University of Southern California Dissertations and Theses

A daily diary approach to compare the accuracy of depressed and nondepressed participants' estimation of positive and negative mood: A test of the depressive realism hypothesis

(USC Thesis Other)

A daily diary approach to compare the accuracy of depressed and nondepressed participants' estimation of positive and negative mood: A test of the depressive realism hypothesis

PDF

Download a page range

Download transcript

Copy asset link

Request this asset

Transcript (if available)

Content A DAILY DIARY APPROACH TO COMPARE THE ACCURACY OF
DEPRESSED AND NONDEPRESSED PARTICIPANTS’ ESTIMATION OF
POSITIVE AND NEGATIVE MOOD: A TEST OF THE DEPRESSIVE REALISM
HYPOTHESIS
by
Scott Jared Cypers
A Dissertation Presented to the
FACULTY OF THE GRADUATE SCHOOL
UNIVERSITY OF SOUTHERN CALIFORNIA
In Partial Fulfillment of the
Requirements for the Degree of
DOCTOR OF PHILOSOPHY
EDUCATION (COUNSELING PSYCHOLOGY)
August 2005
Copyright 2005 Scott Jared Cypers
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
UMI Number: 3196797
Copyright 2005 by
Cypers, Scott Jared
All rights reserved.
INFORMATION TO USERS
The quality of this reproduction is dependent upon the quality of the copy
submitted. Broken or indistinct print, colored or poor quality illustrations and
photographs, print bleed-through, substandard margins, and improper
alignment can adversely affect reproduction.
In the unlikely event that the author did not send 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.
®
UMI
UMI Microform 3196797
Copyright 2006 by ProQuest Information and Learning Company.
All rights reserved. This microform edition is protected against
unauthorized copying under Title 17, United States Code.
ProQuest Information and Learning Company
300 North Zeeb Road
P.O. Box 1346
Ann Arbor, Ml 48106-1346
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Dedication
This dissertation is dedicated with boundless love for my fiancee and family.
Not only have you helped me with school through the years with your acts of
kindness, support, love, and late night self-less reading and re-reading of my work,
you have provided me an understanding of the important areas of life: love, humor,
and balance. I love you all and want you to know that I would not be who I am or
where I am without you.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Acknowledgements
This dissertation is a culmination of the love and knowledge given to me
through the years by my fiancee, family, friends, teachers, and mentors. I want to
thank all of you for your love, kindness, and support throughout my education and I
want all of you to know that I am only here because of you.
I would additionally like to acknowledge and thank my advisor and the
members of my committee. Dr. Goodyear thank you for all your support and
mentorship through my years at USC. You have always been the ship to help me
through the EDCO storm. Dr. Hagedom, you have been nothing short of an angel for
me through my doctoral experience. I am indebted to you in so many ways and will
never be able to fully express my gratitude. Dr. Silverstein, through frantic phone
calls and late night e-mails you have shown nothing but support, helpfulness, and
flexibility. Thank you.
Lastly, I would like to acknowledge Dr. Almeida who helped make this
dissertation possible. For introducing me to the idea of a daily diary six years ago, to
providing me the data to accomplish my dissertation, and to helping me through the
process, I will never be able to say thank you enough. The University of Arizona
experience was truly “something unpredictable and in the end I found the time of my
life.”
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
IV
Table of Contents
Dedication....................................................................................................................ii
Acknowledgements.....................................................................................................iii
List of Tables............................................................................................................... v
List of Figures..............................................................................................................vii
Abstract.........................................................................................................................viii
Chapter 1. Introduction.................................................................................................1
Chapter 2. Literature Review.......................................................................................6
Chapter 3. Methodology.............................................................................................. 56
Chapter 4. Results......................................................................................................... 64
Chapter 5. Discussion...................................................................................................94
References.................................................................................................................... I l l
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
List of Tables
Table 1 Summary of Findings on Depressive Realism................................................. 42
Table 2 Level of Depression and Support for Depressive Realism..............................47
Table 3 Type of Study and Percentage of Support for Depressive
Realism...................................................................................................................48
Table 4 Chi-Square Demographic Differences between
Depressed and Non-Depressed Groups............................................................. 67
Table 5 One Way ANOVA Demographic Differences between Depressed
and Non- Depressed Groups................................................................................ 67
Table 6 Hierarchical Multiple Regression of Gender, Age, Marital Status, and
Diagnosis predicting Weekly Recall of Psychological Distress.....................68
Table 7 Description of Psychological Distress Variables by Depression Level
and Reference Period............................................................................................ 70
Table 8 Hierarchical Multiple Regression of Diagnosis, Peak Day Distress,
Final Day Distress predicting Weekly Recall of Psychological
Distress................................................................................................................... 77
Table 9 Multiple Regression of Peak Day Distress, Final Day Distress
Predicting Weekly Recall of Psychological Distress for Depressed
and Not-Depressed Groups...................................................................................78
Table 10 Hierarchical Multiple Regression of Diagnosis, Weekly Average
Distress Predicting Weekly Recall of Psychological Distress...................... 81
Table 11 Multiple Regression of Weekly Average Distress Predicting
Weekly Recall of Psychological Distress for Depressed and
Not-Depressed Groups........................................................................................82
Table 12 Hierarchical Multiple Regression of Diagnosis, Peak Day
Positive Affect, Final Day Positive Affect predicting Weekly
Recall of Positive Affect..................................................................................... 85
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
VI
Table 13 Multiple Regression of Peak Day Distress, Final Day Distress
Predicting Weekly Recall of Positive Affect for Depressed
and Not-Depressed Groups................................................................................86
Table 14 Hierarchical Multiple Regression of Diagnosis, Weekly Average
Positive Affect Predicting Weekly Recall of Positive Affect....................... 89
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
vii
List of Figures
Figure 1 Difference Between Weekly Psychological Distress and Recall
of Psychological Distress By Depression Status............................................72
Figure 2 Absolute Difference Between Weekly Psychological Distress and
Recall of Psychological Distress By Depression Status.................................72
Figure 3 Difference Between Weekly Positive Affect and Recall of
Positive Affect By Depression Status................................................................74
Figure 4 Absolute Difference Between Weekly Positive Affect and Recall
of Positive Affect By Depression Status.......................................................... 75
Figure 5 Interaction of End Day Distress By Depression Status.................................. 79
Figure 6 Interaction of Weekly Average Distress by Depression Status.................... 83
Figure 7 Interaction of End Day Positive Affect by Depression Status...................... 87
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
ABSTRACT
“Depressive Realism,” the notion that depressed individuals are more realistic
or accurate in their appraisals of the world, has many implications for the core tenets
of psychopathology in general, and more specifically to theories of depression, and
its treatment. This study utilizes a daily diary method to overcome some of the
existing barriers to the depressive realism research, the lack of ability to explore the
differential cognitive processes that could lead to differential reports of accuracy, the
lack of ecological validity of previous studies, the lack of testing the phenomenon on
a depressed sample, and the lack of testing the phenomenon on different populations.
Results failed to support depressive realism, as depressed individuals were less
accurate in their appraisal of their own recall of their psychological distress and
recall of their positive affect. Findings further demonstrated that the lack of accuracy
by depressed individuals may be due to the differential valuing of information within
the two types of cognitive processes explored, peak end and average valuing
cognitive processes.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Chapter 1: Introduction
“Depressive Realism” is the notion that depressed individuals are uniquely
able to understand, perceive, and appraise the world accurately. At the heart of this
notion lies the idea that depressed individuals are free from illusions of control and
the delusions of their importance in the world and therefore able to understand the
“truth” about the world in ways that non-depressed individuals cannot. It may be this
true insight is depressing the individual or it may be that depression itself provides
the key to viewing the world accurately.
