University of Southern California Dissertations and Theses

Marriage and health: The effect of marital status on health and the mechanisms by which marital status affects health among elderly people

(USC Thesis Other)

Marriage and health: The effect of marital status on health and the mechanisms by which marital status affects health among elderly people

PDF

Download a page range

Download transcript

Copy asset link

Request this asset

Transcript (if available)

Content INFORMATION TO USERS
This manuscript has been reproduced from the microfilm master. UMI films
the text directly from the original or copy submitted. Thus, some thesis and
dissertation copies are in typewriter face, while others may be from any type of
computer printer.
The quality of this reproduction is dependent upon the quality of the
copy submitted. Broken or indistinct print, colored or poor quality illustrations
and photographs, print bleedthrough, substandard margins, and improper
alignment can adversely affect reproduction.
In the unlikely event that the author did not send UMI a complete manuscript
and there are missing pages, these will be noted. Also, if unauthorized
copyright material had to be removed, a note will indicate the deletion.
Oversize materials (e.g., maps, drawings, charts) are reproduced by
sectioning the original, beginning at the upper left-hand comer and continuing
from left to right in equal sections with small overlaps.
Photographs included in the original manuscript have been reproduced
xerographically in this copy. Higher quality 6" x 9” black and white
photographic prints are available for any photographs or illustrations appearing
in this copy for an additional charge. Contact UMI directly to order.
ProQuest Information and Learning
300 North Zeeb Road, Ann Arbor, Ml 48106-1346 USA
800-521-0600
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
MARRIAGE AND HEALTH:
THE EFFECT OF MARITAL STATUS ON HEALTH
AND THE MECHANISMS BY WHICH MARITAL STATUS AFFECTS
HEALTH AMONG ELDERLY PEOPLE
by
Jung Ki Kim
A Dissertation Presented to the
FACULTY OF THE GRADUATE SCHOOL
UNIVERSITY OF SOUTHERN CALIFORNIA
In Partial Fulfillment of the
Requirements for the Degree
DOCTOR OF PHILOSOPHY
(GERONTOLOGY / PUBLIC POLICY)
December 2000
Copyright 2000 Jung Ki Kim
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
UMI Number: 3041484
Copyright 2000 by
Kim, Jung Ki
All rights reserved.
___ __®
UMI
UMI Microform 3041484
Copyright 2002 by ProQuest Information and Learning Company.
All rights reserved. This microform edition is protected against
unauthorized copying under Title 17, United States Code.
ProQuest Information and Learning Company
300 North Zeeb Road
P.O. Box 1346
Ann Arbor, Ml 48106-1346
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
UNIVERSITY OF SOUTHERN CALIFORNIA
The Graduate School
U niversity Park
LOS ANGELES, CALIFORNIA 90089-1695
This dissertation, written by
_______ Jung Ki Mm___________________________
Under the direction o f h j& z .. Dissertation
Committee, and approved by all its members,
has been presented to and accepted by The
Graduate School, in partial fulfillment o f
requirements for the degree o f
DOCTOR OF PHILOSOPHY
Dean o f Graduate Studies
Date December 18, 2000
DISSERTATION COMMITTEE
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Acknowledgements
This dissertation is more than a final requirement for the Ph.D. degree. It
has required a lot of time, efforts, knowledge, assistance, encouragement and
prayer. It had made me realize that my life is and will be an on-going process of
learning. However, this realization comes with confidence, enthusiasm for new
things and gratitude to many people.
I would like to express my indescribable gratitude to Eileen Crimmins, who
has been a great mentor in this study, other works and my life. I am thankful to her
guidance and encouragement. Without her support, advice and encouragement, this
dissertation would not have been completed. I have learned from her what a true
mentor would be like. She is the one who has both cold brain and warm heart. Her
expertise, cheerful personality and incredible patience have always encouraged me.
No words can express my appreciation for her.
I also thank my other committee members. I thank Teresa Seeman, Merril
Silverstein, Bob Knight and Kyu-Taik Sung for their valuable and insightful
comments. It was a privilege to have them in my committee. I also would like to
express my gratitude to David Peterson who has been always in my side and given
me smile and encouragement from the beginning of my study at the Andrus
Gerontology Center.
I am especially grateful to my parents, Jin Suck Kim and Suk Jae Yoo, for
their endless love and prayer for me. Without them, my life would not have
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
started, but more importantly, without them, my life would not have been
meaningful. They have been the source o f empowerment of my life, and will
continue to be.
I thank II Young, my husband, for all his prayer and encouragement. From
the moment I met him to now, he has been such a faithful and sincere supporter for
me. He has helped me enjoy working on my dissertation. He has been willing to
share all of the process of writing dissertation whether it was the happy moment or
struggling time. His love, prayer and encouragement will be always in my mind.
I also would like to thank my brother, Yoon-Ki Kim, sister-in-law, Hee
Young Im, my parents-in-law-to-be, Young Tae Kim and Jung Ja Shin, and my
friends, Aaron Hagedom and Sang-Eun Shin for their prayer and support. I give
special thanks to Chong Wha and Peter Park for their care and love.
Finally, but most importantly, I thank God for being my Lord. He always
shows me His way and His will that He wants to fulfill through me. All equations
and numbers were meaningful only with His greatest plan for my life. I would like
to conclude this acknowledgement with prayer that He uses my degree, job, family
and life for His way.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Table of Contents
Acknowledgem ents................................................................................................................... ii
List of T ables.................................................................................................................. vi
List of Figures................................................................................................................ xii
A bstract...........................................................................................................................xv
Chapter I: Introduction................................................................................................. 1
Background
Importance and Contributions of Studying the Effect of Marital Status
on Health and the Mechanisms by which Marriage Affects Health
Overview of Remaining Chapters
Chapter II: Literature Review and Conceptual Framework....................................... 8
Conceptual Linkages
Mechanisms by Type of Health Outcome
Inconsistent Findings on the Relationship between Marital Status and
Health
Gender Differences
Differences among Unmarried Elderly Persons
Widowhood and Health
Change of Marital Status / Duration of Marital Status
Living Arrangements and Health
Conceptual Frameworks
Research Questions and Hypotheses
Chapter III: M ethods..................................................................................................... 39
Data and Subject Selection
Measures
Analysis
Chapter IV: Description of the Subject Population...................................................64
Chapter V: Marital Status and H ealth........................................................................ 75
Number of People in Each Marital Status
Differences in the Prevalence of Health Conditions without
Controlling Age
Differences in the Incidence of Health Conditions without
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
V
Controlling Age
Differences in Mechanisms by Marital Status
Results from Logistic Regressions
Differential Effect of Marital Status on Health and Differential
Mechanisms by Type of Health Outcomes
Chapter VI: Widowhood and Health.........................................................................209
The Married, the Recently Widowed and the Longer-term
Widowed
Results from Logistic Regressions
Chapter VII: Living Arrangements and Health........................................................240
Living Alone vs. Living with Others
Results from Logistic Regressions
Chapter VIII: Summary and Discussion.................................................................. 273
Summary
Discussion
Limitations of the Study
References....................................................................................................................293
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
vi
List of Tables
2.1 Categorization of Health Conditions by their Characteristics and
Mechanisms.....................................................................................................20
3.1 Description of Variables Used: HRS, AHEAD, and Mac A rthur................ 48
4.1 Description of Study Sample at Baseline (Mean & Percent):
HRS, AHEAD and MacArthur......................................................................68
4.2 Number and Percent of Marital Status by Sex in 1992 and 1994:
H R S...................................................................................................................71
4.3 Number and Percent of Marital Status by Sex in 1995 and 1998:
AHEAD........................................................................................................... 72
4.4 Number and Percent of Marital Status by Sex in 1988 and 1991:
MacArthur........................................................................................................73
4.5 Number and Percent of Marital Status by Sex in 1991 and 1995:
M acArthur........................................................................................................74
5.1 Number of People by Marital Status and Sex among Those Whose
Marital Status Had Not Changed between 1992 and 1994: H R S.............. 76
5.2 Number of People by Marital Status and Sex among Those Whose
Marital Status Had Not Changed between 1993 and 1998: AHEAD .... 76
5.3 Number of People by Marital Status and Sex among Those Whose
Marital Status Had Not Changed between 1988 and 1995:
MacArthur........................................................................................................77
5.4 Percent of Those who Had Functioning Difficulties, Diseases and
Conditions in 1992 by Marital Status and Sex: H R S...................................78
5.5 Percent of Those who Had Functioning Difficulties, Diseases and
Conditions in 1993 by Marital Status and Sex: AHEAD........................... 79
5.6 Percent of Those who Had Functioning Difficulties, Diseases and
Conditions in 1988 by Marital Status and Sex: MacArthur........................80
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
vii
5.7 Percent of Those who Got Functioning Difficulties, New Diseases
and Conditions or Died between 1992 and 1994 by Marital Status
and Sex: H R S....................................................................................................82
5.8 Percent of Those who Got Functioning Difficulties, New Diseases
and Conditions between 1993 and 1998 or Died between 1993 and
1995 by Marital Status and Sex: A H EA D...................................................83
5.9 Percent of Those who Got Functioning Difficulties, New Diseases
and Conditions between 1988 and 1995 by Marital Status and Sex:
MacArthur........................................................................................................ 84
5.10 Mean and Percent of Characteristics of Sample at Baseline (1992)
by Marital Status and Sex: H R S.................................................................... 86
5.11 Mean and Percent of Characteristics of Sample at Baseline (1993)
by Marital Status and Sex: AHEAD.............................................................. 87
5.12 Mean and Percent of Characteristics of Sample at Baseline (1988)
by Marital Status and Sex: M acArthur..........................................................88
5.13 Odds Ratios from Logistic Regression of Having Functioning
Difficulties, Diseases and Conditions (1992) on Marital States
and Age: H R S...................................................................................................91
5.14 Odds Ratios from Logistic Regression of Having Functioning
Difficulties, Diseases and Conditions (1993) on Marital States
and Age: AHEAD............................................................................................ 92
5.15 Odds Ratios from Logistic Regression of Having Functioning
Difficulties, Diseases and Conditions (1988) on Marital States
and Age: MacArthur........................................................................................94
5.16 Odds Ratios from Logistic Regression of Getting Functioning
Difficulties, New Diseases and Conditions between 1992 and 1994 on
Marital States and Age: H R S ..........................................................................96
5.17 Odds Ratios from Logistic Regression of Getting Functioning
Difficulties, New Diseases and Conditions between 1993 and 1998 on
Marital States and Age: AHEAD................................................................... 97
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
viii
5.18 Odds Ratios from Logistic Regression of Getting Functioning
Difficulties, New Diseases and Conditions between 1988 and 1995
on Marital States and Age: M acArthur..........................................................98
5.19 Odds Ratios from Logistic Regression of Having Functioning
Difficulties, Diseases and Conditions (1992) on Marital States and
Other Variables among Males: H R S ............................................................ 105
5.20 Odds Ratios from Logistic Regression of Having Functioning
Difficulties, Diseases and Conditions (1992) on Marital States and
Other Variables among Females: H R S........................................................ 112
5.21 Odds Ratios from Logistic Regression of Having Functioning
Difficulties, Diseases and Conditions (1993) on Marital States and
Other Variables among Males: AHEAD.................................................. 119
5.22 Odds Ratios from Logistic Regression of Having Functioning
Difficulties, Diseases and Conditions (1993) on Marital States and
Other Variables among Females: AHEAD................................................ 125
5.23 Odds Ratios from Logistic Regression of Having Functioning
Difficulties, Diseases and Conditions (1988) on Marital States and
Other Variables: M acArthur.........................................................................132
5.24 Odds Ratios from Logistic Regression of Getting Functioning
Difficulties, New Diseases and Conditions between 1992 and 1994
on Marital States and Other Variables among Males: H R S..................... 160
5.25 Odds Ratios from Logistic Regression of Getting Functioning
Difficulties, New Diseases and Conditions between 1992 and 1994
on Marital States and Other Variables among Females: H R S................. 167
5.26 Odds Ratios from Logistic Regression of Getting Functioning
Difficulties, New Diseases and Conditions between 1993 and 1998
on Marital States and Other Variables among Males: AHEAD...............175
5.27 Odds Ratios from Logistic Regression of Getting Functioning
Difficulties, New Diseases and Conditions between 1993 and 1998
on Marital States and Other Variables among Females: AH EA D 181
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
ix
5.28 Odds Ratios from Logistic Regression o f Getting Functioning
Difficulties, New Diseases and Conditions between 1988 and 1995
on Marital States and Other Variables among Males: MacArthur 188
6.1 Percent and Number of the Married, the Recently Widowed
and the Longer-term Widowed by Sex: HRS, AHEAD and
M acArthur.....................................................................................................210
6.2 Percent of Those who Had Functioning Difficulties, Diseases
and Conditions in 1998 for the Continuously Married, the
Recently Married and the Longer-term Widowed by Sex: AHEAD.. . . 212
6.3 Percent of Those who Had Functioning Difficulties, Diseases
and Conditions in 1991 for the Continuously Married, the
Recently Married, and the Longer-term Widowed by Sex:
MacArthur...................................................................................................... 213
6.4 Percent of Those who Got Functioning Difficulties, New
Diseases and Conditions between 1993 and 1998 for the
Continuously Married, the Recently Married, and the
Longer-term Widowed by Sex: AHEAD.....................................................214
6.5 Percent o f Those who Got Functioning Difficulties, New
Diseases and Conditions between 1988 and 1991 for the
Continuously Married, the Recently Married, and the
Longer-term Widowed by Sex: MacArthur................................................ 215
6.6 Mean and Percent of Proposed Mechanisms (1998) for the
Continuously Married, the Recently Married and the
Longer-term Widowed by Sex: AHEAD.....................................................216
6.7 Mean and Percent of Proposed Mechanisms (1991) for the
Continuously Married, the Recently Married and the
Longer-term Widowed by Sex: MacArthur................................................ 217
6.8 Mean and Percent of Mechanisms for the Continuously Married,
the Recently Married and the Longer-term Widowed in 1995 and
1998 among Males: AHEAD........................................................................221
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
X
6.9 Mean and Percent of Mechanisms for the Continuously Married,
the Recently Widowed and Longer-term Widowed in 1995 and
1998 among Females: AHEAD......................................................................222
6.10 Mean and Percent of Mechanisms for the Continuously Married,
the Recently Married and the Longer-term Widowed in 1988 and
1991 among Males: MacArthur.....................................................................223
6.11 Mean and Percent of Mechanisms for the Continuously Married,
the Recently Married and the Longer-term Widowed in 1988 and
1991 among Females: MacArthur................................................................ 224
6.12 Percent with Adverse Change in Proposed Mechanisms for the
Continuously Married, the Recently Widowed and the
Longer-term Widowed between 1993 and 1995 and between
1995 and 1998: AHEAD.................................................................................225
6.13 Percent with Adverse Change in Proposed Mechanisms for the
Continuously Married, the Recently Widowed and the
Longer-term Widowed between 1988 and 1991 in MacArthur..................226
6.14 Odds Ratios from Logistic Regression of Getting Functioning
Difficulties, New Diseases and Conditions between 1993 and
1998 on Being Recently and Longer-term Widowed and Other
Variables: AHEAD.......................................................................................... 230
6.15 Odds Ratios from Logistic Regression of Having Functioning
Difficulties, Diseases and Conditions (1998) on Being Recently
and Longer-term Widowed and Other Variables: AHEAD.......................234
7.1 Percent (N) of Those Who Lived Alone among Never Married,
Separated/divorced and Widowed: AHEAD...............................................241
7.2 Percent of Those who Had Functioning Difficulties, Diseases and
Conditions in 1993 by Living Arrangements and Sex: AHEAD.............. 242
7.3 Percent of Those who Got Functioning Difficulties, New Diseases
and Conditions between 1993 and 1998 by Living Arrangements
and Sex: AHEAD........................................................................................... 242
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
xi
7.4 Mean and Percent of Characteristics of Sample at Baseline (1993)
by Living Arrangements and Sex: AHEAD............................................. 243
7.5 Odds Ratios of Having Functioning Difficulties, Diseases and
Conditions (1993) by Living Alone and Sex among the Unmarried:
AHEAD.........................................................................................................244
7.6 Odds Ratios of Getting Functioning Difficulties, New Diseases
and Conditions between 1993 and 1998 by Living Alone and Sex
among the Unmarried: AHEAD.................................................................245
7.7 Odds Ratios from Logistic Regression of Having Functioning
Difficulties, Diseases and Conditions (1993) on Living Alone and
Other Variables among Males: A H EA D ....................................................249
7.8 Odds Ratios from Logistic Regression of Having Functioning
Difficulties, Diseases and Conditions (1993) on Living Alone and
Other Variables among Females: AHEAD.................................................255
7.9 Odds Ratios from Logistic Regression of Getting Functioning
Difficulties, New Diseases and Conditions between 1993 and
1998 on Living Alone and Other Variables among Males:
A H EA D...........................................................................................................261
7.10 Odds Ratios from Logistic Regression of Getting Functioning
Difficulties, New Diseases and Conditions between 1993 and
1998 on Living Alone and Other Variables among Females:
A H EA D ...........................................................................................................266
8.1 Percent of Those who Had ADL Difficulties in 1993 and 1995
by Living Arrangements and Sex: AHEAD...............................................284
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
List of Figures
2.1 A Model of the Effect of Marital Status on Health Outcomes.................... 34
2.2 A Model of the Effect of Change of Marital Status on Health
Outcomes.........................................................................................................34
3.1 Sample Construction: H R S............................................................................. 40
3.2 Sample Construction: AHEAD.......................................................................42
3.3 Sample Construction: MacArthur...................................................................44
5.1 Odds Ratios of Having ADL Difficulties by Marital Status:
HRS (1992) and AHEAD (1993).................................................................136
5.2 Odds Ratios of Having LADL Difficulties by Marital Status
and Sex: HRS (1992) and AHEAD (1993)............................................... 137
5.3 Odds Ratios of Having Nagi Difficulties by Marital Status
and Sex: HRS (1992), AHEAD (1993) and MacArthur (1988)............. 138
5.4 Odds Ratios of Having Hypertension by Marital Status
and Sex: HRS (1992) and AHEAD (1993)............................................... 139
5.5 Odds Ratios of Having Diabetes by Marital Status and Sex:
HRS (1992), AHEAD (1993) and MacArthur (1988)...............................140
5.6 Odds Ratios of Having Cancer by Marital Status and Sex:
HRS (1992), AHEAD (1993) and MacArthur (1988)...............................141
5.7 Odds Ratios of Having Lung Conditions by Marital Status
and Sex: HRS (1992) and AHEAD (1993)............................................... 142
5.8 Odds Ratios of Having Heart Conditions by Marital Status
and Sex: HRS (1992), AHEAD (1993) and MacArthur (1988)............. 143
5.9 Odds Ratios of Having Stroke by Marital Status and Sex:
HRS (1992), AHEAD (1993) and MacArthur (1988)...............................144
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
xiii
5.10 Odds Ratios of Having Arthritis by Marital Status and Sex:
HRS (1992) and AHEAD (1993).................................................................145
5.11 Odds Ratios of Having Broken Bone (HRS, 1992) or Hip
Hip (AHEAD, 1993) by Marital Status and S ex ........................................ 146
5.12 Odds Ratios of Having Depression by Marital Status and Sex:
HRS (1992), AHEAD (1993) and MacArthur (1988)...............................147
5.13 Odds Ratios of Getting ADL Difficulties by Marital Status
and Sex: HRS and AHEAD...................................................................... 194
5.14 Odds Ratios of Getting IADL by Marital Status and Sex:
AHEAD.......................................................................................................... 195
5.15 Odds Ratios of Getting Nagi Difficulties by Marital Status
and Sex: HRS, AHEAD and MacArthur.................................................... 196
5.16 Odds Ratios of Dying by Marital Status and Sex:
HRS and AHEAD......................................................................................... 197
5.17 Odds Ratios of Getting Hypertension by Marital Status
and Sex: HRS, AHEAD and MacArthur.................................................... 198
5.18 Odds Ratios of Getting Diabetes by Marital Status and Sex:
HRS, AHEAD and MacArthur.................................................................... 199
5.19 Odds Ratios of Getting Cancer by Marital Status and Sex:
HRS, AHEAD and MacArthur....................................................................200
5.20 Odds Ratios of Getting Lung Conditions by Marital Status
and Sex: HRS and AHEAD.........................................................................201
5.21 Odds Ratios of Getting Heart Conditions by Marital Status
and Sex: HRS, AHEAD and MacArthur....................................................202
5.22 Odds Ratios of Getting Stroke by Marital Status and Sex:
HRS, AHEAD and MacArthur....................................................................203
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
xiv
5.23 Odds Ratios of Getting Arthritis by Marital Status and Sex:
HRS and AHEAD......................................................................................... 204
5.24 Odds Ratios of Getting Broken Bone (HRS) or Hip
(AHEAD) by Marital Status and S ex...........................................................205
5.25 Odds Ratios of Getting Depression by Marital Status and Sex:
HRS, AHEAD and M acA rthur...................................................................206
5.26 The Effect of Being Never Married, Separated/Divorced,
Widowed on a Number of Health Outcomes: HRS, AHEAD
and M acArthur...............................................................................................207
5.27 Mechanisms through which Being Unmarried Affects
Health Outcomes: HRS and AHEAD......................................................... 208
6.1 Odds Ratios of Having (1993 or 1998) or Getting Health
Conditions (between 1993 and 1998) on Being Recently
Widowed Compared to Being Longer-term Widowed:
AHEAD...........................................................................................................239
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
XV
Abstract
The effect of marital status on health and the mechanisms by which marital
status affects health are examined. The effect of being never married,
separated/divorced and widowed on health relative to being married, that of recent
or longer-term widowhood on health relative to continuous marriage, and that of
living arrangements on health among the unmarried are examined. Data come from
the Health and Retirement Survey, the Asset and Health Dynamics among the
Oldest Old, and the MacArthur Research Network on Successful Aging
Community Study.
Overall, the married were healthier than the unmarried. However, in some
cases, the married had worse health status than the unmarried. The relationship
between marital status and health differed by age, gender and type of health:
Younger cohorts, men and functioning, some fatal conditions, mental health and
death were more strongly related to marital status. The widowed and the
separated/divorced had worse health status than the married. The health states of
the never married was between that of the married and the separated/divorced and
the widowed.
The mechanisms by which marital stats related to health were not different
by type o f health outcome, nor by gender. Inclusion of covariates most times
explained the effects of marital status on health.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
xvi
Longer-term widowhood was associated with a higher probability of
acquiring some physical conditions primarily because of age. Recent widowhood
was related to a higher likelihood of developing depression while longer-term
widowhood was not. No proposed mechanisms were associated with the
relationship between widowhood and health.
Those living alone had similar health to those living with others. The
exceptions were that they had worse functioning and depression. This leads us to
conclude that the effect of marriage on health is not the result of living arrangement
differences.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
1
I. Introduction
Background
Marital status has long been documented as a factor affecting health. Many
researchers have been examining the effect of marriage on health outcomes. A
substantial literature has shown that married individuals tend to be healthier and
survive longer than unmarried persons (Goldman et al., 1995; Hu & Goldman,
1990; Verbrugge, 1979; Williams et al., 1992; Zick & Smith, 1991).
Many previous studies suggest that the effect o f marital status on health
outcomes occur through a number of mechanisms. For instance, social support
from a spouse may buffer psychological distress, provide needed physical care, and
promote health behaviors, and thus improve physical as well as mental health
outcomes. Economic resources available to married people may provide more
access to health care services, thus positively affecting health outcomes. However,
the mechanisms are not always clear and in most studies, one or two mechanisms
are selectively studied. This dissertation not only seeks to re-examine the effect of
marital status on health among older persons but to clarify the mechanisms by
which marital status affects health as well. It investigates whether married elderly
people are healthier and live longer than unmarried counterparts, and attempts to
explain why health advantages for the married, if any, occur.
While most researchers examine a few selected health outcomes in relation
to marital status, this study examines the effects of marital status on health using a
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
2
number of health indicators expecting that the effect of marital status on health may
differ by type of health outcome. In addition, different mechanisms may determine
the relationship between marital status and different health outcomes. For example,
more serious conditions tend to require more social support while health care
utilization may be more important for physical conditions than mental conditions.
In examining the disadvantaged health status among the unmarried
compared to the married, it is important to note that it may not be marital status
itself, but the change of marital status from being married to unmarried that plays a
more critical role in deteriorating health. A longer time spent widowed may
increase the likelihood of making adaptations to a potentially adverse health
situation (Keith & Lorenz, 1989). Previous literature shows that recent change in
marital status is associated with a decline in health.
Related to the issue of the relationship between marital status and health is
the effect of living arrangements on health and whether that effect is related to, or
independent from that of marital status. Those who live alone tend to be poorer,
more often women, older and single rather than married. However, the issue is
whether those who live alone are less healthy than those who live with others
because of their living arrangements. While it has been well documented that those
with more disabilities and health conditions are likely to live with others because of
their health problems and the probability of coresiding with children is elevated by
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
3
the absence of a spouse, failing health and advanced age (Lee & Dwyer, 1996),
whether living arrangements affect health needs further investigation.
This study examines the issue of the relationship between living
arrangements and health hypothesizing that part of the effect of marital status on
health may be through living arrangements. This study, therefore, examines to
what extent health differences by marital status can be explained by differences in
living arrangements. If the better health of the married is partly due to the greater
social support and integration provided by living with another person, then the
health of the non-married living alone should be poorer than that of the non
married living with another.
Importance and Contributions of Studying the Effect of Marital Status on Health
and the Mechanisms by which Marriage Affects Health
As life expectancy has increased, there is a growing concern about the
quality of later life. We have achieved longer life, but perhaps at the expense of
more years with diseases and disabilities. Thus, the quality o f life in the oldest ages
has become an important research agenda in the field of aging. In health research,
many researchers have examined mortality, morbidity and active life expectancy in
old ages (Colin, 1999; Crimmins et al., 1997a, 1997b; Hayward et al., 1998; Robine
et al., 1998), and how these relate to demographic, socioeconomic, behavioral and
psychological factors.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
4
This study is an extension of those efforts to clarify factors resulting in
healthy life among older people, specifically by addressing the effect of marital
status on health. By examining different dimensions of marital status and a variety
of possible mechanisms, this study is expected to provide a clearer understanding
of what factors are related to health in later life and how they are related to health
outcomes. It may also provide implications for appropriate health interventions for
older people (e.g. providing alternative social support to unmarried people who
lack spousal support). Marital status is not something that can be changed, but by
intervening in other factors associated with the relationship between marital status
and health, we could improve health outcomes o f those who are potentially
disadvantaged in health status because of their marital status.
The overarching conceptual framework o f this study derives from the
biopsychosocial model. From this perspective, health and disease and health-
related behaviors are viewed as resulting from complex interactions and processes
among biological, psychological, socioeconomic and ecological factors. Since
health outcomes associated with marital status are related to a variety of
interdisciplinary factors, applying this model to the relationship between marital
status and health would extend our explanation o f the relationship as well as
confirm the importance and applicability of this model in the relationship between
marital status and health.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
5
In addition to the benefit in developing causal relationships from an earlier
point in time to subsequent health outcomes (Rogers et al., 2000), utilizing
longitudinal data helps to examine changes in marital status and factors associated
with the relation between marriage and health, and how these changes are related to
health outcomes. Also, multiple datasets employed in this study will enable us to
compare study findings and to compensate for data-specific limitations in order to
make research findings more extensive and generalizable.
Another important contribution of this study is related to the practical utility
of the study findings (Bengtson et al., 1999). The findings of differential effects of
marital status on health outcomes and differential mechanisms by gender, for
example, have implications for health professionals and health care providers.
They can assist older people in improving their health by meeting their differential
social and psychological needs. For instance, if the significant effect of the
availability of more economic resources confounds the relationship between marital
status and health among women, this would imply that unmarried women might
have worse health because of a lack of economic resources to facilitate health care
service utilization. Thus, interventions for unmarried women may focus more on
improving access to appropriate health care services.
Lastly, as mentioned, this study seeks to apply the biopsychosocial
framework to a specified example of the relationship between marital status and
health. By building on previous work and findings on this topic, this study aims to
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
6
develop theoretical constructs in the relationship and clarify the linkages between
them. By exploring a variety of socioeconomic, behavioral and psychological
mechanisms involved in the effect of marital status on health, this study would help
explain not only what the effect of marital status on health is, but also how
marriage and health are related. Inconsistent findings on the effect of marriage on
health may be partly attributed to the lack o f comprehensive and proper
incorporation and linkage of mechanisms. By providing theoretical linkages of
constructs, this study aims to make a “marriage and health” picture more complete
and better explainable and testable (Turner, 1986).
Overview of Remaining Chapters
The study begins with a review of what previous research has revealed
about the relationship between marital status and health. Chapter II provides an
overview of differential effects of marital status by type of health outcome, the
conceptual linkages on how marital status influences health outcomes, findings on
differences among unmarried persons, and how the duration of marriage and
widowhood affect health outcomes, followed by previous studies on living
arrangements and health. Research questions and hypotheses proposed for this
study are also discussed in Chapter II. Chapter III describes the research design,
sample selection and characteristics, measures and methodological approaches to
be used in the study. Chapters IV, V, VI and VII report the results of this study.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
7
Chapter IV describes the characteristics of the sample. In Chapter V, the effect of
marital status on health among four marital groups and the mechanisms by which
marital status affects health are examined. Chapter VI focuses on the effect o f
recent and longer-term widowhood on health and the mechanisms. The
relationship between living arrangements and health is examined in Chapter VII.
Chapter VIII summarizes and discusses the study findings and limitations of the
study as well as future directions for research in this area.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
8
II. Literature Review and Conceptual Framework
Researchers have examined the effect o f marital status on a variety o f health
outcomes. While most research findings show that marriage provides beneficial
effects on health outcomes, the effect of marital status on health varies by type of
health outcomes studied. Review of previous research shows that researchers have
not paid much attention to differentiating health outcomes in the relationship
between marital status and health. Differential effects of marital status by type of
health outcome are not the focus of most studies, but the effect of marital status on
health can be differentiated by type of health indicator examined. Since Farr
(1858) noted the marriage advantage for mortality, a number of researchers have
examined the effect of marital status on mortality. While most researchers found
that the married have lower mortality than the nonmarried (never married,
separated, divorced and widowed) (Berkman & Syme, 1979; Johnson et al., 2000),
House et al. (1982) found the association of being married with lower mortality for
men, but not for women, and Blazer (1982) did not find any relationship between
marital status and mortality.
Hahn (1993) examined the association o f marital status with health among
females. She used self-reported health status to indicate health status. Self-
reported health status is relevant in this discussion because research indicates that
self-reported health measures are effective predictors of mortality (Idler & Angel,
1990; Idler & Kasl, 1991; Maddox & Douglas, 1973), and are also closely
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
9
correlated to chronic diseases (Davis & Ware, 1981). Hahn (1993) found that
married women rated their health higher than never married, divorced, separated
and widowed.
Sherboume and Hays (1990) examined the effect of marital status on health
outcomes of patients with specific diseases. They examined patients with one or
more of four chronic diseases such as hypertension, diabetes, coronary heart
disease and depression. Hypertensives were chosen on the basis of systolic and
diastolic blood pressure reported by physicians; coronary heart disease patients
consisted of those who had suffered a myocardial infarction within 12 months
before screening and/or had congestive heart failure; diabetics were selected on the
basis of physicians’ reports of current diabetes; and depressed patients were
sampled in a two-stage screening procedure (Bumam et al., 1988). Even though
they examined functioning limitation among patients with the above diseases, thus
not directly examining the relationship between marital status and the prevalence or
incidence of those diseases, this study implies how those diseases could be
differently related to marital status. They found a positive effect of being married
on mental health outcome— satisfaction with physical abilities, but no effect on
physical health outcomes.
Health conditions differ in pattern of occurrence, progression and
consequences. For example, some conditions such as stroke and fracture can be
experienced by a sudden event that occurs without any warning symptoms while
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
10
other conditions such as diabetes and arthritis tend to be gradually developing. On
the other hand, some conditions such as myocardial infarction can be incapacitating
and others such as hypertension non-incapacitating.
By their severity, chronic diseases and conditions can be categorized as
more serious and potentially life-threatening conditions and less serious conditions.
Diseases such as hypertension, diabetes, stroke, cancer, lung conditions and heart
conditions are often categorized as more serious and possible to threaten life while
arthritis and fracture are as less serious and hardly threatening life in nature. This is
a useful categorization of health conditions because it is simple enough not to
destroy statistical power for analysis, but still reflects important differential aspects
of different diseases and conditions. Even though Pienta et al. (1999) did not
intend to differentiate the relationship between marriage and health by type of
health condition, their categorization of conditions as ‘fatal’ and ‘nonfatal’
diseases, functional limitations and disability is useful in examining differential
effect of marriage on health outcomes and differential mechanisms by type of
health outcome.
Kim (2000) suggested that the effect of marital status on health outcomes at
the older ages differs by type of disease and gender. For instance, stroke and
diabetes may differ in their pattern of psychological effects (Penninx et al., 1999)
and their association with marital status. For women, marriage adversely affected
some of the health outcomes of stroke patients while it had a positive effect on
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
11
most health outcomes of diabetes patients. For men, being married was not
associated with any health outcomes of stroke patients while it had a negative effect
on kidney or urine problems among diabetes patients. The positive association of
marriage with health outcomes among elderly females with diabetes may be partly
due to the nature of diabetes which is a long-term, progressive disease and, thus,
requires more emotional support, which is often provided by a spouse.
Examination of the differential effect of marital status on different health
outcomes is more important in a gender-specific context. Riessman and Gerstel
(1985) found a clear pattern of differential risk for men and women in the
relationship between marital dissolution and health: Separated and divorced women
were at greater relative risk for the less extreme physical and mental health
problems than their counterpart men while separated and divorced men were at
greater relative risk than their counterpart women for the more extreme physical
and mental health outcomes including death.
This dissertation proposes that all health conditions examined will be
affected by marital status, but some conditions will have a stronger association with
marital status than others. The degree of association is expected to be stronger for
death, functioning, mental health and more serious conditions than for less serious
conditions. Detailed discussion of differential effect of marriage on different health
conditions follows later in this section in conjunction with differential mechanisms
by type of health outcome.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
12
Conceptual Linkages
This study examines the relation between marital status and a number of
health outcomes. As Weakland (1977) suggested, marital status might affect health
by causing physiological change (e.g. susceptibility to illness) or it might affect the
course or outcome of health once health problems exist. Theoretically, the absence
of a spouse may have detrimental physical effects such as weakened immune
system functioning. On the other hand, marital status may affect health through
several mechanisms such as social support, health care utilization, economic
resources and health behaviors. This section discusses what we know about the
relationship between those mechanisms and a marriage-health picture.
Social Support
The social support provided by marriage may enhance physical and mental
health and recovery from diseases. Support from one’s spouse may improve
emotional well-being (Gove et al., 1983), and, thus, help both physical and
emotional recovery (Duncan, 1994; Robertson & Suinn, 1968; Ross et al., 1990).
As the buffering hypothesis suggests, the presence or absence of interaction with a
spouse may affect health outcomes through psychological or behavioral processes
(Cohen & Willis, 1985; Mor-Barak et al., 1991). For instance, the presence of
support from a spouse might encourage one to cooperate with a prescribed medical
regimen, promote adaptive behavioral modifications (Shenkel et al., 1985), or
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
13
provide a sense of well-being that inhibits the negative effects o f stress on health.
Silverstein and Bengtson (1994) found that both instrumental and expressive forms
of social support moderated declines in well-being associated with poor health.
Several studies which examined the relationship between marital status and
psychological or emotional well-being showed that marriage may provide
emotional safety/support, thus married people may have better mental health
(Cassileth et al., 1984; Sherboume & Hays, 1990).
However, some studies show that social support may not have a positive
effect all the time. Cotten (1999) found that social support has little positive effect
on depression and, for married women, higher social support appears to be
associated with more burdens. Kim (2000) argued that the large mortality effects
among married female stroke victims compared to their unmarried counterparts
might be attributed to women’s role as a care provider and its costs. Women tend
to take greater responsibility for a spouse’s health behaviors and provide informal
care. They may provide a spouse with benefits of social support rather than
receiving it (Shye et al., 1995). Thus, the potential of health benefits through
marriage might be reduced or even reversed by the costs that they may pay for
caregiving.
In addition to a spouse, the married tend to have larger social support
networks including children, other relatives and friends. The buffering effects of
social networks as well as social support are particularly beneficial for those elderly
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
14
people who are physically and mentally impaired and need to cope with the
consequences of physical and psychological limitations (Unger et al., 1999). Since
adult children, relatives and close friends often play an important role in helping
older persons, the availability of a wide range of support networks through
marriage and the amount of support received from them might mitigate the adverse
effects of diseases and conditions on physical functioning among the elderly
(Kriegsman et al., 1997).
Health Care Utilization
Researchers have examined the relationship between marital status and the
use of health care services (Evashwick et al., 1984; Joung et al., 1995; Morgan,
1980; Nelson, 1993; Verbrugge, 1979; Wolinsky & Johnson, 1992). Most studies
reported a positive relationship between being married and the use of health care
services. Economic resources available to married people may facilitate their
health care service utilization (Verbrugge, 1979). In addition, the married may be
more careful about their own health and, thus, more readily use medical services
for preventive and curative reasons (Verbrugge, 1979). Thus, if health service
utilization is related to health outcomes (Dighe et al., 1997; Stump et al, 1995) and
the married use health services more than the unmarried, health care service
utilization may be confounded in the relationship between marital status and health
outcomes. However, some studies have demonstrated that being unmarried may
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
15
not necessarily be related to lower health service usage. Wolinsky and Johnson
(1992) reported that becoming widowed did not alter previous patterns of health
care service utilization such as physician or hospital utilization.
Economic Resources
Well-documented associations between marital status and economic well
being and between economic well-being and health or mortality have led
researchers to consider the possibility that socioeconomic status may be a
confounding factor in the relationship between marital status and health.
Researchers suggest that economic factors may be key correlates of health status in
old age (Haan et al., 1989; Hahn, 1993). Study findings show that some of the
effects of marital status on health may be explained by differences in economic
assets and marital acquisitions such as income, home ownership and private
insurance (Hahn, 1993). The high risk of morbidity and mortality related to being
unmarried and to being poor may be associated. Smith and Waitzman (1994)
found that a combination of poverty and being single results in an extremely high
risk of mortality that is greater than that expected from each risk acting
independently.
But Smith and Waitzman’s (1994) findings on the high risk of mortality
related to both being unmarried and poor are primarily limited to non-elderly men
(ages 25-64) and, in fact, they detected weak main and interaction effects for the
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
16
elderly (ages 65-74). It is possible that some unmarried older people may be both
better off financially and have more access to medical care in old age than in their
early life because of programs for financial support and medical care available to
older people in our society (Keith & Lorenz, 1989). These provisions may
moderate the effect of being unmarried on health. Also, those who experience
diminished economic resources only in later life, presumably would retain some of
the benefits of earlier economic resources such as better health and health practices
from earlier years. In these circumstances, some of the potentially negative
effects of financial strain resulting from current unmarried status may be mitigated
(Keith & Lorenz, 1989).
Health Behaviors
Although findings o f the effect of marriage on healthy lifestyles are not
completely consistent, it has been argued that marriage has a generally positive
effect on health behaviors (Anson, 1989; Schone & Weinick, 1998; Umberson,
1987). Marriage encourages one to regulate one’s conduct and adopt healthy
behaviors such as quitting smoking, having a balanced diet, doing exercise,
avoiding excessive drinking, participating in physical or social activities and
wearing a seatbelt. For example, Glass et al. (1999) found that social, productive
and physical activities lower the risk of mortality. Their findings that activities that
entail little or no physical exertion may also be beneficial in survival suggest that
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
17
psychosocial as well as physiological pathways be involved in the association
between activity and mortality. These health-related behaviors, which are more
common among the married than the unmarried (Ross et al., 1990; Umberson,
1987), may prevent the incidence of disease, enhance health outcomes and reduce
mortality (Colsher & Wallace, 1990; Fried et al., 1998; Ruigomez et al., 1995).
Differences in health behaviors by marital status are generally attributed to the
greater opportunity for social and productive activities and the greater social
support that encourages one to engage in healthy behaviors among married people
(Hahn, 1993).
Mechanisms by Type of Health Outcome
While the above mechanisms are proposed to explain the relationship
between marital status and health in general, the role each mechanism plays may
depend on the health outcome studied. While some research has examined the
relationship between specific mechanisms and the total number of chronic
conditions (Joung et al., 1994), most previous research has examined the effects of
selective mechanisms on specific health outcomes (Ewart et al., 1983; Hahn, 1993;
Joung et al., 1994; Medalie & Goldbourt, 1976; Penninx et al., 1999; Reed et al.,
1983; Shye et al., 1995; Unger et al., 1999). Social support including the
characteristics of social networks (e.g. size of network, participation in social
events), family problems, spousal support and communication behaviors was found
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
18
to be related to mortality risk (Shye et al., 1995), prevalence of coronary heart
diseases (Reed et al., 1983), development of angina (Medalie & Goldbourt, 1976),
blood pressure (Ewart et al., 1983), the control of diabetes mellitus (Schwartz et al.,
1986), or physical functioning measured by the Nagi scale (Unger et al., 1999).
Penninx et al. (1999) examined how the characteristics of diseases play a
differential role in the relationship between social support and health outcomes.
Their findings showed disease-specific social support patterns such that greater
feelings of loneliness were mainly found for persons with lung diseases or arthritis,
and receiving more instrumental support was mainly found for persons with
arthritis or stroke.
The relationship of health behaviors to specific diseases has also been
examined. For example, smoking is regarded as a predominant risk factor for the
onset of lung conditions. Obesity in old age is shown to be strongly related to the
development of depression and arthritis, and to increase the likelihood of
myocardial infarction, stroke and several major cancers (Lean, 2000). Rauscher et
al. (2000) found a positive relationship between body mass index (BMI) and lung
cancer risk even among non-smoking population. Recent heavy drinking was
found to be an independent risk factor for stroke (Hillbom et al., 1999).
In addition, a positive effect of the increased use of health care resources
such as doctor visits on survival and lower disability was found among elderly
stroke survivors (Dighe et al., 1997), and socioeconomic status measured by
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
19
household income and education was found to be negatively associated with
depression, and the association increased with age (Miech & Shanahan, 2000).
In thinking about the potential effect of mechanisms on the relationship
between marital status and health outcomes, the categorization of health outcomes
with psychosocial characteristics in consideration is useful to clarify the linkages of
health conditions to psychosocial and economic demands for a wide array of those
