“450' mu: no? "19' 1-3:

A4
nu

in.

~.

‘4“‘6’L‘
1U“:

gill‘m». hm

~. .~

”1

u...

. £2“
"a: ‘

0‘s?"
. 135.
45;;

m

.....
A .

. I:
:3559‘; ".- 3.1:

 

fi;:.?l‘~{“"~: ..

(1 Tf'N‘ .‘ ’.

Ft}: ‘c‘;§‘.s‘.-u 3.5;”
, .

 

1’...‘
’14; a" .u'

- d.
1': $139:

9’
i,
ii

”iii " ..
fifimfd"

ml W‘T D .

i‘

.1
"’fnfvr U:

y :‘f ‘1'}: 'l A
S V “A 4 ,l. ", l .
' E. ‘1 “'4";

. . - 36’ , ,
gzézaé .3 * 1. w
‘ "ft-"{5}; ‘3.-
'. u

:1.
Kit Jug: r.

..i

5‘.

a.

Hug-J". ‘33.. sA‘L
‘x‘JLak' $13." u‘v.

5.1%....
a}? $9.1M
.4- _ "3;" J}?
‘ l m 1‘9 ‘~
. ". mu wiicivfl
‘ ’,;~f~”§$}.:'7"?”';§95‘£§—
* ‘ ‘ 3‘ . "r.

. fl
unwrig'fi ...
. .-.~:.~.~.

~3er 1' 'c. .151; ‘3‘
.9...

‘n‘.

“ L .-
w YE}; 2. .5 guégg;

Jud?“ Myflmw

 

‘ r
m . 9,11%.

 

X‘J‘e‘ié‘:

‘T
1‘).
a.

i“:-
‘1? ":‘I

 

“
-\ “a.
.ugnflé’ 31" 335-. "

l‘ 1421'
.. "’1"
“A o

-

<w-hw‘n
.3“ .. o “i.

:m‘fi~“

 

- u .:it(- 0»
Ta»

‘n‘udmhf";

 

-: “4}
.f , up}; .’
'"2‘;fi~1.- .‘
.‘ul‘h‘llt
. .s
h...
"u .u -

‘5’ lid

 

:f‘. A‘ -:i. c {ig‘ivfa‘
«#1 1623mm... 4

CH GAN STATE UNIVERSITY LIBRARIE

WWW WWll\W.\\\\|\\\\\\\\\\\\\\\\i\|

3 1293 00891 8330 ‘

This is to certify that the

dissertation entitled

THE ABLE ELDERLY:
MEMORY, HEALTH, SOCIAL SUPPORT
AND

PSYCHOLOGICAL WELL-BEING

presented by

Ann Marie Rosavage

has been accepted towards fulfillment
of the requirements for

Ph. D. ‘degree in Psychology

 

flaw in,

Major professor

Date l°/'°_ 10

 

 

 

 

'1:

T LIBRARY
Niobium State

Unlverst
1L F W l

 

 

 

PLACE IN RETURN BOX to remove this checkout from your record.
TO AVOID FINES return on or before due due.

DATE DUE DATE DUE DATE DUE

JUNllza‘sa, l 1

55.;ch a. 9 2.000

—=_fi7

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

fl

MSU Is An Affirmative Action/Equal Opportunity Institmion ‘
cmmud

 

 

 

 

THE ABLE ELDERLY:
MEMORY, HEALTH, SOCIAL SUPPORT
AN D
PSYCHOLOGICAL WELL-BEING
BY

Ann Marie Rosavage

A DISSERTATION

Submitted to
Michigan State University
in partial fulfillment of the requirements
for the degree of
DOCTOR OF PHILOSOPHY
Department of Psychology

1990

ABSTRACT

THE ABLE ELDERLY:
MEMORY, HEALTH, SOCIAL SUPPORT
AN D
PSYCHOLOGICAL WELL-BEING

BY
Ann Marie Rosavage

This research investigated the impact of memory complaints, memory
function, perception of health, actual health, and social support on the
psychological well-being of able elderly, those living independently in the
community. The underlying model was M. Powell Lawton‘s
conceptualization of "The Good Life" for this population. Results strongly
suggest that W as measured by quality of support,
perception of health, and memory complaints rather than Mill
Competence as measured by actual health status and memory function is the
best predictor of W; More specifically, in terms of
social support, the presence of confidants and emotional support were the
best predictors of psychological well-being. Both males and females were
more likely to list their spouse as the provider of emotional support;
however, when asked to name individuals not living with them, participants
indicated that daughters, female friends, and sons were the ones they talked
to most about their troubles or concerns. In addition, females were more
likely to name confidants than males. Memory complaints appear to have a
direct effect on psychological well-being, with memory function having an
indirect effect through its relationship to memory complaints. Although
perception of health and actual health were significant predictors of

 

psychological well-being, perception of health was the best indicator. In
addition, college attendance moderated the effect of perception of health and
memory complaints on psychological well-being. Thus, elderly with more
education, better health, and fewer memory complaints appear to lead
psychologically healthier lives than those with less education, poorer health,
and more concerns about their memory.

The virtual absence of depression or severe impairment of memory
function in the participants suggests that this is a representative sample of
able elderly. The majority viewed their health positively, drove cars, left
their neighborhoods frequently, and had a supportive social network. Even in
this relatively healthy sample, however, perception of health, concerns about
memory, and amount of emotional support were significant predictors of
psychological well-being. Based on the findings of the present study, further
research on what constitutes "The Good Life" for the able elderly appears

warranted.

 

 

To Michael

iv

ACKNOWLEDGMENTS

This dissertation would be incomplete without acknowledging the
support I received from my professional and personal social networks. First
of all, I would like to thank my advisor, Dr. Norman Abeles, for his
continued support throughout my graduate career at Michigan State
University. Dr. Abeles was not only the guiding force behind this dissertation
but also served on my Master's thesis committee, was my clinical supervisor,
and gave me the opportunity to have a significant role in the management of
the Psychological Clinic.

I would also like to express my appreciation to Dr. Bertram Karon, not
only for his support of this dissertation, but also for supervising my analysis
of D. H. Lawrence and thereby giving me the opportunity to combine my love
of literature and psychology.

I would also like to thank Dr. Ralph Levine for understanding my
initial apprehension of the statistical analyses required for this dissertation,
and for his willingness to work with me on completing this portion of the
research before he left for sabbatical.

Dr. Ellen Strommen deserves a special thank you for her invaluable
input regarding the social support aspect of the dissertation. In particular, Dr.
Strommen's interest in the role of friends and siblings throughout the
lifespan helped clarify my thoughts and conclusions regarding the findings in
this research.

A special acknowledgment goes to the original coordinator of this
project, Janice Gutfreund, whose contributions include reviewing articles,
organizing the questionnaires, training assistants, and interviewing
participants. In addition, not only is Janice a good friend, but her invaluable
knowledge and expertice about issues related to aging helped clarify my
thoughts and ideas for the research.

I also would like to acknowledge the many elderly individuals who
agreed to participate in this research. Without them none of this would have
been possible. Their zest and enthusiasm for life have shown me that aging
can be a positive, fulfilling part of my life as well.

A very special thank you goes to my children, Stephanie and Dana
Brail, for their love and emotional support throughout the years I spent in
graduate school. I will always remember Dana's "hang in there" note during
an all-night paper session, and Stephanie's shout of "awesome" when I called
to tell her I had passed my orals.

F’mally, I appreciate the support from so many friends. In particular,
Elizabeth DeRath for her unfailing humor and friendship; Ann Holt for her
love and support, as well as for our shared addiction to mystery novels which
gave me many enjoyable breaks from this project; Maureen Marks whose
friendship proves that age makes no difference; Dan Mosher who has been a
supportive friend since our first days together in graduate school; and most of
all, Michael Lindell for the countless hours of invaluable discussions about
the dissertation, but mainly for patiently listening to my many episodes of

anxiety and quietly reminding me each time: "You can do this."

vi

TABLE OF CONTENTS

Page
LIST OF TABLES ................................................................................................... ix
LIST OF FIGURES ................................................................................................. x
INTRODUCTION .................................................................................................. 1
LITERATURE REVIEW ...................................................................................... 3
Evidence of Depression. ................................................................................. 3
Physical Health and Depression ................................................................... 4
Memory and Depression. ............................................................................... 6
Social Convoys, Social Support and Depression ...................................... 9
Confidants and Depression ........................................................................... 1 1
HYPOTHESES ........................................................................................................ 17
METHODS ............................................................................................................. 20
Subjects/ Procedures ....................................................................................... 20
Measures ........................................................................................................... 21
RESULTS .............................................................................................................. 25
Subject Characteristics .................................................................................. 25
Construction of Scales .................................................................................... 27
Perceived Quality of Life ................................................................................ 28
Behavioral Competence ................................................................................ 30
Psychological Well-being ............................................................................... 30
Bivariate Analysis Results ............................................................................ 30
Gender Differences ......................................................................................... 38
Multiple Regression Results ........................................................................ 39
DISCUSSION ........................................................................................................ 48
Subject Characteristics .................................................................................... 49
Bivariate Results ............................................................................................. 53
Gender Differences .......................................................................................... 59
Multiple Regression Results ......................................................................... 61
Conclusion and Implications ........................................................................ 67

vii

Page
70

APPENDD< A. ........................................................................................................
APPENDD( B ......................................................................................................... 72
REFERENCES ........................................................................................................ 73

viii

LIST OF TABLES

Table Page

1. Prediction of intercorrelation of perceived quality
of life, behavioral competence and psychological

well-being in 165 older able adults ............................................................... 19
2. Social Support Variables ................................................................................. 29
3. Bivariate Analysis Results ............................................................................. 36

4. MultipienRegression Analyses with Psychological
Well- ' g as the Dependent Variable ....................................................... 45

5. Multiple Regression Analysis with Emotional Support
as the Dependent Variable ............................................................................. 47

ix

LIST OF FIGURES

Figure Page
1. Multiple Regression Model ........................................................................... 40

INTRODUCTION

It has been suggested that the "dominant popular image of old age is
one of inevitable decline, senility, and dependency" and that the word aging
is associated with the word "problem" (Hagestad, 1987). Hagestad further
states that, although it would be ill advised to ignore the burdens that such
illnesses as Alzheimer's disease place on individuals; nevertheless, the
majority of the elderly are a "vastly underutilized social resource." Even in g
advanced old age, these elderly not only function independently, but can be
positive contributors to society. Kimmel (1988) notes that, while other
cultures such as the Japanese culture see aging as "positive," our culture sees
it as "negative, or at best ambivalent" (p.175). He further states that the focus
of research on aging should be on "success " aging, thus eliminating "global
stereotypes" that emphasize average declines in such areas as physical and
cognitive functioning (p.177). Given that only 5% of Americans are in
nursing homes (Shanas, 1979) and only 5-10% of Americans suffer from
senile dementia of the Alzheimer's type or any other kind of senility
(Storandt, Botwinick, & Danzinger, 1987), focusing research primarily on the
negative aspects of aging clearly neglects the majority of this population.

In addition to the above, increased longevity among elderly
individuals suggests that when researchers report their findings, they
differentiate between the young-old (early 60's and 70's),old-old (late 70's and
early 80's), and very old (late 80's and above) . As Schaie remarks in his 1988
W article on ageism in psychological research: "Because
of our increased longevity, any research findings that are based on a sample

1

2

described as 'over 60' would have as little validity as a study of child
behavior that described its subjects as being 'under 12'" (p.180). In addition,
demographic data such as gender, education, health, and family status are
crucial in research of the elderly because of their "great diversity" and
individual differences (Schaie, 1988).
If researchers are to focus on "successful aging," M. Powell Lawton's
(1986) conceptualization of "The Good Life" for this population bears
consideration. In order to experience this "good life," the elderly need to feel
positive about four separate domains: 1) mm which
includes physical health, functional health (i.e., degree of disability), cognitive
competence (i.e., memory), and time use and social interaction; 2)
WW which encompasses self-esteem and positive mental
health; 3) W as measured by satisfaction with income,
housing, neighborhood, marriage, family, friends, etc.; and 4) M3
W which is anything outside the person that can be measured or
evaluated with high consensus by a majority of judges, such as housing or
neighborhood quality (p.41-43). According to Lawton, the good life can be
improved by increasing the satisfaction in any of these four sectors. In
addition, although invariably by increasing satisfaction within one sector,
other sectors will be positively affected, it is entirely possible that there may
not be corresponding changes in any or all of the other sectors (p.44).
The purpose of the present research will be to examine aspects of

Lawton's "Good Life," with emphasis on the following three factors:

51° IE I E . 19 ll “1de 11.1”»
king. Bflam‘ :31 ngmtgmg will be measured by memory function and

health status. W will be measured by social
interaction, perception of health, and memory complaints. Psychological

3

mugging will be measured by the amount of depression and other mental
health symptoms experienced by a sample of able elderly adults living
independently in the community.

LITERATURE REVIEW

1. WM

One of the symptoms found to be the most prevalent in the elderly is
depression (Blazer, Hughes, 8: George, 1987). However, findings from the
Epidemiological Catchment Area Project (ECA) seem to indicate that the
prevalence of major depressive episodes in the elderly is lower than in
younger age-groups (Myers et al, 1984). Blazer and Williams (1980) reported a
less than 2% prevalence of primary or secondary depressive disorder in a
sample of almost 1000 community based older adults; however, they also
found that nearly 15% reported "significant dysphoria." Blazer, Hughes, and
George suggest that this discrepancy is likely due to the restriction of
traditional DSM III (Diagnostic and Statistical Manual-APA, 1980) diagnosis:

The diagnostic category of major depressive episodes
captures only the pure and severe depressive disorders, such as
melancholic depression and affective psychoses. Other severe
depressive episodes may be secondary to concurrent psychiatric
disorders, such as organic mental disorders and
schizophreniform disorders. Depressive symptoms also derive
from less severe depressive disorders, such as dysthymia
(depressive neurosis), or from physical illness and other life
stresses. Finally, even relatively severe depressive symptoms
may fall through the typical affective diagnosis and therefore be
classified as atypical depression in DSM III (for example,
intermittent but severe depressive episodes or mixed depressive
and anxiety episodes) (p. 2132).

