University of Southern California Dissertations and Theses

Mood Induction As An Analog To Depression In Older Adults: Aspects Of Depression And Cognitive Effects

(USC Thesis Other)

Mood Induction As An Analog To Depression In Older Adults: Aspects Of Depression And Cognitive Effects

PDF

Download a page range

Download transcript

Copy asset link

Request this asset

Transcript (if available)

Content INFORMATION TO USERS
This manuscript has been reproduced from the microfilm master. UM I
films the text directly from the original or copy submitted. Thus, some
thesis and dissertation copies are in typewriter face, while others may be
from any type of computer printer.
The quality of this reproduction is dependent upon the quality of th e
copy subm itted. Broken or indistinct print, colored or poor quality
illustrations and photographs, print bleedthrough, substandard margins,
and improper alignment can adversely affect reproduction.
In the unlikely event that the author did not send UMI a complete
manuscript and there are missing pages, these will be noted. Also, if
unauthorized copyright material had to be removed, a note will indicate
the deletion.
Oversize materials (e.g., maps, drawings, charts) are reproduced by
sectioning the original, beginning at the upper left-hand comer and
continuing from left to right in equal sections with smalt overlaps. Each
original is also photographed in one exposure and is included in reduced
form at the back of the book.
Photographs included in the original manuscript have been reproduced
xerographically in this copy. Higher quality 6*' x 9” black and white
photographic prints are available for any photographs or illustrations
appearing in this copy for an additional charge. Contact UMI directly to
order.
UMI
A Bell & Howell Information Company
300 North Zeeb Road, Ann Arbor MI 48106-1346 USA
313/761-4700 800/521-0600

