and return prior to the start of the interview. The first part of the questionnaire collected personal background information to provide context for the interview and analysis. The second part explained a preinterview task: to think about one product they consider of good design and one product they consider of bad design. Whilst a detailed interview schedule was developed to enable a certain level of consistency of questioning across interviews, the interviews were semi-structured in order to provide flexibility to follow the lead of the participant and their own terminology, and also to follow up on any unexpected line of thinking. To enable some degree of quantification of responses, a shuffle card exercise was introduced which required each participant to prioritise a set of 30 criteria, written on cards, by dividing them into three piles in terms of importance to them for good product design. An example of a ‘most important’ pile is shown below in Figure 18.1.

Figure 18.1. Shuffle card exercise to prioritise criteria

177

What is Good Design in the Eyes of Older Users

During the interviews, participants were observed using the products and any difficulties encountered with them. The benefits of taking a more ethnographic approach were proven to be as follows: participants were able to demonstrate how they used products in their normal environment and the difficulties they experienced; participants’ non-verbal cues could be observed to add understanding and context, e.g. laughing or tone of voice; and observation revealed gaps between what people said and what happened in reality. Transcripts of all interviews were made in order to provide a complete and accurate record of the data collected and to enable structured analysis. An initial review of the transcripts was carried out to input to the design of the online questionnaire, hosted on SurveyMonkeyTM. This formed the second stage of the study and enabled a certain degree of validation of the stage one findings. The language used and the themes covered in the online questionnaire were based on both the interviewees’ comments and themes from the literature review. The data from the interviews and the online questionnaire were coded and analysed according to key themes, their source and frequency of occurrence, using the qualitative data analysis software NVivo.

18.3 Results Thirteen interview sessions were conducted amongst people aged 65 or over, lasting between 45 and 90 minutes, including a mix of one-to-one interviews, paired interviews with couples and a mini group of three friends. This resulted in a total sample of 22 participants. During the interviews, participants were observed demonstrating the products and any difficulties encountered with them. Details of the age, gender, living arrangements and impairments of the interview participants are summarised in Table 18.1. Table 18.1. Interview participant details Total

Male

Female

Age 65-74 75-84 85+

7 13 2

1 7 1

6 6 1

Living arrangements Live alone Live with partner

7 15

2 7

5 8

Impairments (moderate or severe) Vision Dexterity (arthritis) Mobility Multiple impairments

10 5 4 8

4 2 2 4

6 3 2 4

178

Goddard and Nicolle

For the online questionnaire, a sample of older users was achieved by obtaining permission to post a link on a number of University of the Third Age (U3A) websites across the East Midlands. 72 respondents completed the questionnaire, although only 54 answered the demographic questions at the end of the questionnaire (Table 18.2). Table 18.2. Online respondents (n = 54 out of 72) Total Male Female

54 14 40

Age 55-64 65-74 75-84 85+

19 25 9 1

The key themes that emerged from analysis of the interview data and the responses to the online questionnaire were based on the frequency of responses and are summarised in Figure 18.2. High level themes, which were then broken down into finer categories, included the elements of good and bad design; elements in purchase decision; comments on aesthetics vs. function; role of family, friends and other people; tricks, solutions and adaptations; and reactions to Good Grips products.

Figure 18.2. Key themes from the research

What is Good Design in the Eyes of Older Users

179

The most important criteria for good design amongst older users in this study are that products are easy to use, they do the job they are supposed to well and are simple to understand. The most likely problems that older users experience with products, and which are therefore seen as elements of bad design, are difficulties getting into some types of packaging and the heavy weight of some products. These findings are not surprising; however, the section below discusses some interesting results which shed a new light on these issues.

18.4 Discussion 18.4.1 Aesthetics Versus Usability From detailed analysis of data from both the interviews and open-ended online questions, it would appear that many older people would prefer, ideally, to have products that were not just easy to use but also looked good - but importantly, only if usability has been delivered. This idea links to Maslow’s hierarchy of needs (Maslow, 1987) which is often used to help understand consumers’ requirements from products. According to this model, once the needs at the lower levels, e.g. safety and comfort, have been satisfied, emphasis can shift to needs at the higher levels, that is, towards the more emotional attributes of a product. Jordan adapted Maslow’s model to a Human Factors perspective, creating a new three-level model with ‘functionality’ at the lowest level, ‘usability’ in the middle and ‘pleasure’ at the highest level (Jordan, 2000). Using this model to interpret the current results, users’ basic needs of functionality (performs the tasks for which it is intended) and usability (easy to use) do not appear to be met. Many are experiencing difficulties and are therefore not able to progress to the higher levels to meet their emotional needs. If the more functional needs were being addressed successfully then the ‘pleasure’ needs, including looks, would become increasingly important. The literature review revealed that several authors believe that the focus for design should shift to delivering fulfilment at the higher levels, for both users in general and older users (e.g., Crilly et al., 2004; Lee, 2010). However, the results from this study would indicate that product design may be leaving older users behind by not delivering successfully at the functional level for these groups. These results support the view that some designers may have become overly concerned with the aesthetics of product interfaces, resulting in problems being caused for people with impairments (Noonan, 2007).

18.4.2 Tricks and Solutions for Packaging Problems The current study indicates that difficulties encountered in getting into different forms of packaging are not uncommon, and for most (but interestingly not all) people a source of irritation. In particular, it would appear that many people have problems using can and jar openers and with opening child-safe bottle tops. Difficulties and alternative solutions identified in the current study are very similar

180

Goddard and Nicolle

to those revealed in the packaging research (e.g. Yoxall et al., 2010), that is, using either physical strategies, for example using a tool such as a knife, or social strategies, for example asking a relative (see next section). In some cases participants appeared to be proud of these solutions and were quite animated whilst explaining them. Is it possible that such problem solving could be beneficial in keeping the older mind active and in giving the user a sense of pride? The packaging study mentioned above also refers to the pride and achievement felt by users regarding their ability to solve problems (Yoxall et al., 2010). This is possibly an interesting area for further research, although the negative consequences of failing to solve a poor packaging ‘puzzle’ would almost certainly be greater than the gains of solving that puzzle.

18.4.3 Involvement of Other People It appeared from the current study that a significant influence on the products used by older consumers comes from other people rather than the users alone. Products were often bought for them by others, or were seen being used by others, or were recommended by others. Other studies have found that people in the 76+ age group were more likely than the younger 65-75 age group to rely on others to choose or purchase new products for them (Burrows et al., 2010). For one participant in the current study, when her son gave her a can opener for Christmas he demonstrated to her how easy it was to use, but when she later tried to use it herself she could not use it. Therefore, an important influence on what other people buy for older users would be the perception the purchaser has of what is usable by, or will be liked by, the user - that is, what they believe makes good design for older users if they are not older users themselves. Further research amongst these purchasers would enable a greater understanding. This study indicated that there is a family (or friend) social dynamic involved that may play an important role in the lives of older people, particularly those living alone, for example when a son or daughter is called upon to help out with problems experienced. Burrows et al. (2011) also found that, given the choice, older adults often decide to involve other people in the various stages of their interaction with new technology. There may be a strong need by older people for the social interaction that these ‘difficulties’ bring with them, perceived or otherwise, and the authors believe, in line with Burrows et al. (2011) that more understanding is required about this context. Further research might therefore be valuable to get a better understanding of this role, its implications and the extent to which product design problems and affordances ‘encourage’ interactions between family members or neighbours.

18.4.4 The Effect of Familiarity on Purchase Several participants were using the same products over many years and because they were so used to a product they would plan to buy the same one again. It is

What is Good Design in the Eyes of Older Users

181

possible that it is more typical for the older generation than subsequent more ‘disposable’ generations to hold on to products for as long as possible. As a result of looking for this familiarity with what they already know, it is possible that older people may avoid innovative designs. This poses a challenge for designers: what can they do to ease the transition for these older users from their very familiar products to the unfamiliar, newer product interfaces? Previous experience with similar products is a strong predictor of usability, and those products that help the user make a reference to the same function on another more familiar device should perform better than those that do not (Lewis et al., 2006; Langdon et al., 2007). Furthermore, what innovators see as providing personal benefit may not be seen as such by the older person. The same can be said by the family of an older person if a family member purchases a new product to replace the old, worn out one, the older person might have preferred to ‘battle on’, rather than change the way they do things. How do we encourage users to ‘battle on’ with the new product long enough to recognise the benefits?

18.4.5 Expectations There was some indication from the interviews that, amongst this older generation, expectations for products to work perfectly are low. In addition, a few participants were quite accepting of the fact that in some instances they would not be able to understand how to use the products: “Why should we get our knickers in a twist because we can’t understand everything that’s modern?” Having to find alternatives or make adaptations in such situations does not appear to be troublesome for many participants. One possible explanation is that the older generation are used to having to ‘make do’ and to adapt existing products to make them usable and to last. However, the Baby Boomer generation are just about to enter the 65+ ‘older user’ category. They are considered to have two distinctive characteristics, individualism and liberalism, which are likely to affect their attitudes to products and product design (Huber and Skidmore, 2003). Compared with the previous generation, that is the current ‘older users’, they may be more likely to complain about products and to expect their individual wants and desires to be satisfied. Whilst the older participants in the interviews appeared to be more accepting of design problems, the younger-old are likely to be less tolerant, more demanding and therefore more likely to complain when they have problems.

18.5 Conclusions and Recommendations Despite the volume of research conducted on product usability and accessibility, results from the current study into what makes good design for older users would indicate that the basic functional needs are in fact still not being met for this user group. In particular, older people are experiencing difficulties with ease of use, packaging and weight of products. As explained by Maslow’s hierarchy of needs,

182

Goddard and Nicolle

the lower level, basic needs have to be addressed before users’ higher level emotional needs can be met. Therefore, the belief that appears to exist from the literature review, that the focus for product design needs to move away from the basics to the more emotional aspects of product design, would not yet appear to be advisable when considering older users. The recommendation is for designers to continue to focus on the basics of accessibility and usability of mainstream products for older users, but designs should of course aim to excel in both function and desirability. A number of other key themes emerged from this study as possibly warranting further investigation. There was some indication that the solutions the users had to find in overcoming difficulties getting into product packaging were giving them a sense of pride. Further research might be useful in gaining a greater understanding of the value of this problem-solving for older users, the possible benefits for keeping the older mind active and the implications for product design. In many cases, other people such as friends or relatives are purchasing everyday products for older users. It might therefore be useful to get a greater understanding of what the purchasers think is good design for the users for whom they are buying, particularly when they may be significantly younger than the users themselves. Another aspect relating to other people is the social dynamic that exists when others are purchasing products for older people, or are called upon to assist them with the products they use. Further research might be useful to understand the role such situations play in instigating social interactions and the importance these have in older people’s lives. Another key theme identified is the desire by many older users to buy products with which they are already familiar. Further research might help understand the extent to which this might hinder their acceptance of innovation and what designers can do to minimise this conflict. However, the most important message from the current study is that the fundamental need to get the basics right for older users will be increasingly critical as the current Baby Boomer generation are beginning to enter the ‘over 65’ category. Compared to the current group of older users they are likely to be more demanding, less tolerant and more prone to complain about any shortfall in product design, whether in functionality or style, in meeting their needs and aspirations.

18.6 References Baumers S, Heylighen A (2010) Harnessing different dimensions of space: The built environment in auti-biographies. In: Langdon PM, Clarkson PJ, Robinson P (eds.) Designing inclusive interactions. Springer, London, UK Burrows A, Mitchell V, Nicolle CA (2010) Out-of-box experiences: an opportunity for inclusive design. In: Proceedings of 5th Cambridge Workshop on Universal Access and Assistive Technology, Cambridge, UK Burrows A, Mitchell V, Nicolle C (2011) Designing in social benefits. In: Proceedings of the International Conference on Inclusive Design and Communications (INCLUDE 2011), London, UK

What is Good Design in the Eyes of Older Users

183

Cassim J, Coleman R, Clarkson PJ, Dong H (2007) Why inclusive design? In: Coleman R, Clarkson PJ, Dong H, Cassim J (eds.) Design for Inclusivity. Gower Publishing, Aldershot, UK Coleman R, Topalian A, Clarkson PJ, Dong H (2007) The business case. In: Coleman R, Clarkson J, Dong H, Cassim J (eds.) Design for inclusivity. Gower Publishing, Aldershot, UK Crilly N, Moultrie J, Clarkson PJ (2004) Seeing things: Consumer response to the visual domain in product design. Design Studies, 25(6): 547-577 Demirbilek O, Sener B (2003) Product design, semantics and emotional response. Ergonomics, 46(13/14): 1346-1360 EDC (2011) Inclusive design toolkit. Cambridge Engineering Design Centre, University of Cambridge, Cambridge, UK. Available at: www.inclusivedesigntoolkit.com (Accessed July 2011) Heylighen A, Bianchin M (2010) Can crap design be inclusive? The case for deliberative design. In: Proceedings of 5th Cambridge Workshop on Universal Access and Assistive Technology, Cambridge, UK Huber J, Skidmore P (2003) The new old. Why baby boomers won’t be pensioned off. DEMOS, London, UK Jordan PW (2000) Designing pleasurable products. Taylor & Francis, London, UK Keates S, Clarkson PJ (2003) Design exclusion. In: Clarkson PJ, Coleman R, Keates S, Lebbon C (eds.) Inclusive design: Design for the whole population. Springer-Verlag, London, UK Langdon P, Lewis T, Clarkson J (2007) The effects of prior experience on the use of consumer products. Universal Access in the Information Society, 6(2): 179-191 Lee Y (2010) Development of the social implications of inclusive design and some thoughts on the next steps. In: Proceedings of the 5th Cambridge Workshop on Universal Access and Assistive Technology, Cambridge, UK Lewis T, Langdon PM, Clarkson, PJ (2006) Investigating the role of experience in the use of consumer products. In: Clarkson PJ, Langdon PM, Robinson P (eds.) Designing accessible technology. Springer, London, UK Maslow AH (1987) Motivation and personality, 3rd edn. Harper and Row, NY, US Noonan T (2007) The overlooked consumers: A discussion paper examining the access, challenges and emerging possibilities for consumer electronics and home appliances. Australian Human Rights Commission, Sydney, Australia Norman DA (2004) Emotional design: Why we love (or hate) everyday things. Basic Books, NY, US Pullin G (2009) When design meets disability. MIT Press, Cambridge, MA, US Yoxall A, Langley J, Musselwhite EM, Rodriguez-Falcon EM, Rowson J (2010) Husband, daughter, son and postman, hot-water, knife and towel: Assistive strategies for jar opening. In: Langdon PM, Clarkson PJ, Robinson P (eds.) Designing inclusive interactions. Springer, London, UK

Chapter 19 Equal Access to Information? Evaluating the Accessibility of Public Library Web Sites in the State of Maryland J. Lazar, B. Wentz, C. Akeley, M. Almuhim, S. Barmoy, P. Beavan, C. Beck, A. Blair, A. Bortz, B. Bradley, M. Carter, D. Crouch, G. Dehmer, M. Gorman, C. Gregory, E. Lanier, A. McIntee, R. Nelson Jr., D. Ritgert, R. Rogers Jr., S. Rosenwald, S. Sullivan, J. Wells, C. Willis, K. Wingo-Jones and T. Yatto

19.1 Introduction Public libraries in the United States have a long and proud tradition of providing access to information for all residents. Public libraries have been the equaliser providing access to books, and both printed and electronic information, regardless of race, gender, religion, economic status, or disability. Since the mid-1990s, public libraries have also been providing direct access to the Internet for patrons who come to visit. And since that first burst of access to the Internet within public library buildings, libraries have been pushing to provide access to their resources through the Internet, so that patrons can search library catalogues, reserve resources and renew materials, and even access digital libraries of documents, all wherever the patrons happen to have Internet access - at home, work, or using mobile phones (Jaeger et al., 2011). People with disabilities often use library resources, but since transportation to public library buildings is often a challenge, many people with disabilities may prefer to use the resources of the library in electronic format. Therefore, it is important to examine the accessibility of public library web sites and determine whether or not the web sites are in compliance with public laws related to web accessibility. This chapter presents a research study where the home pages of all 24

P. Langdon et al. (eds.), Designing Inclusive Systems, DOI: 10.1007/978-1-4471-2867-0_19, © Springer-Verlag London 2012

185

186

Lazar et al.

public library system web sites in the state of Maryland were evaluated for accessibility.