 

4

In the previously mentioned study which looked at the epidemiology
of depression in an elderly community population, Blazer, Hughes, and
George (1987) analyzed data from 3000 interviews in 5 geographically
distributed communities in the United States. Their results include the
finding that 8% of the entire elderly sample had a current major depressive
episode, dysthymia, mixed depression and anxiety syndrome, or symptomatic
depression. An additional 19% had less severe symptomatology yet still
exhibited signs of decreased life satisfaction (pp. 283-284). Their conclusion,
therefore, is that traditional DSM III categories fail to identify most depressed
older adults in the community. Furthermore, if we look at the four categories
described above, the 8% prevalence found in their analysis will be "similar to
previously reported frequencies in epidemiologic studies of the severely
depressed elderly population" (p.286).

Given the above data, therefore, depression appears to be a significant
variable that needs to be examined when conducting research on the able

elderly, those functioning within communities.

2. El . l H ill I D .
The relationship of physical health to psychological well-being in the
elderly seems irrefutable. In numerous studies, a strong correlation has been
demonstrated between physical illness and depression in the aged (LaRue,
Dessonville, & Jarvik, 1985). For example, in an elderly study of 231 patients
admitted to the hospital for a "manic or depressive symptom complex," 50%
of the men and 42% of the women with recurrent affective disorders had
"significant medical conditions." In addition, 71% of the men and 65% of the

women who had no previous history of affective disorders were suffering

 

5
from some kind of other medical or health problem. In addition, the men

were more likely to have undergone operations prior to their psychiatric
admission and had unusually high rates of chronic illnesses such as urinary
or cardiovascular conditions and were more likely to have multiple medical
problems (LaRue et al, p. 672).

Other studies have also linked physical health and depression in the
elderly. Murrell, I-Iimmelfarb and Wright (1983) found that physical health
had the strongest association with depression. Rozzini, Bianchetti,
Carabellese, Inzoli, and Trabucchi (1988) examined over 1200 men and
women in northern Italy between the ages of 70 and 75. They found that
depression was linked to somatic symptoms which were in turn linked to
physical impairment. Blazer and Williams (1980) report that 44% of the
respondents in a survey of over 900 people aged 65 and above who had
depressive symptoms were medically ill. Murphy (1982) compared elderly
depressed subjects and normal elderly people in the general population and
found an association between poor physical health and the onset of
depression. In a later follow-up of these 124 elderly depressed patients,
Murphy (1983) further elaborated on the relationship of physical health and
depression. Her findings strongly support the theory that physical health may
influence the outcome of depression through its meaning for the individual
rather than through a "direct organic effect" (p.117). Along similar lines,
Stroller (1984) found that psychological distress was more highly correlated
with self-rated health among older adults than actual physical health status.

Finally, a strong link has been found between self-evaluation of poor
health and loneliness in the elderly (Ryan & Patterson,1987). This finding

assumes even greater significance when coupled with the fact that other

6

research has demonstrated a relationship between loneliness in the elderly
and depression (Berg, Mellstrom, 8: Persson, 1981). .

Given the above data linking physical health to depression in the aged,
research on the psychological well-being of the elderly needs to include
measures of perceived health status of the elderly sample, especially as
compared to others in their cohort.

3. Memmandflenressicn

It has been estimated that half of all persons over age 60 complain of
memory dysfunction, with dementia the commonest cause of such difficulty
(Ciocon Gt Potter, 1988); nevertheless, memory decline can also occur in
depression which later reverts or improves with treatment for the depression
(e.g., McAllister & Price, 1982). In addition, cognitive impairment has been
found to be present only in those depressed individuals who are over 60 years
of age, with severity of depression positively correlated with increased
cognitive dysfunction (Cavanaugh 6: Wettstein, 1983). According to Albert
(1981), difficulty in memory ability is probably the chief complaint that results
in neuropsychological referral. There is a substantial body of research
literature which documents the fact that memory processes can be
functionally impaired by depression (Salzman & Gutfreund, in Poon, 1987).
Examples of functional impairment found in research studies include:
Decreased acquisition and recall of new information (Cohen, Weingartner,
Smallberg, Pickar, & Murphy, 1981) and lowered performance on tests of
consolidation or long-term memory (Glass, Uhlenhuth, & Weinreb, in
Williams, et al., 1987). In addition to research data, Salzman and Gutfreund
point out that clinicians who work with the elderly report that many older

 

7

patients with previous intact cognitive abilities seek medical attention for
memory loss and difficulty thinking clearly following the death of a spouse or
lifelong friend.

One of the most serious consequences of this symptomatology is the
necessity of differentiating between a diagnosis of depression or dementia.
The similarity between the cognitive deficits of dementia and depression has
prompted researchers to coin the term "pseudodementia" meaning "false
confusion" for cases of depression that present as possible true dementia
(Ronsman, 1982). However, as Salzman and Gutfreund (In Poon, 1987) state:
"...pseudodementia is neither pseudo nor dementia. Rather, it is true
impairment of memory secondary to depression without impairment of
other aspects of the mental status." Abandoning the term "pseudodementia,"
therefore, has been suggested by several authors (p. 260). Others have felt that
the main value of this term is that it cautions practitioners to make careful
evaluations of affective states before diagnosing individuals as demented
(Small & Jarvik, 1982).

In contrast to actual deterioration in memory functioning in the
depressed elderly, some researchers have reported that memory complaints,
while correlated with depressed mood, do not necessarily correlate with
memory test scores (Kahn, Zarit, Hilbert, & Niederhe, 1975; Perlmutter, 1978).
Kahn et a1 examined a group of elderly diagnosed as having senile dementia
and found that those subjects without severe memory loss, but who were
depressed, were more likely to complain of poor memory than those who
were not depressed but had suffered significant memory deterioration.

Using a structural interview, tests of memory, and self-report scales for
affective status, Zarit (I982) sampled 79 adults ranging in age from 50 to 82
who were relatively well-educated (mean years of education was 13.25) and

8
not diagnosed as suffering from senile dementia. His results showed that

individuals who had memory complaints generally saw themselves as
having difficulty in virtually all areas, such as names, faces, word lists,
directions, and visual recall. He also found that individual items on the
memory complaints structured interviews did not correlate with
performance on measures of recall. On the other hand, total complaints had
strong positive correlations with both affective measures [the lung
Depression Scale (Zung, 1965) and the Brief Symptom Inventory (Derogatis,
Lipman, Covi, Rickels, at Uhlenhuth, 1970)]. He did find that self-evaluations
of memory were related to actual performance but was particularly impressed
by the finding that, even after controlling for effects of performance, self-
evaluations of memory were related to affective status. He concluded,
therefore, that the relation of depression to the older individual's subjective
complaints of memory disturbance needs to be given proper attention in
further research of this population.

Williams, Little, Scates, and Blockman (1987) in a study of depressed
and non—depressed adults over age 40, using a battery of memory tests, the
Beck Depression Inventory (Beck et al, 1961) and a memory problem
questionnaire found a significant difference between these two groups on
memory tests. They found cognitive deficits in concentration or effort in the
depressed sample which "mildly" reduced performance in tests of verbal
learning and recall. However, even though the depressed subjects performed
worse, all subjects scored in the average—to-superior range. These results are
consistent with other research which found substantially unimpaired
performance for depressed subjects (e. g., Gass 8t Russell, 1986). In addition,
similar to Zarit (1982), Williams et al report that depressed subjects

complained of a high incidence of memory problems in virtually every

9

content area, such as old phone numbers, jokes, correspondence, names,
appointments, forgetting what you're doing, etc. In the overall rating,
depressed subjects indicated they had "moderate-to-severe" problems with
memory; whereas, non-depressed subjects reported "minor-to-no problems"
with memory. They concluded their research with the recommendation that
further studies to differentiate dementia from depression include an
assessment of memory complaints as well as memory test performance.
Based on the above findings, it seems clear that depression affects, at
the very least, subjective perception of memory ability in the elderly and, in
addition, actual memory function as well. However, individuals may still
score within the normal range of functioning. Continued investigation of

this phenomenon appears warranted.

4-5 .1: S .15 | “2 .
One of the psychosocial factors that has been shown to be a predictor of

emotional health status or psychological well-being is the support found
within social networks (Israel, 1982). According to Blazer (1983), study of the
social network of older depressed adults is especially critical because the
networks of the older adult are "often more fragmented and less supportive
than at other stages of life" (p. 162). It is important, however, to distinguish
first between social networks and social support. Israel states that a social
network is "person-centered" and refers to the W in the overall
network (e. g. size), M - nature of the linkages themselves (e. g.,
frequency and intensity of interaction), and fungflgns that networks provide
(e. g., affective support). Thus, according to Israel "...a social network refers to

human interactions, some or all of which may or may not provide social

10

support" (p. 65). Other researchers have used the term "convoy" which is
similar to "network" but refers to the "dynamic system of relationships that
envelope an individual over a lifetime" rather than one specific time
(Depner & Ingersoll-Dayton, 1988; p.348). In addition, although convoys
usually provide support, interactions do not necessarily result in such
support. Depner and Ingersoll-Dayton further suggest that m (the
presence or absence of social relationships), finds of relationships (children,
spouses, friends, etc), ms of support (e. g., affective or emotional), and
m of support all be examined empirically.

According to Rivicki and Mitchell (1986) much of the research on social
support and the elderly has focused on quantity measures (i.e., how many
individuals in the convoy) rather than quality of (i.e., satisfaction with)
interpersonal relationships. They suggest that this is unfortunate because, in
fact, the size of one's convoy does not necessarily co-vary with the
"sufficiency" of social support (p. 233). In this regard, it is useful to look at
affective or emotional support because this type of support appears to be an
important determinant of health and well-being in older people. For
example, Krause (1987) found that those older adults with more emotional
support tended to report fewer symptoms of depression than those elderly
with less emotional support. Also, individuals who were W with
the emotional support from their convoy reported less favorable overall
health status and more psychological distress. In addition, Krause found that
frequency of support (quantity) was not associated with self-rated health;
however, satisfaction with (quality) social support m significantly related to

this variable.

ll

5. W211

One form of emotional support which may be especially important is
the existence of a confidant- someone with whom individuals can share
their troubles or concerns. Elderly people without such a confiding
relationship have been found to be at higher risk for depression than elderly
who have such a relationship (Esser 8: Vitaliano, 1988; Murphy, 1982; LaRue
et al, 1982). Murphy (1983) suggests that the elderly in poor health who have
a confidant are less vulnerable. to depression. In the follow-up study of
depression cited above, Murphy found that depressed individuals in the poor
income group were more likely to report no confidin g relationships. Finally,
Surtees (1980) found that a close confiding relationship buffered individuals
against the effects of social adversity.

Since confidants appear to be significant members of elderly
individuals' social convoys, it is important to look at who these individuals
are. It is generally accepted that social convoys consist of spouses, children,
siblings, other relatives, and friends (Depner & Ingersoll-Dayton, 1988). Ward
(1985) maintains that children typically "dominate" the support networks of
older individuals. Depner and Ingersoll-Dayton also assert that it is family
members, especially children, who are cited by the elderly as primary sources
of support. However, these children may not be confidants but may provide
other kinds of support such as health care or financial assistance.

In a study by Babchuk (1978) on aging and primary relationships,
married respondents rarely included all of their adult children as confidant
relatives. In fact, very few listed any young relatives in this capacity.
Research seems to point strongly to the fact that there is little or no significant
relationship between the emotional well-being of elderly persons and the

1 2
frequency with which they interact with their adult children (Lee &
Ellithorpe, 1982; Spreitzer & Snyder, 1974). Arling (1976) looked at the
relationships of the elderly widow to her family, neighbors and friends and
found no relationship between frequency of contact with children, or children
living nearby, to morale. Interestingly, visiting other relatives, such as a
sibling, grandchild, niece, or nephew, slightly reduced their feelings of
loneliness. In another study looking at social interaction, loneliness and
emotional well-being among the elderly, interaction with adult children and
grandchildren had only marginally significant correlations with loneliness
(Lee & Ishii-Kuntz, 1987). Perhaps, as stated above, elderly adults turn to their _.

 

children for other sources of support. For example, some studies have shown
that older adults first turn to daughters and then to sons for assistance when
they are ill (Wood & Robertson, 1978). Finally, there may be a gender
difference in the elderly and the amount of support they receive from adult
children. Depner and Ingersoll-Dayton (1988) found that women received
more emotional and health support from their children than did men.

Other relatives may also be important sources of emotional support.
For example, according to Shanas (1980), increased longevity has resulted in
many elderly having brothers and sisters still living. These individuals could
be additional sources of emotional support or confiding relationships.
Shanas reports that three-fourths of never-married elderly adults in her
sample had seen a brother or sister the week prior to being interviewed. Also,
Depner and Ingersoll-Dayton (1988) found that the lower number of siblings
in the convoys of older adults was associated with decreased exchange of
emotional support. Wood and Robertson (1978) suggest that the sibling
relationship may be "especially important if other age peers are unavailable"
(p. 372-373). As a result of reviewing studies on the role of siblings in later life,

l3

Cicirelli (1982) concluded that "the majority of siblings tend to remain
affectively close throughout adulthood and old age, and may grow still closer
in the last decades of life" (p.276). Finally, in more recent research, Cicirelli
(1989) found that closeness to a sister was related to less depression in both
males and females. In addition, if women perceived "conflict or indifference"
in their relationship with sisters, depression increased.