MOOD INDUCTION AS AN ANALOG TO
DEPRESSION IN OLDER ADULTS:
ASPECTS OF DEPRESSION AND COGNITIVE EFFECTS
by
Lauren Stephanie Fox
A Thesis Presented to the
FACULTY OF THE GRADUATE SCHOOL
UNIVERSITY OF SOUTHERN CALIFORNIA
In Partial Fulfillment of the
Requirements for the Degree
MASTER OF ARTS
(Psychology)
December 1995
Copyright 1995 Lauren Stephanie Fox
UMI Number: 1379582
UMI Microform 1379582
Copyright 1996, by UMI Company. All rights reserved.
This microform edition is protected against unauthorized
copying under Title 17, United States Code.
UMI
300 North Zeeb Road
Ann Arbor, MI 48103
UNIVERSITY O F S O U T H E R N CALIFORNIA
THE GRADUATE SCHOO L
UNIVERSITY PARK
LOS ANGELES, CALIFO RNIA II0 0 0 7
This thesis, •written by
under the direction of h S C .....T hesis Committee,
and approved by all its members, has been pre~
seated to and accepted by the Dean of The
Graduate School, in partial fulfillm ent of the
requirements fo r the degree of
oP Pcffc iia
Dian
THESIS COMMITTEE
■ r | t _ a | y * I
ACKNOWLEDGEMENTS
The author acknowledges the assistance o f Laura Filip in data management and
Travis Fogel, Brian Kaskie, Tara Rose, Michelle Rice, and Deanna Barton in data
collection. I would also like to thank Liz Zelinski for her helpful input. I owe a
special note o f gratitude to my advisor Bob Knight, whose generous guidance and
advice I am very thankful for.
TABLE OF CONTENTS
Acknowledgements ii
List of Tables/List o f Figures iv
Abstract v
Background 1
Depression and Depressed Mood: Research and Theory 1
Theories o f Depression 4
Mood Induction 5
Cognitive Effects of Depression 9
Theories of Cognition in Depression 12
Summary and Hypotheses 15
Methods 19
Subjects 19
Measures 21
Procedure 25
Results 28
Subject Characteristics 28
Efficacy of Induction 30
CES-D Subscales 35
Cognitive Performance 40
Discussion 45
Subjects 45
The Induction and Its Effects 46
Cognitive Performance 51
Limitations of the Study 53
Implications and Recommendations 56
Conclusions 58
References 59
iv
LIST OF TABLES/ LIST OF FIGURES
Table 1: Subject Characteristics 21
Figure 1: Naturally Depressed Subjects’ Subgroup
Means on CES-Ds 29
Figure 2: CES-D Means Pre- and Post-Induction
1st Session 31
Figure 3: CES-D Means Pre- and Post-Induction
2nd Session 31
Figure 4: DACL Means Pre- and Post-Induction
1st Session 33
Figure 5: DACL Means Pre- and Post-Induction
2nd Session 33
Table 2: CES-D Subscale Scores & Oneway ANOVA Results
1st Session 36
Table 3: CES-D Subscale Scores & Oneway ANOVA Results
2nd Session 36
Figure 6: Mean CES-D Subscale Scores, 1st Session 37
Figure 7: Mean CES-D Subscale Scores, 2nd Session 37
Table 4: Correlations Between CES-D Subscales 40
Figure 8: Naturally Depressed Subjects:
Subgroups’ CES-D Subscale Means, 1st Session 41
Figure 9: Naturally Depressed Subjects:
Subgroups’ CES-D Subscale Means, 2nd Session 41
Table 5: Performance on Cognitive Measures 43
Appendix A: CES-D Subscale Items 66
ABSTRACT
To validate the use o f a mood induction procedure with older adults, 20 mood
induced depressed elderly subjects were compared with 17 depressed and 22 non
depressed elderly subjects. The depressed mood induction procedure, using self
referent statements and music, produced a state that was comparable to the depressed
group in the first session, but was weaker than the depressed group in the second.
Post-induction scores on a mood measure were compared on four subscales: mood,
interpersonal isolation, psychomotor retardation, and lack of well-being. The induced
group’s mean subscale scores were generally between the means of the depressed and
non-depressed groups. Mood and psychomotor retardation were most influenced by
the induction. Groups were compared on 10 cognitive tasks but no pattern of
cognitive deficit was significant. The mood induction procedure’s utility with older
adults was demonstrated, but the relationship between mood and cognitive functioning
in the elderly was not supported.
1
BACKGROUND
Depression and Depressed Mood: Research and Theory
Clinical depression is a term for the spectrum of disorders that includes major
depressive disorder, bipolar disorder, dysthymia, adjustment disorder with depressed
mood, and depressive disorder not otherwise specified. The commonalities of clinical
depression include a number of symptoms such as sad mood, psychomotor retardation,
sleep and eating disturbances, loss of pleasure from activities, feelings of guilt, and
interpersonal isolation. A widely used self-report measure of depression is the 20 item
Center for Epidemiological Studies Depression Scale (CES-D), used both in clinical
and research settings. It has been factor analyzed and consistently shows four main
subscales that comprise the experience o f depression: depressed mood, psychomotor
retardation, lack of well-being, and interpersonal isolation (Clark, et al., 1981; Hertzog
et. al., 1990; Liang, Tran, Krause, & Markides, 1989; Mullen, Orbuch, Featherman, &
Nessleroade, 1988; Silberg et al., 1989; Gatz & Hurwicz, 1990; Radloff & Ten, 1986).
Depressive episodes occur when a specified number o f symptoms are present for
more than two weeks; when episodes occur without manic or hypomanic phases, the
person is diagnosed as having major depressive disorder. Bipolar disorder is diagnosed
when people have depressive episodes interspersed with manic or hypomanic episodes.
Research that is interested in clinical depression uses strict criteria, either the DSM IV
or the Research Diagnostic Criteria (RDC), to select eligible subjects.
2
Many people experience a number of symptoms o f depression but do not fit
enough criteria to be diagnosed with major depressive disorder. These people may
have another depressive spectrum disorder, such as dysthymia. Those who do not
meet criteria for a depressive spectrum diagnosis but have some symptoms of
depression are termed "subclinically" depressed. Researchers often identify
subclinically depressed subjects by giving a measure o f depression such as the CES-D
or the Beck Depression Inventory (BDI) to a group o f people (usually undergraduates)
and selecting subjects from those who score above a cutoff score suggesting
depression. The cutoff score most often used with the CES-D is 16. Research on the
suitability of this analog to depression has been done with the most common source of
subjects, undergraduates (Vredenburg, Flett & Krames, 1993; Hill, Kemp-Wheeler &
Jones, 1987). No research could be found that examines the suitability o f analogues to
depression in older samples. Undergraduates are a nonrandom population who are
quite distinct from the elderly. The existing research therefore cannot be generalized
to older adults without studies addressing validation.
Researching depression or its effects inherently involves some problems.
Selection biases may be a confounding factor in any differences found between
depressed and non-depressed samples. One example is that the rate of
co-morbidity of depression and disease or disability is very high, so differences seen in
psychomotor speed, for example, may be due to depression or due to the higher rate of
physical problems in the depressed sample. Another confound might be that women
tend to have higher rates o f depression than men, causing a random depressed sample
3
to have a higher female to male ratio than a random non-depressed sample. Another
potential issue is that the depressed mood itself is out o f the experimenter’s control.
The experimenter is not witness to or manipulator o f the external environment o f the
subjects that may influence mood considerably. Depression also decreases motivation
to participate in research.
Depressed people probably have a different experience in the world than non
depressed people have. This is an important issue that may bias research on depressed
and non-depressed groups. Depressed people have lower energy and disturbed
sleeping patterns; they may have lower activity levels than non-depressed people,
thereby reducing their opportunities for pleasurable events. Depressed people are not
fun to be around, so others may avoid them; their social interaction is decreased and
may be negative when available. It isn’t clear whether these things can make
depressed people even more depressed in a downward spiraling fashion. Somatic
slowness or discomfort, problems in interpersonal interactions, and a lack o f positive
life events and outlook could possibly cause any normally non-depressed person to feel
quite depressed. It isn’t clear if these signs cause depression or are effects o f it or
both. Without controlling all o f these factors (an impossible task), we can’t be sure.
It is clear that these differences make research on depressed and non-depressed mood
full o f potential problems. When in the lab, the two groups have vastly different
experiences to reflect on and associate with current stimuli.
The same confounds that exist with younger depressed subjects are even more
prevalent in older adults with a sometimes life-long history o f depression. Clinically
4
depressed and non-depressed elderly samples may be even more different from one
another in terms o f experiences than younger samples would be. Older adults have
different patterns o f social interaction than younger adults, and the experience o f
emotional information is different in the elderly (Carstensen, 1992; Carstensen &
Turk-Charles, 1994; Labouvie-Vief et al., 1989). It is important to validate research
on younger adults with older adults before using the information and assuming it is
applicable.
Theories o f Depression
One theory o f depression uses the idea o f associations as a key component in
maintenance o f depressed mood. The semantic network theory was first advanced by
Bower (1981). Each emotion has a memory node, and around that node are the
emotion's "associated autonomic reactions, standard role and expressive behaviors...and
descriptions o f standard evocative situations....Activation o f an emotion node also
spreads activation throughout the memory structures to which it is connected, creating
subthreshold excitation at those event nodes" (Bower, 1981, p. 135).
A related idea is Beck’s (1967) schema theory, which assumes people in a
particular mood state have a schema with which they interpret stimuli. For example,
depressed people would have a prevailing depressive schema for organizing
information and directing retrieval. The difference between network and schema
theory lies in the fact that network theory assumes spreading activation (Ingram,
1984). Beck’s cognitive theory o f depression proposes that stressors activate a
prevailing schema which selectively encodes negative information, effectively
5
maintaining the depression. Beck argued that depressed people focus on task irrelevant
or mood-congruent thoughts, maintaining the depressed state (for review see Hartlage,
et al. 1993). Depression-related thoughts may intrude on other cognitive processes,
• disrupting attention, or the focus o f attention may be narrowed or shifted to
depression-related thoughts. This would explain the mood congruent memory
phenomenon, where depressed people show better memory for negatively valenced
stimuli (Salovey & Singer, 1985; Teasdale & Russell, 1981; Teasdale & Taylor, 1981;
Teasdale, Taylor & Fogarty, 1980). Both Beck’s and Bower’s theories do not provide
an adequate rationale for how depression is started, or how people spontaneously
remit; their theories can explain how depression is maintained or increased. Bower’s
theory especially has been used by cognitive researchers as a paradigm to organize and
explain memory functioning in depressed mood, and much research has focused on the
predicted effects o f the network theory (e.g., mood congruent memory). These
memory effects are seen by many researchers to be potential agents that maintain or
deepen depression itself (Singer & Salovey, 1988). Bower’s theory has been supported
by findings o f mood congruency in depressed subjects, but is more frequently studied
in mood induced depressed subjects.
Mood Induction
The limitations o f research on clinically depressed samples and the questions
about depressed mood and its effects on memory led to the development o f mood
induction procedures, so researchers could randomly assign non-depressed subjects to
either a neutral or mood induction condition. Research using induction procedures is
6
mainly interested in mood, not the full syndrome o f depression o f which mood is only
a part. The procedure allows careful controlled study o f the mood phenomena, its
intensity, and its effects.
Random assignment o f groups to conditions as well as control over the mood
induction itself eliminates many o f the confounds o f using clinically depressed
subjects. It also allows control over when the mood is induced, allowing new research
avenues such as the effects o f mood on encoding and retrieval. The desire for
controlled study o f the effects o f depression in the elderly makes mood induction a
good option for researchers. Mood induction has been shown to be capable o f causing
a variety o f cognitive effects in younger subjects (Matt, Vasquez, & Campbell, 1992).
Cognition and aging is an important area o f study, and mood induction would be a
good analog to use in the study o f cognition and emotion in older adults.
Surprisingly, no studies exist which used only older adults as subjects in mood
induction research.
Mood induction procedures include self-referent statements, sad music, guided
imagery, or hypnotic suggestion. The most widely used and strongest acting induction
is the Velten technique (1968), which consists o f self-referent affectively valenced
statements which subjects read and concentrate on for 15 seconds each. The Velten
procedure is effective for about 50% of people. The musical mood induction
procedure consists o f affectively valenced music which is played throughout an
experimental session. Musical mood induction is weaker but more reliable than the
Velten, acting on about 75% o f participants (Martin, 1990). Mood induction has its
7
own limitations; as with any analog, there may be differences between induced and
natural depression that limit the generalizability of results to naturally depressed
people.
Mood induction procedures have been criticized for eliciting demand
characteristics. Subjects are instructed to try to maintain the mood, and studies have
shown that even instructing subjects to behave as if they were in a depressed or happy
mood is enough to elicit a mood congruent memory effect (Perrig & Perrig, 1988).
However, enough evidence has been amassed to refute the claims that effects of mood
induction procedures are solely due to demand characteristics. Many effects of
depressed mood have been noted that subjects either cannot fake or would have
difficulty purposely producing (Martin, 1990). These range from decreased writing
speed in induced depressed subjects (Alloy, Abramson, & Viscusi, 1981; Hale &
Strickland, 1976; Natale, 1977a, 1977b, 1978; Natale & Hantas, 1982; Velten, 1968) to
differences in eye movement, electromyographic activity (EMG), heart rate, and skin
conductance (Natale & Gur, 1980; Teasdale & Bancroft, 1977; Haney & Euse, 1976).
Mood induction is a laboratory procedure that does not change the pre
experiment experiences of the subject. It changes how the subject thinks in the
laboratory, and does not interfere with the actual experiences of the subject in the
external world. Psychomotor retardation, interpersonal isolation, and lack o f well
being, all symptoms o f depression that are normally not problematic for a non-
depressed person, are not intentionally altered by the induction procedure. The mood
itself is the target o f the induction. Scales like the Depressive Adjective Checklist
8
(DACL), Multiple Affect Adjective Checklist (MAACL), or visual analog scales with
continuum measures o f depressed mood are used to operationalize mood change.
Researchers have not used measures typically associated with the assessment o f clinical
depression, such as the CES-D. As mentioned earlier, the CES-D measures four
aspects of depression. It seems that mood induction research has focused on the mood
aspect and not the others. However, it would be interesting to know if the induction
procedure has an effect on the other subscales as well. Does the induction cause
subjects to perceive the other areas to be problematic even if they are normally not?
Does mood induced depression look similar to clinical depression on all four o f the
aspects seen in the syndrome o f clinical depression? In a review o f mood induction,
Martin (1990, p.685) theorized how inductions work:
It may be that mood induction generally involves the
construction o f a vicious circle which involves a reciprocal
relationship between emotion and cognition similar to that in
naturally occurring emotional states. However, mood induction
procedures do not necessarily all have their primary focus at the
same point o f the vicious circle. Three different components
may be distinguished: cognitive, via thought processes; somatic,
via bodily sensations; and emotion, via an intention on the behalf
o f a subject to feel a specified emotion. The cognitive route
involves increased accessibility o f negative cognitions....The
somatic route involves changes in bodily state as a result of
physical manipulation or verbal instruction. The emotion route
may be a separate route from the cognitive and somatic routes
or, in practice, it may simply be a free choice on the part o f the
individual to use the cognitive and/or somatic routes.
Martin (1990) proposed that self-statement and musical mood induction
techniques focused on all three o f the components o f mood: cognitive, somatic,
and emotional. It is not clear how music does this, or what it means in terms
9
o f the resulting symptomatology reported. The Velten statements do relate to
various symptoms o f clinical depression. Many focus on depressed mood,
somatic symptoms, and interpersonal issues. The mood induction procedure
would have to change perceptions of somatic symptoms, interpersonal
situations, and sense of well-being as well as mood itself for these things to be
reported as problematic by an induced subject. This is the proposed action o f
mood induction procedures, but this proposition has not been explicitly tested.
Because induction procedures do not actually change subjects’ experiences
outside o f the laboratory, only their perceptions and cognitions in the
laboratory, there is a disparity between the life experiences of induced and