19.2 Related Literature 19.2.1 Libraries and Accessibility Public libraries in the United States actually have a long history of involvement in the area of accessibility. In the past, this was often focused on physical issues, such as wheelchair access to library buildings, and alternative formats for printed materials (Lazar et al., 2011). For instance, the Library of Congress’s National Library Service (NLS) for the Blind and Physically Handicapped, founded in 1931, focuses specifically on development of and distribution of library resources in alternative formats, such as large-print, braille, and audio (Dziedzic, 1983). The American Library Association notes that equal access to information is extremely important and a significant challenge facing modern libraries (American Library Association, 2011). So there is a history of concern about both physical building accessibility, as well as the accessibility of library content material. However, there is less of a documented tradition about web accessibility for public library resources. This concern needs to include both the public library web sites, and any digital libraries offered through the library web site to library patrons (Bertot et al., 2006). These digital libraries and other services, since they are typically provided by outside firms, are sometimes the hardest to manage in terms of accessibility, since they are out of the control of the public library (Byerley and Chambers, 2002). Recent research documents that many of these digital libraries continue to be inaccessible (Tatomir and Durrance, 2010). Typically the only way to enforce web site accessibility for externally-managed digital libraries is through closelymonitored procurement/acquisition processes. There is surprisingly little data about the accessibility of public library web sites in the United States. The one previously-published study examining the accessibility of public library web sites, during the year 2000, found that only 14/74 of public library home pages were fully accessible (Lilly and VanFleet, 2000), and that data is more than 10 years old. There have been more published studies about the accessibility of university library and school library web sites, such as Spindler (2002) and Comeaux and Schmetzke (2007), who both found that a majority of the home pages that they evaluated were not accessible. The largest annual data collection effort related to US public libraries and the Internet since 1994, known as the Public Libraries and the Internet (and more recently known as the Public Library Funding and Technology Access Survey), does not include any questions related to technology accessibility (http://www.plinternetsurvey.org/). Therefore, the authors decided that it was important to examine public library web site accessibility within the state of Maryland.

Equal Access to Information?

187

19.2.2 Library Structure in Maryland In the state of Maryland, the Division of Library Development and Services is a part of the Maryland State Department of Education, serves a central coordination role, and administers state and federal programs related to libraries. However each of the 24 public library systems is administered by the county in which it is located (Maryland State Archives, 2011). Note that Baltimore City actually is not part of a county, so it is considered a separate administrative entity, and has a separate library system (the Enoch Pratt Free Library) from the Baltimore County Public Library. Typically, each county public library system has multiple branches, but one web site. There are actually multiple laws in place that require public library web sites to be accessible. For instance, the Maryland Information Technology Nonvisual Access (MD IT NVA) Regulatory Standards require all state government web sites to be accessible (Maryland State Department of Information Technology, 2011), and Title II of the Americans with Disabilities Act (a federal law) covers state and local government services. The technical specifications for web sites, from the Maryland Information Technology Nonvisual Access regulations, are the same as the Federal Section 508 regulations, which cover Federal information technology.

19.2.3 Website Accessibility Users with perceptual and motor impairments often use alternate input or output approaches to access web sites. For instance, low-vision users may utilise screen magnification. Blind users typically access web sites through the use of text-tospeech (or screen reader) software, which reads the content of a web page in an audible manner to the user in a linear fashion. Examples of screen reader software include JAWS, WindowEyes, NVDA, and VoiceOver. Screen reader software is the dominant method of access because Braille literacy is extremely low among blind people (National Federation of the Blind, 2011) . Users with limited dexterity may use alternative keyboards, no pointing device, or speech recognition. Users who have spinal cord injuries and/or paralysis may utilise speech recognition or eye tracking. Deaf or hard of hearing users may rely on captioning for videos, or transcripts for audio. Web developers, as well as public policies, tend to focus on the accessibility needs of people with perceptual and/or motor impairment, rather than cognitive impairment, because, while there is nearly a 30-year history of interface design for people with perceptual and/or motor impairment, there is a much smaller history of interface design for people with cognitive impairment, and there are currently no design standards for such people (Lazar, 2007). The international design standards for web content are the Web Content Accessibility Guidelines (WCAG), from the World Wide Web Consortium, and the current version is WCAG 2.0. Web Accessibility regulations in countries such as the US, Canada, UK, and Australia, are all based on the WCAG. Problems arise when web site interfaces are not designed to standards (such as Section 508 in the US or the

188

Lazar et al.

Equality Act in the UK). Poor design results in web site content being inaccessible to users with disabilities, creating exclusion and discrimination.

19.3 Research Methodology Expert inspections of web site interfaces were the methodology chosen for this research. Automated software (such as A-Prompt, Deque WorldSpace, and WebXACT) is often used for accessibility testing, but automated evaluations are not as accurate as expert inspections or user-based testing. This is because some of the most common accessibility problems (such as link titles or alternative text that do not fit the context) cannot be determined in an automated fashion. The most effective form of accessibility evaluation is done with a screen reader, where experts without visual impairments evaluate compliance with technical and legal standards for web sites (Mankoff et al., 2005). The accuracy of this method increases when multiple evaluators inspect the same web site interface (Lazar et al., 2010). Expert inspections tend to focus on technical accessibility, rather than true usability. On the other hand, usability testing, involving people with disabilities, provides a higher understanding of ease of use. Usability testing is task-focused, where users attempt to complete a series of tasks. However it is not inspection focused. As a result, usability testing may not ascertain compliance with the law, since the users may not actually be able to test all of the various components of the regulations as a part of their tasks. Due to the various strengths and weaknesses, the combination of expert inspections and usability testing with people with disabilities is ideal, but it is not performed often enough, due to the time and cost involved. The expert evaluations that were conducted utilised the web site accessibility standards of Section 508 (1194.22) of the US Rehabilitation Act, which comprise 16 guidelines, identified as paragraphs “a” through “p” (which are the same technical standards as from Maryland state law). Expert inspection consisted of an expert with vision inspecting a web site using a screen reader and a check for compliance with each specific guideline from Section 508. The process was guided by the “Absolute Minimum Accessibility Inspection” which was developed by one of the authors and used effectively in other accessibility evaluations (Lazar et al., 2010). Each home page was evaluated separately by 5 individuals using JAWS 12. None of these individuals were blind, but all had previous experience evaluating web sites for accessibility using screen readers. All evaluators then met, discussed their individual evaluations, and compiled one meta evaluation, which typically has a higher level of validity than one only. Each paragraph of the guidelines was weighted equally. There was no measurement of the number of times that a specific paragraph was violated, only the existence of a paragraph violation, because the number of paragraphs violated is considered to be a more accurate measurement of accessibility (Lazar et al., 2003; Lazar et al., 2010). Table 19.1 presents a list of the guidelines, along with a short description of each (note that the descriptions are those of the authors, and not a part of the US regulations).

Equal Access to Information?

189

Table 19.1. Description of each of the 16 paragraphs of the Section 508 web guidelines (a) Text Equivalent (have a text equivalent for any graphical elements) (b) Synchronised Equivalent Alternatives (have captioned video, transcripts of any audio, or other alternatives for multimedia) (c) Use of Colour (colour should not be used as the only method for identifying elements of the web page or any data) (d) Organisation (style sheets are encouraged, but users should still be able to utilise a web page when style sheets are turned off) (e) Redundant Text Links on Server-Side Image Map and (f) Client-Side Image Maps (redundant clickable links for server-side image maps, and accessible client-side image maps are preferred) (g) and (h) Row and Column Headers (use appropriate headers and mark up to allow easy navigation of a table) (i) Frames (title all frames and label all frames for easy identification and navigation, e.g., use “navigation” “main content” and “search” rather than “top” or “bottom”) (j) Screen Flicker Frequency (limit or eliminate the use of flickering, which can provoke seizures) (k) Text-Only Page Default (if a web page cannot be made accessible, provide an equivalent text-only page, and make sure it is kept up to date) (l) Scripting Languages (make sure that equivalents for any non-accessible scripting are included, e.g., for those who are not using pointing devices) (m) Linked Plug-In or Applet (if any plug-ins are required, make sure to provide a link to an accessible version of the plug-in) (n) Online Electronic Forms (all forms must be properly labelled and accessible) (o) Method to Skip Repetitive Navigation Links (all web pages should have a link which allows a user to skip directly to the main content, bypassing any site navigation information) (p) Alerts on Timed Responses (if any page responses are timed, the user should be given the opportunity to indicate that more time is needed)

19.4 Results 19.4.1 Section 508 Paragraph Violations The evaluations of the home page of each of the 24 county library web sites revealed that all the home pages violated at least two or more paragraphs of the Section 508 Guidelines, and some library home pages violated as many as six paragraphs. Table 19.2 illustrates the category violation for each library home page.

Lazar et al.

190

Table 19.2. Section 508 paragraph violations of Maryland County Library websites Paragraph:

a

b

Alleghany Anne Arundel Balt. City

X

X

Balt.County

X

Calvert

X

Caroline

X

c

d

e f

g h

X

i

j

k

l

m

n

o

X

X

X

X

X

X

X X X

X

Carroll

X

Cecil

X

Charles

X

Dorchester

X

Frederick

X

Garret

X

Harford

X

Howard

X

Kent Montgomer y Prince George’s Queen Anne’s Somerset

X

X

St. Mary’s

X

Talbot

X

X X

X

X

X

X

X

X

X

X

X

X

X

X

X

X X X

X

X

Worcester

X

X X

X

X

X

X

X

X

Wicomico

X

X X

X

Washington

p

X X

X

X

X

X

X

X

X

X

X X X

X

X X

X

X

X

X

The most common violations were paragraphs “a,” “n,” and “o.” Paragraph “a” requires there be a text equivalent for a graphical element on a web page. In other words, if an image or graphic is used or included on a web page, it should have alternate text that will be available for screen reader users or any other user who cannot see the image. Paragraph “n” requires that web site forms and form fields be accessible and properly labelled. The most common violation of paragraph “n” occurs when there are not clearly understandable labels associated with form input fields on a web page. The frustration of inaccessible forms could be illustrated by imagining filling out an order form with no idea which text box requires the name,

Equal Access to Information?

191

address, phone number, or credit card information. Paragraph “o” requires that there is a method to skip repetitive navigation links. With web pages often having many navigational links before the main content is reached, a screen reader that is reading the content to a user in a linear manner will have to read through each link before reaching the main content. This can be frustrating and time-consuming. The common solution is to include a link at the top of every page that will allow users to skip to the main content of the page. It is important that this link is not hidden since it can also assist in navigation for users with motor impairments.

19.4.2 Examples of Violations All of the violations in each category are violations of a particular accessible design standard. This can best be illustrated through examples of violations that were discovered on the web sites of various Maryland county library systems. For example, the Alleghany County public library had three videos on its home page that had no captions and could not be accessed by the keyboard alone and therefore had an inaccessible plug-in (violating paragraphs “b” and “m”). Anne Arundel county’s library system violated paragraph “l” with a drop-down navigation menu that is not accessible by using a keyboard alone (Figure 19.1).

Figure 19.1. The drop-down menu is not accessible without the use of a mouse

Carroll, Charles, Frederick, and Harford county public libraries all had web sites that displayed flash slide show content that was inaccessible (with no alternative). Howard and St. Mary’s county public libraries violated paragraph “n” with an online form that is improperly labelled. Figure 19.2 shows an accessibility problem identified on the Howard County Public Library’s home page, where rotating links are not accessible with a keyboard alone.

Lazar et al.

192

Figure 19.2. The rotating flash events and links are not accessible

Washington County’s web site displayed an online event calendar that was not accessible, and Wicomico county’s web site used a table that contained no headers (a violation of paragraphs “g” and “h”). Many of these violations could prevent individuals with disabilities from accessing all or part of the content that is available on the library home pages.

19.5 Discussion The data from this study illustrate the inequity of access to public library home pages, through an accessibility evaluation of home pages of public library systems in one state of the United States. It is likely that similar results might be discovered if the public library web sites were evaluated in other states. With both state and federal laws requiring web accessibility, and with the American Library Association highlighting equal access to information as something of significant importance, this is one area in which a higher level of accessibility is clearly important and could easily be achieved. The majority of the violations fell within the categories of alternate text for graphics, labelling of form fields, and the lack of

Equal Access to Information?

193

skip navigation links. These are all very simple changes that could be implemented with little to no cost on the part of the public libraries. However, the authors acknowledge that in the current environment, with budgets being cut while there is a higher demand from patrons for services due to the weak US economy, any new efforts or initiatives, even simple changes to public library home pages to improve accessibility, may not be implemented quickly due to the heavy demand on library staff. The authors believe that the key to improving web accessibility is awareness and transparency. A good place to start is by posting a web accessibility statement on the home page, noting what accessibility features exist on the web site, what design guidelines were used for ensuring accessibility, what ongoing evaluation (user testing, expert reviews, or automated reviews) is used to ensure ongoing accessibility, and who users should contact if they experience any problems related to accessibility. Research has previously established a relationship between strong accessibility statements and actual higher levels of web accessibility at the state level (Rubaii-Barrett and Wise, 2008), although this relationship was not observed in federal web sites (Olalere and Lazar, 2011). Only three of the county public library web sites have accessibility statements: Baltimore City, Carroll County, and Montgomery County. The Baltimore City and Montgomery County accessibility statements reference applicable laws and policies, but the Carroll County Public Library web site is extremely vague and merely mentions an effort to make the web site accessible to everyone. Two other counties, Dorchester and Wicomico, mention general library accessibility, but there is no mention of web site accessibility. There are other ideas for improving public library web site accessibility. Potentially, web accessibility training could be provided at future meetings or conferences where many employees of public libraries in Maryland are present. In addition, we suggest that while expert inspections were used in this study, it would also be useful to do usability testing involving library patrons with disabilities. Year-by-year comparisons of web site accessibility can also be helpful to determine if progress is being made in improving public library web site accessibility. All of these steps can help raise awareness and hopefully help improve public library web site accessibility.

19.6 References American Library Association (2011) Access. Available at: http://www.ala.org/ ala/issuesadvocacy/access/index.cfm (Accessed 9 November 2011) Bertot J, Snead J, Jaeger P, McClure C (2006) Functionality, usability, and accessibility: Iterative user-centered evaluation strategies for digital libraries. Performance Measurement and Metrics, 7(1): 17-28 Byerley S, Chambers M (2002) Web-based library databases for non-visual users. Library Hi-Tech, 20(2): 169-178 Comeaux, D, Schmetzke A (2007) Web accessibility trends in university libraries. Library Hi Tech, 25(4): 457-477

194

Lazar et al.