Depressive realism is presumed to manifest itself in various ways, including
judgments of control or contingency, recall of self-perceived performance, self-
evaluation of performance, comparing oneself to another person, and to the
prediction of future events. For example, a depressive realist’s judgment of control in
a craps game would resemble the objective mathematical probability of how well
one can do in a craps game. A depressive realist’s recall of his or her performance
would consist of accurately perceiving the amount of positive or negative feedback
he or she received in a performance. A depressive realist’s self-evaluation of
performance would reflect the true nature of his or her performance as evaluated by
an outside observer or a standardized measure. A depressive realist’s comparison of
oneself to another person would accurately reflect the same view that an unbiased
observer or measure would also have. A depressive realist’s prediction of future
events will actually occur.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
2
This depressive realism hypothesis captivates researchers, theoreticians, and
practitioners alike, as it has implications for the core tenets of psychopathology in
general, for theories of depression etiology and relapse, and subsequent treatment of
depression (Dobson & Franche, 1989; Alloy & Abramson, 1988). Most theories of
psychopathology rest on the assumption that psychopathology stems from a
distortion of reality and that accurate reality testing maintains mental health (for a
review see Taylor & Brown, 1988). If depressed individuals were more accurate than
other people in their perceptions of reality, then it would be necessary to reformulate
this core tenet of psychopathology.
In regard to the more specific domain of depression, the validity of
depressive realism would undermine the foundations of one of the major cognitive
theories of depression, that of Aaron Beck’s Cognitive Theory (Hollon & Beck,
2000; Beck, Rush, Shaw, & Emory, 1979; Beck, 1967). This theory builds on the
notion that depression is caused or maintained by misperceptions, misinterpretation,
or dysfunctional interpretations of situations (Beck & Weishaar, 2000). But, whereas
Beck’s theory has garnered strong empirical support for most of its tenets (Williams,
Watts, MacLeod, & Mathews, 1988; Coyne & Gotlib, 1983), one that has eluded
empirical support is the inaccuracy of depressive thinking (for a detailed review see
Haaga, Dyck, & Ernst 1991). This further raises the possibility that depressive
realism may exist. If so, that would demand a reformulation of Beck’s cognitive
theory.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
3
Depressive realism may hold one of the keys to providing a marker for the
etiology of depression as well as explain the high rates of relapse for depression.
Whereas no study has attempted to explore this link, the current depression literature
is replete with studies exploring trait-like characteristics that extend beyond the
depression state (Segrin, 2000; Teasdale, Williams, Soulsby, Segal, Ridgeway, &
Lau, 2000; Illardi, Craighead, & Evans 1997; Hedlund & Rude, 1995; Rohde,
Lewinsohn and Seely, 1990; Segrin, 2000). Depressive realism theory postulates that
the ability to see the world more accurately may be one of the markers that both
cause depression as well as predispose depressed individuals to relapse. This further
justifies the continued exploration of the depressive realism theory.
Lastly, the notion of depressive realism challenges the basic goals of
cognitive therapies for depression, which are “to correct faulty information
processing and to help patients modify assumptions that maintain maladaptive
behaviors and emotions” (Beck & Weishaar, 2000, p. 254). Cognitive therapy has
been remarkably effective in the treatment of depression in the short term (Dobson,
1989) and in long-term follow-ups (Elkin, Shea, Watkins, Imber, Sotsky, Collins,
Glass, Pilkonis, Leber, Docherty, Fiester, & Perloff, 1989). So it clearly works, but
perhaps for the wrong reason. It may be that cognitive therapy actually teaches
depressed individuals methods to disengage from accurate reality testing. Thus, its
effectiveness may lie in its ability to teach depressed people how to distort reality in
positive ways.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
4
Because the construct of depressive realism has such strong and far-reaching
implications, researchers continue to explore its validity. Haaga and Beck (1995)
noted, for example, that the first study that indicated the possibility of depressive
realism, that of Alloy and Abramson’s 1979 sadder but wiser experiments, “is cited
30 times a year, more than a decade after publication, in journals covered by the
Social Sciences Citation Index (page 42).” Major reviews have been conducted to
integrate the findings and explore the veracity of the depressive realism claim
(Ackermann & DeRubeis, 1991; Dobson & Franche, 1989).
Dobson and Franche (1989) reviewed 45 studies on depressive realism and
concluded that “depressive realism does exist as a phenomenon, although claims
about its strength and pervasiveness should be reexamined with caution. Future
research needs to incorporate a strong emphasis on high external validity, as
reflected by the use of clinically depressed samples, and high ecological validity,
through the employment of personally meaningful stimuli and experimental
pargadigms (p 430).” Ackerman and DeRubeis (1991) discounted some of the
studies utilized by Dobson and Franche for failing to have an objective measure of
reality and thus their review contained only 34 studies. They found mixed results for
the depressive realism phenomenon and concluded too that more research needs to
be conducted with depressed samples and that have more ecological validity.
Since the last major review, an additional 18 studies have been conducted
that have explored the depressive realism phenomenon. Each of these studies
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
5
employed the criteria as Ackerman and DeRubeis (1991) suggested as necessary to
test depressive realism: a verifiable, objective standard of reality and a measure of
the participant’s depression. As these studies have come out since the last major
review of depressive realism, an update is necessary to determine if the previous
reviews’ recommendations for future research have been met and to determine if any
of these subsequent studies have been able to explain the current mixed results in the
depressive realism literature. Finally, this review will specifically look for any
studies that directly explore the cognitive problem solving approaches utilized by
depressed and non-depressed groups. This review will focus on these studies as it is
this studies speculation that differential cognitive problem solving approaches to the
same paradigm lead to differential rates of accuracy.
The literature that follows will explore the current state of the depressive
realism research in order to determine the different ways the effect has been found,
to determine the current limitations in the research, as well as to verify that these
studies failed to explore the cognitive processes that lead to differential results within
the studies. Then, the study will provide a novel approach to reconcile many of the
noted limitations and demonstrate that depressed individuals and non-depressed
individuals engage in different cognitive processes within the same paradigm. This
study speculates that it is the differential processes that each group engages in within
each study that leads to differential rates of accuracy. In essence, this study wants to
answer the questions: Does depressive realism exist? If so, why?
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
6
Chapter 2: Literature Review
A Review o f the Depressive Realism Research
To be included in the analysis, each study had to meet two criteria. First, it
had to have assessed for participant depression. This is a necessary criterion, as
knowledge of depression level is vital to determine whether depressed individuals
are accurate in their perceptions of the world. Second, it had to include an objective
assessment of reality with which to compare depressed individuals or groups. This
objective assessment is vital since it allows for a test of accuracy, an absence of
which would undermine the ability to determine for which group a distortion or
perception bias lies.
Studies were identified through three means: Psych Info, the premier
psychological database, past reviews (Ackerman and DeRubeis, 1991; Dobson &
Franche, 1989; Alloy & Abramson, 1988), and the literature reviews of the obtained
articles. Studies from Psych Info were obtained by using different combinations of
the following key terms: depression, accuracy, depressive realism, mood, judgment,
control, self-evaluation, recall, and cognitive distortion. Past reviews provided the
foundation as they included studies that met the necessary criteria for this review and
were found in the past to have important results related to the phenomenon. Lastly,
the introduction and references of every obtained article was examined to ensure that
all studies that examined the phenomenon were obtained.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
7
Overall, 74 studies or parts of studies met the minimum criteria for inclusion
in this review. These studies are sorted by degree of depression (mild versus severe),
and then sub-divided by the task demands employed in each experiment as follows: a
perceived judgment of control, a recall of self-perceived performance, a self-
evaluation, a self-other social comparison, or a prediction of future events.
This review classifies studies by degree of depression because differences
consistently have been found between mildly depressed college students and
clinically depressed populations (for a review see Kendall, Hollon, Beck, Hammen,
& Ingram, 1987; Coyne & Gotlib, 1983). This is particularly important because
studies that employ college samples necessarily lack a non-depressed, clinical
control sample, a design feature that allows researchers to understand whether their
findings are due to the effects of depression or the nature of psychopathology in
general. Most studies that use clinical samples therefore contain this added control
group dimension. Because it is typical for these studies to employ more rigorous
criteria for a diagnosis of depression, they are more likely to have a truly depressed
group rather than a mildly dysthymic group. Thus, studies that employ clinical
samples may be a stronger test of depressive realism than studies done with college
populations that are only mildly depressed.