In terms of spouses as confidants or sources of emotional support, clear

gender differences appear in the research literature. Married males appear to

confide in their wives and rely on them for emotional support (V aux, 1985;
Preston & Grimes, 1987); whereas, women tend to list other individuals such
as friends. Babchuk (1978) in his research on aging and primary relationships
found that a very high proportion of the married men in his sample were on
confiding terms with only one individual, their wives. Many of the men
listed their children as primary but reported that they did not confide in them
or in any other relatives with whom they felt close (p. 147).

One additional source of emotional support and confiding
relationships for the elderly can be found in their interactions with friends.
Lee and Ishii-Kuntz (1987) found that, with respect to loneliness, having
friends was more important than being married. These researchers have
theorized that the symbolic meanings of the interaction with friends versus
relatives may be different and thus affect the psychological rewards received
from these sources. Adult children or other kin may feel obligated to interact
with the elderly adult. Relatives may not perceive one another as desirable or
even the relationship as worthwhile, yet they may feel required to maintain
ties with each other. Some sociologists suggest that if older persons have a
single good friend, they are more able to cope with old age than those with "a
dozen grandchildren but no peer group friends" (Wood & Robertson, 1978; p.

14

367). Arling (1976) in his study of the elderly widow found that widows who
had neighbors or friends they could visit were less likely to feel lonely. What
is also interesting is that frequency of contact with such individuals did not
appear significant; rather, having them available per se elevated morale.

Babchuk (1978), in his study of 800 elderly individuals, found that
almost 700 of them listed primary ties to friends, with many of these friends
going back to childhood. Eighty-seven percent of those cited as primary
friends had been close to the respondent for at least three years. Half of the
sample reported seeing one or more of these friends weekly. What is also
noteworthy is the finding that single individuals, rather than the married
elderly, were more likely to claim not having at least one close friend. Age
was also significant. For example, individuals over the age of 70 were more
likely to be without a primary friend. In addition, gender differences were
noted by Babchuk. Women were more likely to have confidant friends than
men. Sixty-seven percent of women had at least one confidant friend as
compared to 54.2% of the men (p. 147).

Vaux (1985) in his review of social support, age, gender differences, and
ethnicity reports that gender differences are much more likely to emerge in
support from friends than from family. Antonucci and Akiyama (1987) in a
sample of 214 men and 166 women ranging in age from 50 to 95 found that
women consistently report receiving more support from both children and
friends. These differences were most evident in the affective or emotional
area such as "confiding, reassuring, and talking when upset" (p. 744). Finally,
in a sample of 234 individuals 70 years of age or older, Powers and Bultema
(1976) found men had more frequent social contacts; however, a smaller
proportion of these were classified as intimate friends in comparison to

women whose contact with intimate friends contributed significantly to their

15

total interaction scores. These researchers concluded that the social worlds of
aged men and women are vastly different, with men not only limiting their
interaction to less intimate friends and family but also less likely to replace
lost friends. Women, on the other hand, have a "diverse social world and
many intimate ties outside the immediate family" (p. 746).

One other variable that is important to consider when looking at
emotional support is marital status. As already alluded to above, research
suggests that married individuals have more primary relationships than
unmarried. Longino and Lipman (1981) found that not only do married
elderly have more primary relationships, there is also a gender difference,
with married women having the most, and spouseless men the least.

Finally, types of contact, such as letter writing, phone calls, or face-to-
face contact as well as physical distance can affect psychological well-being.
Revicki and Mitchell (1986) found that more frequent telephone contacts
between elderly individuals and family or friends led to reports of fewer
mental health problems, greater life satisfaction and self-esteem, and
concluded that telephone contact is an important source of social support for
the elderly. With this fact in mind, the research of DeWit, Wister, and Burch
(1988) on physical distance and social contact between elders and their adult
children is noteworthy. They found that physical distance is a "potent
determinant" of several types of social contact. For example, at proximate
distances, face-to- face and telephone conversations occur frequently; whereas,
as physical distance gets farther than 1 / 2 day's drive, face-to-face contact
decreases. Also, females have considerably more telephone conversations
than males. Female children are also more likely to contact parents than
male offspring.

l 6
In conclusion, given the fact that social support, and specifically,
emotional support has been shown to be a predictor of emotional health
status or psychological well-being, further investigation of the social convoy
of the elderly with emphasis on the role of confidants appears warranted.

HYPOTHESES

In light of the above literature review on depression, physical health,

memory functioning and complaints, and social support in the able elderly,

the following hypotheses are postulated:

1.

Both perception of health and actual health will have an effect
on psychological well-being; that is, positive perception of health
as well as better overall health will be positively correlated with
psychological well-being.

Memory complaints will have a stronger relationship to
psychological well-being than will memory function; although
there will be a significant correlation between memory function
and psychological well-being.

The presence and perceived helpfulness of a confidant will act as
a buffer or moderator between perception of health and
psychological well-being.

Friends and siblings will provide more emotional support than
children, especially for females.

Social support variables and their effect on psychological well-
being will be in the following order of importance:

Emotional support > number of confidants > number of phone
calls per year > number of visits to and from individuals in the
social network per year > total number of persons in the network

> physical distance from individuals in the network.

17

l 8
In addition, gender differences in type of confidant is hypothesized;
that is, when asked to name one confidant, females and males will differ in
whom they believe is someone they can talk to about their troubles or
concerns:
FEMALES: Friends > Spouse > Siblings > Children
MALES: Spouse > Children > Siblings > Friends
Finally, Table 1 lists the expected significant correlations between such
variables as psychological well-being, social support, and perception of health.

19

.32 8. @ Became u ..

coma—9:00 PEN u c

coca—3.80 95392 n -
coma—8.80 Esteem u +

*+

*+

333
6.5

*+

e+
*+
o

t...

.mmsm
Etom

mmD

.3:an

.3500
to: 508m: fog—oz

e... e... o
e+ e... o
c- - o
+ + o
e- e- c

c+ o
e+ +
o +

dogma 5.3m
5:82 323. Baum

meshes:
.. “83205me

e + «comm—3m 38m
+ swam: 5:333:
.. c2655 E252
+ $529530 foam—2

.. acumeoumm 53mm

.- swam sacs.
.. con—833m
.835 we ewe:
scram
«o mw<

$50.4. Bm< mmDAO mow 2H Damn—Jam; A<U~UOAOIU>mm

OZ< mUZmHmn—EOU AgOgémm .mBA m0 §<DO Dm>~mUMmm mO ZOPSmMy—Oumm—E m0 20:80.53

a mam—<8

METHODS

W

One-hundred sixty-five participants aged 55 and older were recruited
from senior citizen groups, senior nutrition sites, plus retired faculty and staff
from Michigan State University (MSU). Phone calls and subsequent site
visits were made to the senior citizen groups and nutrition sites. A brief 20
minute explanation of the project was presented and individuals were
invited to sign up for a structured interview, memory test battery, and mental
health assessment inventories. Follow-up phone calls were made to schedule
actual appointments. In lieu of signing up after the presentation, individuals
were given the option of contacting the coordinator by phone and scheduling
an appointment at that time. Retired staff and faculty from MSU were sent
flyers describing the project along with a cover letter which instructed them
to call the MSU Psychological Clinic to arrange appointments if they were
interested in participating in the research (see Appendix A).

Individuals who agreed to participate either came to the MSU
Psychological Clinic or a home visit was arranged if travel to the university
was either not possible or not desirable. Interviews and tests were
administered by trained undergraduate and graduate students from the
Department of Psychology at MSU. All interviews were tape-recorded and all
participants signed a permission form which gave them the option to drop
out of the research at any time (see Appendix B). The entire structured
interview, memory tests, and mental health assessment inventories took

approximately 1-1 1 / 2 hours to administer.

20

21

In addition to the above, feedback either in person or by telephone was
given to every participant in the research. This feedback was designed to give
information regarding performance on the tests of memory function;
however, if significant depression or other mental health concerns were
found, this was also conveyed to the individual along with proper counseling
referrals. If memory tests indicated a possible problem, participants were
advised to seek medical attention and/ or were asked to come back for
retesting in 8 monthsto 1 year.

Measures

1) MW (MAI)

(Lawton, M. P., Moss, M., Fulcomer, M., Kleban, M. H., 1982).

The MAI is an assessment instrument capable of measuring the well-being of
the aged in the areas of behavioral competence (health, cognition, time use,
social interaction, etc), psychological well-being, and perceived
environmental quality. The performance of 590 older people in groups
composed of community residents, in—home service clients, and people
awaiting admission to an institution was determined. All respondents were
interviewed in their homes. Interviewers received three days of intensive
training.

W (of summary rating scales/ full length domain indices and
subindices): Summary rating scales in seven domains were completed by
using an interviewer and a "reader-rater" for 484 of the 590 respondents and
by an interviewer and interview observer for the remaining 106. According
to Lawton et al,: "In the sample of 484, interviewers and reader-raters agreed
with either a 0 or 1 point discrepancy in 95% of all instances: intraclass

correlations ranged from .88 (Activities of Daily Living) to a low .58 (Social

22

Interaction). Reliability (Alpha) of the final MAI scales were all .61 or above
except for health behavior (.39) and personal security (.57). Retest reliability,
done at a 3 week interval on 22 cases, was acceptable, with the exception of the
physical self-maintenance subindex, where the variability was very low, the
majority receiving a perfect score." (p. 95)

mm was determined by doing summary ratings, multiple
correlations, and by constructing a "dummy variable" representing
independent versus dependent groups. The only domain that appeared
somewhat problematic was the social domain. Lawton et al conclude that the
reliabilities and validities of the MAI indices and subindices were affirmed by
several different approaches.

W: Because analysis has been done separately for
each domain, portions of the MAI need not be used if desired. The authors
maintain that physical health, cognition, and activities of daily living are the
strong domains. They acknowledge that the social interaction scale is
somewhat deficient psychometrically and suggest using the friends and
family subindices separately.

This researcher has separated the support people into specific categories
(e.g. brothers, sisters, sons, daughters, etc). In addition, two questions were
added pertaining to the presence and helpfulness of confidants.

2) W (BDI)
(Beck, A. T., Ward, C. H., Mendelson, M., Mock, J., & Erbaugh, J., 1960).

This inventory has been studied for reliability and validity in the elderly
(Gallagher, Nies, 6: Thompson, 1982; Gallagher, Breckenridge, Steinmetz, 8:
Thompson, 1983). The BDI is a 21 item, 4-point discrimination scale that rates
the intensity of depression symptoms. Psychometric properties of the scale are

23

reviewed in Beck (1967). Results of studies with the elderly show that the
BDI has respectable internal consistency and stability for use in research with
this population. For example, congruence between conventional cut off
scores on the BDI and selected diagnostic classifications of the Research
Diagnostic Criteria (RDC) (Spitzer, Endicott, Gr Robins, 1978) for detection of
Major and Minor Depressive Disorders in 102 elderly outpatients was high.
Only 16.7% were misclassified by customary BDI cutoff scores (Gallagher et al,
1983). These results suggest that the BDI can also be used as a screening
instrument for identification of clinically depressed elders.

3) W (351)

(Derogatis 6: Spencer, 1983). This instrument is a 53 item abbreviated version
of the SCL-90 (Derogatis, 1977). According to Derogatis and Spencer,
correlations between the two range from .92 to .99. The 851 measures
psychopathology along nine dimensions, including depression. Norms are
available for the elderly grouped by gender (Hale, Cochran, & Hedgepeth,
1984). Internal consistency ranges from alpha = .77 on the Interpersonal
Sensitivity Scale to .91 on the Depression Scale. One difference noted between
the elderly and younger adults is that the elderly tend to report higher levels
of distress on most symptom dimensions (Hale et al, 1984). Hale et al suggest,
therefore, that age-relevant norms be used when the B81 is employed with
the elderly population.

4) WW (SDAT)
(Storandt, Botwinick, Danzinger, Berg, Gr Hughes, 1984). The SDAT battery is a

brief (10 minute) battery of four psychological tests used to assess actual
memory functioning. The battery consists of Mental Control and Logical

24

Memory from the Wechsler Memory Scale (Wechsler, 1955), Trailmaking A,
which is part of the Halstead—Reitan neuropsychological battery, and the
Word Fluency test which requires people to list as many words as possible,
beginning with a specific letter, in 60 seconds. Eight-four individuals, 38 men
and 46 women, half of whom were diagnosed as having mild SDAT and half
normal controls were recruited from local physicians or in response to
widespread community announcement of the project. Those with mild
SDAT ranged in age from 63 to 81 years (mean :1: SD, 71.5 :t 5.0 years). Years of
education ranged from 8 to 21 (12.4 :t 4.1 years). Healthy matched controls
ranged in age from 64-82 years (71.6 13.2 years). Their range of education was
8 to 20 (12.4 :1: 5.2 years).

The sample was partitioned into two subsamples by selecting every
other matched pair in the order enrolled in the study. An initial
discriminant analysis was performed on one set of 21 pairs. The regression
weights were then applied to the second set. The first discriminant analysis
correctly classified all persons on the basis of the four variables described
above. Applying the weights to the second set, all but two persons (one
demented, one not) were classified correctly (98% correct).

Standardized coefficients were used for the final analysis. Canonical
scores greater than or equal to 0 were classified as demented; those with

canonical scores less than 0 were classified as normal.

RESULTS

SUBJECT CHARACTERISTICS

Demographic data were collected for the following categories: Age,
gender, marital status, and education. In addition, information regarding
health, memory, housing, and psychological well-being was also obtained.

AM One hundred sixty-five participants ranging in age
from 55 to 91 (Mean, Median, Mode = 71; SD. = 6.8) agreed to take part in the
research. Although the majority of the participants (109) were female;
nevertheless, there is still a representative sample of males (56). The
percentage of females to males is 66% to 34%.