Dziedzic D (1983) Public libraries. In: Cylke F (ed.) That all may read: Library service for blind and physically handicapped people. US Library of Congress, Washington DC, WA, US Jaeger P, Thompson K, Lazar J (2011). Research in practice: The Internet and the evolution of library research: The Perspective of One Longitudinal Study. The Library Quarterly (in press) Lazar J (2007) Introduction to universal usability. In: Lazar J (ed.) Universal usability: Designing computer interfaces for diverse user populations. John Wiley & Sons, Chichester, UK Lazar J, Beavan P, Brown J, Coffey D, Nolf B, Poole R et al. (2010). Investigating the accessibility of state government web sites in Maryland. In: Langdon PM, Clarkson PJ, Robinson P (eds.) Designing inclusive interactions. Springer, London, UK Lazar J, Beere P, Greenidge K, Nagappa Y (2003) Web accessibility in the Mid-Atlantic United States: A study of 50 web sites. Universal Access in the Information Society, 2(4): 331-341 Lazar J, Jaeger P, Bertot J (2011) Persons with disabilities and physical and virtual public library settings. In: Bertot J, Jaeger P, McClure C (eds.) Public Libraries and the Internet: Roles, Perspectives, and Implications. Libraries Unlimited, Santa Barbara, CA, US Lilly E, VanFleet C (2000) Measuring the accessibility of public library home pages. References and User Services Quarterly, 40(2): 156-165 Mankoff J, Fait H, Tran T (2005) Is your web page accessible? A comparative study of methods for assessing web page accessibility for the blind. In: Proceedings of the 23rd ACM Conference on Human Factors in Computing Systems, Portland, OR, US Maryland State Archives (2011) Maryland libraries at a glance. Available at: http://www. msa.md.gov/msa/mdmanual/01glance/html/library.html (Accessed 9 November 2011) Maryland State Department of Information Technology (2011) Maryland information technology nonvisual access regulatory standards. Available at: http://doit. maryland.gov/policies/pages/nva.aspx (Accessed 9 November 2011) National Federation of the Blind (2011). How many children in America are not taught to read? Available at: http://www.nfb.org/nfb/braille_initiative.asp (Accessed 9 November 2011) Olalere A, Lazar J (2011) Accessibility of US federal government home pages: Section 508 compliance and site accessibility statements. Government Information Quarterly, 28(3): 303-309 Rubaii-Barrett N, Wise L (2008) Disability access and e-government: An empirical analysis of state practices. Journal of Disability Policy Studies, 19(1): 52-64 Spindler T (2002) The accessibility of web pages for mid-sized college and university libraries. References and User Services Quarterly, 42(2): 149-154 Tatomir J, Durrance J (2010) Overcoming the information gap: Measuring the accessibility of library databases to adaptive technology users. Library Hi Tech, 28(4): 577-594

Chapter 20 Clustering User Data for User Modelling in the GUIDE Multi-modal Set-top Box P.M. Langdon and P. Biswas

20.1 Introduction The EU GUIDE project is aimed at developing a research-based prototype of an inclusive and accessible set-top box running on conventional hardware. It utilises advanced user modelling and simulation in conjunction with a single layer interface that permits a wide range of input devices and modalities and output formats and modes. This paper addresses part of the research behind the development of an advanced user model in order to develop the software framework, namely the user profile clustering. A range of user centred design techniques, including focus group, survey and several iterative stages of design trials have been used overall to create the requirements specification for the technology framework under development. However, a key module is the inbuilt user model that allows the system to classify users on the basis of their impairment level, user interface (UI) behaviour, preferences and context. These profiles are pre-generated using an advanced cognitive, perceptual and movement simulation, using parameters clustered from the actual user data. This chapter examines the contextual background literature, briefly describes the inclusive user centred design process and shows how this process generated the data necessary for clustering for cognitive, perceptual and motor-impaired user modelling.

20.1.1 Background and Motivation GUIDE primarily addresses end users who have mild to moderate impairments. This target group is identified within the context of literature concerning inclusive design, accessibility and the digital barriers excluding older technology users from access to digital technology. An inclusive approach considers the visual, hearing and touch perceptual capabilities of users in combination and attempts to quantify capability variation. This approach is well suited to multimodal interface design

P. Langdon et al. (eds.), Designing Inclusive Systems, DOI: 10.1007/978-1-4471-2867-0_20, © Springer-Verlag London 2012

195

196

Langdon and Biswas

where capability impairment in one modality may be compensated for using other modalities (e.g. hearing impaired users may use visual captioning, avatars and sign language) or where multiple impairments can be addressed using the performance gains arising from multimodal interfaces. For instance, gesture, speech and face recognition for input can be used with sound, touch and graphical output displays to enhanced effect.

20.1.2 User Centred Design and the Design Process End user requirements for ageing people and those with impairments were collected using a mixed methods approach based on the advantages of triangulation of data sources (Langdon et al., 2003; Flick, 2006). In essence, this approach does not commit to a single source of data or a single data collection approach. Instead data is collected from multiple approaches, for example: literature review, quantitative analysis of data from forums, user trials, user surveys, and questionnaires, qualitative analysis of observational data from user forums or interviews, video from user trials and usage ethnography (Langdon et al., 2003; Flick, 2006). In the case of GUIDE, the framing of the project as a design problem constrains the triangulation and assists it by directing the focus of comparison on the design of the final interactions between user, system, and technology and usage context. The particular methods used are shown in Figure 1. The sampling strategy employed was opportunistic, and stratified, choosing data sources according to convenience and resource limitations. In particular, much of the work took advantage of empirical and observational trials. However, the combination of multiple sources permits triangulation and thus increased validity and reliability of qualitative findings (Miles and Huberman, 1994).

Figure 20.1. Schematic representation of development of requirements

We focus on the lower region of Figure 20.1, on the data sources and processes of clustering user data from: objective performance data from user trials; user screening survey data; existing data from input device validation trials. The clustering of data from these sources was intended to define groupings that could be traced by membership to specific capability levels in specific modalities.

Clustering User Data for User Modelling

197

20.1.3 User Centred Design and the Design Process The User Model in GUIDE runtime system is incorporated into the framework as a set of pre-calculated user profiles based on the user data clusterings mapping user characteristics. It is developed using the modelling (Benyon and Murray, 1993) but does not include the detailed perception, cognition and motor process models. It consists of the following two parts: • •

A set of sample profiles developed using the clusters described here. The user profile will be described in a machine-readable and human-readable format in accordance to a standard set of variables. A table mapping user characteristics to interface and adaptation parameters developed offline using the simulator and user trials.

To address the limitations of existing user modelling systems, we have developed a simulator (Biswas et al., 2011). The Environment model contains a representation of an application and context of use. It consists of the Application model containing a representation of interface layout and application states for the current task undertaken by a user that will be simulated by breaking it up into a set of simple atomic tasks. The Context model represents the context of use such as background noise, illumination and so on and a Device model decides the type of input and output devices to be used by a particular user and sets parameters for an interface.

Figure 20.2. Architecture of the simulator

The User model simulates the interaction patterns of users for undertaking a task analysed by the task model under the configuration set by the interface model. It uses the sequence of phases defined by the GOMS Model Human Processor (Card, Moran, Newell, 1983). • • •

the perception model simulates the visual perception of interface objects. It is based on the theories of visual attention; the cognitive model determines an action to accomplish the current task. It is more detailed than a GOMS model but not as complex as other cognitive architectures; the motor behaviour model predicts the completion time and possible interaction patterns for performing that action. It is based on statistical analysis of screen navigation paths of disabled users.

198

Langdon and Biswas

The details about users are stored in xml format in the user profile following an explicit ontology.

20.2 Data Origins and Clustering Approach There were several phases to the data clustering approach: • • • • •

obtain and collate survey data and user trial data; cluster the survey data by modality: Vision, Hearing, Cognition, Physical; reduce the dimensionality of the data set by eliminating non-significant variables in a k-means clustering; take the resulting clusterings and characterise the cluster centres in terms of the combined contributions to the clusters; repeat for User trial data.

20.2.1 Sampling Strategy For the purpose of devising a working sampling strategy for early pilot trials it was necessary to adopt a stratified sampling strategy, screening participants who took part in the GUIDE user survey and allocating them to age and capability ranges of interest and that were required for analysis. Initially participants were screened into groups according to the severity of their perceptual, cognitive and motor limitations using conventional standardised tests, such as the US 4.2.1 AMA Uniform Measurement Scales (Figure 20.2). For example, the following classification is taken from Dementer and Anderson, (2003). Table 20.1. AMA General ability ranges

199

Clustering User Data for User Modelling

The relative proportions of participants in sampling categories were therefore initially decided by opportunistic sampling. This process was further informed by an Inclusion and Disability analysis based on exclusion calculations of the proportion of an exemplar EU population (UK GB) who might be expected to fall into the categories of high, medium or low impairment, and who consider themselves disabled in some capability area. These figures were derived using calculations based on the data from the only known complete data set of capability variation publicly available through the UK Office of National Statistics (ONS). This representative national data was from a UK disability survey (Disability Follow-up Survey, DFS, 1997). DFS carried out in 1997 and intended to establish the prevalence and severity of quality of life problems arising from functional impairments (Grundy et al., 1999). Table 20.2 exemplifies the result of this approach for vision and hearing alone but figures were also generated for cognition, and physical movement. Table 20.2. Population and proportion excluded for each of the age ranges and impairment levels (hearing and vision only) TYPE OF CAPABILITY

SUBTYPE CAPABILITY

Vision Perception

Hearing

OF

LEVEL OF IMPAIRMENT

PROPORTION DISABLED POPULATION EXCLUDED THIS LEVEL 40-60

OF AT

PROPO RTION DISABLED POPULATION EXCLUDED THIS LEVEL 60-90

High

0.04%

0.21%

Medium

0.44%

2.5%

Low

1.86%

7.41%

No impairment

3.21%

11.88%

High

0.09%

0.55%

Medium

0.42%

2.49%

Low

2.28%

10.99%

No impairment

4.1%

18.69%

OF AT

Finally, the gender balance for the population age ranges and the proportion of individuals experiencing some impairment for each gender were also calculated using the exclusion estimation from the UK inclusive design exclusion calculator (IDtoolkit, 1999; Waller et al., 2010). Because the sampling strategy of older and impaired users in the age groups of interest was, by necessity, opportunistic, these calculations were used as a broad indication of the relative sizes of the samples required in each category. This meant, for example, that roughly twice as many individuals would be sampled in the older age group compared with the younger age group and that the sample should be biased towards female rather than male participants. In the actual sampling these proportions were only partially achieved, the predominant population being older female, although the age group sampling was partially successful. The extent which this reflects the sample from Northern Spain was unknown but comparisons with other populations in the UK and Germany will be available in the second iteration of user trials to come.

200

Langdon and Biswas

20.2.2 Survey Data Analysis The full Stage 1 survey data set contained 46 users and a very large number of variables. The full set is too large to be illustrated. The variables were then subjected to k-means cluster analysis for each modality, again because of the large number of variables, and this subsequently allowed the removal of variables that did not significantly contribute to the high, medium and low clustering. The trial data contained a prohibitively large number of missing data points for analysis. 20.2.2.1 K-Means Clustering After collecting survey data from them we have defined the profiles of low, medium and high levels of severity. In defining the profiles we have used k-means clustering with a value of k =3. K-means clustering algorithm works like an Expectation Maximisation (EM) algorithm, which partitions n observations into k clusters in which each observation belongs to the cluster with the nearest mean. Each cluster centre defined a profile and only the significant variables (p<0.01) were used to define the profiles. 20.2.2.2 Removal of Non-significant Variables Removal of the variables that did not significantly contribute to the high, medium or low clusters gave the following set. This set of variables appears to be consistent with literature and has face and completeness validity. Table 20.3. Final significant variables for each modality SURVEY DATA REMOVING NON-SIGNIFICANT VARIABLES VISION

HEARING

COGNITION

MOTOR

Close vision: able to read perfectly close to

Able to hear a sound of 500Hz

TMT test of cognitive function

Diagnosed mobility impairment

Diagnosed visual impairment

Able to hear a sound of 2000Hz

Amount of muscular weakness

General eyesight

Able to distinguish a conversation from a noisy back ground

Able to write

Able to see at night

Amount of limb tingling Amount of limb rigidity

Clustering User Data for User Modelling

201

20.2.3 Characterising Cluster Centres The following section uses the result of the k-means clustering to characterise what a typical member of a high medium or low impairment group would be capable of, based on the calculated cluster centres. The profiles of users will ultimately be formed of a combination of all modalities simultaneously such that a specific grouping may represent capability levels for users perceptual, cognitive and motor capability ranges. Clearly, a particular user may present, for example, low vision, medium cognition and high motor capability levels and it is this that will be initially used to index the GUIDE user profiles following a successful initialisation of the application. This may entail a short visual acuity, cognition, physical or hearing test, requested by GUIDE. Following this, adaption may utilise more accurate quantitative data to further locate the user’s profile. 20.2.3.1 Cognitive Tests Before presenting the clustering result it is necessary to describe the cognitive measures used in more detail. Measures for perception and movement are essentially self-explanatory. However, cognitive measures were less obvious. AVLT: The Auditory verbal learning test (AVLT) consists of 15 words which have to be learned during 5 trials. After every trial, the respondent is asked to recall as many words as possible from the list. Two measures were used in GUIDE: 1. AVTL 1: this is the short term memory required to recall as many words as possible. 2. AVLT 5 - 1: this is the learning potential variable. The results of this test are interpreted taking into account the age of the respondent: • <74 years old: the expected average on the first trial is 4; the expected average on the AVLT 5-1 is 5 • 75-84 years old: the expected average on the first trial is 3; the expected average on the AVLT 5-1 is 5 • > 84 years old: the expected average on the first trial is 3; the expected average on the AVLT 5-1 is 3 DST: The Digit-symbol test (from WAIS): It measures information processing speed (Weschler et al., 1999), requiring rapid matching of learnt digit to symbol pairings. • •

55-69 years old: 36-43 symbols in 2 minutes expected >70 years old: 18-22 symbols in 2 minutes expected

TMT: The Trail Making Test (TMT) measures Cognitive executive function. This can be interpreted as the ability to use working memory and the number of tasks that can be kept in mind and dealt with at once. The measure is the time needed to carry out the task of connecting the drawn circles together in numerical or symbolic order, with a pencil. (TMT, 1944). • •

Less than 50 seconds: no impairment 50-60: low impairment

202

Langdon and Biswas

• •

60-70 : = medium impairment > 70: = high impairment.

However, the normalised test literature occasionally conflicted with the clusterings resulting from the k-means approach. In some tests some of the scores interpreted as a level of impairment in the clustering table were not indicated as impairment in the test rubric. This was assumed to be attributable to differences in sampling and inevitable inaccuracies of clustering with small samples. 20.2.3.2 K-Means Cluster Centres The following tables show what levels of the variables characterise the cluster centres for each modality. Table 20.4. K-Means cluster centres for the visual and hearing variables Vision – Cluster Centres LOW

MEDIUM

HIGH

Close Vision: able to read until line Distant_Vision: able to read until line General eyesight

20/20 5/5 good

20/60 5/5 excellent

20/80 5/20 normal

Seeing at distance Seeing at night Colour perception

good normal good

poor poor bad

poor poor bad

Hearing – Cluster Centres LOW

MEDIUM

HIGH

Able to hear a sound of 500Hz? Able to hear a sound of 1Khz?

