wendy e. pentland - time use research in the social sciences

Time Use Research in theSocial Sciences


Edited by

Wendy E. Pentland

Andrew S. Harvey

M. Powell Lawton

Queens UniversityKingston, Ontario, Canada

St. Marys UniversityHalifax, Nova Scotia, Canada

Philadelphia Geriatric CenterPhiladelphia, Pennsylvania

and

Mary Ann McCollQueens University

Kingston, Ontario, Canada

Kluwer Academic PublishersNew York, Boston, Dordrecht, London, Moscow

47155-8

306-45951-5

!""# "$"%

& '

() *(*)**""

"&" " )

+ ,)

-. (//00

1 (//000

Contributors

Mike Barns School of Architecture, University of Auckland, Private Bag

Andrew S. Harvey Department of Economics, St. Marys University,

M. Powell Lawton Polisher Research Institute, Philadelphia Geriatric

92006, Auckland 1020, New Zealand.

Halifax, Nova Scotia, Canada B3H 3C3.

Center, Philadelphia, Pennsylvania 19141.

Mary Ann McColl Division of Occupational Therapy, School of Reha-bilitation Therapy, Queens University, Kingston, Ontario K7Z 3N6.

William Michelson Centre for Urban and Community Studies, Univer-sity of Toronto, Toronto, Ontario, Canada M5S 2G9.

Wendy E. Pentland Division of Occupational Therapy, School of Reha-bilitation Therapy, Queens University, Kingston, Ontario K7L 3N6.

John P. Robinson Department of Sociology, University of Maryland,

Jerome F. Singleton School of Health and Human Performance, Dal-

College Park, Maryland 20742.

housie University, Halifax, Nova Scotia, Canada B3H 3J5.

versity of Waterloo, Waterloo, Ontario, Canada N2L 3G1.

tion, University of Guelph, Guelph, Ontario, Canada N1G 2W1.

Bryan J. A. Smale Department of Recreation and Leisure Studies, Uni-

Joseph A. Tindale Department of Family Relations and Applied Nutri-

K. Victor Ujimoto Department of Sociology and Anthropology, Univer-sity of Guelph, Guelph, Ontario, Canada N1G 2W1.

v

vi CONTRIBUTORS

Gail Whiteford School of Occupational Therapy, Auckland Institute of

Ann Wilcock School of Occupational Therapy, University of South Aus-

Jiri Zuzanek Department of Recreation and Leisure Studies, University

Technology, Private Bag 92006, Auckland 1020, New Zealand.

tralia, Adelaide, South Australia 5000.

of Waterloo, Waterloo, Ontario, Canada N2L 3G1.

Preface

Despite the fact that, for most of us, time is a central focus of our lives, theexamination of what we do with our time and why has received limitedattention as a method for understanding human behavior in the socialsciences. Humans' view and use of time shows tremendous variation,including across cultures and with age, lifestyle, and gender. For many ofus, a sense of time is ever-present. We speak of time as a commodity, aresource, an ally, an enemy, and a gift. It may be on our side, on our hands,with us, or against us. We perceive it to change speeds (dragging vs. flyingvs. standing still) and lest it get away on us, we attempt to harness andcontrol it with clocks, schedules, and deadlines. We describe our use oftime in a myriad of ways: we spend it, save it, waste it, kill it, give it, take it,and grab it.

The impetus for this book grew from a three-day research symposiumwhere established time use researchers from a variety of disciplines fromCanada, the United States, Finland, Australia, and New Zealand gatheredtogether to merge their knowledge and resources to collaborate in examin-ing the relationship between human time utilization and health and well-being. We gratefully acknowledge the financial support for the sympo-sium received from the Government of Canada's Program for Interna-tional Research Linkages and M. Powell Lawton, without whose supportand encouragement this book would likely not exist.

The purpose of this book is both to instruct and to demonstrate the use and variety of applications of time use methodology. It is multidisciplin-

ary, multinational, and multicultural. The contributors are experts in their fields. Students of research in the social sciences will find that the breadth and detail of the text make it a particularly useful research resource. Similarly, the book will appeal to experienced research scholars who may

vii

viii PREFACE

be less familiar with this particular methodology and its potential applica-tion to their research questions.

The first section presents information for planning and conductingtime use research and the analysis of the data. In Chapter 2, AndrewHarvey discusses various important considerations and provides clearguidelines for sampling and diary construction, including how to includevarious subjective and contextual variables. Coding and file setup of time-diary data are critical steps in this research process and both are covered indetail in this chapter. Various analysis issues are then reviewed, includingthe calculation of descriptive characteristics of time use (duration, partici-pation, activity sequences) and contextual features (deriving activity set-tings through the use of hypercodes, episode sampling).

In Chapter 3, John Robinson provides a brief history of the use of timeuse methodologies in Canada and the United States. He provides in-depthinformation regarding the reliability, validity, and limitations of varioustypes of time use data. Multivariate analysis is then outlined and theexamination of contextual variables is demonstrated.

William Michelson, in Chapter 4, focuses on less conventional appli-cations of time use research. He illustrates how time use data and associ-ated contextual information can be used to try to understand and explainhow humans everyday contexts impact on their time use. Examples aregiven, including the influence of physical environment contextual vari-ables (housing, neighborhoods, urban infrastructure) on socialization andunderstanding working mothers' time use by including subject emotionvariables (perceived tension, perceived control, perceived available timeor "busyness").

The intent of the second section is to expose readers to the widevariety of existing applications of time use methodology and, we hope,stimulate researchers' thinking and awareness of how the methodologymay be a valuable tool to apply for answering their research questions. InChapter 5, Powell Lawton, long recognized for his work in gerontology,discusses the unique issues and considerations when applying time-budget methods to research with the elderly. He then gives detailed guide-lines and illustrations of the use of time budgets to understand the el-derly's psychological perceptions of time use and its influence on their lifesatisfaction and well-being.

Time use data are powerful illustrators of role behavior and lifestyle.Examples of such illustrations are discussed in Chapters 68. Jiri Zuzanekand Bryan Smale describe how they used time use data to examine therelationships between life cycle, daily time use, and weekly rhythms of everyday life. In Chapter 7, Joseph Tindale examines the temporality of family and the life cycle and the adaptations people must make to different

PREFACE ix

stages. He outlines the changes in the meaning of time that result from theimpacts of various life-cycle events such as employment, unemployment,and becoming a parent. In Chapter 9, Wendy Pentland and Mary AnnMcColl discuss important considerations for conducting time use researchwith persons with disabilities and provide an example of research compar-ing time use between persons with, and those without, severe physicaldisabilities.

The impact of culture on lifestyles and behavior can be examined withtime use methodologies. Chapters 911 provide examples of time useresearch from Australia and New Zealand, and with elderly people fromethnic minorities. Ann Wilcock discusses the influence of biological neces-sity and sociocultural influences on our use of time and considers thepotential benefits of applying quantitative and qualitative methods to theunderstanding of human time use.

Each culture has its own view and concept of time. This is an essentialconsideration before planning to conduct time use research in countrieswhere routines and the time of day are marked not by clocks and minutes,but by the sun, the moon, the tides, or body needs (hunger, fatigue). Timeuse data collection methods should be considered ethnospecific in thattechniques that work in one culture may fail miserably in others unlessmajor methodological modifications are made. Chapter 10 by Gail White-ford and Mike Barns clearly illustrates this issue. In their chapter, themeaning of time both now and prior to colonial contact is presented fromthe perspective of the Maori, who are the indigenous people of Aotearoa(New Zealand). They outline the inherent dangers in applying traditionaltime use data collection techniques across cultures and suggest recommen-dations for methodological modifications.

In Chapter 11, Victor Ujimoto also addresses cross-cultural issuesrelated to time use research in gerontology. Examples of issues and refine-ments are given through the presentation of a number of his studiesexamining aging of persons from ethnic minorities in Canada. Jerry Sin-gleton, in Chapter 12, provides an example of the application of time usemethods to a very specific population; leisure behaviors in the elderly andthose with Alzheimer's disease. Cognitive deficits in this latter populationpresent significant challenges to data validity and reliability and the reader will find the strategies outlined very useful.

Finally, in Chapter 13, Pentland and Harvey outline the current issues, challenges, and future trends for time use research. The intent of this final chapter is to allow readers to put what they have learned about time use research into context and see both its strengths and those areas where time use research needs to develop further.

We hope that this book will clearly show the unique, adaptable, and

x PREFACE

cross-discipline applications of time use research methods and providereaders with the basics to begin research with this method. It is our beliefthat as we learn more about the complexity of human behavior, individu-ally and socially, time use research techniques will become an increasinglyinvaluable tool that researchers in the social sciences cannot overlook.

