HIHHHHHWHIHIIWIIHUNI

I
I
r

WIHUHHWIWI

   
 

THESIS

llllllllllll!lllllllllllllllllllllllllllllllllllllllllllll

1293 01016 9542

This is to certify that the
thesis entitled
MEMORY COMPLAINTS, DEPRESSION, AND MEMORY
FUNCTIONS IN THE ABLE ELDERLY
presented by

JAMES DENISON ‘ JONES

has been accepted towards fulfillment
of the requirements for

~lltl,t\\de,gree in W

ale/We

Major professor

Date April 6, 1995

0-7639 MS U is an Affirmative Action/Equal Opportunity Institution

 

LIBRARY
Mlchlgan State
Unlverslty

 

 

 

PLACE it RETURN BOX to romovo this checkout from your rooord.

TO AVOID FINES rotum on or botoro doto duo.

DATE DUE DATE DUE DATE DUE

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

MSU to An Affirmative Action/Emu! Opportunity Institution

 

MEMORY COMPLAINTS, DEPRESSION, AND MEMORY
FUNCTIONS IN THE ABLE ELDERLY

BY

James Denison Jones

A THESIS

Submitted to
Michigan State University
in parital fulfillment of the requirements
for the degree of

MASTER OF ARTS

Department of Psychology

1 995

ABSTRACT

MEMORY COMPLAINTS, DEPRESSION, AND MEMORY
FUNCTIONS IN THE ABLE ELDERLY

BY

James Denison Jones

The present study investigated the relationship between memory
complaints, self-reported depression, and objective memory performance in a
sample of community dwelling older adults (N = 121 ). Participants were
administered a depression measure (Geriatric Depression Scale), an everyday
memory task (Rivermead Behavioural Memory Test). a traditional laboratory
memory task (Logical Memory I), and a memory complaint questionnaire
(Memory Assessment Clinics Self-Rating Scale). Contrary to expectations, given
methodological refinements over some previous research, the everyday memory
scores did not account for the most variance in the memory complaint scores.
Results indicate that the depression scores accounted for the most variance (r2 =
.195, p < .001) in the memory complaints scores, followed by the everyday
memory task scores (r2 = .127, p < .001 ). In addition, there were no significant
gender differences found in these measures. Implications of these data are
discussed.

In memory of my dad, Paul Thomas Jones

ACKNOWLEDGMENTS

I would like to thank my thesis chair, Dr. Norman Abeles, for his valuable
contributions and guidance throughout this project. I would also like to thank the
other members of my thesis committee, Dr. Robert Caldwell and Dr. John Hurley,
for their constructive feedback and guidance. I would also like to express my
gratitude to Dr. Greg Lamberty for his guidance and patient support. Finally, I
would like to acknowledge Debra Cole and the rest of my family for their

enthusiasm and support throughout this process.

TABLE OF CONTENTS

LIST OF TABLES ....................................................................................... vi
INTRODUCTION ........................................................................................ 1
DEPRESSION AND AGING ............................................................ 3
GENDER DIFFERENCES ............................................................... 6
MEMORY COMPLAINTS, DEPRESSION, AND
OBJECTIVE PERFORMANCE ............................................. 7
HYPOTHESES ........................................................................................... 17
METHOD .................................................................................................... 18
Participants ...................................................................................... 18
Measures ......................................................................................... 18
Geriatric Depression Scale ................................................... 18
Memory Assessment Clinics Self-Rating Scale .................... 19
Rivermead Behavioural Memory Test ................................... 20
Storandt Brief Dementia Battery ........................................... 22
Procedure ........................................................................................ 23
RESULTS ................................................................................................... 24
Descriptive Statistics ........................................................................ 24
Gender Effect Hypothesis ................................................................ 24
Memory Complaint and Performance Hypotheses .......................... 24
DISCUSSION ............................................................................................. 34
LIST OF REFERENCES ............................................................................. 39

Table 1
Table 2
Table 3
Table 4

Table 5

LIST OF TABLES

Means and Standard Deviations of Measures ....................... 25
Intercorrelations Among the Measures .................................. 26
Gender-Based Tests of Mean Differences ............................ 27

Hierarchical Multiple Regression Analyses Predicting
MAC-S Ability Raw Score ...................................................... 32

Hierarchical Multiple Regression Analyses Predicting
MAC-S Frequency of Occurrence Total Raw Score .............. 33

INTRODUCTION

Discussion of age effects on memory is no simple task. One of the most
difficult problems that must first be overcome is defining memory itself and, more
importantly, describing the actual processes that are involved in "memory." Since
memory and learning are closely associated, LaR ue (1992) offers a model (pp. 55-
60), that divides memory into four different stages based on information-
processing. The first is sensory memory (SM) which is described as a short-lived,
relatively complete representation of an external stimulus. This information is
subject to almost immediate decay unless it is transferred into the second stage,
short term memory (STM). Research indicates that there are age-related deficits in
the operation of STM (Ceralla, Poon, & Fozard, 1982; Walsh, Till, & Williams,
1978). The second stage, STM, is a temporary, limited capacity store, which
requires rehearsal and conscious mental effort. Many years of research indicates
that the capacity for STM is seven (12) "pieces" of information which are held for a
relatively short period of time (approx. 10-30 seconds). The investigations seem to
support that STM is relatively unaffected by normal aging. The third stage is called
long term memory (LTM) which is the, "relatively permanent, unconscious
repository of acquired information that is not restricted in its capacity" (LaRue, p.
56), and includes both retrieval and recognition of the information stored.
Research in laboratories and in clinical work seems to show that LTM is most
affected by the aging process. The final stage of memory is tertiary memory (TM)
or the ability to recall remote information, events, and data from earlier in people's

lives. "Studies of TM have had variable outcomes, but most suggest some age

related decline" (LaRue, p. 61). Memory decline, and cognitive decline more
generally, seems to be a very complex subject of study. It is clear that there is no
simple causative factor that leads to the observed decline. The etiology and
mechanisms involved in the deficits are equivocal and beyond the scope of this
paper. Williams, Denney, and Schadler (1983) stated that, "with respect to
memory, the reports of elderly adults correspond well with the results of memory
research, which indicates that most memory abilities tend to decrease with
increasing age" (p. 147). These deficits are thought to be attributed to age-related
changes in the encoding and retrieval processes themselves (Shaw & Craik,
1989). More specifically, memory performance as reflected in objective memory
measures declines significantly with age. "The relationship is strong and linear
across the entire adult range" (Herzog & Rodgers, 1989, p. 180).

Additional research regarding memory decline has also pointed to older
persons' beliefs about memory and its relationship to actual performance, which
will be discussed below. Their responses converge on the conclusion that they
perceive that their memory is worsening with age (Cavanaugh, 1987). It is
somewhat unclear why this may be happening and what variables are involved
besides simply getting older. Memory rated by older people themselves has been
found to be strongly and positively related to physical health and negatively related
to depression (Herzog & Rodgers, 1989). Cavanaugh further postulates that part
of the subjective feeling of memory decline may be due to the acceptance of the
stereotype of memory decline with age (p. 275). Lachman, Steinberg, and Trotter
(1987) indicated that "there is evidence that elderly individuals and those that are
depressed...are less likely to make adaptive, optimistic attributions, but rather see
success as due to external, unstable, and specific causes" (p. 266).

Personality has also been found to be an important individual difference

factor that should be taken into account in understanding memory performance

(Cavanaugh & Murphy, 1986). Specifically, anxiety and depression have been
found to be two especially important aspects of personality involved in memory
(West, Boatwright, & Schleser, 1984).

What complicates this paradigm further is that memory performance
influences knowledge about memory (metamemory), since information about how
memory works can be acquired through its use (Cavanaugh & Perlmutter, 1982). It
has been found that self-evaluations may influence performance and subsequent
evaluation (Lachman, Steinberg, & Trotter, 1987). Cavanaugh and Murphy (1986)
state that "as long as one defines metamemory as statable knowledge about
memory...results have shown...that what people know about their memory is
related to how well they perform" (p. 386). It has been found that the aged often
appear deficient in spontaneously producing appropriate and effective memory
strategies when asked to remember, even though these strategies are helpful if
older adults are induced to use them (Murphy, Sanders, Gabriesheski, & Schmitt,
1 981 ).

One further problem with the general topic of memory decline in the
literature is that the assessment of memory decline is not well understood and
"that no comprehensive and well documented memory battery has been
developed" (Erickson, Poon, & Walsh-Sweeney, 1980, p. 379). Age sensitive
norms are only beginning to be developed. Data from metamemory measures
and measures of memory complaint indicate that memory beliefs are only
moderately accurate (Herrmann, 1982). Given this, the variable nature of
memory decline in older adults remains somewhat elusive.

