Handbook of the Psychology of Aging -

eBook PDF
Visit to download the full and correct content document:
https://ebooksecure.com/download/handbook-of-the-psychology-of-aging-ebook-pdf/
HANDBOOK OF THE PSYCHOLOGY OF AGING
NINTH EDITION
 THE HANDBOOKS OF
AGING
Consisting of Three Volumes

Critical Comprehensive Reviews of Research Knowledge,

Theories, Concepts, and Issues

Editors-in-Chief
Laura L. Carstensen and
Thomas A. Rando

Handbook of the Biology of Aging, 9th Edition

Edited by Nicolas Musi and Peter J. Hornsby

Handbook of the Psychology of Aging, 9th Edition

Edited by K. Warner Schaie and Sherry L. Willis

Handbook of Aging and the Social Sciences, 9th Edition

Edited by Kenneth F. Ferraro and Deborah Carr
HANDBOOK OF THE
PSYCHOLOGY OF
AGING
NINTH EDITION

Edited by

K. WARNER SCHAIE
Department of Psychiatry & Behavioral Sciences, University of Washington, Seattle,
WA, United States

SHERRY L. WILLIS
Department of Psychiatry & Behavioral Sciences, University of Washington, Seattle,
WA, United States

Associate Editors
BOB G. KNIGHT
BECCA R. LEVY
SUSAN M. RESNICK
Academic Press is an imprint of Elsevier
125 London Wall, London EC2Y 5AS, United Kingdom
525 B Street, Suite 1650, San Diego, CA 92101, United States
50 Hampshire Street, 5th Floor, Cambridge, MA 02139, United States
The Boulevard, Langford Lane, Kidlington, Oxford OX5 1GB, United Kingdom
Copyright © 2021 Elsevier Inc. All rights reserved.
No part of this publication may be reproduced or transmitted in any form or by any means, electronic or mechanical, including photocopying, recording,
or any information storage and retrieval system, without permission in writing from the publisher. Details on how to seek permission, further information
about the Publisher’s permissions policies and our arrangements with organizations such as the Copyright Clearance Center and the Copyright Licensing
Agency, can be found at our website: www.elsevier.com/permissions.
This book and the individual contributions contained in it are protected under copyright by the Publisher (other than as may be noted herein).

Notices
Knowledge and best practice in this field are constantly changing. As new research and experience broaden our understanding, changes in research meth-
ods, professional practices, or medical treatment may become necessary.
Practitioners and researchers must always rely on their own experience and knowledge in evaluating and using any information, methods, compounds, or
experiments described herein. In using such information or methods they should be mindful of their own safety and the safety of others, including parties
for whom they have a professional responsibility.
To the fullest extent of the law, neither the Publisher nor the authors, contributors, or editors, assume any liability for any injury and/or damage to persons
or property as a matter of products liability, negligence or otherwise, or from any use or operation of any methods, products, instructions, or ideas con-
tained in the material herein.

British Library Cataloguing-in-Publication Data

A catalogue record for this book is available from the British Library
Library of Congress Cataloging-in-Publication Data
A catalog record for this book is available from the Library of Congress
ISBN: 978-0-12-816094-7
For Information on all Academic Press publications
visit our website at https://www.elsevier.com/books-and-journals

Publisher: Nikki Levy

Editorial Project Manager: Barbara Makinster
Production Project Manager: Swapna Srinivasan
Cover Designer: Matthew Limbert
Typeset by MPS Limited, Chennai, India
 Contents

List of contributors xi Multi-timescale processes and research designs 24

About the editors xiii Measuring and modeling multiscale processes from
intensive time series 25
Foreword xv Dynamical systems and attractors 25
Preface xvii Dynamical systems models in aging 26
Challenges 28
I Planning for the long term
Privacy, linking, and data sharing
28
28
Concepts, theories, methods Incompleteness and attrition 29
Reliability versus sensitivity 30
Measurement invariance over time 30
1. History of adult cognitive aging research Emerging opportunities 30
Intrinsic capacity and functional ability 31
K. WARNER SCHAIE
Conclusion 31
References 31
Introduction 3
Adult stages of intellectual development 4
A stage model of adult cognitive development 4
Assessment of intellectual functions
A test of intelligence
6
6
II
The nature of intelligence 7 Bio-psychosocial factors in aging
Intelligence as multiple abilities 7
Practical or everyday intelligence 8
Intelligence and age 8 3. Functional connectivity in aging 37
Crystallized and fluid intelligence 9
FRANZISKUS LIEM, LINDA GEERLIGS, JESSICA S. DAMOISEAUX AND
Early cross-sectional studies 9 DANIEL S. MARGULIES
Longitudinal studies 10
Cohort differences 10 General introduction and outline of chapter 37
Magnitude of change 12 Functional connectivity 38
Frequency of decline 13 The default mode network 38
Personality and lifestyles 13 Functional connectivity in aging 38
Summary 14 Within-network connectivity 38
References 15 Between-network connectivity 40
Further reading 17 Whole-brain connectivity 41
Dynamics 42
2. Measurement and models for multi-timescale Cognition and functional connectivity 43
psychological processes in aging research 19 Functional connectivity in neurodegenerative disorders 44
STEVEN M. BOKER AND JOHN R. NESSELROADE
Functional connectivity as a biomarker 45
Functional connectivity as a general biomarker
Introduction 19 for brain health 45
Matching the research design to the question 20 Connectivity as an early marker for neurodegenerative
Modeling review 21 disorders 45
Cross-sectional models 21 Methodological considerations relevant to aging research 46
Autoregressive and crosslag models 22 Summary and future directions 47
Latent growth curves and multilevel models 22 Acknowledgments 47
The changing data landscape 24 References 47

v
vi Contents

4. Memory: behavior and neural basis 53 Consequences of pain in older adults 88

CINDY LUSTIG AND ZIYONG LIN
Pain management in older adults 88
Pain assessment 88
What is memory, and what is aging? 53 Pain management 90
Brain aging and memory: a complex and Summary 93
dynamic relationship 54 References 93
Control and association: major influences on age Further reading 98
differences in memory 55
The controlled processing paradox: important and impaired 55 7. Sleep, neurocognition, and aging, including
Associational memory may be especially impaired in aging 57 secular trends in older adult sleep 99
Default network dysregulation 57 CATHERINE A. MCCALL AND NATHANIEL F. WATSON
Modifying factors: qualitative and quantitative effects 58
Interventions: hope for improvement? 59 Introduction 99
Summary and conclusions 61 Sleep through the life span 100
References 62 Reductions in slow wave activity 101
Sleep spindle changes 101
5. Executive functions and neurocognitive aging 67 Alterations of sleep parameters 102
Circadian rhythm changes 102
PATRICIA A. REUTER-LORENZ, SARA B. FESTINI AND
Etiologies of sleep changes over the life span 103
TIFFANY K. JANTZ
Etiologies of sleep disturbances in older adults 103
Overview 67 Primary sleep disorders 104
Measuring executive functions 68 Medications, substances, sleep, and cognition 105
Executive deficit theories of cognitive aging 68 Neurocognitive disorders and sleep 107
Inhibitory deficit theory (Hasher & Zacks) 69 Benefits and challenges of treating sleep problems
Goal maintenance deficit (Braver & West) 69 in older adults 109
Production deficit hypothesis 69 Conclusions 110
Frontal lobe hypothesis of cognitive aging 70 References 110
A current perspective on executive deficit theories of
cognitive aging 70 8. The final challenge of aging: Death and dying 119
Do executive functions show the earliest and disproportionate MOLLY MAXFIELD, EVA-MARIA STELZER AND JEFF GREENBERG
decline? 70
Cross-sectional evidence for disproportionate executive Terror management theory 120
functions decline? 70 Implications of terror management theory for older adults
Longitudinal evidence for earlier executive functions approaching death 121
decline? 71 Aging and death-related anxiety 123
Do brain regions linked to executive functions show the The medical and lifestyle pragmatics of the dying process
earliest and disproportionate decline? 72 for older adults 125
Aging and the neural substrates of executive functions 72 Denial and avoidance in communicating
Relative degree and onset of prefrontal cortex decline 73 life-limiting diagnoses 127
Executive functions and prefrontal cortex processes as Preparation for end-of-life 128
compensatory and protective 74 The hospice and palliative care movement 129
Executive functions, the frontal lobes, and lifelong aging 75 Death with Dignity 130
References 76 Summary and future directions 130
References 131
6. Pain in older adults 83
ANN L. HORGAS AND AMANDA F. ELLIOTT
III
Overview of pain 83
Behavioral processes
Definitions of pain 83
Theories of pain 84
Pain and aging 85 9. Smartphones, robots, and social media:
Pain in older adults 85 aging with communication technologies 139
Epidemiology of pain in older adults 85 CHENG CHEN, MICHAEL L. KRIEGER AND S. SHYAM SUNDAR
Cross-national estimates of chronic pain 86
Age differences in pain 86 Introduction 139
Factors that contribute to pain in older adults 87 Aging with communication technology 140
 Contents vii

Smartphones 140 12. Financial decision-making and capacity in

Social media 142 older adults 193
Video gaming 143
DANIEL C. MARSON, DEBORAH L. KERR AND DONALD G. MCLAREN
Virtual reality 145
Smart home technology 146 Introduction: what is financial capacity? Legal, clinical and
Robots 147 ethical perspectives 194
Concluding remarks 149 The financial capacity problem: cognitive aging and
References 150 disorders of aging 195
Clinical warning signs of diminished financial capacity 196
Challenges in modeling financial capacity 197
10. Self-perceptions and awareness of aging: Clinical model of financial capacity 198
past, present, and future 155 Approaches to assessing financial capacity 198
MANFRED DIEHL, ALLYSON F. BROTHERS AND HANS-WERNER WAHL Assessing premorbid financial capacity 200
Self and informant-based assessment 200
Introduction 155 Performance-based assessment 200
A conceptual framework for subjective aging research 156 Clinical interview assessment of financial capacity 201
Subjective age: strengths and limitations 157 Empirical studies of financial capacity 202
Self-perceptions and awareness of aging: the past 158 Financial capacity in patients with mild and moderate
Self-perceptions and awareness of aging: the present 158 Alzheimer’s disease 202
Self-perceptions of aging: conceptual and measurement Financial capacity in patients with mild cognitive impairment 203
considerations 159 Neuropsychological predictors of financial capacity 204
Awareness of age-related change: conceptual and Neuroimaging studies of financial capacity 205
measurement considerations 159 Neuroimaging approaches to studying financial capacity 205
Empirical findings on the role of self-perceptions and Neuroimaging studies of financial capacity in a mild cognitive
awareness of aging 161 impairment and Alzheimer’s disease 205
Secular trends/cohort differences in self-perceptions and Neuroimaging studies of financial decision making in
awareness of aging 167 cognitively normal older adults 206
Self-perceptions and awareness of aging: the future 169 Summary 206
Focus on pathways and mechanisms 169 Noncognitive contributions to financial capacity in aging 206
Elaborating developmental antecedents, correlates, Psychological and psychiatric contributions to financial
and trajectories 170 capacity 207
Examining the modifiability of self-perceptions of aging 171 Relationship of physical dependence/medical frailty to
Changing the societal context of self-perceptions of aging 172 financial capacity in aging 208
Conclusion 173 Noncognitive contributions to financial capacity in aging 208
Acknowledgment 174 Cultural and social contributions to financial capacity
References 174 in aging 208
Resilience and financial capacity in aging 209
Future research directions 210
11. Gender diversity in later life 181 Detecting very early financial declines in cognitively normal
elderly 210
JENNIFER L. O’ BRIEN AND SUSAN KRAUSS WHITBOURNE
Study of financial capacity in other older adult clinical
Introduction 181 populations 210
Gender identity 181 Study of noncognitive contributions to financial capacity in
Terms 182 aging 210
Redefining the gender binary 182 Evolution of financial capacity in a technological society 210
Gender identity throughout the life span 183 References 211
Research findings 183
Gender affirmation 184
Aging and gender identity 185 13. The older worker: gender and age discrimination
Unique challenges faced by gender diverse elders 186 in the workplace 215
Improving outcomes for gender diverse elders 187 STEPHANE P. FRANCIOLI AND MICHAEL S. NORTH
Focus on resilience 188
Policies and steps to promote inclusivity in healthcare practice 188 The what, where, and when of discrimination against older
Affirmative care 189 workers 216
Conclusion 190 Hiring discrimination 217
References 190 On-the-job discrimination 218
viii Contents

Discrimination at job separation 220 Cultural context 267

Why does discrimination against older workers occur? 220 Family factors 268
From stereotypes of, to discrimination toward older workers 221 Individual factors 269
The powerful prescriptive effects of age norms 223 Future directions and conclusions 270
At the intersection of age and gender: the case of female Implications of a life-span model of family care exchanges 270
older workers 225 Clinical and policy implications 270
Do older female workers have it worse? 225 References 271
Gendered ageism and “Lookism” 226
Conclusion and opportunities for future research 227
References 228 17. Stress, coping, and aging 275
CAROLYN M. ALDWIN, LORIENA YANCURA AND HYUNYUP LEE

14. Remembering the personal past across Stress 276

adulthood 237 Foundational concepts in the study of stress 276
NICOLE ALEA, SUSAN BLUCK AND SHUBAM SHARMA Trauma 276
Life events 277
Cognitive psychology tradition: ecological, everyday memory 238 Hassles 277
The personal past: empirical research in cognitive psychology 239 Chronic role strain 277
Gerontological tradition: reminiscence and life review 240 Does stress change with age? 278
The personal past: empirical research in gerontology 240 Coping 278
Personality psychology tradition: autobiographical Foundational concepts in the study of coping 278
reasoning and the life story 241 Changes in coping with age 279
The personal past: empirical research in personality Theoretical models of stress, coping, and aging 280
psychology 241 Age-related changes in motivation 280
Bridging traditions 242 Increase in coping efficiency 281
Functions of autobiographical memory 243 Summary 281
Conclusion 243 References 282
References 244 Further reading 286

IV 18. Resilience in midlife and aging 287

KAARIN J. ANSTEY AND ROGER A. DIXON
Complex processes
Introduction 287
Resilience and the life course 288
15. Psychological assessment of neurocognitive Concepts and definitions: distinguishing resilience from
disorders 251 related concepts 288
BRIAN P. YOCHIM AND BENJAMIN T. MAST
Measurement of resilience 291
Scale approaches to measuring resilience 292
Introduction 251 Trajectory approaches to defining and detecting resilience 292
Neurocognitive disorders 252 Epidemiological contributions 293
Assessment of neurocognitive disorders 252 Resilience in midlife 294
Cultural considerations in the assessment of late life Domain specific resilience in aging 294
neurocognitive disorders 253 Comparison of cognitive resilience to related concepts 295
Establishing etiologies 254 Physical, psychological, and general resilience 296
Capacity assessment 255 Conclusions and future directions 296
Emergence of positive psychology and person-centered care 256 Acknowledgments 297
Nontraditional, unobtrusive measures of cognition 257 References 297
Conclusion 258
References 258
Further reading 262
19. Successful aging: an obscure but
obvious construct 301
RACHEL PRUCHNO
16. Family care exchanges across the life span 263
CYNTHIA A. BERG, CAITLIN S. KELLY AND REBECCA L. UTZ Semantics regarding successful aging 302
The bedrocks of successful aging 303
The life-span model of family care exchanges 264 Early science 303
Common risk and protective factors for family care exchanges 266 The perspectives collide 304
 Contents ix