Contents

I. INTRODUCTION

Chapter 1. Time Use Research . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3

Andrew S. Harvey and Wendy Pentland

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3What Is Time Use Research? . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3

Historical Development of Time Use Studies . . . . . . . . . . . . . . . . . . . 5

Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14

Chapter 2. Guidelines for Time Use Data Collection and

Andrew S. Harvey

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19Collection Guidelines . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20Data-File Editing and Creation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25Analysis Issues . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27

Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 42References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 42

47

Applications of Time Use Data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8

Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19

Chapter 3. The Time-Diary Method: Structure and Uses . . . . . . . . . .

Uses of Diary Data. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 48

Features of the Diary Method. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 54The Zero-Sum Character of Time . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 55

John P. Robinson

xi

xii CONTENTS

Measuring How People Spend Time . . . . . . . . . . . . . . . . . . . . . . . . . . 56

Diary Studies in Canada . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 67

697780

Previous Time-Diary Studies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 64

Activity Coding . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 68

Procedures to Analyze Time-Diary Data . . . . . . . . . . . . . . . . . . . . . . .A Larger Multivariate Model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .Methodological Properties of Time Diaries . . . . . . . . . . . . . . . . . . . . .Summary and Conclusions. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 83

of Time Projects . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 84Appendix: Methodology of the 1965-1985 Americans' Use

References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 87

Chapter 4. Analysis and Exploration of Meaning and Outcomes

William Michelson

Introduction. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 91Two Purposes of Time Use Research: Descriptive Patterns and

Indicators of Meaning . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 92

Four Ways to Study Meaning and Outcomes . . . . . . . . . . . . . . . . . . .Concluding Comments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 103

in Connection with Time Use Data . . . . . . . . . . . . . . . . . . . . . . 91

93

References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 104

II . USING TIME USE RESEARCHTO EXAMINE LIFESTYLE VARIABLES

A . Quality of Life

Chapter 5. Methods and Concepts for Time-Budget Research

M . Powell Lawton

Introduct ion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 107

on Elders . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 107

Methodological Issues. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 108Psychological Aspects of Time Use. . . . . . . . . . . . . . . . . . . . . . . . . . . 117Conclusions. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 122References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 123

B . Roles and Lifestyles

Chapter 6. Life-Cycle and Across-the-Week Allocation of Time

Jiri Zuzanek and Bryan J . A . Smale

Review of Literature . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 128

to Daily Activities . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 127

CONTENTS xiii

Statement of the Research Problem .. . . . . . . . . . . . . . . . . . . . . . . . . . 135

Database and Operationalization of the Variables. . . . . . . . . . 135Data Analyses. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 137

Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 145Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 149References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 151

Chapter 7. Variance in the Meaning of Time by Family Cycle.Period. Social Context. and Ethnicity . . . . . . . . . . . . . . . . . . . . .

Joseph A . Tindale

The Times of Our Lives . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 155

Work and Family: Balancing Time .. . . . . . . . . . . . . . . . . . . . . . . . . . . 158

Becoming a Parent: Cohort Changes in Appropriate Timing . . . 159Responses to Unemployment in Different Family Life-Cycle

Periods. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 161162165

References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 167

Chapter 8. Application of Time Use Research to the Studyof Life with a Disability . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

Wendy Pentland and May Ann McColl

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 169

What Is Meant by Time Use?. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 170

The Relationships among Time Use, Health and Well-Being. . . 171What Is Known about Time Use and Variables Relevant to

Persons with Disabilities?. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 173

Relationships between Disability and Use of Time . . . . . . . . . . . . . . 173Methodological Considerations When Examining Use of Time

in This Population . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 178References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 183

C . Culture

Chapter 9. Biological and Sociocultural Perspectives on Time

Ann Wilcock

155

Getting On and Getting Along . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .Conclusions about Family Life-Cycle Flexibility . . . . . . . . . . . . . . . .

169

Use Studies. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .... . . . . . 189

Biological and Sociocultural Temporality. . . . . . . . . . . . . . . . .Research Paradigms. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 194Integrating Research Approaches .. . . . . . . . . . . . . . . . . . . . . . . . . . . . 198

190

Data Collection Methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 200

xiv CONTENTS

Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 208References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 208

Chapter 10. Te Ao Hurihuri: New Zealand's First Time . . . . . . . .

Gail Whiteford and Mike Barns

211

Traditional Time . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 212The Colonial Experience . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 215

Zealand . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 220

Ka Awatea: The Dawning . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 229

Research Ethnicity and Aging . . . . . . . . . . . . . . . . . . . . . . . . . . .

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 231

Time-Budget Methodology for Gerontology . . . . . . . . . . . . . . . . . . . .Refinements in Time-Budget Methodology. . . . . . . . . . . . . . . . . . . . .Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 240References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 241

Aotearoa Now: Time and Time Use in Contemporary New

227

Chapter 11. Time Budget Methodology in Social Science

K. Victor Ujimoto

231

236238

III. CONCLUSION

Chapter 12. Lessons from Leisure-Time Budget Research

Jerome F. Singleton

Time Budget . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 246247

Therapeutic Recreation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 249

Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 255

Implications for Practice. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 245

Background . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 245

Time Budget and Older Persons . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .Implications for Practice . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 248

Alzheimer's Disease . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 251

References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 255

Chapter 13. Future Directions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 259

Wendy Pentland and Andrew S. Harvey

Overview . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 259Applications . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 261

xvCONTENTS

Theoretical and Methodological Issues. . . . . . . . . . . . . . . . . . . . . . . .References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 267

264

Index 269. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

IIntroduction

1Time Use Research

Andrew S. Harvey

and Wendy E. Pentland

INTRODUCTION

Time use methodology can provide a window on actual lifestyles, therebypermitting a rich, objective, and replicable basis on which to make empiri-cal judgments. By providing contextual information about what people do(nature of tasks performed, social and physical environmental context,level of stress experienced, perceived satisfaction), the methodology cangenerate invaluable information for understanding human behavior prob-lems and be used to guide planning and policy development. This chapterintroduces the concept of time use studies and presents a brief overview oftheir historical development and application.

WHAT IS TIME USE RESEARCH?

Time use studies show how people use their time. Minimally, theyshow what activities people do week to week or day to day. Maximally,they show what people are doing, where they are, who they are with, andhow they feel minute to minute.

Andrew S. Harvey Department of Economics, St. Mary's University, Halifax, NovaScotia, Canada B3H 3C3. Wendy E. Pentland Division of Occupational Therapy,School of Rehabilitation Therapy, Queen's University, Kingston, Ontario, Canada K7L 3N6.

Time Use Research in the Social Sciences, edited by Wendy E. Pentland, Andrew S. Harvey, M.Powell Lawton, and Mary Ann McColl. Kluwer Academic/Plenum Publishers, New York,1999.

3

4 ANDREW S. HARVEY AND WENDY E. PENTLAND

Time use studies employ a wide variety of methodologies. Qualitativemethods, including ethnography, can provide rich time use data, but arecostly and can suffer from lack of scientific validity. Direct observation isarguably the most accurate; however, it is extremely expensive, and sub-jects' behavior may be altered by an observer's intrusion. Similarly, thismethod can be prone to a high, nonrandom refusal rate by subjects notcomfortable with being observed.

Activity frequency and duration surveys, which are often stylizedlists of activities for which subjects provide frequency and duration ofparticipation information, provide another measurement approach. Theseinclude labor force surveys, shopping studies, travel studies, readershipsurveys, and studies of general leisure time use and particular facets of itsuch as television viewing habits. In general, these studies require therespondent to complete a checklist showing the extent of participation indefined activities. In order to be accurate and consistently interpreted byall subjects, the activities must be very narrowly defined. This results inlong lists of activities, and there is danger of poor subject compliance dueto the resulting tedium. Other information concerning the activity mayalso be sought, such as satisfaction and preferences. Alternatively, a re-spondent may be asked to keep a log of specific activities such as shop-ping, television viewing, or travel. In this case, it is possible to get informa-tion on duration, sequence, and various other dimensions for the activitybeing logged. Activity data can be collected from the supply side as well.On the supply side, data from which activity participation and time usecan be inferred are collected on the use of various facilities such as li-braries, museums and theaters. Another approach to collecting time use oractivity data uses a beeper. Respondents carry a beeper and a protocol tobe completed each time the beeper sounds. The protocol can collect a widerange of objective and subjective data. While each of the foregoing canprovide some insights into how people use their time, the reports areepisodic and typically taken out of context.

The time diary provides a more comprehensive means of collectingtime use data. In contrast to the previous methods outlined. the time diaryplaces activities in context. It "is a log or diary of the sequence and dura-tion of activities engaged in by an individual over a specified periodmost typically a 24-hour day" (Converse, 1968). All activities during the specified period are recorded, including time of start and completion ofeach activity. A broad range of subjective and contextual data can be collected at the same time (where, who with, perceived satisfaction, con-trol, stress, etc.). An advantage of time diaries is that subjects can use their own terminology., which can later be coded according to classification criteria. The focus of this volume is on diary-based time use studies.