DEPRESSION AND AGING

Clinical experience and epidemiological studies have consistently found

depression to be the second most common mental disorder in old age, exceeded

only by dementia (O'Conner, Pollitt, Roth, Brooks, & Reiss, 1990). Estimates of

the prevalence of depression vary, reflecting differences in sample
characteristics, criteria for case identification, and methods of evaluation.
However, Major Depression is believed to occur in approximately 1-8% of older
community-living adults (e.g., Blazer, Hughes, & George, 1987; Murrell & Meeks,
1991). Similarly, estimates from Scandinavian, United States, and European
studies indicated depression in roughly as many older adults (Murrell & Meeks,
1991). When milder forms of depression are included, estimates ranged from 10—
30% (e.g., Blazer, Hughes, & George, 1987; O'Conner, Pollitt, Roth, Brooks, &
Reiss, 1990).

Estimates dramatically increase when one looks at medically ill and those
older adults living in retirement and nursing homes. Jenike (1988) suggested 20-
30% of elderly patients with medical problems are significantly depressed.
Additionally, he wrote that "while those over age 65 account for about 11% of the
United States population, they commit about one quarter of all suicides" (p. 89).
In another striking study, it was suggested that depressive symptoms among the
medically ill ranged from 12-58% of older adults, depending on the diagnostic
procedure employed (Karanci, 1988). Stewart et al. (1991) found that in long-
term and acute care facilities, from 10% to 20% of elderly people have a major
depression.

Bieliauskas (1993) distinguishes primary depression and the depressive-
Iike symptoms among the elderly. He draws on Freud's (1917), Mourning and
Mela_ncholia. by citing that a fall in self-esteem is "the critical difference between
genuine pathological depression and the depressive-like symptoms which
accompany grief reactions" (p. 125). He further extrapolates this concept to
differentiating these similar, albeit qualitatively different, groups of symptoms for

the aged by stating that:

...without the key presence of lowered self-esteem, we are not
as likely to be dealing with primary psychological depression in
elderly patients as we are faced with reactive grief or mourning
and interpersonal loss, and complications of medical and
neurological illness (p. 125)

Schneider (1992) reported that there are different clinical manifestations,
course, outcome, and family history that distinguish depression first occurring
earlier in life (i.e., prior to ages 50-60). He reported that early onset depression
has been characterized by higher rates of depression in families, recurrent
episodes, and suicide attempts compared to late-onset depressives. Late-onset
depression was characterized by longer episodes, poorer outcome, and
chronicity. Complaints of concentration and memory difficulties are symptomatic
of depression and have been shown to be associated with aging (Herzog &
Rodgers, 1989) and being depressed was strongly associated with elevated
levels of other mental symptoms and with poor physical health (Lewinsohn,

Rohde, Seeley, & Fischer, 1991). The authors stated further that

although a number of provocative differences in the magnitude

of association with depression in the young-old and old age groups
were found, almost all the variables known from previous research
to be associated with depression in younger people were also
related to being depressed in older individuals (p. 258)

Given the prevalence of depression in the elderly, assessment of its
presence has become increasingly important. Researchers, clinicians, as well as
elderly persons, have also become increasingly concerned about the possible
presence of dementia, such as Alzheimer's disease and other organic dementing
processes. Memory is most often the cognitive function that is initially, and most
severely, affected by such processes (Bolla, Huff, Nebes, Holland, & Boller,
1988). Older persons generally report that their memory functioning decreases

as they advance in years. Therefore, it is important to elucidate whether memory

complaints from the elderly population are evidence for a dementing process or
due to other factors, such as depression.
GENDER DIFFERENCES

The literature suggests that there are consistent differences between
elderly women and men for both rates of depressive disorders and depressive
symptomatology and for rates of cognitive decline, including memory. However,
this finding is not universal.

Depressive symptomatology and depressive disorders have quite often
been found to occur more frequently and with greater prevalence in women.
Atchley (1976) reported that, in his sample, women showed an increased
prevalence of depressive disorders. In their Italian study of 317 community living
elderly, Carpiniello, Carla, and Rudas (1989) found a similar trend demonstrating
a higher prevalence of depression in elderly women. Lewinsohn et al. (1991)
found that, for women, old age was more highly associated with feelings of
dissatisfaction with important life roles, with experiencing physical disease, and
with marital conflict. In elderly people with some cognitive impairment, Fuhrer et
al. (1992) found that women tended to report more depressive symptomatology.

Some differ on this reported gender difference. Hale and Cochran (1983),
for instance, found that in a retirement community "older men and women have
distinct patterns of symptom expression even when gender groups are matched
on marital status and are similar in education and perceived health, it is worth
noting that the groups did not differ in level of depression" (p. 649).

Linn, Hunter, and Harris (1980) found depressive symptoms more prevalent in
older females, but they found that this gender difference was reduced to a
statistically insignificant level when marital status was controlled.

Cognitive decline, including memory abilities, has consistently been shown

to occur differently between the sexes. Wlederholt et al. (1983) found women to

show slower cognitive decline than men in a community sample of 1,692 elderly
subjects. Zelinski, Gilewski, and Schaie (1993) also found that elderly women
consistently out-perform elderly men on cognitive tasks. Specific to memory
abilities, the literature rather consistently shows that elderly women achieve
higher scores on measures of verbal selective reminding tasks, immediate recall
(Wrederholt et al., 1993), and general verbal memory procedures (e.g., Trahan &
Quintana, 1990) than do men. Interestingly, Herzog and Rodgers (1989) found
that women rate their own memory performance lower than men rated theirs.
They did find, contrary to the above literature, that there were no sex differences
on objective memory performance.

MEMORY COMPLAINTS. DEPRESSIONJ AND OBJECTIVE PERFORMANCE

Clinicians will recognize that memory complaints increase with age.
Researchers have been testing to see if memory complaints, in and of
themselves, were good predictors of actual (objective) memory performance.
The relationship between increased memory complaints and objective memory
decline is less than well established (Lamberty & Bieliauskas, 1993).

Some investigators have found no relationship between subjective
memory problems and objective measures. Additionally, researchers have found
that subjective memory complaints are more closely related to the presence of
depression. Chandler and Gemdt (1988) found non-significant differences
between those with and without memory complaints on objective memory
measures in a sample of psychiatric patients. Similarly, memory complaints were
significantly related to the presence of depression (Niederehe & Yoder, 1989)
regardless of objective memory performance (Kahn, Zarit, Hilbert, & Niederehe,
1975). O'Conner, Pollitt, Roth, Brooks, and Reiss (1990) found only a poor
correlation between memory complaints and objective performance in both

normals and those with depression. In their study, they found one-third of the

normals reported memory deterioration, although no (objective) evidence was
found on testing.

Williams, Little, and Scates (1987) found that "depressed adults
complained of greater memory problems than did non-depressed adults while
memory test performances of both groups were in the Average to Above Average
range" (p. 595). Memory complaints in the depressed elderly compared to both
younger adults (Chandler & Gemdt, 1988) and to demented adults (O'Conner,
Pollitt, Roth, Brooks, & Reiss, 1990) were not found to be indicative of actual
memory deficits as measured by objective memory tests. Sunderland, Watts,
Baddeley, and Harris (1986) found that interviewer ratings of the participants'
memory complaints correlated with ratings of depression but not with
performance on verbal memory tests. A more recent study by O'Boyle, Amadeo,
and Self (1990) found that, in each of their groups (depressed and
pseudodemented patients), memory complaints correlated strongly and positively
with depression scores and not with a general measure of cognitive ability,
though the pseudodemented group had a significantly higher cognitive complaint
score than did the depressed group. One major methodological flaw with this
study was that memory complaints were part of the diagnostic criteria for
pseudodementia, so the latter finding is not surprising. Similarly, some studies
have shown that as clinical depression lifts, memory complaints decrease
(O'Conner, Pollitt, Roth, Brooks, & Reiss, 1990; Plitkin, Mintz, & Jarvick, 1985),
though some have not found this (Niederehe, 1976). O'Conner, Pollitt, Roth,
Brooks, and Reiss felt that, "although elderly persons' assessment of their own
abilities are of considerable interest, they have little validity and should not be
used to make even tentative diagnoses of dementia” (p. 311). These findings
suggest that, although memory complaints can and do represent real memory

decline in some older adults, they seem to relate more to levels of depression

rather than to actual deficits in memory. The implication of these findings is that
memory complaints alone, without objective performance measures, should not
be used as a diagnostic tool with older adults.

However, other investigators have found that memory complaints did
predict actual memory performance on objective measures. Niederehe (1976)
found that the depressed subjects he tested did not make distorted self-
appraisals and were no more likely to underestimate performance than control
subjects. Cavanaugh and Murphy (1986) found that metamemory and
personality variables accounted for significant portions of the variance in memory
performance on two different tasks (free-recall list learning and free-recall prose;
p. 386).

Similarly, Cavanaugh and Poon (1989) found that performance on tasks
that are more meaningful (e.g., prose) were better predicted by their
metamemory questionnaire than performance on traditional laboratory memory
tasks. Tests of immediate and delayed recall of verbal memory were found to
significantly correlate with a memory complaint questionnaire (Sunderland,
Watts, Baddeley, & Harris, 1986). Taylor, Miller, and Tinklenberg (1992), in their
longitudinal study of community-dwelling older adults, observed decreased
performance roughly parallel at the group level with self-reported memory
decline.