health conditions (Rolland, 1987). In order to refine the effects of four sets of
mechanisms by type of health condition, health conditions can be categorized into
more serious physical conditions, less serious physical conditions, functioning
limitations, mental condition and death. More serious physical conditions are those
diseases that can eventually lead to death (fatal) such as hypertension, diabetes,
cancer, lung conditions, heart conditions and stroke while less serious physical
conditions include such conditions as arthritis and broken hip/bone that are not
directly related to death (nonfatal). Another category of health conditions is
functional limitations such as ADL, LADL and Nagi difficulties, which may or may
not be related to the above diseases and conditions. Mental condition such as
depression is another health indicator which has different characteristics from
physical conditions, and thus potentially results from different mechanisms. Death
is a final health outcome that diseases, conditions and functioning difficulties can
directly or indirectly result in.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
20
Table 2.1. Categorization o f Health Conditions by their Characteristics and
Mechanisms
—----------Mechanisms Social Health Care Economic Health
Health Conditions ------- Support Utilization Resources Behaviors
Fatal Physical Conditions
Hypertension y S y y
Diabetes S y y y
Cancer S y y y
Lung Conditions y y y y
Heart Conditions y y y y
Stroke y y y y
Nonfatal Physical Conditions
Arthritis y
Broken Bone/Hip y
Functioning Limitations
ADL Difficulties y y y
IADL Difficulties y y y
Nagi Difficulties y y y
Mental Condition
Depression y
Death y y y y
Table 2.1 shows the matrix of health conditions grouped into five types of
health outcomes in conjunction with potential mechanisms affecting each
condition. This framework facilitates conceptualizing the linkages between the
mechanisms linked to marital status and health conditions. All mechanisms are
likely to be related to all health conditions, but this table shows the linkages of
relatively strong association between each condition and mechanisms. For
example, social support is also likely to affect nonfatal conditions such as arthritis
and fracture. However, social support is more likely to affect fatal conditions,
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
21
functioning limitations, depression and death which often require a lot of
instrumental and emotional help from caregivers. On the other hand, economic
resources are likely to be more strongly related to fatal physical conditions and
death which require more expensive medical treatments than other conditions.
Health care utilization is more likely to be related to fatal physical conditions,
functioning limitations and death which require consistent medical treatments and
whose risk can be reduced by regular check-ups and examinations. Nonfatal
conditions such as arthritis and fracture are less likely to be related to health care
utilization because fracture often occurs by a sudden accident and arthritis
gradually develops with misconception that it is a normal aging symptom, leaving
little room for intervening by preventive care service usage. Health behaviors are
more likely to be related to physical conditions and death on which consistent self
management is critical. Thus, while this study hypothesizes that married people are
likely to be doing better than unmarried people in health outcomes in general, the
difference may not be consistent across health outcomes, and the mechanisms by
which marital status affects health may differ by type of health condition.
Inconsistent Findings on the Relationship between Marital Status and Health
While a majority of previous studies have reported a positive effect of being
married on health outcomes, some studies show neither a direct nor indirect
relationship between marital status and physical health outcomes (Korenman et al.,
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
22
1990; Sherboume & Hays, 1990). For example, Sherboume and Hays (1990) argue
that many of the previous studies that suggest beneficial effects of being married on
a variety of physical health outcomes for subjects recovering from and coping with
serious physical illness and injury did not control for initial health status, which has
a very important influence on disease outcomes. Even though their findings may
be unique to the sample with chronically ill persons, selecting highly prevalent
chronic conditions such as hypertension, diabetes, coronary heart disease and
depression makes the findings applicable to the general population.
Inconsistent findings on the effect of marriage on health may be partly
attributed to the lack of comprehensive models incorporating a variety of the
mechanisms by which marital status affects health. The relationship between
marital status and health has often been addressed by examining a selective, rather
than comprehensive, set of confounding factors. Anson (1989) reported that the
positive effect of being married on health might be partly through living with a
proximate adult, which is associated with social ties and social regulation. While
her study concentrates on one possible mechanism of the relationship of marital
status with health outcomes, it has a weakness in that it excludes other potential
confounding factors such as the availability of resources for health care usage and
health behaviors triggered by marital status. Goldman et al. (1995) examined the
effect of social environment (e.g. family support and involvement in formal social
network such as church) and economic status of the elderly on the relationship
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
23
between marital status and health outcomes. Their study is also limited to two of
the potential mechanisms suggested by previous studies. While these studies are
useful in demonstrating and clarifying possible mechanisms whereby marital status
affects health, they suffer from not examining an exhaustive set of confounding
factors. When all factors are considered together, the effect of any one factor may
change.
The underlying assumption of this study is that if the introduction of various
possible mechanisms by which marital status affects health cancels out the effects
of marital status on health, we may argue that the effects of marriage on health, in
fact, occur through those mechanisms explained above. But if the model on the
relationship between marital status and health incorporates various possible
mechanisms, but still shows the effect of marital status on health, we may be able
to argue that marriage per se may have an effect on health.
Gender Differences
In addition to examining the effect of marital status on health outcomes,
researchers have tried to reveal how the effect differs by gender. Studies of gender
differences in the relationship between marital status and health have generally
found that men get more health benefits from marriage than women (Berkman &
Syme, 1979; Goldman et al., 1995; Gove et al., 1983; Hu & Goldman, 1990;
Lillard & Waite, 1995; Ross et al., 1990; Schone & Weinick, 1998). The negative
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
24
effect of widowhood on health is often reported to be more detrimental to health of
men than that of women (Stroebe & Stroebe, 1983). For example, Goldman et al.
(1995) found that widowed men had higher risks of being disabled and dying than
did married men while similar effects for women were smaller and statistically
insignificant. Riessman and Gerstel (1985) reported gender-linked effects of
marital dissolution on health. The divorced exhibit higher mortality than their age-
matched married peers and this relationship is stronger for men than for women
(Riessman & Gerstel, 1985). They argue that the higher risk of having severe
health problems including mortality among separated/divorced men compared to
their female counterparts implies that marriage may protect men more than women.
Mechanisms by Gender
Overall, studies on gender differences in the factors associated with the
relationship between marital status and health suggest that differential health status
of men by marital status is more attributable to social support mechanisms while
that of women is more to financial issues. For example, the strains and resources
that affect health status and are associated with being single or experiencing marital
disruption (e.g. divorce, widowhood) may be closely linked to gender differences.
Berkman and Syme (1983) argue that the effect of marital ties is weaker for women
than for men, and the reason is that they get protection from social supports based
on kin and friends, much more than men. Thus, unmarried women may be less
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
25
isolated than unmarried men (Litwak et al., 1989). Men who experienced marital
disruption may suffer more from loss of the social support whereas those women
may be more likely to experience vulnerability as a result of financial strain (Choi,
1996; Gerstel et al., 1985; Umberson et al., 1992). Hahn (1993) also suggested that
poorer health of unmarried women be largely because married women have more
economic resources compared to single women, and as a result, better health.
Thus, even though both men and women gain economic well-being and social
support from marriage, marriage may be more of an economic benefit to women
and benefit of social support to men.
It is also possible that men gain more instrumentally out of marriage
whereas women tend to be more emotionally invested in marriage (Gove et al.,
1983). Also, men might experience more social control through marriage, which
might induce more health benefits to men than to women (Umberson, 1992).
Differences among Unmarried Elderly Persons
While researchers often treat unmarried elderly persons as a homogeneous
group and focus on the presence or absence of a spouse, differences among
unmarried people might exist (Cotten, 1999). Many studies use two marital status
groups, the married and the unmarried, in order to investigate the effect of the
presence of a spouse on health (Kim, 2000). However, the various unmarried
groups (never married, separated/divorced, widowed) face different experiences,
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
26
which may lead to different health outcomes. Also, in order to examine the effect
of changing marital status on health over time, dividing unmarried people into
specific groups is necessary.
For example, the never-married elderly are often grouped with other
unmarried elderly, and, thus, they (merged with other unmarried groups) are found
to be less healthy than the married due to their marital status (not having a spouse)
(Stull & Scarisbrick-Hauser, 1989). However, several studies found that the health
status of the never-married elderly was no worse than that of the married and was,
in fact, better than that o f other groups of single elderly, especially the divorced
(Stull & Scarisbrick-Hauser, 1989).
Specifically, among the unmarried, those who are divorced or widowed
suffer the loss of a spouse and change in financial status whereas those who are
single but never married do not experience that loss (Keith & Lorenz, 1989). On
one hand, the divorced or separated may be healthier than the widowed because
while some of the divorced or separated elderly may choose to be unmarried, the
widowed have not had controllability over their marital situation. Also, the
divorced or separated may have released from situations such as conflict with a
spouse and some family members, which might have negatively affected health.
On the other hand, the divorced may suffer more from a lack of an informal support
network due to the stigma associated with divorce and the reduced scope of
relatives. The never-married elderly usually do not have children to depend on in
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
27
times of illness and failing health, but tend to substitute non-family social networks
for family social networks. Considering all possible variations expected among
unmarried people, it would be very important to differentiate them in the
examination of marital status and health.
Widowhood and Health
Widowhood has been the focus o f health research in the field of aging
because of the higher frequency of widowhood at the older ages and its potential
damaging effect on health. Most researchers agree that being widowed contributes
to deteriorating physical and mental health and increasing mortality (Bennett, 1997;
Stroebe & Stroebe, 1983). The effect o f widowhood on health is thought to be
largely due to stress resulting from the marital dissolution as well as the
accompanying decrease in social support: A stressful life event such as the loss of a
spouse may precipitate the occurrence o f a physical or mental condition (Stroebe &
Stroebe, 1983).
However, not all literature is consistent in expecting effects of widowhood
on health. For example, we may think about the effect of widowhood on health
through changes in social support. According to the continuity model, widowhood
is not associated with decrements or increases in expressive or instrumental
support. On the other hand, the compensatory model argues that social support
may increase in an apparent attempt to compensate for the loss of a spouse.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
28
Children and relatives tend to rally around a surviving person to provide a variety
of services and supports at bereavement (Heinemann, 1983). For example
Umberson et al. (1992) argued that widowhood was associated with an increase in
some types of strain, and at the same time, with an increase in some types of
support. If social support did not consistently change with widowhood, the
assumption of the adverse effect of becoming widowed on health due to decreased
social support should be questionable.
While widowhood in general is often found to be negatively associated with
health, the effect of widowhood may differ by age. As a person becomes widowed
at younger age, he/she is likely to face more challenges than older widow(er)s to
make adaptations to widowhood because bereavement at younger ages may be an
unexpected event accompanied by inadequate resources and support. Clayton
(1974) reported that there is increased psychological morbidity such as depressive
symptoms in the younger widowed, but not in older widowed. On the other hand,
by extending Lopata’s 1974 Widowhood Support System Survey, Heinemann
(1983) examined the determinants of the strength of family support systems among
widows in different life stages— young and middle-aged widows (ages less than 60),
young-old widows (ages 60-74) and old-old widow (ages 75 and over)— and found
that among widows, being younger was related to having a strong family support
system, while the old-old widows were most likely to be without family
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
29
involvement. Consequently, weaker support systems among older widow(er)s
could potentially have a more harmful effect on health.
The effect of widowhood may be also related to duration of widowhood. It
may be the change from the married to the widowed status and the timing of being
widowed that lead to a decline in health. Previous studies suggest that it is
important to control for the duration of widowhood when studying the
consequences of widowhood (Avis et al., 1991). Most studies report that the
recently widowed are more likely to have negative health outcomes than the longer-
term widowed (Dennerstein et al., 1998; Ferraro et al., 1984; Hisnanick, 1994).
One study done by De Leon et al. (1994) showed that depression scores increased
during the first year of bereavement, but generally returned to pre-widowhood
levels thereafter.
Change o f Marital Status / Duration of Marital Status
Changes in marital status that reflect marital dissolution may have negative
effects on health. The effect of changed marital status on health may not be simply
due to the protective effects of marriage, but may also be due to the long-term
detrimental effects of marital dissolution on health (Tucker et al., 1999). However,
chronicity or long duration of a situation affords people the opportunity to make
adaptations and to reduce the effect of the situation (Keith & Lorenz, 1989). In
comparisons o f the divorced and the never married elderly, a majority of the
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
30
divorced elderly might be long-term divorced because of their advanced age, and
the never married have obviously been in this state since the beginning of
adulthood. When these long-term divorced and never-married elderly were
compared, there were as many similarities as differences between the two groups
(Choi, 1996). In her study, Choi (1996) found that the never married and the
divorced did not differ from each other in health status, and that any differences
between the divorced and never married were more likely to stem from different
individual attributes (e.g. education, gender, race, home ownership) than from
continuity versus discontinuity in their marital status.
Several studies show that the recently widowed are more likely to have
negative health change than the long-term widowed (Ferraro et al., 1984).
Dennerstein et al. (1998) reported that women widowed in the last 12 months had
lower self-related health and were more likely to report that they were stressed
about their health while women widowed longer than 12 months did not score
significantly differently from married women on these parameters. In another
study, having been widowed in the past 12 months was associated with the
likelihood of moving into lower state of health status or well-being (Hisnanick,
1994).
Umberson (1992) found that individuals who shifted from being married to
unmarried exhibited more negative health behaviors than their counterparts who
remained married, and these effects were greater for men than women.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
31
Furthermore, a comparison to the long-term unmarried suggests that those
individuals who have recently experienced divorce or widowhood exhibit a change
in health behaviors not simply in response to the stress of marital dissolution, but in
response to the absence of a potential source of social control (Umberson, 1992).
These results provide support for the hypothesis suggesting that recent change of
marital status (i.e. becoming unmarried) may be related to negative health effects.
On the other hand, becoming widowed is not totally negative in consequences. In
her study, Lopata (1979) found that in early widowhood, over a third of the widows
recalled that they received help from siblings, and 11% from children. Thus, early
widowhood may improve family relationships and social support to the extent that
they can modify the potential adverse effect o f bereavement on health. However,
for the widowed, the increased social ties upon bereavement may be reduced as
widowhood lasts longer (Heinemann, 1983; Lopata, 1973).
Living Arrangements and Health
One of the widely cited explanations for the ways in which marriage may
enhance health is derived from Durkheim (1897/1952, cited in Cramer, 1993), in
which the social integration and regulation provided by marriage is thought to
enhance health through such processes as greater social interaction, shared beliefs,
commitment, intimacy and mutual obligations and rewards. It has been argued that
if social integration and regulation facilitate health, then those living alone will be
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
32
less healthy than those living with others (Anson, 1989; Hughes & Gove, 1981).
Living with other people is more likely to lead to greater social interaction and
mutual responsibilities than living alone. As the unmarried are more likely than the
married to be living by themselves, the poorer health of the unmarried may be
partly due to living on their own. However, Anson’s (1989) study is limited in that
it compared the health of women either living with or not living with another adult
who was not her child. Consequently, this study did not properly investigate the
connection between health and living alone because women not residing with an
adult may have been living with a child. Furthermore, the effect of being married
was not excluded from the analysis, which contained a large proportion of married
women (Cramer, 1993).
Some studies found that the health status of co-habiters is generally closer
to the married than to unmarried groups for both sexes (Murphy et al., 1997). One
possible mechanism o f the impact of living arrangements on health may be health
service utilization: Those who live alone are more likely to use no services than
those living with others (Webber et al., 1994).
On the other hand, Joung et al. (1994) found that the excess health risks of
the never married, widowed and divorced decrease to some degree after controlling
for living arrangements, but there are still statistically significant morbidity
differences by marital status, in particular the divorced still had higher morbidity
rates than married people. Rogers (1996) also argues that it is not family living
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
33
arrangements that influence the risk of mortality, but rather the family rearranges
itself to deal with ill health and disability.
Cramer (1993) compared the health status of those living alone with that of
those living with another persons in the four unmarried categories of never married,
separated, divorced and widowed for men and women separately. The main
differences found were that women living alone drank more alcohol than those co
habiting or living with a child aged 4 or under, and that widowers living alone were
more psychologically distressed than those living with an adult child, even when
social integration was controlled. The results suggested that living alone was not
generally associated with poor health and so could not account for the difference
between the married and unmarried (Cramer, 1993). But the sample of Cramer’s
study makes it hard to predict the results of this study because it includes younger
adults.
Conceptual Frameworks
Figures 2.1 and 2.2 indicate the conceptual framework underlying the
relationship of marital status and change of marital status to health outcomes
proposed for this study.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
34
Figure 2.1. A Model of the Effect of Marital Status on Health Outcomes
iving Arrangement!
Health
Behavioi Social
Suppo
Marriage
Health Can
Utilization Economic'
Resources
Figure 2.2. A Model of the Effect of Change of Marital Status on Health Outcomes
Health
Outcomes
Change in
Marital Status
Change in Social Support
Economic Resources
Health Behaviors
Research Questions and Hypotheses
There are four primary research questions addressed in this study. Previous
research shows the beneficial effect of being married on health status compared to
being unmarried. But some research shows no relationship. Also, in the
relationship between marital status and health, most previous research findings
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
35
report that men tend to receive more health benefits from marriage than women.
This study re-examines whether marital status affects health outcomes of older
people and whether the relationship differs by gender.
Only limited previous research has specified the type of health risks in the
relationship between marital status and health. Most previous studies have
examined selective health outcomes. While some research findings showed the
differential effect of marital status on physical health and mental health
(Sherboume & Hays, 1990), none of the studies have attempted to show whether
marital status affects various types o f health conditions in different ways. Thus,
this study examines how married people differ from unmarried people (never
married, divorced/separated and widowed) in the prevalence and onset o f a number
of different diseases (hypertension, cancer, diabetes, lung conditions, heart
conditions, stroke, arthritis, broken hip/bone, and depression), functioning
problems and mortality.
While previous research findings demonstrated possible mechanisms
whereby marital status affects health, they do not provide a comprehensive set of
explanatory factors in the relationship between marital status and health. Thus, this
study attempts to incorporate a variety of factors to examine the mechanisms by
which marriage affects health outcomes. The mechanisms are also examined
separately by gender.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
36
As discussed previously, change in marital status may be as important as
marital status per se in health status. This study examines how recent and longer-
term widowhood affects health relative to continuous marriage.
Finally, this study examines how living arrangements affect health. Some
studies have attempted to examine whether part of the effect of marital status on
health occurs through living arrangements. Findings were mixed such that some
studies showed a close connection between the effect of being married and living
with others on health while other studies found that living arrangements did not
influence health differences by marital status. However, many studies were not
methodologically correct in differentiating the effect of marital status and that of
living arrangements. Thus, this study attempts to re-examine the effect of living
arrangements on health independent from marital status.
In order to examine the above four research questions, the following six
hypotheses are developed. The first hypothesis is that married elderly people are
likely to have better health outcomes than their unmarried counterparts. Married
older persons are more likely to have better functioning; less likely to have diseases
and conditions; less likely to get new diseases; and less likely to die than the
unmarried. Specifically, the degree of effect is likely to be greater for fatal
conditions (hypertension, diabetes, cancer, lung conditions, heart conditions and
stroke) than nonfatal conditions (arthritis and fracture); greater for functioning
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
37
(ADL, IADL and Nagi difficulties), mental health (depression) and death than other
physical conditions including fatal and nonfatal conditions.
The second hypothesis is that there are likely to be differences among the
three unmarried groups in health outcomes: The never married are likely to have
better health than the widowed and the separated/divorced. Since the never married
do not experience marital dissolution that the widowed and the separated/divorced
do, which tends to have harmful impact on health status, and the stability of their
marital status over the life course facilitates developing strategic coping
mechanisms in a later life stage when health is worsening, the never married elderly
are more likely to have better health status than other unmarried groups of the
elderly.
The third hypothesis that the relationship between marital status and health
outcomes is likely to be affected by several covariates such as social support, health
care utilization, economic resources and health behaviors. Specifically, different
mechanisms are likely to affect the relationship between marital status and different
health outcomes: Health behaviors are more likely to be associated with the
relationship between marital status and physical conditions (hypertension, diabetes,
cancer, lung conditions, heart conditions, stroke, arthritis, fracture, ADL, LADL and
Nagi difficulties, and death) and death; social support is more likely to be
associated with the relationship between marital status and fatal conditions,
functioning limitations, depression and mortality than nonfatal conditions; health
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
38
care utilization is more for fatal conditions, functioning limitations and death; and
economic resources are more for fatal conditions and death.
The fourth hypothesis is that men are more likely to get health benefits from
marriage than women, and factors in the third hypothesis that contribute to the
effect of marital status on health outcomes will differ for men and for women:
Availability o f financial resources due to marriage and health care utilization are
more likely to be related to better health outcomes for women while social support
and health behaviors are more likely to influence men’s health outcomes.
The fifth hypothesis is that recent change in marital status is likely to be
related to worsening health: Recently widowed elderly people are likely to have
worse health than the continuously married and the longer-term widowed. This
includes better functioning, fewer diseases and conditions, and fewer new diseases
and conditions. Also, the relationships between widowhood and health outcomes
are likely to be affected and explained by several covariates such as social support,
health care utilization, economic resources and health behaviors.
The last hypothesis is that those living alone are more likely to have worse
health outcomes than those living with others regardless of marital status: Those
unmarried who live alone are more likely to have worse health than those
unmarried who live with others whether they are never married, separated/divorced,
or widowed.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
39
III. Methods
This chapter begins with the description of the datasets and the samples
used in this study and discussion of the advantages of using multiple datasets.
Followed is the description of the variables employed and how they are re
constructed for this study. It describes how marital status is constructed for each
analysis, and then description of health outcomes examined follow. After
explaining the construction of each mechanism and changes in mechanisms. It
finally discusses the analysis plan.
Data and Subject Selection
Data for this study come from two waves of the Health and Retirement
Survey (HRS), three waves of the Asset and Health Dynamics Among the Oldest
Old (AHEAD)1 and three waves of the Mac Arthur Research Network on
Successful Aging Community Study.
Health and Retirement Survey (HRS)
The HRS is a national panel study with an initial sample of 12,652 persons
from the 1931-41 birth cohort and their spouses in 7,600 households in 1992. The
baseline interview was conducted face-to-face in the home, and follow-up
interviews by telephone every second year, with proxy interviews after death. Now
1 While waves 1 (1993) and 2 (1995) are the public release data, the third wave (1998), joint data
with HRS, is the preliminary release data.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
40
4 waves (1992, 1994, 1996 and 1998) are available. The survey topics include
demographic background, health and cognitive conditions and status, disability,
family structure and transfers, retirement plans and perspectives, employment
status and job history, income and net worth, and health insurance and pension
plans. The HRS study was merged with AHEAD into a single data collection effort
and instrument, which includes Wave 4 of the HRS and Wave 3 of the AHEAD.
This study employs the data for those 51 to 61 from the first 2 waves with 9,741
respondents for death analysis and 8,862 respondents of the same ages for other
analyses (Figure 3.1).
Figure 3.1. Sample Construction: HRS
Wave 1
12,652
Wave 1
9,741
Waves 1 & 2
8,862
Ages 51 to 61
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
41
Asset and Health Dynamics Among the Oldest Old (AHEAD)
The AHEAD is a nationally representative prospective panel survey of
community-dwelling elderly bom in 1923 or earlier. The AHEAD is a biennial
study which at the 1993 baseline interviewed 8,222 individuals at ages 70 and over
and their spouses. Blacks, Hispanics and residents of Florida were oversampled to
understand the changing demographic composition and geographic distribution of
older persons. Although the same instrument was administered to all self-
respondents, those aged 70-79 were generally interviewed by telephone, while
those aged 80 and over were generally interviewed in person. This survey includes
information on demographic structure, family structure and family transfers,
physical, functional and cognitive health domains, current income and its sources,
assets, housing, employment, use of health care services, and out-of-pocket
expenditures (Soldo et al., 1997).
Among 8,222 people who were interviewed in 1993, about 91% were aged
70 and over. In 1995, 1,271 of these people were not interviewed because they
were either missing (n=529) or dead (n=742), and 89 people were added to the
survey. There were 5,172 respondents who were over age 70 in 1998. Altogether,
5,037 people were eligible for this study who participated in all three waves (Figure
3.2). To examine the effect of four different marital groups on health, this study
uses 4,348 respondents at ages 70 and over in 1993 who were interviewed at all
three times and whose marital status had not changed for analyses of all health
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
42
Figure 3.2. Sample Construction: AHEAD
Over Age 70
Wave 1
8,222
Wave 1
7,443
Waves 1, 2 & 3
5,037
Waves 1 & 3
5,172
Waves 1 & 2
6,231
outcomes except death. The unavailability of death records for third wave limited
death analyses to the first two waves.
For the analyses of widowhood, this study uses information from 4,485
respondents aged 70 and over at baseline who were interviewed at all three times
and who remained married, who became widowed between 1995 and 1998, or who
remained widowed over the study or became widowed between 1993 and 1995.
This means that persons who remained separated/divorced or who became
separated/divorced are eliminated from the analysis.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
43
Only unmarried persons whose marital status had not changed (N=2,218)
were used for the analyses of living arrangements because by doing this, the effect
of living alone independent of marriage would be examined.
MacArthur Research Network on Successful Aging Community Study (MacArthur)
The MacArthur is a longitudinal study of three large, population-based
cohorts aged 70-79 in Durham, North Carolina, East Boston, Massachusetts, and
New Haven, Connecticut. Subjects were drawn from the Epidemiologic Study o f
the Elderly (EPESE) at each site.
In 1988, those between ages 70 and 79 (N=4,030) were screened on the
basis o f six criteria o f physical and cognitive function to identify a high-functioning
cohort. The group represents those in the top third of their age group in functioning
ability. The six selection criteria are: (1) reports of no disability on the seven-item
Activities of Daily Living (ADL) (Katz et al., 1963); (2) reports of no more than
one mild disability on eight items measuring gross mobility and range of motion
developed by Rosow and Breslau (Rosow-Breslau, 1966) and Nagi (Nagi, 1976);
(3) scores of six or more correct on the nine-item Short Portable Mental Status
Questionnaire (Pfeiffer, 1975); (4) being able to remember three or more of six
elements on a delayed recall of short story; (5) being able to hold a semi-tandem
balance for at least 10 seconds; and (6) being able to stand from a seated position
five times within 20 seconds.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
44
At screening, 1,313 subjects met the criteria for the higher functional status
group and 1,189 (90.6%) among them agreed to participate in the study. Baseline
data collection was completed between May 1988 and December 1989 and these
persons were followed over seven years and re-interviewed in 1991 and 1995. O f
the original sample of 1,189 in 1988, 71 respondents (6%) died between 1988 and
1991, and 47 respondents (4%) refused the follow-up interview in 1991. Twenty
three percent (N=273) of the original respondents in 1988 died between 1991 and
1995 follow-up interviews.
Figure 3.3. Sample Construction: MacArthur
Wave 1
1,354
Wave 1
1,189
Waves 1 & 2
1,118
Waves 1, 2 & 3
916
Ages 70-79
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
45
Among those who participated in all of the three waves (N=916) (Figure 3-
3), this study utilizes 686 people whose marital status had not changed throughout
the study term for the analyses on four different marital groups. For widowhood
analysis, 921 people from those who participated both in waves 1 and 2 and
remained married or became widowed between 1988 and 1991, or who widowed
before 1988 are used.
Advantages o f Using Multiple Datasets
Utilizing multiple datasets can be beneficial in two different ways. One
dataset may not provide all information of interest. Each dataset can provide
different information related to its specific focus, which can fill a hole that might
exist with a single dataset. It should also be noted that comparison of the results
from multiple datasets may provide contradictory outcomes, which could impose
challenges in conceptualization, methodological approach and interpretation o f
results.
There are several advantages to supplementing the AHEAD dataset with
information from the MacArthur and HRS data. First, results of analyses with
these three datasets are expected to show interesting comparisons given their
sample and geographic characteristics: AHEAD and HRS as a nationally
representative sample versus MacArthur with a high functioning sample in specific
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
46
local areas; the very old sample o f the AHEAD and MacArthur versus the pre-
retirement-age people of the HRS.
Second, high functioning status at baseline, longer-term tracking of the
sample (7 years) and the low attrition rate among the MacArthur sample would aid
the interpretation of causal relationships among marital status, socioeconomic and
psychological factors and onset or prevalence of diseases/conditions and functional
difficulties (Seeman et al., 1996).
Third, the MacArthur data include biological measures o f some health
indicators (e.g. blood pressure, blood sugar) and various social and psychological
variables that are proposed to be important in explanation of the relationship
between marriage and health. Self-reported measures of health conditions may be
somewhat inaccurate due to misreporting. Biological measures o f health conditions
can provide more accurate status of health conditions. Also, various social support
variables such as instrumental and emotional support in MacArthur can provide
valuable information on areas not covered in HRS and AHEAD.
Fourth, the relatively younger age of the initial HRS sample is appropriate
for analysis of the incidence of health problems because rapid onset o f health
problems including functional difficulties occurs during the years after age 50
(Crimmins & Kim, 2000).
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
47
Measures
The effect of marital status on health outcomes is examined in the context
of gender differences, while controlling for a broad set o f variables that are likely to
affect health outcomes. The descriptions of the variables in the HRS, the AHEAD,
and the MacArthur are presented in Table 3.1.
Marital Status
Marital status is divided into four categories: Currently married, never
married, separated/divorced and widowed. In analyses, these variables are used as
dummy variables with variables for never married, separated/divorced and
widowed and with married as the reference category in order to examine possible
differences in the effect of being in each unmarried group on health outcomes.
For the analyses of widowhood, marital status is divided into three categories:
Continuously married, recently widowed and longer-term widowed. The first
group is those who remained married throughout the period studied. The recently
widowed are those who became widowed within the last two to three years. Recent
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Reproduced with permission o f th e copyright owner. Further reproduction prohibited without permission.
Table 3.1. D escription of V ariables U sed: H R S, AH EAD an d M acA rth u r
V ariables--------- -D ataset H R S AH EAD M acA rthu r
1 . M a rita l S tatu s C u rrently m arried , never m arried , separated/divorcer , w idow ed
2. C hange of M arital S tatu s: R em ained m arried since 1992 R em ained m arried since 1993 R em ained m arried since 1988
R ecency o r D u ration of
Being W idow ed
versus
R ecently w idow ed: Becam e
widowed betw een 1992 an d 1994
versus
R ecently w idow ed: Becam e w idow ed
betw een 1995 an d 1998
versus
R ecently w idow ed: Becam e
widowed betw een 1988 an d 1 9 9 1
L onger-term w idow ed: W idow ed
before 1992
L on ger-term w idow ed: W idow ed
before 1993 or widowed betw een 1993
an d 1995
L onger-term w idow ed: W idow ed
before 1988
Living A lone
Age 51-61 a t baseline 70 an d over a t baseline 70-79 a t baseline
G ender M ales, fem ales
R ace W hites an d others (blacks, H ispanics an d others) W hites an d blacks
E ducation Y ears of education
Self-rated health a t baseline: Excellent, very good, good, fa ir, p oor Self-rated health a t baseline:
Excellent, good, fa ir, poor, b a d
Prevalence H aving diseases/conditions and functioning
In 1992 In 1993 In 1988
Incidence A cquisition of new diseases/conditions an d functioning problem s am ong those initially free of problem s
Between 1992 an d 1994 Between 1993 an d 1998 Between 1988 an d 1995
1 . D iseases/C onditions H as a doctor ever told you th a t you have (had) H ypertension or high blood pressure, C ancer, D iabetes or high blood
su gar, H eart conditions (only m yocardial infarction in M acA rthur), Stroke, B roken Bone (H R S) / H ip (AH EAD,
M acA rth u r)
L ung conditions; A rth ritis or rheum atism
D epression (CESD 8 item s >=3) (Yes/No): (M uch of the tim e during the p ast week) I felt depressed; 1 felt th a t
everything I d id was a n effort; M y sleep w as restless; I was happy (variable reversed); I felt lonely; I enjoyed life
(variable reversed); I felt sad ; I could not get going
■ fc .
oo
Reproduced with permission o f th e copyright owner. Further reproduction prohibited without permission.
T able 3 .1 continued
2. F unctioning
ADL D ifficulties 4 item s
(1-Any difficulty in a t least one
item or don’t know ; 0-No
difficulty)
(H ow difficult is it for you to)
bathe or show er w ithout h elp ;
eat w ithout h elp; dress w ithout
help; get in and out of bed
w ithout help
6 item s
(1-Any difficulty in a t least one item ; 0-
No difficulty)
(Even when som eone helps you; even
w hen using the equipm ent; w ithout any
help or special equipm ent) Do you have
any difficulty w alking across the room ;
getting in an d out of bed
(Even w hen som eone helps you; w ithout
any help) Do you have any difficulty
dressing; bath in g; eating; using the toilet
IA D L D ifficulties 3 item s
(1-Any difficulty in at least one
item ; 0-No difficulty or don’t
do)
(H ow difficult is it for you to)
m ake telephone calls; tak e
m edications; m anage m oney
5 item s
(1-Any difficulty in a t least one item ; 0-
No difficulty)
(1993) A re you able to (w ithout help)
(1998) (Because of a health or m em ory
problem ) Do you have any difficulty
p rep are hot m eals; shop for groceries;
m ake telephone calls; tak e m edications;
do you m anage your m oney such as
paying your bills and keeping tra ck of
expenses w ithout anyone’s help
v o
Reproduced with permission o f th e copyright owner. Further reproduction prohibited without permission.
T able 3 .1 continued
N agi D ifficulties 4 item s
(1-Any difficulty in a t least one
item or don’t know ; 0-No
difficulty or don’t do)
(H ow difficult is it for you to)
pull or push large objects like a
living room ch air; stoop, kneel,
or crouch; lift or carry w eights
over 1 0 pounds like a heavy bag
of groceries; reach or extend
your arm s above shoulder level
4 item s
(1-Any difficulty in a t least one item o r
can’t d o; 0-No difficulty or don’t do)
pulling or pushing large objects like a
living room ch a ir; lifting or carryin g
w eights over 1 0 pounds like a heavy bag
of groceries; stooping, kneeling, o r
crouching (1998 only); reaching or
extending your arm s above shoulder level
(1998 only)
5 item s
(1 -Som e or a lot of difficulty, or
unable to d o; 0-No difficulty)
(H ow m uch difficulty, if any, do
you have) pulling or pushing large
objects like a living room ch a ir;
stooping, crouching or kneeling;
lifting or carrying weights over 1 0
pounds like a very heavy bag of
groceries; reaching or extending
arm s above shoulder level; either
w riting or handling sm all objects
3. D eath D ied betw een 1992 an d 1994 D ied betw een 1993 an d 1995
1 . Social S u p p ort
N um ber of household m em bers
(besides self an d spouse)
Satisfaction w ith neighborhood
Satisfaction w ith friendships
(V ery or som ew hat satisfied vs.
evenly satisfied, som ew hat or
very dissatisfied)
N um ber of living children
N um ber of siblings
N um ber of household m em bers (besides
self an d spouse)
N um ber of living children
C hildren died (Yes/No)
C hildren live nearb y (Yes/No)
N um ber of relatives live n earby
N um ber of close friends n earb y
In stru m en tal su o D o rt
H ow often you count on spouse,
kids, friends/relatives to shop a n d
to give advice
O
Reproduced with permission o f th e copyright owner. Further reproduction prohibited without permission.
T able 3.1 continued
E m otional s u d d o i I
H ow often kids, spouse, friends/
relatives m ake you feel loved;
listen your w orries
(N ever, rarely, som etim es,
frequently, no need)-sum of 6
item s divided by num ber of item s
answ ered
2. H ealth C are U tilization A t least one doctor visit in the past 1 2 m onths (Yes/No)
3. Econom ic R esources Incom e
Incom e p er adult equivalent (in
the sam e household, the first
adult -a value of 1.0 an d all
subsequent adults-a value of 0.8)
Incom e per adult equivalent (1993)
R espondents’ an d their spouses’ incom es
from Social Security, Supplem ental
Security Incom e an d V eterans Benefits in
the past m onth * 1 2 (1995,1998)
R espondents’ and their spouses’
incom e for the past year, including
incom e from all sources such as
wages, salaries, Social Security o r
retirem en t benefits, help from
relatives, ren t from property, etc.
($ 1000, $3500, $6000, $8500,
$12500, $17500, $22500, $27500,
$32500,$37500,$42500,$47500,
or $52500)
A vailability of supplem ental h ealth insurance
4. H ealth B ehaviors C u rren t sm oking (C u rren t sm oker or not)
H eavy drinking (3 or m ore drink s a day) H eavy drinking (2 or m ore liquor
d rin k s a t one tim e)
O besity (B M I greater th a n 30)
Physical activity/E xercise:
P articipating in light physical
activity such as w alking,
dancing, gardening, golfing,
bow ling, etc. 3 or m ore tim es a
week
Physical (leisure) activities:
W hether they ever do leisure
activities like taking walks, going
swim m ing or golfing
W earing a seatbelt (1995, 1998) (Yes/No)
52
widowhood is defined as becoming widowed between 1992 and 1994 in HRS,
between 1995 and 1998 in AHEAD, and between 1988 and 1991 in MacArthur.
Those who became widowed between 1993 and 1995 in AHEAD were regarded as
the longer-term widowed. The longer-term widowed are those who were widowed
before the first wave of each dataset and stayed widowed since then or those who
were widowed between the first and the second wave (AHEAD).
For the analyses of the relation between living arrangements and health,
three unmarried groups were used in order to eliminate the possible interaction
effect between being married and living arrangements on health outcomes and to
see whether living arrangements affect health independently from marital status.
Dummy variables indicating being never married and separated/divorced were used
in the model and being widowed was omitted as the reference.
Health Outcomes
Three sets of health outcomes are used in this study: Diseases and
conditions, functioning and death. This study includes a number of indicators to
measure health. Even though using various conditions is complex and less
parsimonious, it is important to note its advantage in that the relation between
marital status, health and gender may vary according to the type of health problems
(Riessman & Gerstel, 1985). The prevalence (at baseline) and incidence (between
the baseline and final wave) of eight physical chronic diseases and conditions such
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
53
as hypertension,1 cancer, diabetes, lung conditions (only in HRS and AHEAD),
heart conditions, stroke, arthritis (only in HRS and AHEAD) and broken hip (in
AHEAD and MacArthur) or broken bone (in HRS) are examined in order to see
differences in health status for the recently widowed, the longer-term widowed and
the continuously married. For widowhood analysis, the prevalence of conditions at
the end of study was examined. For incidence of diseases/conditions and
functioning problems, only those who did not have these conditions at baseline
were used for analysis. Depression is a mental health outcome which can be
measured in all three samples. It is indicated by whether a respondent has
symptoms addressed in a short version of the CESD scale (8 items).2
Functioning problems are indicated by having difficulty in performing at
least one selected ADL, LADL or Nagi item because of a health problem in HRS
and AHEAD and at least one Nagi item in MacArthur.3 Questions used to indicate
functioning difficulties are somewhat different in the three datasets: Functioning
difficulties in HRS and AHEAD are measured by performance difficulty with or
1 The self-reported hypertension question used in M acA rthur 1995 data was not asked in the Duke
site. Thus, 354 people were missing in self-reported hypertension question in 1995. In order to
eliminate the potential bias made in the analysis with the missing respondents omitted, several tests
were done: (1) Assuming all missing cases in self-reported hypertension in 1995 as having
hypertension; (2) Assuming all missing cases in self-reported hypertension in 1995 as not having
hypertension; and (3) Using measured high blood pressure combining sum of 3 sitting systolic and
diastolic blood pressures (>=140/90). The comparison with these three different ways of estimating
hypertension prevalence in 1995 did not change the results of logistic regression.
2 In order to unify the cutoff point, the closest num ber, three, to the top quartile of the distribution
for both men and women in all datasets was used. In this study, depression refers to depressed
mood or higher depressive symptoms rather than clinical depression.
3 Distribution of ADL difficulties among the M acA rthur respondents is extremely skewed, thus
ADL difficulties were excluded for analysis in M acA rthur.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
54
without help while need for help or inability in performing activities is asked in
MacArthur. In HRS, ADL difficulties were asked under the wording of ‘without
help’ in 1992 and 1994. In AHEAD, IADL questions at baseline excluded the
ability to perform activity with help while follow-up surveys asked the need for
help in separate questions. Thus, baseline IADL performance ability in AHEAD
may be somewhat underestimated and ADL difficulties in HRS and the onset of
IADL difficulties between 1993 and 1998 in AHEAD overestimated. Death is used
as a health outcome in HRS and AHEAD,4 but not in MacArthur due to the
restriction on usage of death as an outcome in the survey.
Mechanisms by which Marital Status Affects Health: Social Support, Health Care
Utilization, Economic Resources and Health Behaviors
This section describes four proposed mechanisms such as social support,
health care utilization, economic resources and health behaviors, and how changes
in those mechanisms are made.
(1) Social Support
The number of people in the household in HRS and AHEAD is an indicator
of social support. The number in the household does not include a spouse, meaning
that the number includes household members other than a respondent and his/her
spouse. The number of living children and siblings is also used in AHEAD as
4 Because of an unavailability of a tracker file for wave 3 (1998) of AHEAD, death data for
AHEAD were drawn from the death records for the years between 1993 and 1995.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
55
measures of social support. Indications of social integration are included in HRS
such as degree of satisfaction with neighborhood and friendships. Five response
categories such as very satisfied, somewhat satisfied, evenly satisfied, somewhat
dissatisfied and very dissatisfied were divided into two categories, the first two as
satisfied group, the other three categories as less satisfied or unsatisfied group. In
MacArthur, various types of social support such as whether children live nearby,
the number of children and the number of relatives and close friends who live
nearby, and the degree of instrumental and emotional support received from a
spouse, children, other relatives and friends were used to indicate social support.
Specifically, questions used for instrumental support are: How often can
you count on your spouse to help with daily tasks like shopping, giving you a ride
or helping you with household tasks; how often does your spouse give you advice
or information about medical, financial, or family problems; how often can you
count on your child(ren) to help with daily tasks like shopping, giving you a ride or
helping you with household tasks; how often does/do your child(ren) give you
advice or information about medical, financial, or family problems; how often can
you count on your close friends and relatives to help with daily tasks like shopping,
giving you a ride or helping you with household tasks; and how often do these
friends and relatives give you advice or information about medical, financial, or
family problems. The same procedure was done for emotional support with 6
questions: How often does your spouse make you feel loved and cared for; how
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
56
often is your spouse willing to listen when you need to talk about your worries or
problems; how often does/do your child(ren) make you feel loved and cared for;
how often is/are your child(ren) willing to listen when you need to talk about your
worries or problems; how often your close friends and relatives make you feel
loved and cared for; and how often are these friends and relatives willing to listen
when you need to talk about your worries or problems. The response categories are
never (1), rarely (2), sometimes (3), frequently (4) and no need (5). The values of 1
and 5 (never and no need) are recoded to 0, and 2 to 1, 3 to 2, and 4 to 3, and then
the values were summed and divided by number o f items answered. The score for
the instrumental and emotional support measures ranges from 0 to 3, the higher
score indicating more support.
Utilizing three datasets allowed various dimensions of social support to be
examined in this study: Structural support such as the number of household
members, living children and siblings, close friends and relatives, instrumental and
emotional support. By merely assessing the frequency or amount of social support
that has been exchanged, the potentially important role played by variations in
individual need for support may be overlooked (Krause, 1987). Even though
functional support5 such as satisfaction with friendships and neighborhood were
used in this study, more variables assessing perceived adequacy of social support
would have been helpful. Also, beneficial effects of providing social support were
5 Functional support refers to a person’s appraisal of available support or the extent to which one’s
interpersonal relationships provide particular resources
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
57
not tapped in this study. Giving support to others or mutual support might provide
as important health benefits as receiving social support (Krause et al., 1992).
(2) Health Care Utilization
For health care utilization, whether a respondent had at least one doctor visit
in the past 12 months is used in HRS and AHEAD. This is not available in
MacArthur. This study originally attempted to examine the effect o f health care
utilization in the relationship between marital status and health. The logic is that
preventive care such as routine medical and dental exams and stress management or
earlier health care usage are likely to reduce health risks. However, this is difficult
because doctor visits are not an appropriate measure of the use o f preventive care
that helps to monitor health: More doctor visits are related to more health problems.
Ever use of doctor in the past year may simply indicate whether one could be told