TIME USE RESEARCH 5

Chapters 2 and 3 provide considerable insight into the method, content,and analysis of time use studies. The remaining chapters illustrate vividlywhat can be learned from time diary studies.

HISTORICAL DEVELOPMENT OF TIME USE STUDIES

How Working Men Spend Their Time (Bevans, 1913) and Round about aPound a Week (Pember-Reeves, 1913) both published in 1913, the former inthe United States and the latter in the United Kingdom, appear to be theearliest published accounts of time use. During the second decade of thiscentury, time use research emerged in Europe in conjunction with early studies of living conditions of the working class in response to pressures generated by the rise of industrialization. In the United States, household time-allocation studies date from 1915 in the U.S. (Bailey, 1915). The various time use studies examined shares of activities such as paid work, house-work, personal care, leisure, and so on, in the daily, weekly, or yearly time budget of the population. They also examined how the time use varied among population groups such as workers, students, and housewives, and in the use of leisure time. Most often, respondents were asked, through stylized questions, to estimate how much time they allocated to various activities.

The bulk of preWorld War II diaries originated in the Soviet Union, Great Britain, and the United States, with a number of others in France and Germany. The earliest sophisticated study was that of S. G. Strumlin in the Soviet Union in 1924. The study was undertaken for use in governmental and corrununal planning. A student of his, G. A. Prudensky redid Strumlin's study 35 years later on a similar sample. In 1923, the Osaka City office carried out the first time-budget study in Japan, "The Study of LeisureLife" (Tanaka, 1978). In the United States, home economists started usingtime use studies in the mid- to late 1920s to study farm and rural women(Avery, Bryant, Douthitt, & McCullough, 1996; Kneeland, 1929). Work began at Cornel1 during the 1920s on a program to study household output in terms of time use (Walker & Woods, 1976). Since that time, there has been extensive work focusing on household time use in the United States. In the early 1930s, a Westchester County survey launched a whole new era of studies of leisure (Lundberg & Komarovsky, 1934). Later in the 1930s, Sorokin and Berger (1939), in their Time Budgets of Human Behavior, pro-vided some fascinating insights into psychological and sociological moti-vations through an analysis of time-diary data. A number of small studies were carried out through the 1930s and 1940s in the United Kingdom, and in 1938, the audience research department of the BBC conducted the first of


several surveys providing time use information (Gershuny, 1983a). In theUnited Kingdom, some small-scale time use studies were carried out byMoser in 1948 (Gershuny, 1983b). In the decade following 1957, a largenumber of time budget surveys were conducted in Russia, approximately100, between 1958 in 1968, with three institutions playing the leading role intime budget research (Zuzanek, 1980). Since the 1950s, extensive time useresearch in Japan has explored the long working time of paid workers,houseworkers, media, and other leisure time use, as well as other issues.

The most ambitious time use study was the Multinational Time UseStudy in the mid-1960s directed by Alexander Szalai. That study, whichstill stands as a landmark in cross-national survey research, was unques-tionably the most significant time-diary undertaking. The study arose outof ideas generated in a conference held at Yale in 1963," The Use of Quanti-tative Political, Social and Cultural Data in Cross-National Comparisons"(Rokkan, 1966). Under the aegis of the study, data were collected in 13countries and 16 different survey sites. Unfortunately, data for one of thesites, Cuba, was never analyzed. That study has had a profound andlasting effect on the collection of time use data ever since. First, the codingscheme used in that study helped shape most all national time study rangeof subsequent venues (Elliott, Harvey, & Procos, 1976; Gonzalez & Gomez,1985; Yano, 1995). Third, the report on that study, The Use of Time, presenteda wide range of analyses using the time use data, thus broadening thescope of data collectors and data analysts.

Canadian time use studies date at least from the mid-1960s and early1970s. In 1965, Meissner undertook a study of industrial workers in PortAlbernie, British Columbia (Meissner, 1971). The first general populationsurvey, following the methodology of the Multinational Time Use Studywas undertaken in Halifax, Nova Scotia, in 19701971. The Halifax studywas, in fact, a time-space study that captured not only what people weredoing but also where they were, coded to a one-tenth kilometer grid(Elliott et al., 1976). The first nationwide time use study in Canada wasconducted in 1981 (Kinsley & O' Donnell, 1983). As a part of that study, over450 respondents to the 1971 Halifax study completed diaries, thus provid-ing a 10-year panel of time use (Harvey & Elliott, 1983). Statistics Canada, as part of its General Social Survey program, collected diaries for approx-imately 9,000 Canadians in 1986,1992, and 1998 (Frederick, 1995; Harvey, Marshall, & Frederick, 1991).

The United States has never conducted an official national time usestudy. However, the Bureau of Labor Statistics piloted a study in 1997. The major national studies in the United States have been undertaken by the Institute of Social Research (ISR) at the University of Michigan (Juster, 1985) and by the Survey Research Center at the University of Maryland (Robinson & Godbey, 1997).

TIME USE RESEARCH 7

Table1.1. Regular Data Collection Regimens

Country Sponsor Comparable Years

Canada Statistics Canada 1986, 1992, 1998 Holland Social Cultural Planning Bureau 1975, 1980, 1985, 1990, 1995Japan NHK 1960, 1965, 1970, 1975, 1980, 1985,

Japan Prime Ministers Office 1976, 1981, 1986, 1991, 1996Korea KBS 1983, 1985, 1987, 1990, 1995

Norway Statistics Norway 1970, 1980, 1990

1990, 1995

Since the early 1960s, time-diary studies have flourished. National time use studies have been conducted in virtually all Eastern and Western European countries. Many countries including Japan, The Netherlands, Canada, Korea, Finland, and Norway, conduct recurring studies every 5 to 10 years (see Table 1.1).

Of particular note are the time use studies of the Nippon Hoso Kyokai (NHK) in Japan, which have been carried out every five years since 1960. In the last decade, national time use studies have been carried out, or are being planned, by central statistical agencies in many countries including Austria, Australia, Canada, the Dominican Republic, Finland, Germany, India, Italy, Japan, Korea, New Zealand, Norway, South Africa, and Sweden.

Recently, a renewed multinational effort, the Eurostat Time Use Pro-ject, has started to take shape. This project emanated from a meeting of the Eurostat Working Party on Social Indicators in March 1992, which approved work toward a European time use survey. Subsequent meetings were held in 1992 and 1993, giving form to the project. That work formed the basis for subsequent development and piloting. In 1996 and 1997,18 countries participated in pilot time use studies undertaken under the auspices of EUROSTAT.

The EUROSTAT project can potentially provide research oppor-tunities similar to those provided by the Multinational Time Use Study. Other opportunities are provided by archival data from many countries. The Multinational Time Budget Data Archive, developed by Jonathan Gershuny at Essex University, contains comparably organized and format-ted person-level data for over 30 countries. Andrew Harvey has devel-oped a similar but less extensive archive of episodal time use data at Saint Marys University in Halifax, Canada. Given the high costs associated with time-diary data collection, these archives offer exciting opportunities for secondary analysis.

Time use methodology has come a long way since its early applica-


tions. In the last decade, there have been significant advances in the field,including innovative applications to nontraditional topics of inquiry, newand more sophisticated data collection methods, a variety of analysisstrategies, and increasing interdisciplinary collaboration. There is an Inter-national Association for Time Use Research (IATUR), with an annualscientific conference. Time use methodology is now used around the worldfor research by a broad range of disciplines. The literature on time use hasbeen remarkable in reflecting the interests of many different fields, includ-ing economics (Juster & Stafford, 1991; Goldschmidt-Clermont, 1987), busi-ness administration (Das, 1991; Grossin, 1993a, 1993b; McGrath & Rotch-ford, 1983), gerontology (Harvey & Singleton, 1989, Moss & Lawton, 1982),urban planning (Chapin, 1974; Gutenschwager, 1973), political science andoccupational therapy (Larson, 1990; McKinnon, 1992; Pentland, Harvey &Walker, 1998; Stanley 1995), nursing and medicine (Frankenberg, 1992),recreation and physical and health education (Rosenthal & Howe, 1984;Ujimoto, 1985), sociology/anthropology (Andorka, 1987; Elchardus & Glo-rieux, 1993,1994; Garhammer, 1995) and psychology (Block, 1990; Lawton,Moss, & Fulcomer, 1987). In turn, scholars, policymakers, professionals,and students have become sensitized to the power inherent in this ap-proach to examining and understanding human behavior.