Zelinski, Gilewski, and Thompson (1980) found that participants'
subjective memory ratings were related to performance on laboratory tasks.
Larrabee, West, and Crook (1990) found that self-rated memory and objective
performance shared 27.9 to 29.4 percent of the variance in healthy community-
dwelling people.

The investigations of the relationship between subjective memory

complaints, objective performance, and depression are not in agreement.

10

Review of the studies cited above yielded four general distinctions that can be
made which differentiate these studies and may solve the problem of conflicting
results. The first distinction is the different types of subjective memory
questionnaires/questions used in prior research. The second distinction is the
populations used in each of the studies; many were psychiatric samples as
opposed to community-dwelling older adults. Thirdly, the majority of the studies
used depression measures that may not be appropriate for the older adult
populations being considered. Lastly, the memory and cognitive measures
employed in nearly all of the investigations cited above may also have not
adequately assessed memory domains considered in the memory complaint
questionnaires. All four distinctions will be discussed more fully below.

The first potential problem associated with some of the studies that found
no relationship between complaints and performance is with the types of
subjective ratings or questions the older people have been asked. Cavanaugh
(1987) argued that "if memory changes with age, and if people have accurate
knowledge of their memory, then these changes should be reflected in self-
reports" (p. 271). As reflected above, many have not found this relationship. The
discrepant results may be a function of the self-report measures.

The metamemory questionnaires have evolved from asking very general
questions about perceived memory decline to more specific questions.
Niederehe and Yoder (1989) found that depression was associated most strongly
with reports of generalized difficulty, reflecting relatively nonspecific complaints.
Similarly, Zelinski, Gilewski, and Thompson (1980) stated that multiple indices of
memory evaluation definitely improved prediction. Given this, studies that used
questionnaires that have asked more general or non-specific questions regarding
memory would tend to find higher correlations with depression measures than

with performance on objective measures. In the Chandler and Gerndt (1988)

11

study, the investigators asked only three very global questions of their patients.
These questions were "(1) Do you have trouble remembering?, (2) Are there any
activities which you are prevented from participating in as a result of these
memory problems? and (3) Have you sought medical help or taken medication
for this memory problem?" (p. 85). They were scored as having significant
memory complaints if they answered yes to at least two of the three questions.
The result of this study was that "memory complaint was significantly more
common in depressed elderly subjects than in nondepressed subjects... [and]
there was no difference in memory testing scores between the patients with and
without memory complaint" (p. 88). Kahn, Zarit, Hilbert, and Niederehe (1975)
used subjects' spontaneous comments and Only one global question to rate
memory complaints in their sample. Similarly, O'Conner, Pollitt, Roth, Brooks,
and Reiss (1990) asked only three very general questions to rate memory
complaint. Both the previous studies attained results similar to those of Chandler
and Gemdt, that complaint was related more to level of depression than objective
performance. All three may have obtained their results because of the general
and nonspecific nature of their assessment of memory complaint.

There are studies that did use more comprehensive memory complaint
questionnaires that found only a poor relationship between complaint scores and
objective performance. For example the investigations of Popkin, Gallagher,
Thompson, and Moore (1982) and O'Boyle, Amadeo, and Self (1990) both
utilized more comprehensive measures of memory complaint but found no
relationship or a small relationship, respectively, so their negative findings cannot
be explained away as easily.

All of the studies cited that found small, but significant, or robust
correlations between memory complaint measures and objective performance

used more comprehensive and specific measures of memory complaint

12

(Cavanaugh & Poon, 1989; Larrabee, West, & Crook, 1990; Niederehe, 1976;
Sunderland, Watts, Baddeley, & Harris, 1986; Taylor, Miller, & Tinklenberg, 1992;
Zelinski, Gilewski, & Thompson, 1980).

The author's hypothesis is that memory complaints may, in fact, reflect
measurable memory problems, so this study will utilize the Memory Assessment
Clinics Self-Rating Scale (MAC-S; Crook & Larrabee, 1990) which was
developed by its authors to compensate for and improve on the questionnaires
available at the time [see METHOD section].

The second distinction between the investigations is the populations used
in each. Most of the studies that found a higher relationship between depression
measures and memory complaints or only weak or no relationship with
complaints and memory performance were conducted with psychiatric samples
(Chandler & Gerndt, 1988; Kahn, Zarit, Hilbert, & Niederehe, 1975; O'Boyle,
Amadeo & Self, 1990). This was not pervasive, however, as the studies of
Popkin, Gallagher, Thompson, and Moore (1982) and O'Conner, Pollitt, Roth,
Brooks, and Reiss (1990) did use community-based samples. Results from
studies utilizing older adults with significant psychiatric problems, significant
enough to be admitted to psychiatric units, may not be appropriate or
generalizable to community-dwelling older adults.

Without exception, the studies that did find significant relationships
between memory complaints and memory performance used community-dwelling
older adults. The commonly held belief that memory complaints should not be
used as a diagnostic tool for cognitive or memory dysfunction, because
complaints are assumed to be reflective of depression, may not be appropriate
for the significant majority of older adults that are living in their communities. This
study will utilize "able" older adults that are living independently in order to make

the results more applicable for the majority pf older persons.

13

All of the studies which looked at presence of depression, with the
exception of one, used self-report depression scales that may be inappropriate
for use with older adults. Traditionally, the researchers in this area have
employed depression measures such as the Beck Depression Inventory (BDI;
Beck, Ward, Medelson, Mock, & Erbaugh, 1961) and the Hamilton Rating Scale
for Depression (HDRS; Hamilton, 1967). Others have also used the Research
Diagnostic Criteria (RDC) for depressive disorder and depression criteria from
the Diagnostic and Statistical Manual of Mental Disorders, 3rd edition (1980).
The drawback with these depression measures and criteria is their inclusion of a
great many items that reflect somatic and/or vegetative symptoms of depression.
There is no question that somatic complaints and vegetative symptomatology are
part of the constellation of symptoms in depressive disorders. However, the
vegetative symptoms assessed on these depression inventories (e.g., sleep
disturbances, loss of libido, fatigue, etc.) are common among older adults,
irrespective of true depression. With increased age comes a greater likelihood of
physical problems that may have very similar symptoms as those reflected in
some depression measures. Clinical experience suggests that using these types
of items when assessing elderly populations may not be as useful as a measure
that focuses more predominantly on dysphoric mood. Norris, Gallagher, Wilson,
and Wrnograd (1987) reported that "due to several physical and content attributes
of the BDl, the Geriatric Depression Scale (GDS; Yesavage et al., 1983) was
found to be more desirable" (p. 994), especially when screening for depression in
elderly medical patients (Karanci, 1988). Given this assessment, prior
investigations that utilized measures that included somatic/vegetative symptoms
. may actually have been reporting exaggerated levels of depressive

symptomatology when used with older populations.

14

These findings call into question the results of some of the studies that
used the BDI or HDRS to assess the presence of depression in older adults. If
indeed these measures exaggerate levels of depression, the higher correlations
between subjective complaints and depression may be nothing more than a
methodological artifact. For these reasons, this study will utilize the Geriatric
Depression Scale (GDS), which was developed specifically to screen older adults
for depression, in order to get a more accurate picture of the presence of
depression [see METHOD section].

Finally, the lack of consistently strong relationships between subjective
complaints and objective performance may reflect inadequate approximations
between the types of objective memory measures and the subjective memory
complaint measures. Traditional investigations have used laboratory-based
memory tasks that are often different from memory tasks in everyday life, and are
unrelated to the context of the individuals' lives (Herzog & Rodgers, 1989). Some
of the laboratory tasks used in prior research include delayed recall of four
unrelated words (Chandler & Gemdt, 1988), paired associate learning (Kahn,
Zarit, Hilbert, Brooks, & Reiss, 1990), several immediate and delayed list-learning
tasks (Cavanaugh & Poon, 1989; Niederehe & Yoder, 1989; Popkin, Gallagher,
Thompson, & Moore, 1982; Taylor, Miller, & Tinklenberg, 1992; Zelinski,
Gilewski, & Thompson, 1980) and several prose or text passage memory tasks
(Cavanaugh & Poon, 1989; Larrabee, West, & Crook; 1990; Niederehe & Yoder,
1989; Popkin, Gallagher, Thompson, & Moore, 1982; Zelinski, Gilewski, &
Thompson, 1980). Some research has indicated that prose recall (Sunderland,
Watts, Baddeley, & Harris, 1986) or tasks that "are more meaningful"
(Cavanaugh & Poon, 1989) or "ecologically valid" (Zelinski, Gilewski, &
Thompson, 1980) were better predicted than performance on other laboratory

based tasks. These are the memory tasks that clinicians and researchers have

15

used with older adults in order to assess their memories. However, the memory
complaint questionnaires have not assessed these memory domains. For
example, people are never asked to rate how well they will remember or if they
have had any trouble remembering paired words or lists of unrelated items. The
lack of more consistent findings may be the result of the incompatibility between
the subjective memory complaint questionnaires and the memory tasks utilized in
many of the studies.