Housing, One indication of the independence of these participants is
that 75% lived in single family homes and 21 % resided in apartments. Only
4% were in housing specifically for the elderly. In addition, the majority
lived with at least one other person, with only 39% percent living alone.

MW While the majority of the sample was
married, the percentage was not overwhelming (58%), with 31% widowed
and the remaining 11 % either divorced or never married. As expected, most
individuals were not in the work force; however, almost 25% were employed
at least part-time.

Edugag'gn. This is a relatively well-educated sample with 76% having
attended college. Of these, about 80% obtained at least an Associate's degree
and over 45% earned a Master's degree or higher.

Health, In general, this sample of older adults viewed their overall
health as good or excellent (83%) with only 3% rating themselves in poor

25

26

health. An overwhelming majority (95 %) left their neighborhood at least
once per week and were able to drive a car. In addition, 80% felt their health
was the same or better than 3 years ago with over 90% percent stating their
health either had none or virtually no impact on their ability to engage in
activities.

In addition, when asked to rate their overall health compared to others
of the same age, over 50% rated their health as better than most people.
Within the past 12 months, 83% had not been in the hospital; over 70% had
been sick in bed one day or less; and the average number of reported visits to
the doctor in the past year was 5.5 (SD = 7.2) with the median 3 and the mode
2.

The total number of illnesses in the past year ranged from 0 to 11 with
the mean and median 2 (SD = 1.7) and the mode 1. Of the illnesses, arthritis
was most frequently listed (54%), followed by hypertension (33%), and
cataracts (20%). Eighty-four percent took at least one medication with the
range 0 to 11, the mean and median 3 (SD = 2.1), and the mode 2.

m Twenty-six percent indicated they experienced major
memory problems in the past and 70% of these indicated memory problems
were still present. Eleven percent admitted to difficulty in the past with
knowing the time of day, day of the week, or month of the year, with over
60% of these continuing to have this problem.

Actual memory functioning as measured by the SDAT battery
suggested that 11 % of the sample needed further follow-up to determine if
they were showing signs of early dementia; the rest of the participants scored
within the normal range on this test (Scores ranged from -6.380 to +2096 with
the mean -1.908 and 8.0. 1.625).

27

W Results on the two tests of psychological
well-being indicated that this sample is relatively well-adjusted with few
reporting significant depression or other mental health symptoms. For
examwe, scores on the Beck Depression Inventory (BDI) ranged from 0 to 33
with the mean score 6.5 (SD = 5.3). Scores less than or equal to 10 are
considered normal for this measure of depression. Brief Symptom Inventory
(BSI) scores for all subscales were lower than the norms for the elderly except
for the Obsessive-Compulsive and Interpersonal Sensitivity scales which
were slightly higher than those reported in Hale et al (1984). Overall scores on
the BSI ranged from .00 to 1.62 with the Mean .38 (SD = 0.32).

CONSTRUCTION OF SCALES
As stated earlier, the purpose of this research is to look at the

relationships among mm W and
W W Life is measured by social
support, memory complaints, and perception of health; thagigral
Competence is measured by actual health and memory functioning;
W is measured by reported mental health status.

In order to form scales that would measure these constructs, factor
analysis using Principle Axis Factoring (PAF) with a varimax rotation was
performed on items from the Multi—level Assessment Instrument (MAI)
pertaining to memory complaints, perception of health, social support, and
actual health. In addition, items on the MAI that pertain to depression and
other symptoms of mental health status along with scores on the Brief
Symptom Inventory (BSI) and Beck Depression Inventory (BDI) were factor
analyzed to form a scale measuring mental health status.

28

Based on the above analyses, four scales were developed: Momory

m 1 ° ' al a1 Heal and Mental Hoalth.
The following are the scales that form the constructs Pergeivfi Quality

L_e.fLif Wmdwk 1 'al ll-Be' :

PERCEIVED QUALITY OF LIFE

W- Given that none of the Social Support questions
loaded significantly on any one factor, it was decided to add the total number
of relevant items from the MAI (38b through 38h) to establish this scale. This
included items which listed types of people within the social network
(daughters, sons, friends, etc), frequency of visits to and from these
individuals, number of phone calls, physical distance of support people,
number of confidants, and the ability of confidants to provide emotional
support. (See Table 2 for the actual categories). Note: All the items in this table
refer to individuals not living with the subject; hence, there is no category for
"spouse." (Higher scores indicate more social support).

WW5 scale was established by summing the raw
scores of items 24, 24a, 25, and 25a from the MAI (factor loadings ranged from
.69 to .77 ). The internal consistency of the scale was found to be more than
satisfactory (Cronbach's alpha = .85, item-total correlations ranged from .67 to
.72). (Higher scores indicate more memory complaints).

W This scale was established by summing the
standardized scores of items 14 and 15 from the MAI (factor loadings ranged
from .44 to .46 ). The internal consistency of the scale was satisfactory
(Cronbach's alpha = .59, correlation between these two items =.42). (Higher

scores indicate more positive perception of health).

29

TABLE 2
SOCIAL SUPPORT VARIABLES

V A LE E MEAN .D. MEDIAN M DE %
Daughters 0-4 1.03 1.06 1 . 0
Sons 0—5 0.85 1.02 1 0
Grandchild 0-4 0.18 0.63 0 0
Sisters 0-3 0.41 0.73 0 0
Brothers 0-4 0.28 0.62 0 0
Female 0-6 1.65 1.61 1 0
Friends

Male 0-6 0.75 1.23 0 0
Friends

Total 0-8 6.20 1.98 7 8
Persons

Number of 0-8 5.30 2.28 5 8
Confidants

Phone 0-1152 346.8 255.5 290 230
Calls/ Year

Visits per 0-3740 658.0 517.1 518 128

Year

39

9O

33

0.6

3 O

BEHAVIORAL COMPETENCE

Wealth, This scale was established by summing the standardized
scores of items 17, 18, 19, 30b, and the total number of illnesses listed from
items 2a through 22t of the MAI (factor loadings ranged from .32 to .61). The
internal consistency of the scale was satisfactory (Cronbach's alpha = .62, item-
total correlations ranged from .24 to .47). (Lower scores indicate better overall
health).

mm Memory Function was measured by the overall
score on the SDAT battery described above. (Lower scores indicate better
memory functioning).

PSYCHOLOGICAL WELL-BEING

mm This scale was created by summing the standardized
scores of the BDI and the BSI (factor loadings = .78 and .96). The internal
consistency of the scale was more than satisfactory (Cronbach's alpha = .85,
correlation between these two subscales = .73). (Lower scores indicate greater
Psychological Well-being).

The hypotheses were tested using two forms of analyses: Bivariate
Analysis which examines the relationship between individual variables, and
Multiple Regression Analysis which examines a set of independent
(Predictor) variables as they relate to one single dependent (Criterion)

variable.

BIVARIATE ANALYSIS RESULTS

The reader will recall Table 1 which lists the expected direction of the
relationship between the following variables (these results can also be found
in Table 3):

31

mm It was hypothesized that age would be significantly
correlated with Actual Health and Social Support. The correlation between

Actual Health and age was non-significant. The only variable in the scale
Actual Health that was significantly correlated with age was total number of
illnesses in the past year (1 = .16; p < .02). Two support items were
significantly correlated with age: Number of sons (3; = .13; p < .04), and
number of female friends (1; =- .12; p < .02).

W: In addition, Memory Function (1 = .20; p < .004) and
Memory Complaints (3; = .15; p < .03) were significantly correlated with age .

My: Older individuals listed more sons and female friends, and
as age increased, memory functioning decreased and memory complaints
increased, as did number of illnesses. However,
other indicators of health (number of days in bed, etc) were not significantly
correlated with age.

MW It was hypothesized that Actual Health
would be significantly correlated with the following variables: Age,
Perception of Health, Memory Complaints, Social Support, and Psychological
Well-Being.

As noted above, age was not significantly related to Actual Health;
however, the other hypotheses were all confirmed: Perception of Health (1; = -
.49; p < .000); Memory Complaints (r = .31; p < .000); Psychological Well-Being
(; a .43; p < .000).

Social Support was significantly correlated with Actual Health in the
predicted direction (except for number of grandchildren). The Social Support
items that were significant are as follows: Number of grandchildren (1: = .16; p
< .027); number of daughters (1: = -.15; p < .037); total persons (r = -.17; p < .020);
phone calls (r = -.15; p < .034); number of confidants (r; = -.18; p < .013). All

32

other measures of Social Support were not significantly related to Actual
Health. '

Wm In addition to the predicted variables, the following
were also significantly correlated with Actual Health: Memory Function Q =
.22; p < .003); number of years in school (r = -.15; p < .032); attendance at
college (1; = -.15; p < .034).

Summgy: Participants reporting better overall health were more
likely to have attended college or have spent more time in school. They
performed better on memory tests, had fewer memory complaints, fewer
mental health problems, and viewed their health more positively. They also

listed more people and more confidants in their social network.

MW It was predicted that Actual Health,
Memory Complaints, Psychological Well-being, and Social Support would be

significantly correlated with Perception of Health. All of the hypotheses were
confirmed: Actual Health (r = -.49; p < .000); Memory Complaints (r = -.30; p
< .000); Psychological Well-being (r; = -.45; p < .000).

Individuals' perception of health was correlated with the following
social support variables: Number of children (r; = .17; p < .017); number of
friends (1: = .13; p < .05); total number of persons (1 = .19; p < .008); number of
confidants (r a: .18; p < .010); amount of emotional support (r = .17; p < .016).

W: In addition to the predicted variables, Memory
Function was also significantly correlated with Perception of Health (r = -.26; p
< .000).

m: Participants with more positive perception of health
reported fewer memory problems and mental health concerns as well as

greater actual health. They also performed better on memory function tests.

33

In addition, these individuals listed more children, friends, and total number
of people to whom they felt close, in whom they could confide, and from
whom they felt emotional support.

WWW It was predicted that Perception of
Health, Memory Function, and Psychological Well-being would all be

significantly correlated with Memory Complaints. These hypotheses were
confirmed: Perception of health (r = -.30; p < .000); Memory Function (r = .27;
p < .000); and Psychological Well-being (r = .30; p < .000).

W: Amount of emotional support (r = -.21; p < .022) and
number of confidants (r = .18; p < .01) were also significantly correlated with
Memory Complaints.

5mm: Complaints of memory were associated with an increase in
reported mental health problems and a decrease in subjects' perception of
their health as well as their reported actual health status. In addition,
presence of emotional support and number of confidants were correlated

with fewer memory complaints.

‘Wflmmmrh It was hypothesized that Memory
Complaints and Psychological Well-being would be significantly correlated

with Memory Function. Both of these predictions were confirmed: Memory
Complaints (see above), and Psychological Well-being (l; = .22; p < .003).

W: As noted above, Perception of Health and Actual
Health were significantly correlated with Memory Function. Other variables
include attendance at college (1: =- -.25; p < .001); years in college (r = -.19; p <
.017); and overall years in school (r = -.19; p < .009).

34

In addition, the following social support variables were significantly
correlated with Memory Function: Number of daughters (r = -.13; p < .052);
number of brothers (1 = -.13; p < .051); total number of persons (g = -.14; p <
.038); number of confidants (1; a -.26 p < .001); and amount of emotional
support (r; = -.21 p < .004).

5311313311: As Memory Function scores increased, so did actual health
and perception of health; whereas, memory complaints and mental health
- problems decreased. In addition, attendance at college, number of years in
college, overall years in school, as well as having people to confide in, feel
close to, and receive emotional support from increased the likelihood of
better memory function.

WW2. It was hypothesized that Actual
health, Perception of Health, Memory Complaints, and Memory Function

would be significantly correlated with Psychological Well-being. As described
above, all of these hypotheses were confirmed.

The following social support variables were significantly correlated
with Psychological Well-being: Number of daughters (1 = -.20; p < .006);
number of sons (1 = -.15; p < .030); number of female siblings (r = -.13 p < .050);
number of confidants (r; = -.28; p < .000); and amount of emotional support (1;
= -.24; p < .001). Visits, phone calls, or physical distance from the social
network were not significantly related to Psychological Well-being; however,
number of years in school (1 = -.21; p < .004); and college attendance (r = -.24; p
< .001) were significantly correlated with Psychological Well-being.

53mm: Participants in better overall health who had few if any
memory complaints, performed well on memory tests and saw their health

more positively, reported fewer mental health concerns. In addition, fewer

35

daughters, sons, sisters, and confidants, as well as less emotional support were
all negatively correlated with Psychological Well-being. Whether or not there
was frequent contact via phone calls or visits, or how far away these support
people were did not affect perception of mental health; rather, the presence of
supportive individuals and feeling emotional support were the significant
factors. Finally, number of years in school and having attended college were

positively correlated with Psychological Well-being.