Time use methodologies are recognized as capable of providing criti-cal information for a variety of purposes ranging from practical socialplanning concerns (electrical power consumption patterns, traffic flow) tocomplex theoretical problems (understanding behavior or social structurewithin a given society). Time use methodology provides hard, replicabledata that are the behavioral output of decisions, preferences, attitudes, andenvironmental factors. It can be used to examine, describe, and comparerole performance (Hasselkus, 1989; Ross, 1990); cultures and lifestyles(Chapin, 1974; Nakanishi & Suzuki, 1986); demands for goods and services(Juster & Stafford, 1985); poverty (Douthitt, 1993); needs of special groupssuch as the elderly, working mothers, and persons with disabilities (Baltes,Wahl & Schmidt-Furstoss, 1990; Leccardi & Rampazi, 1993); household andcommunity economies (Knights & Odih, 1995); and more recently elusivesocial indicators such as quality of life and well-being (Japan, Ministry ofEconomic Planning, 1975).

APPLICATIONS OF TIME USE DATA

There is virtually no avenue of human endeavor that is immune to some dimension of temporality. How long? When? Before or after? How often? At least one of these questions is likely to be relevant to any given

TIME USE RESEARCH 9

endeavor, issue, or policy. Consequently, time use data are highly relevantacross a virtually unlimited array of concerns, an array far too broad toexhaust. It is useful, however, to identify some of the applications of timeuse data to provide general insight into their use. Many of the chapters inthis volume provide more specific insight into particular issues and appli-

cations. Some general areas where time use data have been fruitful areshown here.

Economic Accounts

National economic accounts are under heavy criticism for their failureto include nonmarket production. Traditional economic variables inac-curately measure total productive activity (Goldschmidt-Clermont, 1987;Juster & Stafford, 1991). A corollary of this is concern over the lack ofknowledge of the structure of unmeasured productive activity, often deno-ted as informal economy (Urdaneta-Ferran, 1986). Failure to fully under-stand both the size and structure of an economy's total productive activityleads to the conception and implementation of, at best, many uselesspolicies, and at worst, harmful policies (Berio, 1986). The measurement oftime allocation provides a major data source for upgrading the accounts(Harvey & MacDonald, 1976). Productive work includes, in addition to

paid work, the only component included in the current national accounts,domestic activity, child care, time allocated to shopping and services, andeducation as a student. All of these activities are fundamental to theprovision of goods and services. Paid work time, when placed in thisperspective, although the largest single component of total productivework time, is less than half of all productive activity. Clearly, the exclusionof the other components yields a faulty view of total production in theeconomy. The Beijing Women's Conference, in 1995, called on nations tomeasure and document nonmarket production and to collect time use datato support such measurement. This proclamation has generated the great-est pressure yet for universal collection of time use data.

Labor Force Analysis

Current labor force statistics are weak on two counts. First, they appear not to measure well that which they purport to measure. It has been shown that the typical labor hours data collected do not truly reflect the reality of hours contributed to paid work (Niemi, 1983; Robinson & Godbey, 1997; Stafford & Duncan, 1976). A study based on Finnish data found that the direct interview question generated a weekly hours worked figure about 1 hour longer than that generated by the time diary (Niemi,


1990). There were differences among groups of workers, with self-employedand salaried occupations appearing to generate the greatest divergence.

Second, focusing only on paid work activity, they fail to account for allproductive activity and for constraints and opportunities related to the useof time. Time allocation studies give a far more complete picture of thenation's use of labor resources by focusing on all time use rather thansimply employment time. It is insufficient to characterize persons as em-ployed, unemployed, or not in the labor force. It is equally necessary to beconcerned with the use of time not identified as market production. Whatis the relationship between market and nonmarket production? Does thischange with economic conditions? Does the relationship play an impor-tant part in the ability of the economy to adjust to change? How canindividuals best be prepared to fulfill their roles in the two sectors? In thecase of shortages of labor, what is the potential labor pool available interms of expanded work time by current and potential labor-force partici-pants? How can production and working conditions be organized tooptimize the flexibility required to permit desired involvement in bothsectors? In times of surplus of labor in the economy, to what ends are thesurplus hours directed? Are these fully utilized, or can they be made moreproductive? What are the effects of nonstandard work times, flexible workhours, work sites, or other work arrangements? All of these questions canbe more adequately addressed with appropriate time use data.

Social Change

Time use data can be useful to government and business in imple-menting and evaluating change in such areas as working hours and pat-terns, shopping time, communications, and advertising. How much, where, and when do people work? Increasingly, work is less tied to specific places and times. Analysis carried out based on the Canadian 1981 Pilot Study provided insight into work patterns at that time. It showed that male workers dominated those working a traditional workday, that is, morning and afternoon hours only. Working morning, afternoon, and evening hours was about equally shared by men and women, and female workers dominated all other work patterns (Harvey, Elliott, & MacDonald, 1984). Traditional data on work hours fail to accurately reflect changes in the extent and pattern of work time. These are fully captured in time use data. Time use data help provide information on both the constraints and opportunities attendant with various work patterns and thus can be used to evaluate the impacts of alternative schemes. When and where do people work? How do individuals allocate their time with respect to the media?

TIME USE RESEARCH 11

How much, when, and where do they watch television, read papers, orlisten to the radio? Such information is captured by time use data.

As governments develop policies to discourage people from spendingtime in certain activities (smoking) or to encourage them to participate inothers (physical fitness programs), it is necessary to have evaluative mea-sures. Statistics Canada pointed out as early as 1974 that the idea of a timeuse survey is to develop a single survey vehicle from which a number ofmeasurements can be obtained on the effectiveness of government pro-grams (Statistics Canada, 1974).

Women's Concerns

Many general and specific concerns of women are directly address-able with time use data. In general terms, the issue of mismeasurement ofeconomic activity is particularly relevant to women, since women's activ-ities overwhelmingly dominate the nonmarket and informal sector in boththe more developed and developing countries (United Nations Interna-tional Research and Training Institute for Women [INSTRAW], 1995).

Specific concerns in developed countries include domestic work(Eichler, 1983; Vanek, 1974; Walker & Woods, 1976), child care (Michelson &Ziegler, 1982; Stone, 1972), the sexual division of labor (Gershuny, & Robin-son, 1988; Harvey & Clark, 1976; Meissner, Humphreys, Meis, & Scheu,1975); time and technology (Stafford & Duncan, 1985), shopping behavior(Hawes, Gronmo & Arndt, 1978; Wilson & Holman, 1984), travel (Jannelle& Goodchild, 1988, Pas & Harvey, 1997), and children's use of time(Medrich, Milos, Reizen, & Rubin, 1978; Timmer, Eccles, & OBrien, 1985).In developing countries, specific concerns include unpaid work (IN-STRAW, 1995), the sexual division of labor (Dixon-Mueller, 1985), house-hold production (Quizon, 1978; INSTRAW, 1996), nutrition (Berio, 1986),measurement of living standards (Acharya, 1982; Chernichovsky, Lucas, &Mueller, 1985), impact of technology (Carr & Sandhu, 1987), and the eco-nomic value of children (Cain, 1980; Minge-Kalman, 1977). Knowledge ofhow time is allocated is indispensable in attempts to understand all theforegoing issues.

Quality of Life

Growing concern with the quality of life has led to a search for valid, reliable, and economical quality measures or social indicators. Time use data provide the opportunity to develop a large number of indicators covering many life domains, such areas as health, education, working time, social interaction, leisure, and use of physical environment. Time use


studies have proven extremely useful in studying the impact of transitionstoward a market economy being experienced in Eastern Europe (Artemov,Rostovtsev, & Artemova, 1997). Minimally, they provide indicators of in-volvement in a broad or complete range of activities engaged in by mem-bers of subject groups. Thus, for example, indicators of involvement inmarket-oriented economic activity, housework and child care, education,and free time can be developed (Aas, 1982; Harvey, 1995). The collection oflocation of individuals, and who they were with, facilitates many moreindicators. Diary data can also provide opportunities to develop indicatorsof mobility, infrastructure use, and sociability, among others. Thus, manydifferent indicators, defined in terms of various life domains and forsignificant target groups, can be developed.

Behavioral indicators can be combined with subjective indicators toidentify possible connections and or explanations (Zuzanek, 1998). Forexample, social interaction data from the time use module, in conjunctionwith a subjective measure of happiness collected concurrently in the 1986Canadian General Social Survey, show a clear relation between time spentalone and the degree of happiness of respondents, particularly for theelderly. Elderly persons, indicating they were Very happy, averaged slightlyover 5 hours per day alone. Those who were Very unhappy averaged nearly9 hours alone. In contrast, for the youngest age group, the Very happy andthe Very unhappy recorded about equal amounts of time alone (Harvey, n.d.)