naturally depressed people. A procedure which influences all of these
dimensions o f the syndrome o f depression would make a suitable analog that
could be used for depression research. The hypothesis proposed here is that
inductions do not produce a similar profile to clinical depression. Specifically,
inductions do not significantly affect the other aspects besides mood, and this
may make induced mood different from naturally occurring depressed mood,
and therefore an unsuitable analog for the study o f clinical depression.
Cognitive Effects of Depression
The effects o f clinical depression on cognition have been well
documented. Deficits have been shown on a number o f psychometric tasks in
clinically depressed groups. A recent meta-analysis on a wide range o f
cognitive abilities done by our laboratory (Fox, Knight, & Zelinski,
10
unpublished data) found significant deficits in depressed samples on verbal
fluency, psychomotor speed, visuo-construction tasks, reasoning, full scale IQ,
list recall, story recall, visual memory, and attention. Abilities that did not
show significant depression effects were crystallized intelligence, simple motor
speed, word list recognition, visuo-spatial attention, and arithmetic. Research
on cognitive deficits in older adults is a popular area; the differentiation
between reversible and non-reversible cognitive deficit (e.g., due to depression
versus dementia) is a crucial research topic. Clinical researchers have done
studies looking at older depressed and non-depressed samples on many tasks.
The relationship between age and the cognitive effects o f depression is still
unclear. A meta-analysis by Burt, Zembar & Niederehe (1995) on recall and
recognition tasks found that younger samples had larger depression deficits,
while our meta-analysis found a larger mean effect size in older samples.
Mood induction research on induced depressed subjects has found
deficits in psychomotor speed and memory but these types o f studies are not
typical o f this area. The paradigm is mainly used to support or disconfirm
Bower’s hypotheses such as the mood congruency effect or the mood
dependence effect. The mood-congruent memory effect predicts a facilitation
in memory for words o f the same valence as the induced mood and, while not
a very robust effect, has been supported by many studies (e. g., Salovey &
Singer, 1985; Teasdale & Russell, 1981; Teasdale & Taylor, 1981; Teasdale,
Taylor & Fogarty, 1980). The mood dependence effect predicts that material
11
learned in one affective state will be best remembered in that affective state,
regardless of the valence o f the material. This effect has not been consistently
found. Even Bower has questioned his hypotheses after a lack of support for
mood dependence (Bower & Mayer, 1985).
A consequence of clinical depression is the cognitive deficit seen on
effortful tasks such as memory (e.g., Lyness, Eaton, & Schneider, 1994;
Cassens, Wolfe, & Zola, 1990; Richards & Ruff, 1989) or fluid abilities such
as the Block Design task from the Wechsler Adults Intelligence Scale Revised
(WAIS-R) (e.g., Sackheim, et al., 1992; Gray, et al., 1987; Austin, et al.,
1992). If this deficit is replicable in mood induced subjects, the similarity of
induced to naturally occurring depression may be supported. It would also
support the theory that depressed mood itself is responsible for the cognitive
deficit, not other variables associated with depression such as severity of the
illness, medication, or hospitalization. These variables have been shown to be
related to cognitive deficit, and some researchers have said that they can
account for most of the deficit (Cole & Zarit, 1984). Our meta-analysis (Fox,
Knight, & Zelinski, unpublished data) found that studies using inpatient
samples had a larger mean effect size o f cognitive deficit than studies using
outpatient samples, but the effect size for outpatients was still significantly
larger than zero (inpatient studies had a mean d o f .54 with a confidence
interval of .27 to .81, while outpatient studies had a mean d o f .18 with a
confidence interval of .08 to .28). Finding this magnitude of deficit in induced
12
subjects would further clarify the relationship between mood and cognitive
functioning by eliminating the confounds of using depressed subjects.
Theories of Cognition in Depression
There are several theories that attempt to explain how depressed mood
affects cognition. Mood induction research is a useful paradigm in the search
for support of one or more of these theories. Bower’s and Beck’s theories have
been described above. These two theories of cognition or memory are used to
explain the maintenance o f depression, although they can also explain memory
functioning itself irrespective of the role they play in the maintenance o f
depression. The mood congruency effect supports these theories, but deficits
in memory or fluid ability are not explained by Beck or Bower’s theories.
Resource allocation models were primarily described by Hasher &
Zacks (1979) and by Ellis & Ashbrook (1988, 1989). The resource allocation
model was primarily developed to explain cognitive functioning when deficits
were documented in clinically depressed subjects but not well understood.
Hasher & Zacks’ work on effortful versus automatic processes posited a
reduced cognitive capacity in depression that results in decreased available
faculties for cognitive tasks. The theory would predict that tasks that require
more cognitive capacity or effort would show greater deficit than the more
automatic tasks requiring less cognitive capacity. This is supported by the
results o f our meta-analysis; the more effortful tasks showed deficits while the
more automatic tasks did not. Ellis & Ashbrook’s resource allocation model
13
"considers the relation between the emotional state o f a person and the
encoding demands o f the cognitive task being performed in terms of the
capacity demands of both the emotional state and the cognitive task," and
assumes that "emotional states regulate the amount o f capacity that can be
allocated to some criterion task" (Ellis & Ashbrook, 1988, 1989). The more
cognitive capacity is used in maintaining or processing the emotion, the less is
left over for the cognitive task. This theory would predict that the severity of
the depression would correlate positively with the absolute size o f the cognitive
deficit.
Hartlage, Alloy, Vasquez, & Dykman (1993) in a comprehensive
review, posit that a combination o f the two major theories may be in operation,
which they call the capacity-reduced, negative-focus hypothesis. Studies o f
effortful processing of valenced material show that depressed people selectively
attend to and process negative information, and they remember depression-
congruent information better (Hartlage, et al. 1993). Depressed subjects have
shown the mood congruent memory effect while simultaneously exhibiting
reduced overall recall (Denny & Hunt, 1992). Both the network/schema theory
and the resource allocation model attempt to explain the cognitive changes that
occur in depressed people which result in impaired speed, retrieval, learning,
memory of semantic and spatial information, and spontaneous use of strategy.
It may be that future research avoids the idea o f competing theories and instead
14
works toward an integrative theory such as Hartlage et al.’s which
accommodates previous and future findings.
The two theories of cognition would predict different though not
mutually exclusive research findings. The resource allocation model would
predict deficits in recall, psychomotor speed, visuo-construction, and other
effortful abilities in people with depressed mood, while sparing more automatic
tasks such as attention or crystallized abilities that require fewer resources.
Alternatively, the network or schema theory would predict a facilitation effect
for mood congruent material (mood congruency); depressing stimuli would be
remembered better by people in a depressed mood than those in a non-
depressed mood since they fit depressed peoples1 prevailing schema and are
more easily activated. Both o f these effects have been found in research on
people in depressed moods. Mood induction research has tended to focus on
the mood congruency phenomenon, while research on clinical samples most
often documents deficits instead of facilitation for mood congruent material.
It is desirable to use mood induction with an elderly sample to
determine what tasks show deficits due to depressed mood, see if there are
differences compared to younger samples, and to learn more about depression
itself in older adults. Inducing any amount o f depressed mood in non-
depressed people supports the idea that depressive symptomatology and mood
lie on a continuum and that it is possible to influence mood. Theories about
emotional regulation argue that emotional processing changes with age;
15
Labouvie-Vief et al. (1989) say that older adults gain affective regulation with
age, and Carstensen & Turk-Charles (1994) demonstrate an increased emotional
focus or salience in old age. Older adults may react differently than younger
adults to mood induction procedures. The direction of the difference may
support one or the other of these emotion theories. Future researchers should
pursue a comparison of older and younger adults to test these ideas, but before
that can be done, the technique must be shown to produce comparable effects.
Validation of the mood induction procedure with older adults is a crucial step
to further use o f the paradigm in research that attempts to understand emotion
and cognition in depressed or non-depressed older adults.
Summary and Hypotheses
Mood induction is a useful analog to depression that has been used with
younger adults but thus far, not with older adults. The paradigm allows for
controlled study of the qualities o f depressed mood itself and the effects of
depressed mood on how people think. While there is research that examines
cognition in clinically depressed elderly and in younger induced subjects, no
literature could be found which examines older subjects in an induced
depressed mood. Successful induction of the elderly that resembles depression
would be valuable in the study o f depression and cognition in depression as
well as in the generalization o f existing findings with younger adults to the
older population. The first step to using mood induced older adults in studies
or applying existing research to the elderly is validating the procedure for use
16
in this population. The first aim of this study was to show that a mood
induction procedure designed to induce a depressed mood state works with an
elderly sample. The second aim was to compare the resulting mood state to
natural depression by comparing induced and naturally depressed subjects on
four aspects of depression: sad mood, psychomotor retardation, lack of well
being, and interpersonal isolation. The last aim of the study was to compare
the cognitive performance o f naturally depressed, induced depressed, and non-
depressed subjects on a variety of cognitive tasks tapping memory,
psychomotor and motor speed, and fluid ability. It was expected that the mood
induction acts on mood itself much more than the other aspects o f depression
since the induction does not change the subjects’ pre-experiment experiences.
If mood is most affected by inductions, and inductions show reliable cognitive
effects, then the mood itself is responsible for the cognitive effects. This
procedure is then a suitable analog when studying the cognitive effects of
depression even without subjects’ experiencing the other aspects o f depression.
Showing that depressed mood is related to cognitive effects allows mood
induction research to be validly applied to thinking about cognition in naturally
occurring depressed mood.
My hypotheses were as follows:
a) It was expected that the depressed mood induction subjects would
show significant change on total CES-D scores when post-induction scores are
compared to pre-induction scores. It was expected that the neutral and the
17
depressed samples would show no change on total CES-D scores. In other
words, it was expected that the depressed mood induction procedure would
work on the older sample, and that the neutral mood induction procedure would
not cause any change in the two groups receiving that procedure.
b) The overall CES-D scores were expected to be highest in the
depressed sample. It was expected that non-depressed comparison subjects
would have significantly lower total scores than the other two groups.
c) It was expected that the mood induced group’s sample mean for the
Depressed Mood subscale would be higher than the depressed group’s sample
mean, indicating a more depressed mood in the induced subjects than the
depressed subjects immediately post-induction. The focus o f the induction
procedures is directly on mood itself, and while not all subjects are expected to
respond strongly to the Velten, the musical induction procedure was expected
to influence the subjects who do not respond to the Velten procedure, as well
as maintain the effects o f the Velten. The depressed sample was not expected
to have consistently strong depressed mood at testing; their mood was expected
to show some variability due to natural fluctuation. It was expected that their
mood would therefore be less depressed on average than the induced group.
Both depressed and mood induced groups were expected to have higher mean
depressed mood scores than the non-depressed comparison group.
d) The Psychomotor Retardation, Interpersonal Isolation, and Lack of
Well-Being subscales were expected to be higher in depressed subjects than
18
mood induced subjects, indicating more somatic symptomatology, social
isolation and decreased life satisfaction in the depressed. This would be the
expected result o f a long term dysphoric experience and interacting in the social
world for an extended period o f time while depressed, an experience mood
induced subjects would not have. Both groups were expected to show higher
mean scores on these three subscales than non-depressed comparison subjects.
e) Since mood induction studies assume that mood is most related to
cognitive and memory deficits, it was expected that the neutral group would
show the best performance and the group with the highest sad mood subscale
score (predicted to be the induced depressed group) would show the lowest
performance on the following cognitive measures: 1) neutral sentence
memory, which would confirm previous findings o f semantic memory
impairment in depression and depressed mood; 2) motor/psychomotor speed,
evidenced by reduced cross off speed, visual search speed, and WAIS-R Digit
Symbol; 3) WAIS-R Block Design, a measure o f visuo-spatial ability; 4)
WMS-R Visual Reproduction I and II, a measure o f visuo-spatial memory; and
5) WMS-R Logical Memory I and II, a measure o f memory for organized
prose. 6) I expected greater memory for depressive sentences, which would
confirm the mood congruent memory effect.
f) A negative correlation was expected between depressed mood
subscale scores and raw scores on cognitive measures; greater depressed mood
was expected to be related to lower scores on cognitive tasks, regardless of
19
which experimental group the subject was in. This correlation was expected to
be stronger than correlations between cognitive measures and the other three
subscales on the CES-D.
METHODS
Subjects
In this analysis, 17 elderly naturally depressed subjects were compared
with 22 mood induced depressed elderly subjects, as well as a control group of
20 neutral induction elderly subjects. Non-depressed elderly subjects (for both
mood induction conditions) were recruited through the Andrus Gerontology
Center Volunteers and through a list o f U.S.C. alumni who have been active in
on-campus research in the past. Depressed subjects were recruited through the
Tingstad Older Adult Counseling Center (TOACC) (N=5) and the Caregiver
Resource Center (CRC) (N=5). Seven depressed subjects were recruited
through the same means as non-depressed subjects but scored above 16 on the
depression screening measure and so were reassigned to the depressed
condition. Depressed subjects were therefore a mixture o f outpatient
psychotherapy clients, help-seeking caregivers who were recruited due to
elevated CES-D scores at a previous measurement, and subjects who scored
above a cutoff at the time of testing.
Before beginning the procedures, subjects are screened for clinical
depression; they completed the CES-D and the Inventory to Diagnose
20
Depression (IDD) to assure that mood induction subjects are not significantly
depressed, and to assure that depressed subjects are still experiencing their
depression. Depressed subjects were included in that group only if they scored
17 or above on the CES-D; neutral and induced subjects were switched to the
depressed condition if they scored above 17. Four o f the 17 depressed subjects
met RDC criteria for probable depression (3 met DSM IV criteria for major
depressive disorder), as measured on the Inventory to Diagnose Depression
(IDD). The source o f these subjects was evenly distributed; one scored above
the CES-D cutoff, one was a client o f the Tingstad Older Adult Counseling
Center, and two were clients o f the Caregiver Resource Center. The remainder
o f the depressed group had symptoms of depression but did not meet criteria
for any diagnosable depressive disorder except possibly Depressive Disorder
Not Otherwise Specified (NOS, defined as a depressive spectrum disorder with
fewer or different criteria than those specified in the DSM IV). The
differential diagnosis between major depressive disorder and dysthymia was not
possible since the IDD does not assess the total duration of symptomatology.
All subjects were also screened for dementia using the Kahn/Folstein Mini
Mental Status Exam (MMSE); possibly demented subjects who scored below
23 were excluded (N=l). Subjects were told that they were free to terminate
the experiment at any time should they begin to experience discomfort. A
profile of each groups’ characteristics can be seen in Table 1.
21
Table 1
Subject Characteristics
Neutrally Induced
N=20
Depressed Induced
N=22
Clinically Depressed
N=17
Age Mean (sd)=74.15
(4.86)
Mean (sd)=73.77
(6.49)
Mean (sd)=69.53
(8.89)
Years o f
Educ
Mean (sd)=16.55
(2.01)
Mean (sd)=16.18
(2.72)
Mean (sd)= 15.38
(2.53)
Gender 50% Female
50% Male
73% Female
27% Male
88% Female
12% Male
Ethnicity 90% Caucasian