AGE
AH

MC
MF
PWB

COLY
DGHTR
SON

F.FR.
M.FR.
BRO
SIS

DS
L

agesasfl

36

CODE FOR TABLE 3

AGE OF PARTICIPANT

ACTUAL HEALTH

PERCEPTION OF HEALTH

MEMORY COMPLAINTS

MEMORY FUNCTION

PSYCHOLOGICAL WELL-BEING

TOTAL YEARS IN SCHOOL

COLLEGE ATTENDANCE

TOTAL YEARS IN COLLEGE

NUMBER OF DAUGHTERS LISTED

NUMBER OF SONS LISTED

NUMBER OF CHILDREN LISTED

NUMBER OF FEMALE FRIENDS LISTED
NUMBER OF MALE FRIENDS LISTED
NUMBER OF BROTHERS LISTED

NUMBER OF SISTERS LISTED

NUMBER OF GRANDCHILDREN LISTED
TOTAL NUMBER OF PERSONS LISTED

TOTAL NUMBER OF PHONE CALLS PER YEAR
TOTAL NUMBER OF CONFIDANTS

AMOUNT OF EMOTIONAL SUPPORT
NUMBER OF MEDICATION S TAKEN PER DAY
TOTAL NUMBER OF ILLNESSES IN PAST YEAR

37

 

TABLE 3
BIVARIATE ANALYSES RESULTS

AGE AH PH MC; MF PWB
AGE -
AH N.S. -
PH N.S. -.49"" -
MC .15" .31"" -.30"" -
MF .20“ .22“ -.26"" .27“ -
PWB N.S. .43“ -.45"" .30“ .22“ -
SCH N.S. -.15"" N.S. N.S. -.19"" -.21""
COL N.S. -.15" N.S. N.S. -.25"" -.24""
COLY N.S. N.S. N.S. N.S. -.19"" N.S.
DGHTR N.S. -.15" N.S. N.S. .13" -.20""
SON .13" N.S. N.S. N.S. N.S. -.15""
CH N .S. N.S. .1 7" N .S. N.S. -.24""
F.FR. .12" N.S. N.S. N.S. N.S. N.S.
M.FR. N.S. N.S. N.S. N.S. N.S. N.S.
BRO N.S. N.S. .13" N.S. N.S. N.S.
SIS N.S. N.S. N.S. N.S. N.S. -.I3"
GC N.S. .16" N.S. N.S. N.S. N.S.
TP N.S. -.17" .19" N.S. .14"" -.27""
PC N.S. -.15" N.S. N.S. N.S. N.S.
CF N.S. -.18"" .18"" -.I8"" -.26"" -.28""
ES N .S. N .S. .17" -.21" -.21"" -.24""
MEDS N.S. + -.34"" .18" N.S. .34“
TILL .16" + -.35"" .18" 23"" .33“

 

"P (.05 ""P<.01

+ part of scale

38

GENDER DIFFERENCES

In addition to the bivariate analyses reported above, cross tabulation of
sex by marital status by confidant was performed to test the following
hypothesis:

When asked to name a confidant, (who may or may not live with the
subject) females and males will differ in who they believe is someone they
can talk to about their troubles or concerns:

FEMALES: Friends '> Spouse > siblings > children

MALES: Spouse > children > siblings > friends

Results of the cross tabulation analysis showed that 89% of the males
and 83% of the females listed a confidant. 92% of these males were married as
compared to 41% of the females. The remaining 8% of males were widowed.
More widowed women (44%) than married women named a confidant. In
addition, of those women who named a confidant, 9% were divorced and the
remaining 6% had never been married.

Given that the original hypothesis included listing a spouse as
confidant, the following gender difference analysis is between married males
and females. The result is statistically significant at the <.000 level:

MARRIED FEMALES: Spouse (59%) > Friend (30%) > Child (11 %)

MARRIED MALES: Spouse (87%) > Friend (6%) > Child (2%) =

Sibling (2%) = Other (2%)

The original hypothesis was not confirmed. Both, genders were more
apt to name their spouse as confidant; however, more married men (87 %) as
compared to married women (59%) did so, and only 6% of married men as
compared to 30% of married women listed a friend as their confidant.

Given that 44% of those women who named a confidant were

widowed, a post hoc analysis evaluated the response of widowed women:

39

WIDOWED WOMEN: Child (50%) > Friend (37%) > Other relative
(7%) > Sibling (2%) = Other (2%)
Thus, half the widowed women in this sample who named a confidant

listed one of their children, and over one-third of them named a friend.

MULTIPLE REGRESSION RESULTS

In order to test the formal hypotheses, forward multiple regression was
performed on the following predictor variables with Psychological Well-being
as the criterion variable: Perception of Health and Actual Health; Memory
Complaints and Memory Function; and Social Support variables. Once the
significant predictors of Psychological Well-being were identified, the
demographic variables (age, gender, marital status, education) were then
regressed onto these "predictors" of Psychological Well-being in a separate
analysis. These same demographic variables were also entered as predictors
of Psychological Well-being. Once the significant demographic variables were
identified, the original multiple regression analyses were run again using
both the significant demographic variables and the original predictor
variables (Memory Complaints, Memory Function, Actual Health, etc.) with
Psychological Well-being as the criterion. (See Figure 1)

Finally, the demographic variables and Social Support individuals (e. g.
daughters, sons, friends) were entered as predictors of amount of emotional
support. The same model of regression analysis described above was
performed with social support individuals (e.g. sons, daughters) as the
predictors and the amount of emotional support as the criterion.

(See Figure 1)

4o

 

 

 

 

 

 

 

 

 

DEMOGRAPHICS
Age
Gender
Education
Marital Status
VARIOUS PREDICTORS fl PSYCHOLOGICAL WELL-BEING
(e.g. Memory Complaints, , or
Actual Health, EMOTIONAL SUPPORT
Social Support)
M
Figure 1

MULTIPLE REGRESSION MODEL

41

See Table 4 for the statistical data on the following results.

HYPOTI-IESIS #1: Both Perception of Health and Actual Health will
have an effect on Psychological Well-being; that is, positive perception of
health as well as better overall health will be positively correlated with
Psychological Well-being.

Results of multiple regression analysis confirm the above hypothesis.
Both of these variables have a unique effect on Psychological Well-being.
That is, if Actual Health (or Perception of Health) is held constant, Perception
of Health (or Actual Health) continues to be a significant predictor of
Psychological Well-being whiCh suggests that Perception of Health is
measuring something that Actual Health is not and vice versa.

This will be discussed below. In addition, the combination of Actual Health
and Perception of Health is a better predictor of Psychological Well-being than
either of these variables alone; however given that Perception of Health
entered the multiple regression equation first, it is likely that this variable has
a greater effect on Psychological Well-being than Actual Health.

Results of entering the demographic variables in the multiple
regression analysis indicate that attendance at college (which is a significant
predictor of Psychological Well-being) has an effect over and above that of
Actual Health and Perception of Health. Thus, those who attended college,
had better overall health and had a more positive perception of their health
were more likely to score better on measures of psychological well-being. No
demographic variables directly predicted either Perception of Health or Actual
Health.

HYPOTHESIS #2: Memory complaints will have a stronger
relationship to Psychological Well-being than will Memory Function;

42

although there will be a significant correlation between Memory Function
and Psychological Well-being.

This hypothesis was confirmed. If Memory Function is held constant,
Memory Complaints continues to have a significant effect on Psychological
Well-being; however, if Memory Complaints is held constant, Memory
Function ceases to be significantly related to Psychological Well-being; thus,
given that Memory Complaints and Memory Function are significantly
correlated with each other, Memory Function appears to have an indirect
effect on Psychological Well-being through its direct effect on Memory
Complaints.

Attendance at college is the only demographic variable that continues
to be a significant predictor of Psychological Well-being when Memory
Complaints is entered into the multiple regression equation; that is, those
who attended college and have fewer memory complaints are more likely to
have positive psychological well-being. No demographic variables
significantly affected either Memory Complaints or Memory Function when
they were analyzed via multiple regression. In addition, Memory Function
which appears to directly affect Psychological Well-being when analyzed by
simple Pearson Product Moment Correlation, does not directly relate to
Psychological Well-being; rather, in the Multiple Regression Analysis, it
appears to affect one's perception of memory (Memory Complaints) and this
perception in turn affects Psychological Well-being.

HYPOTHESIS #3 The presence and perceived helpfulness of a
confidant will act as a buffer or moderator between Perception of Health and
Psychological Well-being.

43

The above hypothesis was confirmed for presence of confidants but not
for perceived helpfulness of confidants; however, attendance at college had a
stronger effect on Psychological Well-being than did presence of confidants.
That is, those who attended college had a more positive perception of their
health and if they also had confidants, their overall Psychological Well-being
was significantly greater than those who had not attended college or had
fewer confidants. An added factor appears to be gender in that females
reported significantly more confidants in their social network than did males;
therefore gender through its effect on number of confidants appears to have
an indirect effect on Psychological Well-being.

HYPOTHESIS #4: Friends and siblings will provide more emotional
support than children, especially for females.

This hypothesis was not confirmed. Results of the multiple regression
analysis reveal that, of individuals who subjects felt close to and who did not
live with them, the following is the order of importance in terms of
emotional support:

Daughters > Female friends > Sons > Brothers > Sisters > Male friends

In addition, females were more apt to name daughters and female
friends; whereas, males were more apt to name male friends. Also,
attendance at college and marital status may have an indirect effect on
emotional support from daughters in that these variables were predictors of
number of daughters in the network. Attendance at college was also a
predictor of emotional support from daughters and female friends.

Given the above results, the original hypothesis was not confirmed in
that daughters are most helpful followed by female friends and then sons.
However, females in general appear to be more helpful than males. In

44

addition, there appears to be two distinct groups with the most helpful group
being daughters, female friends, and sons (in that order) and the less helpful
group being brothers, sisters, and male friends (in that order). The least
helpful of all are male friends which were listed more by males; whereas, the
two most helpful (daughters and female friends) were listed more by females.
(See Table 5)

HYPOTHESIS # 5: Social support variables and their effect on
psychological well-being will be in the following order of importance:

Emotional support > number of confidants > number of phone calls
per year > number of visits to and from individuals in the social network per
year > total number of persons in the network > physical distance from
individuals in the network

Results of the multiple regression analysis indicate that, while both
emotional support and number of confidants were significant predictors of
Psychological Well-being, none of the other variables have any significant
effect. In addition, when the demographic variables are analyzed for their
effect, attendance at college appears to be significantly related to number of
confidants; therefore, it appears that, for this sample, it doesn't matter how
often there is contact with supportive people in the network or how close one
lives to these people; rather, the number of confidants and perceived
helpfulness of them are the significant correlates of Psychological Well-being
and if individuals have attended college, they are more likely to have such
individuals in their network (Table 4).

45

TABLE 4

MULTIPLE REGRESSION ANALYSES
WITH PSYCHOLOGICAL WELL-BEING
AS THE DEPENDENT VARIABLE

PREDICTORS BETA
Fér'céfiti'o'n-Sf'fiéfit'h""""""""""'""?0’.§§? """""
Actual Health 0.26""

College Attendance -0.16"

R2 0.30

F(3, 145) 21.04""
Fér'cé'fititit'ai-fiéfith""""""""""""-TdI§7-*7 """"""
College Attendance -0.19""

Number of Confidants -0.16""

Emotional Support -0.15(n.s.)

R2 0.25

F(3, 153) 17.05""
Memo-ry-Cdn-rpliihts """"""""""""" (its: “““““
Memory Function 0.09(n.s.)

College Attendance 024""

R2 0.14

F(2,155) 12.63""

4 6
TABLE 4 (cont'd)

PREDICTORS BETA
ETmBTiBHa'l'§tI§pB'§t-""""'""""""""?0'.§4?‘ --------
Number of Confidants n.s

Phone Calls per Year n.s

Visits per Year n.s

Number of Persons in n.s

Network

Physical distance from n.s

Network

R2 0.06

F(1, 155) 9.80""
CBEEg-e-At-tend—ahze----—---—------__-—_—-—-_:0_.24-"; """""""
Number of Confidants -0.21""

Emotional Support n.s

Phone Calls per Year n.s

Visits per Year n.s

Number of Persons in n.s

Network

Physical distance from n.s

Network

R2 0.12

F(2, 153) 10.47""

:.---<-751- “““““““““““““““““““““““““

47

TABLE 5

MULTIPLE REGRESSION ANALYSIS
WITH EMOTIONAL SUPPORT AS
THE DEPENDENT VARIABLE

PREDICTORS BETA

bitigitér's ---------------------------- 0 '55:? """""""
Female Friends 0.46""

Sons ' 0.38""

Brothers 0.24""

Sisters 0.17""

Male Friends 0.15"

R2 0.34

F(6,153) 13.12""

rat-5231' """"""""""""""""""""""""

DISCUSSION

The purpose of this research was to examine the correlates of
psychological well-being in the able elderly; that is, those individuals over age
55 who were living independently in the community. The frame of reference
was Lawton's (1986) conceptualization of "The Good Life" for this population
(see Introduction). mm W as measured by Perception of
Health, Memory Complaints, and Social Support; havi r m n as
measured by Actual Health and Memory Function; and MEI—9832131111:
mg as measured by the amount of mental health concerns and symptoms
were the three aspects of this "Good Life" analyzed by this researcher.

It should be noted that Lawton does not differentiate between
Memory Function and Memory Complaints in his model and subsumes both
of these under W. Given that Memory Complaints is a
subjective evaluation of cognitive functioning, while Memory Function is an
objective measure, this researcher has opted to separate these two constructs.
Lawton's conceptualization of Memory Function as a measure of Bohavioral
Qompotongo appears warranted; however, Memory Complaints can be argued
to be more of a subjective evaluation (a perception) of Major-31
W and therefore, more a measure of m W.

In addition, Lawton includes both Perception of Health and Actual
Health as measures of W. As in the case of memory
discussed above, it is reasonable to assume that perception of health is more
of a cognitive gyoluoo'oo of actual health and thus a measure of Porooivoo
W rather than one of ha ' ral m t n e. Therefore, the

decision was made to separate these two constructs as well.

48

49

Finally, even though Lawton conceptualizes number of relatives,
friends, etc as part of the oneoo'yo Enyironment, given that subjects in this
research were asked to name those individuals they "felt close to", these
variables also appear to be a better measure of W and
are thus considered part of this aspect of the "Good Life." The ijootiyg
Eovironmont portion of Lawton's model was not examined in the present

research.

SUBJECT CHARACTERISTICS

Specific characteristics of the participants who agreed to take part in the
research project can be found in the Results Section. In general, there is little
doubt that the men and women who volunteered were a representative
sample of able elderly, those living independently in the community. Only
4% of them lived in housing specifically for the elderly and three—fourths
lived in private homes. Virtually all (95%) left their neighborhood at least
once per week and were able to drive a car.