Successful aging: a multidimensional concept or a fuzzy one? 305 Cohort and longitudinal findings from Seattle Longitudinal
Successful aging: a new paradigm 307 Study: Age-related change from 50 to 80 years in
Successful aging: a public health imperative 309 2 cohorts in old age (1886
1913; 1914
48) 326
Interventions promoting successful aging 310 Earlier born versus later born cohorts 327
Future directions 310 Cohort differences in rate of cognitive aging 327
Conclusions 312 Cohort differences in terminal cognitive decline:
References 312 Two cohorts in old age (1886
1913; 1914
48) 327
Assessing compression of morbidity: Time to death metric 329
Cohort differences 329
20. Cognitive functioning and dementia Recent changes in prevalence and incidence of dementia 330
prevalence in baby boomers and adjacent cohorts: Age specific versus cohort analyses of dementia prevalence 330
Longitudinal and cohort effects 319 Possible explanations 331
SHERRY L. WILLIS AND K. WARNER SCHAIE
Dementia prevalence and incidence in non-Hispanic
Black elderly 331
Increase in US aging population (2.22.2021) 320 Discussion 332
Changes in diversity: Race and ethnicity of older US Cohort differences in various cognitive abilities:
population 320 Mean level and variability 332
Significant contribution of boomers to increases in Cohort differences and cognitive change in from 50 to
US aging population 320 80 years in 2 cohorts in old age (1886
1913; 1914
48) 333
US Hispanic baby boomers 321 Cohort differences in terminal decline 333
Boomer generations world wide 322 Changes in prevalence of dementia and possible cohort
Cohort and longitudinal effects: Comparison of cognition differences 334
in cohorts in midlife with special focus on the Future directions 334
baby boomers 322 References 334
Study of cognition in boomers 322 Further reading 337
Memory: Cognitive change trajectory in midlife 325
Cohort differences in cognition in old age: Differential Author index 339
findings 326 Subject index 359
This page intentionally left blank
 List of contributors
Carolyn M. Aldwin Program in Human Development and
Family Sciences, Oregon State University, Corvallis, OR,
United States
Nicole Alea Department of Psychological and Brain
Sciences, University of California Santa Barbara, Santa
Barbara, CA, United States
Kaarin J. Anstey UNSW Ageing Futures Institute,
University of New South Wales, Sydney, NSW, Australia
Cynthia A. Berg Department of Psychology, Consortium
for Families and Health Research, University of Utah, Salt
Lake City, UT, United States
Susan Bluck Department of Psychology, University of
Florida, Gainesville, FL, United States
Steven M. Boker Department of Psychology, The
University of Virginia, Charlottesville, VA, United States
Jennifer L. O’ Brien Massachusetts Institute of Technology
(MIT), Medical Department, Cambridge, MA, United States
Allyson F. Brothers Department of Human Development
and Family Studies, Colorado State University, Fort
Collins, CO, United States
Cheng Chen Media Effects Research Laboratory, Donald P.
Bellisario College of Communications, Pennsylvania State
University, State College, PA, United States
Jessica S. Damoiseaux Department of Psychology and
Institute of Gerontology, Wayne State University, Detroit,
MI, United States
Manfred Diehl Department of Human Development and
Family Studies, Colorado State University, Fort Collins,
CO, United States
Roger A. Dixon Department of Psychology, University of
Alberta, Edmonton, AB, Canada
Amanda F. Elliott Department of Biobehavioral Nursing
Science, University of Florida, College of Nursing,
Gainesville, FL, United States
Sara B. Festini Center for Vital Longevity, University of
Texas at Dallas, Dallas, TX, United States
Stephane P. Francioli New York University, Stern School
of Business, New York, NY, United States
Linda Geerligs Donders Institute for Brain, Cognition and
Behaviour, Radboud University, Nijmegen, The
Netherlands
xi
Jeff Greenberg Department of Psychology, University of
Arizona, Tucson, AZ, United States
Ann L. Horgas Department of Biobehavioral Nursing
Science, University of Florida, College of Nursing,
Gainesville, FL, United States
Tiffany K. Jantz Department of Psychology, University of
Michigan, Ann Arbor, MI, United States
Caitlin S. Kelly Department of Psychology, Consortium for
Families and Health Research, University of Utah, Salt
Lake City, UT, United States
Deborah L. Kerr Department of Neurology, University of
Alabama at Birmingham, Birmingham, AL, United States;
Department of Psychiatry and Behavioral Neurobiology,
University of Alabama at Birmingham, Birmingham, AL,
United States
Michael L. Krieger Media Effects Research Laboratory,
Donald P. Bellisario College of Communications,
Pennsylvania State University, State College, PA, United
States
Hyunyup Lee Department of Psychology and
Management, Korea Military Academy, Seoul, Republic of
Korea
Franziskus Liem University Research Priority Program
“Dynamics of Healthy Aging”, University of Zurich,
Zürich, Switzerland
Ziyong Lin Department of Psychology, University of
Michigan, Ann Arbor, MI, United States
Cindy Lustig Department of Psychology, University of
Michigan, Ann Arbor, MI, United States
Daniel S. Margulies Integrative Neuroscience and
Cognition Center (UMR 8002), Centre National de la
Recherche Scientifique (CNRS) and Université de Paris,
Paris, France
Daniel C. Marson Department of Neurology, University of
Alabama at Birmingham, Birmingham, AL, United States;
Alzheimer’s Disease Center, University of Alabama at
Birmingham, Birmingham, AL, United States
Benjamin T. Mast Department of Psychological & Brain
Sciences, University of Louisville, Louisville, KY, United
States
xii
Molly Maxfield Edson College of Nursing and Health
Innovation, Center for Innovation in Healthy and
Resilient Aging, Arizona State University, Phoenix, AZ,
United States
Catherine A. McCall Department of Pulmonary, Critical
Care, and Sleep Medicine, VA Puget Sound Health Care
System, Seattle, WA, United States; Department of
Psychiatry and Behavioral Sciences, University of
Washington Sleep Medicine Center, Seattle, WA, United
States
Donald G. McLaren Department of Neurology,
Massachusetts General Hospital, Boston, MA, United
States; Harvard Medical School, Boston, MA, United
States
John R. Nesselroade Department of Psychology, The
University of Virginia, Charlottesville, VA, United States
Michael S. North New York University, Stern School of
Business, New York, NY, United States
Rachel Pruchno New Jersey Institute for Successful Aging,
Rowan University School of Osteopathic Medicine,
Stratford, NJ, United States
Patricia A. Reuter-Lorenz Department of Psychology,
University of Michigan, Ann Arbor, MI, United States
K. Warner Schaie Department of Psychiatry & Behavioral
Sciences, University of Washington, Seattle, Washington,
DC, United States
List of contributors
Shubam Sharma Department of Psychology, University of
Florida, Gainesville, FL, United States
Eva-Maria Stelzer Department of Psychology, University of
Arizona, Tucson, AZ, United States
S. Shyam Sundar Media Effects Research Laboratory, Donald
P. Bellisario College of Communications, Pennsylvania State
University, State College, PA, United States
Rebecca L. Utz Department of Sociology, Consortium for
Families and Health Research, University of Utah, Salt
Lake City, UT, United States
Nathaniel F. Watson Department of Neurology, University
of Washington Sleep Medicine Center, Seattle, WA,
United States
Hans-Werner Wahl Network Aging Research, Heidelberg
University, Heidelberg, Germany
Susan Krauss Whitbourne University of Massachusetts
Boston, Boston, MA, United States
Sherry L. Willis Department of Psychiatry & Behavioral
Sciences, University of Washington, Seattle, Washington,
DC, United States
Loriena Yancura Department of Family and Consumer
Sciences, University of Hawai’i, Manoa, Honolulu, HI,
United States
Brian P. Yochim VA Saint Louis Health Care System,
St. Louis, MO, United States
 About the editors
K. Warner Schaie holds an appointment as affiliate
Professor of Psychiatry and Behavioral Sciences at the
University of Washington. He is also the Evan Pugh
Professor Emeritus of Human Development and
Psychology at the Pennsylvania State University. He
received his PhD in clinical and developmental psy-
chology from the University of Washington, a honor-
ary DPhil. from the Friedrich-Schiller University of
Jena, Germany, and a honorary ScD degree from West
Virginia University. He received the Kleemeier Award
for Distinguished Research Contributions and the
Distinguished Career Contribution to Gerontology
Award from the Gerontological Society of America, the
MENSA lifetime career award, and the Distinguished
Scientific Contributions award from the American
Psychological Association. He is a past president of the
APA Division of Adult Development and Aging and
currently represents that Division on the APA Council
of Representatives. He is author or editor of more than
60 books including the textbook Adult Development and
Aging (5th edition, with S.L. Willis) and of all previous
editions of the Handbook of the Psychology of Aging
(with J.E. Birren or S.L. Willis). He has directed the
Seattle Longitudinal Study of cognitive aging since
1956 and is the author of more than 300 journal articles
and chapters on the psychology of aging. His current
research interest is in the life course of adult intelli-
gence, its antecedents and modifiability, the impact of
cognitive behavior in midlife upon the integrity of
brain structures in old age, the early detection of risk
for dementia, as well as methodological issues in the
developmental sciences.
Sherry L. Willis is a research professor in the
Department of Psychiatry and Behavioral Sciences at
the University of Washington. She previously held an
appointment as Professor of Human Development at
the Pennsylvania State University. Her research has
focused on age-related cognitive changes in later adult-
hood. In particular she is known for her work on
behavioral interventions to remediate and enhance
cognitive performance in community-dwelling normal
xiii
elderly. She was a principal investigator on the
ACTIVE study, a randomized controlled trial to
examine the effects of cognitive interventions in the
maintenance of everyday functioning in at-risk com-
munity-dwelling elderly, funded by NIA. She has been
the codirector of the Seattle Longitudinal Study. In
addition to her cognitive intervention research, she has
conducted programmatic research on changes in
everyday problem-solving competence in the elderly
and cognitive predictors of competence. She and col-
leagues have developed several measures of Everyday
Problem Solving. She is the coauthor of the textbook
Adult Development and Aging (with K. Warner Schaie,
now in its 5th edition). She has edited more than 10
volumes on various aspects of adult development and
cognition and has authored over a hundred publica-
tions in adult development. She has served as
President of Division 20, Adult Development and
Aging, American Psychological Association. She was a
Fulbright Fellow in Sweden. She received a Faculty
Scholar Medal for Outstanding Achievement and the
Pauline Schmitt Russell Distinguished Research Career
Award from the Pennsylvania State University, and
the Paul and Margret Baltes award from Division 20 of
the American Psychological Association. She currently
has funding from NIA for a 20-year follow-up of the
ACTIVE randomized cognitive training trial and fund-
ing to archive the Seattle Longitudinal Study.
Bob G. Knight, PhD, is a professor of Psychology and
Counseling at the University of Southern Queensland
in Australia. Prior to that, for 28 years he was on the
faculty of the University of Southern California School
of Gerontology, a world leader in the study of aging.
Dr. Knight has published extensively in mental
health and aging, including Psychotherapy with older
adults (Sage, 3rd ed. 2004, available in French, Dutch,
Japanese, and Chinese translations). He coauthored
with Nancy Pachana PhD (University of Queensland),
Psychological assessment and therapy with older adults
(Oxford University Press, 2015). His writing on psy-
chotherapy with older adults includes development of
xiv About the editors
the Contextual Adult Life Span Theory for Adapting
Psychotherapy (CALTAP), which applies life span
developmental principles to therapy with older adults.
Becca R. Levy is a professor of Epidemiology at Yale
School of Public Health and Professor of Psychology at
Yale University. She received her PhD in psychology
from Harvard University. Her research explores psy-
chosocial factors that influence older individuals’ cog-
nitive and physical functioning, as well as their
longevity. She is credited with creating a field of study
that focuses on how positive and negative age stereo-
types, which are assimilated from the culture, can
have beneficial and adverse effects, respectively, on
the health of older individuals. Her studies have been
conducted by longitudinal, experimental, and cross-
cultural methods. This research has led to receiving a
number of awards including the Baltes Distinguished
Research Achievement Award from the Adult
Development and Aging Division of the American
Psychological Association, the Richard Kalish
Innovation Publication Award from Gerontological
Society of America, the Brookdale National Fellowship
for Leadership in Aging and the Ewald W. Busse
Research Award for Excellence in Social Behavioral
Sciences from the International Association of
Gerontology and Geriatrics. Her research has been
supported by the National Institute on Aging, the
National Heart, Lung and Blood Institute, the National
Science Foundation, and the Donaghue Medical
Research Foundation. Dr. Levy has given invited testi-
mony before the United States Senate on the effects of
ageism and contributed to briefs submitted to the
United States Supreme Court in age-discrimination
cases.
Susan M. Resnick, PhD, Senior Investigator, Chief,
Laboratory of Behavioral Neuroscience.
Dr. Resnick received her PhD in Differential
Psychology and Behavioral Genetics from the
University of Minnesota and completed a postdoctoral
fellowship in Neuropsychology and Neuroimaging at
the University of Pennsylvania. She was a research
assistant professor of Psychology in Psychiatry at the
University of Pennsylvania prior to joining the
Intramural Research Program of the NIA in 1992. She
is Chief of the Laboratory of Behavioral Neuroscience
and the Brain Aging and Behavior Section at the NIA.
Dr. Resnick initiated the neuroimaging substudy of the
Baltimore Longitudinal Study of Aging (BLSA) in 1994
and studies cognitive and brain aging in health and
disease, with a focus on early markers of cognitive
impairment and Alzheimer’s disease. She also serves
as NIA Project Officer for the Women’s Health
Initiative Memory Study (WHIMS) Suite of Studies,
which assess cognitive decline and risk for dementia in
women previously randomized to postmenopausal
hormone therapy versus placebo through the WHI ran-
domized clinical trials.
 Foreword
Since the inaugural publication of the Handbooks of
Aging in 1976, the series has played a key role in pro-
moting and guiding gerontological science. By preserv-
ing foundational knowledge and illuminating
emerging areas, the series has served as a core
resource for established researchers and an inspiration
for students of gerontology. From its inception, geron-
tological science has been cross-disciplinary. The three-
volume series has played a key role in maintaining
cohesion in a science that spans dozens of disciplines.
The need to understand aging only increases in
importance over time. The global population has now
passed an important tipping point, moving from a
world where children predominate to one in which
there are more older people than youth. This reshap-
ing of the age distribution in the population demands
grand investments in the science of aging.
Thankfully, the science of aging is also growing fas-
ter than ever across social and biological sciences.
Along with phenomenal advances in the understand-
ing of the biology of aging as well as genetic influences
on aging trajectories, and susceptibility to age-related
diseases has come the awareness of the critical impor-
tance of the physical and social environments in which
people age and the psychological factors that modulate
and sometimes alter genetic predispositions.
The Handbooks of Aging series, comprised of the
Handbook of the Biology of Aging, the Handbook of the
Psychology of Aging, and the Handbook of Aging and the
Social Sciences, is now in its ninth edition. The
Handbook of Aging and the Social Sciences and the
Handbook of the Psychology of Aging have long provided
conceptual anchors and frameworks to the social and
behavioral sciences while also addressing emerging
topics that did not exist decades ago, such as the fluid-
ity of race and gender, groundbreaking insights into
the role of sleep in cognitive aging, and the ways that
xv
smartphones, robots, and social media can modify the
experience of aging. The handbooks also provide
cutting-edge updates to the understanding of genetics,
built environments, and intergenerational commit-
ments. The 9th edition of the Handbook of the Biology of
Aging introduces geroscience, a discipline that did not
exist 10 years ago and is now among the most vibrant
in all of science. This edition also provides updates on
the exciting advances in the genetics and integrative
genomics of aging and longevity as well as the biology
and therapeutic opportunities afforded by the studies
of cellular senescence.
What has not changed over the editions is the
superb synthesis of the field. The editors of the 9th edi-
tion extend a long tradition of giants in the field giving
generously of their time and knowledge to produce
consistently excellent volumes. Their thoughtful selec-
tion of topics and recruitment of deeply knowledge-
able authors is reflected throughout the series. We are
most grateful to Nicolas Musi and Peter J. Hornsby,
editors of the Handbook of the Biology of Aging, Kenneth
F. Ferraro and Deborah S. Carr, editors of the Handbook
of Aging and the Social Sciences, and K. Warner Schaie
and Sherry Lynn Willis, editors of the Handbook of the
Psychology of Aging.
We also express our deep appreciation to our pub-
lishers at Elsevier, whose profound interest and dedi-
cation to the topic has facilitated the publication of the
Handbooks through many editions. We remain eternally
grateful to James Birren, for establishing the series and
shepherding it through the first six editions that
played a profound role in establishing the tradition of
multidisciplinary science in the field of aging.
Thomas A. Rando and Laura L. Carstensen
Stanford Center on Longevity, Stanford University,
Stanford, CA, United States
This page intentionally left blank
 Preface
The Handbook of the Psychology of Aging provides a
basic reference source on the behavioral processes of
aging for researchers, graduate students, and profes-
sionals. It also provides perspectives on the behavioral
science of aging for personnel from other disciplines.
The ninth edition of the Handbook continues to reflect
both the continuing interest of the scientific commu-
nity as well as the needs and worldwide growth of the
older portion of the population as well as the increase
in active life expectancy. The growth of the research
literature provides new opportunities to replace chro-
nological age as the primary variable with other vari-
ables that represent causal mechanisms and hence
present the potential for control or experimental modi-
fication. Both academic and public interests have been
contributing to the emergence of the psychology of
aging as a major subject in universities and research
institutions. Issues of interest to the psychology of
aging touch upon many features of daily life, from the
workplace and family life to public policy matters cov-
ering, health care, retirement, social security, and
pensions.
The psychology of aging is complex and many new
questions keep being raised about how behavior is
organized and how it changes over the course of life.
Results of the markedly increasing number of longitu-
dinal studies are providing new insights into the
casual factors in behavior changes associated with
adult development and aging and to the differences in
longitudinal trajectories for various factors and cohort
differences in longitudinal trajectories for the same fac-
tor. They are contributing to our understanding of the
role of behavior changes in relation to biological,
health, and social interactions. Parallel advances in
research methodology, particularly directed toward
xvii
the problems of studying change, allow us to explicate
in greater detail, patterns and subpatterns of behavior
over the life span. Facing the rapidly accelerating
growth of the relevant research literature, the editors
once again have had to make choices about what new
topics should be included in the handbook. But the
growth in research activity does not occur uniformly
across all fields. Hence, some topics covered in earlier
editions of the Handbook are not included in the pres-
ent edition.
The chapters are organized into four divisions:
Section I, Concepts, Theories, Methods; Section II,
Biopsychosocial Factors in Aging; Section III,
Behavioral Processes; and Section IV. Complex
Processes. Section I covers understanding theoretical
perspectives of psychology of aging in a life span con-
text and methods in the study of adult development
and aging. Section II includes all new chapters on
functional connectivity in aging, pain assessment and
management in older patients, and a continued cover-
age of sleep, and death and dying. Section III includes
new chapters on the usage of smartphones, robots, and
social media, self-perceptions of aging, age and gender
identity, age and gender discrimination in the work-
place, and autobiographical memory. Section IV
includes new coverage of clinical psychological testing,
family caregiving across the life span, stressors in
aging and coping, resilience in aging, successful/opti-
mal aging, and cognition, neural structure and pathol-
ogy. The chapters for the ninth edition were underway
before the advent of the pandemic; future editions will
need to examine the effect of this event on the psychol-
ogy of aging.
Sherry L. Willis and
K. Warner Schaie
This page intentionally left blank
 S E C T I O N I

Concepts, theories, methods

This page intentionally left blank
 C H A P T E R

1
History of adult cognitive aging research
K. Warner Schaie
Department of Psychiatry & Behavioral Sciences, University of Washington, Seattle, Washington, DC, United States

O U T L I N E

Introduction 3 Early cross-sectional studies 9

Adult stages of intellectual development 4 Longitudinal studies 10
Cohort differences 10
A stage model of adult cognitive development 4
Magnitude of change 12
Assessment of intellectual functions 6 Frequency of decline 13
A test of intelligence 6 Personality and lifestyles 13
The nature of intelligence 7
Summary 14
Intelligence as multiple abilities 7
Practical or everyday intelligence 8 References 15
Intelligence and age 8 Further reading 17
Crystallized and fluid intelligence 9

Introduction
Mental abilities have long been valued in Western
culture as the basis for learning, problem solving, and
adjustment. Thus intelligence and cognitive develop-
ment quickly became one of the major concerns of the
earliest psychologists. Great efforts were made to
define intelligence, to measure it, and even to try to
increase it. The study of cognition has had a long and
often stormy history. Indeed, the controversies are no
less stormy today, as perhaps should be expected
when dealing with an ability so highly esteemed (cf.
Bengtson & Sattersten, 2016; Gerstrof et al., 2020)
Some of the current controversies have to do with the
role of intelligence in everyday life. Are intelligent people
more successful in school than less intelligent people?
Are they more successful in life? Can they repair a car
more efficiently? Can they run a business more effec-
tively? What other factors besides intelligence are
involved in personal competence? What is the relation-
ship between “academic” and “practical” intelligence? Is
competence in the elderly the same as competence in
young adults? In the following sections, I discuss the dif-
ferent theories of intelligence and research on adult intel-
lectual development. The questions that I seek to answer
are fairly simple, such as “Does intelligence increase or
decline with age?” The answers, however, are more com-
plicated; they vary with age, the specific intellectual func-
tion we are considering, and even the year in which the
individual was born. This latter influence, which com-
prises a number of “generational” or “cohort” factors, is
considered in some detail.

Handbook of the Psychology of Aging.