Leisure

The measurement of leisure has long been a fertile area of studyaddressed by time use research (Lundberg & Komarovsky, 1934; Robinson& Godbey, 1997; Zuzanek, 1980). Works based on the Multinational TimeUse Study highlighted the utility of the time-diary approach for studyingleisure (Ferge, 1972; Skorzynski, 1972), The work of Young and Willmott(1973) and Shaw (1986) shows both the utility and necessity of a time-diaryapproach. Their work raises some questions about previous analyses ofleisure based on time-diary data but indicates the need for time use data tomeasure and analyze it. They have shown that, contrary to the generallyaccepted approach of defining leisure in terms of selected " leisure-like"activities, virtually any activity may be perceived as a leisure activity for some specific person or group, or for a given individual under some conditions, but not under others. In short, subjectively, the concept of what leisure is varies from person to person and from time to time. Thus, it is inappropriate to limit the study of leisure to a circumscribed set of activ-ities. It is incumbent upon the researcher to capture all activity and analyze it in a manner that appropriately reflects the leisure reality.

Time use information juxtaposes leisure, work, and personal time in a


manner that provides considerably more information than do traditionalpencil measures. They enable researchers to fit leisure into the life patternboth quantitatively and qualitatively. Similarly, they make it possible to fitthe components of leisure time into aggregate leisure patterns. Based onthe General Social Survey time use module in 1986, Canadian men hadslightly more leisure (339 minutes a day)measured in terms of residualnonwork, nonpersonal activitiesthan Canadian women (318 minutes).The additional time appeared to accrue primarily from greater time allo-cated to viewing television155 minutes for men and 126 minutes forwomen (Harvey n.d.). In other respects, there was relatively little differ-ence between men and women in leisure time allocation. The only excep-tion being a tendency for women to allocate relatively more time to organi-zational activities (Harvey, n.d.).

The value of the time allocation data for leisure analysis was furtherfound in a simple analysis of time allocated to media using the 1986 timeuse data. A variety of variables (demographic, temporal, and activity)were used to segment time allocated to the media. It was found that allthree types of variables were automatically introduced into the analysis invery early stages. Work status (demographic) provided the first breakvariable. Employed persons (1) and students (3) were grouped togetherautomatically by the grouping algorithm, as were persons looking forwork and others, and persons keeping house, and persons not stating amain activity. Retired persons stood alone, averaging the greatest amountof leisure time, 331 minutes a day to media-related activity (Harvey, n.d.).In contrast, employed persons and students averaged only 146 minutes ofmedia time per day. At the second level, both time and activity variableswere introduced. Paid-work time was the most significant explanatoryvariable for the worker-student grouping. There was a strong inverserelationship between paid-work time and media time. This suggests atendency for work time to crowd out media time. For the group keepinghouse and not stated, an activity variable, number of trips, showed asimilar strong inverse relationship. Persons looking for work and others,with fewer constraints on leisure time, exhibited a trade-off between enter-tainment time and media time. The point here is not to carry out anelaborate analysis of time allocated to media. The purpose is to show thecomplexity of the phenomena and the flexibility for analysis offered by thetime use data for capturing behavior and change.

Travel Behavior Research

Time use data provide a valuable input into the study of travel behav-ior. In the mid-l970s, some travel behavior researchers recognized the need to develop an enriched understanding of day-to-day travel behavior.


This gave rise to the study of the "activity systems" approach, whichstrives to develop both a theoretical and empirical base for modeling dailybehavior (Chapin, 1974; Jones, Dix, Clarke, & Heggie, 1983). The approachbridged the gap between the time-space approach of Hagerstrand (1970)and the more familiar travel survey approach. The major difference be-tween the activity-based approach and the trip-based approach is in thetreatment of time (Pas & Harvey, 1997). The activity-based approach incor-porates the fullness of the time dimension, calling for a full accounting oftime use. Travel researchers need to understand the interplay betweenwhat is done, where, and with whom. The information needed can bereadily gleaned from time use studies (Harvey et al., 1997).

SUMMARY

Time use research has been used throughout this century to examineboth objective and subjective aspects of human behavior in a wide varietyof fields and purposes. In recent years, the methodology has becomeincreasingly sophisticated, both in terms of collection and analysis. Theintent of this volume is both to illustrate applications of the time usemethod and to provide guidelines for those wishing to incorporate it intotheir research.

REFERENCES

Aas, D. (1982). Designs for large scale, time use studies of the 24-hour day. In Z. Staikov, (Ed.),

Its about time: International Research Group on time budgets and social activities (pp. 1753).Sofia: Institute of Sociology at the Bulgarian Academy of Sciences, Bulgarian Sociological

Association.

Acharya, M. (1982). Time use data and the living standards measurement study: LSMS WorkingPapers, No. 18. Washington, DC: World Bank.

Andorka, R. (1987). Time budgets and their uses. Annual Review of Sociology, 13,149164.Artemov, V., Rostovtsev, P., & Artemova, O. (1997, October 810). Conditions and characteristics

in dynamic time use comparative analyses. Paper presented at the 19th International Asso-ciation for Time Use Research Conference, Stockholm, Sweden.

Avery, R. J., Bryant, W. K., Douthitt, R. A., & McCullough, J. (1996). Journal of Family and Economic Issues, 17(3/4), 409418.

Bailey, I. (1915). A study of management of farm homes. Journal of Home Economics, 7,3738.Baltes, M., Wahl, H., & Schmidt-Furstoss, U. (1990). The daily life of elderly Germans: Activity

patterns, personal control, and functional health. Journal of Gerontology, 45(4), 173179. Berio, A. J. (1986). The use of time allocation data in developing countries: From influencing

development policies to estimating energy requirements. Time Use Studies: Dimensions and Applications, 128, 3658.

Bevans, G. E. (1913). How working men spend their spare time. New York Columbia University Press.


Block, R. A. (1990). Models of psychological time. In R. A. Block (Ed.), Cognitive models ofpsychological time (pp. 135). Hillsdale, NJ: Erlbaum.

Cain, M. T. (1980). The economic activities of children in a village in Bangladesh. In H. P.

Binswanger (Ed.), Rural household studies in Asia (pp. 218247). Singapore: SingaporeUniversity Press.

Carr, M., & Sandhu, R. (1987). Women, technology and rural productivity. UNIFEM Occa-

sional Paper, No. 6, pp. 166. New York United Nations Development Fund for Women.

Chapin, F. S. (1974). Human activity patterns in the city: Things people do in time and in space.New York: Wiley.

Chernichivsky, D., Lucas, R. E. B., & Mueller, E. (1985). The household economy of rural

Botswana: An African case. In World Staff Working Papers (No. 7l5, pp. 1227). Washing-ton, DC: World Bank.

Converse, P. E. (1968). Time budgets. In International encyclopedia of the social sciences (pp. 4247). Crowell Collier & Macmillan.

Das, T. K. (1991). Time: The hidden dimension in strategic planning. Long Range Planning,

Dixon-Mueller, R. (1985). Division of labour by task. In Women's work in Third World agricul-ture: Concepts and indicators (pp. 151). Geneva, Switzerland: International Labour Office.

Douthitt, R. (1993). The inclusion of time availability in Canadian poverty measures. In

Instituto Nazionale di Statistica (ISTAT) (Ed.), Time use methodology: Toward consensus (pp.8892). Rome: ISTAT.

Eichler, M. (1983). Changing patterns in household management. In families in Canada today(pp. 140167). Toronto: Gage.

Elchardus, M. & Glorieux, I. (1993). Towards a semantic taxonomy: Classifying activities on

the basis of their meaning. In Instituto Nazionale di Statistica (ISTAT) (Ed.), Time usemethodology: Toward consensus (pp. 250276). Rome: ISTAT.

Elchardus, M. & Glorieux, I. (1994). The search for the invisible 8 hours: The gendered use of

time in a society with a high labour force participation of women. Time and Society, 3(1),

Elliott, D. H., Harvey, A. S., & Procos, D. (1976). An overview of the Halifax time-budget

study. Society and Leisure, 3, 145159.Ferge, S. (1972). Social differentiation in leisure activity choices: An unfinished experiment. In

A. Szalai (Ed.), The use of time: Daily activities of urban and suburban populations in twelvecountries (pp. 213227). The Hague: Mouton.

24(3), 4957.

527.

Frankenberg, R. (1992). Time, health and medicine. London: Sage.Frederick, J. A. (1995). As time goes by . . . time use of Canadians: General social survey. Ottawa:

Statistics Canada; Housing, Family and Social Statistics Division.

Garhammer, M. (1995). Changes in working hours in Germany: The resulting impact on

everyday life. Time and Society, 4, 167203. Gershuny, J. (1983a). Changing use of time in the United Kingdom: 1937-1975,the self-service

era. Studies of Broadcasting, 19, 7191.Gershuny, J. (1983b). Time budget research in the UK. Sussex: Science Policy Research Unit,

University of Sussex.

Gershuny, J. & Robinson, J. (1988). Historical changes in the household division of labour.

Demography, 25(4), 537552. Goldschmidt-Clermont, L. (1987). Economic evaluations of unpaid household work: Africa,

Asia, Latin America and Oceania. In Women, Work and Development (Vol. 14), Geneva:International Labour Organization.

Gonzalez, S., & Gomez, J. C. R. (Eds.). (1985). El uso del tiempo en Bogot. Bogota, Colombia: Anif.