The majority of participants were either married or widowed, with
only 11 % divorced or never married. This may represent a cohort effect as
remaining single or being divorced was not as acceptable when these
individuals were young adults.

As might be expected, the majority of participants were retired,
although almost 25% continued to work at least part-time.

The educational background of these individuals was obviously of an
advanced level, with over 75% having attended college. In addition, of those
who attended college, over 45% obtained a Master's degree or higher. It
seems likely that this result is due to the the majority of volunteers coming
from the retired staff or faculty of MSU (as well as their spouses).

50

A criticism of this research, therefore, might be that these participants
are not representative of the overall population of elderly, and as a result, the
findings are limited in generalizability. While this may be a valid point, one
can look at this aspect of the subject population from another perspective.
Schaie (1988) has argued that, in fact, there is a "great diversity" in today's
elderly. In fact, he cautions researchers not to assume that all elderly over age
60 are alike merely because they are classified as such. If his comments bear
consideration, then examining different groups of elderly demonstrates
responsible research, and one subset of the elderly that investigators need to
study is the highly educated individuals who are both physically and
psychologically healthy and can and should continue to be positive
contributors to society.

That this is a group of able elderly is best seen in the data regarding
health, memory and psychological well-being. Only 3% rated themselves in
poor health and over 80% saw their health as good or excellent. Only 20%
thought their health was worse than three years ago and virtually all (90%)
stated their health had little or no impact on their ability to engage in
activities. Most of them had not even been sick in bed one day in the past 12
months, and the most common illnesses listed were typical consequences of
the aging process (hypertension, arthritis, cataracts), rather than life
threatening conditions such as cancer or strokes.

Concerns with memory do not appear to be a major issue with this
sample of elderly. For example, only about 25% felt they had major memory
problems in the past, with about 70% continuing to have this problem. In
addition, only 11% admitted to problems with the time of day, day of week, or
month of year, with only 60% of these individuals feeling this was a problem
at the time of the interview. It may be, however, that had there been more

51

questions about memory or that the the word "major" had not been chosen,
more individuals might have expressed concerns about this aspect of their
functioning. One possibility would be to use a scale like the one proposed by
Crook and Larrabee (1990) which has a wide range of memory items, a large
normative base, and an "extremely stable factor structure" (p.53).

The test of actual memory function revealed that this sample of
elderly did not have severe memory problems (only 1 1 % needed further
follow-up to screen for early signs of dementia); therefore, even with more
detailed memory complaint questions, it is likely that the percentage of
subjects admitting to problems would remain small. On the other hand, it
should be recognized that there is no normative data which relates memory
complaints in the highly educated able elderly to performance on memory
tests. Until data of that nature is collected, the relationship between memory
complaints and actual performance is an open question.

Despite the suggestion by professionals that depression may be a major
concern for the elderly (Blazer, Hughes, & George, 1987); few of the elderly
who took part in this project reported significant depression or other mental
health problems. However, even though moderate to severe psychological
symptoms were relatively absent in this sample of elderly; nevertheless,
results discussed below indicate that even relatively mild symptoms are
affected by such things as memory complaints and social support.

There are several possible explanations for the virtual absence of
depression in these participants: Newman (1989) suggests that elderly people
who participate in most community based research are likely to be
"disproportionately represented by W" (p.162). Given that this was
the very sub-population of elderly the current study sought to investigate, the

52

absence of major depressive symptoms in the sample can be looked on as a
positive sign that these elderly were indeed "optimal agers."

Another "explanation" can be found in the recent longitudinal study
by Aldwin, Spiro, Levenson, and Bosse (1989) who examined the question of
whether mental health changes with age. They found that psychological
symptoms were "essentially stable and increased at a rate of one new
symptom every 28 years" (p.303). In addition, they found no linear
relationship between age and psychological symptoms; although they did
report non-linear changes in the number of symptoms with symptom
reporting increasing in young adulthood, remaining stable in mid-life and
then increasing slightly in older age. It is possible, therefore, that this sample
of able elderly were also relatively well-adjusted throughout their lives. In
support of this explanation is the fact that there may be an underestimation of
depression in the general population because research has relied on
traditional psychiatric diagnosis (Blazer, Hughes, 8: George, 1987) rather than
standard screening scales to estimate depression (Newman, 1989). Given that
the present study did in fact utilize two standard screening scales (The Brief
Symptom Inventory and the Beck Depression Inventory) and failed to find
significant psychological disturbance, it is likely that this sample is biased
towards the healthy individual, which as stated above, was the intent of this
research.

Gatz and Hurwicz (1990) also report no significant age difference in
level of depression and state: "...although conclusions must be tempered by
their not being based on longitudinal data, there is not compelling
confirmation of the common belief that depression inexorably increases
decade by decade" (p.289). Gatz and Hurwicz also report that there was no
relationship between their "Lack of Well-being" subscale and other

53

depression sub-scales in the oldest age group. They suggest that "expressions
of lack of well-being may represent a construct different from depression" and
that "older people may not be unduly depressed but rather coming to terms
with some of the realistic constraints of their lives" (p.289).

Finally, Lee and Ishii-Kuntz (1988) in their research on social
interaction, loneliness, and emotional well-being in the elderly report that
those with more education have a "higher morale" than others. Given the
educational data reported above, this may also be a significant reason for the
virtual lack of psychological dysfunction in the present sample.

What appears likely, therefore, is that the present research has a
disproportionate sample of individuals who are not only "optimal agers" but
also have been psychologically healthy throughout their lives. While they
may not be experiencing moderate or severe depression, they may be
experiencing some discomfort over the inevitable changes that aging brings
about. Further research is needed to examine this hypothesis.

BIVARIATE RESULTS

Even though the bivariate results reported above were not part of the
formal hypotheses examined in this research, comments about these findings
appear warranted.

mom: Given Schaie's (1988) caution to avoid classifying all
individuals in aging research as "over 60", it is, at first glance, somewhat
surprising that age was not a significant demographic variable in this study.
Indeed, age did not enter at all into any of the Multiple Regression equations
that will be discussed below. Age only accounted for 3% of the variance in

54

total number of illnesses, 4% of the variance in Memory Function, and less
than 2% of the variance in Memory Complaints. In addition, only about 1 %
of the variance in number of sons and female friends could be attributed to
age.

The most likely explanation for the lack of differentiation by age in
this population is that the sample was truncated; that is the mean age was 71
(SD. 6.8). This means that almost 90% of the sample could be classified as
"Young-Old". Therefore, the logical explanation for a lack of age effect is that
this is a homogeneous sample of elderly and the actual age differences among
them have virtually no impact on the variables in the study due to the lack of

variance in age.

WM: Not surprisingly, Perception of Health had the
strongest correlation with Actual Health and accounted for almost 25% of the
variance. Thus, as would be expected, the healthier people are, the more
positively they view their health.

It also appears that Actual Health and Psychological Well-being are
significantly related to one another which is consistent with prior research
(e.g. Larue, Dessonville, & Jarvik, 1985; Blazer & Williams, 1980). In fact,
some researchers found that poor physical health had the strongest
association with depression (Murrell, I-Iimmelfarb & Wright, 1983). More
will be said about this relationship when the regression analyses are discussed
below.

Given that it is not unusual for some illnesses to affect memory
performance (e. g. diabetes, circulation problems), the finding that the
memory scales and Actual Health were significantly related to each other is

not surprising.

55

Although educational background was not hypothesized to be
significantly correlated with Actual Health, the finding that more years in
school and having attended college were related to better health makes sense
if we can assume that these events are likely also correlated with higher socio-
economic status. Unfortunately, data was not collected specifically about
socio-economic status, but it is reasonable to expect that a sample of adults
that came primarily from retired staff and faculty of a major university is
largely from at least the middle if not upper-middle class.

Finally, the expectation that more social support would be positively
correlated with better health was confirmed; however, it apparently does not
matter if face-to face contact takes place in that only number of phone calls
was associated with better health. In addition, daughters, total number of
people, and confidants were significantly (albeit weakly) associated with better
health. Interestingly, number of grandchildren was negatively correlated with
Actual Health. There does not appear to be an easy answer for this finding,
unless perhaps younger people have more time and energy to spend with
their sick elderly relatives. On the face of it, younger off-spring are initially
more mobile before they have children. It might be helpful to examine the
relationship of the elderly and their grandchildren in future research on
social networks.

MW: Perception of Health and its relationship
to Actual Health was already discussed above. It was also expected that

Perception of Health and Psychological Well-being would be significantly
correlated with one another. This was confirmed by the bivariate analysis and
is consistent with previous research which found that the meaning of health
to individuals rather than actual health status may be what actually affects

56

psychological well-being (Murphy, 1983; Stroller, 1984). This hypothesis was
examined in the multiple regression analysis to be discussed below. Of
course, the assumption being made here is that the direction of the
correlation is known, when in fact, it may be that depression affects the
elderly’s perception of health rather than vice-versa.

The finding that memory complaints and memory function are
significantly related to perception of health in this sample of able elderly is
likely due to the fact that cognitive deficits may signal poor health and this in
turn would affect one's perception of health in general.

The finding that many of the social support variables were significantly
correlated with Perception of Health (e.g. children, friends, presence of
confidants and emotional support) is consistent with research by Krause
(1987). However, Krause differentiated between quantity and quality of
support and found that quality rather than quantity was the significant factor.
The present research mainly looks at quality of support in that subjects were
told to list only those individuals in their network to whom they felt close.
However, given that as the number of individuals who were listed increased,
the more times participants could respond to the questions about confidants
and amount of emotional Support, it appears that quano'ty of emotional
support must also be taken into account in these findings. Therefore, it may
not be just quality of support but also the quantity of this type of support that
affects such factors as Perception of Health.

W: In addition to the correlations discussed
above, Memory Complaints were also significantly related to Psychological

Well-being, Memory Function, amount of emotional support and number of

confidants.

5 7
The relationship between Memory Complaints and Psychological

Well-being found in the present study is consistent with previous research
(Zarit, 1982; Williams, Little, Scates, 8: Blockman, 1987). In addition, although
correlations between Memory Function and Memory Complaints have been
reported (see Zarit), the overall conclusion is that Memory Complaints are
more correlated with Psychological Well-being than with actual function.
Though again we caution that there have been few normative studies on
memory among the highly educated, well-functioning elderly. Given that
this sample of elderly is highly educated and many of them may have made a
living by the use of their cognitive abilities, it may be that even small changes
in memory function are noticeable and therefore results in complaints of
memory deterioration which could be real memory decrements.

Post hoc analysis of soda] support and Memory Complaints reveals an
interesting relationship. Those who complained of memory difficulties listed
fewer confidants and less emotional support. Further research looking at the
relationship between memory complaints and sodal networks is needed
before any reliable explanation of this finding can be postulated. It may be,
however, that individuals who are experiendng concerns about their
competence, be it in the area of cognitive abilities or physical health, may be
more aloof from their network or perceive others as non-supportive of their
fears and concerns. Or, they may not confide their fears in others‘even if they
feel "close" to them. Again, the memory complaints may be tool in the sense

of actual memory impairments not found by gross testing of memory

functions.
W: In addition to the correlations discussed

above, Memory Function was also found to be significantly related to

58

Psychological Well-being, overall years in school, attendance at college and
number of years in college. In addition, amount of emotional support,
number of confidants, daughters, brothers, and total number of persons
subjects felt close to were significantly related to Memory Function.

There are two possible explanations that may account for the
correlation between Memory Function and Psychological Well=being. First of
all, given that Memory Complaints and Memory Function were significantly
correlated with one another, Memory Function may be affecting Psychological
Well-being through its effect on Memory Complaints (this will be discussed
below with the Multiple Regression Analyses).

Secondly, the explanation given above for the relationship of memory
complaints to memory function in this sample may also apply to memory
function and its significant correlation to Psychological Well-being; that is, if
even a small loss in memory function occurs in someone who is relatively
well-educated, this loss could result in increased psychological symptoms as
well. The fact that college attendance, years in college, and overall years in
school also increase Psychological Well-being may be another indicator of the
importance of cognitive functioning in this sample of well-educated elderly.

It is interesting that Memory Function was also significantly related to
such sodal support variables as number of daughters, number of confidants,
and emotional support. If these people are functioning well cognitively, they
may be reaching out to people more and thus receiving more support from

others.

W The relationship between
Psychological Well-being, Actual Health, Perception of Health, Memory

59

Complaints, Memory Function and education has already been discussed
above.

The fact that number of confidants and amount of emotional support
are related to positive mental health is consistent with prior research on the
soda] network of the elderly (Krause, 1987; Esser 8r Vitaliano, 1988; Murphy,
1982). As discussed above, it appears that quality of support is the major factor
that relates to Psychological Well-being. The fact that children and female
siblings were the specific types of relatives is consistent with other findings
(e.g. Ward, 1985; Cidrelli, 1989).

Contrary to what might be expected, however, face-to-face or other
types of contact were not significantly related to Psychological Well-being;
thus, it doesn't seem to matter if the elderly visit or talk with their confidants,
merely having them available per se appears to be the significant factor.

Thus, Mm one has support and someone to talk with about troubles or
concerns may be more important than actually sharing with that person.

GENDER DIFFERENCES

It was hypothesized that there would be a significant difference
between males and females regarding whom they would name in answer to
the question: "Is there someone you can talk to about your troubles or
concerns (i.e. a confidant)." Contrary to what was expected, m males and
females were more apt to name their spouse, although the percentage of
males (87%) compared to females (59%) was greater, with almost all married
men naming their wives, and only slightly over half of the women naming
their husbands. It was thought that women would name a friend more often
than any other individual, and although this was not confirmed, friends were

the second most frequently named individual (30%) among married women.