5% African-
American
5% Unknown
77% Caucasian
14% African-
American
9% Other/Unknown
71% Caucasian
18% African-American
6% Asian
6% Other/Unknown
IDD Score Mean (sd)=5,0 (5.67) Mean (sd)=5.67
(5.15)
Mean (sd)= 19,18
(10.95)
Past 2
weeks
CES-D
Score
Mean (sd)=5.84
(4-55)
Mean (sd)=4.23
(2.89)
Mean (sd)=24.0 (7.61)
Measures
Relevant dependent measures in this study are the following:
CES-D completed pre- and post-induction (session 1 and 2)
DACL completed pre- and post-induction (session 1 and 2)
Sentence Memory
Cross Offs
Visual Search (F’s and [’s)
WMS-R Logical Memory I
WMS-R Visual Memory I
WAIS-R Block Design
WAIS-R Digit Symbol
WMS-R Logical Memory II
WMS-R Visual Memory II
22
The CES-D is a valid and reliable measure o f depressive symptoms,
widely used in research with all ages of adults (Radloff, 1977; Radloff & Teri,
1986). Split halves correlation and coefficient alpha have been reported to be
.85 to .92 in a large epidemiological study (CMHA study; Radloff, 1977), and
internal consistency is high in age, sex, geographic, and racial/ethnic subgroups
(Radloff, 1977), The CES-D has both discriminant and construct validity
(Radloff & Teri, 1986). The version used in this analysis is slightly altered: all
verbs in past tense are changed to present tense, and subjects are instructed to
answer the scale "as (they) feel right now". The justification for this change is
that the CES-D is administered directly after the mood induction, so the mood
to be assessed is this immediate mood, not the mood before the induction. As
there are no changes to the content o f the items, it is believed that this change
does not compromise the reliability or validity o f the measure. The alpha
reliabilities o f the present-tense CES-Ds in this experiment ranged from .90 to
.93. The twenty CES-D items have been found to load fairly consistently onto
four factors now commonly used as subscales (Radloff & Teri, 1986; Clark, et.
al., 1981; Hertzog et. al., 1990; Liang, Tran, Krause, & Markides, 1989;
Mullen, Orbuch, Featherman, & Nessleroade, 1988; Silberg et. al., 1989; Gatz
& Hurwicz, 1990). The subscales are Depressed Mood, Psychomotor
Retardation, Lack o f Well-Being, and Interpersonal Isolation. The subscales’
alphas were reported in Gatz & Hurwicz (1990) to be .85, .75, .78, and .57
respectively; these figures were similar to those found by Krause (1986) and
23
Krause and Markides (1985). The items in each subscale and the wording o f
these items can be found in Appendix A.
The Depressive Adjective Checklist is a measure often used in mood
induction studies to check for induction o f the depressed mood. It consists o f
32 affectively valenced adjectives, both positive and negative. Subjects check
off words that describe their current state and scoring counts both negative
items checked and positive items left unchecked. There are four versions of
the DACL which are parallel and scored identically in order to reduce practice
or retest effects. These were counterbalanced in their distribution. Reliability
of the DACL was reported by Lubin (1981) to range from .80 to .93 depending
on the form used. Split half, alternate form, and internal consistency
reliabilities were comparable.
All Wechsler Adult Intelligence Scale-Revised (WAIS-R) and Wechsler
Memory Scale-Revised (WMS-R) subscales were scored according to the
manuals (Wechsler, 1981 & 1987). The WAIS-R Block Design task is a
measure o f fluid ability, and more specifically, o f visuo-construction. Alpha
reliability is reported to be .84 for ages 65-69 and .87 for ages 70-74
(Wechsler, 1981). Snow et al. (1989) report a retest reliability o f .84 in their
older sample (mean age = 67.1). WAIS-R Digit Symbol is a psychomotor
speed test. The WAIS-R manual estimates the reliability to be .82 for adults
over age 65, and Snow et al. (1989) found a retest reliability o f .91. The
Logical Memory I subscale o f the WMS-R assesses immediate verbal memory
24
for two organized prose passages, and has a reported reliability of .76 for ages
65-69 and .79 for ages 70-74. Logical Memory II assesses longer term
memory for the same passages. Its reliability is .78 and .85 for the same age
groups (Wechsler, 1987). Visual Reproduction I and II, measurements of
immediate and delayed memory for 4 line drawings, have relatively low
reliabilities. The reliabilities for Visual Reproduction I are .53 (age 65-69) and
.51 (age 70-74), and those for Visual Reproduction II are .51 (age 65-69) and
.41 (age 70-74).
The sentence memory task is designed to assess short term memory and
is scored according to propositionalizing the sentences for content rather than
exact wording. This task assesses the immediate memory of sentence content,
the "gist" o f the sentence rather than the actual word content, which will be
assessed by Logical Memory I. The sentence memory task yields four subset
scores: low-frequency neutral sentences, high-frequency neutral sentences, low-
frequency sad sentences, and high-frequency sad sentences. While both low-
and high-frequency sentences are grammatically correct, the low-frequency
sentences are organized with nested phrases, to sound unusual. Subjects
received one o f two versions o f the sentence memory task in counter-balanced
order; the low- and high-frequency organization was switched in the two forms.
The Cronbach’s alpha for the sentence memory task as a whole was .79 in this
study.
25
The Cross Offs task is a motor speed measure which is scored as the
number o f crosses made in 45 seconds. Subjects are presented with a piece of
white paper with horizontal dash marks and are told to make the dashes
into plusses (+) as fast as they can. The Visual Search is a timed letter
cancellation task designed to measure psychomotor speed and concentration.
The speed measure of interest in this study is the number o f seconds to circle
all o f the F’s and [’s on the 2 pages, and the concentration measure is the
number of F’s or [’s missed. F’s are mixed with ten other letters, and [*s are
mixed with ten other symbols on the pages. This task has been used by other
researchers (e.g., Feehan, Knight & Partridge, 1991), but no reliability or
validity statistics have been reported, as is common with unstandardized
laboratory tasks.
Procedure
Subjects were recruited over the telephone, and two 2 hour sessions
were scheduled. Most subjects were tested at U.S.C. in a sound-proofed room.
Five clinically depressed subjects recruited through the Caregiver Resource
Center were tested in their own homes, and all efforts were made to reduce
distractions and interruptions. All subjects were told about confidentiality and
their right to terminate at any time and then signed consent forms. Subjects
completed a number of premeasures, including a demographic information
questionnaire, the CES-D, IDD, and MMSE. Subjects then read laminated
cards with typed Velten statements to themselves and then out loud, at a rate
26
of one every 15 seconds. The depression induction includes statements such as,
"Sometimes I think my life is not worthwhile", or "I feel listless and tired; I
think I’ll take a nap when I get home". Along with the neutral control group,
all clinically depressed subjects read neutral Velten statements, to avoid
confounds due to the induction procedure itself. The neutral induction
procedure contained non-mood-related statements. Examples include, "There is
a large rose garden near Tyler, Texas," and "Black and white pictures are
arranged in ten sections."
Mood-congruent music was turned on immediately post-induction, and
played throughout the rest o f the session. Music served to remind subjects of
the mood they were experiencing, as well as to strengthen the induction by
providing a second technique. The music for the neutral condition was
harpsichord concertos by Bach. The induced depression subjects listened to
Prokofiev’s "Russia Under the Mongolian Yoke", Berber’s Adagio, Albinoni’s
Adagio, and Sibelius’ "The Swan of Tuonela". The piece by Prokofiev has
been used by mood induction researchers at Oxford as a sad mood induction
(Clark & Teasdale, 1985; Clark, Teasdale, Broadbent & Martin, 1983; Teasdale
& Spencer, 1984), and the other pieces were chosen by consensus o f our lab as
suitable sad or neutral music.
Due to the large number of dependent measures, subjects participated in
two 1.5 to 2 hour sessions, with identical mood induction procedures. After the
induction, subjects completed a CES-D (with all past tense verbs changed to
27
present tense) and a DACL to assess depressed mood, and were told to focus
on "how they feel right now". They also completed other mood checks
throughout the session (DACL’s). Subjects then complete a number o f memory
and cognitive tasks, some o f which were not included in the present analysis.
The order o f presentation did not vary; session one included sentence memory
and a task designed to elicit tip-of-the-tongue experiences by asking difficult
trivia questions. DACL’s were completed after the sentence memory task and
after the trivia question task. Session two began with a mood check (CES-D
and DACL), then the mood induction, another mood-check, then Digit Span,
Logical Memory I, Visual Memory I, Block Design, Digit Symbol, Visual
Memory Span, Logical Memory II, Visual Memory II, Cross Offs, Visual
Search, and the Stroop task. The DACL was administered as a mood check
after Block Design, after Visual Search, and after the Stroop task. At the end
o f each session, induced depression subjects listened to a happy debriefing tape
with marching band music, and subjects were given another present tense CES-
D to be sure the depressed mood has lifted.
Subject Characteristics
RESULTS
28
A description o f subject characteristics and their means on premeasures
can be found in Table 1. The three groups did not significantly differ by age,
years o f education, or ethnicity, but they did differ by gender (the neutral group
has significantly more men than the depressed group, p<.05). As expected, the
IDD was different among the groups [F(2,55)=20.72, p<.001]; the depressed
group had significantly higher scores than the other two groups, who were
comparable. The past-tense CES-D showed the same pattern [F(2,54)=79.91,
E<.00I], The IDD allowed scoring to determine if the depressed subjects fit
necessary criteria for "probable" or "definite" major depressive disorder (if they
fit 4 criteria or 6 criteria points, respectively). Only 4 o f 17 (24%) depressed
subjects met RDC criteria for "probable major depression", and none met
criteria for "definite depression". The remaining 13 subjects had symptoms of
depression but did not have any other diagnosable depressive spectrum disorder
(except possibly Depressive Disorder NOS). Three o f these "probably
depressed" subjects (18%) met DSM IV criteria for major depressive disorder.
This sample is less severely depressed than other published studies’ samples,
and is predominantly subclinically depressed. There were three sources of
recruitment o f this depressed group: TOACC, CRC, and a cutoff group.
Subjects from these three sources did not differ substantially in level o f
depression, as can be seen in Figure 1, although CRC clients report the highest
30 -
Figure 1
Naturally Depressed Subject1: Subgroups’ Means on CES-Ds
Volunteers, CES-D>16
N=7
S S 3 TOACC Clients
N=5
CRC Clients
N=5
to
VO
1 = Past 2 weeks
2 = Pre-Induct (1)3 = Post-Induc. (1)
4 = Pre-Induc. (2) 5 = Post-Induc.(2)
30
levels of depression overall. These subjects can be identified as showing
depressive symptoms, and are a mix of clinically and subclinically depressed
people. Calling them simply "depressed” in this report does not adequately
capture the mixture of subjects in this group, but this label will be used
throughout as a way to avoid inappropriate labelling.
Efficacy of Induction
To see if pre-induction and post-induction scores were significantly
different, planned paired T-tests were performed to test the hypothesis that
post-induction scores were equal to pre-induction scores. The mood induction
significantly increased depressed mood in the sample that underwent the
depressed mood induction. In the induced depressed group, CES-D scores
from the first session were significantly higher post-induction
[t(21)—2.52; £=.02], as were scores from the second session [t(20)=-
3.21;e =.004]. The other two groups, neutrally induced and depressed, showed
insignificant differences between pre- and post-induction scores on the CES-D
[neutral group, session one t(19)=1.38, £=.18; session two t(19)=1.86, £=08;
depressed group, session one t(16)=1.99, £=.07; session two t(16)=-,13, £=.90].
A visual representation o f the groups’ change on CES-Ds can be found in
Figures 2 and 3. Another method of testing to see if the mood induction
worked was to perform an analysis o f covariance (ANCOVA) on only the
neutral and induced subjects (not the depressed) CES-D scores after the
induction and covary out their pre-induction scores. The significant differences
Figure 2
CES-D Means
Pre and Post Induction, 1st Session
31
25 —
c
E
s
G
r
o
u
P
M
e
a
n
20.88
20 -
15 -
10 -
Neutrally Induced Depressed Induced Clinically Depressed
Pre-induction K \N Post Induction
C
E
S
G
r
o
u
P
M
25 —
2 0 -
15 -
10-
Figure 3
CES-D Means
Pre and Post Induction, 2nd Session
21.18 21-35
Neutrally Induced Depressed Induced Clinically Depressed
Pre-induction Post Induction
32
indicate the efficacy of the induction at influencing change [F (1,39)=8.118,
£=.007 for the main effect of the induction for session one; for session two,
F(I,38)=12.381, £=.001]. In both sessions, the induced depressed group had
higher post-induction CES-D scores than the neutral group.
Planned paired T-tests on the DACLs from both sessions also supported
the hypothesis that mood change occurred in the induced group but not the
other two. Pre-induction scores were compared with post-induction scores to
see if they were different. In the induced depressed group, DACL scores were
significantly higher post-induction in session one [t(21)=3.93, p=.001] and
session two [t(21)=3.47,£=.002]. T values for neutral and depressed groups
were nonsignificant in session one [neutral subjects t( 19)=-1.24, £=23;
depressed subjects t(16)=-1.07, £=.30], and in session two [neutral subjects
t(19)=.16, £=.87; depressed subjects t(16)=1.05, £=.31]. Mean DACL scores
for the three groups can be seen in Figures 4 and 5. It is worth noting that
CES-D and DACL scores were highly correlated; the two are not independent
tests, as they both tap depressive mood. The correlations between post
induction CES-D and DACL’s from sessions one and two ranged from r=.70 to
r=.83 (all £*s <.001).
The depressed group was expected to show the highest total CES-D
score, and the induction was expected to cause some amount o f increase in the
induced groups’s score so that their total score would fall in between the
Figure 4
DACL Means
Pre and Post Induction, 1st Session
33
D
14
A
C
12
L
10
G
r
8
0
u
6
P
4
M
e 2
a
D 0
Pre-induction y/ A Post Induction
1 = Neutral
2 = Induced Depressed
3 = Naturally Depressed
D
A
C
L
G
r
o
u
p
M
e
a
a
Figures
DACL Means
Pre and Post Induction, 2nd Session
Pre-induction y/ A Post Induction
1 = Neutral
2 = Induced Depressed
3 = Naturally Depressed
34
depressed and neutral groups’. To test the hypothesis that all three groups
would show comparable total post-induction CES-D scores, oneway analyses of
variance (ANOVAs) were done which indicated that there are significant
differences between the 3 groups in both sessions [session one F(2,56)= 10.36,
j)=.0001; session two F(2,55)=16.32, p<.0001]. Omnibus tests followed by
pairwise comparisons control the analysiswise type one error rate while
allowing specific pairs of groups to be contrasted; if the omnibus test is
significant, pairwise comparisons can occur with a protected alpha level (Cliff,
1987). Tukey’s Honestly Significant Difference (HSD) is recommended by
many as the best of the many available multiple comparison techniques and is
slightly more conservative than pre-planned contrasts or the Newman-Keuls
procedure (Howell, 1992). In both sessions, the ordering of total post-induction
CES-D scores from lowest to highest was neutral subjects, induced depressed
subjects, and then clinically depressed subjects. In the first session, differences
significant at the .05 level were found between neutral and clinically depressed
subjects and between neutral and induced depressed subjects, but not between
induced and clinically depressed groups. In the second session, all three groups
were significantly different from one another.
Another measurement o f post-induction mood was the DACL. Post
induction DACL scores were significantly different among the three groups in
session one [F(2,56)=13.12, p<.0001] and in session two [F(2,56)=12.55,
j><.0001]. In both sessions, Tukey’s HSD comparisons indicated that the
35
neutral group reported significantly less depressed mood than the other two
groups at the .05 level, but that the induced and naturally depressed groups did
not differ from one another.
CES-D Subscales
Subscale scores on the CES-D (Gatz & Hurwicz, 1990) were calculated
as mean scores for the items on each subscale, due to unequal numbers o f items
in each o f the four subscales. MANOVAs for each sessions’ post-induction
CES-D subscale scores as dependent variables confirmed that there were
significant differences among the three groups (pc.001). The univariate F tests
showed that all subscales were significantly different in both sessions except for
the Interpersonal Isolation subscale which was insignificant for session two. To
further compare group differences among the subscales, oneway ANOVAs were
done with Tukey’s HSD procedure to find significant pairwise differences.
These results are displayed in Tables 2 and 3 and Figures 6 and 7.
The hypothesis that induced depressed subjects would report greater
depressed mood and that naturally depressed subjects would report greater
psychomotor retardation, lack o f well-being, and interpersonal isolation was
only partially supported. Except for one case (Interpersonal Isolation, session
1), all four subscale scores were highest for naturally depressed subjects,
followed by induced depressed and then neutral subjects. In session one, the
induced and depressed groups did not differ significantly on any o f the 4
Table 2
36
CES-D Subscale Scores and Oneway ANOVA Results
Session 1
Neutrally
Induced
(Grp. I)
Induced
Depressed
(Group 2)
Depressed
(Group 3)
Omnibus
F (df)
and p
SignlL