DOI: https://doi.org/10.1016/B978-0-12-816094-7.00017-9 3 © 2021 Elsevier Inc. All rights reserved.
4 1. History of adult cognitive aging research
Adult stages of intellectual development
What is the nature of intelligence in adults? How is
it similar to the intelligence of young persons and how
does it change? If IQ tests are to be constructed that
are fair to older people, we must know more about
adult cognition; and in particular we need to know in
what sense people might increase their competence as
they grow older.
The famous Swiss psychologist Jean Piaget (Flavell,
1963). described the ways in which children’s intelli-
gence increases as they develop. Infants are said to be
at a relatively primitive, sensory-motor stage of intel-
lectual development. They learn simple but basic ways
of perceiving and reacting to the world. With the onset
of speech, children enter a stage in which they grow
primarily in the “conceptual-symbolic rather than
purely sensory-motor arena.” This stage, called preop-
erational, is succeeded around the age of 6 by the stage
of concrete operations. In Piaget’s theory, operations
are the mental routines that transform information in
some way, for example, adding two numbers to get a
third or categorizing, as in placing all red objects
together. The stage of formal operations is entered
around the age of 12 and is defined by the ability to
use mental operations on abstract material.
Intellectual development, of course, is not complete
at the age of 12 when the average child enters the stage
of formal operations, but Piaget does not provide us
with much detail on later development. Although we
can assume that there are advances in the use of for-
mal operations, no new Piagetian stages were specified
for adulthood (Flavell, 1970; Piaget, 1972).
Psychologists who focus on adult development find
this child-centered approach restrictive and wish to
expand it so as to delineate those changes in the qual-
ity of intellectual function that they observe in adult
study participants. As Erik Erikson and Daniel
Levinson expanded the psychoanalytic stages of ego
development to the adult years, these psychologists
have done the same for Piaget’s stages of intellectual
development (Commons, Sinnott, Richards, & Armon,
1989; Commons, Miller, & Giri, 2014; Commons &
Ross, 2008; Sinnott, 1996)
A stage model of adult cognitive development
I have used findings from research on adult intellec-
tual development to formulate six adult stages (Schaie
& Willis, 1999, 2000; Schaie, 1977-78). Piaget’s
I. Concepts, theories, methods
childhood stages describe increasing efficiency in the
acquisition of new information. It is doubtful that
adults progress beyond the powerful methods of sci-
ence (formal operations) in their quest for knowledge.
Therefore if one is to propose adult stages, they should
not be further stages of acquisition; instead they
should reflect different uses of intellectual abilities.
In young adulthood, for example, people typically
switch their focus from the Acquisition stage to the
application and integration of knowledge, as they use
what they know to pursue careers and develop their
families. This is called the Achieving stage. It repre-
sents most prominently the application of intelligence
in situations that have profound consequences for
achieving long-term goals.
These situations are not the hypothetical ones posed
on IQ tests or encountered in classroom studies, nor
are they the problems of childhood, whose solutions
are closely monitored by parents and society. Instead,
they are problems that the adult must solve for him-
or herself, and the solutions must be integrated into a
life plan that extends far into the future. The kind of
intelligence exhibited in such situations is similar to
that employed in educational tasks, except that it
requires more careful attention to the possible conse-
quences of the problem-solving process.
Young adults who have mastered the cognitive
skills required for monitoring their own behavior and,
as a consequence, have attained a certain degree of
personal independence, will next move into a stage
that requires the application of cognitive skills in situa-
tions involving social responsibility.
Typically, this Responsible stage occurs when a
family is established and the needs of spouse and off-
spring must be met. Similar extensions of adult cogni-
tive skills are required as responsibilities for others are
required on the job and in the community (Hagestad
& Neugarten, 1985).
Some individuals’ responsibilities become exceed-
ingly complex—the Executive stage. Such indivi-
duals—presidents of business firms, deans of
academic institutions, officials of churches, and a num-
ber of other positions—need to understand the struc-
ture and the dynamic forces of organizations. They
must monitor organizational activities not only on a
temporal dimension (past, present, and future), but
also up and down the hierarchy that defines the
organization.
They need to know not only the future plans of the
organization, but also whether policy decisions are
being adequately translated into action at lower levels
of responsibility. Attainment of the Executive stage, as
 A stage model of adult cognitive development
an extension of the Responsibility stage, depends on
exposure to opportunities that allow the development
and practice of the relevant skills (Smith, Staudinger,
& Baltes, 1994).
In the later years of life, beyond the age of 60 or 65,
the need to acquire new knowledge declines even more
and executive monitoring is less important because fre-
quently the individual has retired from the position that
required such an application of intelligence. What, then,
is the nature of competence in an elderly adult? There is
a transition from the childhood question “What should I
know?” through the adult question “How should I use
what I know?” to the question of later life “Why should
I know?” This Reintegration stage, corresponds in the
life course to Erikson’s stage of ego integrity.
The information that elderly people acquire and the
knowledge they apply is, to a greater extent than earlier
in life, a function of their interests, attitudes, values,
and physical health (Bowen & Staudinger, 2013;
Westerhof et al., 2014; Heidemeier & Staudinger, 2015).
It requires, in fact, the Reintegration of all of these.
The elderly are less likely to “waste time” on tasks that
are meaningless to them (Berg & Klacynski, 1996;
Staudinger & Glueck, 2011).
This stage also frequently includes a selective reduc-
tion of interpersonal networks in the interest of reinte-
grating one’s concern in a more self-directed and
self-supportive manner (Carstensen, 1993; English &
Carstensen, 2014; Scheibe, English, Tsai, & Carstensen,
2013).
Striving to feel good: ideal affect, actual affect, and
their correspondence across adulthood. Such efforts
are likely to involve a reduction in information-seeking
activities while increasing the importance of emotional
regulation (Carstensen, Gross, & Fung, 1997; Schaie &
Carstensen, 2006).
The original stages were formulated some 30 years
ago. Since that time we have learned a lot about the
differentiation of our older population into distinct life
stages. In the research literature distinctions are now
commonly made between the young-old, the old-old,
and the oldest-old (or very-old). I have therefore
added two further stages.
In the reintegrative stage, a major effort is now
required to Reorganize one’s life in order to replace
the earlier engagement with family-raising and job
responsibilities to meaningful pursuits for the last part
of life. In addition, efforts must be directed toward
planning how one’s resources will last for the remain-
ing 15
30 years of postretirement life that are now
characteristic for most individuals in industrialized
societies.
I. Concepts, theories, methods
5
These efforts include active planning for that time
when dependence upon others may be required to
maintain a high quality of life in the face of increasing
frailty. Such efforts may involve changes in one’s hous-
ing arrangements, as well as making certain of both
familial and extrafamilial support systems. The activi-
ties include making or changing one’s will, drawing
up advanced medical directives, and durable powers
of attorney (Schaie & Willis, 2000).
Although some of these activities involve the similar
cognitive characteristics of the Responsible stage, we
think that the objectives involved are generally far more
centered to current and future needs of the individual
rather than the needs of their family or of an organiza-
tion. Efforts must now be initiated to reorganize one’s
time and resources to substitute a meaningful environ-
ment, often found in leisure activities, volunteerism, and
involvement with a larger kinship network.
Eventually, however, these are activities that are
also engaged in with the finitude of life in clear view,
for the purpose of maximizing the quality of life dur-
ing the final years. The unique objective of these
demands upon the individual represent an almost
universal process occurring at least in the industrial-
ized societies, and designation of a separate
Reorganizational stage is therefore warranted.
The skills required for the Reorganizational stage
require the maintenance of high levels of cognitive
competence, which is increasingly exercised within the
parsimonious principles of selection, optimization, and
compensation. In addition, maintenance of flexible
cognitive styles is required, in order to be able to
restructure the context and content of life after retire-
ment, to relinquish control of resources to others, and
to accept the partial surrender of one’s independence
(Schaie, 1984, 1996).
More and more older persons reach advanced old
age in relative comfort and often with a clear mind
albeit they are coping with a frail body. Once the
Reintegrative efforts described above have been suc-
cessfully completed, and perhaps temporally overlap-
ping with them, there is yet one last stage that is
frequently observed. This last phase, which I have des-
ignated as the Legacy creating stage is concerned with
cognitive activities of many, if not all of the very old,
that occur in anticipation of the end of their life. This
stage often begins with a life review. For the highly lit-
erate and those successful in public or professional life
this will often include writing or revising an
autobiography.
An approximate time line for the Schaie stage model
is provided in Fig. 1.1. But it should be stressed that the
6 1. History of adult cognitive aging research

FIGURE 1.1 Schaie’s stages of adult cognitive development. Source: From: Schaie, K.W. & Willis, S.L. (2000). A stage theory model of adult cog-
nitive development revisited. In R. Rubinstein, M. Moss & M. Kleban (Eds.), The many dimensions of aging: Essays in honor of M. Powell Lawton,
(pp. 175-193). New York: Springer Publishing Co.

precise chronological age at which these stages occur
may be quite variable in different societies as well as
for individuals at different levels of intellectual compe-
tence and personal engagement. What is important is
the sequential process of these developmental stages.
Assessment of intellectual functions
Psychological tests were originally developed to iden-
tify intelligent people. Francis Galton believed that
human intelligence is mostly inherited. But he wondered
how could the most intelligent people be identified? A
test of intelligence would have to be created. Galton took
on the job and in 1883 published the first intelligence test.
A test of intelligence
Influenced by British philosophers who considered
intelligence to be based on the ability to process sen-
sory information, Galton (1883) devised a series of
tasks designed to measure how well a person could
see, hear, smell, taste, and feel.
Galton’s “mental test” (as he called it) was not very
successful; it showed only trivial correlations with
measures of intellectual competence in the real world,
such as scholastic performance (Wissler, 1901).
Almost 20 years later, a French psychologist by the
name of Alfred Binet tried again to construct a test of
intelligence. He had been given a much more practical
problem to solve by the French Ministry of Public
Instruction. They needed a test to distinguish students
of low ability (mentally retarded) from those of ade-
quate ability but low motivation.
Binet and Simon (1905) held a more traditional view of
intelligence than Galton, believing, for example, that play-
ing chess was a better indicator of intelligence than smell-
ing vinegar. He decided to assess “reasoning, judgment,
and imagination” by a series of cognitive problems.
Because Binet’s miniature tasks were quite similar to
those that children are expected to face in school, scores
on his test were highly correlated with scholastic perfor-
mance. First published in 1905, Binet’s test (Binet &
Simon, 1905) was quickly translated into other languages.
In the United States, his test was translated and revised
by Stanford psychologist Lewis Terman and became
known as the widely used Stanford
Binet Intelligence
Scale.
The background information on intelligence testing
is relevant for our discussion of adult intelligence for
two reasons. The first is to show that the testing move-
ment in psychology began in practical circumstances—
there was a need to predict the potential for scholastic
success.