by a doctor that he/she has a disease, thus showing the reverse causation. Thus,
this study reported percent of those who had at least one doctor visit in the past
year by marital status, but did not include health care utilization in logistic
regression models.
(3) Economic Resources
Economic resources are measured by income in all o f the three datasets and
health insurance such as Medigap or other supplement plans in HRS and AHEAD.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
58
As shown in Table 3.1, the HRS, AHEAD and MacArthur utilized different
methods of income calculation. The HRS used income per adult equivalent (LAE)
following Fuchs (1986). LAE provides a better measure of differences among
households by allowing for variations in household size and composition and for
economies of scale in consumption (Easterlin et al., 1993). Household’s total
income from all sources was taken and divided by the number of household
members, converting to an adult equivalent basis. The number of adult equivalents
in a household was obtained by giving the first adult in a household a value of 1.0
and all subsequent adults a value of 0.8. The same conversion to per adult
equivalent basis was used for income in the first wave of the AHEAD (1993). For
the 1995 and 1998 income, the calculated measure indicating total household
income from all sources is not yet available, thus income from only specified
sources such as Social Security, Supplemental Security Income and Veterans
Benefits was used (details follow below). IAE was not able to be calculated in
MacArthur, thus MacArthur used total income from all sources such as wages,
salaries, Social Security or retirement benefits, help from relatives and rent from
property. Income for 1995 and 1998 in AHEAD and income in MacArthur were
that of a respondent’s and his/her spouse’s.
Also, while the HRS and AHEAD asked the exact amount o f income, the
question in MacArthur was asked with 13 response categories such as less than
$1,999, $2,000 to $4,999, $5,000 to 6,999, etc. The mid-point of each response
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
59
category was assigned to respondents to represent each category (e.g. $ 1,000 for
less than $1,999, $3,500 for $2,000 to $4,999, etc.). Economic resources measured
by income and private health insurance in this study might have been better
represented by adding other indicators such as assets (e.g. home ownership) and
pension.
(4) Health Behaviors
Health behaviors include current smoking status, heavy drinking and
obesity indicated by a body mass index greater than or equal to 30 (World Health
Organization, 1977). The HRS/AHEAD and MacArthur use different measurement
to indicate heavy drinking: The HRS and AHEAD define heavy drinkers as those
who drink any alcoholic beverages such as beer, wine or liquor three or more times
a day while in MacArthur, heavy drinking is defined as two or more liquor drinks
at one time. Physical/leisure activities are an additional measure of health behavior
in HRS and MacArthur. In HRS, light physical activities such as walking, dancing,
gardening, golfing and bowling were used as an indicator o f physical activity. The
question originally asked how often they participated in the above activities, but for
this study, the response was divided into two categories indicating participation in
activities three or more times a week or participating less than that. In MacArthur,
a respondent was asked whether they ever do leisure activities like taking walks,
going swimming or golfing (yes/no).
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
60
The use of a broad range of health behavior measures would have better
indicated health behaviors. Environmental safeguards such as wearing a seatbelt
and equipping home safety feature in addition to smoking, drinking, obesity and
exercise would have represented a fuller range of health behaviors (Schone &
Weinick, 1998).
(5) Change in Mechanisms
Change in the above mechanisms over time is also examined in order to see
how changes occur upon the loss of a spouse and how they affect health status.
Changes in health behaviors include becoming a current smoker, becoming a heavy
drinker or becoming obese. In AHEAD, decrease of seatbelt use is additionally
used to indicate changing health behavior.6 Changes in economic resources are
measured by loss of coverage by any supplemental health insurance and reduced
income (in AHEAD, changes in income only from specified sources). In AHEAD,
for income, both baseline income and income change between 1995 and 1998 were
used in logistic regression models. For income change, a respondent’s and his/her
spouse’s income from Social Security, Supplemental Security Income and Veterans
Benefits was used because of the unavailability o f computed total household
income like baseline income as yet. Even though income from these three sources
is not complete, the majority of respondents were already retired and thus, most of
6 Question about wearing a seatbelt, which is not available in wave 1 of AHEAD, is added to the
second and third waves. Thus, this variable was used only for widowhood analysis.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
61
their income had come from retirement benefits and welfare if any. In fact, Smith
and Kington (1997) argued that at the advanced ages of the AHEAD sample,
retirement income offered the best hope of estimating an income effect on health.
Also, controlling for baseline income that includes all other sources of household
income would reduce the bias that might result from using selective sources of
income. While per capita income is a better measure of income, some of the
income sources such as Social Security and Supplemental Security Income are
adjusted by marital status, thus income of the 1995 and 1998 AHEAD and income
of MacArthur, which included only a respondent’s and his/her spouse’s income,
were not converted into per capita income.
A decrease in the number of people in the household, living children and
living siblings, and whether child(ren) was dead were used as indicators of decrease
in social support. Changes in these mechanisms were used in the analyses of
widowhood and health in order to address whether recent and longer-term
widowhood influences health through changes in specific mechanisms.
Background variables such as age, education and race were also included.
To control for the confounding effects of initial health differences on health
outcomes and to prevent a possible marriage selection effect, information on health
status at baseline, self-rated health status, was included in the analysis. As
discussed, self-reported health is an appropriate measure of baseline health status,
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
62
which is not a direct measure of mortality or chronic diseases, but is closely related
to them.
Analysis
In order to determine the relationship of marital status to health outcomes
and the effects of possible confounding factors on the relationship, logistic
regression analysis was used. Except for the analyses on changes in marital status
and widowhood, analyses were done with those whose marital status did not
change between 1992 and 1994 (HRS), 1993 and 1998 (AHEAD), and 1988 and
1995 (MacArthur). For the four marital groups, the married were used as the
reference to compare with three unmarried groups— the never married, the
separated/divorced and the widowed. Several logistic regressions were developed
by introducing additional variables. The first equation included only marital status,
marital status with age was next, and marital status with age, self-rated health,
education and race followed, and then marital status with all other variables
including proposed three sets of mechanisms.
For the widowhood analyses, those who remained married throughout the
term of the study were used as the reference to compare to those who became
widowed recently and who were widowed for longer period. Because of the small
sample size of the recently widowed, the analyses were done with gender included
in the model, and interactions of widowhood with gender were used in order to see
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
63
the gender difference by duration of widowhood. Changes in three sets of
confounding factors were added to the logistic regression models to see their
effects on the relationship between marital status and health.
For the living arrangements analyses, the widowed were used as the
reference group in comparison to the never married and the separated/divorced.
Three groups o f proposed mechanisms by which marital status affects health were
also introduced into the model, first separately, and then together to examine which
mechanisms influence outcomes and how they affect health outcomes when
considered together.
This chapter described the data, sample, measures of variables and plans for
analysis. While there is a room for better measurements for some factors and
construction of variables, the best selection and construction o f variables were
made within the availability of measurements and data.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
64
IV. Description of the Subject Population
This chapter describes the characteristics of the sample that this study is
based on. It first shows mean and percent of all variables used in this study by
gender. And the number and percent in each marital state by sex and year for each
dataset are presented, showing the actual sample that each analysis in the following
chapters are using.
Table 4.1 describes the characteristics of the HRS, AHEAD and MacArthur
samples used for analysis in this study. The mean age of the HRS respondents was
about 56 while that of the AHEAD and the MacArthur respondents was about 76-
77 and 74, respectively. A little less than half of the HRS sample were males while
about 37% of the AHEAD sample and about 40% of the MacArthur sample were
males. About 80% of sample were non-Hispanic whites in all three datasets. The
MacArthur sample consisted of only whites and blacks. Men tended to report
better health in HRS, AHEAD and MacArthur. As expected, the high functioning
MacArthur sample had better health status than the AHEAD sample in self-rated
health, prevalence of diseases and functioning limitations.7 On the other hand, the
MacArthur tended to have higher percentage of developing health conditions,
implying that this high functioning sample might delay the onset of health
problems to later in life relative to the AHEAD sample.
7 In fact, comparison between the HRS, AHEAD and M acArthur should be cautiously m ade given
the different definitions of variables and time period in each dataset.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
65
At baseline, more people in AHEAD than in MacArthur were married
among both men and women; while more people in MacArthur were widowed,
higher percent of the AHEAD respondents experienced recent loss o f a spouse than
those of the HRS or the MacArthur. Among younger respondents from the HRS, a
disproportionate percentage of women (compared to men) were already widowed,
indicating that women experience widowhood earlier than men and, thus, they tend
to experience longer-term widowhood than men. More women lived alone than
men at baseline. A little less than half of women in AHEAD lived alone. Sample
characteristics by marital status are described in other chapters in greater detail.
The number and percent in each marital status by sex is presented in Table
4.2 (HRS), Table 4.3 (AHEAD), and Tables 4.4 and 4.5 (MacArthur). The shaded
parts indicate those whose marital status had not changed throughout the term of
the study. About 82% o f males and 66% of females in HRS remained married.
The percent of those who remained married was much lower among the AHEAD
and MacArthur respondents. About 68% of males in AHEAD and 63% of males in
MacArthur (Table 4.5) were continuously married, as were 28% of females in
AHEAD and 18% of females in MacArthur (Table 4.5).
Only a small number of people in HRS recently experienced the loss of a
spouse: 32 males (0.62%) and 79 females (1.49%). More AHEAD and MacArthur
respondents experienced recent widowhood: 5.18% of males and 5.78% of females
in AHEAD, and 6.94% o f males and 7.82% of females in MacArthur (Table 4.5).
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
66
The difference among the HRS, AHEAD and MacArthur in recent widowhood is
likely to be largely attributable to the age of the sample. However, it may be also
partly attributed to the different definitions of recent widowhood.
Less than 20% of the AHEAD and MacArthur male respondents were
widowed before 1995 and 1991, respectively. On the other hand, over 55% of the
females in the two datasets were longer-term widows. About 10% of female
respondents and 1.5% of males in HRS were longer-term widow(er)s. As
discussed later, in analysis of widowhood, the AHEAD sample who became
widowed before 1995 (N=209) were included with as the longer-term widowed as
shown in small boxes below Table 4.3. The recently widowed are those who
became widowed between 1995 and 1998 (N=2,130).
Table 4.4 shows the number and percent of marital status in 1988 and 1991
in MacArthur. Comparison with Table 4.5 (marital status in 1991 and 1995)
reveals that more people died between 1991 and 1995 than between 1988 and 1991
probably because of the longer-term and increased age. Although three waves of
the data were used in analysis and more cases of recent widowhood between 1991
and 1995 would make analysis statistically more powerful, widowhood analysis
utilized only the first two waves of MacArthur (1988 and 1991) because many
social support variables used are not available in wave 3 (1995). The small box
below Table 4.4 describes the sample structure for widowhood analysis in
MacArthur. Those who became widowed between 1988 and 1991 were used as
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
67
recent widow(er)s (N=34) while those who became widowed before 1988 and
remained widowed until 1991 were the longer-term widowed (N=422). This
number did not come from Table 4.4, which was based on the respondents who
participated in all three waves. On the other hand, the number in the small box
comes from those who participated in waves 1 and 2 whether they participated in
wave 3 or not.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Reproduced with permission o f th e copyright owner. Further reproduction prohibited without permission.
T able 4.1. D escription of Study Sam ple a t Baseline (M ean & P ercen t): H R S, AH EAD and M acA rth u r
H R S (N=8,862)________ A H EAD (N=5,037) M acA rthu r (N=91 6 )
M ales Fem ales M ales Fem ales M ales Fem ales
Age (Baseline) 55.90 55.96 76.16 77.15 73.82 74.35
% Sex 47.53 52.47 37.20 62.80 39.52 60.48
% R ace
W hites 83.86 81.72 82.18 79.51 83.66 79.27
B lacks 8.64 10.43 10.89 1 4 .0 1 16.34 20.73
H ispanics 5.50 5.52 5.55 5.34
—
O th ers 2.00 2.33 1.39 1.14
_ _
E ducation (0-17) 12.26 11.95 11.10 10.94 10.84 10.67
% Self-rated H ealth (Baseline)
E xcellent
24.64 23.31 12.60 11.36 25.14 16.19
V ery G ood
29.37 30.48 25.20 25.83
— _
G ood 27.44 25.29 33.37 30.74 54.97 56.70
F a ir
11.54 13.77 21.04 22.48 19.89 25.72
P oor
7.01 7.15 7.79 9.59 0.00 0.91
B ad — --
0.00 0.00
M arital C haracteristics
% M arital Status (Baseline)
M arried
83.15 68.08 77.37 37.92 71.82 29.84
Never M arried 3.96 3.82 2.20 3.42 5.80 6.87
Separated/D ivorced 11.19 17.25 4.30 6.02 4.70 7.41
W idow ed
1.70 10.85 16.13 52.63 17.16 55.88
% R ecently W idow ed am ong M arried 0.75 2.18 5.1 5.5 2.08 3.76
at Baseline
Living A rrangem ents (Baseline)
% Living Alone
10.05 14.77 19.01 47.01 N /A N/A
H ealth O utcom es (Initial W ave)
Prevalence of Diseases, C onditions and
Functioning D ifficulties
% H ypertension 39.74 39.33 43.83 52.04 43.09 49.46
ON
0 0
Reproduced with permission o f th e copyright owner. Further reproduction prohibited without permission.
T able 4.1 continued
H R S AH EAD M acA rth u r
M ales Fem ales M ales Fem ales M ales F em a le:
% C ancer 3.09 7.59 13.39 11 .7 1 15.04 17.93
% D iabetes 10.98 10.94 11.75 11.45 12.22 11.78
% H ea rt C onditions 15.14 11.06 32.92 26.00 13.93 5.97
% S trok e 3.24 2.59 6.78 6.29 1.95 1.45
% B roken Bone/H ip 13.16 14.26 2.93 4.62 1.67 1.99
% L ung C onditions 7.50 8.56 11.12 8.41 N/A N/A
% A rth ritis 31.07 44.71 19 .9 1 29.51 N/A N /A
% D epression (CESD8>=3) 8.52 12.03 15 .2 1 24.88 7.81 6.68
% A DL Difficulties (4 item s in H R S ; 8.81 10.56 11.32 18.35 — —
6 item s in A H EAD )
% IA D L Difficulties (3 item s in N/A N/A 9.55 15.82 N/A N/A
H R S ; 5 item s in AH EAD)
% N agi Difficulties (4 item s in H R S ; 41.05 62.10 16.44 38.13 11.60 18 .4 1
2 item s (1993) an d 4 item s (1998)
in A H EA D ;5 item s in M acA rthu r)
O nset of Diseases, C onditions an d
Functioning D ifficulties
% H ypertension 6.92 7.32 16.12 21.55 17.02 15.77
% D iabetes 2.64 2.01 5.40 5.06 15.98 14.00
% C ancer 1 .6 1 1.56 8.99 4.24 20.35 11.64
% H ea rt C onditions 4.80 3.50 17.08 12 .0 1 10.32 7.51
% Stroke 0.86 0.68 6.52 6.93 6.82 8.10
% B roken Bone/H ip 3.41 4.39 1.16 3.15 2.66 2.17
% L ung C onditions 2.46 2.19 4.66 4.40 N/A N/A
% A rth ritis 9.80 14.34 47.51 58.47 N/A N/A
% D epression (CESD8>=3) 10.47 17.36 16.80 19.82 49.06 63.12
% A DL Difficulties (4 item s in H R S; 3.34 3.17 26.17 30.44 — —
6 item s in A H EAD )
% IA D L Difficulties (5 item s) N/A N/A 17.27 20.08 N/A N/A
o\
vo
Reproduced with permission o f th e copyright owner. Further reproduction prohibited without permission.
T able 4.1 continued______________________________________________________________________
H R S AH EAD M acA rth u r
M ales Fem ales M ales Fem ales M ales Fem ales
% N agi (4 item s in H R S; 2 item s in 17.39 23.25 22.41 36.18 32.44 50.72
A H EA D ; 5 item s in M acA rthu r)
% D eath (b/wWl andW 2) 2.03 1.28 12.53 9.52 —
. . .
Social Support (Baseline)
% C hildren Live N earby N /A N/A N/A N/A 12.35 1 3 .5 1
N um ber of R elatives N earby — N/A N /A 3.45 2.99
N um ber of F rien ds N earby — — N/A N /A 6.71 4.78
N um ber of H ousehold M em bers 0.90 0.90 0.28 0.37 N/A N /A
(except self & spouse)
N um ber of Living C h ild ren - - 2.89 2.53 2.46 2.32
N um ber of Living Siblings - - 2.17 2.09 N/A N/A
% C hildren D ead N/A N/A 2.35 3.61 N/A N/A
% Satisfied w ith N eighborhood 86.55 85.32 N/A N/A N/A N/A
% Satisfied w ith F riendships 92.07 92.36 N/A N/A N/A N /A
In stru m en tal S u pp ort (0-3) N /A N/A N /A N /A 1.65 1.59
E m otional Support (0-3) N/A N/A N /A N/A 2.42 2.52
H ealth C are U tilization (Baseline)
% D octor Visit (>=1) 74.46 82.90 88.12 89.56 N/A N/A
Econom ic R esources (Baseline)
A nnual Incom e P er A dult E quivalent 25462 21555 17651 15027 N/A N/A
A nnual Incom e (self & spouse) - -- - - 19770 12590
% H ealth In surance 75.14 71.59 80.90 75.50 N/A N/A
H ealth Behaviors (Baseline)
% C u rren t Sm oking 38.09 47.56 12.17 8.19 12.35 1 3 .5 1
% H eavy D rinking
(>=3 d rin k s a day) 8.59 1 .8 1 4.55 0.79 N/A N/A
(>=2 liquor drin ks a t one tim e) N /A N/A N/A N/A 13.50 3.59
% O besity (BM I>=30) 20.53 23.14 11.05 1 5 .7 1 12.53 19.89
% Physical Activities 55.71 49.84 N/A N /A 86.46 79.60
o
71
Table 4.2. Number and Percent o f Marital Status by Sex in 1992 and 1994: HRS
M ales
1994
Frequency
Percent
1992
Never : Married : Separated/
married : Divorced
Widowed Total
Never
manned
Married
Separated/
1 : 3
..................................... 0.05
7 i 28
1
0.03
................... 32
0.62
recently
widowed
.......................... 9
0.16
134
3.95
..........3324’
83.58
........... 394’
10.84
Widowed I : 2 i 5
0.02: 0.04: 0.11
M M
64
1.64
Total 137 3276 405 98
4.01 82.63 11.06 2.30
3916
100.00
Females
1994
Frequency
Percent
1992
Never : Married : Separated/
married : : Divorced
Widowed Total
Never •^•.;.;i:5^i 2 ■ 9 6 176
Married
Separated/
Divorced
Widowed
0 p § ||§ i |§ i 43
0.00 0.93
P w S t
14 i 24
0.27 : 0.47
7 [ ............. 6 ’ i 23
0.12 i 0.17 : 0.39
79
1.49
recently
widowed
................... 53
1.05
3182
68.30
...........790’
17.07
...........486’
10.81
Total 180 3092 774 588
3.96 66.56 16.73 12.75
4634
100.00
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
72
Table 4.3. Number and Percent of Marital Status by Sex in 1995 and 1998: AHEAD
M ales
1998
Frequency
Percent
1995
Never • Married ■ Separated/
married : : Divorced
• Widowed Total
Never ' 0 ; 0 : 0 41
Married
Separated/
Divorced
Widowed
0.00 0.54
........................
0.05 : 0.05
2 • 13 : 8
0.11 : 0.70: 0.43
96
: 5.18
i recently
• widowed
i..................... io '
i 0.54
1369
73.92
..............81*
4.37
361
19.49
Total 44 1277 87 444
2.38 68.95 4.70 23.97
1852
100.00
Females
1998
Frequency
Percent
1995
Never Married | Separated/
married : ; Divorced
Widowed Total
Never married
Married
s' '• ?
. ‘ - 3; o5~ o
0.00
w im m ,
0 i 2
.00 : 0.06
10
0.32
108
3.43
jg g : 5
^ °-16
i» T T T ‘
182
5.78
recently
1066
33.86
Separated/
Divorced
Widowed
0.06 : 0.03
6 ; 5 : "30
0.19 : 0.16 i 0.95
1.33 5.78
..........1792
56.93
Total 104 885 174 1985
3.30 28.11 5.53 63.06
3148
100.00
1993 1995 1998 N
Continuously Married Married Married Married 2130
Recently Widowed Married Married Widowed 268
Longer-term Widowed Married Widowed Widowed 209
Longer-term Widowed Widowed Widowed Widowed 1878
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
73
Table 4.4. Number and Percent of Marital Status by Sex in 1988 and 1991: MacArthur
M ales_____
1991
Frequency
Percent
1988
Never
married
Married
Separated/
Divorced
Widowed
Total
Never
married
H . -iv '
Married
0.00
o 'W i
o.oo
0 i
0 . 0 0 :
0 :
0 . 0 0 i
~2 \
6.10
243
70.64
Total Widowed Separated/
Divorced
72.07
recently
widowed
n v > .
69
20.06
344
100.00
Females
1991
Frequency
Percent
1988
Never
married
Mamed Separated/
Divorced
Widowed Total
Never married 0
0.00
22
4.13
recently
widowed
ii
2.06
Mamed
29.27
Separated/
Divorced
Widowed
56.66
Total
533
100.00 25.52 61.54
1988 1991 N
Continuously Married Married Married 465
Recently Widowed Married Widowed 34
Longer-term Widowed Widowed Widowed 422
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
74
Table 4.5. Number and Percent of Marital Status by Sex in 1991 and 1995: MacArthur
M ales
1995
Frequency
Percent
1991
Never
married
Married Separated/
Divorced
Widowed Total
Never
married
Married
Separated/
Divorced
Widowed
m m
o
0.00
o i
0.00 :
0
0.00
" " ' 6'
o .o o
.9:9.9 .
I
0.32
0
0.00
22
6.94
recently
widowed
0
0.00
19
5.99
” 224’
70.66
9
2.84
—'65"
20.50
Total 19
5.99
202
63.72
11
3.47
85
26.81
317
100.00
Females
1995
Frequency
Percent
1991
Never
married
Mamed Separated/
Divorced
Widowed
Never
married
Married
26.34
recently
widowed
Separated/
Divorced
Widowed
60.70
Total
329
67.70
486
100.00
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
75
V. Marital Status and Health
Using the sample and the variables constructed, the previous chapter
described the characteristics of the samples. Now, this chapter turns into the
analyses of how marital status affects a variety of health outcomes. First, it shows
the number of people for each marital status in Tables 5.1 (HRS), 5.2 (AHEAD)
and 5.3 (MacArthur). The next section shows the difference of the prevalence by
marital status without controlling age in Tables 5.4 (HRS), 5.5 (AHEAD) and 5.6
(MacArthur), and the difference of the incidence in Tables 5.7 (HRS), 5.8
(AHEAD) and 5.9 (MacArthur). Then, it describes differences in baseline
mechanisms by marital status in Tables 5.10 (HRS), 5.11 (AHEAD) and 5.12
(MacArthur). The following sections discuss the results from the logistic
regressions of the prevalence and incidence of health conditions on marital status
and age, and on marital status and all other variables, for males and females and for
the HRS, AHEAD and MacArthur, in order (Tables 5.13 to 5.28 and Figures 5.1 to
5.25). Then, this chapter is concluded with two figures (Figures 5.26 and 5.27)
discussing how marital status differentially affects different health outcomes and
what mechanisms are related to the relationship of marital status to different health
outcomes.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
76
Number of People in Each Marital Status
Tables 5.1, 5.2, and 5.3 show the number of respondents by marital status
and sex among those whose marital status had not changed in HRS, AHEAD, and
MacArthur, respectively. Because of the small sample size of some categories of
marital status particularly in MacArthur, some analyses in this study (e.g. analyses
with different unmarried groups, recent widowhood) were not technically possible.
The majority of missing cases in HRS were those living with partners (Table 5.1).
Table 5.1. Number of People by Marital Status and Sex among Those Whose
Marital Status Had Not Changed between 1992 and 1994: HRS (N=8,148)
Males Females
Number Number
Married 3245 3060
Never Married 129 159
Separated/Divorced 350 699
Widowed 56 450
Missing 95 49
Total 3780 4368
Table 5.2. Number of People by Marital Status and Sex among Those Whose
Marital Status Had Not Changed between 1993 and 1998: AHEAD (N=4,348)
Males Females
Number Number
Married 1256 874
Never Married 41 96
Separated/Divorced 66 137
Widowed 269 1609
Missing 1 0
Total 1632 2716
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
77
Table 5.3. Number of People by Marital Status and Sex among Those Whose
Marital Status Had Not Changed between 1988 and 1995: MacArthur (N=686)
Males Females
Number Number
Married 198 88
Never Married 19 31
Separated/Divorced 9 26
Widowed 54 261
Missing 1 0
Total 280 406
Differences in the Prevalence of Health Conditions without Controlling Age
Tables 5.4, 5.5 and 5.6 show the prevalence of diseases, conditions and
functioning problems by marital status and sex in HRS (1992), AHEAD (1993),
and MacArthur (1988). In HRS (Table 5.4), overall, married respondents in 1992
were less likely to have functioning problems, diseases and conditions than
unmarried counterparts, and the separated/divorced and the widowed were
significantly worse in several health outcomes than the married. A higher level of
functioning problems and diseases among unmarried groups compared to the
married for both sexes was particularly noticeable in the prevalence of ADL
difficulties and depression. Differential prevalence of health conditions by marital
status was particularly strong for separated/divorced and widowed females
compared to their counterpart males.
Table 5.5 shows the percent with functioning problems, diseases and
conditions at baseline (1993) among the AHEAD respondents. Fewer significant
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Reproduced w ith permission o f th e copyright owner. Further reproduction prohibited without permission.
Table 5.4. Percent of Those who Had Functioning Difficulties, Diseases and Conditions in 1992 by Marital Status and
Sex: HRS
Married_________ Never Married______Separated/Divorced________ Widowed
Males Females Males Females Males Females Males Females
ADL Difficulties 7.30 7.58 13.95* 8.18 15.14** 16.31** 14.29* 18.44**
Nagi Difficulties 39.04 60.82 49.61* 65.41 47.71** 63.95 44.64 65.33+
Hypertension 39.54 35.10 41.86 44.03* 39.71 48.07** 51.79+ 49.56**
Diabetes 10.66 9.41 13.18 9.43 10.86 13.45** 19.64 15.78**
Cancer 3.17 7.06 2.33 7.55 2.57 9.59* 3.57 7.56
Lung Conditions 6.93 7.16 7.75 4.40 10.86* 11.44** 14.29 11.78**
Heart Conditions 14.98 9.58 13.95 10.69 17.14 13.59** 17.86 14.44**
Stroke 2.84 2.09 4.65 2.52 4.29 3.72* 5.36 3.56
Arthritis 29.95 41.63 34.88 47.80 34.00 50.21** 41.07 50.89**
Broken Bone 13.14 13.43 10.08 12.58 15.43 17.31* 5.36* 16.44
Depression 6.04 8.24 13.95* 14.47* 20.86** 18.88** 23.21** 21.11**
**p<.01, *p<.05, +p<.10 (com pared to th e m arried )
o o
Reproduced w ith permission o f th e copyright owner. Further reproduction prohibited without permission.
Table 5.5. Percent of Those who Had Functioning Difficulties, Diseases and Conditions in 1993 by Marital Status and
Sex: AHEAD
Married__________ Never Married______Separated/Divorced________Widowed
Males Females Males Females Males Females Males Females
ADL Difficulties 9.81 13.61 12.20 14.58 13.64 22.06* 18.96** 20.44**
IADL Difficulties 9.01 11.52 10.26 14.58 9.09 10.45 10.82 19.31**
Nagi Difficulties 14.65 35.13 26.83+ 38.54 16.67 35.77 24.16** 39.78*
Hypertension 43.55 49.54 63.41* 51.04 47.69 43.07 42.75 54.05*
Diabetes 12.18 10.76 7.32 7.29 15.15 13.97 12.69 12.00
Cancer 12.98 11.44 9.76 12.50 9.09 10.22 16.36 12.38
Lung Conditions 9.88 7.67 14.63 6.25 15.15 16.06* 15.13* 8.58
Heart Conditions 33.04 22.43 41.46 21.88 34.85 24.09 30.86 28.42**
Stroke 5.65 4.69 2.44 5.21 7.58 6.57 8.18 7.09*
Arthritis 17.99 26.20 21.95 28.13 39.39** 32.85 25.00* 30.97*
Broken Hip 3.18 3.32 2.44 5.21 1.52 8.03+ 4.09 5.22*
Depression 9.87 17.28 21.95+ 17.71 30.30** 29.20** 28.62** 28.59**
**pc.01, *p<.05, +p<.10 (com pared to the m arried )
'-j
V O
Reproduced w ith permission o f th e copyright owner. Further reproduction prohibited without permission.
Table 5.6. Percent of Those who Had Functioning Difficulties, Diseases and Conditions in 1988 by Marital Status and
Sex: MacArthur
Married___________Never Married______Separated/Divorced_______Widowed
Males Females Males Females Males Females Males Females
Nagi Difficulties 10.10 28.41 10.53 22.58 11.11 15.38 14.81 15.77**
Hypertension 44.44 44.32 42.11 41.94 55.56 34.62 46.30 49.62
Diabetes 12.63 20.45 5.56 6.45+ 22.22 7.69 5.56 8.88**
Cancer 15.15 14.77 0.00+ 22.58 44.44* 26.92 12.96 19.69
Heart Conditions 14.65 9.09 11.11 0.00+ 11.11 7.69 22.22 5.38
Stroke 1.52 3.41 0.00 0.00 0.00 0.00 5.56+ 0.77+
Broken Hip 1 .01 3.41 0.00 3.23 11.11* 3.85 1.85 1.54
Depression 6.11 6.02 5.88 13.33 0.00 13.04 12.00 6.94
**p<.01, *p<.05, +p<.10 (com pared to the m arried)
o o
o
81
differences between the married and each unmarried group were observed than in
HRS. Widows were noticeably more likely than married women to have
functioning problems, diseases and conditions with the exception of diabetes,
cancer and lung conditions. On the other hand, there were few significant
differences found among marital groups in the prevalence of health conditions
among the MacArthur respondents (Table 5.6). Results suggest a higher
prevalence of depression among the never married and separated/divorced females,
but not widows, compared to married females. A higher percentage of widowers,
but not other unmarried groups, had depression compared to married males even
though these are not statistically significant.
Differences in the Incidence of Health Conditions without Controlling Age
Tables 5.7, 5.8 and 5.9 indicate the incidence or the percent of acquiring
functioning problems, new diseases and conditions since the baseline. The
incidence analysis was based on those who did not have the specified conditions at
baseline (1992 in HRS, 1993 in AHEAD and 1988 in MacArthur), and examined
those who got functioning problems, new diseases and conditions or died between
1992 and 1994 (HRS), between 1993 and 1998 (1995 for death) (AHEAD) and
between 1988 and 1995 (MacArthur). The chi-square of the cross-tab of incidence
analyses did not indicate consistent differences between the married and the
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Reproduced with permission o f th e copyright owner. Further reproduction prohibited without permission.
Table 5.7. Percent of Those who Got Functioning Difficulties, New Diseases and Conditions or Died between 1992 and
1994 by Marital Status and Sex: HRS _____________
Married Never Married Separated/Divorced Widowed
Males Females Males Females Males Females Males Females
ADL Difficulties 2.83 2.38 0.65 3.93 5.55** 4.21* 15.62** 5.26**
Nagi Difficulties 17.54 22.94 30.77** 21.82 16.94 20.24 16.13 23.72
Hypertension 6.47 7.44 9.81 6.71 10.64 8.11 22.69 9.43
Diabetes 2.72 1.67 2.96 3.06 2.14 3.95 4.84 2.75
Cancer 1.53 1.48 2.86 1.58 2.85 2.59 2.30 1.10
Lung Conditions 2.30 1.93 2.59 3.34 4.78 3.03 5.43 3.94
Heart Conditions 3.55 2.68 1.49 3.54 5.13 3.52 4.66 3.71
Stroke 0.74 0.53 0.77 0.94 2.65 1.26 1.13 0.90
Arthritis 9.79 14.15 10.36 16.60 10.45 15.64 1.03 7.18
Broken Bone 4.21 5.52 3.42 5.31 5.14 7.06 18.01 15.41
Depression 8.29 13.50 13.63* 15.53 22.59** 21.07** 22.57** 22.02**
Death 1.57 0.67 1.35 0.27 1.70 1.49* 4.91* 2.04**
'p<.01, *p<.05, +p<.10 (com pared to the m arried )
o o
to
Reproduced w ith permission o f th e copyright owner. Further reproduction prohibited without permission.
Table 5.8. Percent of Those who Got Functioning Difficulties, New Diseases and Conditions between 1993 and 1998
or Died between 1993 and 1995 by Marital Status and Sex: AHEAD ______________
Married__________ Never Married_____ Separated/Divorced_______Widowed
Males Females Males Females Males Females Males Females
ADL Difficulties 23.84 24.74 29.17 38.57* 30.77 23.26 29.81 33.08**
IADL Difficulties 17.43 16.32 20.00 18.75 8.33+ 20.34 17.75 21.71**
Nagi Difficulties 20.90 31.92 36.67+ 33.90 20.00 26.14 30.39** 39.32**
Hypertension 15.83 21.23 33.33 27.66 5.88 14.10 19.46 20.27
Diabetes 4.82 4.63 5.41 7.87 3.57 5.98 6.01 4.98
Cancer 9.09 4.92 8.11 5.95 6.67 8.20 10.81 3.62
Lung Conditions 4.63 4.10 5.71 2.22 5.36 6.09 5.33 4.24
Heart Conditions 17.84 11.37 18.18 9.72 0.00** 14.42 20.11 12.50
Stroke 6.34 5.62 5.13 8.89 1.67 3.17 7.00 8.28*
Arthritis 48.46 54.73 58.06 58.21 29.73* 56.04 45.92 60.63*
Broken Hip 0.99 1.66 2.50 4.40+ 0.00 1.60 1.55 3.80**
Depression 14.49 17.84 15.63 17.72 23.91+ 20.62 20.83* 18.19
Death 12.34 8.25 24.24** 11.26 14.63 7.73 17.75** 12.17**
p<.01, *p<.05, +p<.10 (com pared to the m arried )
o o
U )
Reproduced w ith permission o f th e copyright owner. Further reproduction prohibited without permission.
Table 5.9. Percent of Those who Got Functioning Difficulties, New Diseases and Conditions between 1988 and 1995
by Marital Status and Sex; MacArthur
Married_________ Never Married_____Separated/Divorced_______ Widowed
Males Females Males Females Males Females Males Females
Nagi Difficulties 25.84 47.62 35.29 37.50 50.00 54.55 52.17** 52.97
Hypertension 18.18 10.45 23.53 8.33 16.67 9.52 21.88 18.00
Diabetes 14.65 14.77 15.79 3.23+ 0.00 7.69 18.87 13.85
Cancer 19.19 7.95 5.26 16.13 11.11 7.69 22.22 12.36
Heart Conditions 10.61 7.95 0.00 12.90 0.00 11.54 9.26 6.54
Stroke 7.58 7.95 5.26 6.45 22.22 3.85 5.66 7.31
Broken Hip 2.02 2.27 0.00 3.23 0.00 3.85 5.56 1.54
Depression 47.24 62.86 33.33 58.33 75.00 63.16 52.63 62.79
p<.01, *p<.05, +p<. 1 0 (com pared w ith the m arried )
o o
4 * .
85
unmarried groups. However, in general, lower percentage of married people
acquired new conditions since the baseline relative to unmarried people.
Separated/divorced and widowed persons for both sexes and never married
men in HRS had a much higher incidence of depression between 1992 and 1994
compared to the married: 8.29% among married men versus about 23% among
separated/divorced and widowed men, and 13.63% among never married men;
13.5% among married women versus 21-22% among separated/divorced and
widowed women (Table 5.7). The never married AHEAD male respondents were
about twice as likely to die compared to married counterparts between 1993 and
1995 (12.34% versus 24.24%) (Table 5.8). As in the case of the prevalence of
health conditions shown in the previous table (Table 5.6), there was not much
difference in the incidence of health conditions between the married and each
unmarried group in MacArthur (Table 5.9). Tables 5.4 to 5.9 show that while the
separated/divorced and widowed generally had a higher percent with health
conditions, the never married often more closely resembled the married.
Differences in Mechanisms by Marital Status
Tables 5.10, 5.11 and 5.12 show the characteristics of samples at baseline
by marital status and sex in HRS, AHEAD and MacArthur, respectively. Tables
5.10 and 5.11 show that unmarried females lived with more people than unmarried
males did. Among the HRS respondents (Table 5.10), never married,
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Reproduced with permission o f th e copyright owner. Further reproduction prohibited without permission.
Table 5.10. Mean and Percent of Characteristics of Sample at Baseline (1992) by Marital Status and Sex: HRS______
Married________ Never Married Separated/Divorced______ Widowed
Social Support
Males Females Males Females Males Females Males Females
Mean # Household Members 0.96 0.78 0.53** 0.97+ 0.55** 1.10** 0.75 1.31**
% Satisfied w/ Neighborhood 88.20 87.88 80.95* 84.31 79.71** 76.51** 80.36+ 84.82+
% Satisfied w/ Friendships
Economic Resources
93.18 93.39 86.51** 90.20 87.83** 90.52** 80.36** 92.19
Mean Per Capita Income 25862 24651 27035 18700** 25871 15473** 17703** 14264**
% Health Insurance
Health Behaviors
78.72 76.76 54.09** 66.03** 58.27** 58.29** 51.47** 55.34**
% Current Smoking 34.33 43.66 44.95* 41.34 56.64** 52.63** 48.78* 59.10**
% Obesity 20.86 20.81 19.42 28.03* 18.63 25.69** 25.65 32.15**
% Heavy Drinking 7.56 1.93 7.95 0.49 13.93** 1.53 11.19 1.77
% Physical Activities
Health Care Utilization
56.89 51.78 64.04+ 50.49 50.62* 51.42 40.88** 47.41+
% Doctor Visits 74.61 82.97 74.42 79.87 75.43 82.55 71.43 83.78
**pc.01, *p<.05, +p<.10 (com pared to the m arried )
0 0
O s
Reproduced with permission o f th e copyright owner. Further reproduction prohibited without permission.
Table 5.11. Mean and Percent of Characteristics of Sample at Baseline (1993) by Marital Status and Sex: AHEAD
Married________ Never Married Separated/Divorced______Widowed
Social Support
Males Females Males Females Males Females Males Females
Mean # Household Members 0.24 0.18 0.44 0.66** 0.24 0.36* 0.48** 0.47**
Mean # Living Children 2.94 2.72 0.80* 0.55** 3.47+ 1.82** 2.71 2.54*
Mean # Living Siblings
Economic Resources
2.14 2.19 3.07* 1.81+ 2.32 1.91+ 2.23 2.06
Mean Per Capita Income 16791 15803 17372 14232 21895 13013* 20676 14750
% Health Insurance
Health Behaviors
85.51 86.04 56.10** 67.71** 46.97** 59.12** 78.07** 72.41**
% Current Smoking 11.23 5.95 7.32 9.38 24.24** 13.87** 15.99* 8.70*
% Obesity 10.91 16.70 9.76 11.46 12.12 16.79 8.92 15.91
% Heavy Drinking
Health Care Utilization
4.31 1.14 0.00 0.00 3.03 2.92+ 7.06+ 0.44*
% Doctor Visits 89.78 90.54 82.05 83.33* 83.08+ 88.06 88.06 89.93
**p< 01, *p<.05, +p<.10 (com pared to the m arried )
o o
Reproduced with permission o f th e copyright owner. Further reproduction prohibited without permission.
Table 5.12. Mean and Percent of Characteristics of Sample at Baseline (1988) by Marital Status and Sex: MacArthur
Married_________Never Married_____Separated/Divorced______ Widowed
Males Females Males Females Males Females Males Females
Social Support
Mean # Living Children 2.43 2.32 0.00** 0.29** 3.22 1.69+ 3.19+ 2.45
% Children Nearby 75.82 65.88 — 25.00+ 86.67 80.77 84.62 77.13*
Mean # Relatives Nearby 3.29 3.18 2.83 3.94 2.70 2.42 3.71 2.81
Mean # Friends Nearby 7.17 4.03 5.27 5.83 5.25 4.26 6.44 5.15+
Instrumental Support 1.75 1.68 1.05* 1.66 1.61 1.33* 1.50* 1.59
(0-3)
Emotional Support 2.44 2.50 2.45 2.71* 2.36 2.47 2.36 2.51
m .
Economic Resources
Mean Income
Health Behaviors
22309 19257 20970 14887+ 10731** 11258** 14373** 10195**
% Current Smoking 9.52 11.36 11.76 12.12 33.33** 28.57* 18.87+ 12.99
% Obesity 11.82 23.08 5.26 11.76 21.43 27.03 12.73 20.23
% Heavy Drinking 15.85 5.00 11.76 7.14 12.50 4.35 10.42 2.98
% Leisure Activities 87.68 86.81 94.74 91.18 93.33 75.68 80.36 80.38
**p<.01, *p<.05, +p<,10 (com pared to the m arried )
o o
0 0
89
separated/divorced and widowed females had significantly higher per capita
income at baseline than married females while among males, the difference with
married males was only found among widowers. Having supplemental health
insurance significantly differed for the married and each unmarried group (p.<01)
in HRS and AHEAD. Unmarried males and separated/divorced females showed
lower satisfaction with neighbors and friendships than their married counterparts in
HRS (Table 5.10), and a higher percent of the unmarried, particularly the
separated/divorced and the widowed, currently smoked than married counterparts
in HRS, AHEAD and MacArthur (Tables 5.10, 5.11 and 5.12). For health
behaviors, the separated/divorced and widowed showed more risky health
behaviors than the married while the never married showed less difference from the
married. Fewer differences were noticed between the married and three unmarried
groups in the characteristics presented for the MacArthur sample (Table 5.12).
Results from Logistic Regressions
Prevalence o f Health Conditions on Marital Status and Age
Odds ratios from logistic regressions o f having functioning difficulties,
diseases and conditions on marital status and age are presented in Tables 5.13, 5.14
and 5.15. In model 1 of Table 5.13, being never married among males predicted a
greater likelihood of having ADL and Nagi difficulties and depression compared to
being married, and for females, being never married increased the probability of
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
90
having arthritis and depression. Those effects remained in model 2 in which age
was controlled. Separated/divorced and widowed males were about 4 to 5 times
more likely to have depression than married males while separated/divorced and
widowed females were about 3 times more likely to have depression than married
females. Among females, being separated/divorced or widowed was associated
with greater odds of having all health conditions examined except Nagi difficulties,
cancer and stroke. Unmarried groups of males were more likely to have health
conditions compared to the married, but being unmarried was related to the
prevalence of fewer conditions among men than among women.
Among the AHEAD respondents (Table 5.14), being widowed was related
to a higher probability of having diabetes, stroke and depression among females in
the model with age controlled (model 2). Among males, being unmarried, whether
being never married, separated/divorced or widowed, increased the probability of
having depression by 2 to 4 times relative to being married. On the other hand,
among females, being never married was not related to the prevalence of having
depression, but separated/divorced and widowed females were about 2 times more
likely to have depression than married females. Males had a stronger adverse effect
o f being unmarried than females on depression as well as on other conditions. Also,
as in HRS (Table 5.13), the separated/divorced and the widowed, compared to the
never married, were more likely to be different from the married in the prevalence
o f health conditions. While the HRS and AHEAD results did not show any
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
91
beneficial effect of being unmarried on the prevalence of health conditions, the
results from the MacArthur showed that widows were less likely to have Nagi
difficulties and diabetes than married females (Table 5.15). Analysis was
technically impossible for the male MacArthur respondents for cancer, stroke,
broken hip and depression, and, for females, for heart conditions and stroke
because none of the never married or the separated/divorced had stroke, none of the
never married males had cancer or broken hip, none of separated/divorced males
had depression, and none of never married females had heart conditions. This
necessitated the analysis with gender included in model instead of separate models
for each sex in MacArthur.
Table 5.13. Odds Ratios from Logistic Regression of Having Functioning Difficulties,
Diseases and Conditions (1992) on Marital States and Age: HRS_________________
M odel 1 Model 2
M arital Status (M arried as Reference) M ales Females M ales Females
ADL Difficulties Never M arried 1.78* —
1.81*
—
Separated/Divorced 1.82*** 2.25*** 1.86*** 2.26***
W idowed —
2.36***
—
2.30***
Age 1.06**
—
Nagi Difficulties Never M arried 1.71** —
1.73*** —
Separated/Divorced 1.25* — 1.27*
—
W idowed — —
1.77*
—
Age 1.04*** 1.03**
Hypertension Never M arried — — — —
Separated/Divorced — 1.54*** — 1.57***
W idowed —
1.61***
—
1.53***
Age 1.03* 1.05***
Diabetes Never M arried — — — —
Separated/Divorced —
1.51**
—
1.52**
Widowed 1.98* 1.80***
—
1.71***
Age 1.08*** 1.05**
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
92
Table 5.13 continued
Model 1 M odel 2
Males Females Males Females
Cancer Never M arried
— — — —
Separated/Divorced — — — —
Widowed
— — — —
Age 1.11*** 1.06**
Lung Conditions Never M arried — — — —
Separated/Divorced 1.49* 1.76*** 1.52* 1.78***
Widowed 2.79** 1.88*** 2.55** 1.79***
Age 1.08** 1.05**
H eart Conditions Never M arried — — — —
Separated/Divorced — 1.39* — 1.42***
Widowed — 1.36* — —
Age 1.07*** 1.09***
Stroke Never M arried — — — —
Separated/Divorced — — — —
Widowed — — — —
Age 1.10** —
A rthritis Never M arried — 1.51* — 1.53**
Separated/Divorced — 1.33*** — 1.35***
Widowed — 1.36** — 1.29*
Age 1.06*** 1.06***
Broken Bone Never M arried
— — — —
Separated/Divorced — 1.35** — 1.37*
Widowed
— 1.36* — —
Age 1.05*** 1.07***
Depression Never M arried 2.19** 1.70* 2.17** 1.70*
Separated/Divorced 4.07*** 2.52*** 4.05*** 2.51***
Widowed 5.17*** 3.00*** 5.35*** 3.04***
Age — —
***p<0.00l, **p<0.01, *p<0.05
— not significant
Table 5.14. Odds Ratios from Logistic Regression of Having Functioning Difficulties,
Diseases and Conditions (1993) on Marital States and Age: AHEAD_____________
Model 1 M odel 2
M arital Status (M arried as Reference) Males Females Males Females
ADL Difficulties Never M arried
— — — —
Separated/Divorced — 1.80* — 1.70*
Widowed
Age
2.15*** 1.63*** 1.87***
1.06*** 1.07***
IADL Difficulties Never M arried
— — — —
Separated/Divorced — — — —
Widowed
Age
— 1.84***
1.13*** 1.11***
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
93
Table 5.14 continued___________________________________________________
M odel 1 M odel 2
M ales Females M ales Females
Nagi Difficulties Never M arried 2.14*
—
2.07*
—
Separated/Divorced — — — —
Widowed 1.86*** 1.22* 1.69**
—
Age 1.05*** 1.03***
Hypertension Never M arried 2.25* — 2.29* -
Separated/Divorced — — — —
Widowed
—
1.20*
— —
Age 0.98* —
Diabetes Never M arried — — — —
Separated/Divorced — — — —
Widowed
— — —
1.33*
Age — 0.96***
Cancer Never M arried — — — —
Separated/Divorced — — — —
Widowed — — — —
Age — -
Lung Conditions Never M arried — — — —
Separated/Divorced — 2.30**
— 2.37**
Widowed 1.70** — 1.77**
—
Age — 0.97*
H eart Conditions Never M arried — — — —
Separated/Divorced — — — —
Widowed — 1.37**
— —
Age — 1.03***
Stroke Never M arried — — — —
Separated/Divorced — — — —
Widowed —
1.55*
— 1.04**
Age — —
A rthritis Never M arried — — — —
Separated/Divorced 2.96*** — 2.99*** —
Widowed 1.52** 1.26* 1.47*
—
Age — —
Broken H ip Never M arried — — — —
Separated/Divorced —
2.54*
—
2.29*
Widowed
—
1.61*
— —
Age 1.09***
Depression Never M arried 2.57* — 2.51* —
Separated/Divorced 3.97*** 1.97*** 4.09** 1.96**
Widowed 3.66*** 1.92*** 3.35*** 1.84***
Age 1.04** —
***p< 001; **fX.0I; *p<.05
— not significant
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
94
Table 5.15. Odds Ratios from Logistic Regression of Having Functioning Difficulties,
Diseases and Conditions (1988) on Marital States and Age: MacArthur____________
Model 1 M odel 2
M arital Status (M arried as Reference) M ales Females Males Fem ales
Nagi Difficulties
Hypertension
Diabetes
Cancer
H eart Conditions
Broken H ip
Depression
Never M arried
Separated/Divorced
W idowed
Age
Never M arried
Separated/Divorced
W idowed
Age
Never M arried
Separated/Divorced
W idowed
Age
Never M arried
Separated/Divorced
Widowed
Age
Never M arried
Separated/Divorced
Widowed
Age
Never M arried
Separated/Divorced
Widowed
Age
Never M arried
Separated/Divorced
W idowed
Age
0.47** 0.46**
0.82*
0.38* 0.36**
N/A N/A
N/A N/A
N/A
N/A
N/A
N/A
***p<0.001, **p<0.01, *p<0.05
— not significant
Analysis o f stroke was not technically possible for both sexes because o f the small sample size and the skewed
sample distribution.
Incidence o f Health Conditions on Marital Status and Age
Tables 5.16, 5.17 and 5.18 present the odds ratios of the probability of
acquiring functioning problems and new diseases and conditions on marital states
and age between 1992 and 1994 in HRS, between 1993 and 1995 in AHEAD, and
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
95
between 1988 and 1995 in MacArthur, respectively. Table 5.16 shows that in HRS,
being unmarried, particularly being separated/divorced and being widowed, were
strongly related to acquiring ADL difficulties and depression for both sexes,
hypertension for males, and death for females. Widows were about 3 times more
likely to die than married women. Relative to married males, widowers were 6
times more likely to experience the onset of ADL problems, and being
separated/divorced and widowed for males increased the likelihood of getting
depression about 3 times. Never married males were more likely to acquire Nagi
difficulties and depression than married males.
Fewer effects of being unmarried on the incidence of health conditions were
found in the AHEAD and MacArthur samples. Among the AHEAD respondents
(Table 5.17), being widowed increased the probability of getting a stroke for
females while among males, being separated/divorced decreased the probability of
acquiring arthritis, being widowed increased the probability of getting Nagi
difficulties, and being never married that of death when age was controlled (model
2 in Table 5.17). Being widowed was related to a higher probability of getting
Nagi difficulties among males in MacArthur (Table 5.18). In general, when age
was controlled in all data, the effects of marital status still remained, implying that
marital status affects health outcomes regardless of the age of respondents.1
1 The range of age within each dataset may not be large enough to differentiate outcomes by age.
But age differences among the respondents in three datasets seem to differentiate outcomes by age,
which is discussed later.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
96
Table 5.16. Odds Ratios from Logistic Regression of Getting Functioning Difficulties,
New Diseases and Conditions between 1992 and 1994 on Marital States and Age: HRS
Model 1 Model 2
M arital Status (M arried as Reference) M ales Females Males Females
ADL Difficulties Never M arried
— — — —
Separated/Divorced 2.02** 1.80* 2.05** 1.80*
Widowed 6.36*** 2.28** 6.04*** 2.29**
Age - —
Nagi Difficulties Never M arried 2.11** — 2.17** —
Separated/Divorced — — — —
Widowed
— — — —
Age 1.05** —
Death Never M arried — — — —
Separated/Divorced — 2.23* — 2.27*
Widowed 3.23* 3.07** —
2.79**
Age 1.10* 1.11*
Hypertension Never M arried — — — —
Separated/Divorced 1.68* — 1.67* —
Widowed 4.83*** — 5.21*** —
Age — —
Diabetes Never M arried — — — —
Separated/Divorced — 1.78* —
1.81*
Widowed — — — —
Age — 1.11**
Cancer Never M arried — — — —
Separated/Divorced 2.09* — 2.15* —
Widowed — — — —
Age 1.09* —
Lung Conditions Never M arried — — — —
Separated/Divorced 1.96* ~ 1.97* —
Widowed — 2.28** —
2.18**
Age — —
H eart Conditions Never M arried — — — —
Separated/Divorced — 1.82** — 1.84**
Widowed —
1.98**
—
1.91*
Age 1.08* —
Stroke Never M arried — — — —
Separated/Divorced 3.04** 2.93* 3.21** 2.97*
Widowed — — — —
Age 1.17** —
A rthritis Never M arried — — — —
Separated/Divorced — — — —
Widowed — — — —
Age 1.07*** —
Broken Bone Never M arried
— — — —
Separated/Divorced — — — —
Widowed — — — —
Age — —
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
97
Table 5.16 continued
Model 1 M odel 2
Males Females Males Females
Depression Never M arried 1.86* — 1.84* —
Separated/Divorced 3.46*** 1.78*** 3.44*** 1.77***
Widowed 3.16** 1.88*** 3.27** 1.92***
Age — —
***p<0.001, * * p < O .O t, *p<0.05
— n ot significant
Table 5 .1 7 . Odds Ratios from Logistic Regression of Getting Functioning Difficulties,
New Diseases and Conditions between 1993 and 1998 on Marital States and Age: AHEAD
M odel 1 M odel 2
M arital Status (M arried as Reference) M ales Females Males Fem ales
ADL Difficulties Never M arried — 1.91* — —
Separated/Divorced — — — —
Widowed — 1.50*** — —
Age 1.04** 1.10***
L A D L Difficulties Never M arried — — — —
Separated/Divorced — — — —
Widowed
—
1.42**
— —
Age 1.10*** 1.11***
Nagi Difficulties Never M arried 2.19*
— — —
Separated/Divorced — — — —
W idowed 1.65** 1.38** 1.46* —
Age 1.05*** 1.07***
D ea th ® Never M arried 2.27** — 2.25** —
Separated/Divorced — — — —
Widowed 1.53** 1.54*** — —
Age 1.11*** 1.08***
Hypertension Never M arried — - - -
Separated/Divorced - -- — —
Widowed — — — —
Age — —
Diabetes Never M arried — — — —
Separated/Divorced — — — —
W idowed — — — —
Age — 0.93***
Cancer Never M arried — — — —
Separated/Divorced — — — —
Widowed — — — —
Age - —
Lung Conditions Never M arried — — — —
Separated/Divorced — — — —
Widowed — — — —
Age — —
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
98
Table 5.17 continued
Model 1 M odel 2
M ales Females M ales Females
H eart Conditions Never M arried
— —
Separated/Divorced N/A — N/A —
Widowed — —
Age 1.04**
Stroke Never M arried — — — —
Separated/Divorced — — — —
Widowed — 1.52* — 1.49*
Age - -
Arthritis Never M arried — — — —
Separated/Divorced 0.45* — 0.45* —
Widowed
— 1.27* — —
Age - 1.02*
Broken H ip Never M arried — —
Separated/Divorced N/A 2.35** N/A —
Widowed
— —
Age 1.09***
Depression Never M arried - — - -
Separated/Divorced - — — -
Widowed 1.55*
— — —
Age 1.03* —
***p<0.001, **p<0.01, *p<0.05
— not significant
“Death occurred between 1993 and 1995
Table 5.18. Odds Ratios from Logistic Regression of Getting Functioning Difficulties,
New Diseases and Conditions between 1988 and 1995 on Marital States and Age:
MacArthur
Model 1 M odel 2
M arital Status (M arried as Reference) M ales Females Males Females
Nagi Difficulties Never M arried — — — —
Separated/Divorced — — — —
Widowed 3.13***
— 2.85** —
Age 1.12* —
Hypertension Never M arried - — - - -
Separated/Divorced — — - - —
Widowed — — — —
Age -
Diabetes Never M arried — —
Separated/Divorced N/A — N/A —
Widowed
— —
Age —
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
99
Table 5.18 continued
Cancer
Stroke
Broken Hip
Depression
M odel 1 Model 2
M ales
H eart Conditions
Never M arried
Separated/Divorced
Widowed
A ge
Never M arried
Separated/Divorced
Widowed
A ge
Never M arried
Separated/Divorced
Widowed
A ge
Never M arried
Separated/Divorced
Widowed
A ge
Never M arried
Separated/Divorced
Widowed
Age
N/A
N/A
Females M ales
0.87*
N/A
N/A
Females
1.19***
***p<0.00l, **p<0.01, *p<0.05
— not significant
Prevalence o f Health Conditions on Marital Status and Other Variables
So far, logistic regression analysis was done only with age controlled.
However, the adverse effect of being unmarried on health shown may be related to
respondents’ baseline health, race, education as well as other covariates proposed
as mechanisms by which marriage affects health. By including this baseline
information in equations, this section attempts to control a possible marriage
selection. Self-rated health is used as an indicator o f initial health status given that
it is not a direct measure of health outcomes examined but closely correlated. An
indicator of health care utilization, doctor visits, was omitted in logistic regression
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
100
analysis as it will be in later analysis because of the causation issue. In fact, since
the indicators of health conditions were asked “has your doctor ever told you,”
doctor visits might simply reflect health conditions.
Odds ratios from logistic regression of having functioning difficulties and
diseases and conditions on marital states and other variables are presented in HRS
(Tables 5.19 and 5.20), AHEAD (Tables 5.21 and 5.22), and MacArthur (Table
5.23). Table 5.19 shows that being unmarried among the HRS male respondents
was related to the prevalence o f health conditions such as Nagi difficulties, lung
conditions, heart conditions, broken bone and depression. In many cases, being
unmarried, whether being never married, separated/divorced or widowed, was
related to adverse health outcomes with selected mechanisms controlled in the
model: Being unmarried, relative to being married, increased the probability of
having certain diseases and conditions. However, being never married and being
widowed were not beneficial in some types of health conditions. For instance, the
never married were less likely than the married to have heart conditions, and the
widowed were also less likely than the married to have broken bone. When all the
proposed mechanisms were considered, the effects of marital status observed in
previous models on the prevalence of some health outcomes (e.g. being never
married on Nagi difficulties, being widowed on broken bone, being
separated/divorced on depression) remained significant for HRS male respondents,
but the effects of others disappeared (model 5 in Table 5.19).
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
101
The association of unmarried status with the prevalence of depression was
higher than with physical health conditions among the HRS male respondents.
Even though the effect on depression disappeared by controlling all mechanisms
among the widowed, the effect of being separated/divorced on the prevalence of