Grossin, W. (1993a). Le temps de la vie quotidienne. Paris: Mouton.


Grossin, W. (1993b). Technology evolution, working time and remuneration. Time and Society,

Gutenschwager, G. A. (1973). The time budget: Activity systems perspective in urban re-

Hagerstrand, T. (1970). What about people in regional science? Papers and Proceedings of the

Harvey, A. S. (n.d.). Time use module of the General Social Survey (Working Paper No. 8, p. 33).

Harvey, A. S. (1995). Canadian time-use in a cross-national perspective. Statistics in Transition,

Harvey, A. S., & Clark, S. (1976). The sexual division of labour: The use of time. Atlantis, 2(1),

Harvey, A. S., & Elliott, D. H. (1983). Time, and time again. In M. C. Casserly (Ed.) & B. L. Kinsley (Vol. Ed.), Explorations in time use series (Vol. 4). Ottawa: Employment and Immigration Commission.

Harvey, A. S., Elliott, D. H., & MacDonald, S. W. (1984). Activities and settings. In M. C. Casserly (Ed.) & B. L. Kinsley (Vol. Ed.), Explorations in time use series (Vol. 6), p. 88.Ottawa: Canada Employment and Immigration Commission.

Harvey, A. S., & MacDonald, W. S. (1976). Time diaries and time data for extension of

economic accounts. Social lndicators Research, 3, 2125. Harvey, A. S., Marshall, K., & Frederick, J. A. (1991). Where does time go? Ottawa: Statistics

Canada; Housing, Family and Social Statistics Division.

Harvey, A. S., & Singleton, S. (1989). Canadian activity patterns across the life span: A time

budget perspective. Canadian Journal on Aging, 8(3), 268285. Harvey, A. S., Taylor, M. E., Ellis, S., & Aas, D. (1997). Final Report 24-Hour Society and passenger

travel. Rotterdam, the Netherlands: Transport Research Centre, Ministry of Transport, Public Works and Water Management.

Hasselkus, B. R. (1989). The meaning of daily activity in family caregiving for the elderly.

American Journal of Occupational Therapy, 43(10), 649656. Hawes, D. K., Gronmo, S., & Arndt, J. (1978). Shopping time and leisure time: Some prelimi-

nary cross-cultural comparisons of time budget expenditures. In H. K. Hunt (Ed.),Advances in consumer research (Vol. 52, pp. 151159). Symposium conducted at the 8th Annual Conference of the Association for Consumer Research. Ann Arbor, MI: Associa-

tion for Consumer Research.

Jannelle, D. G., & Goodchild, M. E (1988). Spacetime diaries and travel characteristics for

different levels of respondent aggregation. Environment and Planning Annual, 20,891-906.Japan Ministry of Economic Planning. (1975). An analysis of structure of living time: The pattern

of use of time and the quality of life. Tokyo: Author. (in Japanese) Jones, P. M., Dix, M. C., Clarke, M. I., & Heggie, I. G. (1983). Understanding travel behviour.

London: Gower.

Juster, E T. (1985). A note on recent changes in time use. In E T. Juster and E P. Stafford (Eds.),

Time, goods, and well-being (pp. 313332). Michigan: Institute for Social Research. Juster, E T., & Stafford, E P. (Eds.). (1985). Time, goods and well-being. Ann Arbor Survey

Research Center, Institute for Social Research, University of Michigan.

Juster, E T., & Stafford, E P. (1991). Comment. Public Opinion Quarterly, 55, 357359. Kinsley, B., & O'Donnell, T. (1983). Marking time. Ottawa: Canada Employment and Immigra-

Kneeland, H. (1929). Women's economic contribution in the home. Annals of the American

Knights, D., & Odih, P. (1995). It's about time!: The significance of genered time for financial

2,157167.

search and planning. Journal of the American Institute of Planners, 39(6), 378387.

Regional Science Association, 24, 721.

Ottawa: Statistics Canada.

2(4), 595610.

4666.

tion Commission.

Academy of Political and Social Science, 243, 3340.

services consumption. Time and Society, 4(2), 205231.


Larson, K. B. (1990). Activity patterns and life changes in people with depression. AmericanJournal of Occupational Therapy, 44(10), 902-906.

Lawton, M. P., Moss, M., & Fulcomer, M. (1987). Objective and subjective uses of time by older

people. International Aging and Human Development, 24(3), 171188.Leccardi, C., & Rampazi, M. (1993). Past and future in young womens experience of time.

Time and Society, 2(3), 353379. Lundberg, G. A., & Komarovsky M. (1934). Leisure: A suburban study. New York: Columbia

University Press. McGrath, J. E., & Rotchford, N. L. (1983). Time and behaviour in organization. Research in

Organizational Behaviour, 5, 57101.McKinnon, A. L. (1992). Time use for self-care, productivity, and leisure among elderly

Canadians. Canadian Journal of Occupational Therapy, 59(2), 102110. Medrich, E. A., Milos, M,, Reizen, B., & Rubin, V. (1982). The serious business of growing up: A

study of childrens lives outside of school (Vol. 1). Berkeley: University of California Press. Meissner, M. (1971). The long arm of the job: A study of work and leisure. Industrial Relations,

Meissner, M., Humphreys, E. W., Meis, S. M., & Scheu, W. J. (1975). No exit for wives: Sexual

division of labour and the cumulation of household demands. Canadian Review of Soci-

Michelson, V., & Ziegler, S. (1982). Childcare under constraint. In Z. Staikov (Ed.), Its about time (pp. 218226). Sofia: Bulgarian Sociological Association, Institute of Sociology at the Bulgarian Institute of Sciences. Symposium conducted at the meeting Proceedings of the

International Research Group on Time Budgets and Social Activities.

Minge-Kalman, W. (1977). On the theory and measurement of domestic labour intensity.

American Ethnology, 4(2), 273284. Moss, M. S., & Lawton, M. P. (1982). Time budgets of older people: A window on four

lifestyles. Journal of Gerontology, 37(1), 115123. Nakanishi, N., & Suzuki, Y. (1986). Japanese time use in 1985. Tokyo, Japan: NHK Public

Opinion Research Division.

Niemi, I. (1983). Systematic bias in hours worked? Statistical Review, 4, 326330. Niemi, I. (1990; July 9-13).Systematic error in behavioural measurement: Comparing results from

interview and time budget studies. Paper presented to meetings of the International Asso-ciation for Time Use Research, Madrid, Spain.

Pas, E. I., & Harvey, A. S. (1997). Time use research and travel demand analysis and model-

ling. In P. Stopher & M. Lee-Gosselin (Eds.), Understanding travel behaviour in an era of change (pp. 315338). New York: Pergamon.

10, 239260.

ology and Anthropology, 12(4, pt.1), 424439.

Pember-Reeves, M. (1913). Round about a pound a week. London: Bell. Pentland, W., Harvey, A. S., & Walker, J. (1998). The relationships between time use and health

and well-being in men with spinal cord injury. Journal of Occupational Science, 5(1), 1425. Quizon, E. K. (1978). Time allocation and home production in rural Philippine households.

Philippine Economic Journal, 36(1/2), 185201. Robinson, J. P., & Godbey, G. (1997). Time for life: The surprising ways Americans use their time.

State College: The Pennsylvania State University Press.

Rokkan, S. (1966). Comparative cross-national research: The context of current efforts. In R. L.

Merritt, & S. Rokkan (Eds.), Comparing nations. New Haven, CT: Yale University Press. Rosenthal, L., & Howe, M. (1984). Activity patterns and leisure concepts: A comparison of

temporal adaptation among day versus night shift workers. Occupational Therapy in Mental Health, 4(2), 5978.

Ross, M. M. (1990). Time-use in later life. Journal of Advanced Nursing, 15, 394399.Shaw, S. M. (1986). Leisure, recreation or free time: Measuring time usage. Journal of Leisure

Research,18(3),177189.


Skorzynski, Z. (1972). The use of free time in Torun, Maribor and Jackson. In A. Szalai (Ed.), The use of time: Daily activities of urban and suburban populations in twelve countries (pp. 266 289). The Hague, Paris: Mouton.

Sorokin, P., & Berger, C. (1939). Time-budgets of human behaviour. London: Harvard University Press.

Stafford, E P., & Duncan, G. J. (1976). Market hours, real hours and labor productivity.

Economic Outlook, 103119.Stafford, E P., & Duncan, G. J. (1985). The use of time and technology by households in the

United States. In E T. Juster & E P. Stafford (Eds.), Time, goods and well-being (pp. 245288). Ann Arbor: University of Michigan Survey Research Center, Institute for Social

Research.

Stanley, M. (1995). An investigation into the relationship between engagement in valued

occupations and life satisfaction for elderly South Australians. Journal of Occupat ional Science: Australia, 2(3), 100114.