I. Concepts, theories, methods

 Assessment of intellectual functions
IQ tests are age graded, that is, the average score
for each age level is given the score of 100. A ques-
tion such as “Who has the higher IQ, an average
10-year-old or an average 70-year-old?” is meaning-
less. They both may have IQs of 100, the average for
each age group. However, as we will see other kinds
of comparisons can be made that inform us as to
how intelligence changes from childhood into
advanced old age.
The nature of intelligence
From the very beginning, there has been a great
deal of debate about the nature of intelligence and
whether there may be different kinds of intelligence. Is
intelligence a single, general ability or are there several
different intellectual abilities? Binet favored the idea of
a “general ability” (sometimes called the “g” factor),
but later researchers have favored the notion of several
factors in intelligence.
Some intelligence tests have a number of subtests
covering different content. The Wechsler Adult
Intelligence Scale (WAIS-R) is the test most frequently
used by clinical psychologists for the individual assess-
ment of adult intelligence (Wechsler, 1997).
The fact that there are slightly different subtests on
an intelligence test is of course no guarantee that these
subtests actually measure different intellectual abilities;
they may simply be different ways of measuring a sin-
gle ability: “general intelligence.” Further exploration
has therefore taken the form of factor analysis, a statis-
tical procedure that identifies the number of basic
dimensions or factors in a set of data.
Factor analysis will tell us if intelligence is a one-
dimensional construct or whether it is a construct with
multiple dimensions. The answer to this question is
both. In a factor analysis of the WAIS subtests, for
example, the major dimension was found to be that of
general intelligence, a large factor that accounted for
about half of the information contained in the
test. Three other factors appear to be important for
some purposes. For example, an individual high in
perceptual-organizational abilities might do better on
the block design subtest than we would expect from
his or her general intelligence alone (Cohen, 1957).
One finding of interest in this study is that the
memory factor, a relatively weak factor among young
study participants, became a major factor for persons
over the age of 60. This means that specific memory
abilities vary more among older people and affect
scores on more of the subtests.
I. Concepts, theories, methods
7
Intelligence as multiple abilities
If one’s goal is to map the broad scope of intelli-
gence and not simply that of the WAIS, many different
intellectual tasks must be administered to a large num-
ber of people. Factor analysis of a wide variety of intel-
lectual tasks has regularly turned up between 6 and
12 primary mental abilities. These abilities have some-
times been described as the “building blocks” or basic
elements of intelligence. (Thurstone, 1962). The “pur-
est” tests of these factors are sometimes administered
as tests of the “primary mental abilities.” A more
recent adult version of these tests is called the
Schaie
Thurstone Adult Mental Abilities Test
(STAMAT; Schaie, 1985, 1996, 2013).
But what is the nature of the relationship between
such elementary building blocks of intelligence and
the tasks that people face in real life? To find out, per-
formance on the different primary mental abilities was
supplemented in a sample of over 1000 persons by
administering real-life tasks such as interpreting medi-
cine bottle labels, reading street maps, filling out
forms, and comprehending newspaper and yellow
page advertisements.
The researchers found a substantial correlation
between abilities and performance on tasks; correla-
tions varied, however, depending on the task.
Furthermore, it was found that the composite perfor-
mance on the real-life tasks could be predicted by sev-
eral abilities, particularly reasoning, but also by verbal
knowledge to a lesser extent. This also suggests a
strong relationship between “building blocks” of intel-
ligence and perceived real-life competence.
We have come, then, from the view of intelligence
as primarily a single trait to the view of intelligence as
a number of distinct abilities. As we shall see, the dis-
tinctions of several different abilities are vital for the
study of intellectual development in adults.
Relevance of test instruments to stages of
intellectual development
The simple tasks in the traditional IQ tests are well
suited to measure progress in the performance of
many basic skills through the stages of knowledge
acquisition described by Piaget (Humphreys &
Parsons, 1979). But they are decidedly less adequate
for the assessment of adult competence.
Even a test that was constructed explicitly for
adults, the WAIS, is deficient in several respects. First,
the test was designed with the intent of measuring
cognitive dysfunctions in clinically suspect individuals,
and second, it was originally normed on young adult
8 1. History of adult cognitive aging research
samples, those who in our conceptual scheme would
be classified as being in the achieving stage, although
norms for midlife and older adults are now available.
What we need, therefore is to construct adult tests of
intelligence relevant to competence at different points
in the life span, just as the traditional test is relevant to
the competencies of children in school settings.
Practical or everyday intelligence
Some would argue that intelligence in adults should
be studied by asking well-functioning people how they
go about solving their everyday problems (Sternberg &
Lubart, 2001). This is what is known as a “naive” the-
ory of intelligence; that is, it is not derived from objec-
tive analyses of experts, but rather from the collective
perceptions of laypersons. Perhaps it is indeed the con-
ceptions of adults about their own competence that
ought to be the basis for defining intelligence. But
there is the distinct danger that in this process
we would confuse intelligence with socially desirable
behavior. Moreover, the attributes of intelligence
obtained in this manner may be characteristic only of
the specific group of persons interviewed or may be
governed by time-specific and/or context-specific
conceptions.
We would be remiss, then, if we were to discard the
objective knowledge of mental functioning that is now
in hand and is directly applicable to adult intelligence
(Schaie & Willis, 1999; Willis & Schaie, 2005). Instead,
we may wish to consider how the basic intellectual
processes that are important at all life stages relate to
everyday tasks (also see Diehl, Willis, & Schaie, 1995;
Marsiske & Margrett, 2006; Wettstein, Wahl, & Diehl,
2014).
There have been a number of efforts to develop
objective measures of people’s abilities to engage in
effective problem solving and to perform tasks
required for daily living (see Marsiske & Willis, 1995;
Willis, 1996, 1997). For example, the Educational
Testing Service (1977) developed a test to assess
whether high school graduates had acquired the neces-
sary information and skills to handle everyday pro-
blems. This test includes tasks such as interpreting bus
schedules, tax forms, labels on medicine bottles, adver-
tisements, and understanding instructions for the use
of appliances, and the meaning of newspaper opinion/
editorials. The test has been given to large samples of
adults ranging in age from the 20s to the 80s (Schaie,
1996, 2013). The test correlates with a number of the
primary mental abilities; in fact, most of the individual
I. Concepts, theories, methods
differences on the test can be predicted from knowl-
edge of scores on the basic abilities tests.
Another effort to measure everyday problem solv-
ing was a test constructed to assess the skills that old
people are thought to need to function independently
in the community. These skills, called the instrumental
activities of daily living (IADL; Lawton & Brody,
1969), include the ability to engage independently in
food preparation, housekeeping, medication use, shop-
ping, telephone use, transportation, and financial man-
agement activities. Obviously, each of these activities
requires the exercise of practical intelligence.
Marsiske and Willis (1995) collected written materi-
als (e.g., medication labels, bus schedules, telephone
instructions, mail order forms, appliance instructions,
etc.) that are actually used for each of the seven types
of activities. These items were rated as to their rele-
vance by professionals working with older people, and
then a test was constructed that measured proficiency
with the information to carry out each activity of daily
living independently. The validity of these measures
was validated further by observing individuals in their
homes actually using these materials to engage in
activities such as measuring out medications, using a
microwave, and so forth. Again, individual differences
on this everyday problems test could be explained in
large part by the performance of individuals on the
basic abilities.
Intelligence and age
What happens to intelligence with age? This is a key
question in this chapter. It is argued by some that intelli-
gence enters a process of irreversible decline in the adult
years, because the brain becomes less and less efficient,
just like the heart and lungs and other physical organs.
Others contend that intelligence is relatively
stable through the adult years, with the human brain pro-
viding more than enough capacity for anything that we
would want to contemplate until serious disease, patholo-
gies, and declines in sensory functions set in late in life.
Another view is that intelligence declines in some
respects (in perceptual or mental speed, for example)
and increases in others (in knowledge about life, or wis-
dom, for example, see Staudinger & Glueck, 2011). Some
argue that individual differences can be explained by
compensatory experiences for those who age well, while
others place emphasis on the above-average maintenance
of physiological and psychological resources for the
more favored individuals (Salthouse, 1999, but see
Schaie, 2009).
 Intelligence and age
Crystallized and fluid intelligence
One of the most prominent theories of “hold” and
“don’t hold” tests was formulated by Raymond Cattell
and elaborated by John Horn. In factor analyses of
cross-sectional studies of several intellectual tasks (not
from the WAIS), Cattell and Horn repeatedly discov-
ered that the tests on which older adults do well com-
pared to younger adults show up as a factor that they
call crystallized intelligence (Gc).
As represented by tests of general information and
vocabulary, crystallized intelligence is said to reflect
the mental abilities that depend on experience with the
world—on education in the broad sense, including
both formal schooling and informal learning experi-
ences in everyday life. The “don’t hold” tests show up
as another factor, termed fluid intelligence (Gf).
Fluid intelligence is more akin to what Wechsler called
“native mental ability,” reflecting presumably the quality
of one’s brain: the speed of neural transmission, how
well organized are the neurons involved in associations,
pattern recognition, and memory (Horn & Hofer, 1992).
Adult intellectual development, viewed in terms of
the Gc
Gf theory, implies progressive deterioration in
the neural structures underlying intelligence and thus
systematic decline in fluid intelligence. Crystallized
intelligence, as long as we do not require speedy
responses, should not be affected as much; it may even
increase as a result of adult educational experiences.
The theory is a popular one, for it more clearly speci-
fies the intellectual tasks that can be used to represent
each type of intelligence. Indeed, it is a form of the
general notion that in some respects older adults are
not as sharp as they once were, but in other respects
they are as knowledgeable as ever, perhaps even
wiser.
Differential decline of intelligence that supports the
Gc
Gf theory also comes from a variety of longitudinal
studies in various Western countries that show greater
decline for measures of fluid abilities.
However, as will be shown later, in the detailed dis-
cussion of our longitudinal studies, this pattern may
not hold for all abilities, and in addition may be atten-
uated in advanced old age, when crystallized abilities
also show substantial decline.
Early cross-sectional studies
If we were to administer a typical IQ test such as
the WAIS to various groups representative of people
their age, we would find that each group has an aver-
age IQ of 100. By definition, because the average IQ
I. Concepts, theories, methods
9
for any age group has been set to 100. This automatic
adjustment disguises the fact that a 20-year-old must
earn a higher raw (unadjusted) score to be assigned an
IQ of 100 than does a 50-year-old. On one test, a raw
score of 80 places a 50-year-old slightly above average
for his age group, with an IQ of 101; but a 20-year-old
with the same score would be decidedly below aver-
age with an IQ of 87.
If we compare raw scores across age groups, we
would find that average scores are highest in young
adulthood and early middle age and systematically
decline with advancing age. The interpretation of early
cross-sectional studies reporting such findings was
straightforward: an individual’s intellectual abilities
gradually but inexorably decline over the adult years.
David Wechsler, creator of the WAIS, believed that
the “decline of mental ability with age is part of the
general senescent process of the organism as a whole.”
Wechsler believed that mental ability deteriorated in
ways similar to the decline of lung capacity, reproduc-
tive function, and other physical abilities (Wechsler,
1972, p. 30). It was soon apparent, however, even from
cross-sectional studies, that intellectual decline was not
as pronounced on some tasks as on others.
Researchers have noted that certain subtests on the
WAIS declined less than others. Wechsler (1972) pro-
posed to use the term “hold subtests” for those subt-
ests on which older adults do about as well as younger
adults in contrast to “don’t hold” subtests that showed
a greater decline; as discussed previously with regard
to fluid and crystallized intelligence. In general, the
verbal subtests “hold” and the performance subtests
“don’t hold”). It has been shown that there is improve-
ment on the WAIS from 40 to 61 years on the informa-
tion, comprehension, and vocabulary subtests; mixed
change on picture completion (improvement on easy
items and decline on difficult items); but decline on
the digit symbol and block design subtests Why
should some tasks show almost no decline, and others
show the older persons doing much more poorly
than the younger study participants? One possible
explanation is that the subtests in which older people
do poorly are all speeded tests; the scores reflect the
time that it takes for the person to solve the problem,
or it reflects the number of responses in a given time
interval. One might conclude older people are just
slower, but not necessarily less able. This hypothesis,
however, has been partially discredited by research in
which the participants were given unlimited time to
solve problems and only the number correct was
counted. Older people still did less well on “don’t
hold” tasks, relative to younger persons.
10 1. History of adult cognitive aging research
This type of research, of course, changes the “prob-
lem” from “How long does it take you to solve it?” to
“Can you solve it at all?” In fact, the slower average
speed of solution among older persons may be contrib-
uted to in large part by those who do not solve the
problem at all.
Longitudinal studies
Widespread use of intelligence tests among college
freshmen began in the United States in about 1920. By
1950, therefore it was possible to find a sizable group
of 50-year-olds who had taken an IQ test some 30
years earlier. Several psychologists, seeing their chance
to run a relatively inexpensive longitudinal study,
seized the opportunity by retesting these middle-aged
individuals.
No one expected results different from those found
in cross-sectional studies, which suggested a marked
decrease in IQ scores after the age of 25 or 30. Thus it
came as somewhat of a surprise to find that not only
did the longitudinal studies observe virtually no
decline in IQ by middle age; instead they showed an
increase! The average person seemed to have gotten
smarter with age, at least up to age 50 (Owens, 1953).
Later follow-ups showed that the participants in the
Owens study actually maintained their intellectual
abilities into their 60s (Cunningham & Owens, 1983).
Most of the early longitudinal studies tested highly
educated people (college graduates), whose professional
careers required continuing use of academic skills—
mathematics, extensive reading, and formal reasoning.
Later studies of people at all levels of intelligence
and education in all walks of life showed the increase
in IQ scores to late midlife and beyond to be character-
istic in large part for the highly educated part of the
population. But evidence for the absence of significant
declines in young old age, except for slowing of
response time, during midlife and into early old age
has been replicated again and again.
Our large-scale study combined features of both
cross-sectional and longitudinal designs (Gerstorf
et al., 2019; Gerstorf, Ram, Hoppmann, Willis, &
Schaie, 2011; Schaie, 1994, 1996, 2005, 2013; Schaie &
Zuo, 2001; Willis & Schaie, 1986). In 1956 people rang-
ing in age from 22 to 70 were tested in a cross-
sectional study. In 1963 as many of the original study
participants as could be found and convinced to par-
ticipate once again were retested. This procedure was
repeated a third time in 1970, a fourth in 1977, a fifth
in 1984, a sixth in 1991, a seventh time in 1998, and
I. Concepts, theories, methods
another follow-up of previously tested participants in
2005. Thus the researchers had seven cross-sectional
studies in addition to longitudinal data covering a
period of up to 49 years.
The cross-sectional studies showed the typical pat-
tern of intellectual decline in the adult years; the longi-
tudinal data, however, told a quite different story.
The cross-sectional data show a peak at age 39 for
most abilities, followed by a relatively sharp decline in
some abilities such as inductive reasoning and verbal
memory. In striking contrast, the longitudinal data
suggest increases in abilities such as verbal ability/
comprehension until 53 or 60, with a small decline
thereafter; even at age 74 the estimated performance is
better than at 25.
The data just described came from a single longitu-
dinal study over the period from 1956 to 1963. More
recently, longitudinal data for five mental abilities
have been reported from the expanded investigation
covering 1956
2005, a period of 49 years.
Representative findings, shown in Fig. 1.2, suggest
little if any decline in these abilities until the age of 60;
in several instances, increases occur during the adult
years, with peaks in midlife. Even after age 60, average
decline is slight until age 74 or 81.
Other investigators also have observed the fact that
even in fairly advanced age, change in abilities pro-
ceeds quite slowly if no pathology, and in fact is diffi-
cult to document in studies that extend only over 2 or
3 years. Once the high 80s and 90s are reached, how-
ever, declines become more rapid and extend across
most abilities because of the increasing failures of sen-
sory capacities and other physiological infrastructures,
as well as pathology.
However, no adjustments were made for the well-
documented decline in perceptual speed that puts older
people at successively greater disadvantage. Nor were
any of the participants removed who were later known
to have been in the early stages of dementia, thus per-
haps underestimating mean levels for the normal elderly.
In a study by Schaie, the contribution of perceptual
speed was removed statistically from the scores of 838
adults ranging in age from their 20s to their 80s. This
adjustment removed most of the observed age decre-
ment for highly practiced tasks and markedly reduced
aging effects for novel tasks.
Cohort differences
What accounts for the difference between the cross-
sectional and longitudinal results, with the latter not
only showing relatively modest decline through midlife
 Longitudinal studies
55
50
T-Score means
45
Inductive reasoning
40
Spatial orientation
Perceptual speed
35 Numeric facility
Verbal comprehension
Verbal memory
30
25 32 39 46
FIGURE 1.2 Effects of age on six mental abilities in longitudinal studies. Source: From: Schaie, K.W. (2013). Developmental influences on
adult intellectual development: The Seattle longitudinal study (2nd ed.). New York: Oxford University Press.
and early old age, but also, in some cases, clear increases
in intellectual abilities? Why do the longitudinal studies
give us such a different picture from the earlier, cross-
sectional investigations? The answer is largely “birth
cohort” differences—differences among generations. The
reason longitudinal studies give different results from
cross-sectional studies is that cross-sectional studies com-
pare people of different ages and of different cohorts.
Many of the differences that have been attributed to age
must, for the most part, be relegated to differences
among groups of people differing in year of birth. Cross-
sectional studies make it appear that intelligence declines
steeply over the years, but much of this apparent decline
is an illusion. Longitudinal studies suggest generally that
each generation performs at a higher level than the pre-
vious one, on at least some abilities.
Why is one cohort more advantaged on intelligence
measures than another? Why is it that people born more
recently earn higher scores on some intelligence mea-
sures than their parents or grandparents? Various
answers may be suggested. Over the last several genera-
tions in most countries, years of education has increased
across generations. In the United States, among the mem-
bers of the oldest cohorts now living, the majority may
have achieved no more than a high school diploma and
relatively few have had college experience. The spectrum
of occupations has shifted from manual labor to a
I. Concepts, theories, methods
11
Total sample
53 60 67 74 81 88 95
Age
majority of jobs requiring higher educational skills
(Schooler, Mulatu, & Oates, 1999). Nutrition has vastly
improved in the last 70 or 80 years, and so has medical
care; the physical condition of the brains of the more
recent older cohorts may also differ.
The use of tests like those for IQ has burgeoned, and
thus later generations may be better than earlier genera-
tions at performing well on such instruments because
of their added “test-taking” experience. Because many
different experiences may be relevant to differential per-
formance across different cohorts, it is still uncertain
to know which of these experiences are involved in
every psychological mechanism. Laboratory studies
have employed popular recreational activities such as
crossword puzzles or jigsaw puzzles to determine
whether regular performance on these activities is
related to age differences on related abilities. Crossword
puzzles have been related to verbal (crystallized) ability
and jigsaw puzzles to spatial (fluid) ability (Allard
et al., 2014; Pilail et al., 2011). However, whether prac-
tice on crossword puzzles either reduced age difference
on fluid abilities or enhanced age-related increases on
crystallized abilities is less clear.
Cohort differences in intelligence are not uniform
across different abilities. Fig. 1.3 shows the change in
cohort level in percent of the performance of the earli-
est cohort for 11 cohorts born from 1903 to 1973.
12
15
Inductive reasoning
Perceptual speed
Verbal comprehension
10
Cumulative mean T-score change
–5
–10
1903 1910 1917 1924
Almost continuous gain occurred for each successive
cohort for the primary mental abilities of inductive rea-
soning and spatial orientation. However, gain peaked
in 1952 for verbal meaning. Number ability peaked in
1924 and then went down below the level of the oldest
cohort. These differential findings suggest that older
cohorts are at a particular disadvantage on the fluid
abilities, but may have an advantage with respect to
number skills (Schaie, 1996, 2013; Willis, 1989).
The cohort differences data described above come
from the study of groups of unrelated individuals, but
generational differences of very similar magnitude
have been observed also in studies comparing parents
and their adult children when compared at the same
ages. This kind of data can, of course, be collected only
in studies carried on for long periods of time (Schaie,
2005, 2008, 2013).
There is some evidence that the differences between
generations have begun to turn in favor of the earlier-
born cohorts (e.g., numerical ability). Average test scores
on the Scholastic Aptitude Test have been declining since
1962; before 1962, averages were stable or increasing.
The decline has been blamed on many factors, but
chief among them are poorer educational standards
in our schools, more students taking the SAT, and
“the passive pleasure, the thief of time”—television
(Hanford, 1991).
1. History of adult cognitive aging research
FIGURE 1.3 Cumulative cohort
Spatial orientation differences from oldest to youngest
cohort for five mental abilities. Source:
Numeric facility Schaie, K.W. (2013). Developmental
Verbal memory influences on adult intellectual devel-
opment: The Seattle longitudinal
study (2nd ed.). New York: Oxford
University Press.
1931 1938 1945 1952 1959 1966 1973
Cohort
For whatever reason, the youth of today are doing
somewhat less well on some IQ measures than their
elders did at the same age, and this fact will eventually
show on cross-sectional studies of intelligence. It will
then appear that people get smarter as they grow old-
er; a conclusion that will be no more justified than the
one based on present cross-sectional studies that peo-
ple decline in intelligence.
These differential data suggest that older cohorts are
at a particular disadvantage on the fluid abilities, but
may have an advantage with respect to number skills.
Magnitude of change
Psychologists often become absorbed in trying to
demonstrate the presence or absence of change. They
frequently ignore the question of whether or not a
change “makes a difference,” that is, whether a change
is substantial enough to be important in the everyday
lives of the people who are being discussed. One way
to estimate the meaningfulness of change is to look for
the degree of overlap between a particular older age
group and a younger group if the overlap is large (say
80% or 90%), then the age difference does not affect a
great many people and should not be of much practi-
cal significance.
I. Concepts, theories, methods
 Longitudinal studies
Even an overlap of only 50% still means that many
older people will perform as well as many younger
people. It is only when the overlap begins to shrink to
20% or less, that we can indeed argue that most older
persons do significantly less well than younger com-
parison groups.
Frequency of decline
In addition to knowing the age at which the average
person declines, it is also important to know what pro-
portion of people are likely to decline at a given age.
Such knowledge is useful in at least two ways. First, it
alerts us to the fact that there may be more stability
than change in intellectual aging and that some per-
sons may still grow even at an advanced age.
Second, just as longevity tables permit life insurance
companies to forecast the odds of someone’s dying, a
knowledge of the proportion of those declining at a
given age permits us to determine the probability that
intellectual changes will have important consequences.
For example, such knowledge would permit us to deter-
mine the odds that an elderly president might show
mental decline before completing his term in office.
Frequency distributions were prepared for several
thousand participants in the Seattle Longitudinal Study
to determine what proportion had declined significantly
over each 7-year age range from 25 to 32, 60 to 67, and
so on until age 88
95 (Schaie, 1989, 2013). The research-
ers examined frequencies for the five primary mental
abilities: verbal meaning, inductive reasoning, word flu-
ency, numerical ability, and spatial orientation. The
cumulative percent 7-year decrement is shown in
Fig. 1.4. Note that, although ability varied greatly in the
various areas tested, at least 75% of those studied main-
tained their previous ability level to age 60.
Personality and lifestyles
The participants in our longitudinal study provide
evidence also that changes in lifestyles affect IQ scores.
On the basis of intensive interviews, four types of par-
ticipants were identified. What might be called those
“average” (average social status, intact family, average
involvement with their environment) do quite well
intellectually as they age, maintaining most of their
abilities over the 14 years of testing (Schaie, 2013;
Gerstorf et al, 2019).
The “advantaged persons” (high social status,
with lives that require or allow them to keep
I. Concepts, theories, methods
13
learning new things) do even better, often increas-
ing their test scores over the years. The “spectators”
(average social status, intact family, passive partici-
pation in social activities, declining interest in new
learning situations) generally show a decline in abil-
ities. Finally, the “isolated older women” (poor,
unhappy, likely to be divorced or widowed, isolated
either by choice or circumstances) show the greatest
decline of all.
Favorable consequences for cognitive functioning
of higher education and favorable environments also
have been found in a study of Canadian World War
II veterans over periods as long as 45 years. Equally
interesting is the finding that favorable lifestyles
seem to enhance the development of flexible atti-
tudes. Such attitudes at midlife appear to be highly
predictive of the maintenance of intellectual func-
tioning into advanced old age. But do flexible atti-
tudes maintain intellectual functioning, or do high
levels of intellectual functioning encourage flexible
attitudes? The evidence indicates that flexible atti-
tudes affect intelligence. Correlational information
on the relationship between intelligence and flexibil-
ity has been gathered from longitudinal studies over
as long as 49 years. The data clearly show that corre-
lations between midlife flexibility and intelligence in
old age are much greater than those between intelli-
gence at midlife and flexibility in old age (Schaie,
2013).
Similar evidence indicates that lack of environmen-
tal stimulation leads to cognitive loss. For example, we
know that the greatest risk of cognitive decline occurs
among widowed women who have not pursued a
career and whose environmental stimulation has been
reduced by the death of their spouse.
By contrast, those who had marital partners with
high intellectual levels seem to benefit throughout
their marriage, and there have been several studies
that show demanding job environments tend to
enhance intellectual function in adults.
Certain personality dimensions have also been
identified as influencing level of intellectual func-
tioning in adults. The personality dimensions that
were related to high intellectual functioning included
high Untroubled Adequacy, low Conservatism, low
Group Dependency, and high Openness to new
experiences,
High ratings on the personality trait of openness to
new experience, and midrange ratings on the trait of
introversion
extroversion have also been implicated
as predictors of wisdom-related performance (Fig. 1.4).
14 1. History of adult cognitive aging research

100 FIGURE 1.4 Cumulative hazard

rate of significant decrement in differ-
90 Verbal meaning ent abilities occurring to successive
ages from 32 to 95 year. Source:
Cumulation percent 7-year decrement

80 Spatial orientation Schaie, K.W. (2013). Developmental

Inductive resoning influences on adult intellectual devel-
70 opment: The Seattle longitudinal
Number study (2nd ed.). New York: Oxford
60 University Press.
Word fluency

50

40

30

20

10

0
32 39 46 53 60 67 74 81 88 95
Age

Summary show a major factor called “general intelligence” or

“g.” Slightly different analyses have turned up two
1. Intelligence is usually defined as the ability to learn or major factors called crystallized and fluid
manipulate symbols. Intelligence is an inference from intelligence, about 6
12 factors called primary
competence demonstrated in several situations, but mental abilities, and 120 factors in the most complex
competence involves more than simply intelligence— model of intelligence.
motivation, for example. The particular abilities and 4. Intelligence tests provide information on the basic
motivations necessary to do well on IQ tests, which mechanics of intellectual functioning. For the
were designed to predict scholastic performance, may appraisal of performance of adults on everyday
bias these tests in favor of young people. problems these tests may need to be supplemented
2. Adult stages of intellectual development include the by measures of practical intelligence. However,
Achieving stage, which involves planning and performance on such measures is highly correlated
periodic assessment of programs designed to with the basic ability measures. Practical
achieve major goals in career, family, and life in intelligence represents a combination of basic skills
general. In the Responsible stage, abilities are applied to specific problems of everyday
applied to both short- and long-term concerns for experience.
family units, coworkers, and community groups. 5. Early cross-sectional studies of the relationship
Some progress to the Executive stage, planning and between age and intelligence showed decline after
assessment abilities are applied to the organizations the age of 20 or 30. Some tests, usually of simple
for which one is at least partly responsible. In later information processing abilities that make up fluid
life, the Reorganizational stage requires making intelligence, showed sharper decline than others;
choices mandated by the transition from work to tests of general information, vocabulary, and the
retirement and planning the economic bases for the like (crystallized intelligence) sometimes even
rest of life. The Reintegrative stage requires a increased.
reintegration of abilities, interests, and values, 6. Longitudinal studies, the first of which were
and the Legacy-leaving stage involves disposing of published around 1950, showed that the results
ones assets and creating legacies for the next obtained in cross-sectional studies had been
generation. seriously misinterpreted. Although there is
3. IQ tests were first developed by Alfred Binet in the significant average decline on most abilities
early 1900s Factor analyses of such tests typically beginning in the mid-60s, many individuals