depression remained significant: The relative likelihood of separated/divorced men
having depression was 2.88 times greater than that of the married men while the
positive effect of being widowed on having depression disappeared when
considering all mechanisms together (model 5 in Table 5.19).
Table 5.20 shows the results of the prevalence of health conditions for
women in HRS. For women, being unmarried was associated with a higher
prevalence of ADL difficulties, hypertension, lung conditions, arthritis, broken hip
and depression, but it was associated with a lower prevalence of Nagi difficulties.
Being widowed was positively related to having ADL difficulties, but the effect
disappeared when economic resources (model 3) or all mechanisms (model 5) were
taken into account, indicating that the harmful effect of being widowed on ADL
difficulties was partly explained by income, or by current smoking, drinking status
and activity participation when all other proposed mechanisms were considered
together. Being widowed or separated/divorced also increased the probability of
having lung conditions, but when current smoking status was controlled in models
(models 4 and 5), the effect disappeared, showing that the relationship of marital
status to the prevalence of lung conditions was related to smoking behaviors.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
102
Income and obesity were related to the adverse effect of being never married on
having arthritis (model 5). As in males’ case in Table 5.19, being unmarried
showed a strong positive relationship with the prevalence of depression. In model
5 in which all the proposed mechanisms were included, the effects of being
widowed on depression were still significant while the effect of being never
married or separated/divorced disappeared. Current smoking status affected the
relationship between being never married and having depression (models 4 and 5),
and higher satisfaction with friendships and neighborhood explained the adverse
effect being separated/widowed on the prevalence of depression only when they
were considered with other mechanisms proposed (model 2 vs. model 5). Current
smoking partly explained a higher probability of having depression among the
separated/divorced and the widowed, and higher satisfaction with friendships or
neighborhood partly explained the adverse effect of being widowed on depression
(model 5).
How about older people? They showed similar results. Tables 5.21 and
5.22 provide the results of the same analysis for the AHEAD respondents. As in
HRS, while being unmarried was, in many cases, positively related to having health
conditions, some negative effects were also noticed: Separated/divorced women
were less likely to have hypertension than married women (Table 5.22). As in
HRS, there was a strong positive relationship between being separated/divorced
and widowed and having depression among the AHEAD respondents for both
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
103
sexes. However, being never married was related to the prevalence of depression
among males, but not among females. All of the effects of marital status on health
remained with all mechanisms considered among the AHEAD male respondents
(model 5 in Tables 5.21)— the positive effect of being widowed on ADL and Nagi
difficulties, lung conditions and depression, that of being never married on
hypertension, and that of being separated/divorced on arthritis and depression. On
the other hand, among females (Table 5.22), some effects of marital status on the
prevalence of health conditions disappeared with selected mechanisms controlled.
In terms of the confounding factors in the relationship between being
unmarried and health conditions, some proposed mechanisms influenced the
relationships. For example, a lower probability of having hypertension among
separated/divorced females compared to married females in AHEAD was, in fact,
through current smoking status and obesity; a higher probability of having
depression among separated/divorced females was related to the number of
household members, health insurance and smoking; a higher likelihood of having
depression among separated/divorced females was partly explained by more
household members, no coverage of supplemental health insurance, or current
smoking (model 5 in Table 5.22).
Among the MacArthur respondents (Table 5.23), much fewer effects of
being unmarried on the prevalence of health conditions were noticed. The
prevalence of three health outcomes such as Nagi difficulties, diabetes and cancer
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
104
was related to being unmarried, but in different ways. The widowed were less
likely than the married to have Nagi difficulties with social support controlled in
model (model 3), and diabetes except when all mechanisms controlled (model 5).
However, being separated/divorced increased the likelihood o f having cancer even
when income or health behaviors were controlled. The only mechanism involved
was the number of living children in the relation of being widowed on the
prevalence of diabetes (model 5 in Table 5.23).
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Reproduced with permission o f th e copyright owner. Further reproduction prohibited without permission.
T able 5.19. O dds R atios from Logistic R egression of H aving Functioning Difficulties, Diseases and C onditions (1992) on M a rita l
States an d O th er V ariables am ong M ales: H R S____________ ____________ ____________ ____________ ______
M odel 1 M odel 2 M odel 3 M odel 4 M odel 5
ADL D ifficulties Never M arried
— — — — —
Separated/D ivorced
— — — - - - -
W idow ed
— — - - — —
Age — — - - — —
Baseline SR H 0.32*** 0.32*** 0.33*** 0.32*** 0.35***
W hite — — — —
E ducation — — — — —
H ousehold M em bers
- - —
Satisfaction w ith
N eighborhood
— —
Satisfaction w ith
F riendships
0.43*** 0.34***
Incom e 1.00* 1.00**
H ealth Insu ran ce — —
C u rren t Sm oking
— —
H eavy D rinking
— —
O besity — —
P hysical A ctivity 0.64** 0.67**
N agi D ifficulties N ever M arried 1.56* 1.50* 1.54* 2.23*** 2.01**
Separated/D ivorced — — — - - —
W idow ed
- - — — — —
Age
— — — — —
Baseline SR H 0.49*** 0,49*** 0.50*** 0.51*** 0.52***
W hite 1.28*
— —
1.31*
—
E ducation — — — — —
H ousehold M em bers — 0.90*
Satisfaction w ith
N eighborhood
— —
o
Reproduced w ith permission o f th e copyright owner. Further reproduction prohibited without permission.
T able 5.19 contin uer
M odel 1 M odel 2 M odel 3 M odel 4 M odel 5
N agi Difficulties Satisfaction w ith
F riendships
- - —
Incom e 1.00*** 1.00***
H ealth In surance
— —
C u rren t Sm oking — —
H eavy D rinking 1.34*
—
O besity 1.49*** 1.51***
Physical A ctivity 0.82* 0.84*
H ypertension Never M arried
— — - - - - ~
Separated/D ivorced — — — —
W idow ed - - — — — —
Age 1.02* 1.03*
— - - —
Baseline SR H 0.62*** 0.63*** 0.62*** 0.63*** 0.62***
W hite — — — — —
E ducation 1.05*** 1.05*** 1.05*** 1.06*** 1.06***
H ousehold M em bers — —
Satisfaction w ith
N eighborhood
— —
Satisfaction w ith
F riendships
— - -
Incom e — —
H ealth In su ran ce —
1.30*
C u rren t Sm oking 0.62*** 0.62***
H eavy D rinking 1.54** 1.50**
O besity 1.62** 1.63***
Physical A ctivity — —
D iabetes Never M arried — — — — —
Separated/D ivorced — — — ~ —
W idow ed — — — — —
o
O N
Reproduced with permission o f th e copyright owner. Further reproduction prohibited without permission.
T able 5.19 continuec
M odel 1 M odel 2 M odel 3 M odel 4 M odel 5
D iabetes Age 1.06*** 1.07*** 1.06*** 1.07** 1.07**
Baseline S R H 0.47*** 0.49*** 0.47*** 0.49*** 0.51***
W hite 0.69* 0.71* 0.69* 0.68* 0.69*
E ducation 1.07*** 1.07*** 1.08*** 1.11*** 1.11*
H ousehold M em bers — —
Satisfaction w ith
N eighborhood
—
Satisfaction w ith
F riendships
" —
Incom e
- -
H ealth In su ran ce — —
C u rren t Sm oking — —
H eavy D rinking 0.52* 0.56*
O besity 2.60*** 2.70***
Physical A ctivity
— —
C ancer Never M arried — — - - — —
Separated/D ivorced — — — —
W idow ed — — — — —
Age 1.10** 1.10** 1.10** 1.08* 1.08*
Baseline SR H 0.67*** 0.70*** 0.69*** 0.65*** 0.70***
W hite — — — — —
E ducation — — — —
1.09*
H ousehold M em bers — —
Satisfaction w ith
N eighborhood
- - —
Satisfaction w ith
F riendships
— —
Incom e
— —
H ealth Insu ran ce - -
107
Reproduced with permission o f th e copyright owner. Further reproduction prohibited without permission.
T able 5.19 continu er
M odel 1 M odel 2 M odel 3 M odel 4 M odel 5
C ancer C u rren t Sm oking — - -
H eavy D rinking —
O besity
— —
Physical Activity — - -
L ung C onditions Never M arried — ~ — — —
Separated/D ivorced
- - — — — —
W idow ed 2.15*
. .
2.09*
. . —
Age — — —
1.06* 1.06*
Baseline SR H 0.51*** 0.50*** 0.52*** 0.50*** 0.51***
W hite 3.47*** 3.21*** 3.60*** 3.82*** 3.76***
E ducation 0.95** 0.95*
- - —
0.94**
H ousehold M em bers — - -
Satisfaction w ith
N eighborhood
— --
Satisfaction w ith
F riendships
0.53** 0.55**
Incom e — —
H ealth In surance
— —
C u rren t Sm oking 1.59** 1.51**
H eavy D rinking — —
O besity 0.67*
—
Physical Activity
- -
H eart C onditions Never M arried — — —
0.54*
—
Separated/D ivorced — — — — —
W idow ed — — — — —
Age 1.05*** 1.05** 1.05*** 1.05** 1.04*
Baseline SR H 0.44*** 0.44*** 0.43*** 0.43*** 0.42***
W hite 1.86*** 1.80*** 1.85*** 1.71** 1.73**
E ducation 1.06*** 1.06*** 1.06** 1.06** 1.06**
o
oo
Reproduced with permission o f th e copyright owner. Further reproduction prohibited without permission.
T able 5.19 con tinu er
M odel 1 M odel 2 M odel 3 M odel 4 M odel 5
H eart C onditions H ousehold M em bers — - -
Satisfaction w ith
N eighborhood
— —
Satisfaction w ith
F riendships
— —
Incom e — —
H ealth In surance — —
C u rren t Sm oking 0.78*
—
H eavy D rinking • • —
O besity
- - —
Physical Activity 1.27*
—
S trok e N ever M arried — — . . — - -
Separated/D ivorced — — . . — - -
W idow ed
- - — — —
Age 1.08* 1.09* 1.08* 1.09*
—
Baseline SR H 0.42*** 0.45*** 0.46*** 0.37*** 0.41***
W hite
— — — 0.49* —
E ducation 1.09*
—
1.13*** 1.09* 1.09*
H ousehold M em bers 1.25* 1.22*
Satisfaction w ith
N eighborhood
- - —
Satisfaction w ith
F riendships
— —
Incom e 1.00*
—
H ealth Insurance 0.63*
—
C u rren t Sm oking — —
H eavy D rinking
— —
O besity
—
0.49*
Physical Activity — —
Reproduced w ith permission o f th e copyright owner. Further reproduction prohibited without permission.
T able 5.19 continu er
M odel 1 M odel 2 M odel 3 M odel 4 M odel 5
A rth ritis Never M arried — — — - - —
Separated/D ivorced — — — — . . .
W idow ed — — — — —
Age 1.05*** 1.05*** 1.05*** 1.03*
—
Baseline SR H 0.65*** 0.64*** 0.64*** 0.65*** 0.65***
W hite 1.53*** 1.54*** 1.52*** 1.61*** 1.61***
E ducation 0.96** 0.96** 0.97** 0.96** 0.97*
H ousehold M em bers — —
Satisfaction w ith
N eighborhood
-- —
Satisfaction w ith
F riendships
— - -
Incom e
- - —
H ealth In su ran ce 1.21*
. .
C u rren t Sm oking 0.79** 0.81**
H eavy D rinking
— —
O besity 1.29** 1.31**
Physical A ctivity — —
B roken B one Never M arried — — — — . .
Separated/D ivorced — — — —
W idow ed 0.31* 0.28* 0.29*
—
0.30*
Age 1.05** 1.05** 1.05** 1.05** 1.05**
Baseline SR H 0.86*** 0.85*** 0.87** 0.86** 0.88**
W hite — 1.39* ~ — —
E ducation — — — — —
H ousehold M em bers
—
Satisfaction w ith
N eighborhood
0.70* 0.64**
no
Reproduced with permission o f th e copyright owner. Further reproduction prohibited without permission.
T able 5.19 con tin u er
M odel 1 M odel 2 M odel 3 M odel 4 M odel 5
B roken Bone Satisfaction w ith
F riendships
— --
Incom e — - -
H ealth In surance —
0.76*
C u rren t Sm oking
— —
H eavy D rinking — . .
O besity
— —
Physical Activity 1.37** 1.38**
D epression Never M arried
— — - - —
Separated/D ivorced 3.35*** 3.37*** 2.89*** 3.29*** 2.88***
W idow ed 3.74*** 3.07** 3.33*** 3.07**
—
Age 0.95** 0.94** 0.94**
— - -
Baseline SR H 0.40*** 0.40*** 0.42*** 0.42*** 0.43***
W hite
— — — — - -
E ducation — — —
0.95*
- -
H ousehold M em bers — —
Satisfaction w ith
N eighborhood
-- -
Satisfaction w ith
F riendships
0.29*** 0.25***
Incom e — —
H ealth In su ran ce 0.53*** 0.58**
C u rren t Sm oking 1.44* 1.46*
H eavy D rinking — —
O besity — —
Physical Activity — —
***p<.001; **p<.01; *p<.05
-- not significant
Reproduced with permission o f th e copyright owner. Further reproduction prohibited without permission.
T able 5.20. O dds R atios from Logistic R egression of H aving Functioning Difficulties, Diseases an d C onditions (1992) on M arital
States and O th er V ariables am ong F em ales: H R S__________ ____________________________________________
M odel 1 M odel 2 M odel 3 M odel 4 M odel 5
A DL D ifficulties Never M arried — ~ — — —
Separated/D ivorced
— — — — —
W idow ed 1.57** 1.53*
- -
1.65*
—
Age — - - — — —
Baseline SR H 0.29*** 0.30*** 0.31*** 0.29*** 0.30***
W hite — — — — —
E ducation — „ - - — —
H ousehold M em bers — —
Satisfaction w ith
N eighborhood
--
—
Satisfaction w ith
F riendships
- - —
Incom e 1.00*
—
H ealth In su ran ce
— —
C u rren t Sm oking 0.71* 0.70*
H eavy D rinking —
0.23*
O besity — —
Physical Activity 0.69* 0.68*
N agi Difficulties Never M arried — — — — —
Separated/D ivorced — — — — --
W idow ed 0.77* 0.75* 0.73** 0.73* 0.67*
Age — — — — —
Baseline SR H 0.47*** 0.47*** 0.47*** 0.47*** 0.46***
W hite 1.34** 1.36** 1.39** 1.48** 1.61**
E ducation ~ — — — - -
H ousehold M em bers — —
Satisfaction w ith
Neighborhood
- - —
112
Reproduced with permission o f th e copyright owner. Further reproduction prohibited without permission.
T able 5.20 con tinu er
M odel 1 M odel 2 M odel 3 M odel 4 M odel 5
N agi D ifficulties Satisfaction w ith
Friendships
— -
Incom e 1.00*
—
H ealth Insu ran ce
—
0.80*
C u rren t Sm oking 0.80* 0.80*
H eavy D rinking
— —
O besity 1.97*** 1.93***
Physical Activity 0.63*** 0.63***
H ypertension Never M arried — — — — —
Separated/D ivorced — — —
1.40** 1.33*
W idow ed 1.25* — — — . . .
Age 1.05*** 1.05*** 1.05*** 1.04* 1.04**
Baseline SR H 0.57*** 0.57*** 0.57*** 0.61*** 0.60***
W hite 0.57*** 0.58*** 0.57*** 0.62*** 0.63***
E ducation 1.03* 1.03* 1.04*
— —
H ousehold M em bers — —
Satisfaction w ith
N eighborhood
— —
Satisfaction w ith
F riendships
— —
Incom e — —
H ealth In surance — - -
C u rren t Sm oking 0.73** 0.72**
H eavy D rinking 2.17** 2.17**
O besity 2.29*** 2.23***
Physical Activity
— —
D iabetes N ever M arried — — — — —
Separated/D ivorced — — — — —
W idow ed — - - — —
Reproduced with permission o f th e copyright owner. Further reproduction prohibited without permission.
T able 5.20 continuec
M odel 1 M odel 2 M odel 3 M odel 4 M odel 5
D iabetes Age 1.04* 1.05* 1.04* — - -
Baseline S R H 0.42*** 0.48*** 0.50*** 0.53*** 0.52***
W hite 0.65** 0.73* 0.65**
— —
E ducation — — — — —
H ousehold M em bers 1.13** —
Satisfaction w ith
N eighborhood
— —
Satisfaction w ith
F riendships
— - -
Incom e
—
1.00*
H ealth Insu ran ce
—
1.59**
C u rren t Sm oking 0.55*** 0.52***
H eavy D rinking —
O besity 2.59*** 2.60***
Physical Activity — —
C ancer Never M arried — — — — —
Separated/D ivorced — — — — —
W idow ed — — — — —
Age 1.05** 1.04* 1.05**
— —
Baseline SR H 0.67*** 0.67*** 0.66*** 0.64*** 0.64***
W hite 1.77** 1.73** 1.73**
- - —
E ducation — — — — —
H ousehold M em bers
— —
Satisfaction w ith
N eighborhood
— —
Satisfaction w ith
F riendships
«
- -
Incom e — - -
H ealth In su rance — —
Reproduced w ith permission o f th e copyright owner. Further reproduction prohibited without permission.
T able 5.20 continuec
M odel 1 M odel 2 M odel 3 M odel 4 M odel 5
C ancer C u rren t Sm oking ~ —
H eavy D rinking
- - —
O besity
— —
Physical A ctivity
— —
L ung C onditions Never M a rried
— — — — —
Separated/D ivorced 1.37*
- -
1.37*
— - -
W idow ed 1.50* 1.46* 1.50*
— —
Age — — — — —
Baseline SR H 0.47*** 0.47*** 0.46*** 0.48*** 0.48***
W hite 2.24*** 2.35*** 2.21*** 2.21*** 2,39***
E ducation —
0.96* 0.95*
—
H ousehold M em bers — —
Satisfaction w ith
N eighborhood
— - -
Satisfaction w ith
F riendships
- - - -
Incom e — —
H ealth In su rance — —
C u rren t Sm oking 1.45* 1.46*
H eavy D rinking
— —
O besity
— - -
Physical Activity —
H ea rt C onditions Never M arried — — — — —
Separated/D ivorced — - - — — —
W idow ed — — — — —
Age 1.07*** 1.08*** 1.07*** 1.07** 1.08***
Baseline SR H 0.44*** 0.44*** 0.43*** 0.46*** 0.45***
W hite 1.49** 1.53** 1.48*
— —
E ducation — — - —
Reproduced w ith permission o f th e copyright owner. Further reproduction prohibited without permission.
T able 5.20 con tin uec
M odel 1 M odel 2 M odel 3 M odel 4 M odel 5
H ea rt C onditions H ousehold M em bers — —
Satisfaction w ith
N eighborhood
0.67** 0.62**
Satisfaction w ith
F riendships
— - -
Incom e
— —
H ealth In surance
— —
C u rren t Sm oking 0.69** 0.69*
H eavy D rinking 0.25*
—
O besity — —
Physical Activity
— —
S troke Never M arried — — — —
Separated/D ivorced
— — - - — —
W idow ed — — — — —
Age
— — — — —
Baseline SR H 0.43*** 0.43*** 0.44*** 0.49*** 0.51***
W hite
— — — — —
E ducation
— — - - — —
H ousehold M em bers — —
Satisfaction w ith
N eighborhood
- - —
Satisfaction w ith
F riendships
- - - -
Incom e
— —
H ealth Insurance
— —
C u rren t Sm oking — - -
H eavy D rinking
— - -
O besity — - -
Physical Activity - —
Reproduced with permission o f th e copyright owner. Further reproduction prohibited without permission.
T able 5.20 continuec
M odel 1 M odel 2 M odel 3 M odel 4 M odel 5
A rth ritis Never M a rried 1.58** 1.58** 1.56** 1.67*
—
Separated/D ivorced — — — — —
W idow ed — - - — — —
Age 1.05*** 1.05*** 1.05*** 1.04* 1.04**
Baseline S R H 0.56*** 0.56*** 0.57*** 0.55*** 0.55***
W hite 1.39*** 1.4*0** 1.42*** 1.58*** 1.67***
E ducation — — — — —
H ousehold M em bers — —
Satisfaction w ith
N eighborhood
- - —
Satisfaction w ith
F riendships
— - -
Incom e
—
1.00**
H ealth Insu ran ce — —
C u rren t Sm oking — —
H eavy D rinking — —
O besity 1.56*** 1.53***
Physical A ctivity
— —
B roken Bone N ever M arried — — — — —
Separated/D ivorced 1.35** 1.32* 1.30*
— . . .
W idow ed — — — —
Age 1.07*** 1.07*** 1.07*** 1.08*** 1.08***
Baseline SR H 0.83*** 0.82*** 0.84*** 0.81*** 0.82***
W hite 1.63*** 1.59** 1.63***
- -
1.44*
E ducation — — — —
H ousehold M em bers — —
Satisfaction w ith
N eighborhood
-- —
117
Reproduced w ith permission o f th e copyright owner. Further reproduction prohibited without permission.
T able 5.20 continuec
M odel 1 M odel 2 M odel 3 M odel 4 M odel 5
B roken B one Satisfaction w ith
F riendships
— —
Incom e — —
H ealth In su ran ce — - -
C u rren t Sm oking 1.27*
—
H eavy D rinking
— —
O besity
— —
Physical A ctivity
— —
D epression Never M arried 1.88* 2.00* 1.75*
— - -
Separated/D ivorced 1.88*** 1.83*** 1.71*** 1.47*
- -
W idow ed
2 4i***
2.52*** 2.19*** 2.51*** 2.57***
Age 0.96** 0.95** 0.96*
—
0.95*
Baseline SR H 0.43*** 0.45*** 0.45*** 0.45*** 0.48***
W hite 1.39* 1.41* 1.42* 1.62* 1.72**
E ducation 0.89*** 0.88*** 0.91*** 0.88*** 0.89***
H ousehold M em bers — —
Satisfaction w ith
N eighborhood
0.66** 0.66*
Satisfaction w ith
F riendships
0.41*** 0.37***
Incom e — —
H ealth In su ran ce
— —
C u rren t Sm oking 1.40* 1.38*
H eavy D rinking — —
O besity
— —
Physical A ctivity -
***p<.001; **p<.01; *p<.05
- not significant
o o
Reproduced with permission o f th e copyright owner. Further reproduction prohibited without permission.
T able 5.21. O dds R atios from Logistic R egression of H aving Functioning Difficulties, D iseases an d C onditions (1993) on M arital
M odel 1 M odel 2 M odel 3 M odel 4 M odel 5
A DL D ifficulties N ever M arried — - - — —
Separated/D ivorced — - - — — —
W idow ed 1.80** 1.78** 1.85** 1.83** 1.85**
Age 1.05** 1.04* 1.05** 1.05** 1.05**
Baseline S R H 0.37*** 0.37*** 0.37*** 0.37*** 0.36***
W hite — „ — — —
E ducation — — — — —
H ousehold M em bers 1.25** 1.25**
Living C h ild ren
— —
Living Siblings
— - -
Incom e — - -
H ealth Insu ran ce — —
C u rren t Sm oking
— —
H eavy D rinking
— —
O besity 1.96** 2.02**
IA D L Difficulties Never M arried - - — — — —
Separated/D ivorced — — — — - -
W idow ed
— — — - - —
Age 1.13*** 1.13*** 1.13*** 1.14*** 1.14***
Baseline SR H 0.40*** 0.40*** 0.40*** 0.40*** 0.40***
W hite — — - - - - —
E ducation 0.91*** 0.91*** 0.93** 0.91*** 0.93*
H ousehold M em bers 1.20*
—
Living C h ildren — —
Living Siblings
— —
Incom e
— . .
H ealth Insu ran ce — —
C u rren t Sm oking — —
Reproduced with permission o f th e copyright owner. Further reproduction prohibited without permission.
T able 5.21 continued
M odel 1 M odel 2 M odel 3 M odel 4 M odel 5
IA D L Difficulties H eavy D rinking — —
O besity
— —
N agi D ifficulties Never M a rried — — - - - - —
Separated/D ivorced - - - - — —
W idow ed 1.57* 1.56* 1.59* 1.57* 1.57*
Age 1.05*** 1.05** 1.05*** 1.05*** 1.05**
Baseline SR H 0.39*** 0.39*** 0.39*** 0.39*** 0.39***
W hite 0.65* 0.66* 0.63*
- -
0.65*
E ducation — — — . . - -
H ousehold M em bers — —
Living C h ild ren
- - —
Living Siblings
— —
Incom e — —
H ealth In su ran ce
— —
C u rren t Sm oking — —
H eavy D rinking — —
O besity
— —
H ypertension Never M a rried 2.12* 2.23* 2.13* 2.05* 2.13*
Separated/D ivorced
— — - - — —
W idow ed
— — — . . —
Age 0.97* 0.97** 0.97** 0.98* 0.97**
Baseline S R H 0.77*** 0.77*** 0.77*** 0.76*** 0.76***
W hite 0.67** 0.66** 0.67** 0.65** 0.65**
E ducation
— « — - - —
H ousehold M em bers
— —
Living C hildren
— —
Living Siblings — —
Incom e — —
H ealth Insu ran ce -
Reproduced with permission o f th e copyright owner. Further reproduction prohibited without permission.
T able 5.21 continued
M odel 1 M odel 2 M odel 3 M odel 4 M odel 5
H ypertension C u rren t Sm oking 0.70* 0.69*
H eavy D rinking
— —
O besity 1.53* 1.55**
Diabetes N ever M arried — — — — —
Separated/D ivorced
— — - - — —
W idow ed — — — — —
Age — — - - — —
Baseline SR H 0.61*** 0.61*** 0.61*** 0.61*** 0.61***
W hite — —
0.63*
—
0.64*
E ducation — — — — —
H ousehold M em bers — —
Living C hildren
— —
Living Siblings
- - —
Incom e
- - - -
H ealth In surance — —
C u rren t Sm oking — —
H eavy D rinking — —
O besity 1.74* 1.71*
C ancer N ever M arried — — — — —
Separated/D ivorced — — — — —
W idow ed — — — — —
Age 1.03* 1.04* 1.04* 1.04* 1.05**
Baseline SR H 0.70*** 0.70*** 0.70*** 0.72*** 0.71**
W hite
— - - — — —
E ducation 1.13*** 1.14*** 1.11*** 1.14*** 1.13***
H ousehold M em bers
— —
L iving C hildren
— —
L iving Siblings
— - -
Incom e — —
Reproduced with permission o f th e copyright owner. Further reproduction prohibited without permission.
T able 5.21 continued
M odel 1 M odel 2 M odel 3 M odel 4 M odel 5
C ancer H ealth In surance 2.14** 2.20**
C u rren t Sm oking
— —
H eavy D rinking — —
O besity 1.90** 1.92**
L ung C onditions N ever M arried
— — . . —
Separated/D ivorced
— — — - - - -
W idow ed 1.79** 1.75** 1.78** 1.71** 1.65*
Age
— - - — — —
Baseline SR H 0.51*** 0.51*** 0.50*** 0.51*** 0.50***
W hite 2.03** 2.06** 1.74* 2.18** 1.91*
E ducation — — — — —
H ousehold M em bers — —
Living C hildren — —
Living Siblings — —
Incom e
_ _ - -
H ealth In surance 1.69* 1.76*
C u rren t Sm oking — —
H eavy D rinking 2.20* 2.23*
O besity
— - -
H eart C onditions N ever M arried
— — — — —
Separated/D ivorced
— — — — —
W idow ed — — — — —
Age
- - . . . — —
Baseline SR H 0.53*** 0.53*** 0.53*** 0.53*** 0.52***
W hite 2.78*** 2.78*** 2.65*** 2.78*** 2.69***
E ducation
— — — — —
H ousehold M em bers
— —
L iving C hildren — —
Living Siblings — —
Reproduced with permission o f th e copyright owner. Further reproduction prohibited without permission.
T able 5.21 continued
M odel 1 M odel 2 M odel 3 M odel 4 M odel 5
H ea rt C onditions Incom e
— —
H ealth In su ran ce
- - —
C u rren t Sm oking 0.55** 0.56**
H eavy D rinking
— —
O besity
- - —
S trok e N ever M arried — — - - — —
Separated/D ivorced — — — — —
W idow ed — — - - —
Age — - - — — - -
Baseline SR H 0.46*** 0.47*** 0.46*** 0.47*** 0.46***
W hite
— « — — —
E ducation — — — - -
H ousehold M em bers
— —
Living C h ildren
— - -
Living Siblings — - -
Incom e — - -
H ealth In su rance
—
C u rren t Sm oking — —
H eavy D rinking
— —
O besity
— —
A rth ritis Never M arried — — — — —
Separated/D ivorced 2.49** 2.57*** 2.50** 2.70*** 2.84***
W idow ed
— — — — —
Age — - - — — . .
Baseline S R H 0.62*** 0.62*** 0.61*** 0.61*** 0.61***
W hite 0.53*** 0.55*** 0.50*** 0.51*** 0.51***
E ducation
— — — — —
H ousehold M em bers
— . .
Living C hildren — —
Reproduced w ith permission o f th e copyright owner. Further reproduction prohibited without permission.
T able 5.21 continued
M odel 1 M odel 2 M odel 3 M odel 4 M odel 5
A rth ritis Living Siblings
. . . —
Incom e 1.00* 1.00*
H ealth In su ran ce — —
C u rren t Sm oking 0.60* 0.59*
H eavy D rinking — —
O besity 1.94*** 2.00***
B roken H ip Never M a rried — — - - — —
Separated/D ivorced — — - - — —
W idow ed
— - - — — —
Age
— — — — —
Baseline S R H
—
0.76*
—
W hite
— — — — —
E ducation
— — - - . . —
H ousehold M em bers
— —
Living C h ildren
—
Living Siblings
— . .
Incom e — —
H ealth In surance
- - —
C u rren t Sm oking 2.18* 2.13*
H eavy D rinking
— —
O besity — —
D epression Never M arried
— — —
2.39* 2.33*
Separated/D ivorced 3.66*** 3.60*** 3.64*** 3.58*** 3.49***
W idow ed 3.33*** 3.36*** 3.37*** 3.23*** 3.28***
Age
—
1.03* —
1.04* 1.04*
Baseline SR H 0.51*** 0.51*** 0.51*** 0.51*** 0.51***
W hite — — ~ — —
E ducation
— — — — —
H ousehold M em bers - -
Reproduced with permission o f th e copyright owner. Further reproduction prohibited without permission.
T able 5.21 continued
M odel 1 M odel 2 M odel 3 M odel 4 M odel 5
D epression Living C h ildren
— —
Living Siblings — —
Incom e
— —
H ealth In su ran ce — —
C u rren t Sm oking
„ —
H eavy D rinking — —
O besity - -
***p<.001, **p<.01, *p<.05
-- not significant
T able 5.22. O dds R atios from L ogistic R egression of H aving Functioning Difficulties, D iseases an d C onditions (1993) on M arital
M odel 1 M odel 2 M odel 3 M odel 4 M odel 5
ADL D ifficulties Never M arried — — - - — —
Separated/D ivorced
— —
1.65*
W idow ed - - — — — —
Age 1.06*** 1.06*** 1.06*** 1.06*** 1.07***
Baseline SR H 0.44*** 0.44*** 0.44*** 0.45*** 0.44***
W hite - - — — —
E ducation — — - - —
H ousehold M em bers
— - -
Living C h ildren — —
Living Siblings
— - -
Incom e — —
H ealth Insu ran ce
— —
C u rren t Sm oking — —
H eavy D rinking - - —
Reproduced w ith permission o f th e copyright owner. Further reproduction prohibited without permission.
T able 5.22 continued
M odel 1 M odel 2 M odel 3 M odel 4 M odel 5
A D L D ifficulties O besity 2.10*** 2.08***
IA D L Difficulties Never M a rried
— — . . .
Separated/D ivorced — — — - - —
W idow ed — — — — —
Age 1.11*** 1.11*** 1.10*** l.n*** 1.11***
Baseline S R H 0.49*** 0.48*** 0.49*** 0.49*** 0.49***
W hite
— — — — - -
E ducation 0.90*** 0.91*** 0.91** 0.90*** 0.92***
H ousehold M em bers 1.24*** 1.23***
Living C hild ren 1.07* 1.06*
Living Siblings
— —
Incom e — —
H ealth In su ran ce 0.62** 0.66**
C u rren t Sm oking
- - —
H eavy D rinking — —
O besity — —
N agi D ifficulties Never M a rried — — — —
Separated/D ivorced — — — — —
W idow ed
— — - - — —
Age 1.02* 1.02** 1.02* 1.02* 1.02**
Baseline SR H 0.52*** 0.52*** 0.52*** 0.52*** 0.60***
W hite 1.56** 1.51*** 1.53*** 1.46** 1.56***
E ducation 0.97*
— —
0.97*
—
H ousehold M em bers — —
Living C h ildren 1.07** 1.07**
Living Siblings
— —
Incom e
— —
H ealth In surance — —
C u rren t Sm oking — —
S)
O n
Reproduced with permission o f th e copyright owner. Further reproduction prohibited without permission.
T able 5.22 continued
M odel 1 M odel 2 M odel 3 M odel 4 M odel 5
N agi Difficulties H eavy D rinking
—
O besity — —
H ypertension N ever M a rried — — — — —
Separated/D ivorced 0.65*
—
0.65* 0.67*
—
W idow ed — — — —
Age — — — — - -
Baseline SR H 0.72*** 0.72*** 0.72*** 0.72*** 0.72***
W hite 0.66*** 0.66*** 0.67*** 0.70** 0.71**
E ducation — — - - —
H ousehold M em bers
— —
Living C h ild ren — —
Living Siblings
- - —
Incom e
— —
H ealth In su ran ce
— —
C u rren t Sm oking 0.54*** 0.53***
H eavy D rinking — —
O besity 1.64*** 1.62***
D iabetes Never M arried — - - — — —
Separated/D ivorced
— . . — — —
W idow ed — « — — —
Age 0.94*** 0.94*** 0.94*** 0.94*** 0.94***
Baseline SR H 0.53*** 0.54*** 0.54*** 0.54*** 0.54***
W hite 0.58*** 0.62** 0.61** 0.64** 0.70*
E ducation
— — —
0.96*
- -
H ousehold M em bers
— —
L iving C hild ren
— —
Living Siblings
— —
Incom e
— —
H ealth In surance —
-
N)
- j
Reproduced with permission o f th e copyright owner. Further reproduction prohibited without permission.
T able 5.22 continued
M odel 1 M odel 2 M odel 3 M odel 4 M odel 5
D iabetes C u rren t Sm oking 0.44** 0.44**
H eavy D rinking — —
O besity 1.68*** 1.66***
C ancer Never M arried — — - - —
Separated/D ivorced — — — — —
W idow ed — — — — —
A ge — — — — - -
Baseline SR H 0.87** 0.86** 0.87** 0.86** 0.86**
W hite 1.87** 1.85** 1.82** 1.84** 1.79**
E d ucation 1.06** 1.06** 1.06** 1.06** 1.06**
H ousehold M em bers — —
Living C h ild ren — —
Living Siblings ~ —
Incom e — - -
H ealth In suran ce
. . —
C u rren t Sm oking — —
H eavy D rinking
— —
O besity — —
L ung C onditions N ever M arried — — — — —
Separated/D ivorced 2.62*** 2.73*** 2.63*** 2.51** 2.55**
W idow ed — — — — —
Age 0.95*** 0.95*** 0.95*** 0.96** 0.96**
Baseline SR H 0.60*** 0.59*** 0.60*** 0.60*** 0.59***
W hite 3.80*** 3.88*** 3.98*** 3.84*** 4.01***
E ducation — — — — —
H ousehold M em bers
— —
Living C hildren
— —
Living Siblings
—
Incom e - -
Reproduced with permission o f th e copyright owner. Further reproduction prohibited without permission.
T able 5.22 continued
M odel 1 M odel 2 M odel 3 M odel 4 M odel 5
L ung C onditions H ealth Insurance
— - -
C u rren t Sm oking — —
H eavy D rinking 3.69* 3.95*
O besity
— —
H eart C onditions Never M arried — — — — —
Separated/D ivorced — — — - - —
W idow ed
— — . . —
1.28*
Age 1.02* 1.02* 1.02*
- -
1.02*
Baseline SR H 0.49*** 0.49*** 0.49*** 0.49*** 0.48***
W hite 1.74*** 1.70*** 1.60*** 1.76*** 1.58**
E ducation
— — — — —
H ousehold M em bers
— —
Living C hildren —
Living Siblings
— —
Incom e
- - —
H ealth In surance 1.31* 1.31*
C u rren t Sm oking 0.58** 0.59**
H eavy D rinking — —
O besity
— —
Stroke Never M arried
— — — — —
Separated/D ivorced — — — — —
W idow ed
— — — — —
Age
— — —
Baseline SR H 0.52*** 0.53*** 0.53*** 0.52*** 0.52***
W hite 1.59* 1.68* 1.63*
— —
E ducation
— — — — —
H ousehold M em bers — —
Living C hildren
— - -
Living Siblings —
Reproduced with permission o f th e copyright owner. Further reproduction prohibited without permission.
T able 5.22 continued
M odel I M odel 2 M odel 3 M odel 4 M odel 5
S trok e Incom e
—
H ealth In su ran ce — —
C u rren t Sm oking — —
H eavy D rinking — —
O besity 0.56* 0.56*
A rth ritis Never M a rried — — - - - -
Separated/D ivorced — — — — —
W idow ed
— — — - - —
Age
- - — — — —
Baseline SR H 0.61*** 0.60*** 0.61*** 0.61*** 0.61***
W hite 0.51*** 0.49*** 0.53*** 0.54*** 0.54***
E ducation
— - - - - — —
H ousehold M em bers
- -
L iving C h ildren 1.07** 1.06**
L iving Siblings — - -
Incom e — —
H ealth In su ran ce
— —
C u rren t Sm oking — - -
H eavy D rinking — —
O besity 1.71*** 1.69***
B roken H ip N ever M arried
— — — — —
Separated/D ivorced 2.47* 2.59* 2.74** 2.51* 2.92**
W idow ed
— - - — — —
Age 1.09*** 1.09*** 1.09*** 1.09*** 1.09***
Baseline S R H 0.76*** 0.76*** 0.74*** 0.75*** 0.74**
W hite 2.00* 2.04*
—
1.98*
- -
E ducation
— — — —
H ousehold M em bers — —
Living C h ild ren — -
Reproduced with permission o f th e copyright owner. Further reproduction prohibited without permission.
T able 5.22 continued
M odel 1 M odel 2 M odel 3 M odel 4 M odel 5
B roken H ip L iving Siblings
— —
Incom e 1.00* 1.00*
H ealth In suran ce — —
C u rren t Sm oking
— —
H eavy D rinking
- - —
O besity — —
D epression Never M arried — — — — —
Separated/D ivorced 1.80** 1.77* 1.66* 1.77*
—
W idow ed 1.74*** 1.81*** 1.70*** 1.69*** 1.73***
Age — — — — - -
Baseline SR H 0.51*** 0.50*** 0.51*** 0.51*** 0.51***
W hite — — 1.36*
— —
E ducation 0.96* 0.96**
—
0.96*
«
H ousehold M em bers 0.84** 0.82***
L iving C hildren — —
Living Siblings
— —
Incom e —
H ealth In su ran ce 0.62*** 0.60***
C u rren t Sm oking 1.60** 1.63**
H eavy D rinking
- - —
O besity — -
***p<.001,**p<.01,*p<.05
-- not significant
Reproduced with permission o f th e copyright owner. Further reproduction prohibited without permission.
T able 5.23. O dds R atios from Logistic R egression of H aving F unctioning Difficulties, D iseases and C onditions (1988) on M a rita l
States an d O ther V ariables: M acA rth u r
M odel 1 M odel 2 M odel 3 M odel 4 M odel 5 M odel 6
N agi D ifficulties Never M a rried — — — — - - - -
Separated/D ivorced — - - — —
W idow ed — - -
0.43**
— — —
Fem ale 2.27** 2.30** 2.73** 2.22** 2.06*
—
Baseline S R H — —
0.65*
„ — —
W hite — — — — — —
E ducation
— — — — — —
Age — - - ~ — —
Living C h ildren — - -
C hildren N earby
- - —
R elatives N earby
— —
F riends N earby — - -
In stru m ental S u p p ort — - -
E m otional S u p p ort — —
Incom e
- - —
C u rren t Sm oking
. . . .
H eavy D rinking — —
O besity — - -
L eisure A ctivities
H ypertension Never M arried
— — — —
Separated/D ivorced
- - — - - —
W idow ed — — — —
Fem ale
— — - -
Baseline SR H 0.61* 0.61*** 0.61* 0.56***
W hite 0.49** 0.49** 0.46*** 0.44***
E ducation — - - — —
Age — — —
Living C h ildren N/A N/A
132
Reproduced with permission o f th e copyright owner. Further reproduction prohibited without permission.
T able 5.23 continued
M odel 1 M odel 2 M odel 3 M odel 4 M odel 5 M odel 6
H ypertension C hildren N earby
R elatives N earby
F riends N earby
In strum ental S u p p ort
E m otional S u p p ort
Incom e —
C u rren t Sm oking
—
H eavy D rinking
—
O besity
L eisure Activities
—
D iabetes Never M arried
— — - - — —
Separated/D ivorced — — — — — —
W idow ed 0.34*** 0.33*** 0.35** 0.30*** 0.39** —
Fem ale — — — — — —
Baseline S R H 0.67* 0.68*
- - — — —
W hite — — — — — —
E ducation 0.89** 0.89** 0.84**
—
0.90* 0.83**
Age — — — . . - -
Living C hildren 0.80* 0.73*
C hildren N earby
- - —
Relatives N earby — - -
F riends N earby
— —
In strum ental S u pp ort
— —
E m otional S u pp ort
— - -
Incom e — —
C u rren t Sm oking — —
H eavy D rinking — —
O besity — —
L eisure Activities —
Reproduced with permission o f th e copyright owner. Further reproduction prohibited without permission.
T able 5.23 continued
M odel 1 M odel 2 M odel 3 M odel 4 M odel 5 M odel 6
C ancer Never M arried
— — — —
Separated/D ivorced 2.35* 2.36* 2.86* 3.42**
W idow ed
— — — - -
Fem ale — — — —
Baseline SR H
— — - -
W hite 2.06* 2.06*
— —
E ducation
— — — —
Age
— — — —
L iving C hild ren N /A N/A
C hildren N earby
R elatives N earby
F rien ds N earby
In stru m en tal S u p port
E m otional S u p p ort
Incom e —
N/A
C u rren t Sm oking
—
H eavy D rinking —
O besity —
L eisure Activities
H eart C onditions Never M arried — — - - —
Separated/D ivorced
— — —
W idow ed
— — — —
Fem ale 0.33*** 0.33*** 0.29*** 0.33**
Baseline SR H
— — — —
W hite
—
1.96*
— —
E ducation
— — — —
Age — — —
Living C h ild ren
C hildren N earby N/A N/A
Reproduced with permission o f th e copyright owner. Further reproduction prohibited without permission.
T able 5.23 continued
M odel 1 M odel 2 M odel 3 M odel 4 M odel 5 M odel 6
H ea rt C onditions R elatives N earby
F rien ds N earby
In strum ental S u p p ort
E m otional S up port
Incom e
—
C u rren t Sm oking —
H eavy D rinking —
O besity —
L eisure Activities —
D epression Never M arried
— — — — —
Separated/D ivorced
— — — . . —
W idow ed — - - — — — —
Fem ale — - - — — — —
Baseline SR H — — - - — — —
W hite 0.26*** 0.25*** 0.21*** 0.24** 0,26*** 0.12*
E ducation
— — — — —
Age 1.18**
- -
1.18** 1.18**
—
Living C hildren
— - -
C hildren N earby — - -
Relatives N earby — —
F rien ds N earby — —
In strum ental S u pp ort
— - -
E m otional Sup port
— —
Incom e 1.00** 1.00*
C u rren t Sm oking — —
H eavy D rinking
. . —
O besity — - -
L eisure Activities -
***p<.001, **p<.01, *p<.05; -- not significant
A nalyses of stroke an d broken hip w ere not technically possible because of the sm all sam ple size an d skewed sam ple distribution.
u >
Figure 5.1. Odds Ratios of Having ADL Difficulties by Marital
Status: HRS (1992) and AHEAD (1993)
■ HRS Males
□ HRS Females
□AHEAD Males
□AHEAD Females
Never M arried Separated/Divorced Widowed
6 -
5
4 -
With A ge Controlled
■ HRS Males
□ HRS Females
□AHEAD Males
□AHEAD Females
3
2
m
0 -
■ Iss
Never M arried Separated/Divorced Widowed
With Age, Self-rated Health, Education and Race
Controlled BHRS Males
□ HRS Females
□AHEAD Males
□AHEAD Females
Never M arried Separated/Divorced Widowed
With All M ech anism s Controlled
6 ■ HRS Males
□ HRS Females
□AHEAD Males
□AHEAD Females
-
Never M arried Separated/Divorced Widowed
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Figure 5.2. Odds Ratios of Having IADL Difficulties by Marital
Status and Sex: AHEAD (1993)
□AHEAD Males
□AHEAD Females
6
5
4
3
2
1
Separated/Divorced Widowed Never Mamed
0
6 -
5
4
3 - I
2
1
0 - i
With A ge Controlled
Never Mamed Separated/Divorced
□AHEAD Males
□AHEAD Females
Widowed
5 1
2
1
With A ge, Self-rated Health, Education and R ace
Controlled
□AHEAD Males
□ AHEAD Females
Never Mamed Separated/Divorced Widowed
With All M echanism s Controlled
□AHEAD Males
□AHEAD Females
Never Married Separated/Divorced Widowed
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Figure 5.3. Odds Ratios of Having Nagi Difficulties by Marital Status
and Sex: HRS (1992), AHEAD (1993) and MacArthur (1988)
■ HRS Males
□ HRS Females
□AHEAD Males
■AHEAD Females
□ MacArthur
2
1
Never Married Separated/Divorced Widowed
With A ge C ontrolled ..
■ HRS Males
□ HRS Females
5 - j □ AHEAD Males
I
4 J
□AHEAD Females
□ MacArthur
3 -
* J L ^ — ____________ ■ r a
q J Never Married Separated/Divorced Widowed
With Age, Self-rated Health, Education and Race
6
Controlled
■HRS Males
□ HRS Females
5
□AHEAD Males
4 -
□AHEAD Females
□ MacArthur
3 -
2
1 - I
0 ■
Never Married Separated/Divorced Widowed
6
5
4
3
2
1
0
With All M ech an ism s C ontrolled
■ HRS Males
□ HRS Females
□AHEAD Males
□AHEAD Females
□ MacArthur
Never Married Separated/Divorced Widowed
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Figure 5.4. Odds Ratios of Having Hypertension by Marital
Status and Sex: HRS (1992) and AHEAD (1993)_________
139
6
5
4
3
2
1
0 J
Never M arried Separated/Divorced
■ HRS Males
□ HRS Females
□AHEAD Males
□AHEAD Females
□ MacArthur
j= L
Widowed
With A ge Controlled
I HRS Males
-
□ HRS Females
□AHEAD Males
□AHEAD Females
□ MacArthur
Q F = 1
Never Mamed Separated/Divorced Widowed
With A ge, Self-rated Health, Education and Race
6 -
Controlled
■ HRS Males
5 -
□ HRS Females
□AHEAD Males
4 i □AHEAD Females
3 -
□ MacArthur
2 -
1 . i — i
\ s / s \
o J
Never Mamed Separated/Divorced Widowed
With All M echanism s Controlled
6 BHRS Males
□ HRS Females
□AHEAD Males
- □AHEAD Females
r—i
Never Mamed Separated/Divorced Widowed
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Figure 5.5. Odds Ratios of Having Diabetes by Marital Status and
Sex: HRS (1992). AHEAD (1993) and MacArthur (1988)________
6 i
■ HRS Males
□ HRS Females
5 -
□AHEAD Males
□AHEAD Females
4 □ MacArthur
3
2
1= 1 m ,,
0 -
Never Married Separated/Divorced
i i
Widowed ~ 1
iir .L * _ a ii j ■ HRS Males
With A g e C ontrolled QHRS Females
□AHEAD Males
-
F=1
□AHEAD Females
□ MacArthur
EE) P77J
Never Mamed Separated/Divorced Widowed * ■ —
With A ge, Self-rated Health, Education and R ace
Controlled
■ HRS Males
6
□ HRS Females
5 -
□AHEAD Males
□AHEAD Females
4 -
□ MacArthur
3
2
1
1
0 -
Never Married Separated/Divorced Widowed ^
With All M echanism s Controlled
6 -
■ HRS Males
□ HRS Females
5 - □AHEAD Males
4 -
□AHEAD Females
□ MacArthur
3 - I
2
1 _
0 -
Never Mamed Separated/Divorced Widowed
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Figure 5.6. Odds Ratios of Having Cancer by Marital Status
and Sex: HRS (1992), AHEAD (1993) and MacArthur (1988)
6 n
5
4
3 -
2 -
1
0
n
■ H R S M a le s
□ H R S F e m a l e s
□ A H E A D M a l e s
□ A H E A D F e m a l e s
□ M a c A r th u r
N e v e r M a r r ie d S e p a r a t e d / D i v o r c e d W id o w e d
6
5
4
3
2 ^
1
0
With A ge Ctonrolled
n
N e v e r M a m e d S e p a r a t e d / D i v o r c e d
■ H R S M a le s
□ H R S F e m a l e s
□ A H E A D M a l e s
□ A H E A D F e m a l e s
□ M a c A r th u r
W i d o w e d
With A ge, Self-rated Health, Education and R ace
Controlled b h r s M a le s
□ H R S F e m a l e s
□ A H E A D M a l e s
□ A H E A D F e m a l e s
□ M a c A r th u r
■
N e v e r M a rrie d S e p a r a t e d / D i v o r c e d W id o w e d
5 -
4 -
6
5 ^
4
3
2
With Ail M echanism s Controlled
■ H R S M a l e s
□ H R S F e m a l e s
□ A H E A D M a l e s
□ A H E A D F e m a l e s
0 J
N e v e r M a m e d S e p a r a t e d / D i v o r c e d W i d o w e d
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Figure 5.7. Odds Ratios of Having Lung Conditions by Marital
Status and Sex: HRS (1992) and AHEAD (1993)____________
6 -
5
4
3 H
2
1
0
22
N e v e r M a rr ie d S e p a r a t e d / D i v o r c e d
■ H R S M a l e s
□ H R S F e m a l e s
□ A H E A D M a l e s
□ A H E A D F e m a l e s
W id o w e d
6
5
4
3
2
1
0
With A g e Controlled
N e v e r M a m e d S e p a r a t e d / D i v o r c e d
■ H R S M a l e s
□ H R S F e m a l e s
□ A H E A D M a l e s
□ A H E A D F e m a l e s
W i d o w e d
6
5
4
3
2
1
0
With A ge, Self-rated Health, Education and R ace
Controlled
-
-
r — 1
N e v e r M a m e d S e p a r a t e d / D i v o r c e d
■ H R S M a l e s
□ H R S F e m a l e s
□ A H E A D M a l e s
□ A H E A D F e m a l e s
W i d o w e d
With All M echanism s Controlled
■ H R S M a l e s
□ H R S F e m a l e s
□ A H E A D M a l e s
□ A H E A D F e m a l e s
N e v e r M a m e d S e p a r a t e d / D i v o r c e d W id o w e d
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Figure 5.8. Odds Ratios of Having Heart Conditions by Marital
Status and Sex: HRS (1992), AHEAD (1993) and MacArthur (1988)
■ HRS Males
□ HRS Females
□AHEAD Males
□AHEAD Females
□ MacArthur
F=i i — i l / S A
Never Mamed Separated/Divorced Widowed
With A ge Controlled ■ HRS Males
fi
□ HRS Females
□AHEAD Males
5 □AHEAD Females
4 -
□ MacArthur
3
2
F=1
oJ
Never Married Separated/Divorced Widowed
6
5
4
3
2
1
0
With A ge, Self-rated Health, Education and R ace
Controlled BHRS Males
□ HRS Females
□AHEAD Males
□AHEAD Females
□ MacArthur
Never Married Separated/Divorced Widowed
6
5
4
3
2
1
0
With All M echanism s Controlled
-i ■ HRS Males
□ HRS Females
□AHEAD Males
- □AHEAD Females
-
□ MacArthur
U A H
Never Mamed Separated/Divorced Widowed
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Figure 5.9. Odds Ratios of Having Stroke by Marital Status
and Sex: HRS (1992), AHEAD (1993) and MacArthur (1988)
6
5
4
3
H
1
■ H R S M a le s
] □ H R S F e m a l e s
□ A H E A D M a l e s
□ A H E A D F e m a l e s
J
M -
N e v e r M a r r ie d S e p a r a t e d / D i v o r c e d W id o w e d
6 i
5
4
3 i
2
1
0 J
With A ge Controlled
■ H R S M a le s
□ H R S F e m a l e s
□ A H E A D M a l e s
□ A H E A D F e m a l e s
N e v e r M a m e d S e p a r a t e d / D i v o r c e d W id o w e d
6
5
4
3
2
1
0
With A ge, Self-rated Health, Education and Race
Controlled
■ H R S M a le s
□ H R S F e m a l e s
□ A H E A D M a l e s
□ A H E A D F e m a l e s
N e v e r M a m e d S e p a r a t e d / D i v o r c e d W id o w e d
With All M echanism s Controlled
■ H R S M a le s
□ H R S F e m a l e s
-
□ A H E A D M a l e s
-
□ A H E A D F e m a l e s
N e v e r M a m e d S e p a r a t e d / D i v o r c e d W id o w e d
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Figure 5.10. Odds Ratios of Having Arthritis by Marital Status
and Sex: HRS (1992) and AHEAD (1993)__________________
6
5 1
4 - !
3
2
1
0
■ HRS Males
□ HRS Females
□AHEAD Males
□AHEAD Females
F=1 1 F=^\Nt 7 7 3 I
N e v e r M a m e d S e p a r a t e d /D iv o r c e d W id o w e d
With A ge Controlled _
° BHRS Males
® ' QHRS Females
5 - □ AHEAD Males
4 J 0 AHEAD Females
3
2
1
F = l J 1 .-------T O
N e v e r M a rrie d
— i ------------- - ------------- :----
S e p a r a t e d /D iv o r c e d W id o w e d
With Age, Self-rated Health, E ducation and Race
6 -
Controlled
■ HRS Males
5 - □ HRS Females
4 1
3 1
□AHEAD Males
□AHEAD Females
2 -
1 _ F=1
1
0 J
Never Mamed Separated/Divorced Widowed
With All M echanism s Controlled
6 1 BHRS Males
5 - I DHRS Females
4
3 ^
2
1
□AHEAD Males
□AHEAD Females
q J Never M arried Separated/Divorced Widowed
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Figure 5.11. Odds Ratios of Having Broken Bone (HRS, 1992)
or Hip (AHEAD, 1993) by Marital Status and Sex____________
I HRS Males
6
□ HRS Females
□AHEAD Males
□AHEAD Females
5
4
3
2
^ I
W id o w e d
1
S e p a r a te d /D iv o r c e d N e v e r M a m e d
0
With A g e C ontrolled
I HRS Males
□ HRS Females
6
5
4
3
2
1
0
□AHEAD Males
□AHEAD Females
Separated/Divorced Widowed Never Mamed
With Age, Self-rated Health, Education and Race
Controlled
■ HRS Males
6
□ HRS Females
5
□AHEAD Males
4 □AHEAD Females
3
2
1 . r— i
0
Never Mamed Separated/Divorced ^^/Vidowed
With All M ech an ism s C ontrolled
6 ■ HRS Males
c
□ HRS Females
□AHEAD Males
4 -
□AHEAD Females
3
2
1 . 1
0 -
Never Married ^^Vidowed
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Figure 5.12. Odds Ratios of Having Depression by Marital Status
and Sex: HRS (1992). AHEAD (1993) and MacArthur (1988)
S e p a r a t e d /D iv o r c e d
■HRS Males
□ HRS Females
□AHEAD Males
□AHEAD Females
□ MacArthur
N e v e r M a m e d W i d o w e d
With Age Controlled
S e p a r a t e d /D iv o r c e d
■ HRS Males
□HRS Females
□AHEAD Males
□AHEAD Females
□ MacArthur
N e v e r M a m e d W i d o w e d
With A ge, Self-rated Health, Education and R ace
6 i
5
4
3
2
1
0
Controlled
m
N e v e r M a r r ie d S e p a r a t e d /D iv o r c e d
■HRS Males
□ HRS Females
□AHEAD Males
□AHEAD Females
□ MacArthur
W i d o w e d
6 i
5
4
3
2
1
0 J
With All M echanism s C ontrolled
■ H R S M a l e s
Q H R S F e m a l e s
□ A H E A D M a l e s
□ A H E A D F e m a l e s
N e v e r M a m e d S e p a r a t e d /D iv o r c e d W i d o w e d
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
148
The effects of being never married, separated/divorced and widowed on the
prevalence of each health condition are graphically presented in Figures 5.1 to 5.12.
The first graphs of each figure shows the odds ratios of having a condition on being
never married, being separated/divorced and being widowed compared to being
married. The odds ratios with age controlled were shown in the second graphs,
those with age, baseline self-rated health, education and race controlled in the third
graphs, and those in all factors including mechanisms controlled models in the last
graphs. As discussed earlier, in many cases, the effects of being unmarried
disappeared when age, baseline health, education, race and/or mechanisms were
considered. However, it is noticeable that some effects remained even when
mechanisms were controlled, suggesting that the effect of marital status may not
always occur through covariates.
Influence of marital status on health differed among the never married, the
separated/divorced and the widowed. The graphs show that being
separated/divorced and widowed were more often and more strongly likely to affect
health status than being never married (note the number and height of bars in
separated/divorced and widowed compared to those of never married). It suggests
that, relative to the married status, separated/divorced and widowed states adversely
affect health status while the relative effect of being never married status compared
to that o f the other unmarried statuses may be less.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
149
While being never married, being separated/divorced and being widowed
were found to be associated with selected health conditions, the direction of the
relationship was not consistent as discussed earlier. The bars below the x-axis
indicate that the never married, the separated/divorced or the widowed had better
health status than the married. For example, the widowed were less likely to have
diabetes compared to the married in MacArthur, but widows and
separated/divorced females in HRS were more likely to have diabetes (Figure 5.5).
In addition to differences in the direction of the relationships among three
unmarried groups, age and gender differences were also noticed. Younger cohorts
in HRS had more effects of marital status on health than older cohorts in AHEAD
and MacArthur (compare the number and height of bars in HRS with those of
AHEAD and MacArthur). Also, the effect of being unmarried on the prevalence of
health conditions tended to differ for men and women. In general, more and
stronger effects were found for men.1
Depression shows a good example of age and gender differences in the
effect of marital status on health (Figure 5.12). The probability of having
depression among the unmarried was higher among the HRS respondents than
among the AHEAD respondents, indicating the age difference (HRS vs. AHEAD);
the effects of unmarried status on the prevalence of depression were greater among
1 The obvious exception was the effect on broken hip: Being separated/divorced was strongly related
to a greater likelihood of having a broken bone or hip among females, but not among males (Figure
5.11).
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
150
males than among females both in HRS and AHEAD, indicating the gender
difference.
Examination of the relationship o f three different unmarried statuses in
comparison with the married group showed that even though no one group is
absolutely better or worse than the others in all health conditions in general, the
never married were less different from the married than the separated/divorced and
the widowed in health status, which has been demonstrated in previous research
(Verbrugge, 1979).
Incidence o f Health Conditions on Marital Status and Other Variables
A similar pattern of results was found in the onset analyses (Tables 5.24 to
5.28 and Figures 5.13 to5.25). Among the HRS sample, widowers were more
likely to get ADL difficulties, new hypertension and depression (Table 5.24) while
widows were more likely to die and get new lung conditions, heart conditions and
depression (Table 5.25) between 1992 and 1994. On the other hand, being
separated/divorced was related to a lower probability of getting new diabetes and a
higher probability of getting ADL difficulties, new hypertension, cancer, stroke and
depression among the HRS male respondents (Table 5.24). The positive
relationship between being never married and getting new lung conditions among
the HRS female respondents disappeared by including current smoking status in
model (model 5 in Table 5.25), and a higher probability o f getting new heart
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
151
conditions among separated/divorced and widowed females also disappeared after
taking friendship satisfaction (model 2) and obesity and physical activity (model 5)
into account (Table 5.25). As in the prevalence o f health conditions, being
separated/divorced or widowed was strongly related to getting depression, implying
that unmarried status precipitates a new mental health problem. The onset of ADL
difficulties and hypertension among the HRS widowers, that of Nagi difficulties
among the never married males and that of depression among the
separated/divorced males (model 5 in Table 5.24) and among separated/divorced
females (model 5 in Table 5.25) remained significant when all mechanisms were
considered (Figures 5.13). On the other hand, current smoking status accounted for
the adverse effect of being widowed on getting depression found in models 1, 2 and
3 for females in HRS (models 4 and 5 in Table 5.25), and so did friendship
satisfaction and health insurance as well as current smoking when all mechanisms
were controlled in model 5 (Table 5.25).
In terms of getting new health conditions and diseases in AHEAD (Tables
5.26 and 5.27, and Figures 5.13 to 5.25), the results showed that separated/divorced
men were less likely than the married to get IADL difficulties (Table 5.26 and
Figure 5.14) and new arthritis (Table 5.26 and Figure 5.23) between 1993 and 1998,
and never married men were more likely than married men to die between 1993 and
1995, and this effect remained even when all other covariates were considered
(model 5 in Table 5.26 and Figure 5.16). Widows were about 1.5 times more likely
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
152
to get a stroke than the married women (Table 5.27 and Figure 5.22). Contrary to
the strong effect of being unmarried on the prevalence of depression (Figure 5.12),
no effect was found in the effect of being unmarried on the incidence of depression
among the AHEAD respondents (Tables 5.26 and 5.27 and Figure 5.25). That is,
the onset of depression was not related to unmarried status in AHEAD. No
proposed mechanisms were involved in the relationship between the incidence of
conditions and marital status both for men and women (Tables 5.26 and 5.27).
In MacArthur (Table 5.28), being widowed predicted getting Nagi
difficulties, but contrary to the strong adverse effect of being separated/divorced or
widowed on depression in HRS and AHEAD, being unmarried was not related to a
significantly higher probability of getting depression.
In terms of mechanisms associated with the relationships between marital
status and the incidence of health conditions, several mechanisms were found to
affect the relationships, and different mechanisms were involved in different health
outcomes. Among the HRS male respondents, obesity explained part of the effect
of being separated/divorced on the onset of diabetes and stroke; current smoking
explained part of the effect of being separated/divorced on the onset of cancer; and
satisfaction with neighborhood and friendships explained part of the effect of being
widowed on the incidence of depression (model 5 in Table 5.24). Among the HRS
female respondents, smoking was related to the effect of being widowed on the
incidence of lung conditions and depression; satisfaction with friendships and
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
153
supplemental health insurance was on the incidence o f depression among widows;
and obesity and participation in physical activity was to the effects of being
separated/divorced or widowed on the incidence of heart conditions (model 5 in
Table 5.25). Among the MacArthur respondents (Table 5.28), the only
mechanisms involved were income (model 4) and participation in leisure activity
(model 6) in the relationship between being widowed and the incidence of Nagi
difficulties.
While analyses were done by comparing three unmarried groups with the
married group, a quick analysis using the widowed as the reference in equations
was made in order to make a direct comparison among three unmarried groups.
Overall, as suggested by the results from the previous analyses, the widowed
tended to be more likely to have worse health compared to the never married or the
separated/divorced, and in most cases, the odds of having or acquiring health
conditions were greater among the separated/divorced than among the never
married. However, there were instances in which the separated/divorced and the
never married were worse than the widowed. For instance, in HRS, never married
males were less likely to have depression in 1992 than widows, but there was no
statistical difference between the separated/divorced and the widowed in the
prevalence of depression. On the other hand, separated/divorced males were more
likely to have broken bone than widowers in HRS, but no such effect was found for
the never married. In total, the results from indirect comparison of unmarried
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
154
groups were confirmed such that among unmarried people, the widowed were
worst, and the separated/divorced were next, and the never married were better than
the other two unmarried groups.
Differential Effect of Marital Status on Health and Differential Mechanisms by
Type o f Health Outcome
This study hypothesized that the effect of marital status on health would
differ by type of health conditions: Marital status would be more related to
functioning difficulties, death, fatal conditions and mental condition than nonfatal
conditions. The differential effect of marital status by type of health outcome
appeared to be more obvious in functioning and mental health: The effect o f being
unmarried was much stronger on functioning limitations and the indicator of a
mental health, depression.
Figure 5.26 shows a diagram of the effects of being never married,
separated/divorced and widowed on health categorized as a. functioning problems,
b. death, c. nonfatal health conditions, d. mental health condition, and e. fatal health
conditions in model where all the proposed mechanisms were controlled for both
males and females. Arrows indicate the significant effect of being never married,
separated/divorced or widowed on each health condition. Several things are
noticed. First, all unmarried statuses increased (+) the probability of having or
getting health conditions except between being widowed (in HRS) and Nagi
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
155
difficulties, which was negatively related (-). It appears that functioning and
mental health are more strongly related to marital status than other physical
conditions or chronic diseases.
Second, it was hypothesized that fatal conditions would be more likely to be