Yes Yes

Yes Yes

No Yes

Able to hear a sound of 2Khz?

Yes

Yes

Yes

Able to hear a sound of 3Khz? Able to hear a sound of 4Khz?

Yes Yes

Yes Yes

Yes No

Able to hear a sound of 8Khz? How do you define your hearing..? Hear conversation from a noisy background

Yes excellent

No good

No poor

excellent

normal

normal

Hear movie dialogue

excellent

good

poor

Hear ringing noises Hear phone ringing in a movie

excellent

good

normal

excellent

good

poor

Clustering User Data for User Modelling

203

Table 20.5. K-Means cluster centres for the cognitive variables Cognition – Cluster Centres LOW

MEDIUM

HIGH

TMT seconds AVLT series 1

30 6

49 5

136 3

AVLT series 2

10

7

5

AVLT series 3

11

9

6

AVLT series 4

13

9

6

AVLT series 5

14

10

7

AVLT learning potential

7

5

3

WAIS - digit-symbol test

75

30

20

Table 20.6. K-Means cluster centres for the motor variables Motor – Cluster Centres LOW

HIGH

no never

MEDIUM Hernia,slippeddisc occasional

Able to write

No difficulty

No difficulty

Mild

Able to push a heavy door Able to change a bulb

No difficulty No difficulty

No difficulty No difficulty

Mild Mild

Use of transport (bus, etc.) Amount limb tingling

No difficulty No difficulty

No difficulty Mild

Moderate Mild

Amount limb weakness

No difficulty

Mild

Moderate

Amount limb rigidity

No difficulty

Mild

Moderate

Diagnosed mobility impairment Amount of muscular weakness

no quite a lot

20.3 Discussion This chapter has described the context, antecedents and preliminary results of the GUIDE cluster analysis from a user-centred custom-designed data collection. A data snapshot of survey data was analysed for highly correlated variables and on this basis single variables with utility were chosen and used to represent vision, hearing, cognition and physical movement capabilities. Initial results suggest that the simple k-means clustering approach yields high, medium and low impairment cluster centres that can be characterised on the basis of the significant variables that contribute to them. Both the characterisations and clusterings present face validity and completeness validity, but this initial data set was smaller than the expected final data set. However, we conclude that such a clustering is, in principle, capable of use in the GUIDE framework when combined and indexed within the GUIDE user model. The result will be that the user interface presented to the user will be tailored to their perception, cognition and movement capabilities and that sound, visual display and other outputs will be combined multimodally. Future clustering could take advantage of quantitative parameters of clustering variables to position users in the profile space. Clusterings for the survey data (n = 46) were consistent also with expected distributions based on the users sampled

204

Langdon and Biswas

and age estimates. Further work will continue to enlarge the data sets and employ advanced clustering techniques, such as fuzzy and overlapping clusters, to group both the survey and trial data as candidate sets for user profiling.

20.6 References Benyon D, Murray D (1993) Applying user modeling to human computer interaction design. Artificial Intelligence Review, 7(3/4): 199-225 Biswas P, Robinson P, Langdon PM (2011) Designing inclusive interfaces through user modelling and simulation. International Journal of Human Computer Interaction. Taylor & Francis, 1044-7318 Card SK, Moran TP, Newell A (1983) The psychology of human-computer interaction. Lawrence Erlbaum Associates Publishers, Hillsdale, NJ, US Demeter SL, Andersson GBJ (2003) Disability evaluation, 2nd edn. Mosby, St Louis, MO, US Flick U (2006) An introduction to qualitative research, 3rd edn. Sage Publications Ltd, London, UK Grundy E, Ahlburg D, Ali M, Breeze E, Sloggett A (1999) Disability in Great Britain: Results from the 1996/97 disability follow-up to the family resources survey. Technical Report 94. Department of Social Security, Leeds, UK Inclusive Design Toolkit. Available at: http://www.inclusivedesigntoolkit.com/betterdesign/ downloads/exclusioncalc.html (Accessed 20 November 2011) Langdon PM, Aurisicchio M, Clarkson PJ, Wallace KM (2003) An integrated ethnographic and empirical methodology in a study of knowledge searches in aerospace design. In: Proceedings of the 14th International Conference on Engineering Design (ICED’03), Stockholm, Sweden Miles MB, Huberman AM (1994) Qualitative data analysis. Sage Publications Ltd, London, UK TMT (1944) Army individual test battery. Manual of directions and scoring. War Department, Adjuvant General’s Office, Washington, DC, US Waller SD, Williams EY, Langdon PM, Clarkson PJ (2010) Quantifying exclusion for tasks related to product interaction. In: Langdon PM, Clarkson PJ, Robinson P (eds.) Designing Inclusive Interactions. Springer, London, UK Wechsler D (1999) Adaptación española de la Wechsler Adult Intelligence Scale-III (WAIS). TEA, Ediciones, SA, Madrid

Part V

Designing Inclusive Architecture

Chapter 21 Inclusive Built Heritage as a Matter of Concern: A Field Experiment A. Heylighen

21.1 Introduction Europe’s built heritage is the world’s most diverse and rich patrimony, and an important component of individual and collective identity. Its societal relevance is inextricably linked to sustainability: by opening up built heritage and using it appropriately, its upkeep is best secured and its protection from decline guaranteed (Adriaenssens et al., 1998; Gobyn and Knops, 2000). Integrated conservation therefore strives to give built heritage a contemporary role in society. At the same time, inclusion policy strives for universal participation in society, which requires that environments can be reached, entered, interpreted and used by people with diverse and changing abilities. When built heritage plays a contemporary role in society, and different people participate in society, both meet. Making built heritage inclusive - i.e. reachable, accessible, understandable and usable for as many people as possible - is a highly complex matter, however. Proposals to make historic buildings more inclusive tend to raise objections from conservation authorities, which guard the historic values of built heritage. Current approaches to accessibility do not seem to deal with these concerns well. Particularly telling in this respect is the accessibility legislation for public buildings recently issued in the region of Flanders: monuments that are provisionally or definitely protected, or building sites located in (provisional or definite) conservation areas, are exempt from its application. In this relatively small region, this comes down to no less than 10,000 buildings listed as protected. Built heritage thus remains out of reach, both practically - from the perspective of disabled people - and legally - in terms of building regulation. Is making built heritage (more) inclusive really beyond reach, the proverbial exception to the rule? Or can we address it from a different angle? This paper reports a field experiment that addresses the inclusivity of built heritage in a different way, by allowing a group of people to become concerned with this issue. The context of the field experiment is a European university, whose campus features a considerable number of protected buildings. After introducing the origins and set-up of the

P. Langdon et al. (eds.), Designing Inclusive Systems, DOI: 10.1007/978-1-4471-2867-0_21, © Springer-Verlag London 2012

207

208

Heylighen

experiment, we report what insights it has yielded so far, how these are received by different people and how they impact the real-world situation on campus. We draw on field notes and pictures taken during participation in building visits, analysis reports, and written and oral accounts by different people involved in or affected by the experiment. After comparing the approach explored with more traditional ways of addressing the inclusivity of built heritage, we conclude by questioning its transferability to other real-world contexts.

21.2 Context A university is a very specific and in a sense ‘unique’ institution (Biesta and Simons, 2009) in that it is multiple (id)entities at the same time (Heylighen and Nijs, 2011): an institution of higher education where people study; a research institution where people conduct research; a built environment featuring campuses, buildings and rooms that accommodate students and staff; and an organisation with several departments, including technical services that take care of the accommodation. These multiple identities make a university an ideal setting for a field experiment. A university is first and foremost an educational setting where people study. Since young people are educated as much by example as through teaching, environments that segregate teach acceptance of segregation, whereas inclusive environments teach inclusion (Welch and Jones, 2001). To some extent, a university is also a miniature version of society. Several agents involved in or affected by making built heritage more inclusive are present in its organisation: the building owners, architects and other built environment professionals (e.g. building conservation specialists), services for construction, management and maintenance, and building users, i.e. students, staff, and visitors, both young and moving into old age, with and without disabilities. In line with the exigencies of a ‘real’ experiment, we can thus say that the university offers an ‘ecologically valid’ setting. Of all those active at a university, students, staff and visitors living with diverse abilities and conditions can be considered as “user/experts”, a term introduced by Elaine Ostroff (1997) to denote “anyone who has developed natural experience in dealing with the challenges of our built environment”. Their experiences may offer designers unique and expanded insights - see for instance Pullin (2009) and Helen Hamlyn Centre (2009), yet in relation to built heritage their voices often remain silent. Built heritage is typically approached in an essentialist way that focuses on the built environment in itself. Giving voice to disabled building users in improving its inclusivity, however, shifts the focus from built heritage to how people experience it. Such a relational approach resonates with social conceptions of disability which, unlike prevailing medical conceptions, place the body in its sociomaterial context, recognising the interplay between physiological condition and features of the society one lives in (Butler and Bowlby, 1997). In relation to built heritage, this move to embrace disability as a social issue can be traced in the strategic framework for access to historic and heritage buildings developed by English Heritage (Adams and Foster, 2004). It is this framework which inspired us to conduct a field experiment at

Inclusive Built Heritage as a Matter of Concern

209

our university to involve disabled students, staff and visitors in addressing the inclusivity of built heritage on campus.

21.3 Origins and Set-up The idea of conducting the field experiment arises at a point when the university, in its role as building owner, decides to obtain expert advice from an official accessibility office on the inclusivity of its built patrimony. By way of trial one protected university building is subjected to an accessibility audit. The audit is performed by a professional accessibility advisor - a specialised architect - who assesses the building using a standardised checklist, and formulates a proposal to address the problems identified in a phased way. Rationale. The approach adopted in the field experiment is intended not as an alternative for, but rather as complementary to professional approaches like the accessibility audit. Instead of imposing certain solutions upon architects, it aims to inform them or at most make suggestions to them, offering a basis for making design decisions while leaving the actual design up to them. Buildings. The buildings considered in the field experiment are protected buildings on campus. They are selected in consultation with architects of the university’s technical services. Preference goes to buildings for which works are planned in the near future. So far three buildings have been addressed: the Van Dalecollege, a 16th century college accommodating the university’s student services and student housing; the Arenbergcastle, a 16th century building housing the architecture department; and the Pauscollege, a late 18th century college used as a dorm for 180 students plus a branch of the university restaurant. In an early version, the approach explored in the experiment was also applied to the Grote Aula, a 19th century auditorium used for lectures and music events. Teams. Each building is analysed by five teams. Every team is composed of one user/expert and two Master students in architecture (or, in some cases, one student and one researcher). User/experts include students, staff and visitors with a mobility impairment (using a wheelchair, having difficulty walking), a sensory impairment (blindness, low vision), of a diagnosis on the autistic spectrum. The architecture students attend an elective course on inclusive design. Teams visit the building being considered and identify its qualities and weaknesses from the perspective of the user/expert in the team. Based on an earlier experiment with professional architects, not related to built heritage (Heylighen et al., 2009), we expected that during these visits a particular dialogue would develop between the user/expert on the one hand and the architecture students on the other hand: a dialogue that is embodied in nature, unfolds in situ, and involves a particular knowledge transfer (Heylighen and Nijs, 2011). Through such a dialogue, experience is being framed: both the user/expert and the (student) architect find themselves in a considering stance - considering their experience of the building for the former, considering design practice for the latter. Output. The architecture students write an analysis report summarising the major insights gained during the visit of their team. The report is not normative in that it

210

Heylighen

informs us about how the user/expert in the team experiences the building visited, rather than prescribing what should be altered. It is narrative in that it addresses the building’s spatial qualities and obstacles in a way that respects the intricate relatedness of things in how the user/expert experiences it, rather than point-by-point (as in say a standardised checklist). The report is documented with photos and graphic material that resonate with architects’ visual way of working. Reports are shared and discussed with the other teams analysing the same building, thus augmenting their validity, and with architects and other built environment professionals of the technical services.

21.4 Framing Experience, Nuancing Inclusivity The analysis reports offer a highly nuanced picture of the inclusivity of the university buildings visited, revealing issues that may easily be overlooked in making built heritage more inclusive or that built environment professionals may not be attuned to. Moreover, besides unforeseen issues to address, they also point to unforeseen opportunities for improvement.

21.4.1 Different Needs or Different Reasons The teams’ analysis reports of the building visits show how different the needs put forward by different people - or even by one person - can be. A user/expert who has difficulty walking, sometimes uses a wheelchair, but visits two university buildings (Arenbergcastle and Pauscollege) on foot. During these visits, he sometimes points at aspects that do not raise a problem at this point, but would if he were using his wheelchair; or vice versa. For example, slopes are very handy when in his wheelchair, but on foot he prefers a well dimensioned staircase to a slope, because on the latter he has more difficulty keeping his balance. Inclusivity reveals itself here not as a timeless and invariant feature (as it does in traditional approaches), but instead shows its ambiguity and situational character. However, in addition to the differences in (and at times contradictions between) the needs pointed out by the user/experts, the analysis reports also reveal building aspects that are experienced as problematic by several of them. For example, the Van Dalecollege, Arenbergcastle and Pauscollege each have a courtyard covered with cobblestones. Their unevenness causes problems for several user/experts: for the person having difficulty walking it makes using a cane more difficult and increases the risk of stumbling; for the wheelchair users, it provides a bumpy ride; and for the blind participants it makes walking with a white cane difficult - cobblestones lying in the same direction cannot be felt as a guiding line. Similarly, several user/experts complain about the (lack of) light in the Pauscollege. The architecture students collaborating with a person with low vision notice that the transition from dark to light(er) spaces - and the other way around constitutes a considerable threshold for her. As her eyes need to adjust, she walks less swiftly and less spontaneously through the building. The recognisability of building

Inclusive Built Heritage as a Matter of Concern

211

elements diminishes considerably in dark spaces. For the architecture students it does not make a difference in which corridor they are walking, yet they feel that the user/expert walks more cautiously through the darker corridors. A person with autism also mentions the lack of (sufficient) natural light in the Pauscollege, which he considers especially problematic in the long windowless corridors along the student rooms. Deprived of contact with outside, he does not know on which side of the building he finds himself. It is interesting also that the only corridor he finds beautiful does have sufficient light. This enables him to see better how the space is finished. More generally, several user/experts characterise the Pauscollege as unpleasant. The blind person finds the building “not cosy at all.” It is “way too big” and there is “not much order”. The user/expert with autism has the impression that “lumber is lying everywhere” which he finds disturbing. After the visit, he is happy to be outside again because he dislikes the interior of the building and has an oppressive feeling inside. Asked what he finds unpleasant, he refers to the “prison corridor”. This specific corridor is more spacious than the other ones, but because the “prison feeling” prevails, he finds it particularly unpleasant. In these examples some aspects of buildings are pointed out as problematic by several user/experts, be it for different reasons. By not merely identifying problems, but explicitly describing the different reasons cited for them by the user/experts, the analysis reports can offer architects a better basis for designing more inclusive solutions. Standard or conventional formulations of problems tend to trigger standard or conventional solutions. By contrast, the descriptions of the reasons for these problems in different wordings (e.g. “cosy”, “lumber lying everywhere”, a vague “too big”) or metaphors (“prison corridor”, “prison feeling”) leave designers more degrees of freedom through their semantic openness. Still in the Pauscollege, another space the user/expert with autism finds beautiful is the hall with the old staircase. The hall is light and spacious, and the rustic wood offers a beautiful contrast with the white painted walls. More generally, old staircases in the different buildings are clearly appreciated by several participants. In the Arenbergcastle, for instance, a staircase in the porter’s lodge is praised in several analysis reports for its comfortable dimensions and its handrail. The handrail does not only offer a good grip, its banisters make the staircase clearly recognisable as such to a blind user/expert. Also in the Pauscollege one particular staircase is described as very comfortable and its handrail as offering good grip. These examples demonstrate that the analysis reports include building aspects that are valued by several user/experts, and not only those that are criticised by them. This enables architects designing inclusive solutions to build on strength instead of focusing on faults and weaknesses only. In summary, then, the approach adopted in the field experiment shows that needs may differ considerably depending on the person or situation, thereby unmasking inclusivity of built heritage as ambiguous and situational. For some aspects of buildings, however, several user/experts agree that they are either problematic or valuable. This occasional convergence between different user/experts might suggest a certain ‘universality’ of the problem or solution considered, yet universality or generalisation is not an aim in itself here. More interesting is the fact that architects are offered insight in why user/experts either dislike or value certain building aspects.