I. Concepts, theories, methods

 References
typically show little or no decline in IQ scores until
advanced old age, and some brighter individuals, in
intellectual professions, may actually increase
intellectual abilities in midlife into early old age.
The cross-sectional studies had reflected not age
changes in intelligence but instead reflected birth
cohort or generational differences in average IQ
scores.
7. Interpretations of longitudinal studies have their
own difficulties, not the least of which is the
problem of dropouts and deaths, participants who
usually score somewhat lower than average on IQ
tests. Nevertheless, a “sizable” intellectual decline
that reduces the average of an older group to the
level surpassed by 75% of the younger group was
found for only two of seven measures in a major
longitudinal study for an additional two measures
only at age 81.
8. Why do some people decline in IQ scores over the
age of 60? One possible answer is normal or
pathological biological deterioration (of the brain
and nervous system,). However, diseases, especially
cardiovascular diseases, are a significant factor for
some individuals; life-threatening diseases are
known to produce a “terminal drop” in IQ scores.
Social isolation induced by retirement, deaths
among family and friends, and disabling diseases
can also reduce one’s IQ score. A number of
researchers have shown that appropriate training
can at least partially remediate declines.
References
Allard, M., Husky, M., Catheline, G., Pelletier, A.,
Dilharreguy, B., Amieva, H., . . . Swendsen, J. (2014).
Mobile technologies in the early detection of cognitive
decline. PLoS One, 9(12).
Bengtson, V. L., & Sattersten, R. A. (Eds.), (2016). Handbook of
theories of aging. (p. ). New York: Springer Publishing Co.
Berg, C. A., & Klacynski, P. A. (1996). Practical intelligence
and problem solving: Searching for perspectives. In F.
Blanchard-Fields, & T. Hess (Eds.), Perspectives on cognitive
change in adulthood and aging (pp. 323
357). New York:
McGraw Hill.
Binet, A., & Simon, T. (1905). Methodes nouvelle pour le
diagnostic du niveau intellectuel des anornaux. ^L’Annëe
Psychologique, 11, 191.
Bowen, C., & Staudinger, U. (2013). Relationship between
age and promotion orientation dependence perceived old-
er worker stereotypes. Journals of Gerontology, Series B,
Psychological Sciences and Social Sciences, 68, 59
63.
Carstensen, L. L. (1993). Motivation for social contact across
the lifespan: A theory of social-emotional selectivity. In J.
I. Concepts, theories, methods
15
Jacobs (Ed.), Nebraska symposium on motivation
(pp. 209
254). Lincoln: NE: University of Nebraska Press.
Carstensen, L. L, Gross, J. J., & Fung, H. H. (1997). The social
context of emotional experience. Annual review of gerontol-
ogy and geriarics, 17, 325
352.
Cohen, J. (1957). The factorial structure of the WAIS between
early adulthood and old age. Journal of Consulting
Psychology, 21, 283
290.
Commons, M. L., Sinnott, J. D., Richards, F. A., & Armon, C.
(Eds.), (1989). Beyond formal operations. In: Adolescent and
adult development models (2). (p. ). New York: Praeger.
Commons, M. L., Miller, L., & Giri, S. (2014). A model of
stage change explains the average rate of stage of develop-
ment and its relationship to the predicted average stage
(smarts). Behavioral Development Bulletin, 19, 1
11.
Commons, M. L., & Ross, S. (2008). Editors’ introduction to
the special issue on postformal thought and hierarchical
complexity. World Futures: The Journal of General Evolution,
64, 297
304.
Cunningham, W. R., & Owens, W. A., Jr. (1983). The Iowa
State Study of the adult development of intellectual abili-
ties. In K. W. Schaie (Ed.), Longitudinal studies of adult intel-
lectual development (pp. 20
39). New York: Guilford Press.
Diehl, M., Willis, S. L., & Schaie, K. W. (1995). Everyday prob-
lem solving in older adults: Observational assessment and
cognitive correlates. Psychology and Aging, 10, 477
491.
English, T., & Carstensen, L. (2014). Selective narrowing of
social networks across adulthood is associated with
improved emotional experience in daily life. International
Journal of Behavioral Development, 38, 195
202.
Flavell, J. H. (1963). The developmental psychology of Jean Piaget.
New York: Van Nostrand.
Flavell, J. H. (1970). Cognitive changes in adulthood. In L. R.
Goulet, & P. B. Baltes (Eds.), Life-span developmental psy-
chology: Research and theory. New York: Academic Press.
Galton, F. (1883). Inquiries into human faculty and its develop-
ment. New York: MacMillan.
Gerstorf, D., Hueler, G., Drewelies, J., Willis, S. L., Schaie,
K. W., & Ram, N. (2019). Adult development, and aging in
historical context. Washinton, DC: American Psychologist.
Gerstorf, D., Ram, N., Hoppmann, C., Willis, S. L., & Schaie,
K. W. (2011). Cohort differences in cognitive aging and
terminal decline in the Seattle Longitudinal Study.
Developmental Psychology, 47, 1026
1041.
Gerstrof, D., Hulur, G., Drewelies, J., Willis, S.L., Schaie, K.
W., & Ram, N. (2020). Adult development and aging in
historical context. American Psychologist. In press.
Hagestad, B. O., & Neugarten, B. L. (1985). Age and the life
course. In R. Binstock, & E. Shanas (Eds.), Handbook of
aging and the social sciences (2nd ed., pp. 35
81). New
York: Van Nostrand Reinhold.
Hanford, G. H. (1991). Life with the SAT: Assessing Our Young
People and Our Times. College Board Publications.
Heidemeier, H., & Staudinger, U. (2015). Age differences in
achievement goals and motivational characteristics of work
in an ageing workforce. Ageing and Society, 35, 809
836.
16 1. History of adult cognitive aging research
Horn, J. L., & Hofer, S. M. (1992). Major abilities and devel-
opment in the adult period. In R. J. Sternberg, & C. A.
Berg (Eds.), Intellectual development (pp. 44
99).
Cambridge: Cambridge University Press.
Humphreys, L. G., & Parsons, C. K. (1979). Piagetian tasks
measure intelligence and intelligence tests assess cognitive
development: A reanalysis. Intelligence, 3, 369
382.
Lawton, M. P., & Brody, E. M. (1969). Assessment of older
people: Self-maintaining and instrumental activities of
daily living. Gerontologist, 9, 179
185.
Marsiske, M., & Margrett, J. A. (2006). Everyday problem
solving and decision-making. In J. E. Birren, & K. W.
Schaie (Eds.), Handbook of the psychology of aging (6th ed.,
pp. 315
342). San Diego, CA: Academic Press.
Marsiske, M., & Willis, S. L. (1995). Dimensionality of every-
day problem solving in older adults. Psychology and Aging,
10, 269
283.
Owens, W. A., Jr. (1953). Age and mental abilities: A longitu-
dinal study. Genetic Longitudinal Monograph, 48, 3
54.
Piaget, J. (1972). Intellectual evolution from adolescence to
adulthood. Human Development, 15, 1
12.
Pilail, J. A., Hall, C. B., Dickson, D. W., Busche, H., Lipton,
R. B., & Verghese, J. (2011). Association of Crossword
Puzzle Participation with Memory Decline in Persons
Who Develop Dementia. Journal of the International
Neuropsychological Society, 17, 8.
Salthouse, T. A. (1999). Pressing issues in cognitive aging.
In N. Schwarz, D. C. Park, B. Knaüper, & S. Sudman
(Eds.), Cognition, aging, and self-reports (pp. 185
198).
Psychology Press/Erlbaum (UK) Taylor & Francis.
Schaie, K. W. (1977
78). Toward a stage theory of adult cog-
nitive development. Journal of Aging and Human
Development, 8, 129
138.
Schaie, K. W. (1984). Midlife influence on intellectual func-
tioning in old age. Iˆnternational Journal of Behavioral
Development, 7, 463
478.
Schaie, K. W. (1985). Manual for the Schaie-Thurstone Mental
Abilities Test (STAMAT). Palo Alto, CA: Consulting
Psychologists Press.
Schaie, K. W. (1989). Perceptual speed in adulthood: Cross-
sectional and longitudinal studies. Psychology and Aging, 4,
443
453.
Schaie, K. W. (1994). The course of adult intellectual develop-
ment. American Psychologist, 49, 304
313.
Schaie, K. W. (1996). Intellectual development in adulthood: The
Seattle longitudinal study. New York: Cambridge
University Press.
Schaie, K. W. (2013). Developmental influences on adult intellec-
tual development: The Seattle longitudinal study (2nd ed.).
New York: Oxford University Press.
Schaie, K. W., & Willis, S. L. (1999). Theories of everyday
competence and aging. In V. L. Bengtson, & J. E. Birren
(Eds.), Handbook of theories of aging (pp. 174
195). New
York: Springer Publishing Co.
Schaie, K. W., & Willis, S. L. (2000). A stage theory model of
adult cognitive development revisited. In R. Rubinstein,
I. Concepts, theories, methods
M. Moss, & M. Kleban (Eds.), The many dimensions of aging:
Essays in honor of M. Powell Lawton (pp. 175
193). New
York: Springer Publishing Co.
Schaie, K. W., & Zuo, Y. L. (2001). Family environments and
adult cognitive functioning. In R. L. Sternberg, & E.
Grigorenko (Eds.), Context of intellectual development
(pp. 337
361). Hillsdale, NJ: Erlbaum.
Schaie, K. W. (2005). What can we learn from longitudinal
studies of adult intellectual development? Research in
Human Development, 2, 133
158.
Schaie, K. W., & Carstensen, L. L. (Eds.), (2006). Social struc-
tures, aging and self-regulation in the elderly. (p. ). New
York: Springer Publishing Co.
Schaie, K. W. (2008). A lifespan developmental perspective of
psychological aging. In K. Laidlaw, & B. G. Knight (Eds.),
The Handbook of emotional disorders in late life: Assessment
and treatment (pp. 3
32). Oxford, UK: Oxford University
Press.
Schaie, K. W. (2009). “When does age-related cognitive
decline begin?” Salthouse again reifies the cross-sectional
fallacy. Neurobiology of Aging, 30, 528
529.
Scheibe, S., English, T., Tsai, J., & Carstensen, L. (2013).
Striving to feel good: Ideal affect, actual affect and their
correspondence across adulthood. Psychology and Aging,
28, 160
171.
Schooler, C., Mulatu, M. S., & Oates, G. (1999). The continu-
ing effects of substantively complex work on the intellec-
tual functioning of older workers. Psychology and Aging,
14, 483
506.
Sinnott, J. D. (1996). The developmental approach; Post for-
mal thought as adaptive intelligence. In F. Blanchard,
Fields, & T. Hess (Eds.), Perspectives on cognitive change in
adulthood and aging (pp. 358
383). New York: McGraw-
Hill.
Smith, H., Staudinger, U. M., & Baltes, P. B. (1994).
Occupational settings facilitating wisdom-related knowl-
edge. Journal of Consulting and Clinical Psychology, 62,
989
999.
Staudinger, U., & Glueck, J. (2011). Psychological wisdom
research: Commonalities and differences in a growing
field. In S. Fiske, D. Schacter, & S. Taylor (Eds.), Annual
Review of Psychology (62, pp. 215
241).
Sternberg, R. J., & Lubart, T. I. (2001). Wisdom and creativity.
In J. E. Birren, & K. W. Schaie (Eds.), Handbook of the
psychology of aging (5th ed.). San Diego, CA: Academic
Press.
Thurstone, T. G. (1962). Primary mental abilities for Grades
9
12. Chicago: Science Research Associates.
Wechsler, D. (1972). “Hold” and “Don’t Hold” tests. In S. M.
Chown (Ed.), Human aging. New York: Penguin.
Wechsler, D. (1997). Wechsler intelligence scale (3rd ed.). San
Antonio, TX: Psychological Corporation.
Westerhof, G. J., Miche, M., Brothers, A. F., Barrett, A. E.,
Diehl, M., & Montepare, J. M. (2014). The influence of sub-
jective aging on health and longevity: a meta-analysis of
longitudinal data. Psychology and Aging, 29(4).
 Further reading 17

Wettstein, M., Wahl, H., & Diehl, M. (2014). A multidimen-
sional view of out-of-home behaviors in cognitively unim-
paired older adults: Examining differential effects of
socio-demographic, cognitive, and health related predic-
tors. European Journal of Ageing, 11, 141
153.
Willis, S. L. (1996). Everyday problem solving. In J. E. Birren,
& K. W. Schaie (Eds.), Handbook of the psychology of aging.
San Diego. CA: Academic Press.
Willis, S. L., & Schaie, K. W. (1986). Practical intelligence in
later adulthood. In R. J. Sternberg, & R. K. Wagner (Eds.),
Practical intelligence: Origins of competence in the everyday
world (pp. 236
268). Cambridge: Cambridge University
Press.
Willis, S. L. (1997). Everyday cognitive competence in the elderly:
Conceptual issues and empirical findings. Gerontologist, 36,
595
601.
Willis, S. L. (1989). Cohort differences in cognitive aging: A
sample case. In K. W. Schaie, & C. Schooler (Eds.), Social
structure and aging: Psychological processes (pp. 94
112).
Hillsdale, NJ: Erlbaum.
Willis, S. L., & Schaie, K. W. (2005). Cognitive trajectories in
midlife and cognitive functioning in old age. In S. L.
Willis, & M. Martin (Eds.), Middle adulthood: A lifespan per-
spective (pp. 243
276). Thousand Oaks, CA: Sage.
Wissler, C. (1901). The correlation of mental and physical tests.
The Psychological Review: Monograph Supplements, 3(6), i
62.
Further reading
Sternberg, R. J., & Wagner, R. K. (Eds.), (1985). Practical
intelligence: Origins of competence in the everyday world.
(pp. 236
268). Cambridge: Cambridge University Press.

I. Concepts, theories, methods

This page intentionally left blank
 C H A P T E R

2
Measurement and models for multi-timescale
psychological processes in aging research
Steven M. Boker and John R. Nesselroade
Department of Psychology, The University of Virginia, Charlottesville, VA, United States

O U T L I N E

Introduction 19 Challenges 28
Matching the research design to the question 20 Planning for the long term 28
Privacy, linking, and data sharing 28
Modeling review 21
Incompleteness and attrition 29
Cross-sectional models 21
Reliability versus sensitivity 30
Autoregressive and crosslag models 22
Measurement invariance over time 30
Latent growth curves and multilevel models 22
The changing data landscape 24 Emerging opportunities 30
Intrinsic capacity and functional ability 31
Multi-timescale processes and research designs 24
Measuring and modeling multiscale processes from Conclusion 31
intensive time series 25
References 31
Dynamical systems and attractors 25
Dynamical systems models in aging 26

Introduction Stawski, 2016) from previous editions of this hand-

When Baltes (1987) explored the relations between
general developmental theory and the life span orien-
tation, the several propositions he advanced
highlighted a clear need for new analytic methods and
procedures (and data) to support apt lines of inquiry.
Since then, much has been written about appropriate
methods for use in modeling and understanding pro-
cesses involved in aging. As background overview
reading, we highly recommend starting with the meth-
ods chapters (Ferrer & Ghisletta, 2011; MacDonald &
book. Those chapters provide in-depth views of the
most commonly used methods, models, and caveats
that apply when longitudinal data are sparsely mea-
sured in time, for example, 4
10 occasions of measure-
ment per person. The current chapter extends that
view by providing more detail about analyses that
apply when the data are densely measured in time
(Walls & Schafer, 2005), such as from burst measure-
ment (Nesselroade, 1991a; Stawski, MacDonald, &
Sliwinski, 2015), ecological momentary assessment
(Shiffman, Stone, & Hufford, 2008), smartphones

Handbook of the Psychology of Aging.