60

In addition, women were more likely to name a friend than were men.
Finally, only 11% of females and 2% of males named one of their children.

The results for males, therefore, is consistent with other research (e. g.
Vaux, 1985) in that males most often named their wives. For women,
however, even though friends were often named, nevertheless, husbands
were still in the majority. One possible explanation for this finding is
sampling bias. Many of the volunteers came with their spouse; that is, both
partidpated in the research, and even though the interviews were conducted
separately and neither was obligated to tell the other how they answered any
questions, there may have been some "pressure" to name a husband or wife.
On the other hand, given that these elderly did come together, it is also
possible that the couples were closer than the average married elderly and
honestly felt their spouse was their closest confidant. In addition, perhaps
educational background or sodo-economic status affects this variable. More
research would be needed to better tease out the answer to this question.

Discussion of the Multiple Regression analyses below will include
further elaboration on possible gender differences in sodal support.

Given that 44% of women who named a confidant were widows, these
results are also included. Again, friends were not the source most often
listed; rather, 50% of widows listed one of their children and 37% listed a
friend. This is consistent with results found in other research of women in
general (Depner & Ingersoll-Dayton, 1988), but not similar to a report by
Arling (1976) who found that interaction with children had virtually no
impact on the morale of widows. However, Arling did not spedfically look at
quality of support, but rather frequency of interaction. Thus, the crucial
variable regarding support may very well be quality rather than quantity.

6 1
MULTIPLE REGRESSION RESULTS

_'=e "E010 l‘:--l £11-: rm-.. ._ OIO'HaI'A‘3‘

The results of the Multiple Regression analysis confirm the hypothesis
that both Actual Health and Perception of Health significantly relate to
Psychological Well-Being. If we examine the results more closely, we can state
with a fair amount of confidence that, of the two, perception of health is the
better predictor, but that both together account for the largest percentage of
variance in Psychological Well-Being.

These results are consistent with previous studies which found a
significant relationship between perception of health and psychological well-
being (Gatz & Hurwicz, 1990; Stroller, 1984), and others which reported a
significant relationship between actual health and psychological well-being
(LaRue et al., 1985; Murrell, I-Iimmelfarb 8: Wright, 1983).

Although researchers may find that perception of health is invariably
more salient to elderly people than actual evidence of good versus poor
health; nevertheless, we cannot eliminate the effect of real health problems
anymore than we can ever totally separate nature from nurture in attempting
to explain personality, behavior, etc That is, one does not exist without the
other. Thus, two elderly people, one in good health, the other in fair health
may conceivably 9.239.112 their health in similar manners even though one
person's health is objectively worse than the other's. This perception in turn
may result in similar reports of mental health status.

Finally, the fact that college attendance was the only demographic
variable that entered into the regression equation further emphasizes the
possible significance of the role higher education plays in the lives of
individuals, the elderly in particular. It will be interesting to see how future

62

generations of elderly are affected by having attended college, espedally since
the opportunity to do so is greater now than it was when the present sample
of elderly were young adults. It may be that it will actually cease to be a

predictor of well-being just because it will no longer be a somewhat unusual

experience for people.

P ‘s-sge gm“; 0 s. 1; secs;- sss ano,’ elm

ushering:
Results of the regression analysis confirm the original hypothesis that

H

the presence of confidants can add to the effects of Perception of Health on
Psychological Well-being in the elderly. Whether these "confidants" are also
viewed as "helpful" apparently does not add anything to this effect. Thus, it
seems that if the elderly have a positive perception of their health and also
have people they can talk to about their troubles or concerns, they are more
apt to have greater psychological well-being. This is consistent with other
research (e. g. Krause, 1987).

In addition, if individuals have a somewhat less than positive
perception of their health but also have confidants, they are more likely to be
less psychologically disturbed by a somewhat negative perception of their
health. That is, number of confidants "buffers" the effect of perception of
health on psychological well-being.

Further regression analysis using demographic variables to predict
number of confidants revealed that females were more likely to name such
individuals. This may suggest that a gender difference exists in the effect of
perception of health on psychological well-being; that is, females may be less
affected by a negative perception of health because they have more people to

talk to about problems and concerns than men. However, given that there

6 3

was no significant gender interaction with psychological well-being in this
sample, further research would be needed to test this hypothesis. One other
explanation that comes to mind is that males may not need as many
"confidants" in their overall network to feel psychologically healthy.

Finally, an interesting result of entering the demographic variables in
the original multiple regression equation is that college attendance actually
was a better predictor of psychological well-being than was number of
confidants. The difference was rather small (see Table 4), but nevertheless,
this finding continues to raise an interesting question regarding the role of
higher education in the lives of the able elderly. It appears that those who
attended college were less likely to be psychologically distressed by the
perception that their health was less than perfect. Or to put this another way,
consistent with the findings of Lee 8: Ishii-Kuntz (1988), elderly with more
education and better health appear to lead psychologically healthier lives than
those with less education and poorer health.

trains __!'1!°l trans 0910---Ib 1-10. ’ - 0 0°4C- A '--i‘

The hypothesis that Memory Complaints would have a stronger
relationship to Psychological Well-Being than Memory Function was
confirmed. In fact, if both of these are entered into a Multiple Regression
equation, when Memory Complaints is held constant, Memory Function
ceases to significantly affect Psychological Well-Being. This of course suggests
that it is the elderly individual's subjective porceoo' on of memory function,
not memory function per se that has an appredable effect on mental health.
This is consistent with previous studies reported above.

That Memory Function significantly affects Memory Complaints when

simple Bivariate Correlation analysis is performed, however, suggests that

6 4
Memory Function does affect Psychological Well-being but that it is an

indirect effect through its relationship to complaints of memory.

As mentioned earlier, attendance at college was the only demographic
variable that continued to be a significant predictor of Psychological Well-
being when Memory Complaints was entered in the regression analysis. No
demographic variables directly predicted Memory Complaints using
regression or bivariate analysis. Interestingly, college attendance appears to
buffer the effects of Memory Complaints on Psychological Well-being. Thus,
the role that education plays continues to be unclear, although the trend
seems to be that higher education increases the elderly's chances of being
mentally healthy and happy. Further research is obviously warranted.

W The hypofltesis that
friends and siblings would provide more emotional support than children

was not supported by this research. This sample of elderly named children,
female friends, sons, brothers, sisters, and male friends (in that order).
Furthermore, women were more apt to name daughters and men were more
apt to name male friends. In addition, college attendance predicted number of
daughters and amount of emotional support from daughters and female
friends.

The fact that daughters were the most frequently named providers of
emotional support demonstrates that they continue to be important in the
lives of their elderly parents. This is apparently not merely due to a sense of
obligation. Elderly parents in the present sample see their daughters (and
sons) as people they can "talk to about their troubles or concerns." The
finding that females were more apt to name daughters and female friends is
consistent with other studies (Antonucd 8r Akiyama, 1987). Thus, friends

 

65

assume an important role in the soda] network of the elderly, however, at
least for the present population, they are not as important as daughters, even
though female friends are somewhat more frequently listed than sons.

When the regression results are looked at in more detail, two distinct
groups of support people emerge, with daughters, female friends, and sons
accounting for more emotional support than siblings and male friends.
Siblings have been found in other research to be important individuals in the
soda] network of the elderly (e. g. Cidrelli, 1990). One possible explanation for
the difference found in the present study is that uner Cidrelli's research
which only investigated the role of siblings, other individuals in the network
were also examined. It may be that siblings are important, but that children
and female friends assume a more important role in the lives of the able
elderly. Another possibility is that siblings are "close" to each other, but the
satisfaction with amount of emotional support is less than that with children
or female friends. More research is needed to answer these questions.

F‘mally, given that men were more apt to name males friends than
children, siblings, or female friends, and that male friends had the weakest
assodation with amount of emotional support, it is likely that elderly men
have a smaller base of confidants than elderly women. Thinking back to the
87% of married males who listed a spouse as their one "confidant" as
compared to 59% of married women, and that this research also points to a
gender interaction with number of confidants (women having more than
men), it appears that, if elderly individuals are told to name people they feel
close to othir; than their spouse, men list fewer people, and the ones they do
list tend to be less emotionally supportive.

 

A somewhat surprising finding was the lack of relationship between
actual contact with close family and friends and psychological well-being. As
noted above, frequency of contact, espedally telephone calls have been shown
to be significantly related to fewer mental health problems and greater life
satisfaction in the elderly (Revicki 6: Mitchell, 1986; DeWit, Wister, & Burch,
1988). Results of the regression analysis, however, suggest that the only
significant support variables are number of confidants and amount of
emotional support. Even in the Bivariate results discussed above, there were
no significant correlations between any other support measures and
Psychological Well-being

Given these results, it was dedded to see if contact with people in the
support network possibly had an indirect effect on Psychological Well-being
through its effect on confidants and amount of emotional support. Results of
bivariate correlational analysis suggest that this is, in fact, the case. Number of
phone calls was positively correlated with both number of confidants (; = .38;
p < .000) and amount of emotional support (1; = .38; p < .000); total number of
visits was positively correlated with number of confidants (r = .20; p < .005)
and amount of emotional support (r; = .22; p < .003). A plausible explanation, '
therefore, is that amount of quality support directly affects psychological well-
being in the able elderly, but that this is measured or evaluated by individuals
through the amount of time spent either talking on the phone or visiting
people they feel close to. In addition, the effect of college attendance on
number of confidants may mean that those with higher education will find
their sodal network filled with more people they can talk to about their
troubles or concerns which in turn suggests that they will experience greater
psychological well-being.

67

CONCLUSION AND IMPLICATIONS

How do the above findings contribute to our understanding of what
constitutes "the Good Life" for the able elderly? If we re—examine Lawton's
(1986) model using the results of the present research as predictors of this
"Good Life" one fact stands out above the rest: W as
measured by quality of support, perception of health, and memory complaints
relates strongly to W. B havi ral m as
measured by actual health and memory function has a direct effect on well-
being through actual health status and an indirect effect through memory
function, but nevertheless, a reasonable conclusion is that it is the moaning
such conditions as deteriorating health and memory have for the elderly that
affects their happiness and satisfaction with their lives.

The aging process naturally brings about changes in health and
cognitive functioning. For the able elderly, however, perceiving these
changes as having little or no effect on their ability to engage in activities,
coming from a better than average educational background, and feeling they
can talk to people in their sodal networks about their troubles or concerns all
contribute significantly to less depression and greater life satisfaction.

The above conclusion assumes even greater importance when coupled
with the fact that the present sample of elderly did not have major physical
health, mental health, or memory problems. That is, even when health
and/ or memory may not be significantly impaired, believing they are, or
worrying about it results in more psychological distress for the able elderly.

The role of soda] support must be acknowledged as vital to the well-
being of this population. Having a confidant and feeling these confidants are
good listeners not only contributes to psychological well-being, but also

6 8
buffers the effect that perception of health can have on such symptoms as

depression.

The findings of this research have contributed to our understanding of
how soda] networks help the able elderly. Contrary to expectation, both
married men and married women were more likely to list their husbands
than other relatives or friends; however, when asked to name someone that
was not living with them, daughters and female friends were more often
dted by females and male friends were more often dted by males. Given that
daughters and female friends have the largest correlation with amount of
emotional support, this raises questions as to the vulnerability of elderly men
as compared to elderly women. In support of this concern is the finding that
women were more apt to name a confidant suggesting that even though men
feel close to male friends or others they listed, they do not confide in or feel
emotional support from them. It is their wives who function in this capadty.
The obvious concern then is what happens to these men when their wives
die or leave them. Given that virtually none of the present sample of males
were widowers, this question could not be addressed with the data available.
Clearly one goal of future research on the able elderly would be to examine
more closely the lives of widowed males.

The fact that little relationship between siblings, emotional support
and psychological well-being was found in this research is somewhat
puzzling given other data by such researchers as Cidrelli (1989). This
difference suggests that the role of brothers and sisters may be overshadowed
by others in the network, mainly children and female friends. Future studies
may be more informative if more detailed questions regarding satisfaction

with support were included in questionnaires or interviews.

69

Finally, the virtual lack of depression or other signs of psychological
distress in this sample of able elderly is very encouraging. Given that these
individuals function independently, drive cars, leave their neighborhoods at
least once per week, reach out to friends and family, and feel good about their
overall health is a sign that aging in our society need not be problematic, and
in fact can be a rewarding and fulfilling part of life.

It was uplifting to meet the individuals who agreed to partidpate in
this research project. Some of them rode their bicycles to the interview, they
were animated, interested, and eager to partidpate and help. One former
professor shared fasdnating anecdotes about his department at the University
and what it was like to work there 40 to 50 years ago. One woman, who had a
chronic, severe systemic illness, felt that her health had little or no impact on
her ability to do things, and rated her health as better than most people her
age. She was a single woman who looked up at me when she got to the
question on the BDI that pertained to increased thoughts about death and
said: "Well, of course, I think about death more often than I used to, but not
because I worry about dying or because I‘m depressed, but because I want to
get my life in order so that when I die my niece doesn't have a financial mess
to straighten out." Then there was the woman who rode her bike to the
interview and asked if there was an aerobics fitness program on campus for
the elderly. As our life span increases, vital, active individuals such as these
will increase in number. One crudal area of research, therefore, must be
further investigation into what contributes to the "Good Life" for the able
elderly. Clearly, perception of health, concerns about memory, educational
background, and emotional support stand out in the present research as
significant areas for further study.

APPENDICES

APPENDD( A

70

MICHIGAN STATE UNIVERSITY

 

M ”70006“ QINIC ' ova-rum Of ”YMMV MST WC 0 masses 0 ”I.”
W GI?) ”W

Enclosed is a flyer describing ”Coping with.Aging" sponsored by the
Michigan State University Psychological Clinic. This study will examine
'memory and mood in individuals 55 years of age or older. If you would be
willing to participate in this research. please call 355-9564 and ask for
Sharon Ruggles or the Aging Project.