associated with marital status than nonfatal conditions. In Figure 5.26, c had fewer
arrows than e, confirming the hypothesis.
Third, IADL difficulties, diabetes and cancer were not related to marital
status in final models in which all proposed mechanisms were controlled. The
commonality of them is probably in their severity: Cancer and diabetes are non
incapacitating until the final stage of disease as IADL difficulties are less restricting
in their nature compared to ADL difficulties.
In total, findings suggest that the effect of marital status on health differs by
type of health outcomes, and fatal conditions were more affected by marital status
than nonfatal conditions and functioning and mental health were more affected by
marital status than other physical chronic diseases or conditions. The reasons why
some health conditions were affected while others were not, or why some
conditions were more strongly affected while others were less affected, can be
better understood by seeing what mechanisms were involved in the effect of marital
status on each health outcome.
Using the results from the logistic regression analysis of the HRS and
AHEAD sample, Figure 5.27 draws a diagram showing the relationship between
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
156
three sets of mechanisms and health. Arrows indicate the significant effect of the
mechanisms on the relationship between marital status and health when all the
mechanisms were considered together. Solid lines represent the effect for males
and dotted lines for females. Mechanisms that have more arrows do not necessarily
mean that they are more frequently related because the HRS and AHEAD have
different variables representing each set of mechanisms and a different number of
variables in each set (e.g. 3 different variables representing social support in HRS
and AHEAD— friendship satisfaction, neighborhood satisfaction and number of
people in the same household in HRS, and number of living children, living
siblings and people in the same household in AHEAD; 3 variables representing
health behaviors in HRS and 4 in AHEAD— smoking, drinking and obesity in
AHEAD, and an additional variable, physical activity in HRS).
The proposed mechanisms were found to have some influence on the
relationships between marital status and health, and different mechanisms played a
role in the relationship between marital status and different health outcomes.
Regarding social support, the adverse effect of being widowed among males
disappeared with the consideration of friendship satisfaction: For those widowed
males, satisfaction with friendships partially explained than the effect if being
widowed on the prevalence of lung conditions, and the prevalence and incidence of
depression. The same effect of friendship satisfaction occurred on the prevalence
and incidence of depression among the never married and separated/divorced
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
157
females. Satisfaction with neighborhood was the mechanism by which being never
married or separated/divorced increased the incidence of depression among males,
and the prevalence of depression among females, and the number of same
household members was the one between being separated/divorced and the
prevalence of depression among females. That is, higher neighborhood satisfaction
and more people in the same household were the mechanisms by which being
unmarried adversely affected depression among females. No other social support
variables were found to be associated with the relationship between marital status
and health outcomes in HRS and AHEAD. The number of living children was the
only social support mechanism found to be related to the effect of being unmarried
on health in MacArthur: Those with more children were less likely to have diabetes,
and the number of children partly explained the negative effect of being widowed
on the prevalence of diabetes.
Regarding economic resources, income was the significant mechanism by
which being never married increased the prevalence of arthritis among the HRS
male respondents. On the other hand, the availability of supplemental health
insurance was how being widowed increased the prevalence of depression among
males, and the prevalence and incidence of depression among females.
Health behaviors were the mechanisms that most often affected the
relationship between marital status and health conditions. Heavy drinking was the
mechanism by which being widowed adversely affected the prevalence of ADL
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
158
difficulties and hypertension among females. Current smoking was the mechanism
of the adverse effect o f being separated/divorced on the incidence of cancer, and
that of being widowed on the prevalence of lung conditions and depression while
obesity was on the incidence of diabetes and stroke among males. For females, the
adverse effect of being widowed on the prevalence of hypertension was through
smoking, drinking and obesity, and that of depression through smoking. A higher
probability of getting heart conditions among the separated/divorced and the
widowed disappeared when considering obesity and activity participation. Among
never married females, their higher probability of having arthritis relative to
married females was somewhat explained by obesity. A higher incidence rate of
lung conditions and depression for separated/divorced women compared to their
married counterparts occurred through smoking. In sum, the results showed that
the proposed mechanisms played some role in explaining the effect of marital
status on health. Fatal conditions appeared to be more affected by health behaviors
than other proposed mechanisms, and all three sets of proposed mechanisms such
as social support, economic resources and health behaviors were strongly related to
depression. But other than that, there was no consistent pattern showing that
mechanisms were differentiated by type of health outcomes.
This chapter examined the effect of marital status on health and
mechanisms by which marital status affects health. Overall, married elderly people
were healthier than the unmarried elderly, and the never married were healthier
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
159
than the separated/divorced and the widowed. Differences in direction and size of
effects found in this study were largely attributed to gender, age and type of health
outcomes examined such that unmarried statuses more adversely affected men,
younger cohorts, and functioning problems, mental health and some fatal
conditions.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Reproduced with permission o f th e copyright owner. Further reproduction prohibited without permission.
T able 5.24. O dds R atios from Logistic R egression of G etting Functioning Difficulties, New Diseases and C onditions betw een 1992
an d 1994 on M arital States an d O ther V ariables am ong M ales: H R S___________________________________________
M odel 1 M odel 2 M odel 3 M odel 4 M odel 5
A DL D ifficulties N ever M arried — — - - — —
Separated/D ivorced — — —
1.83*
—
W idow ed 4.35*** 4.17*** 3.93** 4.16** 3.60**
Age — — — — —
Baseline SR H 0.45*** 0.46*** 0.48*** 0.48*** 0.51***
W hite 0.55* — 0.60* 0.51* —
E ducation — - - — —
H ousehold M em bers
— —
Satisfaction w ith
N eighborhood
" —
Satisfaction w ith
F riendships
— - -
Incom e
— —
H ealth In surance — - -
C u rren t Sm oking
- - —
H eavy D rinking
— —
O besity — —
Physical Activity — —
N agi Difficulties N ever M arried 2.12** 1.94* 2.25** 2.53** 2.47*
Separated/D ivorced — - - — — —
W idow ed — — - - — —
Age — — — — —
Baseline SR H 0.65*** 0.68*** 0.65*** 0.66*** 0.70***
W hite 1.45*
—
1.46*
— —
E ducation 0.94*** 0.94** 0.94** 0.94** 0.94*
H ousehold M em bers
— —
Satisfaction w ith
N eighborhood
- -
160
Reproduced with permission o f th e copyright owner. Further reproduction prohibited without permission.
T able 5.24 continued
M odel 1 M odel 2 M odel 3 M odel 4 M odel 5
N agi D ifficulties Satisfaction w ith
F riendships
0.64* —
Incom e — —
H ealth In suran ce — —
C u rren t Sm oking — - -
H eavy D rinking
— —
O besity 1.39* 1.45*
Physical Activity
— —
D eath Never M arried — — — — —
Separated/D ivorced — — — —
W idow ed — — — — —
Age — — - - — - -
Baseline SR H 0.39*** 0.41*** 0.42*** 0.40***
W hite — — - - —
E ducation — — — —
H ousehold M em bers — —
Satisfaction w ith
N eighborhood
— 3.87*
Satisfaction w ith
Friendships
— —
Incom e
- - —
H ealth In surance — —
C u rren t Sm oking — —
H eavy D rinking — —
O besity - -
Physical Activity — - -
H ypertension Never M arried ~ — — — —
Separated/D ivorced —
1.70*
— — —
W idow ed 4.24** 4.70*** 4.17** 3.49* 3.43*
Reproduced w ith permission o f th e copyright owner. Further reproduction prohibited without permission.
T able 5.24 continued
M odel 1 M ode! 2 M odel 3 M odel 4 M odel 5
H ypertension Age
— — — — —
Baseline S R H 0.74*** 0.73*** 0.74*** 0.76*** 0.77**
W hite
— — — — - -
E ducation
— — — —
H ousehold M em bers
— —
Satisfaction w ith
N eighborhood
-- —
Satisfaction w ith
F riendships
— —
Incom e — —
H ealth In suran ce
— —
C u rren t Sm oking
— „
H eavy D rinking
— —
O besity 1.73* 1.79**
Physical A ctivity — —
D iabetes Never M a rried
— — —
Separated/D ivorced 0.28* 0.25* 0.28*
- - —
W idow ed
— — — —
0.32*
Age
— — — — —
Baseline SR H 0.64*** 0.64*** 0.64*** 0.74** 0.69**
W hite 0.48** 0.43** 0.48**
— —
E ducation 0.93* 0.92* 0.93* 0.92* 0.91*
H ousehold M em bers
— —
Satisfaction w ith
N eighborhood
- - - -
Satisfaction w ith
F riendships
— —
Incom e — —
H ealth In su ran ce — —
Reproduced w ith permission o f th e copyright owner. Further reproduction prohibited without permission.
T able 5.24 continued
M odel 1 M odel 2 M odel 3 M odel 4 M odel 5
D iabetes C u rren t Sm oking
— —
H eavy D rinking — —
O besity 2 .2 0 ** 2.15**
Physical A ctivity
. . . —
C ancer Never M a rried ~ — - - — —
Separated/D ivorced 2.16* 1.99* 2.08*
— - -
W idow ed — — —
Age 1.09* —
1.09* 1.12* 1.1 1*
Baseline SR H
— — — —
W hite
— — — —
E ducation
— — — —
H ousehold M em bers — —
Satisfaction w ith
N eighborhood
— - -
Satisfaction w ith
F riendships
— - -
Incom e
— —
H ealth In su ran ce —
C u rren t Sm oking 2.47** 2.50**
H eavy D rinking — - -
O besity
— - -
Physical Activity
— —
L ung C onditions Never M arried
— — — — —
Separated/D ivorced — — — — —
W idow ed
- - — — — —
Age — — — — —
Baseline SR H 0.59*** 0.59*** 0.58*** 0.62*** 0.61***
W hite — — — —
E ducation 0.89*** 0.89*** 0.89*** 0.91** 0.91*
Reproduced with permission o f th e copyright owner. Further reproduction prohibited without permission.
T able 5.24 continued
M odel 1 M odel 2 M odel 3 M odel 4 M odel 5
L ung C onditions H ousehold M em bers
- - ~
Satisfaction w ith
N eighborhood
"
—
Satisfaction w ith
F riendships
— --
Incom e
— - -
H ealth In su ran ce — —
C u rren t Sm oking 2 .6 8 *** 2 .6 8 ***
H eavy D rinking
— - -
O besity — —
Physical A ctivity
— —
H ea rt C onditions Never M arried — — - - — —
Separated/D ivorced — — — — - -
W idow ed — — — - - - -
Age 1.07* 1.08** 1.07* 1.07* 1.1 0**
Baseline SR H 0.62*** 0.61*** 0.61*** 0.63*** 0.62***
W hite — — - - - - —
E ducation — — « - - —
H ousehold M em bers — —
Satisfaction w ith
N eighborhood
" - -
Satisfaction w ith
F riendships
" —
Incom e
— - -
H ealth In surance
—
C u rren t Sm oking 1.94*** 2.14***
H eavy D rinking
— —
O besity — —
Physical Activity — —
Reproduced with permission o f th e copyright owner. Further reproduction prohibited without permission.
T able 5.24 continued
M odel 1 M odel 2 M odel 3 M odel 4 M odel 5
S trok e Never M arried — — — — —
Separated/D ivorced 2.62* 2.41* —
2.48*
W idow ed — — — — —
Age 1.16* 1.15* 1.16* 1.17* 1.16*
Baseline SR H 0.64** 0.64** 0.67* 0 .6 6 * 0.70*
W hite — — — — —
E ducation — — — — . .
H ousehold M em bers
— - -
Satisfaction w ith
N eighborhood
" - -
Satisfaction w ith
Friendships
— - -
Incom e
. . . —
H ealth Insurance
— - -
C u rren t Sm oking — —
H eavy D rinking
- - - -
O besity 3.20** 3.23**
Physical Activity
— —
A rth ritis Never M arried — — — — —
Separated/D ivorced — — — — —
W idow ed — — — —
Age 1.06** 1.07** 1.06** 1.08** 1.08**
Baseline SR H 0.73*** 0.73*** 0.74*** 0.72*** 0.74***
W hite ~ — — ~ —
E ducation 0.96*
— — — —
H ousehold M em bers
— —
Satisfaction w ith
N eighborhood
— —
Reproduced with permission o f th e copyright owner. Further reproduction prohibited without permission.
T able 5,24 continued
M odel 1 M odel 2 M odel 3 M odel 4 M odel 5
A rth ritis Satisfaction w ith
F riendships
— --
Incom e
—
1.0 0 **
H ealth In suran ce
- - —
C u rren t Sm oking — —
H eavy D rinking — —
O besity
—
Physical Activity — —
B roken Bone Never M arried — — — — —
Separated/D ivorced — — — — —
W idow ed — — — — - -
Age - - — — — —
Baseline SR H — — - - — —
W hite — — — — —
E d ucation — — — —
H ousehold M em bers
— —
Satisfaction w ith
N eighborhood
—
—
Satisfaction w ith
F riendships
— —
Incom e
— —
H ealth In suran ce — —
C u rren t Sm oking
- - —
H eavy D rinking
— —
O besity — —
P hysical Activity
. . —
D epression N ever M arried — — —
1.89*
- -
Separated/D ivorced 3.40*** 3.15*** 3.25*** 3.71*** 3.34***
W idow ed 2.76** - 2.61* — -
Reproduced with permission o f th e copyright owner. Further reproduction prohibited without permission.
T able 5.24 continued
M odel 1 M odel 2 M odel 3 M odel 4 M odel 5
D epression Age 0.95** 0.95** 0.95**
— —
Baseline SR H 0.59*** 0.59*** 0.60*** 0.61*** 0.61***
W hite — — — — —
E ducation 0.93*** 0.92*** 0.94** 0.93*** 0.92***
H ousehold M em bers — —
Satisfaction w ith
N eighborhood
0.45*** 0.49***
Satisfaction w ith
F riendships
0.50*** 0.46***
Incom e — —
H ealth In su ran ce — —
C u rren t Sm oking
— —
H eavy D rinking — —
O besity — —
Physical A ctivity — -
***p<.001, **p<.01, *p<.05
-- not significant
T able 5.25. O dds R atios from Logistic R egression of G etting F unctioning Difficulties, New Diseases an d C onditions betw een 1992
M odel 1 M odel 2 M odel 3 M odel 4 M odel 5
A DL D ifficulties Never M arried — — — — —
Separated/D ivorced ~ — — — —
W idow ed — — — — —
Baseline SR H 0.50*** 0.51*** 0.52*** 0.55*** 0.57***
W hite 0.42*** 0.41*** 0.44*** 0.42** 0.39**
E ducation ~ - — — —
o \
Reproduced w ith permission o f th e copyright owner. Further reproduction prohibited without permission.
T able 5.25 continued
M odel 1 M odel 2 M odel 3 M odel 4 M odel 5
A D L Difficulties Age — — — —
H ousehold M em bers — —
Satisfaction w ith
N eighborhood
— —
Satisfaction w ith
F riendships
— —
Incom e —
H ealth In su ran ce —
C u rren t Sm oking — —
H eavy D rinking — —
O besity — —
Physical A ctivity — —
N agi D ifficulties Never M arried — — — —
Separated/D ivorced - - — — - -
W idow ed — — - - — —
Age — — — — —
Baseline SR H 0.61*** 0.61*** 0.60*** 0.64*** 0.64***
W hite — « — — —
E ducation 0.91*** 0.92*** 0.92**
— —
H ousehold M em bers
— —
Satisfaction w ith
N eighborhood
0.65* —
Satisfaction w ith
F riendships
— - -
Incom e
- - —
H ealth In su ran ce — —
C u rren t Sm oking
— —
H eavy D rinking — —
O besity 1.95** 2 .0 1 **
Reproduced w ith permission o f th e copyright owner. Further reproduction prohibited without permission.
T able 5.25 continued
M odel 1 M odel 2 M odel 3 M odel 4 M odel 5
N agi D ifficulties Physical A ctivity
— —
D eath Never M arried — — - - — —
Separated/D ivorced — - - — . . —
W idow ed
—
2.30*
— — —
Age — — — - - —
Baseline SR H 0.43*** 0.44*** 0.47*** 0.45*** 0.50***
W hite — — - - — —
E ducation — — — —
H ousehold M em bers 1.24* 1.34*
Satisfaction w ith
N eighborhood
— —
Satisfaction w ith
F riendships
- - —
Incom e
— . .
H ealth In su ran ce
— - -
C u rren t Sm oking — —
H eavy D rinking — —
O besity — —
Physical A ctivity 0.37*
—
H ypertension Never M arried
— - - — — —
Separated/D ivorced
— — — — —
W idow ed — — - - — —
Age
— — — - - . .
Baseline SR H — — — - - —
W hite 0.54** 0.51** 0.55** 0.47** 0.43**
E ducation
— — — - - —
H ousehold M em bers — - -
Satisfaction w ith
N eighborhood
—
Reproduced with permission o f th e copyright owner. Further reproduction prohibited without permission.
Table 5.25 continued
M odel 1 M odel 2 M odel 3 M odel 4 M odel 5
H ypertension Satisfaction w ith
F riendships
— —
Incom e — —
H ealth In su ran ce
— —
C u rren t Sm oking — _ _
H eavy D rinking — —
O besity 2 .1 2* * 2.09**
Physical A ctivity - - —
D iabetes Never M arried — — — — —
Separated/D ivorced — — — — —
W idow ed — — - - — —
Age 1.1 0* 1.1 0* 1.1 0*
—
1.13*
Baseline S R H 0.57*** 0.57*** 0.57*** 0.63** 0.65**
W hite — — — — —
E ducation — — — — —
H ousehold M em bers
— —
Satisfaction w ith
N eighborhood
— "
Satisfaction w ith
F riendships
- - -
Incom e
- -
H ealth Insu ran ce
— —
C u rren t Sm oking
— —
H eavy D rinking
— . .
O besity — —
Physical Activity — —
C ancer Never M arried - - — — — —
Separated/D ivorced — - - — — —
W idow ed — - — — —
Reproduced with permission o f th e copyright owner. Further reproduction prohibited without permission.
T able 5.25 continued
M odel 1 M odel 2 M odel 3 M odel 4 M odel 5
C ancer Age — — — — —
Baseline SR H
- - — — - - —
W hite
— — — — - -
E ducation — — — _ _
H ousehold M em bers — —
Satisfaction w ith
N eighborhood
— —
Satisfaction w ith
F riendships
— —
Incom e — —
H ealth Insurance
— —
C u rren t Sm oking — —
H eavy D rinking — - -
O besity
. . . - -
Physical Activity
« . .
L ung C onditions Never M arried — — — — —
Separated/D ivorced
— — — — . . .
W idow ed 1.8 6 * 1.90*
- -
2.27*
—
Age
— - - — — - -
Baseline SR H 0 .6 8 *** 0.69*** 0.71*** 0.63*** 0.65***
W hite
— — - - - - - -
E ducation
- - — — — —
H ousehold M em bers
- - —
Satisfaction w ith
N eighborhood
- - —
Satisfaction w ith
F riendships
— —
Incom e
- - —
H ealth Insurance - -
Reproduced with permission o f th e copyright owner. Further reproduction prohibited without permission.
T able 5.25 continued
M odel 1 M odel 2 M odel 3 M odel 4 M odel 5
L un g C onditions C u rren t Sm oking 2.45** 2.34**
H eavy D rinking — —
O besity
— —
Physical Activity
- - —
H ea rt C onditions Never M arried — — — — —
Separated/D ivorced 1.57*
—
1. 6 8 *
— «
W idow ed 1.6 6 * —
1.79*
— —
Age
— — — — —
Baseline SR H 0 ,6 8 *** 0 .6 8 *** 0 ,6 6 *** 0.72** 0.70**
W hite — — — — —
E ducation — — —
H ousehold M em bers
— —
Satisfaction w ith
N eighborhood
— —
Satisfaction w ith
F riendships
0,54* —
Incom e — —
H ealth In suran ce
— —
C u rren t Sm oking
— —
H eavy D rinking — „
O besity
—
1.76*
Physical Activity 0.53* 0.53*
S troke Never M arried — — « — —
Separated/D ivorced
— — — — . .
W idow ed — — — — —
Age — — — — —
Baseline SR H 0.54*** 0.52*** 0.57** 0.55** 0.56*
W hite
— — —
E ducation - - - —
Reproduced with permission o f th e copyright owner. Further reproduction prohibited without permission.
T able 5.25 continued
M odel 1 M odel 2 M odel 3 M odel 4 M odel 5
S trok e H ousehold M em bers — —
Satisfaction w ith
N eighborhood
— —
Satisfaction w ith
F riendships
" - -
Incom e
— —
H ealth In surance
— - -
C u rren t Sm oking
—
H eavy D rinking — —
O besity — —
Physical Activity — —
A rth ritis N ever M arried
— — — — —
Separated/D ivorced — — — — —
W idow ed — — — — —
Age — — — — —
Baseline SR H 0.72*** 0.73*** 0.72** 0.76*** 0.76***
W hite — — — - - —
E ducation — — — — —
H ousehold M em bers
— —
Satisfaction w ith
N eighborhood
— —
Satisfaction w ith
Friendships
0.53** 0.51*
Incom e — —
H ealth In su ran ce — —
C u rren t Sm oking
—
H eavy D rinking — —
O besity 1.94*** 1.93***
Physical Activity 0.70* 0.71*
Reproduced with permission o f th e copyright owner. Further reproduction prohibited without permission.
T able 5.25 continued
M odel 1 M odel 2 M odel 3 M odel 4 M odel 5
B roken Bone N ever M arried — - - — — —
Separated/D ivorced — — — — —
W idow ed - — — - - —
Age — - - — — —
Baseline SR H 0.77*** 0.77*** 0.76*** 0.74*** 0.75**
W hite — — — — - -
E ducation
— — — — - -
H ousehold M em bers
— - -
Satisfaction w ith
N eighborhood
— —
Satisfaction w ith
Friendships
— - -
Incom e — - -
H ealth Insurance 1.49* 1.70*
C u rren t Sm oking — —
H eavy D rinking
— —
O besity
— —
Physical Activity
— —
D epression Never M arried — — — — —
Separated/D ivorced 1.54*** 1.53*** 1.45** 1.62** 1.50*
W idow ed 1.59** 1.56** 1.47**
— —
Age 0.97* 0.97* 0.97*
— —
Baseline SR H 0.64*** 0.64*** 0.65*** 0.64*** 0 . 6 6 ***
W hite 0.72** 0.77* 0.76*
— —
E ducation 0.89*** 0 .8 8 *** 0.91*** 0.87*** 0 .8 8 ***
H ousehold M em bers ~ —
Satisfaction w ith
N eighborhood
0.74* —
174
Reproduced with permission o f th e copyright owner. Further reproduction prohibited without permission.
T able 5.25 continued
M odel 1 M odel 2 M odel 3 M odel 4 M odel 5
D epression Satisfaction w ith
Friendships
0.63** 0.62*
Incom e — —
H ealth Insu ran ce 0.73** 0.76*
C u rren t Sm oking 1.54*** 1.54***
H eavy D rinking — —
O besity — - -
Physical A ctivity — —
***p<.001, **p<.01, *p<.05
-- not significant
T able 5.26. O dds R atios from L ogistic R egression of G etting Functioning Difficulties, New Diseases and C onditions betw een
1993 and 1998 on M arital States and O ther V ariables am ong M ales: AHEAD ___________ ___________ ________
M odel 1 M odel 2 M odel 3 M odel 4 M odel 5
A D L Difficulties Never M arried — — — — —
Separated/D ivorced — — — —
W idow ed
— — — —
Age 1.04** 1.05** 1.04** 1.05** 1.05**
Baseline SR H 0.67*** 0.67*** 0 .6 6 *** 0.67*** 0.67***
W hite — — — — - -
E ducation 0.95* 0.95* 0.94* 0.96* 0.94*
H ousehold M em bers
— —
Living C hildren — —
Living Siblings
— «
Incom e — —
H ealth In su ran ce
. . . —
C u rren t Sm oking -
Reproduced with permission o f th e copyright owner. Further reproduction prohibited without permission.
T able 5.26 continued
M odel 1 M odel 2 M odel 3 M odel 4 M odel 5
A DL Difficulties H eavy D rinking
— —
O besity 1.70* 1.6 6 *
1A D L Difficulties Never M arried — — — —
Separated/D ivorced 0.34* 0.35*
—
0.34*
—
W idow ed — — — - -
Age 1.09*** 1.1 0*** 1.1 0*** 1.09*** 1. 10***
Baseline SR H 0.65*** 0.65*** 0 .6 6 *** 0.65*** 0 .6 6 ***
W hite — — — « —
E ducation 0.91*** 0.91*** 0.92*** 0.91*** 0.92***
H ousehold M em bers — —
Living C h ild ren —
Living Siblings — —
Incom e
— —
H ealth In su ran ce — —
C u rren t Sm oking
— —
H eavy D rinking — —
O besity
—
N agi Difficulties N ever M arried — — — — —
Separated/D ivorced
— — — — —
W idow ed 1.47* 1.47* 1.46* 1.45* 1.45*
Age 1.04** 1.04** 1.04** 1.04** 1.04**
Baseline SR H 0.55*** 0.55*** 0.55*** 0.55*** 0.55***
W hite — — — - -
1.58*
E ducation 0.96* 0.96*
—
0.96*
« ■
H ousehold M em bers —
Living C h ildren
— —
Living Siblings
— —
Incom e
— —
H ealth In su ran ce -
176
Reproduced with permission o f th e copyright owner. Further reproduction prohibited without permission.
T able 5.26 continued
M odel 1 M odel 2 M odel 3 M odel 4 M odel 5
N agi D ifficulties C u rren t Sm oking — —
H eavy D rinking — —
O besity — —
D ea th * Never M arried 2.17* 2.07* 2.24* 2.30** 2.29*
Separated/D ivorced — — — — —
W idow ed — — . . . - - —
Age 1.1 1*** 1.1 1*** l.n*** 1.1 0*** 1.1 1***
Baseline SR H 0.57*** 0.57*** 0.57*** 0.57*** 0.57***
W hite
— — — . .
E ducation — — — — —
H ousehold M em bers 1.2 1 ** 1.2 1 ***
Living C hildren
„ —
Living Siblings — —
Incom e — —
H ealth In su rance
- - - -
C u rren t Sm oking 1.54* 1.58**
H eavy D rinking — - -
O besity — —
H ypertension Never M arried
— — — — —
Separated/D ivorced
— - - — — . .
W idow ed — — — — —
Age
— — — —
Baseline SR H
— — —
W hite
— — —
E ducation
— — —
H ousehold M em bers
— —
Living C hildren 0.90* 0.90*
L iving Siblings
— —
Incom e 1.0 0 * 1.0 0 *
Reproduced with permission o f th e copyright owner. Further reproduction prohibited without permission.
T able 5.26 continued
M odel 1 M odel 2 M odel 3 M odel 4 M odel 5
H ypertension H ealth Insu ran ce — —
C u rren t Sm oking
— —
H eavy D rinking
— —
O besity 1.81* 1.8 6 *
D iabetes Never M a rried — — . . .
Separated/D ivorced - - - - — — —
W idow ed — - - — - - «
Age
- - - - — — —
Baseline SR H
— — — — —
W hite
— — — —
E ducation
— —
0.91*
—
0.92*
H ousehold M em bers 0.50* 0.51*
L iving C h ildren — —
Living Siblings — - -
Incom e
— —
H ealth In su ran ce
— - -
C u rren t Sm oking —
H eavy D rinking — —
O besity 3.23*** 3.26***
C ancer Never M arried
- - — — —
Separated/D ivorced
— — — — —
W idow ed
— — — - -
Age
— - - — - -
Baseline SR H
— — — — —
W hite
— — — — —
E ducation
— — — — —
H ousehold M em bers
— —
Living C h ildren
— —
Living Siblings - —
Reproduced with permission o f th e copyright owner. Further reproduction prohibited without permission.
T able 5.26 continued
M odel 1 M odel 2 M odel 3 M odel 4 M odel 5
C ancer Incom e — —
H ealth In surance
- - —
C u rren t Sm oking — —
H eavy D rinking — —
O besity — —
L ung C onditions N ever M arried
— — — — —
Separated/D ivorced
— — — — —
W idow ed — — — - - —
Age
- - — — — —
Baseline SR H 0.78* 0.77*
- -
0.79*
W hite
— —
2.30*
—
2.32*
E ducation 0.92* 0.92* 0.91* 0.91*
—
H ousehold M em bers —
Living C hildren
— —
L iving Siblings — —
Incom e — —
H ealth In suran ce — —
C u rren t Sm oking — —
H eavy D rinking
. . . —
O besity — —
S trok e N ever M arried
— — — — —
Separated/D ivorced
— — — —
W idow ed — — — — —
Age — — — — —
Baseline SR H — — — — —
W hite
— - - — - - - -
E ducation — — — . . . .
H ousehold M em bers — —
L iving C hildren - -
Reproduced with permission o f th e copyright owner. Further reproduction prohibited without permission.
T able 5.26 con tinu er
M odel 1 M odel 2 M odel 3 M odel 4 M odel 5
Stroke Living Siblings — —
Incom e 1.0 0 * 1.0 0 *
H ealth In suran ce
— —
C u rren t Sm oking - - —
H eavy D rinking — - -
O besity
— —
A rth ritis N ever M arried — — — — —
Separated/D ivorced — 0.48*
— — —
W idow ed — - - — — —
Age — — — — —
Baseline SR H 0.82*** 0.82*** 0.82*** 0.81*** 0.81***
W hite — — — —
E ducation — — — — —
H ousehold M em bers — —
Living C hildren 1.07* 1.08*
Living Siblings
— —
Incom e — —
H ealth In surance
- - —
C u rren t Sm oking — —
H eavy D rinking
- - —
O besity 1.77** 1.76**
D epression Never M arried — - - — - -
Separated/D ivorced — — — — —
W idow ed — — — — —
Age — —
1.03*
—
Baseline SR H 0.76*** 0.75*** 0.76*** 0.75*** 0.75***
W hite — — - - — —
E ducation
— — — — —
H ousehold M em bers — -
Reproduced with permission o f th e copyright owner. Further reproduction prohibited without permission.
T able 5.26 continued
M odel 1 M odel 2 M odel 3 M odel 4 M odel 5
D epression Living C h ild ren — —
Living Siblings — —
Incom e — —
H ealth In suran ce — —
C u rren t Sm oking — —
H eavy D rinking — —
O besity - —
***p<.00l, **p<.01, *p<.05
-- not significant
‘D eath occurred betw een 1993 and 1995
Analyses of h eart conditions and broken hip w ere not technically possible because of the sm all sam ple size and skewed sam ple
distrib ution .
T able 5.27. O dds R atios from Logistic R egression of G etting Functioning Difficulties, New Diseases an d C onditions betw een 1993
an d 1998 on M arital States an d O ther V ariables am ong Fem ales: AH EAD ___________ ______________________
M odel 1 M odel 2 M odel 3 M odel 4 M odel 5
ADL Difficulties Never M arried — — — — —
Separated/D ivorced — — — — - -
W idow ed
- - - - — — —
Age 1.1 0*** 1.10*** 1.1 0*** 1.1 0*** 1.1 1***
Baseline SR H 0.59*** 0.59*** 0.59*** 0.60*** 0.60***
W hite
— — — — —
E du cation 0.96* 0.96*
—
0.96*
—
H ousehold M em bers 1.13* 1.12*
L iving C hildren
—
Living Siblings — - -
Incom e —
oo
Reproduced w ith permission o f th e copyright owner. Further reproduction prohibited without permission.
T able 5.27 con tin uer
M odel 1 M odel 2 M odel 3 M odel 4 M odel 5
A D L Difficulties H ealth In surance
— —
C u rren t Sm oking
— —
H eavy D rinking
— —
O besity 1.75***
1 7 4 ***
IA D L D ifficulties Never M arried — — — — —
Separated/D ivorced
— — — — —
W idow ed — — „ — —
Age 1.1 0*** 1.1 0*** 1.1 0*** 1.1 1*** 1.1 1***
Baseline SR H 0.56*** 0.56*** 0.56*** 0.56*** 0.56***
W hite — — — - -
E ducation
— — — — —
H ousehold M em bers
— —
Living C hildren
— —
Living Siblings
— —
Incom e
— —
H ealth In su ran ce — —
C u rren t Sm oking — —
H eavy D rinking
— —
O besity
— —
N agi Difficulties Never M arried
— — — — —
Separated/D ivorced — — - - — —
W idow ed
— — . . — —
Age 1.06*** 1.06*** 1.06*** 1.07*** 1.07***
Baseline SR H 0.62*** 0.62*** 0.62*** 0.63*** 0.63***
W hite
— — — - -
E ducation 0.96* 0.96* 0.96* 0.96*
—
H ousehold M em bers
—
Living C hildren
— —
Living Siblings - —
Reproduced with permission o f th e copyright owner. Further reproduction prohibited without permission.
T able 5.27 continuec
M odel 1 M odel 2 M odel 3 M odel 4 M odel 5
N agi D ifficulties Incom e
— —
H ealth Insu ran ce — —
C u rren t Sm oking
- - —
H eavy D rinking — - -
O besity 1.62** 1.62**
D ea th * Never M arried
. . — - - — —
Separated/D ivorced
— — — - - —
W idow ed — — - - —
Age 1.08*** 1.07*** 1.07*** 1.08*** 1.07***
Baseline SR H 0.57*** 0.57*** 0.58*** 0.57*** 0.58***
W hite
— — _ _ — —
E ducation
— — — - - —
H ousehold M em bers 1.12* 1.11*
Living C hildren
— —
Living Siblings 0.94* 0.94*
Incom e
— - -
H ealth Insu ran ce 0.63*** 0.64***
C u rren t Sm oking 1.45*
—
H eavy D rinking — - -
O besity
— —
H ypertension Never M arried
— — — — —
Separated/D ivorced
— — — — —
W idow ed
— — — — —
Age
— — — — —
Baseline SR H 0.82** 0.81** 0.82** 0.82** 0.82**
W hite
— — — — —
E ducation
— — — - - —
H ousehold M em bers
— - -
L iving C h ildren - —
Reproduced with permission o f th e copyright owner. Further reproduction prohibited without permission.
T able 5.27 contin uer
M odel 1 M odel 2 M odel 3 M odel 4 M odel 5
H ypertension Living Siblings
— —
Incom e
— —
H ealth In su ran ce
— - -
C u rren t Sm oking — - -
H eavy D rinking — - -
O besity
—
D iabetes Never M a rried — — —
Separated/D ivorced - - — —
W idow ed — — —
Age 0.92*** 0.91*** 0.92***
Baseline SR H 0.84* 0.83*
—
W hite — — —
E ducation 0.94* 0.94*
—
H ousehold M em bers
—
N/A N/A
Living C h ild ren
—
Living Siblings
—
Incom e
—
H ealth In surance
—
C u rren t Sm oking
H eavy D rinking
O besity
C ancer Never M arried — — — — —
Separated/D ivorced — — — — —
W idow ed
— — — - -
Age — — — —
Baseline SR H
— — — - - —
W hite
— — — - - —
E ducation
— — — — - -
H ousehold M em bers -
Reproduced with permission o f th e copyright owner. Further reproduction prohibited without permission.
T able 5.27 continuer
M odel 1 M odel 2 M odel 3 M odel 4 M odel 5
C ancer L iving C hild ren
- - —
Living Siblings
— —
Incom e 1. 0 0 *** 1. 0 0 ***
H ealth In su ran ce
- -
C u rren t Sm oking
- - —
H eavy D rinking — —
O besity
— —
L ung C onditions Never M arried
- - — — — —
Separated/D ivorced
— — — — —
W idow ed — — — — —
Age
- - — — — —
Baseline SR H 0.67*** 0 .6 8 *** 0 .6 8 *** 0.69*** 0.69***
W hite 3.44*** 3.63*** 3.61*** 3.26** 3.53**
E ducation — — - - — —
H ousehold M em bers
—
Living C hildren — —
Living Siblings
— —
Incom e
- - —
H ealth In su ran ce
— —
C u rren t Sm oking 3.79*** 3.71***
H eavy D rinking
- - —
O besity
- -
H eart C onditions Never M arried
— — — — —
Separated/D ivorced
— — - - - - - -
W idow ed
— — — —
Age 1.03* 1.03** 1.03* 1.03* 1.03*
Baseline SR H 0.75*** 0.75*** 0.75*** 0.75*** 0.74***
W hite
— — — — —
E ducation - — — - —
Reproduced with permission o f th e copyright owner. Further reproduction prohibited without permission.
T able 5.27 contin u er
M odel I M odel 2 M odel 3 M odel 4 M odel 5
H ea rt C onditions H ousehold M em bers
— —
Living C h ildren 1.08* 1.07*
Living Siblings
— —
Incom e
— —
H ealth Insu ran ce —
C u rren t Sm oking — —
H eavy D rinking
— —
O besity
- - - -
S trok e N ever M arried — — — — —
Separated/D ivorced — — — - - —
W idow ed 1.50* 1.52* 1.48* 1.45* 1.46*
Age
— - - — - - —
Baseline SR H 0.71*** 0.72*** 0.72*** 0.72*** 0.73***
W hite
— — - - — —
E ducation
— — — — —
H ousehold M em bers — —
Living C h ildren — —
Living Siblings
— —
Incom e — - -
H ealth In su ran ce
— —
C u rren t Sm oking
— - -
H eavy D rinking
— - -
O besity — - -
A rth ritis Never M arried
— — — — —
Separated/D ivorced
— — — — —
W idow ed
— — — - - —
Age — — — —
Baseline SR H 0.75*** 0.75*** 0.75*** 0.75*** 0.76***
W hite - - - - - -
Reproduced w ith permission o f th e copyright owner. Further reproduction prohibited without permission.
T able 5.27 continued
M odel 1 M odel 2 M odel 3 M odel 4 M odel 5
A rth ritis E ducation — — — — —
H ousehold M em bers — —
Living C h ildren — —
Living Siblings
—
Incom e — —
H ealth In su ran ce
- - - -
C u rren t Sm oking
- - —
H eavy D rinking —
O besity 1.52** 1.51**
B roken H ip Never M arried — — — « —
Separated/D ivorced — — — — —
W idow ed — — — - - —
Age 1.09*** 1.08*** 1.09*** 1.09*** 1.08***
Baseline SR H — - - — « —
W hite 2 .8 6 * 3.11** 2.92* 2.63* 2.98*
E ducation — - - — — - -
H ousehold M em bers — —
Living C hildren 0.80** 0.80**
Living Siblings
— - -
Incom e
— . .
H ealth In surance
— —
C u rren t Sm oking
—
H eavy D rinking — - -
O besity
— —
D epression Never M arried - — — — —
Separated/D ivorced — — — —
W idow ed — — — — —
Age — — — — —
Baseline SR H 0.78*** 0.78*** 0.79*** 0.79*** 0.79***
Reproduced with permission o f th e copyright owner. Further reproduction prohibited without permission.
T able 5.27 con tin u ec
M odel 1 M odel 2 M odel 3 M odel 4 M odel 5
D epression W hite — — — — —
E ducation 0.95** 0.95** 0.96* 0.95** 0.96*
H ousehold M em bers — —
L iving C hildren - - —
Living Siblings — —
Incom e — —
H ealth In surance — —
C u rren t Sm oking — —
H eavy D rinking — —
O besity — -
***p<.001,**p<.01,*p<.05
-- not significant
‘D eath occurred betw een 1993 and 1995
Table 5.28. O dds R atios from Logistic R egression of G etting Functioning Difficulties, New Diseases an d C onditions betw een 1988
an d 1995 on M arital States an d O ther V ariables: M acA rth ur
M odel 1 M odel 2 M odel 3 M odel 4 M odel 5 M odel 6
N agi D ifficulties Never M arried
— - - — — — —
Separated/D ivorced
— — — — — —
W idowed 1.84** 1.76**
— —
2.04**
—
Fem ale 1.54* 1.5 4 * 1 .9 1 *
— — —
Baseline SR H 0.72* 0.72* — 0.76* 0.74*
—
W hite — — — — —
Education — — — — —
A ge
—
1.0 7 *
— —
Living Children
- - —
Children Nearby — —
Relatives N earby —
oo
00
Reproduced with permission o f th e copyright owner. Further reproduction prohibited without permission.
T able 5.28 continued
M odel 1 M odel 2 M odel 3 M odel 4 M odel 5 M odel 6
N agi D ifficulties F riends N earby — —
In strum ental S u p p ort — —
E m otional S u p p ort — —
Incom e 1.0 0 *
—
C u rren t Sm oking
— —
H eavy D rinking
- - —
O besity — - -
L eisure Activities 0.47** 0.38**
H ypertension N ever M arried — — — - -
Separated/D ivorced — — —
W idow ed — — — —
Fem ale — —
0.50*
—
Baseline S R H — — — —
W hite — — - - —
E ducation — — — —
Age — — —
Living C hildren
C hildren N earby
R elatives N earby N/A N/A
F riends N earby
In strum ental S u p p ort
E m otional S u p p ort
Incom e
—
C u rren t Sm oking
—
H eavy D rinking
—
O besity
—
L eisure Activities
D iabetes Never M arried — — — - - — —
Separated/D ivorced - - - - — — —
Reproduced with permission o f th e copyright owner. Further reproduction prohibited without permission.
T able 5.28 continued
M odel 1 M odel 2 M odel 3 M odel 4 M odel 5 M odel 6
D iabetes W idow ed — — — — - -
Fem ale
— — — — — —
Baseline SR H —
L
— — - - —
W hite — — — - - - - —
E ducation - - — — — —
Age — — — — —
Living C hildren
— —
C hildren N earby
— —
R elatives N earby — —
F riends N earby
— —
In stru m en tal S u p p ort
— —
E m otional S u p p ort — —
Incom e
— —
C u rren t Sm oking
—
H eavy D rinking
— —
O besity
—
L eisure A ctivities — —
C ancer Never M arried — — — —
Separated/D ivorced — - - — —
W idow ed — — - - —
Fem ale 0.51* 0.52* 0.57* 0.55*
Baseline SR H — — —
W hite 2.31* 2.31* 2.23*
—
E ducation — — — —
Age — — —
Living C h ild ren
C hildren N earby
R elatives N earby N/A N/A
F riends N earby
Reproduced with permission o f th e copyright owner. Further reproduction prohibited without permission.
T able 5.28 continued
M odel 1 M odel 2 M odel 3 M odel 4 M odel 5 M odel 6
C ancer In stru m en tal Su p p ort
E m otional S u p p ort
Incom e
C u rren t Sm oking —
H eavy D rinking —
O besity —
L eisure Activities
- -
H ea rt C onditions Never M arried — — — — —
Separated/D ivorced — — — — — —
W idow ed — — — — — - -
Fem ale
— — — — — —
Baseline SR H — — — - - — —
W hite — — — — —
E ducation — — — — - - —
Age
— — — — —
L iving C hildren
- - —
C hildren N earby
- - —
R elatives N earby
— —
F rien ds N earby — —
In stru m en tal S u p p ort
— —
E m otional S up p ort
- - —
Incom e
- - —
C u rren t Sm oking — . .
H eavy D rinking
- - —
O besity
— —
L eisure Activities
— —
S troke Never M arried — — — —
Separated/D ivorced
— - - — —
W idow ed -
-
— —
Reproduced with permission o f th e copyright owner. Further reproduction prohibited without permission.
T able 5.28 continued
M odel 1 M odel 2 M odel 3 M odel 4 M odel 5 M odel 6
S troke Fem ale — — —
Baseline SR H - — — —
W hite — - - — —
E ducation — — — - -
Age — — —
Living C hildren
C hildren N earby
R elatives N earby N/A N/A
F riends N earby
In strum ental S u p p ort
E m otional S u p p ort
Incom e
—
C u rren t Sm oking —
H eavy D rinking —
O besity 2.24*
L eisure Activities —
D epression Never M arried
— — — —
Separated/D ivorced - - — —
W idow ed — — — - -
Fem ale 1.70* 1.69*
—
1.70*
Baseline SR H 0.52*** 0.52*** 0.53*** 0.54***
W hite 1.71* 1.67* 1.71*
—
E ducation — — — —
Age
— — —
Living C hildren
C hildren N earby
R elatives N earby N/A N/A
F riends N earby
In stru m ental S u p p ort
Reproduced with permission o f th e copyright owner. Further reproduction prohibited without permission.
T able 5.28 continued
M odel 1 M odel 2 M odel 3 M odel 4 M odel 5 M odel 6
D epression E m otional S u p p ort
Incom e 1.0 0 *
C u rren t Sm oking
—
H eavy D rinking —
O besity —
L eisure Activities —
***p<.00l, **p<.01, *p<.05
-- n ot significant
A nalysis of broken h ip was not technically possible because of the sm all sam ple size an d skew ed sam ple d istrib u tion .
Figure 5.13. Odds Ratios of Getting ADL Difficulties by Marital
Status and Sex: HRS and AHEAD
194
7 !
6
5
4
3 J
2 J
o l
Never M arried Separated/Divorced
■ HRS Males
OHRS Females
□AHEAD Males
I
E1AHEAD Females
I . j m .
Widowed
7 i
6 -
5
4
3
2 -<
1
0
With A ge Controlled
Never Mamed Separated/Divorced
■ HRS Males
□ HRS Females
□AHEAD Males
□AHEAD Females
Widowed
With A ge, Self-rated Health, Education and R ace
7 -
Controlled
■ HRS Males
□ HRS Females
6 -
□AHEAD Males
5 -
□AHEAD Females
4 -
■ 1
3 -
■
2
1 . 1
0 J
Never Married Separated/Divorced Widowed
With All M echanism s Controlled
■ HRS Males
7 -
□ HRS Females
6 - □AHEAD Males
5 - □AHEAD Females
4 -
3 1
■
2
i .
0 -
Never Mamed Separated/Divorced Widowed
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Figure 5.14. Odds Ratios of Getting IADL Difficulties by Marital
Status and Sex: AHEAD
7
6
5
4
3
2
1
0
□AHEAD Males
□AHEAD Females
.E 2Z L
Never Mamed Separated/Divorced Widowed
With A ge Controlled •
a □ AHEAD Males
-
□AHEAD Females
Never Mamed Separated/Divorced Widowed
7
6
5
4
3
2
1
0
With A ge, Self-rated Health, Education and Race
Controlled
Never Mamed Separal eparated/Divorced
□AHEAD Males
□AHEAD Females
Widowed
7
6
5 H
4
3 -
2 -
1
With All M echanism s Controlled
□AHEAD Males
□AHEAD Females
Never Married Separated/Divorced Widowed
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Figure 5.15 Odds Ratios of Getting Nagi Difficulties by Marital
Status and Sex: HRS, AHEAD and MacArthur______________
■ HRS Males
7
6
5
4
3
2
1
0
□ HRS Females
□AHEAD Males
□AHEAD Females
□ MacArthur
Separated/Divorced Never Mamed Widowed
■ HRS Males
□ HRS Females
□AHEAD Males
□AHEAD Females
□ MacArthur
With A g e C ontrolled
Separated/Divorced Never Mamed Widowed
W ith S elf-rated H ealth, E du cation and R ace
7 -
6
5 -
C ontrolled ■ HRS Males
□ HRS Females
□AHEAD Males
□AHEAD Females
□ MacArthur
4 -
3
2
1 -
0 -
■ t o m
Never Mamed Separated/Divorced Widowed
With All M ech a n ism s C ontrolled
■ HRS Males
7
6
5
4
3
2
1
0
□ HRS Females
□AHEAD Males
□AHEAD Females
m
Widowed Never Mamed Separated/Divorced
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Figure 5.16. Odds Ratios of Dying by Marital Status and Sex:
HRS and AHEAD
1 6
5
4
3 i
2
1
0 J
Never Married Separated/Divorced
■ HRS Males
□ HRS Females
□AHEAD Males
□AHEAD Females
O f 0 0 .
Widowed
7
6
5
4
3
2
1
0
With A ge Controlled
Never Mamed Separated/Divorced
■ HRS Males
□ HRS Females
□AHEAD Males
□AHEAD Females
Widowed
7
6
5
4
3
2
1
0 J
With Age, Self-rated Health, Education and Race
Controlled
■ HRS Males
□ HRS Females
□AHEAD Males
□AHEAD Females
K L
Never M arried Separated/Divorced Widowed
With All M echanism s Controlled
■ HRS Males
□ HRS Females
□AHEAD Males
-
□AHEAD Females
_
a
Never M arried
»■ i
Separated/Divorced Widowed
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Figure 5.17. Odds Ratios of Getting Hypertension by Marital Status
and Sex: HRS, AHEAD and MacArthur
7
6 - \
5
4
3
2
1
0
■ HRS Males
□ HRS Females
□AHEAD Males
□AHEAD Females
□ MacArthur
Never Married Separated/Divorced Widowed
With A ge Controlled
■ HRS Males
□ HRS Females
7 □ AHEAD Males
6 J HAHEAD Females
5 J m □ MacArthur
4
3
2
1
1
0 J Never Married Separated/Divorced Widowed
7 i
6
5
4
3 H
2
1
0
With A ge, Self-rated Health, Education and R ace
Controlled BHRS Males
□ HRS Females
□AHEAD Males
□AHEAD Females
□ MacArthur
1
Never Mamed Separated/Divorced Widowed
With All M echanism s Controlled
■ HRS Males
^ 1 OHRS Females
6 - □ AHEAD Males
5 - HAHEAD Females
^ □ AHEAD Females
3 -
2 -
1 ,------------------- A
0 - I Never Married Separated/Divorced Widowed
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Figure 5.18. Odds Ratios of Getting Diabetes by Marital Status
and Sex: HRS, AHEAD and MacArthur_________________
■ HRS Males
7 - QHRS Females
6 -
□AHEAD Males
□AHEAD Females
5 1 □ MacArthur
4 -
3
2
1 -
0 -
Never Married Separated/Divorced Widowed
7
6
5
4
3
2
1
0
-
With A ge Controlled
■ HRS Males
□ HRS Females
□AHEAD Males
□AHEAD Females
□ MacArthur
-
Never Married Separated/Divorced Widowed
With A ge, Self-rated Health, E ducation and R ace
Controlled ■ HRS Males
( ~
□ HRS Females
6 - I □AHEAD Males
S -
□AHEAD Females
4 -
□ MacArthur
3 -
2
1 .
0
Never Married ^^rated/Divorced Widowed
With All M ech a n ism s C ontrolled
■ HRS Males
OHRS Females
7 □ AHEAD Males
6 □ MacArthur
5
4
3
2
1
0 Never Married Separated/Divorced Widowed
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Figure 5.19. Odds Ratios of Getting Cancer by Marital Status
and Sex: HRS, AHEAD and MacArthur
200
7
6
5
I 4
i 3
I 2
1
■ HRS Males
BHRS Females
□AHEAD Males
□AHEAD Females
□ MacArthur
Never M arried Separated/Divorced Widowed
7
6
5
4
3 - I
2
With A ge Controlled
Never M arried Separated/Divorced
■ HRS Males
HHRS Females
□AHEAD Males
□AHEAD Females
□ MacArthur
Widowed
7
6
5
4
3
2
1
0
With A ge, Self-rated Health, Education and Race
Controlled
■ HRS Males
□ HRS Females
□AHEAD Males
□AHEAD Females
□ MacArthur
Never Mamed Separated/Divorced Widowed
With All M echanism s Controlled
■ HRS Males
□ HRS Females
□AHEAD Males
-
□AHEAD Females
Never Mamed Separated/Divorced Widowed
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Figure 5.20. Odds Ratios of Getting Lung Conditions by Marital
Status and Sex: HRS and AHEAD
201
7 i
6
5
4 -
3 -
l
2 1
1
0
7
6
5
4
3
2
1
0 J
■ HRS Males
□ HRS Females
□AHEAD Males
□AHEAD Females
Never Married Separated/Divorced
With A ge Controlled
■ HRS Males
□ HRS Females
□AHEAD Males
□AHEAD Females
Never Mamed Separated/Divorced Widowed
With A ge, Self-rated Health, E ducation and Race
7
Controlled
■ HRS Males
6
□ HRS Females
5
□AHEAD Males
4 -
□AHEAD Females
3
2
1 . B
0 ■
Never Mamed Separated/Divorced Widowed
With All M echanism s Controlled
■ HRS Males
□ HRS Females
□AHEAD Males
HAHEAD Females
Never Mamed Separated/Divorced Widowed
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Figure 5.21. Odds Ratios of Getting Heart Conditions by Marital
Status and Sex: HRS, AHEAD and MacArthur________________
■ HRS Males
□ HRS Females
□AHEAD Females
□ MacArthur
5
4
3
2
1
Never Married Separated/Divorced Widowed
With A ge Controlled J J f a , e s .
° BH RS Females
-I □AHEAD Females
-
□ MacArthur
-
Never Mamed Separated/Divorced Widowed
7
6
5
4
3
2
1
0
With A ge, Self-rated Health, Education and R ace
Controlled
F=l
■ HRS Males
□ HRS Females
□AHEAD Females
□ MacArthur
Never Mamed Separated/Divorced Widowed
With All M echanism s Controlled
■ HRS Males
□ HRS Females
□AHEAD Females
-
□ MacArthur
Never Mam'ed Separated/Divorced Widowed
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Figure 5.22. Odds Ratios of Getting Stroke by Marital Status
and Sex: HRS, AHEAD and MacArthur
7
6
5
4
3
2
1 7
0 J
Never M arried Separated/Divorced
■ HRS Males
OHRS Females
□AHEAD Males
□AHEAD Females
□ MacArthur
Widowed
7
6 i
5
4
3
2
1
0
With A ge Controlled
Never Mamed Separated/Divorced
■ HRS Males
□ HRS Females
□AHEAD Males
□AHEAD Females
□ MacArthur
Widowed
7
6
5
4
3 -
2
1
0
With A ge, Self-rated Health, Education and Race
Controlled
Never M arried Separated/Divorced
■ HRS Males
OHRS Females
□AHEAD Males
□AHEAD Females
□ MacArthur
Widowed
With All M echanism s Controlled
■ HRS Males
-
□ HRS Females
-
□AHEAD Males
-
HAHEAD Females
-
V 7 7
Never Mamed Separated/Divorced Widowed
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Figure 5.23. Odds Ratios of Getting Arthritis by Marital Status
and Sex: HRS and AHEAD_______________________________
■ HRS Males
7 -| B HRS Females
g j BAHEAD Males
_ BAHEAD Females
5
4
3
2
1
0
204
-
Never Mamed SeparateJyb'ivorced Widowed
With A ge Controlled
7
6
5
4
3 -
2 -
1
0
Never Mamed :ed/^voi Separated/Divorced
■ HRS Males
BHRS Females
□AHEAD Males
BAHEAD Females
Widowed
7
6
5
4
3 - |
2
1
0 -
With A ge, Self-rated Health, Education and Race
Controlled
■ HRS Males
□ HRS Females
□AHEAD Males
BAHEAD Females
Never Mamed Separated/Divorced Widowed
With All M echanism s Controlled
■ HRS Males
-
□ HRS Females
-
□AHEAD Males
-
BAHEAD Females
Never Mamed Separated/Divorced Widowed
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Figure 5.24. Odds Ratios of Getting Broken Bone (HRS) or Hip
(AHEAD) by Marital Status and Sex
7
6 -
■ HRS Males
□ HRS Females
□AHEAD Females
5 -
4 -
3
2 -
1 _
0 -
Never Married Separated/Divorced Widowed
5
4
3
2
1
0
1
With A ge Controlled
■ HRS Males
□ HRS Females
HAHEAD Females
Never Mamed Separated/Divorced Widowed
With A ge, Self-rated Health, Education and R ace
Controlled
'
■ HRS Males
OHRS Females
-
HAHEAD Females
Never Married Separated/Divorced Widowed
i
i
With All M echanism s Controlled
■ HRS Males
-
□ HRS Females
HAHEAD Females
Never Married Separated/Divorced Widowed
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Figure 5.25. Odds Ratios of Getting Depression by Marital Status
and Sex: HRS, AHEAD and MacArthur
206
7 l
6 -
5 -
4
3
2 -
1 -
0 J
■ HRS Males
□ HRS Females
□AHEAD Males
0 AHEAD Females
□ MacArthur
. 1 k
Never Mamed Separated/Divorced Widowed
7
6
5
4 -
3
2 -
1
0 -
With A ge Controlled J J J ? ales,
a □ HRS Females
□AHEAD Males
0 AHEAD Females
□ MacArthur
^ 1 k
Never Mamed Separated/Divorced Widowed
7
6
5 -
4
3 -
2 -
1
With A ge, Self-rated Health, Education and R ace
Controlled b h r s Males
□ HRS Females
□AHEAD Males
0 AHEAD Females
□ MacArthur
L ,
0 -
Never Married Separated/Divorced Widowed
With All M echanism s Controlled
■ HRS Males
7 -i □ HRS Females
6 - DAHEAD Males
5 J 0 AHEAD Females
4
3
2
1
0 - > Never Married Separated/Divorced Widowed
L
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Reproduced w ith permission o f th e copyright owner. Further reproduction prohibited without permission.
Figure 5.26. The Effect of Being Never Married, Separated/Divorced, Widowed on Health Outcomes: HRS, AHEAD
and MacArthur
a. Functioning Problems e. Fatal Health Conditions
ADL Difficulties Hypertension
IADL Difficulties Diabetes
Nagi Difficulties, Cancer
,ung Conditions
+/- Never Married Heart Conditions
Stroke
b. Death Separated/Divorced
Widowed
c. Nonfatal Health Conditions
Arthritis Mental Health Condition
Broken Bone/Hip •►Depression
207
Reproduced with permission o f th e copyright owner. Further reproduction prohibited without permission.
Figure 5.27. Mechanisms through which Being Unmarried Affects Health Outcomes: HRS and AHEAD
a. Functioning Problems
ADL Difficulties
IADL Difficulties
Nagi Difficulties
e. Fatal Health Conditions
Hypertension < — » .
b. Death
Diabetes
Cancer
Lung Conditions
'Heart Conditions /
Unmarried
Stroke f
Economic
esourc
/
Behavior
c. Nonfatal Health Conditions
Arthritis * '
Broken Bone/Hip
Mental Health Condition
\
4 fiepression
" ► Males ► Females
208
209
VI. Widowhood and Health
The previous chapter examined the effect of static marital status on health.
Marital status in itself was found to be related to health status. Then, how about
changes in marital status? As some researchers argue, changes in marital status
may be as or more important than marital status per se in affecting health outcomes.
This chapter, thus, examines how change in marital status affects health,
particularly being widowed, and attempts to examine whether duration of change in
marital status matters. It begins with tables that show percent and number of the
continuously married, the recently widowed, and the longer-term widowed, and
how they differ in the prevalence and incidence of health conditions. Detailed
discussion follows regarding differences in the proposed mechanisms and their
adverse changes for these three marital groups, which will be added in logistic
regression models to examine whether the effect of changes in marital status are
through changes in mechanisms.
The Married, the Recently Widowed and the Longer-term Widowed
Table 6.1 shows the percent and number of the married, the recently
widowed and the longer-term widowed by gender in HRS, AHEAD and
MacArthur. Samples for the widowhood analysis used only those who remained
married, the recently or longer-term widowed. In AHEAD, among 4,485
respondents, about 74% of men and 31% of women remained married over the
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
210
study period. About 6% of both men and women became widowed between 1995
and 1998 (recently widowed), and 20% of men and 63% of women remained
widowed since 1993. In MacArthur, about 77% of men and 30% of women
remained married throughout the study term. While about 2.48% and 4.63% of
men and women, respectively, became widowed between 1988 and 1991, about
21% of men and 65% of women were widowed before 1988 and stayed widowed
until 1991. Very disproportionate percentage of incidence of widowhood was
noticed particularly among males in HRS: 0.96% recently widowers and 1.68%
longer-term widowers. Thus, the HRS sample is not included in the widowhood
analysis.
Table 6.1. Percent and Number of the Married, the Recently Widowed and the Longer-term
Widowed by Sexa: HRS, AHEAD and MacArthur_______________________________
Males Females
HRS
Percent Number Percent Number
Continuously Married 97.36 3245 85.26 3060
Recently Widowed 0.96 32 2.20 79
Longer-term Widowed
AHEAD
1.68 56 12.54 450
Continuously Married 74.36 1256 31.26 874
Recently Widowed 5.62 95 6.19 173
Longer-term Widowed
MacArthur .. __
20.01 338 62.55 1749
Continuously Married 76.67 309 30.12 156
Recently Widowed 2.48 10 4.63 24
Longer-term Widowed 20.84 84 65.25 338
“ The percent was calculated only with the continuously m arried, the recently widowed (widowed
between 1992 and 1994 in H RS, between 1995 and 1998 in AHEAD, and between 1988 and 1991 in
M acArthur), and longer-term widowed (widowed before 1992 in HRS, before 1995 in AHEAD, and
before 1988 in M acArthur).
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
211
Percents of those who had functioning difficulties, diseases and conditions
and those who got functioning difficulties, new diseases and conditions for the
continuously married, the recently widowed and the longer-term widowed are
presented in Tables 6.2 and 6.4 for AHEAD, and 6.3 and 6.5 for MacArthur.
Overall, compared to those who remained married, the widowed had worse health
outcomes. Both recent and longer-term widowers tended to have a higher
prevalence of most health conditions than the continuously married. In AHEAD,
arthritis and broken hip were particularly problematic among widows, and the
prevalence and incidence o f depression was significantly different between
continuously married persons and widowed ones for both sexes with an exception
of longer-term widows (Tables 6.2 and 6.4). Analysis of depression in MacArthur
was not done due to the unavailability of the same depression scale in 1991.2 In
MacArthur, the prevalence of Nagi difficulties was significantly different between
longer-term widowers and continuously married males (Table 6.3) and so was its
incidence between longer-term widowed and the continuously married for both
sexes (Table 6.5).
2 M easure of depression in 1991 is different from that of 1988 and 1995: The CESD is not available
in 1991.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Reproduced w ith permission o f th e copyright owner. Further reproduction prohibited without permission.
Table 6.2. Percent of Those who Had Functioning Difficulties, Diseases and Conditions in 1998 for the Continuously
Married, the Recently Married and the Longer-term Widowed by Sex: AHEAD_______________________________
Males Females
Continuously
Married
Recently
Widowed
Longer-term
Widowed
Continuously
Married
Recently
Widowed
Longer-term
Widowed
ADL Difficulties 29.70 33.80 39.69** 31.84 40.88* 42.95**
IADL Difficulties 21.46 20.43 24.92 22.20 28.92 30.44**
Nagi Difficulties 51.27 54.74 57.69* 64.76 65.90 69.01*
Hypertension 50.57 51.09 49.54 59.08 61.76 60.94
Diabetes 16.28 17.89 14.84 14.60 11.63 16.57
Cancer 20.35 22.83 22.46 14.84 11.63 14.37
Lung Conditions 12.25 8.79 17.01* 9.71 12.14 10.40
Heart Conditions 32.56 36.17 28.44 24.71 27.91 28.19
Stroke 11.17 19.15* 13.64 9.91 9.15 13.66**
Arthritis 57.38 58.06 57.10 65.73 73.84* 72.55**
Broken Hip 0.96 1.05 2.07 1.95 5.78** 4.12**
Depression 17.52 35.79** 29.59** 25.06 44.51** 27.96
**p<.01, *p<.05 (com pared to the continuously m arried)
212
Reproduced with permission o f th e copyright owner. Further reproduction prohibited without permission.
Table 6.3. Percent of Those who Had Functioning Difficulties, Diseases and Conditions in 1991 for the Continuously
Married, the Recently Married and the Longer-term Widowed by Sex: MacArthur_____________________________
Males Females
Continuously
Married
Recently
Widowed
Longer-term
Widowed
Continuously
Married
Recently
Widowed
Longer-term
Widowed
Nagi Difficulties 17.15 10.00 27.38* 29.49 41.67 34.12
Hypertension 45.31 30.00 48.81 46.10 62.50 54.44
Diabetes 18.45 20.00 12.05 23.53 29.17 18.75
Cancer 24.27 10.00 20.24 21.43 12.50 22.39
Heart Conditions 17.48 20.00 22.62 8.44 8.33 10.98
Stroke 7.12 0.00 8.33 3.90 0.00 3.87
Broken Hip 1.62 0.00 3.57 1.95 0.00 3.27
*p<.05 (compared to the continuously married)
to
u >
Reproduced with permission o f th e copyright owner. Further reproduction prohibited without permission.
Table 6.4. Percent of Those who Got Functioning Difficulties, New Diseases and Conditions between 1993 and 1998
for the Continuously Married, the Recently Married and the Longer-term Widowed by Sex: AHEAD_____________
Males Females
Continuously
Married
Recently
Widowed
Longer-term
Widowed
Continuously
Married
Recently
Widowed
Longer-term
Widowed
ADL Difficulties 23.77 27.42 31.68* 24.78 30.48 33.45**
1ADL Difficulties 17.43 19.05 18.97 16.32 21.53 21.99**
Nagi Difficulties 20.90 23.75 27.76* 31.92 29.36 40.00**
Hypertension 15.83 16.98 17.80 21.23 29.21 20.89
Diabetes 4.82 10.34* 5.02 4.63 3.18 5.36
Cancer 9.09 11.39 8.51 4.92 3.82 3.66
Lung Condition 4.63 6.74 4.88 4.10 4.46 4.33
Heart Disease 17.84 21.05 16.17 11.37 12.88 12.37
Stroke 6.34 7.41 7.54 5.62 5.81 8.05*
Arthritis 48.46 47.22 45.06 54.73 65.60* 60.59*
Broken Hip 0.99 1.06 1.83 1.66 5.36* 3.80**
Depression 14.49 36.14** 21.16* 17.84 39.86** 18.46
**p<.01, *p<.05 (com pared to the continuously m arried )
Reproduced w ith permission o f th e copyright owner. Further reproduction prohibited without permission.
Table 6.5. Percent o f Those who Got Functioning Difficulties, New Diseases and Conditions between 1988 and
1991 for the Continuously Married, the Recently Married and the Longer-term Widowed by Sex: MacArthur
Males Females
Continuously
Married
Recently
Widowed
Longer-term
Widowed
Continuously
Married
Recently
Widowed
Longer-term
Widowed
Nagi Difficulties 14.55 0.00 27.03* 20.54 27.78 32.50*
Hypertension 2.91 0.00 4.76 2.60 0.00 2.66
Diabetes 5.18 0.00 6.02 7.19 16.67 8.28
Cancer 10.36 10.00 10.71 3.90 0.00 5.64
Heart Conditions 4.85 0.00 3.57 2.60 4.17 4.14
Stroke 4.21 0.00 3.57 1.95 0.00 3.25
Broken Hip 0.32 0.00 1.19 0.00 0.00 0.89
*p<.05 (compared to the continuously married)
215
Reproduced with permission o f th e copyright owner. Further reproduction prohibited without permission.
Table 6.6. Mean and Percent o f Proposed Mechanisms (1998) for the Continuously Married, the Recently
Married and the Longer-term Widowed by Sex: AHEAD____________________________________________
Males Females
Continuously
Married
Recently
Widowed
Longer-term
Widowed
Continuously
Married
Recently
Widowed
Longer-term
Widowed
Mean Age
Social Support
80.63 82.82** 83.00** 79.76 81.57** 83.48**
Mean # Household Members 0.22 0.37 0.45** 0.16 0.27* 0.46**