Larrabee, West, and Crook (1991) found, despite some of the above
findings, "significant associations...between self-rated and objectively measured
memory, provided that measures are constructed to enhance isomorphism
between rating scales and objective memory tests" (p. 475). In other words,
when the subjective measure was more closely related to tasks that were to be
objectively assessed, they did find an increase in relationship with objective
performance. Memory complaint questionnaires have traditionally asked people
to rate how well they are able to remember or how much trouble do they have
remembering names, faces, where they have placed objects, and other items
that are very applicable to circumstances that people encounter during their daily
routines. These "everyday" items are exactly the types of problems that concern
older adults and bring them to the attention of professionals.

Given these findings, subjective memory complaints should be more
robustly predicted by objective memory tasks that approximate everyday types of
memory tasks. The Rivermead Behavioural Memory Test (Wilson, Cockbum, &
Baddeley, 1985) was designed to assess everyday memory and will be used in
the present study.

Given the discrepancies in the literature regarding memory complaint and
its relationship with memory performance and levels of depressive

symptomatology, it is the aim of this study to test some of these questions. The

16

present study will utilize an improved subjective memory complaint questionnaire,
a more traditional laboratory-based memory task, a memory test designed to
assess everyday memory abilities, and a predominately mood-oriented
depression measure in a sample of community-dwelling older persons. It is
believed that, with these methodological improvements over past research,
subjective memory complaints will indeed predict actual objective performance.
These results will then be more directly applicable to the older person

community.

HYPOTHESES

1. It is hypothesized that composite scores on the MAC-S (Ability total
and Frequency of Occurrence total) reflecting higher levels of memory complaint
will most be most strongly related to scores on the everyday memory tasks.
Thus, it is hypothesized that the RBMT Standardized Profile score will account
for the greatest amount of variability in the MAC-S scores. Memory performance
on the Logical Memory l task, a more traditional laboratory memory task, will
account for the second greatest amount of variability with the same measure.
Lastly, depression scores from the GDS will account for the third greatest amount
of variability with the MAC-S scores.

2. It is hypothesized that there will be a significant gender effect. It is
predicted that females will tend to score lower on the MAC-S composite scores

(indicating greater levels of memory complaint).

17

METHOD
52mm

The participants were 121 "able" older adults (31 men and 90 women)
living independently in the community. All were aged between 51 and 91 with
mean ages of 71 (SD = 6.4) for men and 69 (SD = 9.4) for women. The
participants in this study were recruited through advertisements and flyers posted
in the community. Each participant volunteered for the study which was
advertised for those experiencing concern about their memory functioning. Each
was given an assessment battery and a 7-session mood and memory workshop
taught by advanced clinical psychology graduate students. This study included
data gathered from the first administration of the battery ("pre").

Measures
1. Geriatric Depression Scale (GDS)

The Geriatric Depression Scale is a 30—item measure that has a yes-no
format, with scores ranging from 0 ( no depression) to 30 (severe depression).
Subjects endorse items related to mood and psychological symptoms associated
with depression. A one-month test-retest stability coefficient of .85 was found
(Parmalee, Lawton, & Katz, 1989). The questionnaire was designed for use with
older people (Parmalee, Lawton, & Katz, 1989), focusing on subjective affective
states rather than vegetative symptoms, cognitive disturbances, thoughts of
death, spending less time on hobbies, and other areas of concern that may be
present in aged populations irrespective of presence of depression. They also
found that "the GDS has a generally clear...factor structure... [which is]

essentially unidimensional (general psychiatric distress; p. 337). The GDS

18

19

appears to be affected by age and sex differences, but these differences were
consistently small (Parmalee, Lawton, & Katz, 1989). In its preliminary validation
study, the GDS was found to be a reliable and valid self-rating depression
screening scale for elderly populations (Yesavage et al., 1983). Additionally, the
GDS was found to have good concurrent validity with the Hamilton Rating Scale
for Depression with the Melancholia Scale (r = 0.77; Salamero & Marcos, 1992)
and with psychiatric diagnoses in an older person community sample (Yesavage
et al., 1983). Zgourides, Spofford, and Doppelt (1989) supported GDS validity,
finding that the GDS significantly discriminated between depressed and
nondepressed elderly day-treatment clients.
2. Mem A ssment Clinics lf-Ratin S ale MA -

The Memory Assessment Clinics Self-Rating Scale (Crook & Larrabee,
1990) was developed to evaluate memory in everyday life. The measure
consists of two composite score factors. The first is the Ability total, which is the
sum of 18 items with questions about how people rate their memory in a number
of everyday situations (e.g., remembering the name of someone just introduced,
the date and day of the week, verbal directions to locations, which door you
entered when shopping in a department store, etc). These items are rated on a
5-point Likert scale, ranging from Very Poor to Very Good. The second is the
Frequency of Occurrence total, which is the sum of 20 items with questions about
how often people find themselves forgetting things (e.g., forgetting what you go
into a room for, forgetting an entire event such as a party, forgetting the name of
a familiar object, failing to recognize a person who recognizes you, etc.). These
items are also rated on a 5-point Likert scale, with ratings from Very Often to
Very Rarely. The measure also contains four global ratings items assessing
overall comparison to others, comparison to the best one's memory has been,

speed of recall, and concern or worry over memory function (p. 41).

20

The MAC-S items reflect everyday memory abilities and skills that are
common to all age groups (Crook & Larrabee, 1992). It has the advantage of
being brief, having a wide range of memory self-report factors, and a large
normative base that covers the adult range of 18 to 92 years (p. 53). The MAC-S
was found to show minimal associations with sex, WAIS Vocabulary, and
educational level, with sex accounting for the largest amount of variance (5.6%).

Crook and Larrabee (1990) found that the GDS correlated significantly, in
a negative direction, with 15 of 16 MAC-S variables, while only seven of these
correlations accounted for greater that 6% of MAC-S test score variance (p. 50).
Their study concluded that "the data suggest that the MAC-S Ability and
Frequency of Occurrence total scores, Frequency of Occurrence factor scores,
and global items are most likely to be associated with affective variation" (p. 52).
3. Rivermead Behavioural Memory Test (R BMT)

The Rivermead Behavioural Memory Test was developed to detect ‘
impairment of everyday memory functioning and to monitor changes following
treatment for memory difficulties (Wilson, Cockbum, & Baddeley, 1985). The
RBMT is "a bridge between laboratory-based measures of memory and
assessments obtained by questionnaire and observations" (Wilson, Cockbum,
Baddeley, & Hiorns, 1989, p. 857). The test was developed to provide analogues
of everyday memory situations (Wilson, Cockbum, Baddeley, & Hiorns, 1989).

The subtests of the RBMT were chosen on the basis of the memory
difficulties reported by subjects in a study of memory problems in a head-injury
population (Sunderland, Harris, & Baddeley, 1983) and observations at the
Rivermead Center, Oxford, England (Wilson, Cockbum, & Baddeley, 1985).
There are twelve components of the RBMT, and are as follows: (I & ll)
Remembering a name (first (I) and last (ll)), (Ill) Remembering a hidden

belonging, (lV) Remembering an appointment, (V) Picture recognition, (VI)

21

Remembering a newspaper article (immediate and delayed), (Vll) Face
recognition, (Vlll) Remembering a new route (immediate), (lX) Remembering a
new route (delayed), (X) Delivering a message, (Xl) Orientation (nine questions
in both time and place), and (Xll) Date. Each component is allocated 2 points
(normal), 1 point (borderline), or 0 points (abnormal) depending on the raw score.
The Standardized Profile Score is the sum of these allocated points across the
components. This enables comparison across items (Wilson, Cockbum, &
Baddeley, 1985).

The validation study conducted by Wilson, Cockbum, Baddeley, and
Hiorns (1989) yielded extremely high inter-rater reliability (100%) between two
raters scoring 40 subject's profiles. The RBMT was designed with four parallel
forms (A,B,C,D). Reliability scores yielded correlations with version A of .86, .83,
and .88 respectively. The test-retest reliability with 118 patients whom were
tested twice yielded a correlation of .85 for the Profile Score.