All adults 55 or older are eligible, so if you know of anyone who
qualifies, please pass this information on. I: is not necessary to have
been an employee of Hichigan State‘University to be a part of the study.

Sincerely.

Ann-Marie Schee:baum,‘H.A.
Coordinator, Coping with Aging Project
14.5.11. Psychological Clinic

ANS/sat

Enclosure

It." ' u e- MIN-snow Anus/Fayed Weed!) ham-“-

71

COPING -WITH
AGING
PROJECT

0 need 100 volunteers
0 55 or older
. 0 need not be retired

 

 

As part of a research project , the MSU Psychological Clinic is
giving memory and mood evaluations to people over 55. The
first 100 who volunteer will be accepted, and people can
participate whether they are retired or not.

Those who participate can find out how they compare to others
in their age group. This can be helpful for those who may be
worried about how well they are doing as they get older.

For more information and scheduling,
please call 355-9564 and ask for Aging Evaluation.

 

 

APPENDIX B

72

RESEARCH CONSENT FORM
Coping with Aging Project

I hereby agree to participate in the assessment portion of the
Coping with Aging Project conducted by the Hichigan State University
Psychological Clinic. I understand that I will be interviewed and be
administered a series of tests. These are designed to measure memory
functioning. mood and aspects of daily living. I also understand that
these interviews wdll be audiotaped. A short time later I will be given
feedback on the results of these tests. If I require further assistance.
I will be provided referrals to appropriate agencies.

I understand that the interview and questionnaire results will be
held in strict confidence. and that I will be completely anonymous.
I understand that my participation in this study is completely voluntary
and I have the right to drop out of the study at any time. Hy right to
further referral and services from the Clinic will not be affected by my
decision to drop out of the study.

 

 

Name Bite

REFERENCES

REFERENCES

Albert, M. S. (1981). Geriatric neuropsychology.1g1_rm.l_gf£Mgngfl
MW 42, 835-850

Aldwin, C. M., Spiro, A., Levenson, M. R., 8: Bosse, R. (1989). Longitudinal
findings from the normative aging study. 1. Does mental health change

with age?P sygho Iggy and Ag’mg, 4(3), 295-306.

American Psychiatric Association (APA). (1980). W statistical
Wm). Washington. DC: American
Psychiatric Association.

Antonucci, T. C. G: Akiyama, H. (1987). An examination of sex differences 1n

social support among older men and women. W 1101/ 12), 737-
749.

Arling, G. (1976). The elderly widow and her family, neighbors, and friends.
lgurnal of Marriage; and the Family, 3_8_, 757-768.

Babchuk, N. (1978-79). Aging and primary relations. Intgrnag'gngl lgurn algf
Wm 2(2)137-151.

Beck, A. T. (1967). Demign: clinicall exgrimental and theoretical asm.
New York: Harper & Row.

Beck, A.T., Ward, C. H, Mendelson,M., Mock, J., &Erbaugh, I. (1961). An

inventory for measuring depression. Argygs gf angral Psyghiagy“, 4
53-63.

Berg, 5., Mellstrom, D., & Persson, G. (1981). Loneliness in the Swedish aged.
lgumal of Gerontology, &, 342-349.

Blazer, D. 6: Williams, C. D. (1980). Epidemiology of dysphoria and

depression in the elderly population. MW
£714), 439-444.

Blazer, D. G. (1983). Impact of late-life depression on the social network.
Ameisaaloumalntfiwhiamb 149(2),162-166.

Blazer, D., Hughes, D. C., & George, L. K. (1987). The epidemiology of
depression in an elderly community population. The Gerontological
men of America, 21(3), 281-287.

Breslow, R., Kocs1s, I & Belkin, B. (1981). The contribution of the depressive
perspective to memory function in depression. American Igurnal of

13mm 1&3, 227-230

 

73

74

Cavanaugh, S. & Wettstein, R. M. (1983). The relationship between severity
of depression, cognitive dysfunction and age in medical inpatients.

American lournal of Psychiagy,__,140 495.

Cicirelli, V. G. (1982). Sibling influence throughout the lifespan. In M. E.
Lamb & B. Sutton-Smith (Eds. ), Sibling Relationships. Thgir ngtgg and

We litespan (pp. 267-284). Hillsdale, NJ: Erlbaum.

Cicirelli, V. G. (1989). Feelings of attachment to siblings and well-being in
later life. Echology and Agg’ g, 3(2), 211-216.

Ciocon,I. 0. dc Patten]. F. (1988). Age-related changes 111 human memory.
Normal and abnormal. 93:23:32. 43(10), 43-48.

 

Cohen, R. M., Weingartner, H., Smallberg, S. A., Pickar, D., 6: Murphy, D. L.
(1982). Effort and cognition in depression. Archives of General

rim :9, 593-597.

Crook, T. H. Gr Larrabee, G. I. (1990). A self-rating scale for evaluating
memory in everyday life. Psycholggy and Agm' g, §(1), 48-57.

Depner, C. E. 6: Ingersoll-Dayton, B. (1988). Supportive relationships in later
life. WW 3(4). 348-357.

Derogatis, L. R. (1977) WWW
procedures for the SCL- 90. Baltimore, MD: Johns Hopkins University
School of Medicine, Clinical Psychometrics Unit.

Derogatis, L. R., Li an, R. 5., Covi, L., Rickels, K, 8: Uhlenhuth, E. R. (1970).
Dimensions 0 outpatient neurotic patholo :Comparison of a clinical
versus an empirical assessment. 9414M

mm :2. 164-171.

Derogatis, L. R. 8: Spencer, P. M. (1983). B51 Manual I: Administradgn and
W. Baltimore, MD: Johns Hopkins University School of
Medicine, Clinical Psychometrics Unit.

DeWit, D. J., Wister, A. V., & Burch, T. K. (1988). Physical distance and social
contact between elders and their adult children. Research on Aging,
19(1), 56-80.

Esser, S. R. 6: Vitaliano, P. P. (1988). Depression, dementia, and social

suppogséaomltgmadg' ngl Iddmg gf Agm' g gd Human Mldpment,

Gallagher, D., Neis, 6., Gt Thompson, L. W. (1982). Reliability of the Beck

Depression Inventory with older adults. W
W 59(1) 152-153

7 5
Gallagher, D., Breckenbridge, J., Steinmetz, J., 8: Thompson, L. (1983). The
Beck Depression Inventory and Research Diagnostic criteria:

Congruence m an older population. lournal of Consulting and Clinical
new 51(6) 945946

Gass, C. S. & Russell, E. W. (1986). Differential impact of brain damage and

depression on memory test performance. MW
Message» 52. 261-263

Gatz, M. 8: Hurwicz, M. (1990). Are old people more depressed? Cross-
sectional data on center for epidemiological studies depression scale

factors. _sxshelozxandm 5(2), 284-290

Hagestad, G. O. (1987). Able elderly 1n the family context: Changes, chances,
and challenges. The Qerontglggical Society of America, 21(4), 417-422.

Hale, W. D., Cochran, C. D. ,,Hedgepeth B. E. (1984). Norms for the elderly on

the Brief Symptom Inventory WWW
Exclaim 52(2). 321-322

Israel, B. A. (1982). Social networks and health status: Linking theory,

research and practice. Patient Counseling and Health Education, g2), 65-
79.

Kahn, R. L., Zarit, S. H., Hilbert, N. M., 6: Niederhe, G. (1975). Memory
complaints and impairment in the aged: The effects of depression and

altered brain functionAr Argv vgs of General Psychiady, 3_, 1569-1573.

Kimmel, D. C. (1988). Ageism, psychology, and public policy. m
Psychologist. 43(3). 175178.

Krause, N. (1987). Satisfaction with social sup rt and self-rated health in
older adults. Thg Qerontglggigal Society 0 America, _2_7(3), 301-308.

LaRue, A., Dessonville, C., 6t Jarvik, L. F. (1985). Aging and mental disorders.

In]. Birren&K.W.Schaie(Edsu)H§m1M1£9_flL¢_EW
pp. 664-702.

Lawton, M. P., Moss, M., Fulcomer, M., 8: Kleban, M. H. (1982). A research
and service oriented multilevel assessment instrument. loumal of

9mm 12(1). 91-99.

Lawton, M P. (1986). Functional assessment. In L. Teri & P. Lewinsohn (Eds)
W New York: Springer

Lee, G. R. & Ellithorpe, E. (1982). Intergenerational exchange and subjective

well-being among the elderly. lgdmal Qf Marriaggan and 1115:le y, 4_,4U
217-224.

7 6
Lee, G. R. 8: Ishii-Kuntz. (1987). Social interactions, loneliness, and

emotional well-being among the elderly. Wang, 2(4), 459-
482.

Longino, C. F. 8c Lipman A. (1981). Married and spouseless men and women
in planned retirement communities: Support network differentials.

We. 43. 169-177

McAllister, T. W. 8: Price, T. R. P. (1982). Severe depressive pseudodementia

with and without dementia. American lournal of ngchiap'y, 1&6), 626-
629.

Murphy, E. (1982). Social origins of depression in old age. Brit_i§h lpumal pf
menu 141. 135-142

Murphy, E. (1983). The prognosis of depression in old age. British lpurnal of
chhiagy, 1_4_2_, 111-119.

Murrell, S. A. Himmelfarb, S., 8: Wright, K. (1983). Prevalence of d ression

and its correlates 1n older adults. Amggcan lpurnal pf Epi 11mdg mid ggy,
_1__,17 173.

Myers, I. K., Weissman, M. M., Tischler, G. L., Holzer, C. E., Leaf, P. I.,
Ovaschel, H., Anthony,I. C. Boyd, I. H., Burke, I. D., Kramer,M., 8:
Stoltzman, R. (1984) Six-month prevalence of psychiatric disorders in

three communities: 1980-1982. WM g, 959-
967.

Newmann, I. P. (1989). Aging and depression. Psychology and Ag’ g, 4(2),
150-165.

Perlmutter, M. (1978). What is memory aging the aging of? W
melon 1.4. 330345.

Powers, E. A. 8: Bultema, G. L. (1976). Sex differences 1n intimate friendships
of old age We 38. 739-747

Preston, D. 8: G1imes, I. (1987). A study of difference in social support.
lournal pf Qerontplpgical nging, _1_3_(2), 36-40.

Revicki, D. A. 8: Mitchell, I. (1986). Social support factor structure in the
elderlY- W 8(2). 232-248.

Ronsman, K. M. (1988, Ianuary/ February). Pseudodementia. gm
Nu_r§ing, pp. 50-52.

Rozzini, R., Bianchetti, A., Carabellese, C., Inzoli, M., 8: Trabucchi, M. (1988).

Depression, life events and somatic symptoms. The Gerdntplggigg
w 28(2) 229-232

77

Ryan, M C. 8: Patterson, I. (1987). Loneliness in the elderly. W
Gerontolpgigal Nursing, 13(5), 6-12.

Salzman, C. 8: Gutfreund, M. I. (1987). Clinical techniques and research
strategies for studying depression and memory. In L. W. Poon (Ed.),
BMW.
Washington, DC. APA Press.

Schaie, K. W. (1988). Ageism in psychological research. American
Psychplpggg' t, £6), 179-183.

Shanas, E. (1979). Social myth as h ypoztheses The case of the family relations
of old pe0ple W ypo12(1) 3-9

Shanas, E. (1980). Older people and their families: The new pioneers.
Iournal of Marriagg and dge Family, 4;, 9-15.

Small, G. W. 8: Jarvik, L. F. (1982). The dementia syndrome. Met,
1443-1446.

Spreitzer, E. 8: Snyder, E. (1974). Correlates of life satisfaction among the
aged. lournal of ficntology,_, 29 454-458.

Storandt, M., Botwinick, I., Danziger, W. L., Berg, L., 8: Hughes, C. P. (1984).
Psychometric differentiation of mild dementia of the Alzheimer's type.

Archives of leogy, fl, 497-499.

Storandt, M., Botwinick, I., Danziger, W. L. (1987). Longitudinal changes.
Patients with mild SDAT and matched healthy controls. In L. Poon (ed)
. 1 f _ - 1,

 

Stroller, E., (1984). Self-assessments of health by the elderly: The importance

of informal assistance. lournal of Health and Social Belgvior, 25, 260-
270.

Surtees, P. G. (1980). Social su p,port residual adversity and depressive
outcome My, 15. 71-80

Vaux, A. (1985). Variations 1n social su pport associated with gender,
ethnicity. and age W40). 89-110

Ward, R A. (1985). Informal networks and well-being in later life: A
research agenda. Thg Eontplpgical meg of America, §(1),55-61.

Wechsler, D. 8: Stone, C. P. (1983). Manual: Wechsler Memogy Scale. New
York: Psychological Corporation.

78

Williams,I. M., little, M. M., Scates, S., 8: Blockman, N. (1987). Memory
complaints and abilities among depressed older adults. lgmal pf

Consulting and Clinical Psychology, 5_5(4), 595-598.

Wood, V. 8: Robertson, I F. (1978). Friendship and kinship interaction.
Differential effects on the morale of the elderly. ldurnal pr mggg Qd

mantel» 19. 367-375

Yesava e, I A. (1987). The use of self-rating depression scales 1n the elderly.

In8 L. W. Poon (Ed), Handbook for clinical memogy assessment of older
adults. Washington, D. C. APA Press, 213-217.

Zarit, S. H. (1982). Affective correlates of self-reports about memory of older
people WW mm 1(2). 25-35

Zung, W. K. (1965). A self-rating depression scale. Archives of Gengal
Psychiag, 1;, 63-70.