Most studies of depressive realism utilize one or more of three common
measures to ascertain depression level: the Beck Depression Inventory (BDI), the
Hamilton Rating Scale for Depression (HRSD), or the Center for Epidemiological
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Studies - Depression Scale (CES-D). Studies utilizing mildly depressed college
students typically use the bare minimum cut-offs to categorize participants as in the
depressed range. For example, most studies of mildly depressed individuals utilize a
BDI score of 12 or above to classify the depressed group and 4 or below to classify
the non-depressed group. According to Groth-Marnat (2003), the following criteria
can be used as a heuristic to use the BDI as cutoffs for depression: 0 to 13 (no or
minimal), 14 to 19 (mild), 20 to 28 (moderate), and 29 to 63 (severe). Studies that
utilize a clinical sample tend to use the same measures yet depressed subjects usually
are required to be two standard deviations above the mean on these measures. As
well, studies that utilize a depressed sample may include a structured clinical
interview to determine depression status.
Sub-grouping studies by type of task makes it possible to explore the
implications that the task and setting can have on participants’ ability to accurately
perceive reality. The underlying assumption is that the type of task demands different
problem solving methods and skills in order to successfully negotiate the task and to
be accurate. It may be that the demands of certain types of tasks promote conditions
of accuracy for depressed individuals and at the same time promote inaccuracy for
non-depressed individuals. As well, previous reviews (Ackerman and DeRubeis,
1991; Dobson & Franche, 1989; Alloy and Abramson, 1988) have noted significant
differences based on these divisions.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
9
Mildly Depressed/Dysthymic Group
Perceived Judgment of Control
Studies were categorized as one employing “perceived judgment of control
tasks” if they required participants to evaluate their level of actual control over the
outcome of a witnessed event or task. As the researchers manipulated the actual
amount of control that the subject had, all contingency studies contained a true
measure of the amount of control. Thus, they are strong avenues to explore
depressive realism and have yielded some of the strongest support for depressive
realism in past reviews (Ackerman & DeRubeis, 1991, Dobson & Franche, 1989).
Alloy and Abramson’s “sadder but wiser” experiments (1979) are the classic
example of this paradigm and the ones which began the depressive realism debate. In
this series of studies, Alloy and Abramson brought students into a lab and had them
rate the degree of contingency between their responses (pressing or not pressing a
button) and the environmental outcome (onset of a green light). The experimenters
controlled the actual degree of contingency by manipulating the percentage of trials
on which the outcome of interest (green light onset) occurred when the subject
pressed the button (indicator one) and the percentage of trials on which green light
onset occurred when the subject did not press the button (indicator two).
In contingent conditions, where the subjects had 25% control (75-50), 50%
control (75-25), and 75% control (75-0), it was found that both depressed and non-
depressed subjects correctly estimated their actual degree of control. In non-
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
10
contingent conditions (the subject had no control) with low (25-25) and high (75-75)
green light onset conditions, it was found that depressed individuals were accurate in
their actual degree of control in both the high and low conditions, while non-
depressed individuals only demonstrated accuracy in the low condition.
In the high condition, non-depressed individuals incorrectly believed they
had more control than they did. These results were replicated even with the
introduction of a valence reward of winning or losing money (experiment 3).
Furthermore, the reintroduction of contingency (subjects had 50% control) with
valence (winning or losing money) had the further effect of causing non-depressed
individuals to underestimate their amount of control in the losing money condition
while depressed individuals still maintained a high level of accuracy in both
conditions.
These findings have been replicated numerous times and in different types of
learning experiments (Msetfi, Murphy, Simpson, & Kombrot, 2005 (experiments 1
and 2); Pacini, Muir, & Epstien, 1998 (under trivial conditions); Rosenfarb, Burker,
Morris, & Cush, 1993 (changing contingencies during experiment); Alloy &
Clements, 1992; Mikulincer, Gerber, & Weisenberg, 1990 (high threat/no mirror
condition, low threat/no mirror conditions); Vazquez, 1987 (experiment 1 and 2 with
a Spanish sample); Benassi & Mahler, 1985 (no-other person present condition);
Martin, Abramson, & Alloy, 1984 (own-control condition); Alloy & Abramson,
1982; Abramson, Alloy, & Rosoff, 1981 (complex solution offered); Alloy,
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
11
Abramson, & Viscusi, 1981 (with a non-depressed group who was induced into a
depressive mood); Golin, Terrell, Johnson, 1977 (own throw condition)) and have
failed to replicate any aspect of depressive realism in four studies (Pusch, Dobson,
Ardo, & Murphy, 1998, experiment #2; Koeing, Clements, & Alloy, 1992; Ford &
Neale, 1985; Bryson, Doan, & Pasquali, 1984). Failure to replicate these findings
may be due to the added use of a computer that could increase one’s belief in control
(Koeing, Clements, & Alloy, 1992), the lack of a truly depressed group (Ford &
Neale, 1985), or the group beliefs of the sample, Canadian vs American (Bryson,
Doan, & Pasquali, 1984), and differences in personality styles (Pusch, Dobson, Ardo,
& Murphy, 1998).
Besides replicating or failing to coincide with Alloy and Abramson’s
findings, this research extends Alloy and Abramson’s work by providing limitations
on depressed individuals ability to accurately judge contingency. These studies seem
to reveal that depressed individuals degree of accuracy diminishes under the
following conditions: the monetary consequences of choices are very high (Pacini,
Muir, Epstein, 1988, consequential conditions group), another person is present in
the room while the experiment is conducted (Benassi & Mahler, 1985, experiments 1
and 2), the solution to produce the light is highly complex and the experimenter
provides no clues as to the solution (Abramson, Alloy, & Rosoff, 1981), the task is
conducted in public (Koeing, Clements, & Alloy, 1992), the reinforcement
negatively coincides with the individuals’ self-schema (Vazquez, 1987), the amount
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
12
of control is in reference to another person conducting the exact experiment (Martin,
Abramson, & Alloy, 1984; Golin, Terrell, & Johnson, 1977), the presence of a mirror
in an uncontrollable task (Mikuliner et. ah, 1990), the addition of a computer to
increase the skill of the task (Koeing, Clements & Alloy, 1992), the length of time
between stimulus presentation (Msetfi. et ah, 2005), the personality characteristics of
the depressed individual (Pusch, Dobson, Ardo, & Murphy, 1998, experiment #2;
Ford & Neale, 1985), and the perception of the task as being a skill task (Golin,
Terrell, & Johnson, 1977).
Two studies to date have examined judgment of control outside the confines
of the laboratory. Glass, McKnight, and Valdimarsdottir (1993) had depressed/
burned-out nurses and non-depressed/non-bumed-out nurses rate their perceived
amount of job control. Compared to a criterion of actual job control, they found that
non-depressed/non-bumed out subjects overestimated their degree of control,
whereas depressed/burned out subjects approached complete agreement. McKnight
and Glass (1995) later replicated this study using a two-year follow-up of 100 nurses
and obtained similar results.
Overall, these contingency studies seem to reinforce the notion of a limited
depressive realism. Mildly depressed individuals are more accurate in their
perceptions of contingency under certain conditions than non-depressed individuals.
Despite these findings, changing the demand characteristics by introducing
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
13
additional elements to the experiment can cause depressed individuals to become
more inaccurate in their appraisal of their degree of control.
A possible explanation for the variation in findings may be that altering the
demands of the experiment changes the conditions significantly enough that
depressed or non-depressed individuals change their problem solving approach. The
different approaches or understanding of the goal would lead to the activation of
different automated procedures to solve the problem accurately. Thus both depressed
and non-depressed participants may have the ability to be accurate in their appraisals
of the degree of control, but they need to recognize the goal of the experiment and
understand the necessary conditions to match the demand characteristics to the
correct automated procedure.
Further evidence for this notion comes from Msetfi et. al.’s (2005) study that
showed that changing the interval length affected non-depressed individuals
accuracy in the judgment of contingency yet had no affect on depressed individuals.
They argue that the reason for their finding the depressive realism effect is that
depressed individuals did not pay attention to contextual processing in their response
pattern and so accuracy was a byproduct of inattention.
Recall of Self-Perceived Performance
Recall of self-perceived performance studies involve the subject running
through a given experiment in which the subject fails or succeeds on a given number
of trials. After the experiment, subjects had to rate the amount of times they were
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
14
correct or incorrect and the amount of time they received positive or negative
feedback. As there exists a predetermined amount of positive and negative feedback,
recall o f self-perceived performance studies provide another avenue to explore
accuracy.