DOI: https://doi.org/10.1016/B978-0-12-816094-7.00011-8 19 © 2021 Elsevier Inc. All rights reserved.
20 2. Measurement and models for multi-timescale psychological processes in aging research
(comScore, 2014), and wearable devices (Patel, Asch, &
Volpp, 2015).
The chapter will first provide a brief tutorial for
some of the fundamental concepts involved in match-
ing analyses and data to the type of question that one
wishes to understand. We will then provide some
pointers to sources for more detailed descriptions of
well-known methods that have been usefully applied
in the context of aging research. We will end the first
section with a look at how the landscape of available
data has changed since the previous edition of this
handbook—a shift towards intensive longitudinal time
series measurement and individual-specific models.
The second section will address multi-timescale mod-
els and analyses. This section will review terminology
and concepts that distinguish dynamical systems mod-
els from models for intraindividual variability and
change that rely on the concept of a trait-state distinc-
tion. Model fitting methods for fitting dynamical sys-
tems models will be reviewed and two example multi-
timescale models discussed. The third section will dis-
cuss a variety of challenges that arise when planning
and analyzing longitudinal studies. The final section
discusses an emerging opportunity that is amenable to
multi-timescale dynamical systems modeling and that
would be difficult to instantiate without the newly
emerging analysis techniques discussed in the present
chapter.
Matching the research design to the question
Consider three research questions which illustrate
important differences in research designs: (1) a policy
maker wants to know how the incidence of dementia
is changing in the population in order to plan for con-
struction of long-term care facilities; (2) a clinician
wants to be able to diagnose the likelihood of the
development of the onset of dementia for each client
that walks into her clinic; and (3) a psychologist wants
to evaluate the effectiveness of an intervention for indi-
viduals with mild cognitive impairment. Each of these
questions requires a different experimental design in
order to obtain meaningful answers.
The first question can be answered using a multi-
wave cross-sectional design. At wave one, a diagnostic
battery is given to a large representative sample of
individuals who are within a wide age range. Five
years later at wave two, the same diagnostic battery is
given to a new large representative sample of indivi-
duals within the same age range. From these data, one
can estimate the risk of dementia in the population
I. Concepts, theories, methods
given each chosen age at wave one and also estimate
how that risk is changing over the 5-year interval
between wave one and wave two. This analysis is
well-suited to the policy maker’s question, since it
allows the prediction of total need for long-term care
facilities at a future time.
However, the results of the first analysis are close to
useless when the clinician wishes to predict the risk of
a particular client developing dementia within some
chosen interval of time. The best one can say is that
given the age of the client there is a risk of dementia at
the current moment. This provides less information
than simply giving the client the diagnostic battery at
the current moment. In order to answer the clinician’s
question about the development of dementia in a par-
ticular client, the clinician needs to know the history of
this client’s scores on the diagnostic battery and how
they compare to the history of scores of a representa-
tive sample of individuals who eventually did and did
not develop dementia. To provide an answer to the
clinician’s question, one needs longitudinally linked
data from a large sample of individuals, including the
client in question. One then needs to use the trajectory
of diagnostic scores from the client to select those in
the study with similar trajectories and then calculate
the risk of onset of dementia at multiple time lags for
those individuals. Then, one can ask: “Given the cur-
rent client’s trajectory, what is the expected risk of
dementia at, for example, 1 year, 5 years, 10 years, and
20 years?” The reasoning behind this analysis is that
different trajectories of diagnostic scores may exhibit
different risks of dementia and this risk may not neces-
sarily vary parametrically with parametric changes in
trajectories.
The third question is even more difficult to answer.
In order to understand the effectiveness of an interven-
tion, one must take into account the possibility that the
intervention is differentially effective for each individ-
ual. Random assignment of the intervention to a sub-
sample of people who were in the study that was used
to answer the clinician’s question and then estimating
the average effect of the intervention will not provide
a definitive answer to the psychologist’s question. The
intervention and placebo may be differentially effective
for individuals. Ideally, the psychologist would want
each individual to be her or his own control so that the
intervention could be given, then withdrawn, then
given again. But in this case, as is common in aging
research, an A
B
A research design is not possible.
The fallback analysis is to parameterize individuals’
trajectories into two parts—pre and post interven-
tion—so that a few coefficients allow for individual
 Modeling review
differences in trajectories with a boundary at the time
of intervention. Then, one may estimate the effect of
an intervention using a mixed effects model or multi-
part latent growth curve (LGC) (see the section below)
so that each individual’s trajectory can be summarized
as a few random variables and the presence of the
intervention moderates these variables. This analysis
can help understand the mean intervention effect and
the variance in individuals’ response to the
intervention.
These three illustrative examples are given in the
hope of helping researchers avoid mistakes in experi-
mental designs and to avoid fallacies in meta-analyses.
All too often, a naive meta-analysis is undertaken with
the thought that more data is always better. Large har-
monized longitudinal data sets [e.g., Integrative
Analysis of Longitudinal Studies of Aging (IALSA)
and Gateway to Global Aging] are promising opportu-
nities, but come with their own risks. Selecting a
research design that fits the question one wishes to
answer is the most important and influential analytic
decision that one can make. The next most important
decision is selecting a method of aggregation that fits
the question one wishes to answer. Only once these
two decisions have correctly been made does more
data imply more precision in answering the desired
question. A very precise answer to the wrong question
must be recognized and avoided.
Modeling review
This section will present some of the main catego-
ries of methods that are currently being used to model
age-related processes. Each method will be briefly
described along with its assumptions, strengths, weak-
nesses, and types of questions for which the model is
appropriate. The models will be presented in relation
to the following five categories of data to which they
can be applied:
1. Cross-sectional designs in which each individual
belongs to an age category.
2. Longitudinal designs where each individual has data
from between 3 and 20 occasions of measurement.
3. Burst designs where several short bursts of
longitudinal samples are separated by longer
intervals of time.
4. Intensive longitudinal designs where each individual
contributes between 20 and thousands of
observation occasions.
I. Concepts, theories, methods
21
5. Intensive burst designs where each individual burst of
data comprises an intensive longitudinal sample.
The first three of these research designs have been
covered extensively in previous editions of this hand-
book and so we will cover models that are restricted to
these designs only in brief overview. The last two
research designs will be introduced in the next main
section entitled “The Changing Data Landscape” and
models especially appropriate for these two designs
will be introduced in the subsequent section entitled
“Measuring and Modeling Multiscale Processes from
Intensive Time Series.”
Cross-sectional models
Cross-sectional models take on two forms: (1) youn-
ger versus older group designs and (2) age-as-
moderator designs. Models for younger versus older
group data are popular, but have severe restrictions
when it comes to interpreting the results. Most often,
ANOVA-style general linear models (see Maxwell &
Delaney, 2003, for a comprehensive treatment) are fit
to these data to determine if there are age-related dif-
ferences between the groups. However, individuals are
not randomly assigned to age groups and so attrition
and other selection effects cannot be ruled out when
using this type of analysis.
Selection is a serious problem that cannot be
ignored since there is at least one selection that is
baked into these data: The participants in the older
group can only include those who have lived to that
age. Thus if the experiment had been run as a longitu-
dinal study, some of those in the younger group
would not be alive to participate in the older group.
To the extent that the variable of interest is related to
longevity, the analysis will be biased by this selection
effect.
Age-as-moderator designs include individuals in a
range of ages and then use age as a predictor of the
variable or variables of interest. While this may pro-
vide a more nuanced view of age-related differences, it
suffers from the same problems due to selection effects
as do the younger versus older group designs. When
estimating between-group differences, random assign-
ment helps prevent selection effects. But age is never
randomly assigned and so to the extent that participa-
tion is related to age and age is related to the variable
of interest, the results of cross-sectional studies pro-
vide inherently biased results.
In order to interpret results from cross-sectional
designs as being due to age-related processes, one
22 2. Measurement and models for multi-timescale psychological processes in aging research
must assume that selection into the experiment is
unrelated to age and longevity, and that between-
person differences are representative of within-person
differences. To the extent that these two assumptions
do not hold, conclusions based on cross-sectional
designs are likely to be either biased or incorrect.
However, if one wishes to estimate prevalence of a
phenomenon in the population relative to age, then
cross-sectional designs can be appropriate. This is
because the prevalence of the age of participants is
part of the research question. If a sample of older indi-
viduals is representative of the population of older
individuals, then one may make an inference to that
population. It is when one wishes to directly compare
younger and older individuals that cross-sectional
designs are likely to fail to meet the necessary
assumptions.
Autoregressive and crosslag models
Autoregressive and crosslag models are a class of
models in which variables measured at an earlier time,
t, are used to predict variables from the same person
at a later time, t 1 Δt. The discrete amount of time, Δt
that elapses between the observations is why this type
of model is termed a discrete time model. Suppose two
variables, x and y, are measured at two occasions in a
sample of participants. A combined autoregressive and
crosslag model may be expressed as
xðt 1 ΔtÞ 5 bxx xðtÞ 1 bxy yðtÞ 1 exðt1ΔtÞ
ð2:1Þ
yðt 1 ΔtÞ 5 byx xðtÞ 1 byy yðtÞ 1 eyðt1ΔtÞ :
This model has the intuitive appeal that it is
expressed as a set of linear regressions that can be eas-
ily estimated by open source structural equation
modeling (SEM) software such as Lavaan (Rosseel,
2012) or OpenMx (Neale et al., 2016).
The usual linear regression assumptions apply to
autoregressive and crosslag models, but there are a
few extra implications of this model. The residuals ex
(t 1 Δt) and ey(t 1 Δt) are assumed to be unrelated to
the predictors, but since the predictors are the same
variables at an earlier time, the residuals must have a
form in which they are unrelated to time. Thus these
residuals are often called “shocks” or “innovations.”
A second implication of this model is often over-
looked. Suppose that the process is continuous. That
is, if one observed the process at time t 1 Δt/2, there
would be a value for that process. Then, if the assump-
tions of linear regression hold, if we were to estimate
the model using the data from times t and t 1 Δt/2 or
I. Concepts, theories, methods
at the times t 1 Δt/2 and t 1 Δt, the discrete elapsed
time would be half as much, Δt/2, and the values of
the b coefficients would be greater than they had been
when the model was estimated with an elapsed time
of Δt. If the process is continuous and linear regression
assumptions hold, then coefficients estimated with the
autoregressive/crosslag model are a function of both
the discrete elapsed time and the true effects of the
variables on themselves and one another. A common
mistake often observed in the literature is that the
elapsed time is set to “1” and ignored. This means that
two experiments with different intervals between
observations will not replicate one another if each
author arbitrarily sets their interval to “1.” Solutions to
this problem are discussed below in the subsection
“Continuous Time Models.”
A third implication of the autoregression/crosslag
model is that of centering. Consider the following sce-
nario: Two time series x and y are simulated by draw-
ing 100 observations from a random normal
distribution with mean 0.0 and standard deviation 1.0.
Clearly the expected values of all four of the b coeffi-
cients in Eq. (2.1) would be 0. There should be no asso-
ciation between x and itself over time or between x
and y over time. But now add 10 to every value of x
and every value of y so that the mean of x and y are
both 10. Now the expected values of all four b coeffi-
cients approach 0.5. Thus the mean value of time series
can have an effect on the coefficients of the autoregres-
sion/crosslag model. The autoregression/crosslag
model is a model of dynamics centered about a point
equilibrium. We will discuss this in greater detail in
the section “Dynamical Systems and Attractors.”
Latent growth curves and multilevel models
LGC models (McArdle & Epstein, 1987) are a form
of multilevel (or mixed effects) model (Mehta & West,
2000) applied to longitudinal data with a few (3
20)
occasions of measurement per person. Fig. 2.1A pre-
sents a path diagram of a latent intercept and slope
model. In this example, each individual is measured at
five occasions so that the data matrix is arranged as
persons 3 occasions. If the loading matrix L is held
constant to
2 3
1 0
6 1 Δt 7
6 7
L56 6 1 2Δt 7
7 ð2:2Þ
4 1 3Δt 5
1 4Δt

A B
1 1
mI mS
I S
L L
x1 x2 x3 x4 x5 x1
u1 u2 u3 u4 u5 u1
then the latent scores for I and S represent the inter-
cept and slope for each row in the data. The variances
of I and S are the same as would be obtained from a
multilevel model where the variable x is predicted by
the elapsed time from the first occasion of measure-
ment and grouped by participant. Similarly, the mean
of the intercept mI and mean of the slope mS corre-
spond to the fixed effects from the multilevel model.
There are two main advantages of LGC models
over multilevel models. The first is that missing data
can be handled appropriately without need for impu-
tation if full information maximum likelihood (FIML)
is used as the optimization method (Duncan &
Duncan, 1994; McArdle & Hamagami, 1992). It is
important to remember that missing data methods
such as FIML assume that missingness can be
accounted for by the variables within the model. Thus
FIML does not fully account for missingness due to
mortality unless a model for mortality is included in
the full model.
The second main advantage is that one may free the
nonzero elements of the second column of the L matrix
in order to estimate an arbitrary nonlinear curve. This
model is called a free basis LGC model. One may then
test whether nonlinearity improves the model fit since
the fixed basis model is nested within the free basis
model. One should note that the row in which there is
a zero in column two of the L matrix is the occasion
of measurement at which the intercept obtains. For
instance, when there is growth to an asymptote it may
be appropriate to assign the last row to contain the
zero so that the intercept is at the last occasion of
measurement.
Often, a third latent variable is added by squaring
the second column of the L matrix and adding it as a
I. Concepts, theories, methods
Modeling review 23
FIGURE 2.1 Path diagrams of a latent
growth curve (LGC) model for data with five
occasions of measurement per person. (A) The
two latent variables I and S, represent the
mI mS mC
latent intercept and slope for each individual.
The loadings connecting I and S to the five x
variables are fixed to values as shown in
I S C
Eq. (2.2). (B) An intercept, slope, and curva-
ture model where the L matrix now has three
columns.
x2 x3 x4 x5
u2 u3 u4 u5
third column. The resulting LGC model is presented in
Fig. 2.1B where C represents the amount of quadratic
curvature. It may also be useful to free the third col-
umn in this model’s L matrix to see if a curve other
than quadratic fits the data better. This type of free
basis function is particularly useful in aging research
where curves such as a negative exponential might fit
better or be more theoretically appealing than qua-
dratic curves. An example of the use of a negative
exponential curve would be a transition from a
stable functional ability in midlife to a linear decline in
old age. For instance, grip strength may be relatively
stable until a transition age when it begins to decline.
Individual differences in the negative exponential
curve would indicate individual differences in how
quickly the transition from maintenance to decline
occurs.
Since LGC models are standard SEM models, one
may also wish to predict the latent intercept and slope
from covariates. For instance, one might wish to pre-
dict cognitive decline given age and education. This is
similar to a multilevel model where there are second
level predictors. Alternatively, the latent intercept and
slope may be used to predict other outcome variables
in the data.
At this point the chapter will switch from focusing
on methods and models where longitudinal data are
limited to longitudinal panel studies or short time
series and introduce models for multi-timescale
dynamics. Until recently multi-timescale models
have not been necessary, since only recently have
intensive longitudinal data and multi-timescale burst
data become widely available for analysis. But
as Dylan famously sang, “The times they are
a-changing.”
24 2. Measurement and models for multi-timescale psychological processes in aging research
The changing data landscape
In order to understand the processes operating across
the life span, the need for longitudinal within-person
data collection and analyses has long been known (e.g.,
Baltes & Nesselroade, 1979a, 1979b). However, until
recently longitudinal data have been difficult and time
consuming to acquire. Decades-long longitudinal studies
are slow to come to fruition (Baltes & Mayer, 2001;
Baltes, Mayer, Helmchen, & Steinhagen-Thiessen, 1993;
Salthouse & Davis, 2006; Schaie, 1996, 2005) but acceler-
ated longitudinal designs can help mitigate this down-
side (Joiner, Bergeman, & Wang, 2018). Recent efforts
have begun to harmonize aging-related longitudinal
studies across projects and countries. For instance, the
Gateway to Global Aging Data (https://g2aging.org) has
harmonized the data from the US Health and Retirement
Study (Burkhauser & Gertler, 1995) and similar longitu-
dinal surveys from nine other countries in Europe, Asia,
and Central America. Another harmonization effort
comes from the IALSA project (http://www.ialsa.org).
Efforts such as these that provide harmonization and can
tie together multiple longitudinal studies are expected to
multiply over the coming years. One must be careful
when using these data sources since it is not always clear
that variables are comparable across studies even after
significant resources have been spent in harmonizing the
coding of the variables. Again we emphasize that the
data and analysis must be appropriate to the question of
interest to the user. All too often the availability of data
drives the question and not the other way around.
While collection of meaningful longitudinal data con-
tinues to pose challenges, recent developments in wear-
able sensors (e.g., Fan, Ruiz, Gong, & Lach, 2016;
Varatharajan, Manogaran, Priyan, & Sundarasekar, 2017),
computer vision (e.g., Dean, Samson, Newberry, &
Mittal, 2018; Theobald, Matthews, Cohn, & Boker, 2007),
and smartphones and smartwatches have radically chan-
ged the possibilities for data acquisition over short time-
scales. Personal and smart devices are now commonly
used for Ecological Momentary Assessment (Shiffman
et al., 2008; Stone & Shiffman, 1994). The market penetra-
tion of smartphones in particular has happened incredi-
bly quickly. The 2013 Pew Foundation’s Tracking for
Health study (Fox & Duggan, 2013) reported that 69% of
Americans track some form of personal health data and
21% of Americans do so on a personal digital device. As
of February 2018, more than 77% of the US population
owned a smartphone (Pew Research, 2018) up from
about 50% in 2014 (comScore, 2014). As of 2018, about
46% of those over 65% and 73% of those between 50 and
I. Concepts, theories, methods
65 own a smartphone. There is almost no difference in
smartphone ownership between white, Black, and
Hispanic segments of the population. Smartphone own-
ership (Pew Research, 2018) is lower among those with
income less than US$ 30,000 (67%) and those who did
not finish high school (57%). Thus while smartphone
owners do not yet comprise a representative sample of
the population, the rapid reduction in cost and increased
availability of these devices is projected to rapidly
improve the representativeness of smartphone samples.
The changing landscape of data collection implies
that better models for integrating short-term dynamics
and long-term effects of aging are needed. Models that
estimate processes that evolve over several timescales
are known as models of multiscale dynamics (Boker,
Molenaar, & Nesselroade, 2009; Nesselroade, 1991b).
Multi-timescale processes and research designs
To illustrate what we mean by multiscale processes,
consider the well-known phenomenon of speed
accu-
racy trade-off (Wickelgren, 1977; Woodworth, 1899).
First, if we measure some phenomenon such as typing
speed and typing accuracy using a cross-sectional design,
we will obtain a positive between-persons correlation as
shown in Fig. 2.2A. In general, faster typists tend to
make fewer mistakes. If on successive trials, we ask peo-
ple to either focus on speed or focus on accuracy, we
obtain the classic negative relationship between speed
and accuracy as shown in Fig. 2.2B. Within-person and
over short timescales, the faster one attempts to type, the
less accurate is the result. These two graphs illustrate
that there can be surprisingly different results from
cross-sectional and longitudinal research designs.
But now consider Fig. 2.2C where a single naive typist
is given training over a longer period of time. While there
are day-to-day differences in typing speed and accuracy,
over the course of the longitudinal data set we see train-
ing effects that can be ascribed to learning and which
result in both improved speed and accuracy. This illus-
trates that one’s results not only depend on within- versus
between-persons designs, but also upon the timescale.
The same individual can exhibit both a positive relation-
ship between speed and accuracy as well as a negative
relationship between speed and accuracy depending on
the timescale over which they are measured.
This problem of timescales is inherent in many of the
phenomena of interest to gerontopsychologists. For
instance, there is ambiguity in the term “resilience.” This
 Multi-timescale processes and research designs
A B
100
100
80
80
60
60
Accuracy
Accuracy
40
40
20
20
0
0 20 40 60 80 100 0 20
Speed
FIGURE 2.2 Simulated data illustrating aspects of the speed
accuracy trade-off in typing. (A) Cross-sectional designs find a positive
between-persons relationship between speed and accuracy. (B) Short timescale within-persons designs find a negative relationship: the faster
one attempts to type, the more mistakes are likely to be made. (C) A longer timescale exposes the effects of learning, once again finding a pos-
itive relation between speed and accuracy as an individual develops typing skill.
ambiguity is in part due to the different timescales at
which resilience is considered to be occurring. Some con-
sider resilience to be a process that occurs over dec-
ades—effects of early adversity are not seen in later life
and so an individual is termed resilient. But others con-
sider resilience to be a short-term regulatory process in
which adverse events are managed and regulated by a
short-term dynamic. We take no stance on which of these
is a better definition. In fact, resilience may be a multi-
timescale process and both definitions may simulta-
neously be observed in the same individual.
In order to collect data that can be used to test theo-
ries of multi-timescale processes, burst designs are
recommended (MacDonald & Stawski, 2016). A burst of
data refers to within-person observations collected at
short intervals between observations on the scale of milli-
seconds to days over a period of time ranging from min-
utes to months. Then an amount of time is allowed to
elapse, typically in the range of weeks to years before the
next burst is obtained. In this way one efficiently obtains
data that allow estimation of both short-term dynamics
and long-term change in the process of interest.
Measuring and modeling multiscale processes
from intensive time series
The past two decades have seen growth in interest in
dynamical systems models of processes involved in
aging. This has helped spur the development of a sub-
stantial body of techniques for estimating such models
from data. The language of dynamical systems theory
has helped disambiguate the many types of age-related
I. Concepts, theories, methods
25
C
100
80
60
Accuracy
40
20
0
40 60 80 100 0 20 40 60 80 100
Speed Speed
change that can occur in important psychological pro-
cesses including affect regulation (Butler, 2017; Chow,
Ram, Boker, Fujita, & Clore, 2005; Cole, Bendezú, Ram,
& Chow, 2017), social interaction (Ashenfelter, Boker,
Waddell, & Vitanov, 2009; Bales, 2017; Hamaker, Zhang,
& Maas, 2009), grief (Bisconti, Bergeman, & Boker, 2004),
and resilience (Boker, Montpetit, Hunter, & Bergeman,
2010; Montpetit, Bergeman, Deboeck, Tiberio, & Boker,
2010; Ong, Bergeman, & Boker, 2009) to name a few.
Dynamical systems and attractors
For those unfamiliar with dynamical systems ideas,
we next provide a few definitions. A system is a group of
variables that have some natural connectedness and that
may influence each other over time. For instance, a sys-
tem could be a person’s health, activity, and diet. Or a
system could be measures of interactions between family
members over time. In the end it is up to the researcher
to decide what is part of the system and what is exoge-
nous, since at the limit we are all systems within systems
within systems up to the system of the universe. Good
advice is to define your system of interest to include suf-
ficient variables to capture the complexity of possible
between-systems differences in within-system trajectories
over time. If the system is defined as a person, then the
logic of interindividual differences in intraindividual
change (Nesselroade, 1991a) maps directly onto the logic
of dynamical systems. The state of a system is the value
of all of its variables at one moment in time.
A dynamical system is a system that evolves over time
and where the variables defined as being part of the
26 2. Measurement and models for multi-timescale psychological processes in aging research
system have some effect (i.e., exhibit some force) on one
another. A deterministic system is one where all of the evo-
lution of the system can be accounted for by the initial
values of the variables and the effect of the variables on
each other over time. A mixed deterministic and stochas-
tic system has both deterministic effects as well as ran-
dom components. These random components are often
considered to be unmeasured exogenous effects similar to
unique factors in a factor model. When estimating models
to account for data from human systems, it is likely that
one will always need stochastic components. Separating
these stochastic components from deterministic effects is
an important job of the optimizer used in estimating para-
meters when fitting a dynamical systems model to data.
Two further terms are useful when talking about
dynamical systems: equilibrium and attractor. A
stable equilibrium is the set of states to which a system
returns when exogenous effects are zero. For instance,
a physical system might be defined as the position,
velocity, and acceleration of a marble when it is rolled
into a round-bottomed mixing bowl. As long as no
movement of the bowl or other interference occurs, the
marble will roll around in the bowl until it comes to
rest at the bottom. When in this equilibrium state, the
marble is at the point at the bottom of the bowl where
the velocity and acceleration are zero. This is called a
stable point equilibrium because if the ball is nudged
away from this equilibrium, there is just one state to
which the ball will roll and come to rest: the point
equilibrium state.
An attractor encompasses not only the set of equilib-
rium states, but also the dynamics by which the sys-
tem returns to equilibrium when the system is
perturbed away from equilibrium. One way to think
about this is as the shape of the mixing bowl: some
bowls might be shallow and others might have steep
walls. This change in shape of the attractor is also
inherent in the deterministic relationship between the
variables. For instance in the case of a parabolic-
shaped bowl, there is a linear relationship between the
position, velocity, and acceleration of the marble that
determines what trajectory the marble will take no
matter where it is initially placed on the bowl.
Dynamical systems models in aging
There are many types of dynamical systems models
that can be used to specify and test theories in aging
research. In fact, some of the most popular models in
aging research can be thought of as special cases
of dynamical systems models in which there is
I. Concepts, theories, methods
age-related change in equilibrium and in which the
dynamics are nondeterministic (i.e., stochastic). For
instance, consider the LGC model in Fig. 2.3A. Each
person has scores on x at five occasions of measure-
ment and latent scores for the intercept and slope that
estimate the change in the equilibrium for that person.
The residual scores have a mean of zero by the
assumptions of regression and so they represent dis-
placement from their equilibrium. The displacement
from equilibrium is stochastic since the residual scores
do not have time-dependency (i.e., they are uncorre-
lated). This example LGC is a multiscale model since
the change in equilibrium is modeled as happening at
a longer timescale than the short-term stochastic
dynamics of the displacement from equilibrium.
While changes in equilibrium value are important
to model, the short-term dynamics around the equilib-
rium are also of interest. Much of dynamical systems
modeling is concerned with how to represent and esti-
mate the time dependency of the displacements from
equilibrium. These models are written in two different
forms: discrete time and continuous time. The most
common discrete time example is the first-order auto-
regressive model which can be written as
ut 5 but21 1 vt21 ; ð2:3Þ
where b is the autoregressive coefficient and vt21 is the
residual at time t 2 1. This first order autoregressive
model is illustrated for the changing equilibrium con-
dition in Fig. 2.3A.
Eq. (2.3) is a form of discrete time model because a
discrete amount of time has elapsed between time t
and time t 2 1. But how much time has elapsed? Is it
“1.0” in some arbitrary units of time? Suppose that in
a first experiment data are sampled once per day and
in a second experiment data are sampled once per
hour. If Eq. (2.3) is used to estimate b in each experi-
ment, the values obtained for b will turn out to be very
different since the value of b is partially dependent on
the time interval between successive occasions of mea-
surement. In order to obtain replicability, it is impor-
tant to remove the confounding of the size of the effect
and the elapsed time between measurements.
Continuous time modeling removes the dependency
between the lagged effects and the time interval of the
lag. In continuous time, the model from Eq. (2.3) can
be rewritten as
dut
5 βut 1 vt ; ð2:4Þ
dt
where β is the coefficient defining the linear relation-
ship between the displacement from equilibrium, ut,
Another random document with
no related content on Scribd:
agog to be at a thing, and then, when you think her settled to half-an-
hour’s pleasant play, off she is like any butterfly. She says her, ‘How
doth the little busy bee,’ dutifully, but when I tell her she is not a bit
like a busy bee, but rather like a foolish, flitting butterfly, I’m afraid
she rather likes it, and makes up to the butterflies as if they were
akin to her, and were having just the good time she would prefer. But
you must come and see the child to understand how volatile she is.
“‘Oh, mother, please let me have a good doll’s wash this
afternoon; I’m quite unhappy about poor Peggy! I really think she
likes to be dirty!’
“Great preparations follow in the way of little tub, and soap, and
big apron; the little laundress sits down, greatly pleased with herself,
to undress her dirty Peggy; but hardly is the second arm out of its
sleeve, than, presto! a new idea; off goes Kitty to clean out her doll’s-
house, deaf to all nurse’s remonstrances about ‘nice hot water,’ and
‘poor dirty Peggy.’
“I’m afraid the child is no more constant to her loves than to her
play; she is a loving little soul, as you know, and is always adoring
somebody. Now it’s her father, now Juno, now me, now Hugh; and
the rain of warm kisses, the soft clasping arms, the nestling head,
are delicious, whether to dog or man. But, alas! Kitty’s
blandishments are a whistle you must pay for; to-morrow it is
somebody else’s turn, and the bad part is that she has only room for
one at a time. If we could get a little visit from you, now, Kitty would
be in your pocket all day long; and we, even Peggy, would be left out
in the cold. But do not flatter yourself it would last; I think none of
Kitty’s attachments has been known to last longer than two days.
“If the chief business of parents is to train character in their
children, we have done nothing for Kitty; at six years old the child
has no more power of application, no more habit of attention, is no
more able to make herself do the thing she ought to do, indeed, has
no more desire to do the right thing, than she had at six months old.
We are getting very unhappy about it. My husband feels strongly that
parents should labour at character as the Hindoo gold-beater labours
at his vase; that character is the one thing we are called upon to
effect. And what have we done for Kitty? We have turned out a ‘fine
animal,’ and are glad and thankful for that; but that is all; the child is
as wayward, as unsteady, as a young colt. Do help us, dear aunt.
Think our little girl’s case over; if you can, get at the source of the
mischief, and send us a few hints for our guidance, and we shall be
yours gratefully evermore.”