212

Heylighen

21.4.2 Use and Organisation When visiting the Van Dalecollege, two visually impaired user/experts point at the lack of clear organisation (Heylighen et al., 2010). A person with low vision has the impression that the building complex is not designed as a whole. He finds that having to search for a room is not very user-friendly, and increases the importance of inclusive signage. A blind person, for her part, finds the building inconveniently arranged. For her to use it independently would require a clear explanation of its appearance, location, orientation and structure. Nevertheless, she is able to find the reception by herself because it is near the entrance, where she would look intuitively (and because of its smell, reminding her of a library or journals). The arrangement of the ground level seems relatively convenient to her, yet overall, she characterises the building as a true labyrinth requiring supreme concentration to navigate. A person with autism also has trouble with the lack of clear organisation. Except in places he is familiar with (because he has been there before), it is difficult for him to locate at which point in the building he finds himself. When entering the main entrance of the Arenbergcastle, a blind user/expert is relatively quickly on to the fact that the building is structured around a courtyard. He derives this from what he hears. He describes the covered entrance as a passageway in between two buildings (he notices an echo), which is followed by an open space. The rectangular shape of the courtyard makes it easy to orientate himself. The secretariat of the department housed by the building, however, is located in an illogical spot: while he would expect it close to the main entrance, it is located in a side wing. A user/expert having difficulty walking suggests changes to the castle’s organisation as well. The entrance to the porter’s lodge would be much more accessible to him when using the back door instead of the door giving on to the courtyard. The same applies for the seminar rooms. By recognising the entrance via the current secretariat as a main entrance to the seminar rooms, people are not obliged to cross the bumpy cobblestones in the courtyard. For a user/expert with autism, the experience of a space seems to be influenced considerably by its use. The entrance hall in the Pauscollege, for instance, has large windows which let in a lot of light. Still he does not find it a pleasant place to wait, as the noises of the drinks machine and of the people passing are too disturbing. Similarly, the big spaces in the Arenbergcastle used by architecture students as design studios probably would not be very suitable for him to work. These spaces may be very busy, with students and staff running in and out. When entering the room you are directly confronted by the people present. For him, he says, the design studios perhaps would be better subdivided in smaller, structured spaces that are more or less separated from each other in terms of view and sound. These examples illustrate how the approach adopted in the field experiment explores and evaluates built heritage as a physical entity, but also considers how it works. The problems experienced by the user/experts turn out to be caused not only by material barriers raised by the historic building itself; major problems - and thus also possible solutions - relate to how the building is used and how this use is organised. This attention to use and variation in use resonates with architects’ core business: the organisation of space rather than the physical building as such.

Inclusive Built Heritage as a Matter of Concern

213

21.4.3 New Weaknesses, Old Strengths In the Arenbergcastle, recently introduced building elements turn out to cause major problems to several user/experts. A case in point are the concrete platforms that were installed a few years ago in front of both entrances to the west wing, and which are experienced as highly problematic by a blind person, a person with low vision and one who has difficulty walking. They find the platforms very dangerous because of the lack of handrails, contrast and marks. The door handle of the new outside door is not easy to find, it is not recognisable as such and is difficult to grasp. Inside the castle, the staircases causing most problems are the most recent ones: the spiral staircase close to the secretariat, and the escalator leading to the staff room above the secretariat. The spiral staircase is experienced as very unsafe because the steps are irregular, a handrail is provided on the narrow side of the steps only, it is unstable and interrupted at two points. Interestingly, user/experts find the old staircases - near the seminar room and in the porter’s lodge - much more comfortable. Conversely, historic elements do not always turn out to be problematic; on the contrary. The different shapes of door handles throughout the castle offer a very pleasant surprise to a user/expert with low vision. Moreover, they mostly contrast very well with the door, which makes them clearly visible. She also appreciates the fact that within the castle and the porter’s lodge the relatively logical structure of spaces is preserved, so that searching for a room does not take long. The rooms are rectangular, which facilitates orientation. A user/expert with autism seems to regret that the original functions are no longer visible in the way the castle is currently used. The building would be more readable to him if some relation existed between old functions, e.g. “kitchens” and “salons”, and new ones. Although he realises that this is impossible as the building now has a completely different function, it would make the functional organisation better understandable for him by offering clues on how to orientate himself in the building. To him, overview and organisation are important in space, but also in time. Overview and organisation in time assist him in handling his environment and anticipating new situations. Together these examples indicate that recent interventions in historic buildings are not always an improvement in terms of inclusivity; conversely, historic elements are not always problematic. This suggests that the idea that inclusivity of built heritage is problematic as a result of its age, and that newer interventions are better, is undeserved. In other words, by shifting the focus from the historic building to how it works for people with diverse abilities and conditions, it becomes clear that built heritage may offer inclusive designers not only a major challenge, but also a source of inspiration.

21.5 Discussion Feedback from built environment professionals of the technical services suggests that they value the analysis reports considerably. As one architect formulates it: “I found the subjective analyses highly interesting and in many respects they actually taught

214

Heylighen

me more than the objective analysis of the accessibility office. I find the added value thus very high.” Compared to the accessibility audit conducted by the professional accessibility advisor, the architect especially seems to appreciate the nuanced approach adopted in the field experiment: “An important aspect is the broadening of the term accessibility by including very diverse disabilities, also and above all those whereby the person is not “entirely” blind or chained to the wheelchair. The experience of the person with autism surprised me in the most positive sense: he uncovered in a very direct way problems (…) which we all do sense but never can point to that well.” Compared to the professional accessibility audit, the field experiment approaches inclusivity of built heritage on campus from a completely different angle. To some extent, inclusivity of the built environment can be - and often is - considered as what Bruno Latour (2005) refers to as a matter of fact. Similar to AIDS, poverty, global warming and equality, it is often something we are detached from, taken care of by state officials or experts, instead of something to which we, as a public, are exposed or attached (Simons and Masschelein, 2009). Accessibility legislation translates inclusivity into facts (or indicators and averages) by fixing maximum heights of thresholds and minimum widths of doors, which in turn can be objectively measured by professional accessibility advisors performing accessibility audits. Reducing inclusivity to the realm of matters of fact, in which accuracy becomes the closing argument of professional experts, leaves those affected by it - the disabled people themselves - seemingly incapable of joining the dialogue because they are supposedly no experts in the field (Heylighen and Nijs, 2011). As pointed out in the introduction, however, inclusivity of built heritage is often not covered by accessibility legislation, or is considered as an exception to the rule. The approach adopted in the field experiment acknowledges that there is hardly any regulation available to address inclusivity of built heritage (including several university buildings), and that the traditional specialisation and available expertise is inadequate to solve this problem. Therefore, the approach allows for a group of people to become concerned with or attached to this issue - architecture students, disabled students, staff and visitors, and staff of the university’s technical services. In other words, inclusivity of the built environment is not presented as a matter of fact. Rather, through analyses of university buildings in collaboration with user/experts, it is made perceptible in the public sphere and gradually becomes a matter of concern (Callon, 2005; Latour, 2005). Feedback from the architecture students suggests that they experience this alternative approach as highly motivating and insightful, but also as very unusual and therefore somewhat confusing. From other courses, they have become used to the fact that teachers have the necessary expertise to offer the (or at least a) solution, and they are surprised to discover that for this issue, this is not the case. In the real-world situation on campus, inclusivity of the built environment presents itself to the students as ambiguous and situational. They learn to be affected in new ways by the same issues (Latour, 2004; Despret, 2004). Through the particular dialogue with user/experts in situ, i.e. in the protected buildings under consideration, the attention of the architecture students is being trained. One user/expert attests to this learning process: “[the student] was open to it. I could clearly notice that by talking to her.” The user/expert enjoyed participating in

Inclusive Built Heritage as a Matter of Concern

215

the building visits, and would like the approach to become a compulsory topic of the students’ program, allowing more people to become attached to the issue: “As an elective course, you only have motivated students. That’s of course an advantage. But then it’s something that is possible but not compulsory. (…) The more people you address, the more obvious it probably becomes. Maybe it’s good to strive for a matter of course instead of a possibility.” Interesting to notice, however, is that even as part of an elective course, the approach has a major impact on local decision making. The insights gained through the visits meanwhile have motivated and enabled the technical services to implement major alterations in some of the buildings visited. As we write, the Grote Aula is undergoing major interventions to improve its acoustic comfort, which are directly motivated by insights gained through the analyses with user/experts. The outcome of these also played a crucial role in the negotiations with and convincing of the conservation authorities. For the Van Dalecollege, input from the user/experts unlocked the impasse the student services had ended up in (Heylighen et al. 2010). The lack of organisation pointed out by several user/experts inspired major organisational interventions to rearrange the student services more logically in the available space so that all students can consult them, and yet interventions which require altering the historic fabric remain limited (ibid.). Interestingly, these organisational interventions come down, to a large extent, to restoring the logic present in the original building, making use of its inherent qualities. By shifting the focus from the protected buildings themselves to how people experience them, it becomes clear that improving their inclusivity does not necessarily require supplementing the present situation with new (material) layers; leveraging concepts already present in historic layers may improve the value of the buildings from an inclusive perspective, while respecting their heritage value.

21.6 Conclusions Is making built heritage (more) inclusive the proverbial exception to the rule? Or can we address it from a different angle? Starting from the real-world situation on a university campus, this paper has demonstrated that it is possible to approach inclusivity of built heritage in a different way. Key to the approach adopted in the field experiment is that it allows for a group of people to become involved with or attached to this complex issue as a matter of concern, rather than considering it as a matter of fact. While the approach originally was not intended to be political, it turns out to have a considerable impact on local decision making, which in turn impacts the inclusivity of built heritage on campus. Concerns do matter - and are (cap)able of mattering - apparently! Therefore it is tempting to suggest transferring the approach to other real-world contexts. Yet we should keep in mind that the field experiment took place in the ‘unique’ context of a university. In order to investigate to what extent the approach is transferable to other real-world contexts, we seek to extend the field experiment to protected buildings off campus. In addition, it would be interesting to investigate how the skills developed by the architecture students involved in the experiment are received in their professional situation after graduation.

216

Heylighen

21.7 Acknowledgements This research has received funding from the European Research Council under the European Community’s Seventh Framework Programme (FP7/2007-2013)/ERC grant agreement No. 201673. The author thanks all who contributed to the field experiment, in particular the user/experts, architecture students, researchers, and staff of the technical services. Special thanks go to Stijn Baumers, Jasmien Herssens, Sam Michiels, Greg Nijs and Iris Van Steenwinkel for their invaluable comments on earlier versions of this paper.

21.8 References Adams J, Foster L (2004) Easy access to historic buildings. English Heritage, Swindon, UK Adriaenssens I, Hendryckx M, Van Alsenoy J (1998) Het geheim achter een open deur [The secret behind an open door]. Koning Boudewijnstichting, Brussels, Belgium Biesta G, Simons M (2009) Higher education and European citizenship as a matter of public concern. European Educational Research Journal, 8(2): 142-145 Butler R, Bowlby S (1997) Bodies and spaces: An exploration of disabled people’s experiences of public space. Environmental and planning D: Society and Space, 15(4): 411433 Callon M (2005) Disabled persons of all countries, unite! In: Latour B, Weibel P (eds.) Making things public: Atmospheres of democracy. MIT Press, Cambridge, MA, US Despret V (2004) The body we care for. Body and Society, 10(2-3): 111-134 Gobyn R, Knops G (2000) Erfgoedzorg in de 21ste eeuw [Built heritage care in the 21st Century]. Koning Boudewijnstichting, Ghent, Belgium Helen Hamlyn Centre (2009) DBA inclusive design challenge. Helen Hamlyn Centre, London, UK Heylighen A, Herssens J, Froyen H (2009). Architecture criticism blindfolded. In: Proceedings of International Conference on Inclusive Design (INCLUDE 2009), Helen Hamlyn Centre, London, UK Heylighen A, Neyt E, Baumers S, Herssens J, Vermeersch P-W (2010) Conservation meets inclusion. Model meets reality. In: Proceedings of the 5th Cambridge Workshop on Universal Access and Assistive Technology (CWUAAT), Cambridge, UK Heylighen A, Nijs G (2011) Studying (architecture) in dialogue with disability. In: Simons M, Decuypere M, Vlieghe J, Masschelein J (eds.) Curating the European University. Leuven University Press, Leuven, Belgium Latour B (2004) How to talk about the body? Body and Society, 10(2-3): 205-229 Latour B (2005) From realpolitik to dingpolitik or how to make things public. In: Latour B, Weibel P (eds.) Making things public: Atmospheres of democracy. MIT Press, Cambridge, MA, US Ostroff E (1997) Mining our natural resources: The user as expert. Innovation, Industrial Designers Society of America (IDSA), 16(1): 33 Pullin G (2009) When design meets disability. MIT Press, Cambridge, MA, US Simons M, Masschelein J (2009). The public and its university: beyond learning for civic employability? European Educational Research Journal, 8(2): 204-217 Welch P, Jones S (2001) Advances in universal design education in the United States. In: Preiser WFE, Ostroff E (eds.) Universal Design Handbook. McGraw Hill Professional, NY, US

Chapter 22 Designing a Virtual Environment Framework for Improving Guidance for the Visually Impaired S. Kammoun, M.J-M. Macé, B. Oriola and C. Jouffrais

22.1 Introduction Electronic Orientation Aids are dedicated to orientation assistance for the visually impaired. They are made of at least 3 essential components: 1) A positioning system (e.g. GPS); 2) A Geographical Information System (GIS) that includes both a digitised map and a software designed to select routes, track the traveller’s path, and provide him with navigation information; 3) A User Interface (UI) that relies on nonvisual (usually auditory or tactile) interaction. These three components could all be the cause of usability issues. The first major issue is error in GPS positioning that is frequently greater than 20 metres (especially in cities), which is really not compatible with VI pedestrian guidance. Secondly, GIS usually contain exclusively road networks, and hence lack pedestrian-related information. Finally, the interaction with an EOA is a key element, and it must be designed from the beginning for visually impaired users on the move. Virtual interactive environments may represent a valuable platform to selectively isolate GIS or UI-related issues for being safely and systematically tested in laboratory before on-site evaluations. With the rise of the number of EOAs, research groups were interested in virtual environments (VE) to assist visually impaired people in learning Orientation and Mobility skills (Sanchez and Tadres, 2010). Those systems have been developed to help construct a mental representation of space from tactile or auditory cues, to increase spatial cognitive abilities (Mereu and Kazman, 1996), and to give visually impaired people a tool for safely exploring and learning about new spaces on their own (Schloerb et al., 2010). Generally, these systems are designed to allow VI users to explore virtual representations of real or abstract (e.g. a labyrinth) spaces, as well as to interact with objects within these spaces (Sanchez and Hassler, 2006). However

P. Langdon et al. (eds.), Designing Inclusive Systems, DOI: 10.1007/978-1-4471-2867-0_22, © Springer-Verlag London 2012

217

218

Kammoun et al.

they never rely on defective positioning, nor on GIS adapted to visually impaired pedestrians. With the purpose of improving autonomy in the mobility of the visually impaired, this study proposes the exploration of a virtual environment (VE), through auditory and haptic interaction. This platform allows the user to navigate in a virtual environment representing an existing space (based on the GIS of a city for instance). The aim of this platform is to systematically test several guidance processes before implementation in an EOA. These tests will help to determine which guidance process is the most efficient to compensate for inaccurate GPS positioning, while improving mobility and orientation at the same time. In the following section, we describe the guidance process generally used in EOAs. In section 22.3 we present the proposed virtual environment framework. We focus on the GIS component, user interfaces and mobility within the VE, and technical issues during design and implementation steps. Finally we discuss how this system will help to benchmark different guidance processes used in EOAs, and evaluate resultant navigation performance and cognitive mapping.