Kennedy and Craighead’s (1988) maze task is a typical recall of self
perceived performance paradigm. The researchers directed subjects to attempt to
learn the correct path through a mental maze consisting of 15 choice points, with
four alternatives per choice, represented by four buttons on a panel in front of the
subject. At each choice point, a light corresponding to correct, incorrect, or neutral
would inform them of accuracy. The subjects had to respond correctly at each choice
point to advance to the next choice point. Subjects were divided into two conditions,
one in which each person received social and monetary rewards for their correct,
incorrect, or neutral responses, or one in which the subject received information
only. In the first experiment, the subjects had to go through two mazes and at the end
rate the amount of feedback they received during the complete experiment. In the
second experiment, subjects had to go through only one extended maze (twelve
choice points) but had to record the amount o f positive and negative feedback they
perceived after each pass through a point and then rate a global assessment of the
amount of feedback they received at the end o f the experiment.
In the first study, all three groups (depressed/anxious, non-depressed/non-
anxious, and non-depressed/anxious) significantly underestimated the amount of
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
15
positive feedback they received but only the non-depressed/anxious group
significantly underestimated the amount of negative feedback they received under all
conditions. The depressed/anxious and non-depressed/non-anxious were accurate in
the amount of negative feedback they received.
Results from the second study indicate that all three groups performed with
equal accuracy on re-calling amounts of positive and negative feedback during the
first nine passes, but differed in the last three passes where extreme differential rates
of feedback existed. In the last three passes, the non-depressed/anxious and the non-
depressed/non-anxious groups were accurate in their perception of positive feedback,
while the depressed/anxious group continued to underestimate the amount of positive
feedback. In the negative feedback conditions, all three groups were accurate in the
amount of negative feedback they had received.
Using different experimental tasks but following the same paradigm, findings
have been replicated that demonstrate that depressed individuals underestimate the
amount of positive feedback they receive but are accurate in the amount of negative
feedback they receive (Horowitz, Abramson & Usher, 1991; Abramson, Alloy, &
Hartlage, 1989; Dobson & Shaw, 1981 (first 40 trials); Dykman, Dykman,
Buchwald, 1977; Nelson & Craighead, 1977 (experiment 1)). The work of Dykman,
Abramson, et al. (1989) and Dykman, Horowitz, et al. (1991) further reveal that the
salience of the information to the depressed person’s schema may affect his or her
ability to recall the performance.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
16
Exceptions to these findings were those of Dykman & Volpicelli (1983),
Craighead, Hickey, and DeMonbreun (1979); and Kuiper (1978). These differences
may be due to the amount of reinforcement (Kuiper, 1978) and to the difficulty of
the task (Craighead, Hickey, and DeMonbreun, 1979, Dykman & Volpicelli, 1983).
Kuiper (1978) had subjects engage in a word association task where the
experimenter controlled the design such that the subjects either failed (20% correct),
succeeded (80% correct), or obtained a score that was neither clearly indicative of
success nor failure (55% correct). After the task, subjects had to estimate the amount
of reinforcement they received. He found that depressed subjects were accurate in
their assessment of the amount of reinforcement they received in the 20% condition,
but underestimated the amount of reinforcement they received in the 55% and 80%
reinforcement conditions.
Craighead, Hickey, and DeMonbreun (1979) presented subjects with a slide
of three syllables shot out of focus. This was done quickly (1/2 second). They then
presented the subjects with a focused slide that contained four options, one of which
was the previous slide. From the four options, subjects had to identify the words just
presented. After each choice, subjects were presented with a slide whose color
denoted the degree of accuracy, which could either be accurate, inaccurate, or
neutral. At the conclusion of the experiment, subjects were asked to rate how many
times they received positive, negative, or neutral feedback on the task. The
researchers found that there were no differences between depressed/anxious, non
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
17
depressed/anxious, and non-depressed/non-anxious in terms of the accuracy of
recalling the amounts of positive, negative, and neutral feedback they had received.
Dykman and Volpicelli (1983) introduced the variable of task difficulty to
examine recall of feedback differences between depressed and non-depressed
individuals. Subjects had to estimate the number of dots on a screen (ranging from
40 to 100 dots) presented to them for a half a second. Subjects went through 40
trials and received random valenced feedback (subjects were trained to believe this
corresponded with good, average, or poor) or ambiguous feedback (subjects had no
knowledge of correspondence) After each trial, subjects rated the level of feedback
they received and then made a global rating at the end of the experiment of the total
amount of feedback they received. Researchers found that in short-term trial
feedback, non-depressed individuals perceived feedback improving over trials where
depressed individuals correctly inferred the randomness of the feedback. In terms of
long-term recall, the researchers found no differences between depressed and non-
depressed individuals in their accuracy of long-term recall of evaluative feedback.
Besides findings that coincide or fail to coincide with Kennedy and
Craighead’s study, recall of evaluation studies have provided limits on depressed
individuals’ ability to accurately recall evaluative information. Five factors have
emerged which significantly affect depressed and non-depressed individuals ability
to accurately recall evaluative feedback: task difficulty (Dykman & Volpicelli, 1983;
Craighead, Hickey, and DeMonbreum, 1979), the length in time of recall of the
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
18
evaluative information (Dykman & Volpicelli, 1983), the amount of reinforcement
received (Kuiper, 1978), and the relevance of the information to one’s own schema
(Dykman, Horowitz, Abramson & Usher, 1991; Dykman, Abramson, Alloy, and
Hartlage, 1989).
Overall, these studies suggest that mildly depressed individuals may be
accurate in their perception of negative feedback, but underestimate the amount of
positive feedback they receive in their recall of evaluative information. These overall
findings may also be bounded, depending on the task difficulty, the length in time of
recall of the evaluative information, amount of reinforcement the person receives,
and the relevance o f the information to one’s own schema.
A possible cognitive explanation for the variation in task performance may be
that different demand characteristics of the experiments caused shifts in the problem
solving approach of the participant. These shifts directed the participant to either
accurately or inaccurately encode and recall information from the experiment. The
demand characteristics that seemed to optimally produce the depressive realism
effect were when the amount of reinforcement was low, the task difficulty was high,
and the information was schema relevant.
Self-Evaluation
Self-evaluation studies involve running participants through a task and
having them evaluate their performance on it. This evaluation is then compared to
either a base-line or an outside rater’s evaluation of the subjects’ performance. Self
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
19
evaluation studies to explore depressive realism have been the most heavily
criticized studies of depressive realism as it is argued that in some of the instances
there is no objective standard which to compare for a judge of accuracy (Ackerman
and DeRubeis, 1991, Dobson & Franche, 1989). As such, these studies should be
interpreted with caution as being strong tests of the depressive realism hypothesis.
Gotlib and Metlzer’s (1987) social interaction paradigm is one of the core
methods used to examine accuracy in self-evaluation. The researchers videotaped
depressed and non-depressed female students interactions with one another in a 15
minute get to know you conversation. After the interaction, both depressed and non-
depressed subjects rated their social competence in the interaction with either the
depressed or non-depressed individual. These interactions were later reviewed and
each subject was rated on the same attributes by an independent observer. In terms of
self-evaluation, it was found that depressed individuals significantly rated
themselves lower than non-depressed individuals and closer to the independent
observer. Although this is true, both groups over-estimated their amount of social
competence compared to the trained observer ratings. As well, non-depressed
individuals interacting with a depressed individual rated themselves as being less
socially competent after the interaction with depressed individuals.
Other social interaction paradigms have replicated the findings that depressed
individuals were more aligned with independent observers o f their performance
whereas non-depressed individuals believed that they performed better than the
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
20
observers rated (Kistner, Balthazor, Risi, & David, 2001; Edison & Adams, 1992;
Altmann & Gotlib, 1988; McNamara & Hakcett, 1986). Edison and Adams (1992)
extended the findings by increasing the number of possible interactions to include
depressed people interacting with depressed people and non-depressed people
interacting with non-depressed people. McNamara and Hackett (1986) extended the
findings by employing a group paradigm rather than a dyad paradigm. Altmann and
Gotlib (1988) extended the findings to another sample and an ecologically valid
context, children during free play at school. Kistner, Balthazor, Risi, and David
(2001) extended the finding to adolescent students, an ecologically valid context, and
to using the students of the class as the objective raters rather than outside observers.