“And now for my poor little great-niece! Her mother piles up

charges against her, but how interesting and amusing and like the
free world of fairy-land it would all be were it not for the tendencies
which, in these days, we talk much about and watch little against.
We bring up our children in the easiest, happy-go-lucky way, and all
the time talk solemnly in big words about the momentous importance
of every influence brought to bear upon them. But it is true; these
naughty, winsome ways of Kitty’s will end in her growing up like half
the ‘girls’—that is, young women—one meets. They talk glibly on
many subjects; but test them, and they know nothing of any; they are
ready to undertake anything, but they carry nothing through. This
week, So-and-so is their most particular friend, next week such
another; even their amusements, their one real interest, fail and flag;
but then, there is some useful thing to be learnt—how to set tiles or
play the banjo! And, all the time, there is no denying, as you say, that
this very fickleness has a charm, so long as the glamour of youth
lasts, and the wayward girl has bright smiles and winning, graceful
ways to disarm you with. But youth does not last; and the poor girl,
who began as a butterfly, ends as a grub, tied to the earth by the
duties she never learnt how to fulfil; that is, supposing she is a girl
with a conscience; wanting that, she dances through life whatever
befalls; children, husband, home, must take their chance. ‘What a
giddy old grandmother the Peterfields have!’ remarked a pert young
man of my acquaintance. But, indeed, the ‘giddy old grandmother’ is
not an unknown quantity.
“Are you saying to yourself, a prosy old ‘great-aunt’ is as bad as a
‘giddy old grandmother’? I really have prosed abominably, but Kitty
has been on my mind all the time, and it is quite true, you must take
her in hand.
“First, as to her lessons: you must help her to gain the power of
attention; that should have been done long ago, but better late than
never, and an aunt who has given her mind to these matters takes
blame to herself for not having seen the want sooner. ‘But,’ I fancy
you are saying, ‘if the child has no faculty of attention, how can we
give it to her? It’s just a natural defect.’ Not a bit of it! Attention is not
a faculty at all, though I believe it is worth more than all the so-called
faculties put together; this, at any rate, is true, that no talent, no
genius, is worth much without the power of attention; and this is the
power which makes men or women successful in life.
“Attention is no more than this—the power of giving your mind to
what you are about—the bigger the better so far as the mind goes,
and great minds do great things; but have you never known a person
with a great mind, ‘real genius,’ his friends say, who goes through life
without accomplishing anything? It is just because he wants the
power to ‘turn on,’ so to speak, the whole of his great mind; he is
unable to bring the whole of his power to bear on the subject in
hand. ‘But Kitty?’ Yes, Kitty must get this power of ‘turning on.’ She
must be taught to give her mind to sums and reading, and even to
dusters. Go slowly; a little to-day and a little more to-morrow. In the
first place, her lessons must be made interesting. Do not let her
scramble through a page of ‘reading,’ for instance, spelling every
third word and then waiting to be told what it spells, but see that
every day she learns a certain number of new words—six, twelve,
twenty, as she is able to hear them; not ‘spellings’—terrible
invention!—but words that occur in a few lines of some book of
stories or rhymes; and these she should know, not by spelling, but by
sight. It does not matter whether the new words be long or short, in
one syllable or in four, but let them be interesting words. For
instance, suppose her task for to-day be ‘Little Jack Horner,’ she
should learn to know, by sight, thumb, plum, Christmas, corner, &c.,
before she begins to read the rhyme; make ‘plum’ with her loose
letters, print it on her slate, let her find it elsewhere in her book, any
device you can think of, so that ‘plum’ is brought before her eyes
half-a-dozen times, and each time recognised and named. Then,
when it comes in the reading lesson, it is an old friend, read off with
delight. Let every day bring the complete mastery of a few new
words, as well as the keeping up of the old ones. At the rate of only
six a day she will learn, say, fifteen hundred in a year; in other words,
she will have learned to read! And if it do not prove to be reading
without tears and reading with attention, I shall not presume to make
another suggestion about the dear little girl’s education.
“But do not let the lesson last more than ten minutes, and insist,
with brisk, bright determination, on the child’s full concentrated
attention of eye and mind for the whole ten minutes. Do not allow a
moment’s dawdling at lessons.
“I would not give her rows of figures to add yet; use dominoes or
the domino cards prepared for the purpose, the point being to add or
subtract the dots on the two halves in a twinkling. You will find that
the three can work together at this as at the reading, and the children
will find it as exciting and delightful as ‘old soldier.’ Kitty will be all
alive here, and will take her share of work merrily; and this is a point
gained. Do not, if you can help it, single the little maid out from the
rest and throw her on her own responsibility. ’Tis a ‘heavy and a
weary weight’ for the bravest of us, and the little back will get a trick
of bending under life if you do not train her to carry it lightly, as an
Eastern woman her pitcher.
“Then, vary the lessons; now head, and now hands; now tripping
feet and tuneful tongue; but in every lesson let Kitty and the other
two carry away the joyous sense of—
‘Something attempted, something done.’

“Allow of no droning wearily over the old stale work,—which must

be kept up all the time, it is true, but rather by way of an exciting
game than as the lesson of the day, which should always be a
distinct step that the children can recognise.
“You have no notion, until you try, how the ‘now-or-never’ feeling
about a lesson quickens the attention of even the most volatile child;
what you can drone through all day, you will; what must be done, is
done. Then, there is a by-the-way gain besides that of quickened
attention. I once heard a wise man say that, if he must choose
between the two, he would rather his child should learn the meaning
of ‘must’ than inherit a fortune. And here you will be able to bring
moral force to bear on wayward Kitty. Every lesson must have its
own time, and no other time in this world is there for it. The sense of
the preciousness of time, of the irreparable loss when a ten minutes’
lesson is thrown away, must be brought home.
“Let your own unaffected distress at the loss of ‘golden minutes’
be felt by the children, and also be visited upon them by the loss of
some small childish pleasure which the day should have held. It is a
sad thing to let a child dawdle through a day and be let off scot-free.
You see, I am talking of the children, and not of Kitty alone, because
it is so much easier to be good in company; and what is good for her
will be good for the trio.
“But there are other charges: poor Kitty is neither steady in play
nor steadfast in love! May not the habit of attending to her lessons
help her to stick to her play? Then, encourage her. ‘What! The doll’s
tea-party over! That’s not the way grown-up ladies have tea; they sit
and talk for a long time. See if you can make your tea-party last
twenty minutes by my watch!’ This failing of Kitty’s is just a case
where a little gentle ridicule might do a great deal of good. It is a
weapon to be handled warily, for one child may resent, and another
take pleasure in being laughed at; but managed with tact I do believe
it’s good for children and grown-ups to see the comic side of their
doings.
“I think we err in not enough holding up certain virtues for our
children’s admiration. Put a premium of praise on every finished
thing, if it is only a house of cards. Steadiness in work is a step on
the way towards steadfastness in love. Here, too, the praise of
constancy might very well go with good-humoured family ‘chaff,’ not
about the new loves, which are lawful, whether of kitten or playmate,
but about the discarded old loves. Let Kitty and all of them grow up
to glory in their constancy to every friend.
“There, I am sending you a notable preachment instead of the few
delicate hints I meant to offer; but never mount a woman on her
hobby—who knows when she will get off again?”
 CHAPTER III