Pairwise
Contrasts
d fo r
Tvs 3
d for
Tvs
2
M«d
S eb u i*
.05 (.10) 28 (.73) .97 (.66) F<226M 224
p-.OOO
1 & 3
L & 2
1.62 .93
P lf tb e e e w
Rtivdattaa
SobKale
2 7 (2 5 ) .73 (.75) .90 (.60) F(2JS)*>5J7
P-.006
I & 3
I & 2
1.05 .76
Lack a t Wttt*
tUJtJ
S lb n lt
23 (.69) .75 (.86) 122 1.66) F(2J6)-&55
px.003
1 & 3 120 27
loWrpttWaai
lafart—
S ib u U
.05(22) 2 2 (.84) 2 5 (25) F{226)-324
p-,047
1 & 2 JO .78
* Numbers within cells are subscale means (standard deviations).
Subscale scores are raieni»r«t u m«mt of items loading on thai subscale.
* d's given for 1 vs 3 refer to the effect size of depression on the subscale sane compared to the neutral group,
d's given for 1 vs 2 refer to U se effect size of induced depression on the subscale score compared to the neutral
group.
Table 3
CES-D Subscale Scores and Oneway ANOVA Results
Session 2
Neutral
(Grp. I)
Induced
Depressed
(Group 2)
Depressed
(Group 3)
Omnibus
F (df)
and p
SlgniL
Pairwise
Contrasts
d for
Tvs
3
d for
I vs 2
Mood-
Sobecdt
.06 (.11) .53 (.63) S 9 (.69) F(226)«14.13
p-.OOO
I & 3
1 & 2
2 & 3
1.75 .89
l*iyfki»»tar
RttirttKka
SetMOl*
2 4 (2 1 ) .67(22) 1.11 (29) F(225)-16.I8
p-.OOO
1 & 3
1&.2
2 & 3
128 .92
UckoTWdJ-
Mat
SatxaUi
.46 (.81) .70 (.70) 1.43 (29) F(226)«721
p a .002
1 4:3
2 & 3
122 20
[UtiqMIHttl
fuJeHiw
Sib*c*U
.025 (.11) •34(22) .47 (1.U7) F(226)=2J1
P-.108
cannot
compart;
Omnibus
insignil
.68 .48
* Numbers within cells are mlwgie m w w (standard deviations).
Subscale scores are as n e m of items in that subscale.
* c l’s for 1 vs 3 refer to effect size doe id depression on subscale score a m pared to neutral group, d’s for 1
vs 2 refer to effect size doe to Induced depressed mood on subscale score compared to neutral group.
Figure 6
Mean CES-D Subscale Scores
Post Induction, 1st Session
37
Subscale
Neutral N X ! Induced Depressed I 1 Naturally Depressed
0.05
0.58
; 0.97
0.27
^ w w w w v v m v m i Q . 7 3
1 0.9
0.33
3 v X X X X X X X X X X X X X X X X X X X X N ^ o : ? 5
I p;35
0.52
1 1.22
0.2 0.4 0.6 0.8
1 = Mood 2 = Psycbomotor Retardation
3 = Lack of Well-Being
4 = Interpersonal Isolation
Figure 7
Mean CES-D Subscale Scores
Post Induction, 2nd Session
1.2 1.4
Subscale
Neutral K X 1 Induced Depressed I I Naturally Depressed
i
0.06 : I
\\X v \X X X x X X \l 0 .5 3
599Q
0 5 4
XX vXXX X\ XXX ^sl 0.67 j
0.46
3 XsXXXXXXXXXXXXXXXXX 0.7
B 0.03__________
\X x x \x x x X i 034
0.47
0.99
1.43
02 0.4 0.6 0.8
1 = Mood 2 = Psychomotor Retardation
3 = Lack of Well-Being
4 = Interpersonal Isolation
1.2 1.4 1.6
38
subscales, but the neutral group was lower than one or both of the other two
groups on all 4 subscales. The induced group was significantly higher than the
neutral group on all but the well-being subscale, and was not different from the
depressed group. The induction therefore produced a comparable state to
depression. Session two scores do not repeat this pattern o f induced and
depressed similarity; the induced group was significantly lower than the
depressed group on 3 o f the 4 subscales (all but Interpersonal Isolation). If a
less conservative multiple comparison procedure was used, such as the
Newman-Keuls procedure, the induced and depressed groups would be
significantly different at the e <.05 level in session one on the Mood and Lack
of Well-Being subscales, but not the Psychomotor Retardation scale.
Interpersonal Isolation was only different between induced depressed and
neutral subjects in session 1 (induced depressed subjects had the highest score
on this subscale) and showed nonsignificant differences in session 2. The
depressed sample is, in general, more depressed than the induced group. The
one subscale hypothesized to show greater scores in the induced group, the
mood subscale, was significantly higher in depressed subjects than induced
subjects in session two but not in session one.
Effect sizes were computed for the effect of mood induced depression
and the effect o f depression on subscale scores. Both groups were compared to
the neutral group as a basis o f measurement. Effect sizes are a measure o f the
distance between groups based on a pooled standard deviation and regardless of
39
sample size. Cohen (1992) defines a small effect size as .2, medium as .5, and
large as .8. The effect sizes (d’s) for the mood induction and for depression
can be found in Tables 2 and 3. The Mood and Psychomotor Retardation
subscales show large effects in both induced and depressed groups. Lack of
Well-Being shows a large effect in the depressed group but only a small to
medium effect in the induced group. Interpersonal Isolation shows small to
medium effects in both groups. Consistently, depression shows a larger effect
than induced depressed mood. The mood and psychomotor retardation scores
are the most affected by the induction. The induction’s effects on the other
two subscales are more unclear. The subscale scores are significantly correlated
with one another in most cases; the correlation coefficients can be found in
Table 4.
To see if the different sources o f subjects in the depressed group played
a role in this groups* subscale scores, the means on the subscales were graphed
for each depressed subject source (see Figures 8 and 9). In session one, the
CRC clients had the highest scores on all but the Interpersonal Isolation
subscale, but in session two the volunteers had the highest scores on those same
subscales. While this pattern is intriguing, it does not suggest any consistent
difference in depression subscale scores based upon recruitment source.
Table 4
Correlations Between CES-D Subscales
40
M
1
PR
1
LWB
1
II
1
M
2
PR
2
LWB
2
II
2
Mood 1 1.0 .619
**
.668
**
.536
**
.815 .610
**
.568
**
.445
+♦
Psychomotor
Retardation 1
.619
**
1.0 .573 .603
**
.552
**
.767 .507 .284
Lack o f
Well-Being 1
.668
**
.573
**
1.0 .326
*
.615
**
.503
**
.723 .366
*
Interpersonal
Isolation 1
.536
**
.603
**
.326
**
1.0 .457
**
.503
**
.294 .479
Mood 2 .815
**
.552
**
.615
**
.457
**
1.0 .735
**
.679
**
.482
**
Psychomotor
Retardation 2
.610
* *
.767 .503
**
.503 .735 1.0 .616 .433
**
Lack o f
Well-Being 2
.568
**
.507
**
.723 .294 .679
**
.616
**
1.0 .441
Interpersonal
Isolation 2
.445
+ *
.284 .366
*
.479 .482
**
.433
**
.441 1.0
* <.05
* * <.01
Cognitive Performance
To examine the effect of group on selected cognitive tasks, oneway
ANOVAs were performed on each of the 13 cognitive tasks, with Tukey’s
HSD contrasts to identify pairs of significantly different groups if the omnibus
test was significant. The means and standard deviations on all measures can be
found in Table 5. The only task that showed significant differences among the
Figure 8
Depressed Subjects
Session 1
41
2 3 4
Subgroups’ CES-D Subscale Means
V olunteer, CES-D>16 E S 3 TOACC Clients C J CRC Clients
1 = Mood 2 = Psychomotor Retardation
3 = Lock of Well-Being
4 = Interpersonal Isolation - -
Figure 9
Depressed Subjects
Session 2
0.6 0.6
2 3 4
Subgroups’ CES-D Subscale Means
Volunteers, CES-D> 16 S 3 TOACC Clients C D CRC Clients
1 » Mood 2 = Psychomotor Retardation
3 = Lade of Well-Being
4 a Interpersonal Isolation
42
groups was Block Design [F(2,56)=4.56, p=.015]. Since we are doing 13
oneway ANOVAs on the cognitive task scores, a Bonferroni adjustment should
be used to control analysiswise alpha. With this strict criterion, even Block
Design would show nonsignificant differences with p=.015. The pairwise
comparisons for Block Design indicate that the clinically depressed subjects did
worse than the neutral subjects (p<.05). All other cognitive tasks showed
nonsignificant differences on the omnibus test, allowing no pairwise
comparisons to occur.
Due to small sample size and low power in the significance tests
reported, effect sizes were computed for the effect of mood induced depression
and the effect of depression, with both groups compared to the neutral group as
a basis o f measurement (see Table 5). Some of these effect sizes paint a more
promising picture than the nonsignificant results suggest. This study was able
to identify the large effect of depression seen on the block design task, but
other tasks with medium or small effect sizes would not be found significant
given the relatively low power o f this study. The pattern o f effect sizes suggest
pure motor speed (such as Cross-Offs or circling F’s) do not show a deficit in
depression or induced depressed mood, while tasks which require more
cognitive effort, such as sentence memory, Logical Memory, and Visual
Reproduction, show some small or medium effects. But for a few exceptions
(Circling ]’s, block design) it would seem that memory tasks are affected by
depressed mood while speed tasks are not. The exceptions are notable
Table 5
Performance on Cognitive Measures
Neutrally
Induced
(1)
Depressed
Induced
(2)
Naturally
Depressed
(3)
d for
1 vs 3
d for
1 vs 2
# of Cross-Offs
In 45 seconds
89.10
(18.67)
88.23
(17.09)
86.18
(23.56)
-.149 -.048
Seconds to
Circle F ’s
121.50
(38.19)
119.05
(54.60)
126.18
(67.87)
-.087 .045
Seconds to
Circle )’s
143.37
(23.21)
155.85
(38.92)
168.94
(83.29)
-.476* -.227*
Digit Symbol 43.85
(10.19)
42.95
(9.94)
44.06
(14.05)
.018 -.079
Block Design 28.90
(7.83)
27.23
(10.28)
19.65
(11.21)
-.944*** -.182
Logical Memory I 24.63
(7.10)
22.36
(6.86)
20.44
(4.84)
-.650** -.380*
Logical Memory
II
18.00
(7.42)
17.77
(8.71)
14.53
(6.11)
-.456* -.030
Visual
Reproduction I
32.00
(5.88)
30.14
(6.10)
27.18
(9.42)
-.676** -.260*
Visual
Reproduction II
23.95
(8.14)
22.18
(10.41)
18.65
(10.34)
-.549** -.183
Sent. Mem.
Low-Frcq.
Neutral
.66
(.19)
.56
(.16)
.58
(.10)
-.523** -.680**
S ent Mem.
High-Freq.
Neutral
.71
(.15)
.62
(.18)
.57
(.23)
-.712** -.472*
S en t Mem.
Low-Frcq. Sad
.68
(.21)
.69
(.22)
.76
(.10)
.389* .019
Sent. Mem.
HEgh-Freq. Sad
.82
(.17)
.77
(-14)
.75
(.19)
-.437* -.357*
44
Table 5 (continued)
• All scores given are raw, not scaled scores. Sentence Memory scores are percentages of
correct idea units remembered.
• d’s given for I vs. 3 refer to the effect size of the deficit in the naturally depressed group
compared to neutral group, d’s given for I vs. 2 refer to the effect size of the deficit in the
induced depressed group compared to the neutral group.
• Negative d values signify the depressed group performed worse than the neutral group (either
lower scores or slower performance). Positive d values signify better performance in the
depressed group.
• * = small effect. ** = medium effect. *** = large effect.
however, especially since the effect o f depression on Block Design was the
only large d found. In general, the depressed group showed larger effects than
the induced group. The mean effect size on all o f the tasks was mean d = -
.404 for the depressed group (variance o f d = 1.54, error variance = .016) and
mean d = -.218 for the induced group (variance o f d = .538, error variance =
.013). These mean effect sizes are larger than the mean effect sizes for
outpatient depressed samples in our meta-analysis (Fox, Knight, & Zelinski,
1995).
To determine the relationship between intensity o f depression and
performance on the cognitive tasks, correlations were computed o f the total
post-induction CES-D scores and their four subscale scores with the cognitive
tasks. The correlations were computed for the whole sample, pooling the three
groups. For the whole sample, most correlations were approximately zero, with
only an occasional coefficient above .2; two coefficients out o f 152 were
significant (the Psychomotor Retardation subscale o f session one correlated
45
with the number o f F’s circled in the visual search task, r = -,338, p<.01; and
the Interpersonal Isolation subscale in session two correlated with the number
o f seconds to circle the ]’s, r = .597, £<.001), but they seem to be only
randomly and inconsistently related and are probably due to chance significance
due to the large number of coefficients computed.
Since it may be the case that non-depressed people do not show a
relationship between mood and cognitive functioning, but that more depressed
people show the effect, it was decided post hoc to compute correlations
separately for each group. The depressed groups would be expected to show
higher magnitudes o f correlations. The non-depressed group may have washed
out the effect in the whole sample. When the CES-D subscales from session
one were correlated with the cognitive measures for the neutral subjects, five o f
76 correlations were significant. However, only one of these correlations was
repeated in the second session [the Interpersonal Isolation subscale from both
sessions correlated with the number o f [’s circled (r = -.826, £<.001)].
When the correlations were computed for the induced depressed
subjects, five correlations were significant. They showed no discemable pattern
except that 4 of the 5 correlations with CES-D subscales were subsets of the
sentence memory task, although these correlations were not consistent across
sessions. The clinically depressed subjects showed six significant correlations
between CES-D subscales and cognitive measures. There was no consistent
pattern to the findings. Again, the significant correlations are probably due to
46
chance. All o f the significant correlations found must be interpreted with
caution. With the number o f coefficients generated, 5 percent or more would
be expected to turn out significant just by chance since the cognitive tasks are
not likely to be independent.
DISCUSSION
The results o f this study were affected by the type o f subjects in the
sample, the procedure, the sample size, and a number o f other factors. An
understanding o f the somewhat unclear results requires discussion o f these
factors, the meaning behind the significant and nonsignificant results, and the
issues this study raises for future research.
Subjects
The depressed group in this study consisted o f mostly subclinically
depressed rather than clinically depressed according to RDC or DSM IV criteria
(hence calling them depressed and avoiding the term clinically depressed).
This indicates that while all o f the subjects were experiencing significant
depressive symptoms over the two weeks prior to the study as measured by the
CES-D, and most were seeking help, these subjects did not all have the
diagnostic syndrome o f depression. The fact that few had syndromal
depression may make significant results harder to find; any differences between
the depressed group and the other groups might be even larger in a clinically
depressed sample. It may also be the case that the induced group is more
47
similar to this subclinically depressed group than it would be to a clinically
depressed group, not just in severity but possibly also in profile. It is not clear
if people with clinical and subclinical depression differ in depressive
symptomatology and on their subscale scores on the CES-D.
The Induction and its Effects
The results indicate that the mood induction procedure does work in an
older sample, significantly increasing depressed mood as measured by the
present-tense CES-D and the Depressive Adjective Checklist. The neutral
procedure did not influence the mood o f the non-depressed and the depressed
subjects who underwent that procedure. Unfortunately, there is no standard to
judge what percentage of subjects experienced a successful induction. Previous
studies have not consistently used any particular measurement o f mood change.
Operationalizations o f mood change have ranged from visual analog scales and
adjective checklists, to writing speed and decision time (Martin, 1990). While
some studies have used the benchmark o f a 10% change on visual analog scales
to operationalize successful induction (Martin, 1990), it is unclear how large a
difference score between pre-induction and post-induction CES-Ds or DACLs
would have to be to consider an individual induction to be successful. The
CES-D has not been used for this purpose before, but was used due to the
subscales that tap different aspects o f depression.
The total CES-D scores for both sessions supported the hypothesis that
depressed subjects would show the highest total CES-D score, followed by
48
induced depressed and then neutral subjects. DACL results suggest that the
mood induction induces a mood state comparable to natural subclinical
depression. Mood induction studies usually use measures such as the DACL or
other gross measures o f depressed mood without analyzing the components o f
the resulting depressed mood. The motivation for looking at different aspects
of the experience o f depression prompted use o f the subscales within the
CES-D.
The induced depressed group experienced a significant change from
their pre-induction depression scores, and the strength o f their depressed state
was comparable to the depressed subjects in session one but not session two.
While the induced group’s subscale scores are lower but generally comparable
to the depressed group’s in session one, they are significantly lower in session
two. The exception to this is the interpersonal isolation subscale; induced
subjects had the highest score on this subscale in session one, but none o f the
three groups differed significantly on this subscale in session two. The results
on the interpersonal isolation subscale are difficult to interpret. The lack of
well-being subscale shows the smallest effect due to the induction, although it
is significantly elevated in the depressed subjects. The induced group’s well
being scores are not different from the neutral group’s in either session. The
two subscales most clearly affected by the induction are mood and psychomotor
retardation. Even though the significance testing on the second session
subscale scores show these subscales to be lower in induced than depressed
49
subjects, the effect sizes demonstrate the induction’s large effects on these
subscales.
When retesting occurs, it is important to consider practice effects.
Subjects may not have responded as well to the induction in session two since
it was no longer novel, thereby causing their scores to be significantly lower
than depressed subjects’. This is supported by the results o f the DACLs in
both sessions. Induced subjects had higher post-induction DACL scores than
depressed subjects in session one, but they were not quite as large (although
still significant) in session two.
These results do not disconfirm the similarity o f induced to depressed
subjects but raise many questions about the nature of the induction’s effects.
Upon visual inspection, the induction looks like a weaker yet similar version of
the depressed group on every subscale except possibly Interpersonal Isolation.