Two social interaction studies failed to replicate the findings (Pusch, Dobson,
Ardo, & Murphy, 1998, experiment 1; Dykman, Horowitz, Abramson, & Usher,
1991). Both studies found that depressive participants were less accurate than non-
depressed individuals in their self-evaluations than rated by either objective
evaluators or by their peers. Dykman, Horowitz Abramson, and Usher (1991) found
that depressed and non-depressed participants evaluations of their performance
differed significantly from the objective raters in that depressed students rated
themselves as worse than the objective raters. One possible reason for the difference
is that in this study the raters were in the room during the experiment whereas in the
other studies the raters made their evaluations out of the room behind a mirror or
after through the use of videotape.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
21
Pusch, Dobson, Ardo, and Murphy (1998, experiment 1) found that after the
dyadic interaction, non-depressed individuals’ estimates of their friends ratings of
them were not significantly different from the actual ratings, but depressed
individuals’ estimates of their friend’s evaluation were significantly lower than the
actual evaluations. There are a number of possible reasons for the difference. It may
be that the friends of depressives did not want to be harsh on their friends so they
rated the interaction as better than was actually the case. Or, the salience of the
friend’s evaluation to the depressed person’s own schema may have caused the
depressed individual to view himself more harshly than was warranted in the
interaction.
Cane and Gotlib’s (1985) self-evaluation study differed from the above
paradigm in that participants engaged in a task rather than a social interaction. The
researchers had depressed and non-depressed college students respond to five audio
taped descriptions of problematic situations and rate their performance knowing that
an observer would be rating their responses as well. After completing the task,
participants received random favorable, unfavorable, or no feedback and completed a
second task. The researchers found that in the first task, non-depressed participants
rated their performance equal to the independent raters whereas depressed
individuals rated their performance less favorably on the task then the raters.
Additionally, the type of feedback received after the first experiment did not increase
depressed students accuracy in evaluations of performance on the next task.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
22
Two studies, Heiby (1981) and Keller, Lipkus, and Rimer (2002) represent
another set of self-evaluation studies that allowed for an examination of accuracy.
These studies were different in this category that they had the participant engage in
some sort of project to self evaluate his or her performance where there existed an
objective performance indicator rather than a rater. Heiby (1981) had depressed and
non-depressed subjects fill out a 15 item analogies test. After each analogy, subjects
rated their accuracy and the degree of self-praise for their performance. Heiby found
that depressed individuals were only accurate in their administration of self-praise if
they believed that they performed poorly on the task. If depressed individuals
believed they performed well, then self-praise was not correlated with accuracy.
Keller, Lipkus and Rimer (2002) had depressed and non-depressed women ages 40
to 60 rate their risk of getting breast cancer and then provided accurate risk factor
estimates based on personal information provided. They found that depressed women
more accurately changed their information from their initial estimates to coincide
with the tailored risk factor estimate whereas non-depressed women maintained their
previous estimates. Interestingly, this meant that the depressed women had to lower
their risk estimate or be more optimistic in order to be accurate. In a second study
with college students, the researchers explored the additional role that mood after
initial feedback had on the follow-up risk estimate to understand the process that
may lead to depressive realism. The researchers were able to replicate the results
such that depressed individuals more accurately changed their risk estimates to
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
23
reflect the provided information as well as found that less positive mood may be the
factor that leads to more accurate assessment.
The mixed findings for depressive subjects accuracy were replicated and
extended in two studies (Stone, Dodrill, & Johnson, 2001; Nelson & Craighead,
1977, experiment 2). Nelson and Craighead (1977, experiment 2) extended the
findings to include differential rates of positive, negative, or neutral reinforcement.
They found that depressed subjects underestimated the amount of accurate feedback
in the high and low feedback conditions, while non-depressed subjects overestimated
the amount of praise in the low feedback condition and underestimated the amount of
praise in the high feedback condition. Stone, Dodrill, and Johnson (2001) extended
the findings to differences between item reports and global evaluation reports. They
found that depressed individuals were more accurate than non-depressed individuals
in their accuracy judgments of individual item reports, however, in the aggregate
judgment, non-depressed students were more accurate in their appraisals while the
depressed students underestimated the amount of accuracy.
Two studies using similar paradigms failed to replicate the mixed findings in
terms of depressive accuracy (Beyer, 2002; Vestre & Caufield, 1986). In total
support of depressive accuracy, Vestre and Caufield (1986) using 15 objectively
rated neutral personality feedback statements found that depressed people accurately
rated the global personality impression as being more neutral, while non-depressed
individuals rated the profile as more positive than reality. However, Beyer (2002)
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
24
could not find any support for depressive accuracy in depressives self-evaluations of
how well they performed on a 30 item gender based multiple choice test which
included questions either about Math (masculine gender based), English (feminine
gender based), or history and geography (neutral based).
Taken together, self-evaluation studies seem to point to a context based
depressive realism in which different factors can influence depressed individuals’
ability for accuracy. The differential conditions allow mildly depressed individuals to
be more accurate in their perceptions of themselves than non-depressed individuals
whereas other conditions foster non-depressed individuals to be more accurate than
depressed individuals. These conditions include the type of self-evaluation task (as
evidenced by the different findings within each type of self-evaluation task), the type
of self-evaluation appraisal within the task (Stone, Dodrill, & Johnson, 2001), the
presence of the evaluator in the room (Dykman, Horowitz, Abramson, & Usher,
1991), the salience of the self-evaluation to one’s identity (Beyer, 2002; Pusch,
Dobson, Ardo, & Murphy, 1998, experiment 1), the number and type of rater
(Kistner, Balthazor, Risi, & David, 2001; Pusch, Dobson, Ardo, & Murphy, 1998,
experiment 1), the mood of the person and type of information (Keller, Tipkus, &
Rimer, 2002), and the type of reinforcement (Nelson & Craighead, 1977, experiment
2).
These studies further reinforce the notion that depressed and non-depressed
individuals pay attention to different contexts or demands within the different
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
25
paradigms. It continues to suggest that the cognitive process that depressed and non-
depressed individuals engage in may be what leads them to differential rates of
accuracy.
Self/Other Social Comparison Studies
Studies were categorized as using self/other social comparison studies if
they demanded that the individual compare himself/herself to another group on
performance, personality, or ability dimensions. Nine of these types of studies were
found.
Tabachink, Crocker, and Alloy (1983) represent one group of self/other
social comparison studies. These researchers had participants come into a lab and fill
out a questionnaire that had each subject rate each item whether it was true of
themselves, whether it was true of the typical college student, the percentage of
college students each item characterizes, and the percentage of college students who
would say that each item describes them. The researchers found that in terms of
accuracy compared to others who had filled out the questionnaire in the past,
depressed individuals were less accurate in their perception of the number of student
who would endorse each item. Although depressed individuals were less accurate in
their perception of how college students’ really saw themselves, depressed people
were more aware of students’ ways to try to present themselves in a positive light.
This knowledge may have led them to their inaccuracy of reporting.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
2 6
But the findings of this study are unique as no self-other social comparison
studies using this paradigm obtained similar results (Crocker, Alloy, & Kayne, 1988;
Ahrens, Zeiss, & Kanfer, 1988; Cambell, 1986). Rather, these studies supported
depressive realism in that non-depressed individuals rated more positive events and
less negative events happening to themselves than to comparison others, while
depressed individuals rated the likelihood of positive and negative events happening
to themselves as equal to comparison others.
Cambell (1986) extended these findings using a similar paradigm but added
measures of opinion and ability. She found that these added dimensions changed the
degree of accuracy in subjects’ reports, such that high opinion questions yielded the
most accuracy where high ability questions yielded the least accuracy for both
depressed and non-depressed students.
Although these studies are reported as they contain some degree of objective
assessment (as the measures were pre-validated on other students to get at the typical
college student), these are the least objective measurements of all the depressive
realism studies. Depressed and non-depressed individuals may not have accurate
knowledge of the entire campus, just knowledge o f their own friends in their own
world. As this is the case, the typical college student would be a different reference
point for each person. Thus, accuracy with these studies can be an illusive construct.