22.2 Guidance in Electronic Orientation Aids When compared to car navigation, it is obvious that pedestrian navigation, especially with visually impaired travellers, imposes additional requirements upon the electronic aids. Studies on human navigation (see e.g. Loomis et al., 2001) report that there are two distinct methods for keeping track of position and orientation during travel. The first mode is called landmark-based navigation. In this mode, visual landmarks provide the traveller with direct feedback regarding current position and orientation. When considering navigation of visually impaired pedestrians, visual landmarks are inoperative, but are replaced by auditory, somatosensory or olfactory landmarks. In the second mode, called path integration, the traveller uses the sense of body motion (kinaesthetic feedback) to update his current position and orientation relative to the starting point. This mode is operative in visually impaired pedestrian navigation, but estimated position rapidly drifts if landmarks for position correction are too sparse. Then an efficient EOA should provide the visually impaired traveller with mobility instructions, but also with frequent and usable landmarks that allow both landmark-based navigation and path integration. In EOAs, the guidance process consists in first identifying the location of a visually impaired user relative to the expected trajectory and then providing her/him with the appropriate direction instructions, or with pertinent information about the surroundings. Hence we can define guidance process according to three main steps: 1) Route selection procedure for computing the optimal itinerary; 2) User tracking for estimating his current position; 3) Display of navigation instructions and spatial descriptions to orientate the traveller and improve his mental representation of the environment (Kammoun et al., 2011). In order to systematically and efficiently test different guidance processes, these three modules should be operating at once, which is very challenging with visually impaired users moving in a real environment. To palliate this difficulty, we designed a controlled virtual environment for testing different guidance processes potentially used in EOAs.

Designing a Virtual Environment Framework

219

22.3 A Platform for Improving Visually Impaired Guidance The system described here is a research tool aimed at designing and evaluating improved guidance for the visually impaired. Our objective is to use this platform to evaluate different guidance processes to be implemented in an EOA. In a second step, the platform will also serve to study the enhancement of cognitive mapping of the visually impaired during guidance. Improved guidance is based on the presence of numerous geolocated pedestrian-related data (pedestrian paths, non-visual landmarks and points of interest) that are annotated in the GIS database and displayed during navigation. In the following sub-sections, we focus on the technical development of the platform.

22.3.1 Adapting GIS Components to Visually Impaired Needs Most spatial databases used in GIS have been developed without considering the needs of the Visually Impaired. Very few studies have aimed to identify their needs and proposed an annotation of geographical objects that should be included in specialised GIS databases. An interesting classification has been proposed (Golledge et al., 1998). They divided pedestrian features into four classes: (1) transportation (e.g. roads, bike paths, walkways, car parking areas, bike parking), (2) buildings, (3) land use (e.g. open space, recreation, vegetation), and (4) other objects (e.g. light poles, telephones, and stairs). This classification is interesting but still misses important data to select the best adapted route for the VI. In our research group, we have adopted a long-term user-centered design approach in collaboration with the Institute for the Young Blind (CESDV-IJA, Toulouse) (see Brock et al., 2010). For this specific project, we interviewed 19 users to define more precisely their needs as well as their degree of autonomy and technological knowledge. The target population comprised 7 females and 12 males with a mean age of 37. For daily mobility, 5 of them use a guide dog and 10 use the white cane. The 4 remaining prefer to have a person to guide them. All of them are legally blind and expressed their motivation and agreement to participate in this project. We had three meetings with four different O&M instructors from the CESDV-IJA. They precisely described the different steps and techniques that they teach to blind persons during O&M training. We also analysed through video and a posteriori interviews the O&M behaviour of two blind users (one with a white cane and one with a dog). We finally performed three brainstorming sessions with at least 4 blind users in which we investigated issues with VI navigation. Taking into consideration these different data, we finally proposed an annotation of geographical data including three main classes: (1) Walking areas that compose the pedestrian network (sidewalks, pedestrian crossing, etc.); (2) Non-visual landmarks corresponding to locations that can be detected by a VI pedestrian on his/her own (e.g. change in the pavement texture or inclination, street furniture, odours or sounds, etc.). These landmarks are either decision points (e.g. when to turn) or confirmation points (e.g. a good choice was operated) points;

220

Kammoun et al.

(3) Points of Interest (POIs) such as places or objects that are potential destinations (e.g. public building, shop, street furniture like mailbox or bus stop, etc.). When POIs are not a destination per se but near the itinerary, they may be useful to figure out how the different elements in a city are spatially organised. In addition, both landmarks and POIs may subserve landmark-based navigation as well as path integration. We used data from Open Street Map (OSM) to construct the GIS database. OSM is an open source project used by a large and dynamic community. The main advantage is that OSM is an open resource, and it is easy to add data and features to the database via a simple editor (JOSM). In addition, the data can be shared online. Hence it is easy to annotate the OSM database according to the proposed classification by editing nodes, ways, metadata tags and relations.

22.3.2 Avatar Mobility We chose to rely on a force feedback joystick to control the exploration, because it is a convenient device to manipulate several dimensions at once (heading, translations and rotations), and its sensitivity and gain can be adjusted to fit real human walking speed. Forward/backward movements of the joystick allowed displacement respectively forward and backward in the VE. Left/right joystick movements controlled body rotation angle. In order to minimise the complexity of input interaction, translation and combination of two movements were not allowed. As the user moved the joystick, the system tracked the direction of the movement and continuously updated the position of a corresponding avatar in the VE while at the same time providing auditory feedback in relation to the displacement (step sounds).

22.3.3 Output User Interface In order to design the output interface, we organised two separate brainstorming sessions with expert and novice users of EOAs. Our goal was to define the type and quantity of information required during guided travel, as well as appropriate modalities that do not interfere with the learned O&M techniques and abilities. These two sessions made it clear that during a navigation task, an EOA should provide two classes of information: (1) Direction instructions, i.e. turn-by-turn instructions, and (2) Space-related information (landmarks for navigation, but also information about the surroundings, description of difficult points, etc.). To display both of them, an adapted interface is required, which must rely on non-visual (e.g. auditory or somatosensory) modalities. Text To Speech (TTS) as well as binaural synthesis - which provides virtual 3D sounds at any desired location in the listener’s space - have been evaluated in real navigation context (Loomis et al., 1998; Gaunet, 2006) and within a virtual environment for orientation and mobility training (Sanchez et al., 2009; Schloerb et al., 2010). In real navigation tasks, the somatosensory modality is another efficient output modality in different situations, e.g. when natural sounds are critical and should not be masked, when ambient noise level is too high, or when user impairment prevents

Designing a Virtual Environment Framework

221

auditory-based interaction (visual impairment is sometimes associated with auditory impairment). Different studies have shown the usability of tactile displays for presenting directions by mapping them onto body locations. Information regarding the expected direction was provided through the activation of a given vibrator mounted in a belt (Pielot et al., 2008), or within a backpack (Ross and Blasch, 2000). Haptic feedback has also been used within virtual environments designed for supporting orientation and mobility training (Schloerb et al., 2010). In the platform, we implemented a feedback editor that allows the addition of single or combined feedback on objects (TTS and/or 2D/3D sound on the audio channel; vibration and/or force on the joystick). We first implemented a footstep feedback that was related to walking speed. Additional feedbacks allowed the display of both direction instructions, obstacles encountered, and spatial configuration information (landmarks, POIs). We suggest that a combination of TTS and binaural synthesis is an interesting solution for guidance as the two kinds of information are easily distinguishable. Spatialised audio was mainly used for directional information. Indeed it allowed the user to hear the direction (Right/Left) of sounds in the VE as if he were standing at the location of the avatar. Spatialised TTS was also used to render the location of environmental features (landmarks and POIs). As binaural estimation of distance is rather limited in humans (Middlebrooks and Green, 1991), TTS or spatialised TTS seemed advantageous to render this cue. We also designed force-feedback effects that we can attach to different textures or obstacles in the VE. Hence the user felt the variations in ground texture and the presence of steps (e.g. sidewalk) or obstacles (e.g. wall) as in a real navigation task using the white cane. These somatosensory cues may also serve as landmarks.

22.3.4 Virtual Environment Implementation As mentioned above, we used the OSM database as the original source of geographical information in the GIS database. In order to extract information from the OSM database, we generated an XML file from a selected area. This XML file was then parsed to get the geometry and attributes of roads, walking areas, buildings, landmarks and POIs. We finally converted the Word Geodetic System coordinates (WGS84) used in OSM database to a Cartesian coordinate system in order to build the virtual environment. Three important classes of objects were included: 1) Polygons that represent the different buildings, walls and car park areas. Each building was defined by its name or function, extracted from the OSM data. 2) Lines that represent different types of roads, walking areas, sidewalks, zebra crossings. Each line was defined by its name (for the streets), or its direction compared to North for differentiating road sidewalks. 3) Points that represent Landmarks and POIs. The platform presented two distinct modes: firstly a Control mode that is used by VE developers, researchers, and Orientation and Mobility instructors. The Control mode allows developers to create and modify VEs. A key feature of the Control mode is the program’s ability to import an XML file from Open Street Map to create a new 3D virtual map and to manually or automatically (via the

Kammoun et al.

222

route selection algorithm from Kammoun et al., 2010) select a path between two points. This makes it possible to easily import maps of different campuses or cities. The control mode includes the feedback editor (tactile and auditory feedbacks). The second Evaluation mode allows researchers and Orientation and Mobility instructors to record and review the user behaviour in the VE. During an experimental session, the system recorded in a text file the avatar’s (user) position, orientation and speed within the VE, along with any interaction between the user and the system. For the experimenter display, different textures were applied according to the type of surface (building, walls, etc.) or the type of line (e.g. tar texture is chosen for roads, and zebra texture for pedestrian crossing). Figure 22.1a shows an example of the University of Toulouse environment including the global map in the lower right hand corner. Figure 22.1b shows a close-up view of the global map with the actual position of the user (red spot) in the virtual space. The platform was implemented in C++ code running under Windows 7. It uses an open-source graphic rendering engine (OGRE 3D), and a cross-platform 3D audio API (OpenAL) appropriate for displaying 3D spatialised sounds via the headphones. A collision detection algorithm based on ray tracing was integrated into the framework so as to allow real time collision detection in complex environments. a)

b)

Figure 22.1. a) A screenshot from the graphic interface of the virtual environment. The global map is on the lower right hand corner. One can observe different buildings, walking areas and car park areas. b) Close-up view of the global map, the red point represents the position of the traveller in the virtual space. Each type of walking area (sidewalk, pedestrian crossing, car park area, and road) is represented by a distinct colour.

22.4 Navigation and Guidance Process Before beginning guidance, whether in real or virtual conditions, route selection is necessary and is usually included in the GIS component. It is defined as the procedure of choosing an optimal pathway between origin and destination. When considering pedestrian navigation, the shortest path might be appropriate but should rely on a GIS database including essential information for pedestrian mobility (e.g. sidewalks and pedestrian crossings). In addition, brainstorming sessions revealed that excluding difficult points (e.g. complicated crossroads) as well as including non-visual landmarks on the itinerary is important when route

223

Designing a Virtual Environment Framework

selection is made for the Visually Impaired. Using the database described in chapter 22.3.1 we were able to select an adapted path for a VI pedestrian by using the algorithm indicated in (Kammoun et al., 2010). When a route was selected, several sections defined by two successive Itinerary Points (IPs) were generated. Each section contained a list of Landmarks and POIs extracted from the GIS. As mentioned in the previous section, virtual 3D sounds, TTS and Spatialised TTS were used to display instructions and environmental information (see Figure 22.2). Based on simple interaction properties attached to the different scene elements it will be possible to design and evaluate many guidance processes. A first instance of direction instruction process was designed by displaying a spatialised sound on the next IP position according to the traveller position and orientation. a)

b)

Figure 22.2. a) The green point represents an intermediary itinerary point (IP) (a virtual spatialised sound is placed at this location). When the point is reached, feedback is given to the subject and the next IP is activated. This guidance process allowed the user to reach his/her destination by small, easily reachable steps. b) Blue points represent landmarks or POIs selected on this path, and extracted from the GIS.

Figure 22.3. a) Representation of a recorded journey around two virtual buildings in the University campus during an evaluation session with a VI user. Green, blue and red points represent, respectively, IPs, landmarks, and POIs that were displayed. The dark line represents the trajectory of the avatar. b) During the evaluation session, navigation speed and orientation of the avatar were also recorded; each arrow represents speed and orientation during navigation within the VE.

Landmarks and POIs in the vicinity of the path, which provide the user with spatial indications about his surroundings were displayed. In this case, a circular activation field (with an adjustable diameter) was attached to each type of element

224

Kammoun et al.

to trigger an information display. Figure 22.3 shows the representation of recorded displacements during two preliminary evaluation sessions.

22.5 Discussion and Future Work This paper presents the design and implementation of a virtual environment framework in order to evaluate tracking and guidance strategies before their implementation in an EOA. Based on a geographical database including specific and essential - information for pedestrian mobility as well as important Landmarks and POIs for the VI users, we built a virtual environment in which several user tracking and guidance strategies will be evaluated safely and systematically in the laboratory. The most promising strategies will then be tested in the real world with a prototype of an EOA sharing a large set of components with the VE (GIS database; route selection, tracking and guidance software; TTS and binaural synthesis output, see Katz et al., 2010).