Statistics Canada. (1974, August 1923). Time Use Surveys----- Summary Paper. Prepared for dis -tribution at the 8th World Congress of the International Sociological Association, Toronto.

Stone, P. J. (1972). Childcare in twelve countries. A. Szalai (Ed.), The use of time: Daily activities ofurban and suburban populations in twelve countries (pp. 249264). The Hague: Mouton.

Tanaka, Y. (1978, July). Time budgets and social activities in Japan. Paper presented at meetings of the Working Group on Time Budgets and Social Activities, International Sociological

Association World Congress, Upsalla, Sweden.

Timmer, S. G., Eccles, J., & O'Brien, K. (1985). How children use time. In E J. Juster & E P.

Stafford (Eds.), Time, goods, and well-being (pp. 353382). Michigan: Survey Research Center, Institute for Social Research, University of Michigan.

Ujimoto, K. V. (1985). The allocation of time to social and leisure activities as social indicators

for the integration of aged ethnic minorities. Social Indicators Research, 17, 253-266.United Nations International Research and Training Institute for Women (INSTRAW). (1995).

Measurementand valuation ofunpaid contribution:Accounting through timeand output. SantoDomingo, Dominican Republic: Author.

United Nations International Research and Training Institute for Women (INSTRAW). (1996).

Valuation of household production and the satellite accounts. Santo Domingo, Dominican Republic: Author.

Urdaneta-Ferran, L. (1986). Measuring and valuing women's participation in the informal sector of the economy. Conference of European Statisticians: Informal Meeting on Statistics, Ge-neva, Switzerland.

Usunier, J. G., & Valette-Florence, P. (1994). Perceptual time patterns ("timestyles"): A psycho-

metric scale. Time and Society , 3(2), 219-241.Vanek, J. (1974, November). Time spent in housework. Scientific American, pp. 116120. Walker, K. E., & Woods, M. E. (1976). lime use: A measure of household production of family

goods and services. Washington, DC: Centre for the Family of the American Home

Economics Association.

Wilson, R. D., & Holman, R. H. (1984). lime allocation dimensions of shopping behaviour.

Advances in Consumer Research, 11, 2934. Yano, M. (Ed.). (1995). Sociology of time budgets: Social time, personal time. Tokyo: University of

Tokyo Press.

Young, M., & Willmott, P. (1973). The symmetrical family: A study of workand leisure in the Londonregion. Harmondsworth, Middlesex, UK: Penguin Books.

Zuzanek, J. (1980). Work and leisure in the Soviet Union: A time-budget analysis. New York: Praeger.

Zuzanek, J. (1998). Tie use, time pressure, personal stress, mental health, and life satisfac-

tion from a life cycle perspective. Journal of Occupational Science, 5(1), 2639.

2Guidelines for Time Use DataCollection and Analysis

Andrew S. Harvey

INTRODUCTION

Time diaries provide an ideal approach to the collection of activity data.Activity data collected by means of stylized questions or activity lists,taken out of the context of daily life, miss many of the objective andsubjective circumstances about participation in activities. Yet often theseare the circumstances that, with personal characteristics, determine actualbehavior. A time diary places activities in their natural temporal context.By its nature, the diary provides a record of all activities during a specifiedperiod (day, week), along with a potentially rich array of contextual infor-mation. This chapter explores the collection and analysis of diary data andspecific opportunities and problems they pose for the researcher.

As indicated in Chapter 1, even the simplest time use studies providecrucial measures of involvement in a broad range of activities engaged inby individualssuch as paid work, housework and child care, education,sleep, eating, socializing, games, sports, media use. If supplementary dataare collected about the location of activities, and whom individuals arewith, many more measures can be generated. These additional data pro-

Andrew S. Harvey Department of Economics, St. Mary's University, Halifax, NovaScotia, Canada B3H 3C3.

Time Use Research in the Social Sciences, edited by Wendy E. Pentland, Andrew S. Harvey, M.Powell Lawton, and Mary Ann McColl. Kluwer Academic/Plenum Publishers, New York,

1999.19

20 ANDREW S. HARVEY

vide an opportunity to develop measures of mobility, infrastructure use,sociability, and other diverse social phenomena. If subjective informationhas also been collected, construction of affective measures of the quality oflife are also possible. Many different, rich measures of the texture ofeveryday life can be developed.

COLLECTION GUIDELINES

In many ways, the collection of time use data differs little from thecollection of other social and economic data. There are, however, a numberof issues that should be addressed to optimize the value and accuracy ofthe final data. While the diary is the preferred data collection method.there are alternatives. Activity lists, logs, continuous or random observa-tion, and beeper studies have all been used at one time or another to collectactivity data (United Nations International Research and Training Institutefor the Advancement for Women [INSTRAW], 1995). The actual approachchosen will depend on a number of issues that can be evaluated in terms ofboth input and output criteria (Harvey & MacDonald, 1976). The sug-gested input criteria are respondent knowledge, respondent cooperation,time and money resources, and processability. Output criteria are validity,reliability, usability, and flexibility. Once an activity capture approach hasbeen chosen, questions regarding data collection remain. Collection meth-odology issues can be classified in terms of sampling, collection, diarycontent, and background variable content (Harvey, 1993b).

Sampling of Respondents

Sampling issues relate to the choice of the respondent population, thesample size, geography, and survey timing. Typically, national statisticalagencies collect data that are nationally representative. The major issuesstatistical agencies face in terms of population are whether to collectdiaries for individuals only or for several or all household members.Additionally, they must set the ages of the respondent population. There isno clear choice. The ages of populations covered have ranged from age 2years in Bulgaria to age 15 in Canada. The current Statistical Office of theEuropean Communities (EUROSTAT) project guidelines are to collect di-aries for all household members aged 10 years and over (EUROSTAT,1996).

Many time use studies have been carried out for particular sub-populations of substantive value to the research design. Michelson (1988) collected data on complete families in Toronto to study the effects of

GUIDELINES FOR TIME USE DATA COLLECTION AND ANALYSIS

maternal employment. Several of the editors of this volume are collecting data for a sample of individuals with spinal-cord injuries (McCall, Pent-land, Harvey, Walker, & Comis, 1993). At a minimum, the sample must be chosen in such a manner as to provide unbiased data for the population it purports to represent.

Sample size must be considered in terms of coverage of both popula-tion and diary days, since both population and behavior are being sam-pled. Consequently, the amount of data collected on particular behavior (e.g., meal preparation) is a function of both how many persons do it and how frequently. The sampling will be particularly affected by the nature of the issues motivating the survey. If one is interested in particular behavior, it is important that the sampling take both the propensity for doing the targeted activity and its frequency of occurrence into account. Eating, sleeping, and television viewing are not a problem, since they are done virtually daily. Sewing and mending, use of services (bank, doctor, etc.) and concert going are done by sufficiently few individuals and with sufficient infrequency that either extremely large, or extremely focused samples, would be required to provide useful analytical data. The geogra-phy of the sample will depend to a great extent on the purpose of the study.Gershuny (1991) suggests that time use estimates are somewhat insensitive to gross locational differences. This is understandable, since measured behavior is a function of the role and context of an individual (Harvey, 1983). If the geographic area is sufficiently large to be representative of a broad range of individual and microareal differences, the time use esti-mates should be relatively stable.

The final sampling issue relates to the time of year for data collection. Practice has varied, ranging from drawing a full sample in only 36 days (Nippon Hoso Kyokai [NHK], 1995) to sampling for over a full year (Niemi, Pkknen, Rajaniemi, Laaksonen, & Lauri, 1991; Statistics Can-ada, 1995). The choice of period is not just of academic concern. To the extent that behavior varies by time of the week, month, or season, it is necessary to ensure that the survey period appropriately reflects the gen-eral or particular behavior of interest. Niemi (1983) showed that time use during OctoberNovember, typically used for short-duration studies, was close to the annual average. Other work, however, found time-of-year did lead to substantial variation in the data (Hill, 1985). One must be sensitive to the interaction of population and season. If the sample includes young children, choice of a school period as representative of an annual average can be misleading both with respect to education and to the behavior of child caregivers. The results may well overestimate time allocated to edu-cation by the students and underestimate parental time spent caring for children.

21

22 ANDREW S. HARVEY

Diary Design

There are a number of diary design issues involving the interviewmode, the focus on "yesterday" versus "tomorrow" diaries, and the choiceof day. Two major options are the choice of precoded versus open-responsecategories and the choice of fixed time versus open intervals.

Open versus Coded Category

Most diary researchers shun the precoded format, opting instead foran open-response diary in which individuals respond in their own words.Precoding, usually limited to relatively few codes, forces excessive and ir-reversible data reduction at too early a stage in the survey process. How-ever, the extreme data reduction accompanying precoding is not absolute.An ongoing study in the Netherlands (Knulst & Schoonderwoerd, 1983)has used a broadly based precoding scheme incorporating a large numberof codes, which is somewhat more flexible later in the process.