The following self/other studies contain stronger findings as they contain more
objective criterion or available other comparison information.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
27
Albright and Henderson’s (1995) study represent a group of studies that were
similar to Tabachink, Crocker, and Alloy (1983), but used the person’s self
evaluation or objective information about the comparison other as the tool to make
the comparison. This slight change allows for a more objective comparison to be
made. The researchers had depressed and non-depressed participants fill out a
personality profile in which they rated their personality on a number of dimensions.
On a separate day, subjects came back into the laboratory and were given a
personality profile that was either (-2 points, +2 points, or the same) on every 8 point
personality dimension. Participants were then given a series of life events and asked
to compare how they would respond to the life event compared to the comparison
other. In regards to comparison other ratings and depressive realism, it was found
that non-depressed individuals ratings for the comparison other were more accurate,
whereas depressed individuals rated the comparison other more favorably than was
merited by the personality profile.
Although using similar paradigms, other studies found different results that
either partially supported depressive realism (Albright, Alloy, Barch, & Dykman,
1993, experiment #2) or fully supported it (Alloy & Ahrens, 1987). Albright, Alloy,
Barch, and Dykman (1993, experiment #2) conducted a similar experiment but had
the personality profile of the comparison other being equal to the profile that the
depressed and non-depressed student had filled out. Participants were given another
type of personality profile and asked to rate themselves and the comparison other on
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
28
the same dimensions. They found that both depressed and non-depressed students
displayed accurate and inaccurate evaluations depending on the relevance of the
personality attribute to the student’s self- schema. Overall, depressed students were
accurate in their attributions for negative and neutral personality attributes, but
underestimated their accuracy for positive attributes. Non-depressed students were
accurate in the appraisal of neutral attributes, but viewed the comparison other as
worse then them on negative attributes and as better then the comparison other on
positive attributes.
Alloy and Ahrens (1987) asked depressed and non-depressed students to
decide whether a hypothetical student would make the honor role or be placed on
academic probation given statistically relevant information about the person’s SAT
score, percent of time healthy, hours spent studying a week, and ease of course
schedule. The researchers found that non-depressed students rated the hypothetical
student more favorably than was warranted by the probabilities and that depressed
students rated the hypothetical student more accurately to the probability of the
situation.
Strack and Coyne’s (1983) study represents one last type of self/pother social
comparison study. These researchers used the dyad paradigm discussed in the self-
evaluation condition but this time had subjects fill out information not about
themselves, but about their perceptions of the social skill of the other person. As
well, participants were told that they were going to share the information with their
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
29
partner or the information would be kept confidential. Researchers found that
depressed individuals correctly anticipated more rejection from the subjects with
whom they interacted than non-depressed individuals. Non-depressed individuals
perceived themselves as being accepted even when this was not the case.
Using the exact same paradigm but adding a measure of anxiety to create
three groups instead of two (depressed/anxious, non-depressed/anxious, and non-
depressed/non-anxious groups), Dobson (1989a) found the reverse finding in terms
of accuracy. Depressed/anxious and non-depressed/anxious groups still perceived
themselves to be rejected more often by their partner, but the partners’ perception of
these groups were more accepting. All three groups had equally high scores of social
skill.
In summary, depressive realism receives partial support in the different types
of self/other studies with mildly depressed people. Whereas some studies show that
mildly depressed individuals are more accurate in their ratings of another person in a
group task, other studies reveal that mildly depressed individuals engage in both
positive and negative distortions of reality just as non-depressed individuals. This
partial support lends credence to the notion of a contingent depressive realism, in
which the type of self-evaluation task, the salience of the task to one’s identity
(Albright, Alloy, Barch, & Dykman 1993, experiment #2; Campbell, 1986) and other
pertinent characteristics foster cognitive processes which lead depressed or non-
depressed individuals to be more accurate.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
30
Part of the distortion in accuracy may be due to the depressed individual’s
lack of general knowledge into the comparison other such that the depressed
individuals ratings of the comparison other are based on limited comparison
information. Another source that may contribute to the ability for depressed
individuals to be accurate is the salience of the study to the subject’s identity
(Campbell, 1986).
Prediction of Future Events
Studies fell under this rubric if they demanded that the participants make
some prediction of likelihood of a future event occurring to them. The basic
paradigm for these studies is one in which individuals are to rate whether events that
are either typical or atypical of events experienced by that group o f people (college
students) will occur for that specific individual. These predictive statements fell
within the following categories: interpersonal relations, psychological functioning,
academics, finances, recreations, and health. After a given time period, the
individuals will come back to the laboratory and disclose which events did occur and
which events did not occur. These experiments are seen as more ecologically valid
tests of depressive realism as they occur outside the confines of a laboratory.
Three studies used this paradigm and found different results. Dunning and
Story (1991) found no support for depressive realism as depressed college students
were less accurate in their prediction of future life events. They found that the reason
for the inaccuracy was mainly due to their greater predictions for less occurring
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
31
events to happen to them. As well, depressed individuals exhibited more unrealistic
optimism than non-depressed respondents. Kapci and Cramer (1997) found mixed
support for depressive realism as depressed British college students were more
accurate in predicting the negative life events they would experience and the positive
life events they would not experience. Shrauger, Mariano, and Walter (1998) found
mixed support for depressive realism as depressed individuals were more accurate in
predicting negative outcomes but less accurate when predicting positive outcomes.
Severely Depressed/Clinical Samples
Studies that utilize a clinical sample tend to use the same measures as utilized
in the mildly depressed studies yet depressed subjects were usually required to be
two standard deviations above the mean on these measures. As well, studies that
utilized a depressed sample may include a structured clinical interview to determine
depression status.
Perceived Judgment of Control
Two studies were found that followed similar versions to the Alloy and
Abramson judgment of control paradigm but used clinical populations instead of
college populations (Dobson & Pusch, 1995; Lennox, Bedell, Abramson, & Foley,
1990).
Lennox, Bedell, Abramson, and Foley (1990) used Alloy and Abramson’s
(1979) judgment of contingency with four male subject groups: depressed,
schizophrenic, schizophrenic with depression, and non-psychological
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
32
medical/surgical problems. They found that all four groups were found to perceive
control accurately when a contingent relationship existed between a response and
outcome and the outcome was not tied to effect. Furthermore, it was found that
participants were only accurate when an active bar pressing was required to produce
the outcome.
Dobson and Pusch (1995) used a computerized version of the Alloy and
Abramson (1979) judgment of non-contingency task (experiment 2) with clinically
depressed, remitted depressed, and never depressed females. They found that all
three groups were substantially over optimistic in their judgment of control, with the
remitted depressed group being the most realistic. Furthermore, they found that these
results could not be accounted for through initial trial practice.
Using an innovative judgment of control paradigm, Golin, Terrell, Weitz, and
Drost (1979) had either depressed or non-depressed inpatients estimate the degree of
control that they would have in a dice game (craps). Participants came into a room
with a craps table and were explained that a roll of a 2, 3, 4, 9, 10, 11, or 12 was a
winning roll, while all other rolls were losing rolls. Thus, the objective probability of
obtaining a winning roll was 44%. As well, either the subject would get to roll the
dice or the croupier would roll the dice. In each condition, the participants rated the
likelihood that they would get a winning roll. In both the player control and the
croupier control conditions, depressed inpatients correctly estimated their degree of
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
33
control while non-depressed participants believed they had more control than they
did in the player control condition but not in the croupier control condition.
Overall, these studies provide mixed support for the notion of depressive
realism in judgment of control situations. Clinically depressed individuals were more
accurate than other clinical groups under certain conditions in their prediction of the
amount of control that they had. However, under other conditions, clinically
depressed individuals were subject to the same biasing errors as non-clinically
depressed subjects.
These clinical studies introduce evidence that depressive realism may be
associated with psychopathology in general, not just depression (Lennox, Bedell,
Abramson, & Foley, 1990). It may be that serious distress creates the conditions that
cause individuals to engage in the process that would lead to more accuracy. As well,
these studies introduce evidence that the depressive realism effect is associated with
mild dysphoria as opposed to clinical depression (Dobson & Pusch, 1995,
experiment 2).