Mean # Living Children 3.04 2.83 2.97 2.77 2.93 2.65
Mean # Living Siblings
Economic Resources
1.97 1.69 1.82 1.87 1.73 1.70+
Mean Income (SS, SSI, Vet
Benefits)
17475 11269** 11417** 16642 11025** 10379**
% Health Insurance
Health Behaviors
73.41 72.63 69.82 70.59 71.10 68.84
% Current Smoking 6.77 6.32 8.58 4.46 8.09* 6.00
% Obesity 10.75 6.32 11.54 15.22 9.83+ 15.15
% Heavy Drinking 5.30 6.49 5.14 1.24 0.69 0.80
% Wearing Seatbelt 83.13 84.04 78.51+ 89.33 86.13 85.86*
**p<.01, *p<.05, +p<.10 (com pared to the continuously m arried)
216
Reproduced with permission o f th e copyright owner. Further reproduction prohibited without permission.
Table 6.7. Mean and Percent of Proposed Mechanisms (1991) for the Continuously Married, the Recently
Married and the Longer-term Widowed by Sex: MacArthur____________________________________
Males Females
Continuously
Married
Recently
Widowed
Longer-term
Widowed
Continuousl
y Married
Recently
Widowed
Longer-term
Widowed
Mean Age 76.86 78.90* 77.43 76.88 77.33 77.65**
Social Support
Mean # Living Children 2.40 2.40 2.92+ 2.36 2.21 2.41
% Children Nearby 72.99 100.00+ 76.62 66.42 86.96+ 73.33
Mean # Relatives Nearby 2.92 3.33 2.38 2.26 1.73 2.32
Mean # Friends Nearby 6.11 17.14 6.46 4.88 4.41 4.48
Instrumental Support 1.90 1.85 1.69* 1.78 1.80 1.84
Emotional Support 2.51 2.73 2.42 2.51 2.81** 2.66**
Economic Resources
Mean Income 23143 19700 14267** 17242 10609** 11192**
Health Behaviors
% Current Smoking 12.34 10.00 18.07 11.04 16.67 10.09
% Obesity 12.30 20.00 12.05 22.08 16.67 19.35
% Heavy Drinking 8.37 22.22 12.50 0.71 0.00 2.31
% Leisure Activities 82.74 90.00 85.71 81.17 79.17 81.07
**p<.01, *p<.05, +p<.10 (compared to the continuously married)
217
218
Tables 6.6 and 6.7 show the description of the proposed mechanisms in
AHEAD and MacArthur. In AHEAD (Table 6.6), the longer-term widowed were
older than the recently widowed and continuously married persons. Longer-term
widowed had more people in the same household than other groups. Recent and
longer-term widowed elderly had a significantly lower level of per capita income
than the continuously married. Health behaviors were not significantly different
for these three groups. It is probable that even though people may engage in
behaviors with high health risks such as smoking and heavy drinking in order to
relieve stress resulting from becoming widowed, they may rely more on family
relationships and support network established while married (Verbrugge, 1979). In
fact, the widowed had more people in the same household than the continuously
married in AHEAD and MacArthur, and more widow(er)s, either recent or longer-
term, in MacArthur, had children who lived nearby than the continuously married
for both sexes (Table 6.7).
Tables 6.8 to 6.13 present means and percents of mechanisms through
which marital status affects health for the continuously married, the recently
widowed and the longer-term widowed between 1993 or 1995 and 1998 in
AHEAD and between 1988 and 1991 in MacArthur. Comparison between 1993
and 1995 columns in Tables 6.8 and 6.9 shows that among the AHEAD sample,
recent widowhood did not necessarily result in negative changes in proposed
mechanisms. Those who became widowed between 1995 and 1998 had more
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
219
people in the same household since they became widowed, and percents of those
who smoked, who were obese, or who drank heavily decreased, and more people
wore a seatbelt after they became widowed (Tables 6.8 and 6.9). Similarly,
recently widowed elderly people increased a healthy behavior (participation in
leisure activity) and decreased risky behaviors (smoking, obesity, heavy drinking)
since they became widowed among the MacArthur sample (Tables 6.10 and 6.11).
Tables 6.12 and 6.13 show changes in proposed mechanisms. For
household members, living siblings, relatives and friends nearby, if a respondent
had smaller number in 1998 (AHEAD) or 1991 (MacArthur) than in 1993
(AHEAD) or 1988 (MacArthur), he/she is the one who experienced adverse change
(decreased number). For income, lower income in 1995 (AHEAD) or 1991
(MacArthur) than at baseline indicates adverse change in income. Decreased
instrumental and emotional support represent those whose score got lower between
1988 and 1991. Changes in other mechanisms were constructed such that a person
who became a smoker, lost insurance supplements, became obese, became a heavy
drinker, did not wear a seatbelt any more, did not have children nearby any more,
or did not participate in leisure activity any more since baseline years, represented
one who experienced adverse change in each mechanism.
The difference between the continuously married and the longer-term
widowed was more noticeable than the difference between the married and the
recently widowed in AHEAD: More longer-term widowed persons had decreased
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
220
social support such as number of household members and living siblings relative to
the continuously married while percent of the recently widowed whose social
support decreased was similar to that of the continuously married among the
AHEAD sample (Table 6.12) except in death of children, which is probably
because the longer-term widowed were older than the recently widowed. On the
other hand, Table 6.13 shows that more recently widowers experienced a decrease
in number of relatives and friends who live nearby among the MacArthur sample.
Recent widowhood was related to significant reduction in income in AHEAD
(Tables 6.8 and 6.9) and in MacArthur (Tables 6.10 and 6.11) while longer-term
widowhood was related to increase in income both in AHEAD and MacArthur. As
described above, those who became widowed between 1995 and 1998 in AHEAD
increased the number of additional household members. This supports the findings
in previous studies that showed no evidence of reduced social support following
widowhood.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Reproduced w ith permission o f th e copyright owner. Further reproduction prohibited without permission.
Table 6.8. Mean and Percent of Mechanisms for the Continuously Married, the Recently Married and the
Longer-term Widowed in 1995 and 1998 among Males: AHEAD_______________________________
Continuously Married Recently Widowed Longer-term Widowed
1995 1998 1995 1998 1995 1998
Social Support
Mean # Household Members8 0.24 0.22 0.14 0.37 0.44 0.45
Mean # Living Siblings8 2.14 1.97 2.12 1.69 2.22 2.04
Economic Resources
Mean Income (SS, SSI, Vet 14935 17475 14392 11269 10590 11417
Benefits)
% Health Insurance 77.55 73.41 83.16 72.63 75.44 69.82
Health Behaviors
% Current Smoking 8.12 6.77 7.37 6.32 10.65 8.58
% Obesity 11.23 10.75 9.47 6.32 12.13 11.54
% Heavy Drinking 6.00 5.30 8.51 6.49 5.03 5.14
% Wearing Seatbelt 81.32 83.13 82.11 84.04 75.07 78.51
a1993 d a ta w ere used instead of 1995 for social su p port variables, w hich a re not available in 1995.
221
Reproduced with permission o f th e copyright owner. Further reproduction prohibited without permission.
Table 6.9. Mean and Percent of Mechanisms for the Continuously Married, the Recently Widowed and Longer-term
Widowed in 1995 and 1998 among Females: AHEAD______________________________________________
Continuously Married______ Recently Widowed______ Longer-term Widowed
1995 1998 1995 1998 1995 1998
Social Support
Mean # Household Members* 0.18 0.16 0.22 0.27 0.45 0.46
Mean # Living Siblings* 2.19 1.87 2.13 1.73 2.06 1.70
Economic Resources
Mean Income (SS, SSI, Vet 14880 16642 14497 11025 9080 10379
Benefits)
% Health Insurance 76.89 70.59 78.61 71.10 75.41 68.84
Health Behaviors
% Current Smoking 5.61 4.46 9.25 8.09 7.32 6.00
% Obesity 16.36 15.22 14.45 9.83 15.04 15.15
% Heavy Drinking 1.37 1.24 1.73 0.69 1.09 0.80
% Wearing Seatbelt 89.22 89.33 83.24 86.13 84.17 85.86
*1993 d ata w ere used instead of 1995 for social su p port variables, w hich are not available in 1995.
222
Reproduced w ith permission o f th e copyright owner. Further reproduction prohibited without permission.
Table 6.10. Mean and Percent of Mechanisms for the Continuously Married, the Recently Married and the
Longer-term Widowed in 1988 and 1991 among Males: MacArthur________________________________
Continuously Married_______Recently Widowed_______Longer-term Widowed
Social Support
1988 1991 1988 1991 1988 1991
Mean # Living Children 2.43 2.40 2.40 2.40 3.04 2.92
% Children Nearby 77.14 72.99 100.00 100.00 85.71 76.62
Mean # Relatives Nearby 3.46 2.92 8.50 3.33 3.65 2.38
Mean # Friends Nearby 7.02 6.11 9.71 17.14 7.04 6.46
Instrumental Support 1.72 1.90 1.67 1.85 1.41 1.69
Emotional Support
Economic Resources
2.40 2.51 2.23 2.73 2.35 2.42
Mean Income
Health Behaviors
21215 23143 22625 19700 13259 14267
% Current Smoking 13.45 12.34 10.00 10.00 19.23 18.07
% Obesity 12.34 12.30 20.00 20.00 12.20 12.05
% Heavy Drinking 13.28 8.37 20.00 22.22 10.53 12.50
% Leisure Activities 84.14 82.74 90.00 90.00 80.95 85.71
to
to
u>
Reproduced w ith permission o f th e copyright owner. Further reproduction prohibited without permission.
Table 6.11. Mean and Percent of Mechanisms for the Continuously Married, the Recently Married and the
Longer-term Widowed in 1988 and 1991 among Females: MacArthur______________________________
Continuously Married_______Recently Widowed______ Longer-term Widowed
Social Support
1988 1991 1988 1991 1988 1991
Mean # Living Children 2.42 2.36 2.21 2.21 2.48 2.41
% Children Nearby 71.94 66.42 82.61 86.96 75.79 73.33
Mean # Relatives Nearby 2.97 2.26 2.53 1.73 2.86 2.32
Mean # Friends Nearby 3.95 4.88 4.40 4.14 5.23 4.48
Instrumental Support 1.65 1.78 1.77 1.80 1.58 1.84
Emotional Support
Economic Resources
2.47 2.51 2.49 2.81 2.56 2.66
Mean Income
Health Behaviors
18040 17242 12804 10608 10506 11192
% Current Smoking 12.67 11.04 17.39 16.67 12.42 10.09
% Obesity 21.57 22.08 20.83 16.67 19.82 19.35
% Heavy Drinking 2.86 0.71 0.00 0.00 3.90 2.31
% Leisure Activities 82.47 81.17 66.67 79.17 78.40 81.07
224
Reproduced with permission o f th e copyright owner. Further reproduction prohibited without permission.
Table 6.12. Percent with Adverse Change in Proposed Mechanisms for the Continuously Married, the Recently Widowed
and the Longer-term Widowed between 1993 and 1995 and between 1995 and 1998: AHEAD____________________
Males Females
Continuously
M arried
R ecently
W idow ed
Longer-term
W idow ed
Continuously
M arried
R ecently
W idow ed
Longer-term
W idow ed
Between 1993 and 1998
Decreased # Household Members 5.8 3.2 11.8** 5.0 5.8 10.6**
Children Died 2.55 3.16 1.78 2.75 2.89 4.41*
Decreased # Living Siblings 25.3 34.0 36.2** 26.9 32.9 31.9**
Between 1995 and 1998
Decreased Income 23.65 85.71** 15.28** 16.88 83.01** 16.14
Decreased Health Insurance 13.0 15.8 15.1 15.1 14.5 16.6
Increased Current Smoking 0.7 1 .1 0.6 0.2 0.0 0.6
Increased Obesity 2.4 1 .1 2.4 3.0 0.6 3.6
Increased Heavy Drinking 2.0 1.3 3.1 0.4 0.0 0.5
Decreased Wearing Seatbelt 4.64 7.45 5.69 4.60 6.94 6.53*
♦*p<.01, *p<.05 (com pared to th e continuously m arried )
225
Reproduced w ith permission o f th e copyright owner. Further reproduction prohibited without permission.
Table 6.13. Percent with Adverse Change in Proposed Mechanisms for the Continuously Married, the Recently
Widowed and the Longer-term Widowed between 1988 and 1991 in MacArthur______________________
Males Females
C ontinuously
M arried
R ecently
W idow ed
L onger-term
W idow ed
C ontinuously
M a rried
R ecently
W idow ed
L on ger-term
W idow ed
Decreased # Living Children 4.56 0.00 8.33 4.55 0.00 7.40
Decreased Children Nearby 8.79 0.00 9.09 7.30 0.00 6.36
Decreased # Relatives Nearby 38.86 80.00 49.18 40.32 31.58 35.66
Decreased # Friends Nearby 46.33 57.14 48.44 45.69 30.00 45.74
Decreased Instrumental
Support
30.54 30.00 27.71 39.61 37.50 27.74**
Decreased Emotional Support 33.44 20.00 32.53 33.12 4.17** 25.76
Decreased Income 23.02 37.50 20.55 34.23 68.18** 12.46**
Increased Current Smoking 1.38 0.00 1.30 1.32 4.35 1.56
Increased Obesity 2.92 0.00 2.47 5.88 4.17 3.68
Increased Heavy Drinking 3.75 11.11 4.55 0.00 0.00 0.35
Decreased Leisure Activity
Participation
9.45 10.00 5.95 10.26 8.33 8.88
**p<.01 (com pared to the continuously m arried )
226
227
Results from Logistic Regressions
This study hypothesized that health status would be affected by changes in
marital status through changes in mechanisms. If this were true, changes in
independent variables are likely to be highly correlated each other, which may
undifferentiate the results by those variables. Correlation matrix was developed to
check the possible multicollinearity among independent variables indicating
changes in mechanisms. Correlation matrix (not shown) showed that there were no
correlations of .20 or more among independent variables, rejecting the possibility
of multicollinearity.
Odds ratios from logistic regressions of the incidence of health conditions
on being recently and longer-term widowed and a number of other variables in
AHEAD are shown in Table 6.14.3 The first model included being recently
widowed, being longer-term widowed and gender, the second model added age into
the first model. Model 3 had interaction terms between gender and being recently
widowed or longer-term widowed, and finally all mechanisms were added in model
4. The longer-term widowed in AHEAD were more likely than the continuously
married to get stroke in model 2. Recent widowhood was related to a higher
probability of getting depression since baseline. That is, the recently widowed
were clearly more likely to develop depression than the continuously married while
that was not the case for the longer-term widowed. Also, model 4 in Table 6.14
3 Because of the small num ber and the skewed distribution of people in each category of m arital
status, logistic regression analysis was not done for the M acArthur sample.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
228
showed that changes in the proposed mechanisms were not associated with the
relationship of widowhood to health outcomes.
In terms of the effect of widowhood on health, age seems to play a role: The
effect of widowhood on physical conditions disappeared with age or the interaction
term of duration of widowhood and gender controlled (e.g. the effect of being
longer-term widowed on the incidence of ADL, IADL and Nagi difficulties, broken
hip and stroke). But the effect o f widowhood on depression remained with age,
interaction and/or all proposed mechanisms controlled. It appears that age is a
critical factor in the effect of widowhood on physical health conditions, but not on
mental health. With age considered, the effect of being recently widowed on
getting depression was almost similar to the effect without controlling age.
Odds ratios of the presence of health conditions among the continuously
married, the recently widowed and the longer-term widowed in 1998 are presented
in Table 6.15. When age was controlled (model 2), the effect of being longer-term
widowed on the prevalence of ADL difficulties and stroke remained significant
while there was no such effect of being recently widowed on the prevalence of any
physical health conditions. Recent widowhood increased the odds of having
depression a little less than 3 times and longer-term widowhood increased about
1.3-1.8 times compared to continuous marriage. When all mechanism changes
were included in model (model 4), the effect of recent widowhood on having
depression weakened (OR=2.01) but remained significant, but that of longer-term
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
229
widowhood disappeared. Baseline income modified the effect o f recent or longer-
term widowhood on the prevalence of depression. Also, even when age, gender
and all mechanisms including smoking were controlled, the longer-term widowed
were more likely to have lung conditions than the continuously married (model 4).
The effect of mechanism changes on the relationship between widowhood
and the prevalence of health conditions was not as much as anticipated. Baseline
income was the only mechanism found in the relation: Lower income explained
part of the adverse effect of being longer-term widowed on having depression.
Interactions between gender and widowhood states were introduced in
model 3, but did not show any significant effect on the prevalence of health
conditions with the exception of interactions of female and longer-term widowhood
on heart conditions and depression: Interaction term between gender and longer-
term widowhood state was significantly related such that the longer-term widowed
were less likely than the continuously married to have depression when they were
women. In sum, comparison of the results on the prevalence and incidence of
depression shows that the recently widowed, compared to those who remained
married, were more likely to have and develop depression while the longer-term
widowed were not likely to get depression, but sustained depression that occurred
previously.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
230
Table 6.14. Odds Ratios from Logistic Regression o f Getting Functioning Difficulties, New Diseases and
Conditions between 1993 and 1998 on Being Recently and Longer-term Widowed and Other Variables: AHEAD
■ 1 .............. ..
Model 1 Model 2 Model 3 Model 4
Continuously married as reference
ADL Difficulties Recently Widowed
— — — —
Longer-term Widowed 1.52*** — — —
Female
— — — —
Age 1.08*** 1.08*** 1.10***
Female*Recently Widowed
— —
Femaie*Longer-term Widowed
— —
Decreased # Household Members
—
Children Died
—
Decreased # Living Siblings
—
Income in 1993
—
Decreased Income
—
Decreased Health Insurance
—
Increased Current Smoking —
Increased Obesity
—
Increased Heavy Drinking
—
Decreased Wearing Seatbelt —
IADL Difficulties Recently Widowed
— — — —
Longer-term Widowed 1.41*** — — —
Female
— — — —
Age 1.09*** 1.09*** 1.08***
FemaIe*Recently Widowed
— —
Female*Longer-term Widowed
— —
Decreased # Household Members
—
Children Died
—
Decreased # Living Siblings
—
Income in 1993 1.00**
Decreased Income
—
Decreased Health Insurance _
Increased Current Smoking
—
Increased Obesity
—
Increased Heavy Drinking _
Decreased Wearing Seatbelt 2.11**
Nagi Difficulties Recently Widowed
— — — _
Longer-term Widowed 1.52***
— — —
Female 1.65*** 1.70*** 1.70*** 2.52**
Age 1.07*** 1.07*** 1.07***
Female*Recently Widowed — —
Female* Longer-term Widowed
— —
Decreased # Household Members
—
Children Died
—
Decreased # Living Siblings
—
Income in 1993
—
Decreased Income
—
Decreased Health Insurance 0.65*
Increased Current Smoking
—
Increased Obesity
—
Increased Heavy Drinking
—
Decreased Wearing Seatbelt —
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
231
Table 6.14 continued
Model I Model 2 Model 3 Model 4
Hypertension Recently Widowed
— — — —
Longer-term Widowed — — _ —
Female 1.38** 1.37** 1.45*
—
Age
— — —
Female*RccentIy Widowed
— —
Female*Longer-term Widowed
— —
Decreased # Household Members —
Children Died —
Decreased # Living Siblings —
Income in 1993 —
Decreased Income —
Decreased Health Insurance
—
Increased Current Smoking
—
Increased Obesity
—
Increased Heavy Drinking
—
Decreased Wearing Seatbelt
—
Diabetes Recently Widowed — — 2.29* —
Longer-term Widowed — — — —
Female — — — —
Age 0.95** 0.95** 0.94**
Female*RecentIy Widowed 0.26* —
Female* Longer-term Widowed
— —
Decreased # Household Members
—
Children Died
—
Decreased # Living Siblings —
Income in 1993 —
Decreased Income 1.68*
Decreased Health Insurance
_
Increased Current Smoking —
Increased Obesity _
Increased Heavy Drinking
—
Decreased Wearing Seatbelt
—
Cancer Recently Widowed — — — —
Longer-term Widowed
— — — —
Female 0.46*** 0.46*** 0.54***
—
Age
— — —
FemaIe*Recently Widowed
— ___
Female*Longer-term Widowed
— —
Decreased # Household Members 2.07*
Children Died
—
Decreased # Living Siblings
—
Income in 1993 —
Decreased Income —
Decreased Health Insurance 0.24**
Increased Current Smoking ~
Increased Obesity
—
Increased Heavy Drinking —
Decreased Wearing Seatbelt
—
Lung Conditions Recently Widowed — — — —
Longer-term Widowed — — — —
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
232
Table 6.14 continued
Model 1 Model 2 Model 3 Model 4
Lung Conditions Female — — — —
Age
— — —
Female*Recently Widowed — —
Female*Longer-term Widowed
— —
Decreased # Household Members —
Children Died —
Decreased # Living Siblings
—
Income in 1993 1.00*
Decreased Income —
Decreased Health Insurance 0.28*
Increased Current Smoking 17.33***
Increased Obesity
—
Increased Heavy Drinking —
Decreased Wearing Seatbelt —
Heart Conditions Recently Widowed — — — —
Longer-term Widowed — — — —
Female 0.65*** 0.66*** 0.61** —
Age 1.04*** 1.04*** 1.05***
Female* Recently Widowed — —
Female*Longer-term Widowed — —
Decreased # Household Members
—
Children Died 2.02***
Decreased # Living Siblings
—
Income in 1993 —
Decreased Income —
Decreased Health Insurance 0.52*
Increased Current Smoking
—
Increased Obesity 2.62**
Increased Heavy Drinking
—
Decreased Wearing Seatbelt
—
Stroke Recently Widowed — — — —
Longer-term Widowed 1.48** 1.38*
— —
Female — — — —
Age — — —
Female*RecentIy Widowed — —
Female* Longer-term Widowed
— —
Decreased # Household Members
—
Children Died ___
Decreased # Living Siblings 1.43*
Income in 1993 _
Decreased Income —
Decreased Health Insurance —
Increased Current Smoking —
Increased Obesity —
Increased Heavy Drinking —
Decreased Wearing Seatbelt
___
Arthritis Recently Widowed
___ — — ___
Longer-term Widowed
— — — ___
Female 1.47*** 1.48*** 1.30** 1.94**
Age 1.02* 1.02*
- —
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
233
Table 6.14 continued
Model I Model 2 Model 3 Model 4
Arthritis Female*Recently Widowed 1.94* —
FemaIe*Longer-term Widowed
— —
Decreased # Household Members —
Children Died —
Decreased # Living Siblings
Income in 1993 1.00*
Decreased Income
—
Decreased Health Insurance
—
Increased Current Smoking
Increased Obesity —
Increased Heavy Drinking —
Decreased Wearing Seatbelt —
Broken Hip Recently Widowed 2.32*
— — —
Longer-term Widowed 2.41***
— — —
Female 2.24** 2.31**
— —
Age 1.09*** 1.09*** 1.09***
Female*Recently Widowed
— —
Female* Longer-term Widowed — —
Decreased # Household Members —
Children Died
—
Decreased # Living Siblings —
Income in 1993 —
Decreased Income —
Decreased Health Insurance
—
Increased Current Smoking —
Increased Obesity
—
Increased Heavy Drinking
—
Decreased Wearing Seatbelt
—
Depression Recently Widowed 3.41*** 3.39*** 3.39*** 2.26*
Longer-term Widowed
— — —
Female
— —
1.31*
Age
— —
Female*Recently Widowed
— —
Female*Longer-term Widowed — —
Decreased # Household Members
___
Children Died
___
Decreased # Living Siblings
___
Income in 1993 —
Decreased Income _ _
Decreased Health Insurance —
Increased Current Smoking —
Increased Obesity
—
Increased Heavy Drinking
—
Decreased Wearing Seatbelt | —
***p<0.001, **p<0.01, *p<0.05
— not significant
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
234
Table 6.15. Odds Ratios from Logistic Regression o f Having Functioning Difficulties, Diseases and Conditions
(1998) on Being Recently and Longer-term Widowed and Other Variables: A H E A D ___________ ___________
1
Model 1 Model 2 Model 3 Model 4
Continuously married as reference
ADL Difficulties Recently Widowed
— — — —
Longer-term Widowed 1.62*** 1.21* — 1.80*
Female
— 1.20* — —
Age 1.08*** 1.08*** 1.09***
Female*Recently Widowed — —
Female*Longer-term Widowed
— —
Decreased # Household Members —
Children Died
—
Decreased ft Living Siblings
—
Income in 1993 1.00**
Decreased Income —
Decreased Health Insurance —
Increased Current Smoking —
Increased Obesity 2.19**
Increased Heavy Drinking
—
Decreased Wearing Seatbelt
1ADL Difficulties Recently Widowed
— — — —
Longer-term Widowed 1.52*** — — 1.75*
Female — 1.19*
—
1.89*
Age 1.11*** 1.11*** 1.09***
Female*Recently Widowed
— —
Female*Longer-term Widowed — —
Decreased # Household Members
—
Children Died
—
Decreased # Living Siblings —
Income in 1993 1.00***
Decreased Income
—
Decreased Health Insurance —
Increased Current Smoking
—
Increased Obesity
—
Increased Heavy Drinking —
Decreased Wearing Seatbelt 2.03***
Nagi Difficulties Recently Widowed
— — —
Longer-term Widowed 1.27***
— —
1.50*
Female 1.72*** 1.76*** 1.86*** 2.20***
Age 1.05*** 1.05*** 1.05***
FemaIe*RecentIy Widowed
— —
Female*Longer-term Widowed — —
Decreased # Household Members
—
Children Died
Decreased # Living Siblings —
Income in 1993
—
Decreased Income
—
Decreased Health Insurance 0.75*
Increased Current Smoking
—
Increased Obesity 2.35**
Increased Heavy Drinking
—
Decreased Wearing Seatbelt —
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
235
Table 6.IS continued
Model I Model 2 Model 3 Model 4
Hypertension Recently Widowed — — — —
Longer-term Widowed
— — — —
Female 1.46*** 1.46*** 1.41*** —
Age
— — —
Female*Recently Widowed
— —
Female*Longer-term Widowed
—
Decreased # Household Members —
Children Died
—
Decreased # Living Siblings
—
Income in 1993 1.00**
Decreased Income
—
Decreased Health Insurance
—
Increased Current Smoking
—
Increased Obesity —
Increased Heavy Drinking —
Decreased Wearing Seatbelt
—
Diabetes Recently Widowed — — — —
Longer-term Widowed — — — —
Female — — — —
Age 0.97*** 0.96*** 0.95***
Female*Recently Widowed — —
Female*Longer-term Widowed
— —
Decreased # Household Members 1.53*
Children Died —
Decreased # Living Siblings
—
Income in 1993 1.00**
Decreased Income
—
Decreased Health Insurance
—
Increased Current Smoking
—
Increased Obesity 2.58**
Increased Heavy Drinking
—
Decreased Wearing Seatbelt
—
Cancer Recently Widowed
— — — —
Longer-term Widowed
— — — —
Female 0.64*** 0.65*** 0.71**
—
Age — — —
Female*Recently Widowed
— —
Female*Longer-term Widowed —
0.49*
Decreased # Household Members 1.51*
Children Died —
Decreased # Living Siblings
—
Income in 1993 1.00*
Decreased Income
—
Decreased Health Insurance
—
Increased Current Smoking
—
Increased Obesity
—
Increased Heavy Drinking —
Decreased Wearing Seatbelt
—
Lung Conditions Recently Widowed
— — — _ _
Longer-term Widowed 1.26* 1.38** 1.73** 2.70***
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
236
Table 6.IS continued
Model I Model 2 Model 3 Model 4
Lung Conditions Female 0.69*** 0.69*** 0.75*
Age 0.97** 0.98** 0.97*
Female*Recently Widowed
— —
Female* Longer-term Widowed
— —
Decreased # Household Members
_
Children Died —
Decreased # Living Siblings
_
Income in 1993
—
Decreased Income —
Decreased Health Insurance
Increased Current Smoking 9.66***
Increased Obesity
—
Increased Heavy Drinking
—
Decreased Wearing Seatbelt
—
Heart Conditions Recently Widowed _ — — —
Longer-term Widowed
— — — —
Female 0.78** 0.79** 0.70***
—
Age 1.04*** 1.04*** 1.04***
Female*Recently Widowed
— —
Female*Longer-term Widowed 1.40*
_
Decreased # Household Members
_
Children Died
—
Decreased # Living Siblings
—
Income in 1993 —
Decreased Income
—
Decreased Health Insurance
—
Increased Current Smoking
—
Increased Obesity 1.96*
Increased Heavy Drinking —
Decreased Wearing Seatbelt
Stroke Recently Widowed — — — —
Longer-term Widowed 1.48*** 1.35** — —
Female
— — —
Age 1.02** 1.02**
—
Female*Recently Widowed
—
Female*Longer-term Widowed
— —
Decreased # Household Members
_ _
Children Died
_ _
Decreased # Living Siblings
_ _
Income in 1993
—
Decreased Income
_ _
Decreased Health Insurance
—
Increased Current Smoking
—
Increased Obesity 2.37**
Increased Heavy Drinking
_ _
Decreased Wearing Seatbelt
—
Arthritis Recently Widowed
— — — —
Longer-term Widowed 1.20* — — —
Female 1.59*** 1.60*** 1.44*** 2.05**
Age 1.02** 1.02** 1.03**
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
237
Table 6.15 continued
Model 1 Model 2 Model 3 Model 4
Arthritis Female*RecentIy Widowed — —
Female*Longer-term Widowed — —
Decreased # Household Members —
Children Died —
Decreased # Living Siblings
—
Income in 1993 1.00***
Decreased Income
—
Decreased Health Insurance
—
Increased Current Smoking —
Increased Obesity
—
Increased Heavy Drinking
—
Decreased Wearing Seatbelt
—
Broken Hip Recently Widowed 2.32*
— — —
Longer-term Widowed 2.41***
— — —
Female 2.34** 2.41** 2.36* —
Age 1.09*** 1.09*** 1.10***
Female’ Recently Widowed — —
Female*Longer-term Widowed — —
Decreased # Household Members 2.21*
Children Died
—
Decreased # Living Siblings
—
Income in 1993
—
Decreased Income —
Decreased Health Insurance
—
Increased Current Smoking —
Increased Obesity
—
Increased Heavy Drinking —
Decreased Wearing Seatbelt
Depression Recently Widowed 2.82*** 2.81*** 2.65*** 2.01*
Longer-term Widowed 1.32*** 1.31** 1.82***
—
Female 1.31*** 1.32*** 1.53*** 1.68*
Age
— —
Female* Recently Widowed
— —
Female* Longer-term Widowed 0.63**
—
Decreased # Household Members —
Children Died
—
Decreased # Living Siblings
—
Income in 1993 1.00**
Decreased Income
—
Decreased Health Insurance
—
Increased Current Smoking —
Increased Obesity —
Increased Heavy Drinking —
Decreased Wearing Seatbelt
***p<0.001, **p<0.01, *p<0.05
— not significant
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
238
So far, analyses were done to compare the recently widowed and the longer-
term widowed with the continuously married. Even though it showed indirect
comparison between the recently widowed and the longer-term widowed by using
the continuously married as the reference of comparison, quick analysis was done
in order to make direct comparison between the recently widowed and the longer-
term widowed. This time, being longer-term widowed was used as the reference in
the equation to compare with being recently widowed and continuously married.
Figure 6.1 presents the odds ratios of the prevalence and incidence of the health
conditions that showed a significant difference between recently widowed and
longer-term widowed elderly people. In most conditions, recently widowed were
less likely to have or get conditions compared to longer-term widowed even though
most differences were not significant. The only exception was depression.
Recently widowed elderly were 3.12 times more likely to acquire depression
compared to the longer-term widowed. Also, at baseline, the recently widowed
who were actually married were less likely to have depression compared to the
longer-term widowed. But after they became widowed in 1998, they were more
likely to have depression than the longer-term widowed. The direct comparison of
recent widowhood and longer-term widowhood confirms the results that the longer-
term widowed were physically worse than the recently widowed while the opposite
was the case for mental health.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
239
Figure 6.1. Odds Ratios of Having (1993 or 1998) or Getting Health
Conditions (between 1993 and 1998) on Being Recently Widowed
Compared to Being Longer-term Widowed: AHEAD
4 i
3 ■ : --------
0 J
New Nagi New Depression IADL93 Depression93 Depression98
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
240
VII. Living Arrangements and Health
So far, whether and how marital status differentiates health has been
discussed. In previous chapters, the effect of the number of people in the same
household on the relationship between marital status and health was examined.
However, the number excludes a spouse, leaving the possibility of the intervention
of living arrangements on health in the relationship to health. Thus, this chapter
examines the relationship between living arrangements and health using a different
approach in order to see whether part of the effect of marital status on health is
related to living arrangements of different marital groups by examining this issue
among unmarried groups. In Table 7.1, percent of the elderly who lived alone in
1993 and 1998 is shown, and in the next three tables (Tables 7.2 to 7.4), differences
in the prevalence and incidence of health conditions and mechanisms for those
living alone and those living with others are presented. Then, logistic regression
analysis is done to see whether and how living alone affects health outcomes
independent from marital status.
Living Alone vs. Living with Others
Table 7.1 shows that among the AHEAD sample, a higher proportion of the
separated/divorced lived alone compared to the never married and the widowed.
Over 65% of the unmarried lived alone both in 1993 and 1998. About 17-18% of
unmarried people lived with somebody else at baseline, but lived alone in 1998.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
241
Table 7.1. Percent (N) of Those Who Lived Alone among Never married, Separated/
divorced and Widowed: AHEAD
1993 1998 Continuously
living alone
Males
Never married 75.61 73.17 75.00 (30)
Became living alone8
Separated/divorced 83.33
0.00
77.27 85.71 (48)
Became living alone
Widowed 75.46
27.27
74.72 78.01 (188)
Became living alone
Total living alone 76.86
19.70
75.00 78.93 (337)
Total became living alone
Females
Never married 65.63
18.39
69.79 69.77 (60)
Became living alone
Separated/divorced 79.56
21.21
76.64 80.80(101)
Became living alone
Widowed 75.08
14.29
71.11 76.50(1071)
Became living alone
Total living alone 74.92
17.75
71.45 76.47(1232)
Total became living alone 17.79
a Became living alone was based on those who lived with others at baseline.
b Only those who lived alone both in 1993 and 1998 were used in analysis.
Table 7.2 shows the percent of those who had diseases, conditions and
functioning problems in 1993 by living arrangements and sex among the unmarried
AHEAD respondents. A significantly higher percent o f the unmarried elderly who
lived alone compared to those who lived with others had functioning difficulties such
as ADL, LADL and Nagi difficulties for both sexes, and hypertension, stroke and
arthritis for females. Table 7.3 shows the percent of those who got new diseases,
conditions and functioning problems between 1993 and 1998 by living arrangements
and sex. As in the prevalence of conditions, a higher proportion of unmarried elderly
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
242
Table 7.2. Percent of Those who Had Functioning Difficulties, Diseases and Conditions
in 1993 by Living Arrangements and Sex: AHEAD____________________________
Males Females
Living
Alone
Living with
Otherfs)
Living
Alone
Living with
Otherfs)
ADL Difficulties 13.91** 28.17 18.32+ 22.22
IADL Difficulties 8.27** 19.12 13.70** 28.19
Nagi Difficulties 18.05** 35.21 37.09* 43.80
Hypertension 43.77 50.70 51.67* 57.78
Diabetes 11.65 15.71 10.89 13.19
Cancer 14.66 18.31 13.23 11.08
Lung Condition 13.64 15.49 9.42 9.50
Heart Disease 33.83 29.58 27.35 27.51
Stroke 5.26 7.04 6.01+ 8.47
Arthritis 26.69 32.86 29.00* 34.83
Broken Hip 3.76 4.23 5.19 6.60
Depression 30.45 22.54 27.76 25.86s
**p<.01, *p<.05, +p<.10 (compared to those living with other(s))
Table 7.3. Percent of Those who Got Functioning Difficulties, New Diseases and
Conditions between 1993 and 1998 by Living Arrangements and Sex: AHEAD
Males________ Females
Living
Alone
Living with
Otherfs)
Living
Alone
Living with
Otherfs)
ADL Difficulties 20.09 20.69 28.09** 40.08
IADL Difficulties 15.61 18.52 18.21* 25.28
Nagi Difficulties 29.36 21.74 36.52 37.56
Hypertension 18.06 20.00 19.39 20.00
Diabetes 6.44 1.69 5.32 4.59
Cancer 7.56 10.34 3.75 4.76
Lung Condition 0.00+ 5.26 4.59 2.62
Heart Disease 13.79 22.45 12.27 12.18
Stroke 5.69 6.06 7.48 8.53
Arthritis 43.92 53.33 58.34 60.42
Broken Hip 1.17 1.47 3.25 3.95
Depression 12.73 22.70 17.64 18.15
**p<.01, *p<.05, +p<.10 (compared to those living with other(s))
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
243
Table 7.4. Mean and Percent of Characteristics of Sample at Baseline (1993) by Living
Arrangements and Sex: AHEAD__________________________________________
Social Support
Males Females
Mean Living Children
Mean Living Siblings
Economic Resources
Mean Per Capita Income
% Health Insurance____
Health Behaviors______
% Current Smoking
% Obesity
% Heavy Drinking_____
Health Care Utilization
% Doctor Visits
Living Living with Living Living with
Alone Otherfs) Alone Otherfs)
2.58 2.77 2.12** 2.77
2.32 2.77 1.97 2.15
19227* 14812 15516** 12073
72.93 69.01 75.24** 60.69
15.79 15.49 9.01 9.50
8.65 12.68 14.77+ 18.73
5.26 4.29 0.57 0.79
87.88 81.16 89.70 88.14
*p<.01, *p< 05, +p<. 10 (compared to those living with other(s))
females living with other(s) had functioning difficulties such as ADL and IADL
difficulties. The means and percentages indicating characteristics o f the sample at
baseline by living arrangements and sex are presented in Table 7.4. About 75% of
the unmarried females who lived alone had supplemental health insurance while
only about 61% of those who lived with other(s) had insurance supplements.
Those living alone had more per capita income than those living with other(s).
Results from Logistic Regressions
Tables 7.5 and 7.6 show the odds ratios from logistic regression of health
outcomes on living alone and age among the unmarried AHEAD respondents.
Those who lived alone were less likely to have ADL, IADL and Nagi difficulties
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
244
among males, and IADL and Nagi difficulties, hypertension and arthritis among
females (Table 7.5). On the other hand, no effect o f living alone on the incidence
of health conditions was found among males, but those females living alone were
less likely to develop ADL and IADL difficulties and less likely to die (Table 7.6).
It is probably because those who live alone among unmarried people do so because
they are healthy enough to live alone. Thus, it is necessary to control baseline
health status in order to prevent the potential reverse causal relationship o f living
alone to health outcomes.
Table 7.5. Odds Ratios of Having Functioning Difficulties, Diseases and Conditions (1993)
by Living Alone and Sex among the Unmarried: AHEAD_________________________
Model 1 Model 2
_______________ Males
ADL Difficulties Living Alone 0.42**
Never Married —
Separated/Divorced
_______________ Age
IADL Difficulties Living Alone 0.38*
Never Married
Separated/Divorced —
_______________ Age
Nagi Difficulties Living Alone 0.42**
Never Married —
Separated/Divorced
_______________ Age
Hypertension Living Alone —
Never Married 2.30*
Separated/Divorced
_______________ Age
Diabetes________ Living Alone
Never Married —
Separated/Divorced
________________ Age_____________________
Females Males Females
0.41
* *
1.06***
0.40*** 0.33** 0.39***
1.14*** 1.11***
0.76* 0.42** 0.76*
1.03*
0.78* - 0.78*
2 . 11*
0.94**
0.95*
**
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
245
Table 7.5 continued
Males Females Males Females
Cancer Living Alone — — — —
Never Married
— — — —
Separated/Divorced — — - - —
Age — —
Lung Conditions Living Alone — — — —
Never Married
— — — —
Separated/Divorced — 2.06** — 1.90*
Age 0.94* - -
Heart Conditions Living Alone — — — - -
Never Married — — — —
Separated/Divorced ~ — — —
Age 0.95* 1.03**
Stroke Living Alone — —
Never Married N/A — N/A
—
Separated/Divorced — - -
Age 1.03*
Arthritis Living Alone — 0.76* — 0.76*
Never Married
— — — —
Separated/Divorced 1.87* - - — —
Age — —
Broken Hip Living Alone - - — — —
Never Married
— — — —
Separated/Divorced — — — —
Age — 1.08***
Depression Living Alone — — — —
Never Married
— — — —
Separated/Divorced — — — —
Age — —
***p<0.001, **p<0.01, *p<0.05
-- not significant
Table 7.6. Odds Ratios of Getting Functioning Difficulties, New Diseases and Conditions
between 1993 and 1998 by Living Alone and Sex among the Unmarried: AHEAD______
Model I Model 2
Males Females Males Females
ADL Difficulties Living Alone — 0.59***
—
0.57***
Never Married — — — —
Separated/Divorced — — —
Age 1.06* 1.09***
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
246
Table 7.6 continued
IADL Difficulties
Nagi Difficulties
Death3
Hypertension
Diabetes
Cancer
Lung Conditions
Heart Conditions
Stroke
Arthritis
Living Alone
Never Married
Separated/Divorced
Age
Living Alone
Never Married
Separated/Divorced
Age
Living Alone
Never Married
Separated/Divorced
Age
Living Alone
Never Married
Separated/Divorced
Age
Living Alone
Never Married
Separated/Divorced
Age
Living Alone
Never Married
Separated/Divorced
Age
Living Alone
Never Married
Separated/Divorced
Age
Living Alone
Never Married
Separated/Divorced
Age
Living Alone
Never Married
Separated/Divorced
Age
Living Alone
Never Married
Separated/Divorced
A ge_______________
Males Females Males
0.66**
1.07**
0.56*
1.91* - 2.38**
1. 12* * *
2.53*
N/A - N/A
N/A - N/A
N/A - N/A
0.41*
Females
0.64**
1. 10* * *
1.05***
0.78*
1.09***
0.91***
2.69*
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
247
Table 7.6 continued___________________________________________________________
_________________________ Males Females Males Females
Broken Hip______ Living Alone N/A — N/A --
Never Married — —
Separated/Divorced — --
_______________ Age 1.08***
Depression______ Living Alone — — — —
Never Married _ _ _ _ _ _
Separated/Divorced _ _ _ _ _
________________Age____________________________________—________-
***p<0.001, **p<0.0l, *p<0.05
-- not significant
“ Death occurred between 1993 and 1995
Odds ratios from logistic regression of having functioning difficulties,
diseases and conditions on living alone and other variables are presented in Table
7.7 for males and in Table 7.8 for females in AHEAD. Among males, the negative
relationship between living alone and the prevalence of ADL, IADL and Nagi
difficulties continued with other variables controlled (models 5 in Tables 7.7 and
7.8), somewhat rejecting the possibility that living alone adversely affects health
outcomes, but still leaving the room for further investigation on this. On the other
hand, those males who lived alone were more likely to have depression than those
who lived with others even when proposed mechanisms were controlled. However,
none of the proposed mechanisms was found to be associated with the relationship
between living alone and health conditions.
While the inverse relationship of living alone on the incidence of ADL and
IADL difficulties was found even with other variables controlled among females
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
248
(Table 7.10), there was no such effect among males (Table 7.9). Elderly males who
lived alone were more likely to develop depression between 1993 and 1998 even
when baseline information such as baseline health, race and education and social
support or economic resources was controlled. However, this effect disappeared
when health behaviors (model 4 in Table 7.9) or all the other proposed mechanisms
controlled (model 5 in Table 7.9) without any significant effect of mechanisms on
the relationship between living alone and the incidence of depression.
The consistently negative relationship between living alone and ADL and
IADL difficulties in baseline health controlled models (Table 7.10) suggests that
the negative effects of living alone on getting ADL and IADL difficulties among
unmarried females found in previous models (Table 7.6) did not totally result from
baseline health status or other factors.
In total, there was little evidence to show that living alone was associated
with poorer physical health among unmarried elderly people, but elderly males who
lived alone were more likely to have and develop depression. Those males who
lived alone were less likely to have all the functioning problems examined, and
those females were less likely to have IADL difficulties and less likely to get ADL
and IADL difficulties. Consequently, the less favorable health of the unmarried in
comparison with the married does not appear to be due to the fact that more
unmarried individuals live alone.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Reproduced with permission o f th e copyright owner. Further reproduction prohibited without permission.
Table 7.7. Odds Ratios from Logistic Regression of Having Functioning Difficulties, Diseases and Conditions (1993) on
Living Alone and Other Variables among Males: AHEAD ____________ ____________ ____________ ____________
Model 1 Model 2 Model 3 Model 4 Model 5
ADL Difficulties Living Alone 0.45* 0.39* 0.45* 0.47* 0.43*
Never M arried -- — -- —
0.25*
Separated/Divorced — — — — —
Age
— « - - — —
Baseline SRH
q 4 4 ***
0.42*** 0.45***
0 4 4 ***
0.43***
W hite — - - — — --
Education
-- - - — — —
Living Children
„ - -
Living Siblings — «
Income
- - —
Health Insurance
— —
C urrent Smoking
— —
Heavy Drinking — —
Obesity
— - -
IADL Difficulties Living Alone
—
0.35* 0.41*
- -
0.38*
Never M arried
— — - - - - —
Separated/Divorced — — — — - -
Age
I 16***
1.15*** 1.16*** 1.17*** 1.16***
Baseline SRH 0.48*** 0,47***
0.49*** 0.50*** 0.48***
White — — — — —
Education 0.87* 0.85**
—
0.88*
Living Children
— —
Living Siblings
— - -
Income — —
Health Insurance — —
249
Reproduced with permission o f th e copyright owner. Further reproduction prohibited without permission.
Table 7.7 continued
Model 1 Model 2 Model 3 Model 4 Model 5
IADL Difficulties C urrent Smoking
— —
Heavy Drinking — —
Obesity
— —
Nagi Difficulties Living Alone 0.40** 0.37** 0.40** 0.41** 0.39**
Never M arried
— - - - - — —
Separated/Divorced
— — — — —
Age — — — — - -
Baseline SRH 0.45*** 0.44*** 0.46*** 0.45*** 0.45***
White
— — - - — —
Education
— — — - - —
Living Children
— —
Living Siblings
- - - -
Income
— —
Health Insurance
— - -
C urrent Smoking
- - —
Heavy Drinking
- - - -
Obesity
—
2.58*
Hypertension Living Alone
— — — - - —
Never M arried 2.07* 2.14* 2.09* 2.09* 2.26*
Separated/Divorced
— — — - - —
Age 0.95* 0.94** 0.95* 0.95* 0.95*
Baseline SRH 0.71*** 0.71** 0.70*** 0.70*** 0.69***
W hite
— — — — —
Education
— — — —
Living Children
— - -
Living Siblings — —
250
Reproduced with permission o f th e copyright owner. Further reproduction prohibited without permission.
Table 7.7 continued
Model 1 Model 2 Model 3 Model 4 Model 5
Hypertension Income — - -
Health Insurance
- - - -
Current Smoking
- - - -
Heavy Drinking 2.98*
—
Obesity — - -
Diabetes Living Alone
— — — — —
Never Married
— — — - - —
Separated/Divorced — — — — —
Age
— — — - - —
Baseline SRH 0.55*** 0.56*** 0.57*** 0.56*** 0.57***
White
— — — — —
Education
— — — — —
Living Children
- - —
Living Siblings
— —
Income
— —
Health Insurance
— —
Current Smoking
- - —
Heavy Drinking
— —
Obesity
— - -
Cancer Living Alone
— — — — —
Never Married
— — - - - - . .
Separated/Divorced
— — — — —
Age
— — — - - - -
Baseline SRH 0.65** 0 .6 6 ** 0.65* 0 .6 6 ** 0 .6 6 **
White — — — — —
Education 1. 1 1* 1.12* 1.12* 1.13* 1.14*
Reproduced with permission o f th e copyright owner. Further reproduction prohibited without permission.
Table 7.7 continued
Model 1 Model 2 Model 3 Model 4 Model 5
Cancer Living Children — - -
Living Siblings —
Income
— —
Health Insurance
— —
Current Smoking
— —
Heavy Drinking
— —
Obesity
- - —
Lung Conditions Living Alone
— - - - - —
Never Married
— — — - -
Separated/Divorced
- - — — — —
Age 0.93* 0.93* 0.92*
— —
Baseline SRH 0.57*** 0.57*** 0.54*** 0.56*** 0.54***
White
— — — — —
Education
— - - — — - -
Living Children
—
Living Siblings — —
Income
— —
Health Insurance — —
Current Smoking
— —
Heavy Drinking
— —
Obesity
— —
Heart Conditions Living Alone
— — — — - -
Never Married
— — — —
Separated/Divorced
— — — — —
Age 0.94* 0.94* 0.94* 0.94* 0.95
Baseline SRH 0.59*** 0.58*** 0.57*** 0.58*** 0.57***
252
Reproduced with permission o f th e copyright owner. Further reproduction prohibited without permission.
Table 7.7 continued
Model 1 Model 2 Model 3 Model 4 Model 5
Heart Conditions White 3.83***
3 7 3 ***
3.27** 4.08*** 3.40***
Education — — - - — - -
Living Children
— —
Living Siblings —
Income — —
Health Insurance
— - -
Current Smoking
— —
Heavy Drinking
- -
Obesity — . .
Arthritis Living Alone — — — — —
Never Married
— — — — - -
Separated/Divorced
— — — — - -
Age
— — — —
Baseline SRH 0.57*** 0.57*** 0.56*** 0.58*** 0.57***
White 0.45** 0.44** 0.41** 0.46* 0.40**
Education
— - - — — —
Living Children
— - -
Living Siblings
— —
Income
— - -
Health Insurance
— —
Current Smoking
— —
Heavy Drinking
- - - -
Obesity
— —
Broken Hip Living Alone
— - - — — —
Never Married — — — - - - -
Separated/Divorced — — — — —
253
Reproduced w ith permission o f th e copyright owner. Further reproduction prohibited without permission.
Table 7.7 continued
Model 1 Model 2 Model 3 Model 4 Model 5
Broken Hip Age
- - — « - - —
Baseline SRH
— — - - — - -
White
— — - - - - - -
Education
— — — — —
Living Children
— —
Living Siblings
— - -
Income 1.0 0 * 1.0 0 *
Health Insurance
— —
Current Smoking
— —
Heavy Drinking
- - —
Obesity
— —
Depression Living Alone 1.97* 2.03* 1.98*
- -
2 .0 2 *
Never Married
— — — —
Separated/Divorced
— — — - -
Age
- - — — — —
Baseline SRH 0.56*** 0.55*** 0.57* 0.57*** 0.56***
White
— - - — —
Education
— - - - - — - -
Living Children — —
Living Siblings
— _ _
Income
— —
Health Insurance
— —
Current Smoking
— - -
Heavy Drinking
— - -
Obesity — —
***p<.001, **p<.01, *p<.05
~ not significant; Analysis of stroke was not technically possible because of the skewed sam ple d istribution.
254
Reproduced with permission o f th e copyright owner. Further reproduction prohibited without permission.
Table 7.8. Odds Ratios from Logistic Regression of Having Functioning Difficulties, Diseases and Conditions (1993) on
living Alone and Other Variables among Females: AHEAD____________ _________________________ ____________
Model 1 Model 2 Model 3 Model 4 Model 5
ADL Difficulties Living Alone
— — - - — —
Never Married
— — — „ - -
Separated/Divorced
— — — — —
Age
1 06***
1.06*** 1.06*** 1.06*** 1.06***
Baseline SRH 0.46*** 0.46*** 0.46*** 0.46*** 0.46***
White — — — —
Education — - - - - - - —
Living Children
- - —
Living Siblings — - -
Income
— —
Health Insurance
- - —
Current Smoking
— —
Heavy Drinking
- - - -
Obesity 2 .2 1 *** 2 .2 1 ***
IADL Difficulties Living Alone 0.46*** 0.47*** 0.46***
0 4 5 ***
0.47***
Never Married — — — —
Separated/Divorced
— - - — — - -
Age
Ill*** 1 ii***
1. 11*** 1. 11***
1 ii#**
Baseline SRH 0.51*** 0.51*** 0.51*** 0.51*** 0.51***
White — - - — — —
Education 0.90*** 0.91* 0.91*** 0.90*** 0.92***
Living Children
— . .
Living Siblings
— —
Income
— —
Health Insurance 0.59** 0.60**
255
Reproduced w ith permission o f th e copyright owner. Further reproduction prohibited without permission.
Table 7.8 continued
Model 1 Model 2 Model 3 Model 4 Model 5
IADL Difficulties Current Smoking
— —
Heavy Drinking
- - —
Obesity
- - —
Nagi Difficulties Living Alone
— — - - —
Never Married
- - — - - - - —
Separated/Divorced
— - - — — —
Age 1.0 2 * 1.03* 1.0 2 * 1.0 2 * 1.0 2 *
Baseline SRH 0.50*** 0.50*** 0.51*** 0.50*** 0.50***
White 1.60** 1.58** 1.67** 1.58** 1.63**
Education
— « — - - - -
Living Children
Living Siblings
- - —
Income
— - -
Health Insurance
— —
Current Smoking
- -
Heavy Drinking
— —
Obesity
— —
Hypertension Living Alone — — - - —
Never Married
— - - — - -
Separated/Divorced 0.67*
- -
0.67*
— —
Age
— — — —
Baseline SRH 0.73*** 0.73*** 0.73*** 0.72*** 0.73***
White 0.69** 0.70** 0.73* 0.75*
—
Education
— — — — —
Living Children
— —
Living Siblings — —
256
Reproduced with permission o f th e copyright owner. Further reproduction prohibited without permission.
Table 7.8 continued
Model 1 Model 2 Model 3 Model 4 Model 5
Hypertension Income 1.0 0 * 1.0 0 *
Health Insurance
- - - -
Current Smoking
0 4 j***
0.40***
Heavy Drinking
— —
Obesity - - —
Diabetes Living Alone
- - — — — —
Never Married
— — - - — —
Separated/Divorced
- - — — — —
Age
0 9 4 ***
0,94***
0,94***
094***
0.93***
Baseline SRH 0.54*** 0.54*** 0.55*** 0.53*** 0.54***
White 0.55** 0.55** 0,57** 0.60* 0.62*
Education
— - - - - - - —
Living Children —
Living Siblings 0.90* 0.90*
Income
— - -
Health Insurance
— - -
Current Smoking 0.29** 0.29**
Heavy Drinking
- - —
Obesity
— - -
Cancer Living Alone
— — — — —
Never Married
— — — - - —
Separated/Divorced
— — — - -
Age
— — — - - —
Baseline SRH 0.84* 0.84* 0.84* 0.84* 0.84*
White 1.74* 1.73* 1.76* 1.77* 1.81*
Education 1.07* 1.07* 1.07* 1.07* 1.07*
257
Reproduced with permission o f th e copyright owner. Further reproduction prohibited without permission.
Table 7.8 continued
Model 1 Model 2 Model 3 Model 4 Model 5
Cancer Living Children
— —
Living Siblings
— - -
Income — —
Health Insurance
— —
Current Smoking
— —
Heavy Drinking
— —
Obesity — —
Lung Conditions Living Alone
- - — — - - —
Never Married
— - - — — —
Separated/Divorced 1.96* 2.04** 1.98* 1.8 6 * 1.92*
Age 0.96* 0.97* 0.96* 0.97*
- -
Baseline SRH 0.60*** 0.60*** 0.61*** 0.60*** 0.61***
White
3 4 9 ***
3.45*** 3.48***
342***
3.34***
Education
— — - - — —
Living Children
— - -
Living Siblings 1.09* 1.09*
Income
— —
Health Insurance
— —
Current Smoking
- - —
Heavy Drinking
- - - -
Obesity
— —
Heart Conditions Living Alone
— — - - — —
Never Married
— — — — —
Separated/Divorced
— — — — - -
Age 1.0 2 * 1.0 2 * 1.0 2 *
—
1.03*
Baseline SRH 0.54*** 0.54*** 0.54*** 0.54*** 0.53***
258
Reproduced with permission o f th e copyright owner. Further reproduction prohibited without permission.
Table 7.8 continued
Model 1 Model 2 Model 3 Model 4 Model 5
Heart Conditions White 1.98***
1 9 7 ***
1.73** 2 .0 1 *** 1.73**
Education
— - - — — - -
Living Children
— —
Living Siblings
— - -
Income — —
Health Insurance 1.48* 1.50**
Current Smoking
— —
Heavy Drinking
- - - -
Obesity — —
Stroke Living Alone
— — —
Never Married
— — - -
Separated/Divorced
- - — - -
Age
— — —
Baseline SRH 0.59*** 0.59*** 0.59***
White 2.13** 2 . 1 2** 2.15*
Education
— — —
N/A N/A
Living Children
- -
Living Siblings
- -
Income
—
Health Insurance
—
Current Smoking
Heavy Drinking
Obesity
Arthritis Living Alone
— - - - - — —
Never Married
— — - - - - - -
Separated/Divorced
—
— — — —
259
Reproduced w ith permission o f th e copyright owner. Further reproduction prohibited without permission.
Table 7.8 continued
Model 1 Model 2 Model 3 Model 4 Model 5
Arthritis Age
- - — - - —
Baseline SRH 0.58*** 0.58*** 0.59*** 0.59*** 0.59***
White 0.45*** 0.45*** 0.47***
0 4 7 ***
0.49***
Education
— — - - „ —
Living Children
— —
Living Siblings
— - -
Income — —
Health Insurance
— —
Current Smoking
— —
Heavy Drinking
— —
Obesity 1.54** 1.53**
Broken Hip Living Alone
- - — — - -
Never Married — — — - -
Separated/Divorced
— - - — - - - -
Age 1 08***
1.09*** 1.08***
Baseline SRH 0.82* 0.82* 0.80*
White 2.31* 2.29*
—
Education
— — —
N/A N/A
Living Children
—
Living Siblings
- -
Income
- -
Health Insurance 1.97*
Current Smoking
Heavy Drinking
Obesity
260
Reproduced with permission o f th e copyright owner. Further reproduction prohibited without permission.
Table 7.8 continued
Depression Living Alone
— — — - -
Never Married
— — - - —
0.51*
Separated/Divorced
— — — —
Age
— — — — —
Baseline SRH 0.51*** 0.50*** 0 51*** 0.51*** 0.51***
White — - -
1.46*
—
1.41*
Education 0.96* 0.96* 0.96*
- -
Living Children — - -
Living Siblings
—
Income — —
Health Insurance 0.61** 0.60***
Current Smoking 1.74** 1.77**
Heavy Drinking — —
Obesity — —
***p<001,**P<01,*p<-05
-- not significant
Table 7.9. Odds Ratios from Logistic Regression of Getting Functioning Difficulties, New Diseases and Conditions between
1993 and 1998 on Living Alone and Other Variables among Males: AHEAD____________ __________________________
Model 1 Model 2 Model 3 Model 4 Model 5
ADL Difficulties Living Alone
— — — — —
Never Married
— — — — - -
Separated/Divorced
— — - - — —
Age 1.07* 1.08* 1.07* 1.08** 1.08*
NJ
o \
Reproduced w ith permission o f th e copyright owner. Further reproduction prohibited without permission.
Table 7.9 continuer
Model 1 Model 2 Model 3 Model 4 Model 5
ADL Difficulties Baseline SRH 0.70* 0.69* 0.69* 0.71* 0.69*
White
— — - - —
Education
— - - —
Living Children
— —
Living Siblings — - -
Income 1.0 0 * 1.0 0 *
Health Insurance
— —
Current Smoking
- - - -
Heavy Drinking — —
Obesity
— —
IADL Difficulties Living Alone
— - - — — - -
Never Married
— — — - - —
Separated/Divorced
— — — - - —
Age 1.08* 1.08* 1.08* 1.09** 1.09**
Baseline SRH 0.71* 0.70* 0.70* 0.71* 0.70*
White
— — — —
Education
— — - - —
Living Children
— —
Living Siblings
— —
Income
—
Health Insurance
- - —
Current Smoking
— —
Heavy Drinking
— —
Obesity
- - —
Nagi Difficulties Living Alone
— - - — — —
Never Married — — — —
Reproduced with permission o f th e copyright owner. Further reproduction prohibited without permission.
Table 7.9 continuer
Model 1 Model 2 Model 3 Model 4 Model 5
Nagi Difficulties Separated/Divorced
— — — — —
Age
— - - — - -
Baseline SRH 0.65** 0.65** 0.64** 0.64** 0.64**
White
— — - - — —
Education
— — — —
Living Children
—
Living Siblings
— - -
Income
- - —
Health Insurance
— —
Current Smoking
- - —
Heavy Drinking
- -
Obesity
— —
Hypertension Living Alone — ~ —
Never Married
— — —
Separated/Divorced
— — - -
Age
— — —
Baseline SRH
- - —
White
— - -
Education 0 .8 8 *
- -
N/A N/A
Living Children
—
Living Siblings
- -
Income
- -
Health Insurance
—
Current Smoking
Heavy Drinking
Obesity
Reproduced with permission o f th e copyright owner. Further reproduction prohibited without permission.
Table 7.9 continuec
Model 1 Model 2 Model 3 Model 4 Model 5
Diabetes Living Alone
— - - - - — - -
Never Married
— — — - - —
Separated/Divorced
— — — — - -
Age
— — — —
Baseline SRH — — — — —
White
— — — — —
Education — — — - - —
Living Children
— —
Living Siblings
- - —
Income
— - -
Health Insurance
— —
Current Smoking
— - -
Heavy Drinking — —
Obesity 4.76* 6.93**
Cancer Living Alone
— — —
Never Married
- - - - - -