22.5.1 User Tracking and Guidance Process Tracking user location is essential for an efficient guidance process in visually impaired navigation. For EOAs, geolocation based on GPS is the most common technique. However, positioning precision with GPS alone is rarely better than 20 metres in many environments. This is particularly the case in cities where urban canyons between buildings prevent direct line of sight with the satellites. Several approaches have been developed to improve the GPS precision in such environments. Dead-reckoning algorithms combine GPS with inertial sensors to improve estimated user positioning. Differential GPS (see e.g. Loomis et al., 1994) can provide an accurate positioning, and hence better guidance, but rely on a network of stations that is not available everywhere. Electronic location identifiers, such as RFID tags, WLAN networks or Bluetooth beacons have also been used in indoor environments such as museums, hospitals, and so on. However, they entail a large scale infrastructure with a specific deployment phase and important maintenance costs, which is very rarely available in cities. In this context, the use of a platform simulating real environment and positioning (with added inaccuracy) seems a promising approach when designing an EOA. Indeed, in this experimental platform, it is very easy to add in systematic and random noise to the positioning of the avatar. The resulting positioning will mimic what is observed with real GPS in different environments. The platform will then be used for testing the robustness and usability of the different tracking and guidance processes implemented.

22.5.2 Sound Localisation and Haptic Feedbacks In the platform that we describe, guidance instructions and environmental description are displayed using a combination of TTS, binaural synthesis and

Designing a Virtual Environment Framework

225

haptic feedback. There is a well-known issue related to the use of binaural sounds: the spatialisation must be defined in the head-centred reference frame of the user, not in that of the external world. However (Katz and Picinali, 2011) showed that head-tracking was not necessary for 3D sound-based navigation in VE; 3D sound synthesis based on the movements of the joystick (and hence on the displacement of the avatar) was sufficient for users to succeed in non-visual navigation tasks. A rich haptic interaction is provided through the use of a force feedback joystick device. Many scene elements can generate vibration and force feedback when they are encountered. This haptic interaction was designed to reproduce the information that a visually impaired person gathers - including ground texture, step and obstacle detection - when moving with a white cane or a dog.

22.5.3 Spatial Cognition in the Visually Impaired Virtual environments appeared to be effective in the transfer of cognitive mapping to real environments, even for visually impaired users (see e.g. Lahav and Mioduser, 2003). Our aim was different in that the platform is designed for evaluating robustness of tracking and improved guidance for the visually impaired (including map adaptation, route selection, direction instruction and space description). Indeed, as we previously stated, this platform will ease the tests of route selection - comparing for instance the automatically selected route with the users’ preference - or robustness of tracking algorithms by adding systematic and/or random noise to the position of the avatar. But, importantly, we will also modify the quality and the quantity of direction instructions, landmarks and POIs mentioned during the journey. We will use different experimental designs to test the efficiency of the guidance itself (number of errors and time to succeed), but also the mental maps acquired during the journey. These maps will be compared with maps acquired during mobility in real environments, as well as with topographic maps (see Lahav and Mioduser, 2008, for indoor real vs. virtual exploration). Our aim is to finally select the strategies that enhance not only the user-centred representations used in turn-by-turn instructions but also the global spatial knowledge of the explored environment.

22.6 References Brock A, Vinot J-L, Oriola B, Kammoun S, Truillet P, Jouffrais C (2010) Méthodes et outils de conception participative avec des utilisateurs non-voyants. In: Proceedings of the 22nd Francophone Conference on Human-Computer Interaction, Luxembourg Gaunet F (2006) Verbal guidance rules for a localised wayfinding aid intended for blindpedestrians in urban areas. Universal Access in the Information Society, 4(4): 338-353 Golledge RG, Klatzky RL, Loomis JM, Speigle JM, Tietz J (1998) A geographical information system for a GPS based personal guidance system. International Journal of Geographical Information Science, 12(7): 727-749

226

Kammoun et al.

Kammoun S, Macé MJ-M, Oriola B, Jouffrais C (2011) Toward a better guidance in wearable electronic orientation aids. In: Proceedings of the 13th IFIP TC13 International Conference on Human-Computer Interaction, Lisbon, Portugal Kammoun S, Dramas F, Oriola B, Jouffrais C (2010) Route selection algorithm for blind pedestrian. In: Proceedings of the International Conference on Control, Automation and Systems (ICCAS 2010), Gyeonggi-do, Korea Katz BFG, Picinali L (2011) Spatial audio applied to research with the blind. In: Strumillo P (ed.) Advances in sound localization. Available at: http://www.intechopen.com/articles/ show/title/spatial-audio-applied-to-research-with-the-blind (Accessed 9 November 2011) Katz BFG, Truillet P, Thorpe S, Jouffrais C (2010) NAVIG: Navigation assisted by artificial vision and GNSS. In: Proceedings of the Workshop on Multimodal Location Based Techniques for Extreme Navigation (Pervasive 2010), Helsinki, Finland Lahav O, Mioduser D (2003) A blind person’s cognitive mapping of new spaces using a haptic virtual environment. Journal of Research in Special Educational Needs, 3(3): 172177 Lahav O, Mioduser D (2008) Haptic-feedback support for cognitive mapping of unknown spaces by people who are blind. International Journal of Human-Computer Studies 66(1): 23-35 Loomis JM, Golledge RG, Klatzky RL, Speigle J (1994) Personal guidance system for the visually impaired. In: Proceedings of the ACM Conference on Assistive Technologies (ASSETS ’94), Marina Del Rey, CA, US Loomis JM, Golledge RG, Klatzky RL (1998) Navigation system for the blind: Auditory display modes and guidance. Presence: Teleoperators and Virtual Environments, 7(2): 193-203. Loomis JM, Klatzky RL, Golledge RG (2001) Navigating without vision: Basic and applied research. Optometry and Vision Science, 78(5): 282-289 Mereu SW, Kazman R (1996) Audio enhanced 3D interfaces for visually impaired users. In: Proceedings of the ACM International Conference on Human Factors in Computing Systems (CHI 1996), Vancouver, Canada Middlebrooks JC, Green DM (1991) Sound localization by human listeners. Annual Review of Psychology, 42: 135-159 Pielot M, Henze N, Heuten W, Boll S (2008) Evaluation of continuous direction encoding with tactile belts. In: Proceedings of the International workshop on Haptic and Audio Interaction Design, Jyväskylä, Finland Ross DA, Blasch BB (2000) Wearable interfaces for orientation and wayfinding. In: Proceedings of the International conference on Computers and accessibility, Arlington, VA, US Sanchez J, Hassler T (2006) AudioMUD: A multi-user virtual environment for blind people. In: Proceedings of the International Workshop on Virtual Rehabilitation, Manhattan, NY, US Sánchez J, Tadres A (2010) Audio and haptic based virtual environments for orientation and mobility in people who are blind. In: Proceedings of the International conference on Computers and accessibility, Orlando, US Sanchez J, Tadres A, Pascual-Leone A, Merabet L (2009) Blind children navigation through gaming and associated brain plasticity. In: Proceedings of the International Conference on Virtual Rehabilitation, Haifa, Israel Schloerb DW, Lahav O, Desloge JG, Srinivasan MA (2010) BlindAid: Virtual environment system for self-reliant trip planning and orientation and mobility training. Haptics Symposium, Massachusetts, MA, US

Chapter 23 Spatial Clues for Orientation: Architectural Design Meets People with Dementia I. Van Steenwinkel, C Van Audenhove and A. Heylighen

23.1 Introduction The physical environment holds great potential to improve the well-being of people with dementia (Calkins et al., 2001; Van Audenhove et al., 2003; Sternberg, 2009; Van Audenhove et al., 2009). However, when designing environments for them, architects are faced with a lack of adequate design knowledge. On the hypothesis that the perspectives of people with dementia have the potential to expand architects’ design expertise (Zeisel, 2001), our research aims to gain a better understanding of their spatial experiences in order to achieve this end. This paper outlines how this overall objective has become more articulated into more specific research questions through preliminary research: • • • • •

by reviewing literature in the fields of anthropology, psychology, phenomenology, human geography, architectural theory, supplemented with a study of (auto)biographies of people with dementia; by visiting 22 residential and care environments for people with dementia in Flanders and one in the Netherlands; by talking to people with dementia, their family and professional care givers and other professionals; by conducting open interviews with five architects, experienced in designing residential and care environments for people with dementia, about the strategies they used in former design projects; by conducting voluntary work, both in a day care centre and in a residential care centre for people with dementia, one day a week during two months, to become familiar with their daily life.

Secondly this paper explains which research methods are chosen to address the proposed research questions, and why.

P. Langdon et al. (eds.), Designing Inclusive Systems, DOI: 10.1007/978-1-4471-2867-0_23, © Springer-Verlag London 2012

227

228

Van Steenwinkel et al.

23.2 Context Dementia is a syndrome of progressive memory impairment and loss of other cognitive functions (American Psychiatric Association, 2000). This causes a very particular way of experiencing and negotiating space (Godderis, 1992). Family, friends and professional care givers often face a lack of insights into the way people with dementia perceive the environment. It can be difficult, for example, to find out why someone unexpectedly - and seemingly without provocation - gets anxious, suspicious, or aggressive, how a person can get lost in a familiar environment (Friel McGowin, 1993), or why people with dementia sometimes just sit and stare for a long period of time seemingly detached from what happens around them (Boden, 1998). At the same time, we expect that there is much to learn from people with dementia. As the main thread of the research conducted by our team, we consider people with different abilities and/or conditions - in this case people with dementia - as experts in perception and use of the (built) environment. They can bring forward the very different ways in which people in a diverse and aging society experience and negotiate their living environment and, thus, they can expand and refine architects’ knowledge on how to design that environment. In case of designing a living environment for people with dementia, the client often differs from the future user, i.e. people with dementia. If concerned with this future user, the client tries to communicate the user’s needs and desires to the architect(s). Although clients often have several years of experience in caring for people with dementia and the (built) environments they live in, interviews with architects reveal that clients can articulate only few insights in the spatial needs of people with dementia. For example, clients tend to emphasise the importance of “normality” and “homeliness”, characteristics which architects find too vague to work with. On the other hand, here lies an important role for architects, since we expect that their spatial knowledge can afford new insights in the daily experiences of people with dementia. In practice, however, architects often do not have or take the time to explore this in depth. Following from the above, we set out to explore the spatial experiences of people with dementia from an architectural point of view. Moreover, we assume that from this exploration we may also learn about the spatial experiences of people in general. Indeed, people with dementia are persons in the first place, who undergo a relative slow dementing process. Their often unrestrained and spontaneous behaviour (Zeisel, 2001) may reveal how other people secretly or unconsciously experience a certain situation.

23.3 Research Focus 23.3.1 Orientation Each person with dementia experiences a particular situation in a unique way. There are several types of dementia and different, but not clearly discriminated,

Spatial Clues for Orientation: Architectural Design Meets People with Dementia

229

stages in the dementing process, which may succeed each other at different rates. The disease can strike people of different ages and all kinds of backgrounds (Godderis, 1992). Since we are faced with such a great diversity of people who often live together in a group housing facility, we do not intend to focus on a delimited target group, e.g. women with early-onset Alzheimer’s. Instead we focus on one important aspect which all people with dementia have in common, i.e. disorientation in time, space and identity, due to memory loss (Godderis, 1992). People with dementia, may find their sense of time becomes upset or even lost. Now and then they do not know which (time of the) day or year it is, as Christine Boden, a women with early-onset Alzheimer’s, witnessed: “I don’t seem to have space in my brain for that sense of ‘Thursday-ness’ (or whatever day it might happen to be), or ‘April-ness’ or ‘1981-ness’.” (Boden, 1998)

One of the possible consequences is that their (daily) routines may get mixed up. A person with dementia may, for example, want to go shopping in the middle of the night (Braam, 2005). Something similar may happen concerning orientation in space. A person with dementia may get lost in space and be baffled when, in the morning, “everything seems new” (Braam, 2005). Additionally, people with dementia show a particular way of knowing people, as Christine Bryden described: “You see, I did not know their name, whether they were married or not, whether they had children, if they had a job. I knew nothing about them, nothing in the ‘normal’ sense of how you know people and recognise them. The way I know people is in a spiritual and emotional way. There’s a knowing of who a person really is right at their core. But I have no idea who they are, in terms of who they are meant to be in your world, of cognition and action, and labels and achievement.” (Bryden, 2005)

The fact that people with dementia sometimes are disoriented in time as well in space and identity may not be a surprise, since - according to several authors time, space and identity are interrelated dimensions. Edward Hall (1969) touched upon the idea that the differentiation of time and space as two distinct dimensions is only an arbitrary one, since in lived experiences they are actually “inextricably bound up in each other” (Hall, 1969). Christian Norberg-Schulz (1971) also explained that “perception mediates a world which could also very well be described as ‘events in a four-dimensional space-time’.” Moreover, Hall (1969) points out that space, and thus time, is also related to identity: “Man’s sense of space is closely related to his sense of self, which is in an intimate transaction with his environment. Man can be viewed as having visual, kinesthetic, tactile, and thermal aspects of his self which may be either inhibited or encouraged to develop by his environment. “ (Hall, 1969)

Because of these interrelations we henceforth use the term ‘orientation in timespace-identity’. Thus, we do not consider orientation merely as a means for wayfinding, i.e. knowing how to go from one geographical location to another. We use

230

Van Steenwinkel et al.

‘orientation’ in a broad sense of the word. Including identity as one dimension of orientation, we also consider questions like: Can I be myself here? Does this place detach me from who I am, my past and my hoped-for future? Can I find a place of my own? Do I feel at home in this place? In fact, ‘home’ is a special reference point in all dimensions of orientation, i.e. time, space and identity. Indeed, one’s home is a particular meaningful place (Norberg-Schulz, 1971; Tuan, 1977; Madanipour, 2003) and it has an important role in colouring one’s identity (Tuan, 1977; Madanipour, 2003; Chapman, 2006). Yi-Fu Tuan (1977) elicits the time dimension of home by writing that “in an ideal sense home lies at the center of one’s life, and center (we have seen) connotes origin and beginning.” Exploring how an entity like home carries a sense of time, space and identity may reveal how, for a particular person, it can serve as a point of reference for orientation. The particular experiences of people with dementia may elicit new insights, or vice versa, such an exploration may reveal why people with dementia sometimes can (not) orient themselves. After all, people with dementia not seldom express the desire to go home, even when in fact they are at the location where they currently live. Based on these considerations, we delineate our research question a little more precisely: How can the physical environment allow or impede a person to orient their self in time-space-identity? In other words: How can physical entities provide a sense of time, space and identity, through which they can be points of reference for orientation? Hence, we look at how one dimension of orientation, i.e. the physical entities of space, contributes to or hinders the overall orientation, i.e. orientation in time-space-identity.