Recall of Self-Perceived Performance
Gotlib’s (1983) study with depressed psychiatric inpatients, non-depressed
psychiatric inpatients, and non-depressed hospital workers is typical of a recall of
self-perceived performance paradigm with clinically depressed individuals. Gotlib
had these different groups interact in pairs in which they discussed difference of
opinions on hypothetical situations. Participants then viewed the video interaction
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
34
followed with the receiving of a fictitious evaluation of themselves that contained
ratings o f 13 7-point bipolar adjectives. Participants asked to rate how favorable they
found the evaluation as well as to reproduce the feedback after five minutes on a
blank sheet. Findings in terms of perception of evaluation will be discussed in self-
evaluation studies. Depressed individuals recalled the evaluation to be significantly
more negative than it was in actuality, whereas the other groups remained accurate in
their evaluations.
The findings of Gotlib (1983) were corroborated and extended by a number
of different studies (Kendall, Stark, & Adam 1990, experiments 1 and 2; Gotlib,
1981; Roth & Rehm, 1980). Kendall, Stark, and Adam (1990 experiments 1 and 2)
extended the findings to a performance task with children in an ecologically valid
context, school. Gotlib (1981) extended the findings to a performance task and
further found that the valence of the reinforcement affected the type of bias exhibited
by depressed individuals, such that self-reinforcement recall exhibited an
underestimation bias and self-punishment recall exhibited an overestimation bias.
Roth and Rehm (1980) corroborated the findings of Gotlib (1981) in a social dyad
paradigm. However, one study failed to replicate these findings (DeMonbreum &
Craighead, 1977).
DeMonbreum and Craighead (1977) employed the same methodologies as
discussed in Craighead, Hickey, and DeMonbreun (1979) with psychiatric depressed
outpatients, non-depressed psychiatric outpatients, and a non-depressed non-
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
35
psychiatric sample with the following changes. In the first 40 trials of the
experiment, subjects received random reinforcement for their answers while in the
second 40 trials, subjects were assigned to either a high rate of acceptable feedback
or a low rate of acceptable feedback. Findings from the first forty trials indicated that
all three groups were accurate in the amount of positive feedback that they received.
Findings from the second forty trials reveal that depressed individuals significantly
underestimated their rate of positive feedback in the high rate condition, but were
accurate in their perception of positive feedback in the low rate condition. On the
other hand, non-depressed psychiatric and non-depressed non-psychiatric patients
■ were accurate in th,eir perception of positive feedback in the high rate condition, but
only non-depressed/non-psyehiatric patients were accurate of the perception of
■positive feedback in the low rate condition. Non-depressed psychiatric patients in the
positive feedback condition overestimated the amount of positive feedback. Recall
accuracy of negative’ feedback is not reported by the authors and thus cannot be
discussed.
Overall, in terms of recall of self-perceived performance, these studies seem
to demonstrate that depressed individuals display a bias in their accuracy of recall of
evaluative information such that depressed individuals underestimate the amount of
positive feedback: they receive especially when the amount of reinforcement is high
and may overestimate the amount of negative feedback they receive. These findings
further lend credence to a contextual depressive realism, one whose limits may be
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
36
affected by the amount of feedback (DeMonbreum & Craighead, 1977), the type of
feedback (Gotlib, 1981; Roth & Rehm, 1980), the level of depression, and whether
the task is performance based or social interaction based.
Self-Evaluation
Lewinsohn, Mischel, Chaplin, and Barton’s (1980) study is one of the self-
evaluation studies most often cited in support of depressive realism. The researchers
had depressed psychiatric patients, non-depressed psychiatric inpatients, and normal
controls from the population at large interact in four 45 minute group interactions at
four different assessment periods. In the group interaction, participants provided a
three-minute monologue. After the monologue, the group was free to discuss
anything for 20 minutes. Participants rated themselves on their social ability within
the interaction and blind observational coders rated participants on the same
measures. The researchers found that depressed individuals accurately rated their
social ability within the interaction while both non-depressed groups rated
themselves as performing better in the interaction than objective raters had observed.
Other self-evaluation studies have found similar results and also extended the
findings of this study (Ducharme & Bachelor, 1993; Gotlib, 1983; Rozensky, Rehm,
Pry, & Roth, 1977). Rozensky, Rehm, Pry, and Roth (1977) extended the findings to
a performance task and the valence of the reward. They found using a word
association task in which depressed, non-depressed referred clinical patients, or
control subjects rewarded or punished themselves for their correct responses on a
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
37
word task, the ratio of correct responses to self-reward was smallest for depressed
individuals than either group, indicating that depressed individuals were more
accurate in their administration of self-reward. Non-depressed referred clinical
patients and controls both over-estimated the amount of self-reward to a much higher
degree. This was true despite all three groups performing equally on the task.
Gotlib (1983) extended the finding to a performance task that demanded
participants understand that a bogus evaluation of their performance was neutral. He
found that depressed psychiatric inpatients, non-depressed psychiatric inpatients, and
non-depressed hospital workers accurately viewed their bogus evaluation as being
neutral in its evaluation of them. This demonstrated that all groups were accurate in
their self-appraisal.
Ducharme and Bachelor (1993) extended the findings to different severities
of depression by using both depressed students and depressed clients. As well, they
added further measures to examine the specific variables of perceived social distress
and self-efficacy expectations. In terms of depressive realism, they found that
although between group differences were not found in terms of social skill deficits,
depressed students more accurately appraised their performance (their ratings were
closer to the objective raters reporting of their social skill). However, depressed
clients believed that they had evaluated themselves better than the “objective”
observers. As well, the researchers found that depressed individuals indicated higher
social distress and avoidance, more negative expectations regarding their social
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
38
competency, and more negative attitudes with regard to their interpersonal skills
before actually attempting the task. Although depressives had these negative
expectations, they did not perceive themselves in the actual social encounter as less
competent than non-depressives.
However, other studies failed to fully support the notion of depressive
realism for clinically depressed participants in self-evaluation studies (McKendree-
Smith & Scogin, 2000; Kendall, Stark, & Adam, 1990, experiment 3; Dow &
Craighead, 1987; Golib, 1981).
Gotlib (1981) found that non-depressed hospital patients and depressed
hospital patients rewarded themselves significantly less than non-depressed hospital
employees despite performing equally. As non-depressed/non-psychiatric were
accurate in the amount of the amount of rewarding behavior compared to the number
of answers they got correct, it was determined that depressed and non-depressed
psychiatric inpatients were inaccurately harsh on their self-evaluative performance.
Dow and Craighead (1987), in a similar interaction paradigm to Lewinsohn,
et. ah, (1980), but employing a confederate group, found that depressed individuals
had a greater discrepancy in the self-evaluation of the groups social skill compared to
the confederate rating of their own social skill more so than any other group.
However, non-depressed/non-anxious individuals had a greater discrepancy in the
ease at being heard and understood to the confederate rating o f the groups ease at
being understood compared to the depressed group.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
39
Kendall, Stark, and Adam (1990, experiment 3) compared depressed and
non-depressed children’s reports of their school performance, popularity, and athletic
prowess to the same report filled out by their teachers. As teachers were seen as the
objective standard, depressed individuals evaluated their performance in these areas
as significantly worse than the teacher’s ratings. In contrast, it was found that non-
depressed individuals were accurate in their evaluation of themselves compared to
teacher’s ratings.
McKendree-Smith and Scogin (2000) expanded the work of Vestre and
Caufield (1986) by the inclusion of a truly depressed sample along with a mildly
depressed sample and control sample. As well, they included the variable of time.
Participants rated their bogus personality profile on two separate occasions in terms
of its accuracy in describing themselves. The researchers found that mildly depressed
subjects were the most accurate in their ratings of their personality profile, while
severely depressed subjects displayed negative distortion biases and non-depressed
individuals displayed a positive distortion biases. As well, the researchers found that
time did not matter in overall ratings of accuracy, since participants rated themselves
the same on each trial.