The validity of the RBMT, in the above cited study, was assessed in four
ways. The first was comparing overall level of performance between brain
damaged patients and controls. The assumption was that brain damaged
patients would have substantially more memory problems. The results showed
that brain damaged patients had substantially lower scores than the controls.
The second attempt was to find concurrent validity with existing standard memory
tests. They found that correlations ranged from .20 to .63 on different measures
and they maintained that the RBMT "can be regarded as measuring overall
memory performance...assessed in terms of existing tests (Wilson, Cockbum,
Baddeley, and Hiorns, 1989; p. 863). Thirdly, since the test was designed to
reflect everyday memory, they compared the RBMT scores with observations of
memory lapses made by therapists. The correlation between RBMT Profile

Scores and the number of memory lapses reported by observers was -.75 for the

22

Standardized Profile Score (p_< .001). Lastly, they compared RBMT scores with
subjective ratings of memory problems by patients and relatives and found
correlations that suggest the RBMT is a valid indicator of everyday memory
problems.
4. Storandt Brief Dementia Battery

The Storandt brief dementia battery is an easily administered battery of
four neuropsychological tests. The battery classified successfully 98% of
patients with mild dementia of the Alzheimer's type and healthy older persons
matched for age, sex, and social position (Storandt, Botwinick, Danziger, Berg, 8.
Hughes, 1984, p. 497). The battery includes the Logical Memory l (LM) and
Mental Control (MC) subtests of the Wechsler Memory Scale (Wechsler, 1945),
Trailmaking, Form A (Armitage, 1946), and a word fluency test (C & P, Thurston
& Thurston, 1949.). The Logical Memory l subtest is a test of auditory-linguistic
memory requiring verbal recall of auditorally presented prose passages (Abikoff
et al., 1987). This subtest will be used as the laboratory-type measure of
memory for this study. The Trailmaking Test (Form A) is a timed visuospatial
tracking measure that involves tracing lines from circled numbers (e.g., 1 to 2, 2
to 3, etc.). The word fluency measure requires the person to name as many
words as possible that begin with the letters S and P in a 60 second interval. The
Mental Control subtest involves recall of well-practiced information (e.g., the
alphabet, counting backward from 20, and counting forward by three's). The
original study tested 84 persons, half diagnosed with senile dementia of the
Alzheimer's type (SDAT) and half were matched normals. Based on preliminary
correlational analyses, the brief battery was chosen from a larger number of
psychological tests. Following this, a discriminant function analysis was carried
out on only half of the sample of subjects with SDAT and matched normal control

subjects. The results were then applied to the second half of the sample. The

 

23

raw scores of each of the four measures are then entered into a computational
formula to yield a canonical score. Canonical scores greater than or equal to 0
are classified as demented; those with scores less than 0 are classified as
normal. The 1984 study found that only two people (one demented and one
normal) were misclassified using this brief dementia screening battery. Storandt
and Hill (1989) used the canonical score to evaluate 41 individuals with very mild
SDAT, 83 healthy participants, and 66 with mild SDAT. Using the canonical
formula and its cutoff point of 0, no one in the control group and only three people
in the mildly demented group were misclassified.
Procedure

Depression and memory measures were administered as part of a larger
battery. Only a portion of the measures used in this battery were under
investigation for this study. The participants were assessed by advanced
graduate students, as mentioned above. The examiners were trained and
observed by a clinical faculty member. The test order was fixed in order to
reduce the possibility of serial effects and fatigue. All measures were
administered and scored according to each measure's published administration
and scoring rules. Each participant's raw data were re-scored for accuracy and

coded by the present investigator.

RESULTS

Descriptive Statistics

The Memory Assessment Clinics Self-Rating Scale (MAC-S), Total,
Rivermead Behavioural Memory Test (RBMT), Logical Memory I score, and the
Geriatric Depression Scale (GDS) were used in testing the hypotheses. Means
and standard deviations for the measures are presented in Table 1.
Intercorrelations between the measures are shown in Table 2.
Gender Effect Hypotheses

The gender hypothesis stated that there would be a significant difference
between males and females on the MAC-S Ability total and Frequency of
Occurrence total composite scores, with females' scores indicating greater levels
of memory complaint; it was not supported. The t-tests on gender revealed no
significant difference between males (M = 57.1) and females (M = 57.0) on MAC-
S Ability Scores, 1(119) = .070, p = .945. The t-tests on gender also revealed no
significant difference between males (_M_ = 63.9) and females (Lil = 64.5) on MAC-
S Frequency of Occurence Scores, 1(119) = -.28, p = .78. Table 3 shows the t-
test results for each variable.
Memory ComplainLand Performance Hypotheses

The hypothesis stated that the everyday memory tests (R BMT) would
account for the greatest amount of variability in the MAC-S composite scores
(Ability total and Frequency of Occurence total). Additionally, the Logical
Memory task and the Geriatric Depression Scale would account for the second

and third greatest amount of variability, respectively. The hypothesis that the

24

Table 1

25

Means and Standard Deviations of Measures (N = 121)

Vgrigble

Age

Education
MAC-St

Ability total

Frequency of Occurence total
RBMT Standardized Profile Score
Logical Memory Ii

GDSI

i The means and standard deviations are based on raw scores

2
§

 

69.49
14.64

57.02
64.35
18.84
17.93

7.17

10.40

10.83
4.65
7.16
5.91

3
3

29
27

91
21

82
92
24
34
24

 

26

Table 2
Intercorrelations Among the Measures (N = 121)

2.. 3
MAC-S
1. Ability .684‘" 356*"
2. Frequency of Occurrence .296"
3. RBMT

4. Logical Memory I

5. Geriatric Depression Scale

* = significant at .05 level
* = significant at .01 level

* = significant at .001 level

I.“

fit

.302

.236"

.568***

IS”

2442*"

-.489***

-.243**

-.202*

 

27

Table 3

Gender-Based Tests of Mea_n Differences

Maria—ble M_ean
MAC-S Ability Total
Males 57.1 29
Females 56.978

MAC-S Frequency of Occurrence Total

Males 63.871

Females 64.511
RBMT Standardized Profile Score

Males 18.226

Females 19.044

Logical Memory l
Males 18.613
Females 17.700
Geriatric Depression Scale
Males 6.194
Females 7.51 1

9.190
10.826

1 1 .427
10.682

4.295
4.773

8.260
6.772

6.129
5.829

.945

.778

.400

. 286

28

RBMT would account for the greatest variance in MAC-S composite scores was not
supported. Each of these analyses will be discussed below.

The hypothesis that the RBMT would account for the greatest amount of
variability in the MAC-S Ability total was not supported. The RBMT did, as
hypothesized, account for a greater amount of variance than did Logical Memory.
The analyses to derive the relative contribution of each of the measures to the
dependent measure were conducted in three steps. The first step was simply finding
the correlation between the GDS, RBMT, Logical Memory scores and the MAC-S
Ability total. The second step was entering the measures two at a time to see the
additional change in variance accounted for using a two-measure model as
predictors. Lastly, all of the variables were entered simultaneously, giving Beta
weights for each measure. The final computation was finding the relative change in
variance accounted for by each measure when added to the two-measure model
found in step two. Results from regression analyses predicting MAC-S Ability total
are summarized in Table 4.

Step One

First GDS, RBMT, and Logical Memory were used to predict MAC-S
Ability total. In Table 4 it can be seen that the GDS score by itself accounted for
nearly 20% of the MAC-S Ability total score variance (R2 = .195, p < .001). The
RBMT Standard Score sum accounted for nearly 13% of the variance (R2 = .127,

p < .001). Finally, Logical Memory score accounted for 9% of the variance (R2 =
.091, p < .01).

Step Two

The second step was a series of two-variable combinations as predictors of

the dependent measure. The first was the combination of GDS and RBMT

(2a) which accounted for 26% of the variance (R2 = .261, p< .05).

29

Adding RBMT to GDS, resulted in an increase of 7% to the variance, (AR2 =
.066, p < .05). Conversely, adding GDS to RBMT yielded an increase of
13% to the variance (AR2 = .134, p = < .01). The second combination of
GDS and Logical Memory (2b) accounted for 24% of the variance (R2 =
.242, p < .001). Adding Logical Memory to GDS yielded an increase of
nearly 5% to
the variance (AR2 = .047, p <05). Conversely, adding GDS to Logical Memory
yielded an Increase of 15% to the variance (AR2 = .151, p < .001). Lastly, the
combination of RBMT and Logical Memory (2c) accounted for nearly 14% of the
variance (n2 = .137, p < .001). Adding RBMT to Logical Memory resulted in an
increase of 5% to the variance (AR2 = .051, p < .05). And adding Logical Memory to
RBMT yielded an increase of nearly 2% to the variance (AR2 = .015, ns).
Step Three

In the final step, all variables (GDS, RBMT, and Logical Memory) were
entered into the model as predictors. 27% of the MAC-S Ability Total was
explained by the three variables (R2 = .270, p < .001; GDS Beta = -0.370, p < .001,
RBMT Beta = 0.203, p < .05, Logical Memory Beta = 0112, ns). Finally, the
incremental change in variance accounted for (AR?) was calculated as the change
in total R2 for the inclusive model (R2 = .270) from the R2 contributions of the
variable combinations entered in Step 2. Adding Logical Memory to GDS and
RBMT (2a) yielded an increase of less than 1% to the variance (AR2 = .009, ns).
Adding RBMT to GDS and Logical Memory (2b) yielded a significant increase of
nearly 3% to the variance (AR2 = .027, p < .05). Finally, adding GDS to RBMT
and Logical Memory (20) yielded a significant increase of nearly 13% to the
variance (AR2 = .128, p< .001).