Separated/Divorced —
—
Age — — —
Baseline SRH
- - — —
White
— - -
Education
—
1.17*
—
N/A N/A
Living Children
—
Living Siblings
—
Income
- -
Health Insurance
- -
Current Smoking
264
Reproduced with permission o f th e copyright owner. Further reproduction prohibited without permission.
Table 7.9 continuer
Model 1 Model 2 Model 3 Model 4 Model 5
Cancer Heavy Drinking
Obesity
Arthritis Living Alone
— — — - - - -
Never Married
— — — - - —
Separated/Divorced
— — — —
Age
— — — — - -
Baseline SRH 0.79*
— — — —
White
— - - — — —
Education
— — - - — - -
Living Children
- - - -
Living Siblings
— —
Income
— - -
Health Insurance
— - -
Current Smoking
- - —
Heavy Drinking
- -
Obesity — - -
Depression Living Alone 2 .6 6 * 2.72* 2.71*
— —
Never Married
- - — — - - —
Separated/Divorced
— — — - - - -
Age
— — — — —
Baseline SRH
— — — - - - -
White
— —
0.41*
— - -
Education
— — — - - - -
Living Children
— - -
Living Siblings
— —
Income — —
265
Reproduced with permission o f th e copyright owner. Further reproduction prohibited without permission.
Table 7.9 continuer
Model 1 Model 2 Model 3 Model 4 Model 5
Depression Health Insurance
— - -
Current Smoking — —
Heavy Drinking - - —
Obesity — - -
***p<.001, **p<.01, *p<.05
-- not significant
"A nalyses of h ea rt conditions, lung conditions, stroke an d brok en h ip w ere not technically possible because of the skew ed sam ple
d istribu tion .
Table 7.10. Odds Ratios from Logistic Regression of Getting Functioning Difficulties, New Diseases and Conditions
between 1993 and 1998 on Living Alone and Other Variables among Females: AHEAD ______________________
Model 1 Model 2 Model 3 Model 4 Model 5
ADL Difficulties Living Alone 0.58** 0.60** 0.58** 0.59** 0.60**
Never Married
— — — — —
Separated/Divorced — — — — —
Age 1.09*** 1.09*** 1.09*** 1.09***
1 10***
Baseline SRH 0.60*** 0.60*** 0.60*** 0.60*** 0.60***
White
— —
1.54*
—
1.59*
Education
— — — — —
Living Children
— —
Living Siblings — —
Income — —
Health Insurance 0.67* 0 .6 8 *
Current Smoking
—
Heavy Drinking
— —
266
Reproduced with permission o f th e copyright owner. Further reproduction prohibited without permission.
Table 7.10 continued
Model 1 Model 2 Model 3 Model 4 Model 5
ADL Difficulties Obesity 1.50*
—
IADL Difficulties Living Alone 0.63** 0.63** 0.63* 0.63* 0.63**
Never Married
— - - - - - - —
Separated/Divorced
— — — - - - -
Age 1. 10*** 1.1 0*** 1.1 0***
1 J 1***
1.10***
Baseline SRH 0.59*** 0.59*** 0.59*** 0.59*** 0.59***
White
— — — —
1.57*
Education
— - - — — —
Living Children
— - -
Living Siblings
— —
Income — - -
Health Insurance
—
0 .6 8 *
Current Smoking
— - -
Heavy Drinking
- —
Obesity
— —
Nagi Difficulties Living Alone — — - - — —
Never Married
— — — — —
Separated/Divorced
— — — - - - -
Age 1.05*** 1.05*** 1.05*** 1.05*** 1.06***
Baseline SRH 0.64*** 0.64*** 0.64*** 0.64*** 0.65***
White
— — - - - - —
Education
— — — — - -
Living Children
— —
Living Siblings
— - -
Income — —
Health Insurance — —
267
Reproduced with permission o f th e copyright owner. Further reproduction prohibited without permission.
Table 7.10 continued
Model 1 Model 2 Model 3 Model 4 Model 5
Nagi Difficulties Current Smoking
— —
Heavy Drinking
— —
Obesity 1.50* 1.51*
Hypertension Living Alone
— — — —
Never Married — — — - -
Separated/Divorced
- - — — — —
Age
— - - - - - - —
Baseline SRH 0.80** 0.80** 0.79** 0.80** 0.79**
White
— — - - - - - -
Education
- - — - - — —
Living Children
— —
Living Siblings
— « _
Income
— —
Health Insurance — - -
Current Smoking
— - -
Heavy Drinking
- -
Obesity —
Diabetes Living Alone
— — —
Never Married
— — —
Separated/Divorced
— — —
Age 0.90*** 0.90*** 0.90***
Baseline SRH
— — —
White
— — —
Education 0.89*** 0 .8 8 ** 0.90* N/A N/A
Living Children —
Living Siblings —
268
Reproduced with permission o f th e copyright owner. Further reproduction prohibited without permission.
Table 7.10 continued
Model 1 Model 2 Model 3 Model 4 Model 5
Diabetes Income —
Health Insurance
- -
Current Smoking
Heavy Drinking
Obesity
Cancer Living Alone
— — — - - —
Never Married
— - - - - — - -
Separated/Divorced 2.65* 2.75* 2.75* 2.48* 2 .6 8 *
Age
— — — - - —
Baseline SRH
— - - — — —
White
— — — — —
Education
— - - - - - - - -
Living Children
— - -
Living Siblings
- - - -
Income
— - -
Health Insurance
- -
Current Smoking
. . - -
Heavy Drinking 6.15* 6.05*
Obesity
— —
Lung Conditions Living Alone
— — —
Never Married — - - —
Separated/Divorced
— - - —
Age
— - - —
Baseline SRH 0.69** 0.69** o
*
*
White 3.41** 3.48** 3.67**
Education — —
N/A N/A
Reproduced with permission o f th e copyright owner. Further reproduction prohibited without permission.
Table 7.10 continued
Model I Model 2 Model 3 Model 4 Model 5
Lung Conditions Living Children
—
Living Siblings
—
Income 1.0 0 *
Health Insurance
- -
Current Smoking
Heavy Drinking
Obesity
Heart Conditions Living Alone
- - — — — —
Never Married
- - — - - — . .
Separated/Divorced
- - - - - - - -
Age
— — - - — —
Baseline SRH 0.80** 0.80** 0.81* 0.79** 0.79**
White
— - - — — —
Education
— — — — - -
Living Children
— —
Living Siblings
— —
Income
- - —
Health Insurance
— —
Current Smoking 0.28** 0.28**
Heavy Drinking
—
Obesity — —
Stroke Living Alone
— — —
Never Married
— — - -
Separated/Divorced
- - - - - -
Age — — - -
Baseline SRH 0.72*** 0.73*** 0.72***
270
Reproduced with permission o f th e copyright owner. Further reproduction prohibited without permission.
Table 7.10 continued
Model 1 Model 2 Model 3 Model 4 Model 5
Stroke White
— —
Education
— — - -
N/A N/A
Living Children —
Living Siblings
—
Income —
Health Insurance
Current Smoking
Heavy Drinking
Obesity
Arthritis Living Alone
— — — - - - -
Never Married
— — - - — —
Separated/Divorced
— — - - — - -
Age — — — — - -
Baseline SRH 0.75*** 0.75*** 0.75*** 0.75*** 0.75***
White
— — — - -
Education
— — — « —
Living Children — —
Living Siblings — - -
Income
- - . .
Health Insurance
— —
Current Smoking
- - —
Heavy Drinking
— - -
Obesity
— —
Broken Hip Living Alone
— — —
Never Married
— — —
Separated/Divorced — —
—
Reproduced with permission o f th e copyright owner. Further reproduction prohibited without permission.
Table 7.10 continued
Model 1 Model 2 Model 3 Model 4 Model 5
Broken Hip Age 1.09*** 1.08** 1.09***
Baseline SRH — —
White 4.46* 4.54* 5.38**
Education — —
1 .12* N/A N/A
Living Children 0.81*
Living Siblings
—
Income
- -
Health Insurance
—
Current Smoking
Heavy Drinking
Obesity
Depression Living Alone
- - — — —
Never Married
— - - — — —
Separated/Divorced
— — „ — —
Age — — — — - -
Baseline SRH 0.80** 0.80** 0.81** 0.80** 0.80**
White
— — — . . - -
Education
— — — — —
Living Children — - -
Living Siblings
- - —
Income
—
Health Insurance
- - —
Current Smoking
— —
Heavy Drinking
— —
Obesity — —
***p<.001, **p<.01, *p<.05
-- not significant
272
273
VIII. Summary and Discussion
This final chapter summarizes the findings of the previous chapters on
effects of marital status on health and the mechanisms by which marital status
affects health focusing on differences by type of health outcome, age and gender,
and effects of widowhood and living arrangements on health. Then, it discusses
major findings and limitations of the study, and makes suggestions for future
studies on this area.
Summary
Using the three different datasets, this study examined the effect o f marital
status on a number o f health indicators and the mechanisms by which marital status
affects health under the context of gender difference. Each dataset has distinct
characteristics: The HRS is a nationally representative sample of pre-retirement
age; the AHEAD is also a nationally representative sample, but o f older ages; and
the MacArthur consists of a smaller number of elderly people representing elders
who were high functioning in specific areas. In general, the findings showed that
the married were healthier than the never married, the separated/divorced and the
widowed: For some health outcomes, the married were less likely to have or to
acquire health conditions than unmarried people. For other health outcomes, there
was no significant difference between married group and each unmarried group in
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
274
the prevalence and incidence of health conditions. However, there were a few
instances in which the married were worse than the unmarried in health status.
Comparison of each unmarried group with the married group as reference
showed that the widowed were worst among unmarried groups, and the never
married were better than the separated/divorced and the widowed in health status.
The results suggest that the widowed were more disadvantaged in health status
relative to the married, and so were the separated/divorced, but to a lesser degree.
Also, health status of the never married was worse than that of the married, but the
degree of difference from the married was much less than that of the
separated/divorced and the widowed. The variation in the existence, direction and
degree of the effect of marital status on health is mainly attributed to type o f health
conditions, age and gender.
Differential Effects o f Marital Status by Type o f Health Outcome
This study hypothesized that marital status is more likely to be related to
fatal conditions (hypertension, diabetes, cancer, lung conditions, heart conditions
and stroke) and death than nonfatal conditions (arthritis and fracture), and more
related to mental health condition (depression) and functioning difficulties (ADL,
IADL and Nagi difficulties) than other physical conditions. The findings
confirmed the hypothesis such that the effect of being unmarried was stronger on
ADL and Nagi difficulties and depression than other physical conditions including
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
275
diabetes, cancer, heart conditions, stroke, arthritis and fracture, and the effect was
stronger on fatal conditions (particularly, hypertension, lung conditions, heart
conditions and stroke) than nonfatal conditions such as arthritis and fracture. In
most cases, unmarried status increased the probability of having or getting
functioning problems and depression. Also, being never married among males (in
AHEAD) increased the probability of dying compared to their married counterparts.
Thus, it appears that functioning and mental health are more strongly related to
marital status than other physical conditions or chronic diseases, and fatal
conditions and death are more related than nonfatal conditions. In sum, the results
of this study showed that the effect of marital status on health differed by type of
health outcome, and the adverse effect of being unmarried on health was found to
be greater for measures of functioning, mental health and some of the fatal
conditions.
Mechanisms
Three sets of mechanisms which intervene on the relationship between
marital status and health were examined in the logistic regression models: Social
support, economic resources and health behaviors. It was found that some
mechanisms were associated with the effect of marital status on health, but in many
cases, the role of the mechanisms was minimal or did not exist. The effects of
marital status on health disappeared in models in which all proposed mechanisms
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
276
were controlled, but consistently strong mechanisms by which marital status affects
health were not found. The effect o f each set of mechanisms was first examined by
introducing them separately in the equation, and then the effects of all the proposed
mechanisms together were examined by putting all in the equation at the same time.
The results showed that most of the effects of the mechanisms significant in the
separate entry remained so when all other mechanisms were considered together.
Among the HRS respondents, friendship satisfaction, an indication of social
support from friends, had the most frequent impact on the relationship between
marital status and health, showing that for those in their 50s to early 60s,
friendships partly explain the adverse effect of being unmarried on health. If they
had the same level of satisfaction with friendships, unmarried males were not more
likely to have or acquire conditions such as lung conditions and depression, and
unmarried females were not more likely to have or acquire depression. The only
other social support factors that affected the relationship between marital status and
health in the HRS and AHEAD samples were higher satisfaction with
neighborhood which reduced the likelihood of depression among males, and more
household members for depression among females. Among the MacArthur
respondents, number of living children and participation in leisure activities were
the mechanisms by which being widowed aversely affected the prevalence of
diabetes and the incidence of Nagi difficulties, respectively.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
277
In terms of economic resources, supplemental health insurance decreased
the probability of having or acquiring depression both for males and females, and
the adverse effect of being never married on prevalence of arthritis was partly
explained by income.
Health behaviors appeared to be more frequently associated with the
relationship between marital status and health among females than among males.
For males, obesity cancelled out the effect of being unmarried on diabetes and
stroke, and smoking did for cancer and depression. On the other hand, among
females, smoking was the mechanism by which unmarried status affected ADL
difficulties, hypertension, lung conditions and depression; obesity was related to
hypertension, heart conditions and arthritis; and activity participation was related to
ADL difficulties and heart conditions.
Mechanism Differences by Type o f Health Outcome
One indirect way of examining the differential effect of marital status on
various health indicators was through what different mechanisms were involved in
the relationship between marital status and different health outcomes under the
assumption that mechanisms show “how” marital status affects health. The
proposed mechanisms were found to have some influence on the relationships
between marital status and health. Health behaviors were the most frequent
mechanisms that affected all types of health conditions studied. However, the
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
278
proposed mechanisms were not differentiated as hypothesized. Different from the
hypothesis that health behaviors would affect physical conditions more than mental
condition, the relationship between marital status and fatal physical health
conditions were mostly explained by health behaviors, but health behaviors were
also related to mental health condition. While it was hypothesized that social
support would play an important role in fatal conditions, functioning limitations
and death, there was no evidence found that social support is more likely to be
associated with the relationship between marital status and more serious conditions
such as fatal conditions, functioning limitations and death than nonfatal conditions.
Nor were economic resources found to be more important for fatal conditions and
death. On the other hand, the adverse effect o f being unmarried on depression was
through all three sets of mechanisms— social support, economic resources and
health behaviors.
Age Differences
Younger cohorts (HRS respondents) had the strongest relationship between
marital status and health. More health conditions in HRS were related to marital
status than in AHEAD, and the effect tended to be stronger among the HRS
respondents than among the AHEAD respondents. For instance, the relative
likelihood of having depression among the unmarried compared to the married was
greater among the HRS respondents than among the AHEAD respondents, which
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
279
indicates the age difference (HRS vs. AHEAD). In addition, the marital status
difference in the prevalence of depression was found both in HRS and AHEAD,
but differential incidence of depression was only found in HRS. That is, younger
unmarried people (HRS) were more likely to develop depression, but older
respondents (AHEAD) were not. It may indicate that single marital status is more
disruptive to younger cohorts than older cohorts.
Why did unmarried status more matter among younger people? This age
difference may partly be explained through differential effects o f mechanisms in
the relationship between marital status on health. For example, at younger ages,
economic resources may affect health status o f unmarried people, but older people
take advantage of social provisions such as retirement benefits and health insurance
(e.g. Medicare), which actually reduces the impact that income changes can make
on health outcomes.
Gender Differences
Men and women differed in the relationship between marital status and
health. In most cases, relatively stronger health benefits o f marriage for men,
hypothesized in this study, were noticed (e.g. the prevalence of lung conditions,
incidence of ADL difficulties and hypertension among the widowed, the prevalence
and incidence of depression particularly among the separated/divorced and the
widowed).
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
280
Depression is a good example of the gender difference. Unmarried males in
HRS and AHEAD were more likely to have or get depression than the
corresponding females. The effect of being unmarried on the prevalence of
depression was greater for males than for females in HRS and AJHEAD, and the
effect on the incidence of depression was greater for males than for females in HRS.
In terms of gender differences in mechanisms, the hypothesis that social
support and health behaviors would be more related to the effect of marital status
on health among males and economic resources would be among females, was
rejected. As shown in Figure 5.27 in Chapter V, no systematic difference of
mechanisms was found between men and women. Differences in mechanisms
were more attributed to type of health outcomes as discussed.
Widowhood
Odds ratios from logistic regressions showed that, with age controlled,
being widowed was not associated with a higher or lower probability of acquiring
any health conditions except stroke and depression. The effect on stroke
disappeared when interactions between gender and duration o f widowhood were
controlled and when all mechanisms were controlled while the effect of
widowhood on depression was somewhat weakened with all mechanisms
controlled, but still remained significant. On the other hand, more relationships
were found between being widowed and the prevalence of health conditions: Being
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
281
widowed was related to a higher probability of having ADL and Nagi difficulties,
lung conditions and depression.
The hypothesis that recently widowed elderly people would be worse in all
health outcomes than the continuously married while the longer-term widowed
would not be much different from the continuously married was not confirmed.
Rather, the findings showed that longer-term widowhood predicted worse physical
health to a greater degree than recent widowhood did while the recently widowed
were worse than the longer-term widowed only in mental health condition. The
longer-term widowed were more likely to have ADL, IADL and Nagi difficulties,
lung conditions and stroke compared to those who remained married, but the
recently widowed were not different from the continuously married in any physical
conditions.
However, recent widowhood was strongly related to the incidence of
depression, but longer-term widowhood was not, while widowhood, either recent or
longer term, increased the prevalence of depression. Comparison of the prevalence
of depression in 1998, the prevalence in 1993, and the incidence between 1993 and
1998 showed very interesting results. Widowhood had a significant adverse impact
on the prevalence in 1998 and the incidence of depression. The recently widowed,
compared to those who remained married, were more likely to develop depression
while the longer-term widowed were not likely to get depression, but sustained
depression that occurred previously.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
282
In terms o f a differential effect of widowhood by age, it was found that
increased age cancelled out the adverse effect of being widowed on several health
outcomes such as the incidence of ADL, IADL and Nagi difficulties and broken hip
among the longer-term widowed, for example. That is, the adverse effect of being
longer-term widowed on those health conditions was through older ages among
them. This finding that age was related to the effect o f widowhood on health is
consistent with the previous findings showing that the older widowed were
potentially disadvantaged in health status compared to the younger widowed
(Heinemann, 1983), indicating that in addition to bereavement per se, aging itself
may contribute substantially to the measured longitudinal effects of widowhood
(Bennett & Morgan, 1992).
There was no evidence that social support or health behaviors deteriorated
following the loss of a spouse, but the recently widowed experienced a significant
decrease in income. The logistic regression analysis showed that there was no role
of covariates in the relationship between widowhood and the incidence of health
conditions, and baseline income was the only mechanism found between
widowhood and the prevalence of health conditions: Part of the adverse effect of
being longer-term widowed on having depression was explained by baseline
income.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
283
Living Arrangements
This study examined whether living arrangements were part of the
explanation of the poorer health status of unmarried people relative to married
people. By examining whether those who lived alone were better or worse in
health status independent from marital status, this study attempted to tease out the
effect of living arrangements in the relationship between marital status and health.
Living alone was not found to be associated with health conditions except
functioning and depression. This finding in models in which marital status was
controlled, suggests that the effect of marital status on health does not occur
through living arrangements. Both elderly males and females who lived alone were
less likely to have or get functioning problems while only males living alone, but
not females, were more likely to have or develop depression. It appears that both
males and females somewhat voluntarily live alone when their health status is good
enough to afford independent living. But the consequence of single residency
seems to be affecting males’ mental health in a negative way, but it was not the
case for females.
Regarding functioning difficulties among those who lived alone, the
findings showed that those living alone performed better functioning than those
living with others. The results may be attributed to reverse causation, indicating
that people live alone because they are functionally able, and those who have
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
284
functional limitations tend to live with other people. The incidence analysis shows
this possibility.
Some gender difference was noticed. Both elderly males and females who
lived alone were more likely to have better functioning ability than those who lived
with others. But, only elderly females who had lived alone since 1993 developed
ADL and IADL difficulties, which was not the case for males. The possibility is
that over the term of the study, 5 years, men might be more likely to become co
resident either in their own house or that o f another because of their functioning
deterioration, while women might stay in their own place alone even with
functioning problems. The comparison o f the percent of those with ADL
difficulties in 1993 and 1998 among those who had continued to live alone sheds
light on this possibility (Table 8.1). Among males, a lower percent of males living
alone had ADL difficulties in 1993 than those living with others, but the percent of
those with ADL difficulties in 1998 were similar for those living alone and those
Table 8.1 Percent of Those who Had ADL Difficulties in 1993 and 1995 by Living
Arrangements and Sex: AHEAD____________________________________________
______________Males___________________________ Females____________
Living Living with Living Living with
Alone Qtherfs) Alone Qther(s)
1993 13.91* 28.17 18.32 22.22
1998_______ 37.31_____________38.78___________37.93*___________ 46.69
*p<.01 (compared to those living with other(s))
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
285
living with others. On the other hand, the opposite thing was found for females.
The ability to perform ADL became significantly different between those living
alone and those living with others as time went by. The findings that the difference
of the percents with ADL difficulties for those living alone and those living with
others became smaller among males, but the difference became greater among
females, suggest that men might change their living arrangements with functioning
limitations while women tend to stay as they are even with functioning limitations.
Discussion
The study findings suggest that the effects of marital status on health
depend on age, gender and type of health conditions. Overall, this study suggests
that the married are better than the unmarried in health, and the role of marital
status is more important in mental health, functioning difficulties, some fatal
conditions and death than nonfatal conditions.
While findings of less health difference of being never married from being
married (compared to health difference of being separated/divorced from being
married) somewhat supported the hypothesis that the duration of marital status
matters in health status, findings on the stronger effect of longer-term widowhood
relative to recent widowhood on physical health did not support the hypothesis that
recent change of marital status has a greater adverse impact on health outcomes.
Even though much attention has been given to adverse effects of recent widowhood
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
286
on physical health, the study findings suggest that longer-term adverse effect of
widowhood on health should not be neglected.
While there were only a few significant differences in physical health
between the widowed and the continuously married, widowhood had a strong effect
on the psychological symptoms of depression. Widowhood had a significantly
adverse impact on the incidence as well as prevalence of depression. In particular,
the recently widowed, compared to those who remained married, were more likely
to develop depression while the longer-term widowed were not likely to get
depression, but sustained depression that occurred previously. Increased likelihood
of getting depression among the recently widowed and high prevalence of
depression among longer-term widowed imply that treatment or prevention of
depression would be appropriate for the recently widowed who are at high risk of
getting depression and longer-term widowed who suffer from chronic depression.
The findings of an adverse impact o f widowhood on depression in contrast to the
minimal effect of widowhood on many physical diseases and conditions should
alert health professionals to the needs for psychological interventions among the
widowed. Particularly, depression following the loss of a spouse should be
carefully monitored, prevented and treated at the beginning of its onset.
Few mechanisms proposed in the study were associated with the
relationship between widowhood and health outcomes. Those who became
recently widowed did not necessarily experience adverse changes in social support
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
287
and health behaviors. In fact, the findings showed that they had more household
members, and a higher percentage o f them adopted healthy behaviors since they
became widowed than when they were married. Perhaps, established lifelong
lifestyles of elderly people do not change significantly upon the loss o f their
spouses (Bennett & Morgan, 1992; Stroebe & Stroebe, 1983), so health is not
affected through changed health habits. Also, intact or even improved social
network of those who were and who became widowed may rather play a role in
preventing deterioration of their health.
On the other hand, the lack o f a definite role for the tested mechanisms on
the effect of widowhood on health does not necessarily mean that the relationship is
not through these or other covariates. It suggests that widowhood may have some
direct effect on health. However, measurements of some of the mechanisms
proposed for this study may need to be improved. For example, use o f preventive
care or earlier health care usage was a more appropriate measure for health care
utilization than doctor visits in the past 12 months.
The results suggest that living alone is not always associated with poorer
health. The results of this study are consistent with those of Anson (1989), Hughes
and Gove (1981), and Cramer (1993) who found few differences in health between
those living with another person and those living alone, with the exception of
functioning difficulties and depression. One implication o f this finding is that any
potential health benefits of living with someone may depend not simply on co
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
288
residence per se but on some other aspect of the relationship, such as on how
supportive that person is seen to be (Cramer, 1993). For example, living alone may
be better healthwise than living with someone who is perceived as being
unsupportive. There is substantial evidence that the quality of social support
covaries positively with physical and psychological well-being (Cramer, 1991;
Schwarzer & Leppin, 1989). Thus, the results of this study may not be surprising
given that the quality of social support was not controlled in analysis.
The results that those unmarried elderly who lived alone were more likely
to be functionally capable may still reflect some reverse causation even though
baseline self-rated health was taken into account. Unmarried persons may have to
live with someone else or in an institution if they are to survive with severe
functional problems while functional independence may enable them to live alone.
Living with someone else may not necessarily be beneficial for older people.
Some efforts to develop multigenerational housing may not provide direct help for
those whose family relationships were not good, or who have been comfortably
adjusted to living alone. Policy and program development efforts of policymakers
and practitioners to improve quality of relationship with family members, to
enhance social networks for those living alone, or to provide better services for
elderly people living alone would be as important as solely working on co-residing
issue.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
289
Analysis with the MacArthur data showed somewhat incompatible results
from those o f the HRS and the AHEAD, particularly in the prevalence and onset of
depression for unmarried people compared to the married: No effect was found in
the relationship between marital status and depression, which, in fact, showed a
very strong association in HRS and AHEAD. Using a continuous depression scale
did not change the results obtained using a dichotomous scale of depression. In
general, fewer effects, or no effect, of marital status on health were found in the
MacArthur sample. Small sample size may have contributed to these results, but it
does not completely explain this difference.
Limitations o f the Study
There are several limitations or suggestions for further study in this area.
First, this study did not incorporate the possible impact o f the shared environment
such as a spouse’s health behaviors, education or occupation on individual health.
When one or the other spouse is faced with a health problem, for example, the
health status o f a spouse may affect the other’s health in that the dysfunctional or
unhealthy spouse may not be able to provide enough support for a sick spouse.
Some previous attempts at establishing paired or couple models in explaining
health and functioning (Lillard & Waite, 1995; Mare & Palloni, 1988; Smith &
Zick, 1994) deserve further attention in future study o f the effect o f marital status
on health.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
290
The information on the exact length of widowhood (when a person became
widowed) was not incorporated in analysis. Recent widowhood has been defined
in various ways in previous research, and also in this study for different datasets.
Inclusion of the exact time of becoming widowed would empower the findings on
the effect o f change in marital status and duration of widowhood on health.
Employing better indicators of mechanisms would improve the exploration
of the relationship between marital status and health. The use of a broader range of
health behavior measures is recommended. For instance, other measures of health
behaviors such as preventive care (e.g. routine medical and dental exams, stress
management)4 and environmental safeguards (e.g. seat belts, avoiding home
accidents) in addition to lifestyle (e.g. exercise, smoking, drinking, nutrients) could
be expected to have an impact on the relationship between marital status and health
(Schone & Weinick, 1998).
In terms of economic resources, income excluding asset and pension
income may not be a fair indicator for people with a high level of income. Social
Security Administration (2000) reported shares of income by quintiles of total
4 Questions of some preventive health medical test or procedures such as a flu shot, a blood test for
cholesterol, monthly check-up of breasts for lum ps, a mammogram or x-ray of the breast to search
for cancer, a PAP smear and an examination of prostate to screen for cancer are available in the
second wave (1995) of the AHEAD. The percent of those who had none of these preventive health
care practices regardless of age was lower among the m arried than among each unm arried group for
males (4.13% among the m arried; 11.58% among the never m arried, 10.95% among the
separated/divorced; and 6.07% among the widowed), but the difference by m arital status was not
significant for females.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
291
income.5 Persons in the lowest quintile have the largest share of total income from
Social Security benefits (82%), and public assistance provides the second largest
share (10%). For those in the highest income quintile, however, Social Security
(18%) and public assistance (less than 0.5%) provide less than one fifth of total
income while earnings (31%), income from assets (28%) and pensions6 (20%)
provide higher shares of income. Even though considering assets in analysis did
not change the results, income excluding asset in this study may not be an accurate
indicator of financial status for people in different income levels.
Comparison of the results from the three datasets used for this study should
be cautiously made in that differences in time period studied and the structure of
questions and inclusion of different variables may or may not be the factors that
created the different results among the datasets.
A lack of findings on difference in health between the married and the
separated/divorced or the never married and between the continuously married and
the recently widowed in some data (particularly in MacArthur study) as well as a
lack o f findings on distinct mechanisms may be partly due to the sample size and
the distribution of covariates and health conditions in the sample. The small
sample size of the separated/divorced and the never married, and the skewed
distribution of the sample in certain mechanisms (e.g. heavy drinking, leisure
5 Income in Social Security Administration report includes Social Security, pensions, asset income,
earnings, public assistance and others while income in this study excludes wealth or asset income.
6 Pensions include private pensions and annuities, government employee pensions, Railroad
Retirement, and IRA, Keogh, and 401(k) payments.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
292
activities) might have led to “false negatives.” Rather than a set of single count of
dummy variables for the conditions used in this study, alternative measures such as
a group of serious (fatal) conditions and nonfatal conditions categorized according
to the International Classification of Diseases (ICD) may help increase the
statistical power for testing rare diseases such as stroke and hip fracture (Ferraro &
Wilmoth, 2000) by making variables more analytical.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
293
References
Anson, O. (1989). Marital status and women’s health revisited: The importance o f a
proximate adult. Journal of Marriage and the Family. 51. 185-194.
Avis, N., Brambilla, D., Vass, K., & McKinlay, J. (1991). Effect of widowhood on
health: A prospective analysis from the Massachusetts Women’s Health Study.
Social Science and Medicine. 33(9). 1063-1070.
Bengtson, V., Rice, C., & Johnson, M. (1999). Are theories of aging important?
Models and explanations in gerontology at the turn of the century. In V. Bengtson,
& W. Schaie (Eds.), Handbook of Theories of Aging (Ch. 1). New York: Springer.
Bennett, K. (1997). Longitudinal study of well-being in widowed women.
International Journal of Geriatric Psychiatry. 12(1), 61-66.
Bennett, K., & Morgan, K. (1992). Health, social functioning, and marital status—
Stability and change among elderly recently widowed women. International Journal
of Geriatric Psychiatry. 7(11), 813-817.
Berkman, L., & Syme, S. (1979). Social networks, host resistance, and mortality: A
nine-year follow-up study of Alameda County residents. American Journal of
Epidemiology. 109. 186-204.
Blazer, D. (1982). Social support and mortality in an elderly community
population. American Journal of Epidemiology. 115. 684-694.
Burman, M., Wells, K., Leake, B., & Landsverk, J. (1988). Development of a brief
screening instrument fo detecting depressive disorders. Medical Care. 26. 775-589.
Cassileth, B., Lusk, E., Strouse, T., Miller, D., Brown, L., Cross, P., & Tenaglia, A.
(1984). Psychosocial status in chronic illness: A comparative analysis of six
diagnostic groups. New England Journal of Medicine. 311. 506-511.
Choi, N. (1996). The never-married and divorced elderly: Comparison of economic
and health status, social support, and living arrangement. Journal of Gerontological
Social Work. 26(1/2), 3-25.
Clayton, P. (1974). Mortality and morbidity in the first year of widowhood.
Archives of General Psychiatry. 30(6), 747-750.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
294
Cohen, S., & Wills, T. (1985). Stress, social support, and the buffering hypothesis.
Psychological Bulletin. 98(2), 310-357.
Colin, M. (1999). Gains in health expectancy from the elimination of diseases
among older people. Disability and Rehabilitation, 21. 211-221.
Colsher, P., & Wallace, R. (1990). Elderly men with histories of heavy drinking:
Correlates and consequences. Journal of Studies on Alcohol. 51(6). 528-535.
Cotten, S. (1999). Marital status and mental health revisited: Examining the
importance of risk factors resources. Family Relations. 48(3). 225-233.
Cramer, D. (1991). Social support and psychological distress in women and men.
British Journal of Medical Psychology. 64. 147-158.
_______ . (1993). Living alone, marital status, gender and health. Journal of
Community and Applied Social Psychology. 3(1). 1-15
Crimmins, E., & Kim, J. K. (2000). Health status and health dynamics among U.S.
and California Residents over Age 50. Paper presented in Conference, Employment
and Health Policies for Californians Over 50, Jan. 6-7, 2000. San Francisco, CA.
Crimmins, E., Saito, Y., & Ingegneri, D. (1997a). Trends in disability-free life
expectancy in the United States, 1970-1990. Population and Development Review.
23(3), 555-572.
Crimmins, E., Saito, Y., & Reynolds, S. (1997b). Further evidence on recent trends
in the prevalence and incidence of disability among older Americans from two
sources: The LSOA and the NHIS. Journal of Gerontology: Psychological Sciences
and Social Sciences. 52B(2). S59-S71.
Davies, A., & Ware, J. (1981). Measuring Health Perceptions in the Health
Insurance Experiment. Santa Monica: Rand.
De Leon, M., Kasl, S., & Jacobs, S. (1994). A prospective study of widowhood and
changes in symptoms of depression in a community sample of the elderly.
Psychological Medicine. 24. 613-624.
Dennerstein, L., Feldman, S., Murdaugh, C., Rossouw, J., & Tennstedt, S. (1998).
Gender and health issues in aging. Australian Journal of Ageing. 17(1). 19-21.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
295
Dighe, M., Aparasu, R., & Rappaport, H. (1997). Factors predicting survival,
changes in activity limitations, and disability in a geriatric post-stroke population.
Gerontologist. 37(4). 483-489.
Duncan, P. (1994). Stroke disability. Physical Therapy. 74. 399-407.
Easterlin, R., Schaeffer, C., & Macunovich, D. (1993). Will the baby boomers be
less well off than their parents? Income, wealth, and family circumstances over the
life cycle in the United States. Population and Development Review. 19(3). 497-
522.
Evashwick, C., Rowe, G., Diehr, P., & Branch, L. (1984). Factors explaining the
use of health care services by the elderly. Health Services Research. 19. 357-382.
Ewart, C., Burnett, K., & Taylor, C. (1983). Communication behaviors that affect
blood pressure: An A-B-A-B analysis of marital interaction. Behavior
Modification. 7, 331-344.
Farr, W. (1858). The influence of marriage on the mortality of the French people.
In G. Hastings (Ed.), Transactions of the National Association for the Promotion of
Social Sciences (pp. 504-513). London: John W. Parker and Son.
Ferraro, K., Mutran, E., & Barresi, C. (1984). Widowhood, health, and friendship
support in later life. Journal of Health and Social Behavior. 25(3). 245-259.
Ferraro, K., & Wilmoth, J. (2000). Measuring morbidity: Disease counts, binary
variables, and statistical power. Journal of Gerontology: Social Sciences. 55B(3).
S173-S189.
Fried, L., Kronmal, R., Newman, A., Bild, D., Mittelmark, M., Polak, J., Robbins,
J., & Gardin, J. (1998). Risk factors for 5-year mortality in older adults: The
Cardiovascular Health Study. Journal of the American Medical Association.
279(8). 585-592.
Fuchs, V. (1986). Sex differences in economic well-being. Science. 232: 459-464.
Gerstel, N., Riessman, C., & Rosenfield, S. (1985). Explaining the symptomatology
of separated and divorced women and men: The role of material conditions and
social networks. Social Forces. 64. 84-101.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
296
Glass, T., de Leon, C., Marottoli, R., & Berkman, L. (1999). Population based
study of social and productive activities as predictors of survival among elderly
Americans. British Medical Journal. 319(7208), 478-483.
Goldman, N., Korenman, S., & Weinstein, R. (1995). Marital status and health
among the elderly. Social Science and Medicine. 40(12), 1717-1730.
Gove, W., Hughes, M., & Style, C. (1983). Does marriage have positive effects on
the psychological well-being of the individual? Journal of Health and Social
Behavior. 24. 122-131.
Haan, M., Kaplan, G., & Syme, S. (1989). Old observations and new thoughts. In J.
Bunker, D. Gomby, & B. Kehrer (Eds.), Pathways to Health (pp. 76-117). Menlo
Park, CA: Henry J. Kaiser Family Foundation.
Hahn, B. (1993). Marital status and women’s health: The effect of economic
marital acquisitions. Journal of Marriage and the Family. 55. 495-504.
Hayward, M., Crimmins, E., & Saito, Y. (1998). Cause of death and active life
expectancy in the older population of the United States. Journal of Aging and
Health, 10(2), 192-213.
Hillbom, M., Numminen, H., & Juvela, S. (1999). Recent heavy drinking of alcohol
and embolic stroke. Stroke. 30(11), 2307-2312.
Heinemann, G. (1983). Family involvement and support for widowed persons. In
T. Brubaker (Ed.), Family Relationships in Later Life (pp. 127-148). Beverly Hills:
Sage Publications.
Hisnanick, J. (1994). Changes over time in the ADL status of elderly U.S. veterans.
Age and Ageing. 23(6), 505-511.
House, J., Robbins, C., & Metzner, H. (1982). The association o f social
relationships and activities with mortality: Prospective evidence from the
Tecumseh Community Health Study. American Journal of Epidemiology. 116.
123-140.
Hu, Y., & Goldman, N. (1990). Mortality differentials by marital status: An
international comparison. Demography. 27. 133-150.
Hughes, M., & Gove, W. (1981). Living alone, social integration and mental
health. American Journal of Sociology. 87. 48-74.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
297
Idler, E., & Angel, R. (1990). Self-rated health and mortality in the NHANES-I
epidemiological follow-up study. American Journal of Public Health. 80, 446-452.
Idler, E., & Kasl, S. (1991). Health perceptions and survival: Do global evaluations
of health status really predict mortality? Journal of Gerontology. 46. S55-S65.
Johnson, N., Backlund, E., Sordie, P., & Loveless, C. (2000). Marital status and
mortality: The National Longitudinal Mortality Study. Annals o f Epidemiology.
10(4), 224-238.
Joung, I., Vandemheen, H., Stronks, K., Vanpoppel, F., & Mackenbach, J. (1994).
Differences in self-reported morbidity by marital status and by living arrangement.
International Journal of Epidemiology. 23(1). 91-97.
Joung, I., Van Dermeer, J., & Mackenbach, J. (1995). Marital status and health care
utilization. International Journal of Epidemiology. 24(3). 569-575.
Katz, S., Ford, A., Moskowitz, R., Jackson, B., & Jaffe, M. (1963). Studies of
illness in the aged— The index o f ADL: A standardized measure of biological and
psychosocial function. Journal of the American Medical Association. 185. 914-919.
Keith, P., & Lorenz, F. (1989). Financial strain and health o f unmarried older
people. Gerontologist. 29(5), 684-691.
Kriegsman, D., Vaneijk, J., Penninx, B., Deeg, D., & Boeke, A. (1997). Does
family support buffer the impact of specific chronic diseases on mobility in
community-dwelling elderly? Disability and Rehabilitation. 19(2). 71-83.
Kim, J. K. (2000). The effect of marital status on health outcomes among older
persons with stroke and diabetes. International Journal of Welfare for the Aged.
2(1), 3-24.
Korenman, S., Goldman, N., & Hu, Y. (1990). Health and mortality differentials by
marital status at older ages: Economics and gender. Mimeographed document,
Princeton University.
Krause, N. (1987). Satisfaction with social support and self-rated health in older
adults. Gerontologist. 27(3), 301-308.
Krause, N., Herzog, A., & Baker, E. (1992). Providing support to others and well
being in later life. Journal of Gerontology. 47(5). 300-311.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
298
Lean, M. (2000). Pathophysiology of obesity. Proceedings of the Nutrition Society.
59(3), 331-336.
Lee, G., & Dwyer, J. (1996). Aging parent adult-child coresidence-Further
evidence on the role of parental characteristics. Journal of Family Issues. 17(1). 46-
59.
Lillard, L., & Waite, L. (1995). ’Til death do us part: Marital disruption and
mortality. American Journal of Sociology. 100. 1131-1156.
Litwak, E., Messeri, P., Wolfe, S., Gorman, S., Silverstein, M., Guilarte, M. (1989).
Organizational theory, social supports, and mortality rates: A theoretical
convergence. American Sociological Review. 54, 49-66.
Lopata, H. (1973). Widowhood in an American City. Cambridge: Schenkman.
_______ . (1979). Women as Widows: Support Systems. New York: Elsevier.
Maddox, G., & Douglas, E. (1973). Self-assessment of health: A longitudinal study
o f elderly subjects. Journal of Health and Social Behavior. 14. 87-93.
Mare, R., & Palloni, A. (1988). Couple models for socioeconomic effects on the
mortality of older persons. Working paper 88-7, Center for Demography and
Ecology, University of Wisconsin.
Medalie, J., & Goldbourt, U. (1976). Angina pectoris among 10,000: II.
Psychosocial and other risk factors as evidenced by a multivariate analysis of a five
year incidence study. American Journal of Medicine. 60. 910-921.
Miech, R., & Shanahan, M. (2000). Socioeconomic status and depression over the
life course. Journal of Health and Social Behavior. 41(2). 162-176.
Morgan, M. (1980). Marital status, health, illness and service use. Social Science
and Medicine. 14A. 633-643.
Mor-Barak, M., Miller, L., & Syme, L. (1991). Social networks, life events, and
health of the poor, frail elderly: A longitudinal study of the buffering versus the
direct effect. Family and Community Health. 14(2). 1-13.
Murphy, M., Glaser, K., & Grundy, E. (1997). Marital status and long-term illness
in Great Britain. Journal of Marriage and the Family. 59(1). 156-164.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
299
Nagi, S. (1976). An epidemiology of disability among adults in the United States.
Milbank Memorial Fund Quarterly. 6, 493-508.
Nelson, M. (1993). Race, gender, and the effects of social supports on the use of
health services by elderly individuals. International Journal of Aging and Human
Development. 37(3). 227-246.
Penninx, B., Van Tilburg, T., Kriegsman, D., Boeke, J., Deeg, D., & Van Eijk, J.
(1999). Social network, social support, and loneliness in older persons with
different chronic diseases. Journal of Aging and Health. 11(2). 151-168.
Pfeiffer, E. (1975). A Short Portable Mental Status Questionnaire for the
assessment of organic brain deficit in elderly patients. Journal o f the American
Geriatrics Society. 23. 433-441.
Pienta, A., Hayward, M., & Jenkins, K. (2000). Patterns of health and marriage in
later life. Journal of Family Issues. 21. 559-586.
Rauscher, G., Mayne, S., & Janerich, D. (2000). Relation between body mass index
and lung cancer risk in men and women never and former smokers. American
Journal of Epidemiology. 152(6), 506-513.
Reed, D., McGee, D., Yano, K., & Feinleib, M. (1983). Social networks and
coronary heart diseases among Japanese men in Hawaii. American Journal of
Epidemiology. 117. 384-396.
Riessman, C., & Gerstel, N. (1985). Marital dissolution and health: Do males or
females have greater risk? Social Science and Medicine. 20. 627-635.
Robertson, E., & Suinn, R. (1968). The determination of rate of progress of stroke
patients through empathy measures of patient and family. Journal of Psychosomatic
Research. 12. 189-191.
Robine, J. M., Mormiche, P., & Sermet, C. (1998). Examination of the causes and
mechanisms o f the increase in disability-free life expectancy. Journal o f Aging and
Health. 10(2). 171-191.
Rosow, I., & Breslau, N. (1966). A Guttman Health Scale for the aged. Journal of
Gerontology. 21. 556-559.
Rogers, R. (1996). The effects o f family composition, health , and social support
linkages on mortality. Journal o f Health and Social Behavior. 37(4). 326-338.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
300
Rogers, R., Hummer, R., & Nam, C. (2000). Living and Dying in the USA:
Behavioral. Health, and Social Differentials of Adult Mortality. San Diego:
Academic Press.
Rolland, J. (1987). Chronic illness and the life cycle: A conceptual framework.
Family Process. 26. 203-221.
Ross, C., Mirowsky, J., & Goldsteen, K. (1990). The impact of the family on
health: The decade in review. Journal of Marriage and the Family. 52. 1059-1078.
Ruigomez, A., Alonso, J., & Anto, J. (1995). Relationship of health behaviours to
five-year mortality in an elderly cohort. Age and Ageing. 24(2). 113-119.
Schone, B., & Weinick, R. (1998). Health-related behaviors and the benefits of
marriage for elderly persons. Gerontologist. 38(5). 618-627.
Schwarzer, R., & Leppin, A. (1989). Social support and health: A meta-analysis.
Psychology and Health. 3. 1-15.
Schwartz, L., Springer, J., Flaherty, J., & Kiani, R. (1986). The role of recent life
events and social support in the control of diabetes mellitus: A pilot study. General
Hospital Psychiatry. 8, 212-216.
Seeman, T., Bruce, M., & McAvay, G. (1996). Social network characteristics and
onset of ADL disability: MacArthur Studies of Successful Aging. Journal of
Gerontology: Social Sciences. 51B(4). S 191 -S200.jungki
Shenkel, R., Rogers, J., Perfetto, G., & Levin, R. (1985). Importance of significant
others in predicting cooperation with diabetic regimen. International Journal of
Psychiatry in Medicine. 15(2). 149-155.
Sherboume, C., & Hays, R. (1990). Marital status, social support, and health
transitions in chronic disease patients. Journal of Health and Social Behavior. 31.
328-343.
Shye, D., Mullooly, J., Freeborn, D., & Pope, C. (1995). Gender differences in the
relationship between social network support and mortality— A longitudinal study of
an elderly cohort. Social Science and Medicine. 41(7). 935-947.
Silverstein, M., & Bengtson, V. (1994). Does intergenerational social support
influence the psychological well-being of older parents— The contingencies of
declining health and widowhood. Social Science and Medicine. 38(7). 943-957.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
301
Smith, J., & Kington, R. (1997). Demographic and economic correlates of health in
old age. Demography, 34(1). 159-170.
Smith, K., & Waitzman, N, (1994). Double jeopardy: Interaction effects of marital
and poverty status on the risk of mortality. Demography. 31(3). 487-507.
Smith, K., & Zick, C. (1994). Linked lives, dependent demise? Survival analysis of
husbands and wives. Demography. 31(1). 81-93.
Social Security Administration (2000). Income o f the Aged Chartbook 1998.
Washington, D.C.: Social Security Administration.
Soldo, B., Hurd, M., Rodgers, W., & Wallace, R. (1997). Asset and Health
Dynamics among the Oldest Old: An overview o f the AHEAD study. Journal of
Gerontology: Psychological and Social Sciences. 52B(Special Issue), 1-20.
Stroebe, M., & Stroebe, W. (1983). Who suffers more? Sex differences in health
risks of the widowed. Psychological Bulletin. 93. 279-301.
Stull, D., & Scarisbrick-Hauser, A. (1989). Never-married elderly: A reassessment
with implication for long-term care policy. Research on Aging. 11(1). 124-139.
Stump, T., Johnson, R., & Wolinsky, F. (1995). Changes in physician utilization
over time among older adults. Journal of Gerontology: Social Sciences. 50B. S45-
S58.
Tucker, J., Schwartz, J., Clark, K., & Friedman, H. (1999). Age-related changes in
the associations o f social network ties with mortality risk. Psychology and Aging.
14(4), 564-571.
Turner, J. (1986). Sociological theorizing. In J. Turner (Ed.), The Structure of
Sociological Theory (Ch. 1). Chicago, IL: The Dorsey Press.
Umberson, D. (1987). Family status and health behaviors: Social control as a
dimension of social integration. Journal of Health and Social Behavior. 28. 306-
319.
_______ . (1992). Gender, marital status and the social control of health behavior.
Social Science and Medicine. 34(8). 907-917.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
302
Umberson, D., Wortman, C., & Kesseler, R. (1992). Widowhood and depression:
Explaining long-term gender differences in vulnerability. Journal of Health and
Social Behavior. 33(1). 10-24.
Unger, J., McAvay, G., Bruce, M., Berkman, L., & Seeman, T. (1999). Variation in
the impact of social network characteristics on physical functioning in elderly
persons: MacArthur Studies of Successful Aging. Journal of Gerontology: Social
Sciences. 54B. S245-S251.
Verbrugge, L. (1979). Marital status and health. Journal of Marriage and the
Family. 41, 267-285.
Weakland, J. (1977). Family somatics: A neglected edge. Family Process. 16. 263-
272.
Webber, P., Fox, P., & Burnette, D. (1994). Living alone with Alzheimer’s disease:
Effects on health and social service utilization patterns. Gerontologist. 34(1). 8-14.
Williams, R., Barefoot, J., Califf, R., Haney, T., Saunders, W., Pryor, D., Hlatky,
M., Siegler, I., & Mark, D. (1992). Prognostic importance of social and economic
resources among medically treated patients with angiographically documented
coronary artery disease. Journal of American Medical Association. 267(4). 520-
524.
Wolinsky, F., & Johnson, R. (1992). Widowhood, health status, and the use of
health services by older adults: A cross-sectional and prospective approach. Journal
of Gerontology: Social Sciences. 47(1). S8-S16.
World Health Organization. (1977). Manual of the International Statistical
Classification of Diseases. Injuries, and Causes of Death: Based on the
Recommendations of the Ninth Revision Conference. 1975. and Adopted bv the
Twenty-ninth World Health Assembly. Geneva, Switzerland: World Health
Organization.
Zick, C., & Smith, K. (1991). Marital transitions, poverty, and gender differences
in mortality. Journal of Marriage and the Family. 53. 327-336.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
A