23.3.2 Movement If we take our research question a step further, the question arises: How does a person develop a sense of time, space and identity? Literature study revealed that ‘movement’ is a very important, if not an essential key to the development of a sense of time, space, and identity. We found strong indications that a focus on movements may yield an understanding of the lived experiences of people with dementia. Tim Ingold (2000) contends that “movement is the very essence of perception” and Tuan (1977) explains the role of movement in orientation in time and space (and therefore also in identity) by writing: “We can have sense of space because we can move and of time because, as biological beings, we undergo recurrent phases of tension and ease. The movement that gives us a sense of space itself is the resolution of tension. When we stretch our limbs we experience space and time simultaneously - space as the sphere of freedom from physical constraint and time as duration in which tension is followed by ease.” (Tuan, 1977)

Movements can be of different kinds and of different scales. For example, the possible bodily movements in a room influence how a person experiences that room. At a bigger scale, the daily commuting between home and office has a part

Spatial Clues for Orientation: Architectural Design Meets People with Dementia

231

in the ‘image’ a person develops of their living environment. Cycles like night and day or the seasons are also movements, which “are embodied, incorporated into our very constitution as biological organisms” (Ingold, 2000). Such movements contribute to orientation. Being out of tune with them may be stressful and/or confusing, for example, when a person with dementia gets up in the middle of the night and thinks he/she needs to buy some food, but finds all stores closed (Braam, 2005), or when, in the late afternoon, a person feels it is time to go home, but is hindered by carers or locked doors. Movements, sequences of tension and ease, should not only be considered as physical, but also cognitive and emotional, e.g. being ‘moved’ by a(n emotionally charged) conversation with friends or family versus a peaceful time on your own. Piaget and Inhelder (1971) describe that a comprehension of the physical environment evolves from internalised actions rather than from mental images. We know things not only by using our brain, but also by incorporating the movements of other parts of our body. The following quote from a person with dementia illustrates that his hands ‘know’ his PIN code when standing in front of the keyboard, while his head could not remember it in advance: “En route, I try to remember the PIN code. […] When I stand in front of the desk, I know.” (Braam, 2005)

The development of a sense of time also requires movements, and more precisely the co-ordination of moving entities (like your own body, that of others, or objects) at different velocities (Piaget, 1969). Norberg-Schulz (1971) - drawing on the work of Otto Friedrich Bollnow (1963) - describes the importance of movement to human beings as follows: “How we get from one place to another is a basic aspect of man’s being in the world. […] Life itself can be understood as movement from one condition to another. This movement is incessant and continuous, but it has rhythm and form. […] Furthermore, man is part of a system of natural rhythms, such as night and day, the change of seasons, and his own ‘ages’. (Norberg-Schulz, 1971)

Therefore, movement seems to be a basic aspect of a person’s well-being or illbeing. In the context of care for people with dementia we thus consider it valuable to investigate their ‘movements’, through time, space and identity and the way their built environment may reveal or hide cycles like night and day or possible other events that mark particular points in time. Consequently, we reformulate our research question as follows: How can the physical environment as a motor medium afford or impede a person’s orientation in time-space-identity? Setting out our exploration by considering the experiences of people with dementia raises several sub-questions: Which activities, pauses, events make up or are left out of the ‘timescape’ of a person with dementia, supporting or hindering orientation in time? Which physical features induce movements that constitute ‘the image’ of the environment for people with dementia and, therefore, operate as clues for orientation in space? How can the physical environment afford a place of one’s own affording orientation in one’s own

232

Van Steenwinkel et al.

identity? And how can the physical environment afford the feeling of coming or being home, supporting orientation in time-space-identity? By analogy with ‘landscape’. Ingold (2000) uses the term ‘taskscape’, yet we prefer ‘timescape’ because it refers more explicitly to the dimension of time. By analogy with ‘the image of the city’, described by Kevin Lynch (1960), though we will not limit our study to only visual aspects of space.

23.4 Methodological Issues Addressing the proposed research questions requires consideration of both the physical environment and people living with dementia and these two in relation to each other. We plan to tackle the research questions in two ways: firstly, through ethnographic research that is a combination of an analysis of the physical environment, interviews and observations, and secondly, through ‘research by design’.

23.4.1 Ethnographical Research Studies on people with dementia typically focus on one particular housing context, e.g. either large scale residential settings (nursing homes or Alzheimer‘s special care units) (Sloane et al., 1998), or small group living facilities ( Elmståhl et al., 1997; Van Audenhove et al., 2003), which limits the transferability of the findings to other contexts. By contrast, our research project aims to study the spatial experience of people with dementia across different contexts and housing schemes. Possibly, these differences can even accentuate how people with dementia can (not) orient themselves in time-space-identity. We take into account persons at their homes - like individual houses, apartments, service flats - and persons in group housing facilities - like small group living facilities or traditional large-scale nursing homes. These settings are analysed in terms of five spatial aspects: • • • • •

spatial articulation, which deals with social aspects of space, like being (semi-)private or public; enclosure, which deals with how spaces are delineated by physical entities like walls, windows, level changes, furniture, signs, etc.; sensory qualities, like colour, sound, smell, warmth, texture, etc.; materials, which will be analysed not only for their sensory qualities but also for the meanings they may evoke; form, measurements and proportions of rooms, openings in rooms, outdoor places, objects, etc.

The list is based on spatial aspects described by Ola Nylander (1999), which have been modified in a study on living environments for older people (without dementia). We do not consider this list of spatial aspects fixed, but allow it to be modified during the research if necessary.

Spatial Clues for Orientation: Architectural Design Meets People with Dementia

233

The spatial analysis is complemented with both interviews and observations. Through interviews we try to explore how entities in a person’s physical environment carry a sense of time, space and identity. All too often, cognitive impairment is assumed to obliterate the ability of people to evaluate and share their evaluations with others (Cotrell and Schulz, 1993). Interviewing people with dementia may be challenging, but it is not impossible. In fact it has been done many times before. Louise Nygård (2006) draws on more than 10 years of ethnographic research to reflect on and make suggestions for how the scientific exploration of the experiences of people with dementia may be undertaken. She suggests that “a combination of qualitative observations and adapted interviews may make it possible for people with dementia to participate as research informants” (Nygård, 2006). Interviews should be adapted to the participant’s capacities, in terms of formulating questions and interview duration (Nygård, 2006). Therefore, the researcher must be very flexible. Number, length, focus and context of interviews may be different with every participant (Nygård, 2006). In our research, the number of interview participants will be decided based on a pilot study. Observations will offer extra material on the daily life and experiences of people with dementia by mapping out (i.e. making notes and drawings) the ‘movements’ by persons with dementia and their environment. These ‘maps’ are made over a period of time and include accounts of what people do and say, when they use a space, where they sit, the way they walk, etc.

23.4.2 Research by Design In a second part, we will confront the perspective and appreciation of people with dementia related to the physical environment with the perspective of architects, and in particular with their designerly ways of knowing (Cross, 2006). The approach used in this work package shows similarities with research by design in the sense that the activity of designing is used as a way to generate knowledge. However, unlike in typical instances of research by design (Pedgley, 2007), the researcher will not perform this design activity herself, but will call in others’ design activities. We plan to involve architects in the research process in two ways: • •

by challenging architects to freely (i.e. without any precondition) reconsider (the design of) space from the perspective of people with dementia, introduced through information from the ethnographic fieldwork; by introducing insights from the ethnographic fieldwork as input for a realworld design assignment.

In this way we call in architects’ expertise to tackle our research questions. The thoughts of these architects are expected not only to offer a glimpse of future possibilities, were architectural design to meet people with dementia, but also to further our understanding of the spatial experience of people with (and without) dementia. Unlike scientists who set out specifically to study a problem/issue,

234

Van Steenwinkel et al.

architects (like other designers) learn about the nature of the problem/issue largely as a result of trying out solutions (Lawson, 1998; Cross, 2006). Moreover, the design outcomes of both approaches are presented to people with dementia, their relatives and/or care givers. Indeed, former research projects have shown that the evaluation of (unrealised) designs by possible users can elicit additional insights in their experiences, like in research projects dealing with the experiences of visually impaired people (Vermeersch et al., 2011).

23.5 Conclusions The general aim of our research is to better understand the spatial experiences of people with dementia, firstly because the physical environment holds great potential to improve their well-being, and secondly because of the assumption that their perspective could expand architects’ design expertise. Based on preliminary research we figured out that - beside space - dimensions of time and identity should be included in the research, since time, space and identity are inextricably bound up with each other. Moreover, people with dementia, because of their memory impairments, find difficulties in orientating in all three dimensions. A literature study elicited useful concepts and theories to explore the experiences of people with dementia. Most importantly, many authors consider movement essential to perception. That is why we focus on how the physical environment as a motor medium assists or impedes people in orientating themselves in time-space-identity. Our research covers several (if not all) types of dementia, several types of housing facilities, and people with different backgrounds. In this way, insights can be transferred to different contexts. Nevertheless, we are aware that other important factors, e.g. socio-cultural background, have a part in how people negotiate space. For that reason material from the fieldwork should be interpreted cautiously with regard to the transferability of insights to different contexts. We hope to add new insights on important aspects of how people with frailty and cognitive impairment negotiate space and that our results will help designers to improve the orientation and wellbeing of people with dementia.

23.6 Acknowledgements The research reported here received support from the European Research Council under the European Community’s Seventh Framework Programme (FP7/20072013)/ERC grant agreement No. 201673. The authors would like to thank the persons working in the residential care centres with whom we talked and who showed us around in the buildings where they work, the architects who participated in interviews, and especially the persons with dementia, their friends, relatives and care givers for sharing time and experiences.

Spatial Clues for Orientation: Architectural Design Meets People with Dementia

235

23.7 References American Psychiatric Association (2000) Diagnostic and statistical manual of mental disorders: DSM-IV-TR. American Psychiatric Publications, Washington, DC, US Van Audenhove C, Declercq A, De Coster I, Spruytte N, Molenberghs C, Van den Heuvel B (2003) Kleinschalig genormaliseerd wonen voor personen met dementie. Garant, Apeldoorn, The Netherlands Van Audenhove C, Spruytte N, Detroyer E, De Coster I. Declercq A, Ylieff M et al. (2009) De zorg voor personen met dementie: Perspectieven en uitdagingen. Koning Boudewijnstichting, Brussels, Belgium Boden C (1998) Who will I be when I die? HarperCollins, Melbourne, Australia Bollnow OF (1963) Mensch und raum. Kohlhammer, Stuttgart, Germany Braam S (2005) Ik heb Alzheimer: Het verhaal van mijn vader. Nijgh & Van Ditmar, Amsterdam, The Netherlands Bryden C (2005) Dancing with dementia: My story of living positively with dementia. Jessica Kingsley, London, UK Calkins M, Sanford JA, Proffitt MA (2001) Design for dementia: Challenges and lessons for universal design. In: Preiser WFE, Ostroff E (eds.) Universal design handbook. McGraw Hill Professional, NY, US Chapman S (2006) A “new materialist” lens on aging well: Special things in later life. Journal of Aging Studies, 20(3): 207-216 Cotrell V, Schulz R (1993) The perspective of the patient with Alzheimer’s disease: A neglected dimension of dementia research. The Gerontologist, 33(2): 205-211 Cross N (2006) Designerly ways of knowing. Birkhäuser Verlag, Basel, Switzerland Elmståhl S, Annerstedt L, Ahlund O (1997) How should a group living unit for demented elderly be designed to decrease psychiatric symptoms? Alzheimer’s disease and Associated Disorders, 11(1): 47-52 Friel McGowin D (1993) Living in the labyrinth: A personal journey through the maze of Alzheimer’s. Thorndike Press, Thorndike, ME, US Godderis J (1992) Handboek geriatrische psychiatrie. Garant, Leuven. The Netherlands Hall E (1969) The hidden dimension. Doubleday, NY, US Ingold T (2000) The perception of the environment: essays on livelihood, dwelling and skill, Routledge, London, UK Lawson B (1998) How designers think, 3rd edn. Butterworth Architecture, Oxford, UK Lynch K (1960) The image of the city. The MIT Press, MA, US Madanipour A (2003) Public and private spaces of the city, 1st edn. Routledge, London, UK Norberg-Schulz C (1971) Existence, space and architecture. Studio Vista, London, UK Nygård L (2006) How can we get access to the experiences of people with dementia? Scandinavian Journal of Occupational Therapy, 13(2): 101-112 Nylander O (1999) The home as architecture. Bok & Bild, Gothenburgh, Sweden Pedgley O (2007) Capturing and analysing own design activity. Design Studies, 28(5): 463483 Piaget J (1969) The child’s conception of time. Routledge and Kegan Paul, London, UK Piaget J, Inhelder B (1971) The child’s conception of space, 4th edn. Routledge and Kegan Paul, London, UK Sloane PD, Mitchell CM, Preisser JS, Phillips C, Commander C, Burker E (1998) Environmental correlates of resident agitation in Alzheimer’s disease special care units. Journal of the American Geriatrics Society, 46(7): 862-869 Sternberg E (2009) Healing spaces: The science of place and well-being. The Belknap Press of Harvard University Press, Cambridge, MA, US

236

Van Steenwinkel et al.

Tuan Y-F (1977) Space and place: the perspective of experience. Edward Arnold (Publishers) Ltd, London, UK Vermeersch P-W, Nijs G, Heylighen A (2011) Mediating artifacts in architectural design: A non-visual exploration. In: Leclercq P, Heylighen A, Martin G (eds.) Designing together. CAADfutures. Les Editions de l’Université de Liège, Liège, Belgium Zeisel J (2001) Universal design to support the brain and its development. In: Preiser WFE, Ostroff E (eds.) Universal design handbook. McGraw Hill Professional, NY, US

Index of Contributors e

Afacan Y .....................85

Cirba N ...................... 43

Goldhaber T.S .......... 105

Akeley C ...................185

Clarkson P.J ....... 55, 105, .................. 115, 133, 145

Goodman-Deane J....... 55

Almuhim M ..............185 Annemans M ................3 Attanayake D ............125 Barmoy P ..................185 Beavan P ...................185 Beck C ......................185 Bhakta B .....................33 Biswas P ...................195 Blair A ......................185 Bortz A .....................185 Bracewell R.H ..........133 Bradley B ..................185

Cobb S ..................... 155

Gorman M ................ 185 Gregory C ................. 185

Conduit G ................... 93 Cooke T ..................... 23 Corso A ...................... 93 Crouch D .................. 185

Henderson H .............. 75 Herriott R ................. 165 Heylighen A ........ 3, 207, .................................. 227

Dalke H....................... 93

Holt R.......................... 33

Dehmer G ................ 185

Horberry T ................. 23

Denholm-Price J ....... 125

Hosking I .................... 55

Dong H ...................... 13

Hunter G ................... 125

Edlin-White R........... 155

Jokisuu E .................. 115

Bradley M....................55

Floyde A .................. 155

Carter M ...................185

Gallagher J.F .............. 33

Ceccacci S ..................65

Germani M ................. 65

Chakraborty S .............75

Goddard N ............... 175

P. Langdon et al. (eds.), Designing Inclusive Systems, DOI: 10.1007/978-1-4471-2867-0, © Springer-Verlag London 2012

Jordan P.W ................. 13 Jouffrais C ................ 217

Kammoun S ............. 217 Kharal N ..................... 43

237

238

Langdon P.M ......55, 105, ...........115, 133, 145, 195 Lanier E ....................185 Lazar J .................75, 185 Levesley M.C .............33

Index of Contributors

Oriola B ................... 217 Patmore J.J ............... 133 Pfluegel E ................. 125

Mengoni M .................65 Mieczakowski A .......133

Riedel J .................... 155 Ritgert D .................. 185 Rogers Jr. R ............. 185 Rosenwald S ............ 185

Wang J ..................... 155 Weightman A.P.H ...... 33 Weir R ........................ 75 Wells J ...................... 185 Wentz B ............. 43, 185 Willis C .................... 185

Mon-Williams M ........33 Moran J .......................43

Vermolen H................... 3 Waller S.D................... 55

Riaz A ........................ 93 McIntee A .................185

Van Steenwinkel I .... 227

Preston N ................... 33

Lewthwaite S ............155 Macé M.J-M .............217

Van Audenhove C........ 3, .................................. 227

Sizemore B ................ 75

Wingo-Jones K ......... 185

Slate M ....................... 43 Nelson Jr. R ..............185

Sullivan S ................. 185

Yatto T ..................... 185

Tenneti R .................... 55

Zitkus E .................... 145

Nickpour F...................13 Nicolle C ..................175