Closed versus Open Interval

The option relates to the closed versus open time intervals. The Multi-national Time Use Study used an open-interval approach (Szalai, 1972),meaning that the respondent reports starting and ending times of eachactivity as part of the diary entry. This approach has been followed by allthe major North American studies mentioned in Chapter 1. However, mostof the European national surveys have opted for fixed-interval diaries,with intervals rangingfrom5 to20 minutes. Workof Lingsom (1979) andofNiemi (1983) suggested little difference between the two approaches. Un-published pilot testing for the 1986 Canadian Time Use Study concludedthat there were no cost savings from fixed time slots. Some work, however,suggests that there may be hidden problems. There is evidence that the useof, and size of, time slots differentially affect various activities (HarveyElliott, & Stone, 1977). The cooperative European time use survey beingfacilitated by EUROSTAT has adopted a 10-minute fixed-interval diary.

Yesterday versus Tomorrow Basis

Time-diary data can be collected on either a yesterday or tomorrowbasis. Yesterday diaries are typically collected by personal or phone inter-view, while tomorrow diaries are left behind by interviewers ("leave-behind diaries") or mailed to respondents. Although tomorrow diariesyield more events, research suggests that the difference in the number of events (an increase on the order of 10%) does not justify the additional cost

GUIDELINES FOR TIME USE DATA COLLECTION AND ANALYSIS 23

of obtaining tomorrow diaries (Juster, 1985b; Robinson, 1977,1985; Szalai,1972). Can diaries be collected for days further back than yesterday? Thereis no clear answer to this question. Research on it (Juster, 1985b; Keller,Kempter, Timmer, & Young-Demarco, 1982; Klevemarken, 1982) givesmixed results. Juster (1985b) argues that people appear to be able to recallFridays through Sundays better than other days. The general view ofexperienced time-diary researchers is, however, that recall should not beattempted for more than 2 days in arrears.

Number of Days

There are choices in the number of days to capture per respondent.While many diary studies collect only 1 day per respondent, it has becomemore common to collect at least 2 days per respondent. It is argued that atleast 2 days provide for greater reliability (Kalton, 1985; Pas, 1986; Sanik,1983). Kalton, however, argues for 2 weekdays, leaving aside the issue oftwo Saturdays or Sundays (Kalton, 1985). The EUROSTAT pilot surveydesign calls for collecting two diary days per respondent, one weekdayand a Saturday or Sunday (EUROSTAT, 1996).

Random versus Convenient Days

The actual days may be designated by random selection or chosen atthe convenience of the interviewer or respondent. While Kinsley andO'Donnell (1983) found no strong argument for either approach, they didfind that designated-day diaries were more likely to contain time spent athome. Juster (1985b) believes that the designated-day approach will en-hance representativeness. Although administrative and cost considera-tions may favor the convenience approach, it is preferable to use adesignated-day approach in order to reliably capture the several dimen-sions of behavior. To reduce sample loss if a respondent is unavailable onthe designated day, the diary day may be set for the same day, one or twoweeks later. Lyberg (1989), following tests with Swedish data, suggestedthat there was little difference in diaries collected " on time" and those"delayed" to the same day the following week or two.

Personal versus Telephone Interview

There are several ways in which the diaries may be administered, including a personal interview, phone interview, drop-off and pick-up, or drop-off and mail-back of time-diary protocols. Research suggests that there is little difference between a yesterday diary completed over the

24 ANDREW S. HARVEY

phone and one completed by personal interview (Kinsley & O'Donnell,1983; Klevemarken, 1982). There is, however, no reported research I knowof that can provide guidance on the choice between drop-off /pick-up anddrop-off /mail-back. However, without considerable follow-up, it wouldappear that the drop-off/mail-back approach is subject to much greaternonresponse, and experience with drop-off /pick-up diaries indicates thata review at the time of pick-up usually leads to revisions and additions tothe diary.

Diary Content

Diary content is driven by three factors: the need for relevant informa-tion in line with the objectives of the study, the need for validity andreliability, and concern for respondent burden. Typically, researchers areinterested in a variety of dimensions of each activity. The vast majority ofnational time-diary surveys collect or report information on what is beingdone (primary activity), what else is being done (secondary activity),where it is being done (location), and with whom it is being done (socialcontact). Collecting such information is important not only for the data,but also because it can add to the validity and reliability of the activitydata. Recalling changes in the several dimensions as one reports the un-folding day serves as a memory jog for other dimensions and adds rela-tively little time to the interview process. Other objective information hasalso been sought. For example, studies focusing on household productionhave sought information on appliances used; other studies have soughtinformation on smokers present (Robinson, Ott, & Switzer, 1996).

Subjective Dimensions

Several researchers have shown the efficacy of, and argue for, thecollection of subjective data (Clark, Harvey & Shaw, 1990; Cullen, Godson,& Major, 1972; Michelson, 1986; Robinson, 1983). The subjective data can beused both to define activities and provide perceptions of activities. Forexample, respondents have provided their own information on whichactivities they view as work and leisure (Shaw, 1986). Alternatively, thesubjective data may be used to measure the respondents feelings aboutactivities ouster, 1985a; Robinson, 1983,1984b). Subjective dimensions ex-plored include satisfaction (Robinson, l977,1983,1984b), liking (Moss &Lawton, 1982), tension (Michelson, 1988), and material benefit from activity(Harvey, 1993a). Recently, attention has turned to gathering motivationalinformation related to "for whom" activities are being done (Blnke, 1994).Such information is being sought in the EUROSTAT pilot survey. This

GUIDELINES FOR TIME USE DATA COLLECTION AND ANALYSIS 25

information takes on primary relevance in studies focused on upgradingeconomic accounts or in studies on volunteer activity.

Background Data

Interpretation of time-diary data is highly dependent on the nature ofthe attendant background data. An individual's role or sociodemographiccircumstances is of central importance in determining time use. The im-portance of sociodemographic characteristics was noted in reporting onthe Multinational Time-Use Study where it was found that individualsoccupying roles defined in terms of sex and employment were more alikeacross sites than they were like individuals occupying other, different basicroles in their own site (Converse, 1972). Aas (1982) argues for the impor-tance of role in the household (child, spouse, parent, other). If diaries andattendant background information are not collected from all members ofthe household, it is important that, at least, employment status of thespouse be obtained, since it can significantly affect the household divisionof labor and other time use as well. Additionally data on socioeconomicstatus, income, life-cycle state, age, education, number and ages of chil-dren, number of other household members, and employment status andurbanization level of household community should also be collected(Harvey,1993).

DATA-FILE EDITING AND CREATION

One of the most challenging aspects of time-diary data analysis is the preparation and organization of the diary data. It is this process, more than any other, that separates the collection and analysis of time-diary data from similar processes in traditional social surveys. At the heart of the editing and coding of the diary data is the coding scheme used to record the reported behavior. There is no standard activity coding scheme. The multinational study established a de facto standard (Szalai, 1972). Most national studies have maintained some comparability to the multinational coding scheme. There are, however, a number of problems with it (Harvey, 1996b). A coding scheme addressing some of these problems was recently proposed (Harvey & Niemi, 1994). The current EUROSTAT time use pro-ject may well establish a new referent.

As with any survey, once completed, the forms need to be edited for accuracy and completeness. The major difference in a diary survey is in the editing and checking of the diary form. It is not sufficient to simply browse for nonresponse to items that should be completed. The diary form itself

26 ANDREW S. HARVEY

must be checked for consistency and completeness by following it throughthe day to ensure that there are no time gaps, that all activities and theirseveral dimensions have been reported, and that several competing activ-ities (7:007:20 A.M., ate breakfast and took daughter to school) have notbeen recorded in one time slot. Often, in the editing process, it will bepossible for the editor to make corrections from the information provided.However, since there may be a need to recontact the respondent and con-firm information, this process should be done immediately followingcompletion of the diary to ensure accurate recall on the part of the respondent.

File Creation

Processing and analysis of the diary data can be facilitated with the construction of three different files: a respondent summa ry file, an episode file equivalent to the activity file (Chapter 3), and a time-points file. Once the questionnaire and diary have been edited and the data entered, it is useful to construct at least two data files, one containing respondent-level infor-mation, and one containing episode-level data (Fraire, 1993; Harvey, 1984). A third file, a time-points file, is also useful for further analyzing episodes and the temporal location of activities (Faire, 1993; Stone, 1984). Because the initial data extracted from diaries are typically time allocation by activity, it is necessary to summarize for each diary day the time allocated to all episodes of a given activity; that is, the total time spent eating at various times of the day must be consolidated into total daily time spent eating. Typically, such aggregations are preformed and a respondent sum-may file is created, with one variable for each activity code, which contains the number of minutes allocated to that activity during the day. These

wendy e. pentland - time use research in the social sciences

Documents