The induced group had slightly lower post-induction CES-D subscale scores in
session two, which may have been enough to make the differences between
them and the depressed subjects significant.
Contrary to our hypotheses, the mood induction seems to be acting on
not just sad mood, but also on somatic symptoms. It may also be affecting
other aspects comprising the experience o f depression, e.g. the interpersonal
isolation and the decrease in positive well-being. Results on these two
subscales were not as strong or clear. While the mood induction does just act
directly on what people think in the laboratory (as a cognitive-mediated
50
procedure) and not directly on the other aspects of depression, it may be that
forcing subjects to ruminate about the other aspects o f depression (e.g. somatic
symptoms, interpersonal isolation) causes them to think that the other factors
are problematic. In a circular fashion this could thereby increase sad mood.
This makes sense given the types of Velten statements we used that deal with
just these topics (e.g. "I feel discouraged and drowsy; maybe I’ll take a nap
when I get home" and "People annoy me; I wish I could be by myself').
Martin (1990) posited that both the self-statement and music methods act on
cognitive, somatic, and emotional aspects of mood, unlike some other
procedures that do not focus on all three. The results suggest that mood
induced depression is a valid analog to natural depression, although the induced
state achieved by subjects in this study was not as strong, and there were slight
differences between induced and depressed groups in the profile o f the four
aspects of depression. The induction procedure has been shown in this study to
influence more than just mood; as suggested by Martin, the induction procedure
o f self-statements and music influenced subjects on a number of dimensions.
Riskind and his colleagues (Riskind, Rholes, & Eggers, 1982; Rholes,
Riskind, & Lane, 1987) found that different Velten procedures, one focusing on
somatic statements and the other focusing on self-devaluative statements,
produced different effects on memory. Riskind (1989) reported that the "self-
devaluative statements tended to influence a subject’s recall of negative as
opposed to positive personal memories, while the negative somatic statements
51
did not." He proposed that "depressed mood may disrupt attention because
depressive cognitions invade attention and distract the individual...on the other
hand, a reduced state of arousal may reduce the effort that the individual has
available to devote." Both of these mechanisms would predict deficits in
memory, although the self-devaluative statements would seem to be more
related to mood congruency than somatic statements given the proposed action
of the somatic statements. These studies were done with younger subjects; it is
unknown if the same relationships between type of statement and cognitive
effect exist for mood induced older adults.
Since we selected our induced group to be non-depressed before the
induction procedure, we may have limited our group to those least likely to
respond well to the induction. Martin (1990) suggests that those with a pre
existing schema for depression can access these thoughts easiest when asked to
do so in an induction procedure. She also suggests that non-depressed people
generally have positive schema, subclinically depressed people have some
positive and some negative schema, and clinically depressed people have
mostly negative schema. Subclinically depressed people can draw on both
types o f schema and use the negative schema if in a depressogenic situation
such as the induction procedure. By selecting our sample for non-depressed
people, we may have reduced our chances of having subjects with some
depressive schema to draw upon, thereby reducing the strength o f the effect of
the induction.
52
It may be that pre-existing depressive affect is related to the ability to
become induced into a depressed mood. This could be examined by correlating
past-tense CES-D scores with the amount of change on CES-D or DACL due
to the induction. A positive correlation would run counter to logical
expectations; the ceiling effect would predict that the higher original scores are,
the less room subjects have to increase (although subjects all had plenty of
room to increase on the CES-D). A positive correlation between original score
and change due to induction would support Martin’s idea. When these
correlations were computed post hoc, they were nonsignificant; past tense CES-
D correlated with CES-D change r = .33 in session one, and r = 0.0 in session
two. The amount of change itself was highly correlated in session one and
two, r = .71, (p<.001). Past tense CES-D correlated with change on the first
session DACL only r = -.10 and with the second session DACL r = -.23.
Again the amount o f change on the DACLs was highly correlated, r = .82
(E<.001), These correlations suggest pre-experiment level o f depressive
symptomatology is not related to the efficacy of the induction at changing
mood. The present sample is small, limiting power to find significant
correlations, but if the null hypothesis is true and there is no relationship, it is
difficult to use Bower’s theory to explain how mood induction works. Bower’s
network theory posits that subjects need to have some existing depressive
material to draw upon and associate with when in an induction that can
maintain the induced thoughts and feelings o f depression and increase sad
53
mood. This hypothesis could be better tested if a measure specifically designed
to tap depressive schema had been included as a pre-measure to correlate with
change scores instead o f the CES-D, a more symptom-based measure.
It may also be the case that the induced group had positive schema that
they utilized to modulate the induction procedure. Older adults have been
shown to have increased affective modulation skills over time. The induction
was shown to be effective immediately post-induction but the subjects may
have recruited competing positive associations (Singer & Salovey, 1988) to
combat the undesirable depressed mood. Support for this idea includes the fact
that mood congruency effects are stronger and more reliable in happy moods
than in sad moods (Salovey & Singer, 1988). People don’t want to stay in sad
moods, so they do what they can to get out o f them. Anecdotal evidence from
this study suggests that this may be the case. Subjects reported not enjoying
the sad moods and focusing on the cognitive tasks to get their minds onto
something else. The tasks then became distractors and mood influences
themselves. Many subjects reported that their emotions, and the words they
checked on the DACL, were influenced by the immediately preceding cognitive
task. Remission o f depression may use a similar mechanism, by deliberate
short circuiting o f the negative spiral.
The data also indicate that for all the groups, in both sessions, Lack o f
Well-Being is the factor that subjects score highest on o f the four; they all
experience more lack of positive emotion than negative emotion. This was
54
found by Gatz & Hurwicz (1990) to be true in the oldest group in their sample
(age 70-98) but not the younger ones. Their youngest subjects (age 20-39)
scored highest on the Depressed Mood subscale. If younger subjects show a
greater amount o f negative affect and older subjects show a greater deficit o f
positive affect in depression, this is evidence o f a real difference in the
experience o f depression with age. These differences may also affect the
cognitive sequelae o f depression if, as is hypothesized in this study, mood is
related to cognitive performance. Alternately, the Well-Being subscale o f the
CES-D has been shown to relate to cognitive functioning in a group o f healthy
elderly men (LaRue, Swan, & Carmelli, 1995). These findings in a healthy
sample may not hold true for a more depressed sample because o f the
previously mentioned differences between depressed and non-depressed people.
The depressed group was hypothesized to have a lower mean score on
the sad mood factor o f the CES-D because it was thought that they would not
consistently be in a depressed mood at the time o f testing. Although the
criterion for inclusion in the depressed group was based on the CES-D
assessing the past two weeks (not mood at testing), 12 o f these 17 subjects
(70%) also met cutoff criteria for significant depressive symptomatology at the
time o f testing. A majority o f the subjects were in a depressed mood at the
time o f testing, but certainly not all. This group did have the highest levels o f
depressed mood, even given the fact that not all o f them were in a significantly
depressed mood at the time o f testing.
55
Cognitive Performance
It was hypothesized that the depressive state would influence subjects’
performance on various cognitive tasks. One hypothesis was that mood itself
was the greatest predictor of cognitive deficit, so the group with the predicted
greatest depressed mood, the induced depressed group, was hypothesized to
show the greatest cognitive deficits. The sad mood was strongest in the
depressed group, however. Consistent with the general hypothesis that mood is
the greatest predictor of cognitive deficit, it would then be expected that this
group would show the greatest cognitive deficit. The one cognitive task that
differed among the groups was Block Design; the depressed group did worse
than the other two groups. Even Block Design would be considered
nonsignificant if we were strictly applying the Bonferroni adjustment, so
interpretation of this result merits caution. None of the other tasks showed
significant differences among any of the groups, although on 8 o f the 12 other
tasks the depressed performed somewhat worse than the induced depressed
who, in turn, performed somewhat worse than the neutral subjects. It was
expected that cognitive performance would reflect depressed mood, so this
pattern o f performance is consistent with the pattern of scores on the depression
measures. The fact that not all o f the tasks showed this pattern o f performance
suggests that maybe the deficit in depressed mood is not a robust phenomenon.
56
More likely, however, is that low power contributed to the nonsignificance of
many findings.
Calculating effect sizes helps to determine the size of the difference
among the groups, regardless o f the sample size. The d ’s are more consistent
with the expected effects of depression than the non-significant ANOVAs
indicate. There are one large, five medium, and four small effects on the
cognitive tasks for the depressed group, and there are one medium and five
small effects for the induced group. Since the depressed group had the
strongest depressed mood, it is not surprising that this group showed the
greatest effect sizes. The tasks which showed deficits tended to be the memory
tasks rather than psychomotor speed, although the strongest effect was for a
timed fluid ability task, Block Design.
Those highest on the mood subscale were expected to show the greatest
cognitive deficit, due to the hypothesized relationship between mood and
cognitive functioning. The correlational analysis showed no discemable pattern
in any of the groups between mood and cognitive performance. If the mood
subscale in the depressed group (the group with the highest mood scores) was
highly negatively correlated with cognitive scores, a stronger relationship could
be inferred for mood itself and cognitive deficit. Unfortunately, a large
negative correlation coefficient (without corresponding large correlations with
the other subscales) would be unlikely due to the high intercorrelations among
the CES-D subscales (see Table 4).
57
The mood congruency effect was not supported by significant findings
in this study, but the effect sizes give us a basis for discussing these results.
The sentence memory task contains two subsets of sad sentences; low
frequency and high frequency. The low frequency sad sentences (which would
be harder to remember given the effort necessary to remember the unique order
o f the words) were best remembered by the depressed subjects as would be
predicted, and were equally remembered by induced and neutral subjects. The
difference between depressed and neutral groups was not large enough to be
significant; the effect size was small (.389). This pattern was reversed for the
high frequency sad sentences, however. Both induced and depressed subjects
remembered less of these sentences than neutral subjects, counter to the mood
congruency hypothesis. These results are inconclusive. With a more powerful
design it may be possible to support or disconfirm this hypothesis.
The correlation matrices obtained for CES-D factors with cognitive
tasks showed no consistency or pattern of significant findings, either for the
whole sample or for each group by itself. It seems that in this sample there is
no consistent relationship between aspects o f depression and cognitive
functioning. As with the other analyses of the cognitive measures, low power
may be contributing to the null findings. This is likely since depressed people
have been shown to have deficits on the tasks used in this study. The
depressed subjects may not be depressed enough to have real deficits.
Induction may cause different effects than naturally occurring depressed mood,
58
such as mainly congruency rather than broad deficits. If these two suppositions
are true, than the results obtained in this study are true and not merely'type two
errors. Further research may determine if this is the case.
Limitations o f the Study
There were several methodological weaknesses that may have
contributed to the low power and nonsignificant findings in this study. Effect
size is the degree to which the null hypothesis (H0 ) is false, indexed by the
discrepancy between H0 and H,. Cohen (1992) states, "My intent was that
medium effect size represent an effect likely to be visible to the naked eye o f a
careful observer. (It has since been noted in effect-size surveys that it
approximates the average size o f observed effects in various fields.)" Cohen
(1992) defines a small effect size for the difference between two independent
means as .20, a medium effect size as .50, and a large effect size as .80. Effect
sizes for cognitive task performance are shown in Table 5. With the small
sample size used in this study, it would be unlikely to find significance for
small or medium effects. Cohen charts the number o f subjects necessary per
group to find small, medium, and large effects with .80 power at .01, .05, and
.10 alpha levels. To find small effects with a .05 alpha level, it would have
been desirable to use 393 subjects per group. Even medium effects would
require 64 subjects per group to get significant results. In fact, with our sample
size o f approximately 20 per group, only very large effect sizes would be found
significant. The argument can be made that small effects are unimportant
59
effects, and that small N studies are justifiable since they only find significance
when the effect is large. The opposing argument, however, is that even
medium or small effects can be very important and should not be missed (and
assumed nonexistent) due to studies with small sample sizes. Unfortunately,
pragmatic considerations made recruitment o f a large sample impossible. It is
worth noting that many other studies using mood induction have very small
sample sizes; the present sample size (total N=59) was comparable to others in
the mood and memory area o f study (e.g., Salovey & Singer, 1985, experiment
1, N=60, experiment 2, N=30; Riskind, Rholes, & Eggers, 1982, N=52;
Teasdale & Russell, N=32; Teasdale & Taylor, 1981, N=37). These published
studies found significant effects supporting mood congruency as a robust
phenomenon, but there are probably many unpublished "file-drawer" studies
with non-significant results given the average sample size in this field. Mood
congruency may also be a larger effect in induced subjects than the other types
o f deficits this study expected to find.
Mood induction is typically done by cognitive researchers. In none o f
their published reports do they mention selecting subjects for induction who
have been screened to be sure they are non-depressed before inducing them.
As clinical psychologists, we were ethically concerned about inducing
depression in anyone who is already experiencing depression. We also wanted
to avoid overlap o f the induced and depressed group on the depression
measures at the starting point. This selection may have actually screened out
60
the most induceable subjects. Inducing happy people may be harder to do than
inducing someone who already has a tendency to feel sad or is sensitive to cues
that prompt sad feelings. This selection bias may have resulted in lower
average depressed mood in the induced group, and therefore fewer significant
effects.
Another limitation which may explain the failure to find significant
cognitive deficits in at least the depressed group was the fact that most o f these
subjects did not fit Research Diagnostic Criteria (RDC) for Major Depressive
Episode. These subjects may not have the same level o f cognitive deficits as
those who fit RDC criteria. Our meta-analysis indicates that more severe
depression as measured by inpatient status (an imperfect operationalization of
severity) or by measures such as the Beck Depression Inventory or Hamilton
Depression Scale is related to cognitive deficit (Fox, Knight, & Zelinski, 1995).
It may also be that cognitive deficits only show up in people who are quite
significantly depressed, and we did not have any subjects experiencing this
level o f severity. If this is the case, other studies’ samples may contain a
mixture o f very depressed people with cognitive deficit and some less
depressed people without deficit. The reported means show group differences
compared to non-depressed subjects but may not capture the individual
variability in the performance of the depressed subjects. The other possibility
is that cognitive deficit lies on a continuum, with less depressed people
61
experiencing smaller cognitive effects. Small effects are difficult to find unless
power is large, and this study had the power to identify only large effects.
The gender balance o f this sample is not equal among the groups. The
neutral group has a higher male to female ratio. The effect of this difference is