The hypothesis that the RBMT would account for the greatest amount of

variability in the MAC-S Frequency of Occurrence total was also not supported.

30

Again, the RBMT did, as hypothesized, account for a greater amount of variance
than did Logical Memory. The analyses to derive the relative contribution of each of
the measures to the dependent measure were conducted in three steps (see
description above). Results from regression analyses predicting MAC-S Frequency
of Occurrence total are summarized in Table 5.
Step One

First GDS, RBMT, and Logical Memory were used to predict MAC-S Ability
Total. In Table 5 it can be seen that the GDS score by itself accounted for 11% of
the MAC-S Ability Total score variance (R2 = .113, p < .001). The RBMT Standard
Score sum accounted for nearly 8% of the variance (R2 = ..078, p < .01). Finally,
Logical Memory score accounted for nearly 3% of the variance (R2 = .028, ns).
Step Two

The second step was the same series, as stated above, of two-variable
combinations entered as predictors of the dependent measure. The first was the
combination of GDS and RBMT (2a) which accounted for 15% of the variance (R2 =
.154, p < .001). Adding RBMT to GDS, resulted in an increase of 4% to the
variance, (AR2 = .041, p < .05). Conversely, adding GDS to RBMT yielded an
increase of nearly 8% to the variance (AR2 = .077, p < .01). The second
combination of GDS and Logical Memory (2b) accounted for 12% of the variance
(R2 = .124, p < .001). Adding Logical Memory to GDS yielded an increase of 1% to
the variance (AR2 = .011, ns). Conversely, adding GDS to Logical Memory yielded
an increase of nearly 10% to the variance (AR2 = .095, p < .001). Lastly, the
combination of RBMT and Logical Memory (20) accounted for nearly 8% of the
variance (R2 = .078, p < .01). Adding RBMT to Logical Memory resulted in an
increase of 5% to the variance (AR2 = .05, p < .05). And adding Logical Memory to
RBMT yielded no increase to the variance (An2 = .00, ns).

31

M
In the final step, all variables (GDS, RBMT, and Logical Memory) were

entered into the model as predictors. Nearly 16% of the MAC-S Frequency of
Occurrence total was explained by the three variables (R2 = .155, p < .001 ; GDS
Beta = -0.286, p < .001, RBMT Beta = 0.216, ns, Logical Memory Beta = 0012, ns).
Finally, the incremental change in variance accounted for (AR 2) was calculated
as the change in total R2 for the inclusive model (R2 = .155) from the R2
contributions of the variable combinations entered in Step 2. Adding Logical
Memory to GDS and RBMT (2a) yielded no increase to the variance (AR2 = .000,
ns). Adding RBMT to GDS and Logical Memory (2b) yielded a significant
increase of 3% to the variance (AR2 = .031, p < .05). Finally, adding GDS to
RBMT and Logical Memory (20) yielded a significant increase of nearly 8% to the
variance (AR2 = .078, p< .01).

32

Table 4

Hirri MltilR inAnI Priin
-Ail'TtRw r

 

T I ml n=121

1. First step entry 5
a. Geriatric Depression Scale .442
b. RBMT sum of std. scores .366
c. Logical Memory .302

2. Second step entry

a. cos & RBMT .511
Adding RBMT to cos
Adding cos to RBMT

b. GDS 81 Logical Memory .492
Adding Logical Memory to GDS
Logical Memory

0. RBMT 8: Logical Memory .370
Adding RBMT to Logical Mem.
Logical Memory

3. Third step entry

9. All Variables .519
Geriatric Depression Scale
RBMT sum

Logical Memory

Adding Logical Mem. t0 G08 8: RBMT (2a)
Adding RBMT to GDS 81 Logical Mem. (20)
Adding GDS t0 RBMT & Logical Mom. (20)

52

.195'“
.127‘“
.091 “

.261 “‘

.24 “*

.137““

.270“"

A82 91
.066"
.134“
.0471
451*“
.0511
.015
0.37
0.2031
-0.112
.009
.0271
.128‘“

1 8 (Beta) is for all variables entered simultaneously and Is without regard to previous order

of entry.
" = significant at .05 level
*1 = significant at .01 level
i“ = significant at .001 level

Table 5

 

'r ' ltil R r inAnl Pr
- r n f rrn Ttl w

T ml n=121

1. Firststepentry B

a. Geriatric Depression Scale .337
b. RBMT sum of std. scores .279
c. Logical Memory .168

2. Second step entry

a. GDS 8 RBMT .393
Adding RBMT to GDS
Adding GDS t0 RBMT

b. GDS 8 Logical Memory .352
Adding Logical Mem. t0 GDS
Adding GDS to Logical Mem.

0. RBMT 8 Logical Memory .279
Adding RBMT to Logical Mem.
Adding Logical Mem. to RBMT

3. Thid step entry

All Variables .393
Geriatric Depression Scale
RBMT sum
Logical Memory

Adding Logical Mem. t0 GDS 8 RBMT (2a)
Adding RBMT t0 GDS 8 Logical Mem. (2b)
Adding GDS t0 RBMT & Logical Mem. (2c)

.1 13'“
.078"
.028

.1541“

.124‘“

.078“

.155'“

.041 "
.077“

.011
.095’“

.050'

-0.286“'
0.216
-0.012

.000
.031'
.07 "

1- 0 (Beta) is for all variables entered simultaneously and is without regard to previous order

of entry.
1 = significant at .05 level
.. .-. significant at .01 level
‘“ = significant at .001 level

DISCUSSION

The results of the present study failed to support the hypothesized
strength of the relationships between subjective memory complaints, objective
performance, and depressive symptomatology. However, the results do support
the hypothesized superior predictive value of memory complaints as they relate
to everyday memory measures when compared to more traditional laboratory-
based memory measure. The results did support significant, but weak,
relationships between the objective memory measures and memory complaint.
Finally, the predicted relationship between gender and memory complaints was
also unsupported. Each result will be discussed in turn followed by a general
discussion and recommendations.

The hypothesis that, given a more targeted method of measuring
depressed mood in older adults, the relationship between memory complaints
and depressed mood would be weaker than when compared to memory
measures, was not supported. Similar to the findings of Kahn, Zarit, Hilbert, and
Niederehe (1975) and Niederehe and Yoder (1989) the depression measure
used in this study did indeed account for more variance in both of the memory
complaint composite totals (Ability total and Frequency of Occurrence total) than
either of the objective memory measures. Interestingly, the correlation between
the Geriatric Depression Scale (GDS) scores and the two composite memory
complaint totals was lower than the reported correlations between depression
and complaint measures in previous studies (e.g., Kahn, Zarit, Hilbert, &
Niederehe, 1975; O'Boyle, Amadeo, & Self, 1990). The use of a depression

34

35

measure without a preponderance of vegetative/somatic symptom ratings, with
their suspected exaggerated ratings of depression in older adults, may have
contributed to the lower correlations between subjective complaint and
depression. It would be helpful for a future study, to include both types of
measures of depression in an analysis, to investigate this possible relationship.

The Rivermead Behavioural Memory Test (RBMT) did not account for
more variance than did the GDS, although as predicted, this everyday memory
measure did account for more variance than did the traditional laboratory
memory test, Logical Memory l. The Logical Memory task is an immediate recall
of two prose passages, whereas the RBMT is a series of memory tasks
approximating what a person would likely encounter during an average day. This
does support the notion of this author, and others (Cavanaugh & Poon, 1989;
Zelinski, Gilewski, 8 Thompson, 1980), that increased isomorphism between
complaint and memory measures would increase the correlation between them.

The more laboratory-based memory measure, Logical Memory l, was
found to account for the least amount of variance in the MAC-S composite
scores. The relative contribution of this measure to the variance was very small
and was not significant with respect to either the MAC-S Ability and Frequency of
Occurrence totals. This, again, is in accord with the prior investigations that
found only a weak or no relationship between complaint and objective
performance.

These analyses revealed no significant differences between males and
females on any of the measures. In fact, the results suggest that mean scores
on these measures were strikingly comparable as a function of gender. One
significant difference along gender lines was, however, revealed. Males (M =
15.7) were found to have a significantly higher level of education than women (M

= 14.3) in this sample (p < .01). Due to this significant education difference,

36

analyses were conducted holding this variable constant. Separate analysis of
covariance (ANCOVA) analyses were conducted, holding education constant.
These ANCOVAs were done for all variables and in no case did the gender effect
become significant.

General Digpfiion

The major finding of this study revealed the stronger association between
memory complaints with depression than with objective memory performance.
There are two general perspectives that may be helpful in understanding this
finding. The first suggests that depressed people exaggerate their symptoms
and negative complaints. The second perspective, the depressive realism
hypothesis, suggests that depressed individuals are, in fact, more accurate and
realistic in their appraisals.