AN EDUCATIONAL EXPERIMENT
You wish me to tell you the story of my little girl? Well, to begin at the
beginning. In looking back through the pages of my journal I find
many scattered notices of Agnes, and I always write of her, I find, as
“poor Agnes.” Now, I wonder why? The child is certainly neither
unhealthy nor unhappy—at least, not with any reason; but again and
again I find this sort of entry:—
“Agnes displeased with her porridge; says nothing, but looks
black all day.”
“Harry upset his sister’s work-basket, by accident, I truly believe,
but she can’t get over it; speaks to no one, and looks as if under a
cloud.”
I need not go on; the fact is, the child is sensible of many injuries
heaped upon her; I think there is no ground for the feeling, for she is
really very sweet when she has not, as the children say, the black
dog on her back.
It is quite plain to me, and to others also, I think, that we have let
this sort of thing go on too long without dealing with it. We must take
the matter in hand.
Please God, our little Agnes must not grow up in this sullen habit,
for all our sakes, but chiefly for her own, poor child; I felt that in this
matter I might be of more use than Edward, who simply does not
understand a temper less sunny and open than his own. I pondered
and pondered, and, at last, some light broke in upon me. I thought I
should get hold of one principle at a time, work that out thoroughly,
and then take up the next, and so on, until all the springs of
sullenness were exhausted, and all supplies from without stopped. I
was beginning to suspect that the laws of habit worked here as
elsewhere, and that, if I could get our dear child to pass, say, six
weeks without a “fallen countenance,” she might lose this distressing
failing for life.
I meant to take most of the trouble of this experiment upon
myself, but somehow I never can do anything without consulting my
husband. I think men have clearer heads than we women; that is,
they can see both sides of a question and are not carried off their
feet by the one side presented to them.
“Well, Edward, our little Agnes does not get over her sulky fits; in
fact, they last longer, and are harder to get out of than ever!”
“Poor little girl! It is unhappy for her and for all of us. But don’t you
think it is a sort of childish malaise she will soon grow out of?”
“Now, have you not said, again and again, that a childish fault, left
to itself, can do no other than strengthen?”
“True; I suppose the fact is I am slow to realise the fault. But you
are right. From the point of view of habit we are pledged to deal with
it. Have you made any plans?”
“Yes; I have been trying to work the thing out on your lines. We
must watch the rise of the sullen cloud, and change her thoughts
before she has time to realise that the black fit is coming.”
“You are right; if we can keep the child for only a week without this
settling of the cloud, the mere habit would be somewhat broken.”
We had not to wait for our opportunity. At breakfast next day—
whether Harry’s porridge looked more inviting than her own, or
whether he should not have been helped first, or whether the child
had a little pain of which she was hardly aware—suddenly, her eyes
fell, brows dropped, lips pouted, the whole face became slightly paler
than before, the figure limp, limbs lax, hands nerveless—and our
gentle child was transformed, become entirely unlovable. So far, her
feelings were in the emotional stage; her injury, whatever it was, had
not yet taken shape in her thoughts; she could not have told you
what was the matter, because she did not know; but very soon the
thinking brain would come to the aid of the quick emotions, and then
she would be sulky of fixed purpose. Her father saw the symptoms
rise and knew what that would lead to, and, with the promptness
which has often saved us, he cried out—
“Agnes, come here, and hold up your pinafore!” and Agnes trotted
up to his side, her pinafore held up very much to receive the morning
dole of crumbs for the birds; presently, she came back radiant with
the joy of having given the birds a good breakfast, and we had no
more sulky fits that day. This went on for a fortnight or so, with fair
but not perfect success. Whenever her father or I was present, we
caught the emotion before the child was conscious of it, and
succeeded in turning her thoughts into some pleasant channel. But
poor nurse has had bad hours with Agnes; there would sit the child,
pale and silent, for hours together, doing nothing because she liked
to do it, but only because she was made. And, once the fit had
settled down, thick and steady as a London fog, neither her father
nor I could help in the least. Oh, the inconceivable settled cloudiness
and irresponsiveness of that sweet child face!
Our tactics were at fault. No doubt they helped so far as they
went. We managed to secure bright days that might otherwise have
been cloudy when we happened to be present at the first rise of the
sullen mood. But it seemed impossible to bring about so long an
abstinence from sullen fits as would eradicate the habit. We pictured
to ourselves the dreary life that lay before our pretty little girl; the sort
of insulation, the distrust of her sweetness, to which even one such
sullen fit would give rise; worse, the isolation which accompanies this
sort of temper, and the anguish of repentance to follow. And then, I
know, madness is often bred of this strong sense of injured
personality.
It is not a pleasant thing to look an evil in the face. Whether or no
“a little knowledge is a dangerous,” certainly, it is a trying thing. If we
could only have contented ourselves with, “Oh, she’ll grow out of it
by-and-by,” we could have put up with even a daily cloud. But these
forecasts of our little girl’s future made the saving of the child at any
cost our most anxious care.
 “I’ll tell you what, Mary; we must strike out a new line. In a general
way, I do believe it’s best to deal with a child’s faults without making
him aware that he has them. It fills the little beings with a ridiculous
sense of importance to have anything belonging to them, even a
fault. But in this case, I think, we shall have to strike home and deal
with the cause at least as much as with the effects, and that, chiefly,
because we have not effects entirely under our control.”
“But, Edward, what if there is no cure? What if this odious temper
were hereditary—our precious child’s inheritance from those who
should have brought her only good?”
“Poor little wife! so this is how it looks to you. You women are
sensitive creatures. Why, do you know, it never occurred to me that it
might be all my fault. Well, I will not laugh at the fancy. Let us take it
seriously, even if, as it seems to me, a little morbid. Let us suppose
that this sad sullenness of which I hear so much and see so little, is,
indeed, Agnes’s inheritance from her mother—may she only inherit
all the rest, and happy the man whose life she blesses! The question
is not ‘How has it come?’ but ‘How are we to deal with it?’—equally,
you and I. Poor things! It’s but a very half-and-half kind of matrimony
if each is to pick out his or her own particular bundle of failings, and
deal with it single-handed. This poor man finds the prospect too
much for him! As a matter of fact, though, I believe that every failing
of mind, body, temper, and what not, is a matter of inheritance, and
that each parent’s particular business in life is to pass his family
forward freed from that particular vicious tendency which has been
his own bane—or hers, if you prefer it.”
“Well, dear, do as you will; I feel that you know best. What it would
be in these days of greater insight to be married to a man who would
say, ‘There, that boy may thank his mother’ for this or the other
failure. Of course, the thing is done now, but more often than not as
a random guess.”
“To return to Agnes. I think we shall have to show her to herself in
this matter, to rake up the ugly feeling, however involuntary, and let
her see how hateful it is. Yes, I do not wonder you shrink from this.
So do I. It will destroy the child’s unconsciousness.”
 “Oh, Edward, how I dread to poke into the poor little wounded
heart, and bring up worse things to startle her!”
“I am sorry for you, dear, but I think it must be done; and don’t you
think you are the person to do it? While they have a mother I don’t
think I could presume to poke too much into the secrets of the
children’s hearts.”
“I’ll try; but if I get into a mess you must help me through.”
The opportunity came soon enough. It was pears this time. Harry
would never have known whether he had the biggest or the least.
But we had told nurse to be especially careful in this matter. “Each of
the children must have the biggest or best as often as one another,
but there must be no fuss, no taking turns, about such trifles.
Therefore, very rightly, you gave Harry the bigger and Agnes the
smaller pear.”
Agnes’s pear was not touched; there the child sat, without word or
sob, but all gathered into herself, like a sea anemone whose
tentacles have been touched. The stillness, whiteness, and brooding
sullenness of the face, the limp figure and desolate attitude, would
have made me take the little being to my heart if I had not too often
failed to reach her in this way. This went on all day, all of us
suffering; and in the evening, when I went to hear the children’s
prayers before bed, I meant to have it out.
We were both frozen up with sadness, and the weary little one
was ready to creep into her mother’s heart again. But I must not let
her yet.
“So my poor Agnes has had a very sad day?”
“Yes, mother,” with a little quivering sob.
“And do you know we have all had a very sad day,—father,
mother, your little brother, nurse—every one of us has felt as if a
black curtain had been hung up to shut out the sunshine?”
The child was sympathetic, and shivered at the sight of the black
curtain and the warm sunshine shut out.
 “And do you know who has put us all out in the dark and the cold?
Our little girl drew the curtain, because she would not speak to any
of us, or be kind to any of us, or love any of us all the day long; so
we could not get into the sunshine, and have been shivering and sad
in the cold.”
“Mother, mother!” with gasping sobs; “not you and father?”
“Ah! I thought my little girl would be sorry. Now let us try to find out
how it all happened. Is it possible that Agnes noticed that her
brother’s pear was larger than her own?”
“Oh, mother, how could I?” And the poor little face was hidden in
her mother’s breast, and the outbreak of sobs that followed was too
painful. I feared it might mean actual illness for the sensitive little
soul. I think it was the right thing to do; but I had barely courage
enough to leave the results in more loving hands.
“Never mind; don’t cry any more, darling, and we will ask our
Father above to forgive and forget all about it. Mother knows that her
dear little Agnes will try not to love herself best any more. And then
the black curtain will never fall, and we shall never again be a whole
long day standing sadly out in the cold. Good-night from mother, and
another sweet good-night from father.”
The treatment seems to answer. On the slightest return of the old
sullen symptoms we show our little girl what they mean. But the grief
that follows is so painful that I’m afraid we could not go on with it for
the sake of the child’s health. But, happily, we very rarely see a sulky
face now; and when we do we turn and look upon our child, and the
look melts her, until she is all gentleness, penitence, and love.
 CHAPTER IV

DOROTHY ELMORE’S ACHIEVEMENT: A

FORECAST

Part I
I know of no happier moment for parents than that when their eldest
daughter returns from school to take her place finally by her mother’s
side. It was two years that very day since we had seen Dorothy
when her father set out for Lausanne to bring her home; and how the
children and I got through the few days of his absence, I don’t know.
The last touches had been put, many times over, to her rooms—not
the plain little room she had left, but a dainty bower for our young
maiden, a little sitting-room opening into a pure nest of a bedroom.
Our eyes met, her father’s and mine, and moistened as we conjured
up I don’t know what visions of pure young life to be lived there, the
virginal prayers to be offered at the little prayer table, the gaiety of
heart that should, from this nook, bubble over the house, and, who
knows, by-and-by, the dreams of young love which should come to
glorify the two little rooms.
Two or three times already had the children put fresh flowers into
everything that would hold a flower. Pretty frocks and sweet faces,
bright hair and bright eyes, had been ready this long time to meet
sister Dorothy.
At last, a telegram from Dover—“Home by five”—and our
restlessness subsided into a hush of expectation.
The sound of wheels on the gravel, and we flew to the hall door
and stood in two files, children and maids, Rover and Floss, waiting
to welcome the child of the house. Then, a lovely face, glad to tears,
looking out of a nest of furs; then, a light leap, almost before the
carriage drew up, and I had her in my arms, my Dorothy, the child of
my heart! The order of the day was “high tea,” to which every one,
down to baby May, sat up. We two, her father and I, gave her up to
the children, only exchanging notes by the species of telegraphy
married folk understand.
“Indubitably lovely!” said her father’s eyes; “And what grace—
what an elegant girl she is!” answered mine; “And do but see what
tact she shows with the little ones;” “And notice the way she has with
us, as if her heart were brimming with reverence and affection.”
Thus, we two with our eyes. For a week or more we could not settle
down. As it was the Christmas holidays, we had not Miss Grimshaw
to keep us in order, and so it happened that wherever Dorothy ran,—
no, she went with a quick noiseless step, but never ran,—about the
house to find out the old dear nooks, we all followed; a troop of
children with their mother in the rear; their father too, if he happened
to be in. Truly we were a ridiculous family, and did our best to turn
the child’s head. Every much has its more-so. Dorothy’s two special
partisans were Elsie, our fifteen years old girl, fast treading in her
sister’s steps, and Herbert, our eldest son, soon to go to college.
Elsie would come to my room and discourse by the hour, her text
being ever, “Dorothy says.” And as for Herbs, it was pleasant to see
his budding manhood express itself in all sorts of little attentions to
his lovely sister.
For lovely she was; there could not be two opinions on that point.
A lilymaid, tall and graceful, without a trace of awkwardness or self
consciousness; the exquisite complexion of the Elmores (they are a
Devonshire family), warm, lovely rose on pearly white, no hint of
brunette colouring; a smile which meant spring and love and other
good things; and deep blue eyes reflecting the light of her smile.
Never, not even during the raptures of early married life, have I
known a month of such joyous exhilaration as that which followed
Dorothy’s return, and I think her father would own as much.
 What a month it was! There was the pleasant earthly joy of going
to town to get frocks for Dorothy; then the bewilderment of not being
able to find out what suited her best.
“Anything becomes her!” exclaims Mdme. la Modiste; “that figure,
that complexion, may wear anything.”
And then, the pleasure of entering a room—all eyes bent upon us
in kindliness; our dear old friends hurrying forward to make much of
the child; the deference and gentleness of her manner to these, and
the warmth with which she was received by her compeers, both
maidens and men; her grace in the dance; her simplicity in
conversation; the perfection of her manner, which was not manner at
all, but her own nature, in every situation. After all, she liked best to
be at home; was more amiable and lovely with father and mother,
brothers and sisters, than with the most fascinating strangers. Our
good child! We had grown a little shy of speaking to her about the
best things, but we knew she said her prayers: how else this outflow
of sweet maiden life upon us all?
I can imagine these ramblings of mine falling into the hands of a
young pair whose life is in each other:—“Oh, only the outpourings of
a doting mother!” and they toss the pages aside. But never believe,
young people, that yours are the only ecstatic moments, yours the
only experiences worth recording; wait and see.

Part II
These happy days had lasted for a month or more, when, one
bright day in February, I remember it well, a little cloud arose. This is
how it was: Dorothy had promised Elsie that she would drive her in
the pony-carriage to Banford to choose a doll for May’s birthday.
Now, it happened that I wanted the little carriage to take to my
“Mothers” at Ditchling the clothing I had bought in London with their
club money. My errand could not be deferred; it must be done that
day or a week later. But I did not see why the children’s commission
would not do as well to-morrow; and so I said, in good faith, as I was
stepping into the carriage, hardly noticing the silence with which my
remark was received.
I came home tired, after a long afternoon, looking forward to the
welcome of the girls. The two seniors were sitting in the firelight,
bright enough just then to show me Dorothy sitting limp and pale in a
low chair, and Elsie watching her with a perplexed and anxious
expression. Dorothy did look up to say, “Are you tired, mother?” but
only her eyes looked, there was nothing behind them.
“You look tired and cold enough, my dear; what has been the
matter?”
“Oh, I’m very well, thank you; but I am tired, I think I’ll go to bed.”
And she held up a cold cheek for the mother’s kiss for which she
offered no return.
Elsie and I gazed at one another in consternation; our fairy
princess, our idol (was it indeed so?) What had come to her?
“What is the matter with Dorothy? Has she a headache?”
“Oh, mother, I don’t know,” said the poor child, on the verge of
tears. “She has been like this ever since you went, saying ‘Yes,’ and
‘No,’ and ‘No, thank you,’ quite kindly, but never saying a word of
herself. Has any one been grieving our Dorothy, or is she going to be
ill? Oh, mother, mother!”
“Nay, child, don’t cry. Dorothy is overdone; you know she has
been out twice this week, and three times last, and late hours don’t
suit her. We must take better care of her, that’s all.”
Elsie was comforted, but not so her mother. I believed every word
I had said to the child; but all the time there was a stir in my heart
like the rustling of a snake in the grass. But I put it from me.
It was with a hidden fear that I came down to breakfast. Dorothy
was in the room already doing the little duties of the breakfast table.
But she was pale and still; her hands moved, her figure hung, in the
limp way I had noticed the night before. Her cheek, a cold “Good-
morning, mother,” and a smile on her lips that brought no light to her
eyes, was all the morning salutation I got. Breakfast was an
uncomfortable, constrained meal. The children wondered what was
the matter, and nobody knew. Her father got on best with Dorothy for
he knew nothing of the evening’s history, so he petted her as usual,
making all the more of her for her pale looks.
For a whole week this went on, and never once was I allowed to
meet Dorothy eye to eye. The children were hardly better served, for
they, too, had noticed something amiss; only her father could win
any of the old friendliness, because he treated her as the Dorothy
who had come home to us, only a little done up.
“We must have the doctor for that child, wife. Don’t you see how
she is losing flesh, and how the roses she brought home are fading?
She has no appetite and no spirits. But, why, you surely don’t think
our dainty moth has burned her wings already? There’s nobody
here, unless it’s young Gardiner, and she would never waste herself
on a gawky lad like that!”
This was a new idea, and I stopped a moment to consider, for I
knew of at least half-a-dozen young men who had been attentive to
Dorothy, all to be preferred to this hobbledehoy young Gardiner. But,
no! I could trace the change from the moment of my return from
Ditchling. But I jumped at the notion of the doctor; it would, at any
rate, take her out of herself, and—we should see.
The doctor came; said she wanted tone; advised, not physic, but
fresh air, exercise, and early hours. So we all laid ourselves out to
obey his directions that day, but with no success to speak of.
But the next was one of those glorious February days when every
twig is holding itself stiffly in the pride of coming leafage, and the
snowdrops in the garden beds lift dainty heads out of the brown
earth. The joy of the spring did it. We found her in the breakfast-
room, snowdrops at her throat, rosy, beaming, joyous; a greeting,
sweet and tender, for each, and never had we known her talk so
sparkling, her air so full of dainty freshness. There was no relapse
after this sudden cure. Our good friend Dr. Evans called again, to
find her in such flourishing health that ten minutes’ raillery of “my
poor patient” was the only attention he thought necessary. But, “H’m!
Mighty sudden cure!” as he was going out, showed that he too found
something odd in this sudden change.
In a day or two we had forgotten all about our bad week. All went
well for awhile. At the end of five weeks, however, we were again
pulled up—another attack of sudden indisposition, so outsiders
thought. What did I think? Well, my thoughts were not enviable.
“Father, I wish you would call at Walker’s and choose me some
flowers for this evening.” It was the evening of the Brisbanes’ dance,
and I had half an idea that Arthur Brisbane had made some
impression on Dorothy. His state of mind was evident enough. But,
without thinking twice, I interrupted with—
“Don’t you think what we have in the ‘house’ will do, dear? What
could make up better than stephanotis and maidenhair?”
Dorothy made no answer, and her father, thinking all was right,
went off at once; he was already rather late. We thought no more of
the matter for a minute or two, when, at the same moment, Elsie and
I found our eyes fixed upon Dorothy. The former symptoms followed
—days of pallor and indisposition, which were, at the same time,
days of estrangement from us all. Again we had in Dr. Evans, “just to
look at her,” and this time I noticed—not without a foolish mother’s
resentment—that his greeting was other than cordial, “Well, young
lady, and what’s gone amiss this time?” he said, knitting his bushy
brows, and gazing steadily at her out of the eyes which could be
keen as well as kind. Dorothy flushed and fidgeted under his gaze,
but gave only the cold unsatisfactory replies we had been favoured
with. The prescription was as before; but again the recovery was
sudden, and without apparent cause.

Part III
To make a long story short, this sort of thing went on, at longer or
shorter intervals, through all that winter and summer and winter
again. My husband, in the simplicity of his nature, could see nothing
but—
 “The child is out of sorts; we must take her abroad for a month or
two; she wants change of air and scene.”
The children were quicker-eyed; children are always quick to
resent unevenness of temper in those about them. A single angry
outbreak, harsh word, and you may lay yourself out to please them
for months before they will believe in you again. Georgie was the first
to let the cat out of the bag.
“Dorothy is in a sulky fit again, mother; I wish she wouldn’t!”
Elsie, who has her father’s quick temper, was in the room.
“You naughty ungrateful little boy, you! How can you say such a
thing of Dorothy? Didn’t she sit all yesterday morning making sails
for your boat?”
Georgie, a little mollified, “Yes, but why need she be sulky to-day?
We all loved her yesterday, and I’m sure I want to to-day!”
Now that the mask was fallen and even the children could see
what was amiss, I felt that the task before me must not be put off. I
had had great misgivings since the first exhibition of Dorothy’s sullen
temper; now I saw what must be done, and braced myself for a
heavy task. But I could not act alone; I must take my husband into
my confidence, and that was the worst of it.
“George, how do you account for Dorothy’s fits of wretchedness?”
“Why, my dear, haven’t I told you? The child is out of sorts, and
must have change. We’ll have a little trip up the Rhine, and perhaps
into Switzerland, so soon as the weather is fit. It will be worth
something to see her face light up at some things I mean to show
her!”
“I doubt if there is anything the matter with her health; remember
how perfectly well and happy she is between these fits of
depression.”
“What is it, then? You don’t think she’s in love, do you?”
“Not a bit of it; her heart is untouched, and her dearest loves are
home loves.”
 My husband blew his nose, with a “Bless the little girl! I could find
it in my heart to wish it might always be so with her. But what is your
notion? I can see you have got to the bottom of the little mystery.
Trust you women for seeing through a stone wall.”
“Each attack of what we have called ‘poorliness’ has been a fit of
sullenness, lasting sometimes for days, sometimes for more than a
week, and passing off as suddenly as it came.”
My dear husband’s face clouded with serious displeasure; never
before had it worn such an expression for me. I had a sense of
separation from him, as if we two, who had so long been one, were
two once more.
“This is an extraordinary charge for a mother to bring against her
child. How have you come to this conclusion?”
Already was my husband become my judge. He did not see that I
was ill, agitated, still standing, and hardly able to keep my feet. And
there was worse to come: how was I to go through with it?
“What causes for resentment can Dorothy conceivably have?” he
repeated, in the same cold judicial tone.
“It is possible to feel resentment, it is possible to nurse
resentment, to let it hang as a heavy cloud-curtain between you and
all you love the best, without any adequate cause, without any cause
that you can see yourself when the fit is over!”
My voice sounded strange and distant in my own ears: I held by
the back of a chair to steady myself: but I was not fainting: I was
acutely alive to all that was passing in my husband’s mind. He
looked at me curiously, inquisitively, but not as if I belonged to him,
and were part and parcel of his life.
“You seem to be curiously familiar with a state of feeling which I
should have judged to be the last a Christian lady would know
anything about.”
“Oh, my husband, don’t you see? You are killing me. I am not
going through this anguish for nothing. I do know what it is. And if