unknown. Gender may play a role in the effects o f the mood induction; some
mood induction researchers report that female subjects may be more susceptible
to the technique than men (Gouaux & Gouaux, 1971) but it may be that women
are more typically familiar with the emotionally laden material, giving them an
advantage in becoming induced. Our induced group had a similar gender
distribution to the depressed group, and the higher proportion o f women may
have helped the "induceability" o f the group. The gender inequalities should be
considered a potential confound in the reporting of depression and its
symptoms. Men generally report less depression than women and are less often
help-seekers which may have influenced our population from which to recruit
depressed subjects. Confounds may also exist in the cognitive performance
results; some evidence suggests men are better in general at spatial tasks and
women better at verbal ones (Lim, 1994; Hyde, 1990).
Implications and Recommendations
This study did not include enough measures that specifically
assessed mood congruency along with deficits. Only four sentences on the
sentence memory task were sad sentences, and these four were split into two
subscores (two were low-frequency and two were high-frequency). A well
62
designed measure using valenced and neutral words in both immediate and
delayed recall as well as recognition would tap different kinds o f cognitive
processing and allow a more developed discussion o f mood congruency versus
deficit models of cognition in depression. Cognitive processing models have
influenced thinking about the etiology and mechanisms of depression. Clear
support for one of the existing models or a synthesis of several may be
clinically useful in treatment o f depression.
It would also be potentially interesting to use a visual memory task with
valenced and neutral pictures as well as the verbal task. Visual and verbal
performance differ among people and use different processes and areas o f the
brain. Gender differences in these abilities may exist in emotional or valenced
processing as well as the general differences seen in non-valenced processing.
It is unclear thus far whether emotion such as depressed mood would
necessarily effect verbal more than visual processing.
Future research should use a minimum o f Cohen’s recommended 64
subjects per group to find medium effects of depression on cognition (Cohen,
1992). A clearly clinically depressed group o f subjects would probably also
increase the difference among the groups, thereby increasing the chance o f
finding an effect. The difficulty in finding such a sample is that other
confounds are usually introduced with increasing severity o f depression such as
medications or hospitalization. It may be of interest to test both subclinically
and clinically depressed groups (separately) along with the induced depressed
63
and non-depressed groups to see if cognitive deficit or mood congruency is
related to the severity o f the natural depression.
While this study did not find a consistent relationship between depressed
mood and cognitive functioning, validating the mood induction procedure and
examining its effects is a preliminary step to further understanding the role
emotion plays in the cognitive functioning o f older adults. Future studies
should replicate mood congruency effects with older adult samples, and include
representative groups or groups o f various ages instead of solely
undergraduates.
Mood induction may prove to be a valuable paradigm for more than the
study o f cognition. This technique may allow further research into depression
itself, for instance regarding the role schemas play in the development and
maintenance o f depression, how people get themselves out o f depression, or the
accuracy o f depressed people’s perceptions. Future mood induction research
may find that the presence o f negative schemas makes people more susceptible
to the mood induction procedure, which may be informative about the qualities
o f people who are prone to depressive episodes. Closely monitoring and
examining mood change as people recover normal mood after being induced
into a depressed mood may be indicative o f the reparative work depressed
people do to become nondepressed (or the failures they encounter in this effort
that keep them depressed). Future research could assess induced subjects’ use
of competing thoughts to remove the induction’s effects and compare this to the
64
behavior o f depressed people, for example. Another valuable aspect of
induction is that normally nondepressed people can be objectively measured on
a variety o f personal constructs, induced into a sad mood, and then measured
on their perceptions o f the constructs. Depressed people make negative
judgements o f themselves and have negative perceptions, but their perceptions
may actually be accurate rather than negatively distorted. The mood induction
procedure allows a test of this hypothesis. Emotional understanding is different
in older adults; in general, older adults have more complex emotional processes
and perceptions of experiences (Labouvie-Vief, DeVoe, & Bulka, 1989). All of
these questions about depression may have different answers for younger and
older adults. An induction technique that is valid for all ages is an important
tool to allow these questions to be answered for people o f various ages.
Conclusions
This study contributes to current knowledge by validating the mood
induction procedure for use with older adults. It shows that the mood induction
procedure works and is a reasonable analog to naturally occurring subclinical
depression. It is unclear whether the technique is a valid analog to clinical
depression. The induction procedure acts on more than just mood itself. It is
effective at changing normally non-depressed older peoples1 perceptions o f their
mood and somatic state, and perhaps also their interpersonal interactions and
well-being. The relationship between symptoms of a depressed state and
cognitive processes is still unclear. Future studies with larger, more distinct
samples may shed light on the nature o f depressed mood and its effects
cognition and memory.
6 6
REFERENCES
Alloy, L. B., Abramson, L. Y., & Viscusi, D. (1981). Induced mood
and the illusion of control. Journal o f Personality and Social Psychology. 41.
1129-1140.
Austin, M.P., Ross, M., Murray, C., O ’Carroll, R.E., Ebmeier, K.P., &
Goodwin, G.M. (1992). Cognitive function in major depression. Journal of
Affective Disorders. 25, 21-30.
Beck, A.T. (1967). Depression: Causes and Treatment. Philadelphia:
University o f Pennsylvania Press.
Beck, A. T. (1967). Depression: Clinical, experimental, and theoretical
aspects. New York: Hoeber.
Bower, G. H. (1981). Mood and memory. American Psychologist. 36.
129-148.
Bower, G. H., & Mayer, J. D. (1985). Failure to replicate mood
dependent retrieval. Bulletin of the Psvchonomic Society. 23. 39-42.
Burt, D. B., Zembar, M. J., & Niederehe, G. (1995). Depression and
memory impairment: A meta-analysis o f the association, its pattern, and
specificity. Psychological Bulletin. 2. 285-305.
Carstensen, L. (1992). Social and emotional patterns in adulthood:
Support for socioemotional selectivity theory. Psychology and Aging. 7. 3,
331-338.
Carstensen, L., & Turk-Charles, S. (1994). The salience o f emotion
across the adult life span. Psychology and Aging. 9. 2, 259-264.
Cassens, Wolfe, & Zola (1990). The neuropsychology o f depressions.
Journal o f Neuropsychiatry. 2. 2, 202-213.
Clark, V. A., Aneshensel, C. S., Frericks, R. R., & Morgan, T. M,
(1981). Analysis o f effects o f sex and age in response to items on the CES-D
scale. Psychiatry Research. 5. 171-181.
Clark, D. M. & Teasdale, J. D. (1985). Constraints o f the effects of
mood on memory. Journal o f Personality and Social Psychology. 48. 1595-
1608.
67
Clark, D. M., Teasdale, J. D., Broadbent, D. E. & Martin, M. (1983).
Effect of mood on lexical decisions. Bulletin of the Psvchonomic Society. 21.
175-178.
Cliff, N. (1987). Analyzing Multivariate Data. New York: Harcourt
Brace Jovanovich.
Cohen, J. (1992). A power primer. Psychological Bulletin. 112. 1,
155-159.
Cole, K., & Zarit, S. (1984). Psychological deficits in depressed
medical patients. Journal of Nervous and Mental Disease. 172. 3, 150-155.
Denny, E.B., & Hunt, R.R. (1992). Affective valence and memory in
depression: Dissociation o f recall and fragment completion. Journal of
Abnormal Psychology. 101, 3, 575-580.
Ellis, H. C. & Ashbrook, P. W. (1988). Resource allocation model of
the effects of depressed mood states on memory. In K. Fiedler & J. Forgas
(Eds.), Affect, cognition and social behavior. Toronto, Canada: Hogrefe.
Ellis, H. C. & Ashbrook, P. W. (1989). The state of mood and memory
research. In D. Kuiken (Ed.), Mood and memory: Theory, research, and
applications. [Special Issue], Journal of Social Behavior and Personality. 4. 2,
1-21.
Feehan, M., Knight, R. G., & Partridge, F. M. (1991). Cognitive
complaint and test performance in elderly patients suffering depression or
dementia. International Journal of Geriatric Psychiatry. 6. 287-293.
Fox, L. S., Knight, B., & Zelinski, E. (1995). Unpublished data, to be
presented as "A meta-analysis o f cognitive deficit in depression across the
lifespan," at the annual meeting o f the Gerontological Society o f America, Los
Angeles, California (November, 1995).
Gatz, M. & Hurwicz, M. (1990). Are old people more depressed?
Cross-sectional data on Center for Epidemiological Studies Depression Scale
factors. Psychology and Aging. 5. 2, 284-290.
Gray, J. W., Dean, R. S., Rattan, G., & Cramer, K. M. (1987).
Neuropsychological aspects of primary affective depression. International
Journal of Neuroscience. 32. 911-918.
68
Gouaux, C. & Gouaux, S. M. (1971). The influence of induced
affective states on the effectiveness o f social and non-social reinforcers in an
instrumental learning task. Psvchonomic Science. 22. 341-343.
Hale, W. D., & Strickland, B. R. (1976). Induction of mood states and
their effect on cognitive and social behaviors. Journal o f Consulting and
Clinical Psychology. 44. 155.
Haney, J. N., & Euse, F. J. (1976). Skin conductance and heart rate
responses to neutral, positive, and negative imagery: Implications for covert
behavior therapy procedures. Behavior
Therapy. 7. 494-503.
Hartlage, S., Alloy, L.B., Vazquez, C., & Dykman, B. (1993).
Automatic and effortful processing in depression. Psychological Bulletin. 113,
2, 247-278.
Hasher, L., & Zacks, R.T. (1979). Automatic and effortful processes in
memory. Journal of Experimental Psychology; General. 108, 356-389.
Hertzog, C., Van Alstine, J., Usala, P. D., Hultsch, D. F., & Dixon, R.
(1990). Measurement properties o f the Center for Epidemiological Studies
Depression Scale (CES-D) in older populations. Psychological Assessment. 2.
64-72.
Hill, A.B., Kemp-Wheeler, S.M., & Jones, S.A. (1987). Subclinical and
clinical depression: Are analogue studies justifiable? Personality and Individual
Differences. 8, 1, 113-120.
Howell, D. C. (1992). Statistical Methods or Psychology. 3rd Edition.
Duxbury Press: Belmont, California.
Hyde, J. S. (1990). Meta-analysis and the psychology of gender
differences. Signs. 16. 1, 55-73.
Ingram, R. E. (1984). Toward an information-processing analysis of
depression. Cognitive Therapy and Research. 8. 443-478.
Krause, N. (1986). Stress and sex differences in depressive symptoms
among older adults. Journal o f Gerontology. 41. 727-731.
Krause, N., & Markides, K. S. (1985). Employment and psychological
well-being in Mexican American women. Journal o f Health and Social
Behavior.26. 15-26.
69
Labouvie-Vief, G., DeVoe, M., & Bulka, D. (1989). Speaking about
feelings: Conceptions o f emotion across the life span. Psychology and Aging.
4, 425-437.
LaRue, A., Swan, G., & Carmelli, D. (1995). Cognition and depression
in a cohort o f aging men: Results from the Western Collaborative Group Study.
Psychology and Aging. 10. 1, 30-33.
Liang, J., Tran, T., Krause, N., & Markides, K. S. (1989). Generational
differences in the structure o f the CES-D scale in Mexican Americans. Journal
o f Gerontology; Social Sciences. 44. S110-S120.
Lim, T. K. (1994). Gender-related differences in intelligence:
Application o f confirmatory factor analysis. Intelligence. 19. 2. 179-192.
Lubin, B. (1981). Depression Adjective Checklists: Professional Manual.
Psychological Assessment Resources, Inc., Odessa, FI.
Lyness, S. A., Eaton, E. M., Schneider, L. S. (1994). Cognitive
performance in older and middle-aged depressed outpatients and controls.
Journal o f Gerontology: Psychological Sciences. 49. 3, P129-P136.
Martin, M. (1990). On the induction o f mood. Clinical Psychology
Review. 10. 669-697.
Matt, G. E„ Vasquez, C., & Campbell, W. K. (1992). Mood-congruent
recall o f affectively toned stimuli: A meta-analytic review. Clinical Psvcholoev
Review. 12. 227-255.
Mullen, J. A., Orbuch, T. L., Featherman, D. L., & Nessleroade, J. R.
(1988). Stability o f the structure o f the CES-D in an elderly sample. Paper
presented at annual meeting o f the Gerontological Society o f America, San
Francisco.
Natale, M. (1977a). Effects o f induced elation-depression on speech in
the internal-external locus o f control. Journal o f Consulting and Clinical
Psychology. 45. 45-52.
Natale, M. (1977b). Induction o f mood states and their effects on gaze
behaviors. Journal o f Consulting and Clinical Psychology. 45. 560.
Natale, M. (1978). Effect o f induced elation and depression on internal-
external locus o f control. Journal of Psychology. 100. 315-321.
70
Natale, M. & Gur, R. (1980). Differential hemispheric lateralization of
positive and negative emotions. International Neuropsychological Society
Bulletin. March, 21.
Natale, M., & Hantas, M. (1982). Effect o f temporary mood states on
selective memory about the self. Journal of Personality and Social Psychology.
42, 927-934.
Perrig, W.J., & Perrig, P. (1988). Mood and memory: Mood-congruity
effects in absence of mood. Memory & Cognition. 16, 2, 102-109.
Radloff, L. S. (1977). The CES-D Scale: A self-report depression scale
for research in the general population. Applied Psychological Measurement. 3.
385-401.
Radloff, L.S., & Ten, L. (1986). Use of the Center for Epidemiological
Studies-Depression Scale with older adults. Clinical Gerontologist. 5. 1/2, 119-
135.
Rholes, W. S., Riskind, J. H., & Lane, J. W. (1987). Emotional states
and memory biases: Effects o f cognitive priming and mood. Journal of
Personality and Social Psvcholoav. 52. 91-99.
Richards, P. M., & Ruff, R. M. (1989). Motivational effects on
neuropsychological functioning: Comparison o f depressed versus nondepressed
individuals. Journal o f Consulting and Clinical Psychology. 57. 3, 396-402.
Riskind, J. H. (1989). Will the field ultimately need a more detailed
analysis o f mood-memory? In D. Kuiken (Ed.), Mood and memory: Theory,
research, and applications. [Special Issue]. Journal o f Social Behavior and
Personality. 4. 2, 1-21.
Riskind, J. H., Rholes, W. S., & Eggers, J. (1982). The Velten Mood-
Induction Procedure: Effects on mood and memory. Journal o f Consulting and
Clinical Psychology. 50. 146-147.
Sackheim, H.A., Freeman, J„ McElhiney, M., Coleman, E., Prudic, J.,
& Devanand, D.P. (1992). Effects of major depression on estimates o f
intelligence. Journal o f Clinical and Experimental Neuropsychology. 14, 2,
268-288.
71
Salovey, P., & Singer, J. A. (1985). The effects o f mood on the recall
o f childhood and recent autobiographical memories. Unpublished manuscript,
presented as "Mood and recall o f autobiographical memories: Failure to find
congruency" at the annual convention o f the American Psychological
Association, Los Angeles, California, August, 1985.
Silberg, J. L., Heath, A. C., Kessler, R., Neale, M. C., Meyer, J. M.,
Eaves, L. J., & Kendler, K. S. (1989, June). Genetic and environmental effects
on depressive symptoms in a general population twin sample. Paper presented
at meeting o f Behavioral Genetic Association, Charlottesville, VA.
Singer, J. A., & Salovey, P. (1988). Mood and memory: Evaluating the
network theory o f affect. Clinical Psychology Review. 8. 211-251.
Teasdale, J. D. & Bancroft, J. (1977). Manipulation o f thought content
as a determinant o f mood and corrugator electryomyographic activity in
depressed patients. Journal o f Abnormal Psychology. 86. 235-241.
Teasdale, J. D., & Russell, M. L. (1981). Differential effects o f induced
mood on the recall o f positive, negative, and neutral words. British Journal o f
Clinical Psychology. 22. 163-171.
Teasdale, J. D. & Spencer, P. (1984). Induced mood and estimates of
past success. British Journal o f Clinical Psychology. 23. 149-150.
Teasdale, J.D., & Taylor, R. (1981). Induced mood and accessibility of
memories: An effect o f mood state or o f induction procedure? British Journal
o f Clinical Psychology. 20, 39-48.
Teasdale, J. D., Taylor, R., & Fogarty, S. J. (1980). Effects o f induced
elation-depression on the accessibility o f memories o f happy and unhappy
experiences. Behavior. Research, and Therapy. 18. 339-346.
Velten, E. (1968). A laboratory task for induction o f mood states.
Behaviour Research and Therapy. 6. 473-482.
Vredenburg, K., Flett, G.L., & Krames, L. (1993). Analogue versus
clinical depression: A critical reappraisal. Psvcholoeical Bulletin. 113, 2, 327-
344.
Wechsler, D. (1981). WAIS-R Manual. The Psychological
Corporation, Harcourt Brace Jovanovich.
72
Wechsler, D. (1987). WMS-R Manual. The Psychological Corporation,
Harcourt Brace Jovanovich.
Appendix A
CES-D Subscale Items
73
Depressed Mood Subscale
3) I feel that I cannot shake the blues even with help from my family and
friends.
6) I feel depressed.
9) I think my life has been a failure.
10) I feel fearful.
14) I feel lonely.
17) I have crying spells.
18) I feel sad.
Psvchomotor Retardation Subscale
1) I am bothered by things that don’t usually bother me.
2) I do not feel like eating, my appetite is poor.
5) I have trouble keeping my mind on what I am doing.
7) I feel that everything I do is an effort.
11) My sleep is restless.
13) I talk less than usual.
20) I cannot get "going".
Lack of Well-Being Subscale
4) I feel that I am just as good as other people.
8) I feel hopeful about the future.
12) I am happy.
16) I enjoy life.
Interpersonal Isolation Subscale
15) People are unfriendly.
19) I feel that people dislike me.
* All items are answered by circling 1 to 4; l=not at all, 2=a little,
3=somewhat, 4=very much. Items on the Lack o f Well-Being Subscale are
reverse scored when totals are calculated.