This above finding may imply that there is some shared factor that both
memory complaints and depression measures assess. It seems reasonable to
assume that the nature of depression and depressed mood in older adults may
also contribute to the increased likelihood of reporting memory problems. Beck
(1971) suggested that exaggeration of symptoms, complaints, or negative
perceptions may reflect an underlying personality structure (i.e., depression).
Similarly, Kahn, Zarit, Hilbert, and Niederehe (1989) posited that "exaggerated
memory complaints is considered a manifestation of a general pattern of
discrepant reporting of symptoms by depressed persons, and may be related to
an underlying personality factor..." (p. 1573). Additionally, Niederehe and Yoder
(1989) stated complaints about memory may be potentiated if the individual
harbors negative perceptions about memory and age changes. Therefore,
though the levels of depression as measured by the GDS did not reflect, on
average, evidence of clinical levels of depression, the increased endorsement of

depressive symptomatology and dysphoric mood also has a positive association

37

with complaints and exaggerations of symptoms in other domains, specifically
regarding memory. The effect is a statement of this sort, "I am feeling very sad,
hopeless, guilty, and am experiencing interferences in my current life, including
worrisome memory problems".

However, there is also literature that suggests that people tend to exhibit
unrealistic optimism about their own ability (Dunning 8 Story, 1991) and tha "...at
times depressed people are 'sadder but wiser' than nondepressed people" (Alloy
& Abramson, 1979, p. 479-480). This fact "runs counter to the traditional
cognitive conception of depression whereby depressed persons are believed to
be overly negative in their perceptions" (Dobson 8 Franche, 1989, p. 419). Given
that it is generally found that memory performance declines significantly with age
(e.g., Herzog & Rodgers, 1989; Williams, Denney, and Schadler, 1983), perhaps
those older adults whose endorsements suggest depressive mood are accurately
perceiving this decline. This perspective would account for the higher levels of
memory complaints in this sub-population of older adults.

Memory measure scores for older individuals are compared to normative
data, which reflect this general decline in memory function on average. Perhaps
it is the case that, given the general decline in memory functioning with
increasing age, those older adults exhibiting more depressive symptomatology
are accurately perceiving this decline and reporting it via the memory complaint
questionnaires. The non-depressed older adults may be over-estimating their
current memory functioning or painting a rosier picture than actually exisits,
thereby not reporting increased memory complaints on these measures. In order
to make this relationship clearer, premorbid data would need to be obtained to
see if an individual's memory functioning had declined from previous levels.
Then, the question of whether memory complaints are accurate would be more
reliably validated.

38

Given the above findings, it seems that assessments of older adults with
memory complaints should also Include measures of depressive
symptomatology. Diagnoses of memory dysfunction, altered brain function, and
other organic processes should be considered tentatively and carefully,
especially before knowing if depression and its symptomatology are present for
an older individual.

The populations used in several of the previous studies reviewed were
mostly psychiatric inpatients 0r outpatients. There was question whether the
results of thses studies could be fairly generalized to the substantial majority of
older adults who are living independently in the community without symptoms
debilitating enough to come to the attention of health care professionals. This
study utilized a group of community-dwelling older adult volunteers in order to
see if the relationships found with previous samples could be applied to the
majority of older adults. The use of this population, with milder levels of
depressive symptomatology than those in psychiatric samples, mirrored the
findings with those populations. This finding lends support to the notion that
depression is not qualitatively different than lower levels of dysphoric mood.
Gradations in quantity separate individuals with dysphoric mood, rather than
discrete entities of psychopathology. Given that the results did indicate the same
general trends as the previous research, this study should add more confidence
to practitioners' diagnostic and treatment considerations with healthy "able" older

adults.

LIST OF REFERENCES

LIST OF REFERENCES

Abikoff, H., Alvir, J., Hong, G., Sukoff, R., Orazio, J., Solomon, S., &
Saravay, S. (1987). Logical Memory subtest of the Wechsler Memory Scale:
Age and education norms and alternate-form reliability of two scoring systems.

ourn l f lini l Ex eriment l Nepropsyphology, 9, 435-448.

Alloy, L. B., 8 Abramson, L. Y. (1979). Judgement of contingency in
depressed and nondepressed students: Sadder but wiser? Journal of
Experimental Psychology, 108, 441-485.

Armitage, S. G. (1946). An analysis of certain psychological tests used
for the evaluation of brain injury. P oh I lo M no r hs 60 1-48.

Atchley, R. C. (1976). Selected social and psychological differences
between men and women in later life. Journal pf Gerontolpgy, 31, 204-211.

Bak, J. S., & Greene, R. L. (1981). A review of the performance of aged
adults on various Wechsler Memory Scale subtests. Jppmgl of Clinical

Psychology, 31, 186-188.

Beck, A. T. (1971). D ressi n: lini l ex erim ntal nd theoretical
asgpts. New York: Paul B. Hoeber, Inc.

Beck, A. T., Steer, R. A., & Garbin, M. G. (1988). Psychometric
properties of the Beck Depression Inventory: Twenty-five years of evaluation.

Clinical Psyphology Review, 8, 77-100.

Beck, A. T., Ward, C. H., Mendelson, M., Mock, J., & Erbaugh, J. (1961).
An inventory for measuring depression. Archivg of General Psychiatry, 4, 561-571.

Bolla, J., Huff, J., Nebes, R., Holland, A., 8: Boiler, F. (1988).
Neuropsychological function in Alzheimer's disease. Archives of Neurolpgy, 45,
263-268.

Carpiniello, B., Carta, M. G., & Rudas, N. (1989). Depression among
elderly people: A psychosocial study of urban and rural populations. Acta

Psychiatrika Scandinavika, 80, 445-450.

Cavanaugh, J. C., & Perlmutter, M. (1982). Metamemory: A critical
examination. Child Devglppmgnt, x, 11-28.

39

4o

Cavanaugh, J. C., 8 Murphy, N. Z. (1986). Personality and metamemory
correlates of memory performance in younger and older adults. Edupational

Garontolpgy, 12, 385-394.

Cavanaugh, J. C. (1987). Age differences in adults' self-reports of
memory ability: It depends on how and what you ask. lntemational Journal

of Aging Ea Hpman Davelppmmt, 24, 271-277.

Cavanaugh, J. C., 8 Poon, L. W. (1989). Metamemorial predictors of
memory performance in young and older adults. Psvcholami and Aging, 4,
365-368.

Ceralla, J., Poon, L. W., 8 Fozard, J. L. (1982). Age and iconic read-out.
qurnal pf Gerontolpgy, 37, 197—202.

Chandler, J. D., 8 Gerndt, J. (1988). Memory complaints and memory
deficits in young and old psychiatric inpatients. Journal of Geriatric Paychiatm

and Neurplpgy, 1, 84-88.

Crook, T. H., 8 Larrabee, G. J. (1990). A self-rating scale for evaluating
memory in everyday life. Psyphplpgy Ed Aging, a, 48-57.

Crook, T. H., 8 Larrabee, G. J. (1992). Normative data on a self-rating
scale for evaluating memory in everyday life. Arphiveé 0f Clinipa

Neprppsmholpgy, 7, 41-51.

Dobson, K. 8 Franche, R. (1989). A conceptual and empirical review of

the depressive realism hypothesis. Canadian Journal pf Behavioral Spience, 21,
419-433.

Dunning, D. &Story, A. L. (1991). Depression, realism, and the
overconfidence effect: Are the sadder wiser when predicting future actions and
events? Journal of P rsonali rid cial P cholo 61 521-532.

Diagnostic and Statistical Manpal pf Marital Digrders, ed 3. (1980).

Washington DC: American Psychiatric Association.

Erber, J. T., Szuchman, L. T., 8 Rothberg, S. T. (1990). Age, gender, and
individual differences in memory failure appraisal. Psychology and Aging, 5, 600-603.

Fuhrer, R., Antonucci, T. C., Gagnon, M., Dartigues, J. F., Barberger-
Gateau, 8 Alperovitch, A. (1992). Depressive symptomatology and cognitive
functioning: An epidemiological survery in an elderly community sample in

France. P§ychglggical Mggigina, 22, 159-172.

Grut, M., Jorm, A. F., Fratiglioni, L., Forsell, Y, Viitanen, M., 8 Winbald, B.
(1993). Memory complaints of elderly people in a population survey: Variation
according to dementia stage and depression. Journal of the American Geriatrics
Society, 41, 1295-1300.

41

Hale, W. D., 8 Cochran, C. D. (1983). Sex differences in patterns of self-

reported psychopathology in the married elderly. qumal of Clinical Psychology,
& 647-650.

Hamilton, M. (1967). Development of a rating scale for primary
depressive illness. British Jpprnal pf Saga ang Clinpial Payphology, 6. 278-296.

Harper, R. G., Kotik-Harper, D., 8 lfirby, H. (1990). Psychometric
assessment of depression in an ederly general medical population. Journal of
Narvppa aria Mental Dififl, 17a, 113-119.

Herrmann, D. J. (1982). Know thy memory: The use of questionnaires to
assess and study memory. Psycholpgical Bulletin. 92. 434-452.