Linked assets

University of Southern California Dissertations and Theses

Conceptually similar

PDF

The effects of East Asian values, coping styles, social support on physical health among Korean American caregivers

PDF

Determinants of life satisfaction among the high-functioning elderly: Implications of effects of psychological factors and activity participation

PDF

Out -of -pocket health expenditures by older adults in relation to age, race, and insurance

PDF

Heart disease among middle -aged and elderly persons in the United States: Trends and a multistate model

PDF

A statistical analysis of the formation and location factors of high -tech centers in the United States, 1950--1997: An evaluation using quasi -experimental control group methods

PDF

Evidence of age-based rationing of health care to the elderly: The effect of age and death on the use of health care services by elderly enrollees in an HMO

PDF

Using network perspective to examine the organization of community -based elder care systems across four communities

PDF

Social support, social change, and psychological well-being of the elderly in China: Does the type and source of support matter?

PDF

Impact of housing arrangements on social support and health status among Chinese American elderly

PDF

The outcomes of trade conflicts and negotiations between the United States and South Korea, 1980--2000

PDF

The extent and effectiveness of nursing home regulation in the 50 states

PDF

Effects of a formulary expansion of the use of SSRIs and health care services by depressed patients in the California Medicaid program

PDF

The construct validity of the Health Utilities Index in patients with chronic respiratory disease in a managed care population

PDF

Alcoholism among the elderly: Fact or fallacy

PDF

Education and intelligence test scores: Predictors of dementia?

PDF

Socioeconomic status and the incidence of heart disease among the elderly

PDF

Service evaluation and quality improvement of a geriatric evaluation and management program

PDF

The determinants of child growth retardation in India

PDF

The measurement of physical functioning in a longitudinal study of Americans 70 years of age and older

PDF

The effects of dietary and other lifestyle behaviors on the risk of colorectal cancer among the Mexican -origin Latino population

Asset Metadata

Creator Kim, Jung Ki (author)

Core Title Marriage and health: The effect of marital status on health and the mechanisms by which marital status affects health among elderly people

Degree Doctor of Philosophy

Degree Program Gerontology/Public Policy

Publisher University of Southern California (original), University of Southern California. Libraries (digital)

Tag Gerontology,health sciences, public health,OAI-PMH Harvest

Language English

Contributor Digitized by ProQuest (provenance)

Permanent Link (DOI) https://doi.org/10.25549/usctheses-c16-129103

Unique identifier UC11334724

Identifier 3041484.pdf (filename),usctheses-c16-129103 (legacy record id)

Legacy Identifier 3041484-0.pdf

Dmrecord 129103

Document Type Dissertation

Rights Kim, Jung Ki

Type texts

Source University of Southern California (contributing entity), University of Southern California Dissertations and Theses (collection)

Access Conditions The author retains rights to his/her dissertation, thesis or other graduate work according to U.S. copyright law. Electronic access is being provided by the USC Libraries in agreement with the au...

Repository Name University of Southern California Digital Library

Repository Location USC Digital Library, University of Southern California, University Park Campus, Los Angeles, California 90089, USA