Linked assets

University of Southern California Dissertations and Theses

Conceptually similar

PDF

Cognitive Assessment Of Reactance Using The Articulated Thoughts In Simulated Situations Paradigm

PDF

Articulated thoughts regarding cognitions toward older adults

PDF

Syntactic Priming As Applied To Syntactic Ambiguity Resolution

PDF

The effects of anxiety and age on attentional processes: Emotional Stroop and dot -probe in younger and older adults

PDF

The Emotional, Physiological, And Cognitive Reactions Of Boys And Girls From High-Conflict And Low-Conflict Homes To Simulated Marital Conflict

PDF

Depression as a risk factor for dementia

PDF

An examination of invariance using cognitive training data of older adults

PDF

Applying the theory of reasoned action to condom use: The effect of immediate consequences on intention-behavior consistency

PDF

History of depression, antidepressant treatment, and other psychiatric illness as risk factors for Alzheimer's disease in a twin sample

PDF

Smoking And Cognitive Functioning In Non-Demented Swedish Twins

PDF

An integrative analysis of racism and quality of life: A comparison of multidimensional moderators in an ethnically diverse sample

PDF

Expectancies for alternative behaviors predict drinking problems: A test of a cognitive reformulation of the matching law

PDF

Articulated thoughts about intentions to commit anti-gay hate crimes

PDF

The Effect Of Positive Affective Arousal On Crossed Categorization Effects

PDF

Rationalizing risk: sexual behavior of gay male couples

PDF

Children's immediate reactions to interparental conflict

PDF

The relationship between alcoholism and crime: autonomic and neurpsychological factors

PDF

A typology of maritally violent men and correlates of violence

PDF

Effects Of Different Treatment Procedures On Reading Ability And Anxiety Level In Children With Learning Difficulties

PDF

An investigation of the pseudohomophone effect: where and when it occurs

Asset Metadata

Creator Fox, Lauren Stephanie (author)

Core Title Mood Induction As An Analog To Depression In Older Adults: Aspects Of Depression And Cognitive Effects

Degree Master of Arts

Degree Program Psychology

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

Tag Gerontology,OAI-PMH Harvest,Psychology, clinical,psychology, cognitive,Psychology, experimental

Language English

Contributor Digitized by ProQuest (provenance)

Advisor Knight, Bob (committee chair), Davison, Gerald C. (committee member), Zelinski, Elizabeth M. (committee member)

Permanent Link (DOI) https://doi.org/10.25549/usctheses-c18-13875

Unique identifier UC11357432

Identifier 1379582.pdf (filename),usctheses-c18-13875 (legacy record id)

Legacy Identifier 1379582-0.pdf

Dmrecord 13875

Document Type Thesis

Rights Fox, Lauren Stephanie

Type texts

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

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

Repository Name University of Southern California Digital Library

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