Herzog, A. R., 8 Rodgers, W. L. (1989). Age differences in memory
performance and memory ratings as measured in a sample survey. Psychology

and Aging, 4, 173-182.

Hultsch, D. F., Masson, M. E. J., 8 Small, B. J. (1991). Adult age differences
in direct and indirect tests of memory. Journal at Gerontplpgy, 46, 22-30.

Jenike, M. A. (1988). Assessment and treatment of affective illness in the

elderly. Journal pf Geriatric Psychiatgy ang Naurplogy, 1, 89-107.

Karanci, N. A. (1988). Patterns of depression in medical patients and
their relationship with causal attributions for illness. Psychotherapy and

Paychospmatics, 50, 207-215.

Kahn, R. L., Zarit, S. H., Hilbert, N. M., 8 Niederehe, G. (1975). Memory
complaint and impairment in the aged. Archivaa of General Payphiatm, 32, 1569-1573.

Koening, H. G., Meador, K. G., Cohen, H. J., 8 Blazer, D. G. (1992).
Screening for depression in hospitalized elderly medical patients: Taking a

closer look. Journal at tha Amaripan Cariatrips ficiety, Q, 1013-1017.

Lachman, M. E., Steinberg, E. S., 8 Trotter, S. D. (1987). Effect of control
beliefs and attributions on memory sail-assessments and performance.

Psycholpgy and Aging, 2, 266-271.

Lamberty, G. J., 8 Bieliauskas, L. A. (1993). Distinguishing between
depression and dementia in the elderly: A review of neuropsychological findings.

Arphivas pf Clinical Naprppsypholpgy, a, 149-170.

Larrabee, G. J., 8 Crook, T. H. (1993). Do men show more rapid age-
associated decline in simulated everyday verbal memory than do women?

Psychplogy and Aging, 8. 68-71.

42

Larrabee, G. J., West, R. L., 8 Crook, T. H. (1991). The association of
memory complaint with computer-simulated everyday memory performance.

qurnal pf Clinpial aria Exparimental Naurppsychplogy, 13, 466-478.

LaRue, A. L. (1992). A in n n r ch l i ent. New
York: Plenum Press.

Lewinsohn, P. M. Rohde, P., Seeley, J. R., 8 Fischer, S. A. (1991). Age
and depression: Unique and shared effects. Psychplpgy and Aging, 6. 247-260.

Lichtenberg, P. A., 8 Christensen, B. (1992). Extended normative data
for the Logical Memory subtests of the Wechsler Memory Scale-Revised:
Responses from a sample of cognitively intact elderly medical patients.

Pgholggical Remus, 71, 745-746.

Linn, M. W., Hunter, K., 8 Harris, R. (1980). Symptoms of depression
and recent life events in the community elderly. qurnal of Clinical Pschology,
16‘ 675-682.

Murphy, M. D., Sanders, R. E., Gabriesheski, A. S., 8 Schmitt, F. A.
(1981). Metamemory in the aged. Jpprnal pf Carpntology, 36, 185-193.

Niederehe, G. (1976). Self-appraisal of memory in depressions of later
life. Disgrtation Abstracts lntgnatipnal, a7, 1921 B.

Niederehe, G., 8 Yoder, C. (1989). Metamemory perceptions in
depressions of young and older adults. Jopmal of Nervous and Mentfl Disease,
177, 4-14.

Norris, J. T., Gallagher, D., Wilson, A., 8 Vlfinograd, C. H. (1987).
Assessment of depression in geriatric medical outpatients: The validity of two

screening measures. Jpprnal pf the Amarican Ceriatrics Sociey, 35, 989-995.

O'Boyle, M., Amadeo, M., 8 Self, D. (1990). Cognitive complaints in elderly
depressed and pseudodemented patients. Psyphplogy aria Aging, 5, 467-468.

O'Conner, D. w., Pollitt, P. A., Roth, M., Brook, P. B., 8 Reiss, B. B.
(1990). Memory compaints and impairment in normal, depressed, and demented
elderly persons identified in a community survey. Archivas pf Cenaral
Pachiatm, 47, 224-227.

Parmelee, P. A., Lawton, M. P., 8 Katz, l. R. (1989). Psychometric
properties of the Geriatric Depression Scale among the institutionalized aged.

Paycholpgical Asggment, 1, 331-338.

Petersen, R. C., Smith, G., Kokmen, E., lvnik, R. J., 8 Talngalos, E. G.
(1992). Memory function in normal aging. Naprolpgy, 42, 396-401.

43

Plotkin, D. A., Mintz, J., 8 Jarvik, L. F. (1985). Subjective memory complaints in
geriatric depression. Amaricap Jppmg pf Psyphiatm, 142, 1103-1105.

Popkin, S. J., Gallagher, 0, Thompson, L. W., 8 Moore, M. (1982).
Memory complaint and performance in normal and depressed older adults.

Exmrimantal Aging Rasearph, a, 141-145.

Salamero, M. 8 Marcos, T. (1992). Factor study of the Geriatric
Depression Scale. Aga Psyphiatrica Scandinavica, 86, 283-286.

Shaw, R. J., 8 Craik, F. l. (1989). Age differences in predictions and
peformance on a cued recall task. Paypholpgy ang Aging, 4, 131 -135.

Stewart, R. B., Blachfield, R., Hale, W. E., Moore, M. T., May, F. E., 8 Marks,
R. G. (1991). Correlates of Beck Depression Inventory scores in an ambulatory
elderly population: Symptoms, diseases, laboratory values, and medications.
The qurnal ofjamilv Practice 32 497-502.

Sunderland, A., Watts, K., Baddeley, A. D., 8 Harris, J. E. (1986).
Subjective memory assessment and test performance in elderly adults. Journal

of Carontplogy, 41, 376-384.

Taylor, J. L., Miller, T. P., 8 Tinklenberg, J. R. (1992). Correlates of
memory decline: A 4-year longitudinal study of older adults with memory

complaints. Psycholpgy aria Aging, 7, 185—193.

Teri, L. 8 Wagner, A. (1992). Alzheimer's disease and depression.

Joprnal 0f @nsplting and Clinical Psychplogy, 60, 379-391.

Thurston, L. L., 8 Thurston, T. G. (1949). Examinar Manual for the SRA
Primfly Mental Abilities Tast. Chicago: Science Research Associates.

 

 

Trahan, D. E., 8 Quintana, J. W. (1990). Analysis of gender effects upon
verbal and visual memory performance in adults. Archives of Clinical

WEE. 325-334-

Wechlser, D. (1945). A stadardized memory scale for clinical use.

Jggrnfl Cf thglggy, 19, 87-95.

Wiederholt, W. C., Cahn, D., Butters, N. M., Salmon, D. P., Kritz-
Siverstein, D., 8 Barrett-Conner, E. (1993). Effects of age, gender, and
education on selected neuropsychological tests in an elderly community cohort.

Jpprnal pf tha Ameripan Cariatrics Socim, 41, 639-647.

Williams, S. A., Denney, N. W., 8Schadler, M. (1983). Elderly adults'
perception of their own cognitive development during the adult years.
Int rn tion urn lofA in Human Develo ment 16 147-158.

44

Williams, J., Little, M. M., Scates, S., 8 Blockman, N. (1987). Memory
complaints and abilities among depressed older adults. Joprnal pf Qnsulting

W $95-$98.

Wilson, 8, Cockbum, J., 8 Baddeley, A. (1985). Tha Rivarmaad
Behavioural Memog Test. Bury St. Edmunds: Thames Valley Test Company.

Wilson, 8, Cockburn,J., Baddeley,A., 8 Hiorns, R. (1989). The
development and validation of a test battery for detecting and monitoring
everyday memory problems. m I f lini l nd Ex rim nt l

Neurppgrchplpgy, 11, 855-870.

Yesavage, J. A., Brink, T. L., Rose, T. L., Lum, 0., Huang, V., Adey, M., 8
Leirer, O. (1983). Development and validation of a geriatric depression

screening scale: A preliminary report. Jppmal of Psychiatrip Ragargh, 17, 37-49.

Zelinski, E.M., Gilewski, M. J., 8 Thompson, L. W. (1980). Do Laboratory
tests relate to self-assessment of memory ability in the young and old?
In L. W. Poon, J. L. Fozard, L. S. Cermak, D. A. Arenberg, 8 L. W. Thompson (Eds),

New girections in mempgy ang aging. Hillsdale, New Jersey: Lawrence Erlbaum
Associates.

Zelinski, E. M., Gilewski, M. J., 8 Schaie, K. W. (1993). Individual
differences in cross-sectional and 3-year longitudinal performance across the

adult life span. Paychology and Aging, a, 176-186.

Zgourides, G., Spofford, M., 8 Doppelt, L. (1989). The Geriatric
Depression Scale: Discriminant validity and elderly day-treatment clients.

Paycholpgical Remrts, Q, 1082.

I'IICH G

littltillltllti