Overall, the self-evaluation studies provide mixed support for depressive
realism. Whereas some studies demonstrate depressive accuracy in evaluation, other
studies reveal that depressed individuals display a negative distortion bias. This
negative distortion bias causes severely depressed individuals to see themselves as
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
40
less capable then they truly are. These findings further lend credence to a contextual
depressive realism, one whose limits may be affected by the type of self-evaluation
task (performance versus social interaction based), the level of depression
(McKendree-Smith & Scogin, 2000; Ducharme & Bachelor, 1993), and the valence
of self-evaluation (Rozensky, Rehm, Pry, & Roth, 1977).
Self/Other Evaluation Studies
It was possible to identify three very different types of self-evaluation studies
with clinically depressed individuals. All three supported some aspect that depressed
individuals are more accurate in their evaluations than non-depressed individuals
(Lovejoy, 1991; Rickters & Pellegrini, 1989; Hoehn-Hyde, Schlottman, & Rush,
1982). As each study was unique in its paradigm and findings, each one will be
discussed.
Hoehn-Hyde, Schlottman, and Rush (1982) had depressed, remitted
depressed, and never depressed control subjects rate nine positive, negative, or
neutral videotaped scenes in which the principal actor in the tapes was either
directing his comments to the participant (self condition) or to another person (other
condition). Participants rated these scenes in terms of 11 bipolar adjectives. In terms
of depressive realism, depressed individuals were the most accurate in rating the
neutral scenes as such in both conditions compared to remitted depressed and control
subjects. Both remitted depressed subjects and controls rated the neutral scenes more
positively than they were.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
41
Rickters and Pellegrini (1989) compared clinically depressed, remitted
depressed, and non-depressed mothers ratings of their children’s problem behavior to
the teacher’s evaluation of the child’s behavior. The researchers found that all groups
were accurate assessors of their child’s behavior.
Lovejoy (1991) had clinically depressed and non-depressed mothers interact
with their children in a two part laboratory session that was being rated by
independent observers in terms of the amount of positive (compliance, play, talk,
question, and request, physical affection, laughter, approval) and negative (yell, cry,
negative commands, ignore, tease, threat, whine, or negativism) child behaviors.
After the two sessions, the mothers also rated the amount of positive and negative
child behaviors in the two paradigms. In support of depressive realism, clinically
depressed mothers accurately perceived the amount of negative child behavior
compared to the independent observers whereas the non-depressed mothers
underestimated the amount of negative child behaviors compared to the independent
observers. However, depressed and non-depressed mothers overestimated the
amount of positive behaviors of their child compared to the observers.
Overall, these findings further support the notion o f a contextualized
depressive realism. It seems that with self/other evaluation studies, the ability for
depressed participants to be accurate was influenced by the valence of the behavior
under observation, such that depressed individuals were most accurate when the
behavior under examination was positive or neutral (Hoehn-Hyde, Schlottman, &
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
42
Rush, 1982; Lovejoy, 1991). These findings also lend credence to the notion that
depressive realism is a state based phenomenon rather than a trait-like characteristic
of people that become depressed (Hoehn-Hyde, Schlottman, and Rush, 1982).
Prediction of Future Events
No studies to date have examined the prediction of future events with clinical
samples.
Conclusions
Overall, of the 73 studies that met the necessary criterion to be an
examination of depressive realism, 28 fully supported the notion, 30 provided mixed
results, and 17 found no support. For a summary review of all the studies, see Table
1 below.
Table 1.
Summary of Findings on Depressive Realism
A uthor/D ate T ype o f Sam ple Support D epressive R ealism
(Mildly Depressed = MD; (Full Support = FS, Mixed
Clinically Depressed = CD) Support = MS; No Support = NS)
Judgm ent of C ontrol Studies
1) Alloy & Abramson (1979) MD FS
2) Golin, Terrell, & Johnson (1977) MD MS
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
43
Table 1-continued.
Summary of Findings on Depressive Realism
A uthor/D ate Type o f Sam ple
(Mildly Depressed = MD;
Clinically Depressed = CD)
Support D epressive R ealism
(Full Support = FS, Mixed
Support = MS; No Support = NS)
3) Abramson, Alloy, & Rosoff (1981) MD MS
4) Alloy, Abramson, & Viscusi (1981) . MD MS
5) Alloy & Abramson (1982) MD FS
6) Martin, Abramson, & Alloy (1984) MD MS
7) Benassi & Mahler (1985) MD MS
8) Vazquez (1987) MD MS
9) Mikulincer, et. al. (1990) MD MS
10) Alloy & Clements (1992) MD FS
11) Rosenfarb, Burker, Morris, Cush (1993) MD FS
12) Pacini, Muir, Epstein (1998) MD MS
13) Pusch, Dobson, et al. (1998 exp #2) MD NS
14) Koeing, Clements, and Alloy (1992) MD NS
15) Ford & Neale (1985) MD NS
16) Bryson, Doan, & Pasquali (1984) MD NS
17) Msetfi et. al (2005) MD FS
18) Glass, et al. (1993) MD FS
19) McKnight & Glass (1995) MD FS
20) Lennox, et. al (1990) CD MS
21) Golin, Terrell, Weitz, & Drost (1979) CD FS
22) Dobson & Pusch (1995) CD NS
Recall of Self-Perceived Perform ance
23) Kennedy & Craighead (1988) MD MS
24) Buchwald (1977) MD MS
25) Nelson & Craighead (1977) MD MS
26) Dobson & Shaw (1981) MD MS
27) Dykman, Abramson, et al. (1989) MD MS
28) Dykman, Horowtiz, et al. (1991) MD MS
29) Kuiper (1978) MD MS
30) Craighead, Hickey, et al. (1979) MD FS
31) Dykman & Volpicelli (1983) MD FS
32) Gotlib (1983) CD NS
33) Kendall, Stark, & Adam (1990) CD NS
34) Gotlib (1981) CD NS
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
44
Table 1-continued.
Summary of Findings on Depressive Realism
A uthor/D ate T ype o f Sam ple
(Mildly Depressed = MD;
Clinically Depressed = CD)
Support D epressive R ealism
(Full Support = FS, Mixed
Support = MS; No Support = NS)
35) Roth & Rehm (1980) CD NS
36) DeMonbreum & Craighead (1977) CD MS
Self-Evaluation Studies
37) Gotlib & Meltzer (1987) MD MS
38) Edison & Adams (1982) MD FS
39) McNamara & Hackett (1986) MD FS
40) Altman & Gotlib (1988) MD FS
41) Kistener et. al (2001) MD FS
42) Dykman, Horowitz, et. al. (1991) MD NS
43) Pusch et. al (1998, exp #1) MD FS
44) Cane & Gotlib (1985) MD NS
45) Heiby (1981) MD MS
46) Keller, Lipkus, & Rimer (2002) MD FS
47) Nelson & Craighead (1977) MD MS
48) Stone, Dodrill, & Johnson (2001) MD MS
49) Vestre & Caufield (1986) MD FS
50) Beyer (2002) MD NS
51) Lewinsohn, Mischel, et al. (1980) CD FS
52) Rozensky, Rehm, Pry, & Roth (1977) CD FS
53) Gotlib (1983) CD FS
54) Ducharme & Bachelor (1993) CD,MD MS
55) Gotlib (1981) CD NS
56) Dow & Craighead (1987) CD MS
57) Kendall, Stark, & Adam (1990) CD NS
58) McKendree-Smith & Scogin (2000) CD, MD MS
Self/Other Social Com parison Studies
59) Tabachink, Crocker, & Alloy (1983) MD MS
60) Crocker, Alloy, & Kayne (1988) MD FS
61) Ahrens, Zeiss, & Kanfer (1988) MD FS
62) Cambell (1986) MD MS
63) Albright & Henderson (1995) MD NS
64) Albright, Alloy, et al. (1993) MD MS
65) Alloy & Ahrens (1987) MD FS
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
45
Table 1-continued.
Summary of Findings on Depressive Realism
Author/Date Type of Sample Support Depressive Realism
(Mildly Depressed = MD; (Full Support = FS, Mixed
Clinically Depressed = CD) Support = MS; No Support = NS)
66) Strack & Coyne (1983) MD FS
67) Dobson (1989) MD NS
68) Hoehn-Hyde et. al. (1982) CD FS
69) Rickters & Pelligrini (1989) CD FS
70) Lovejoy (1991) CD MS
Prediction o f Future
71) Dunning and Story (1991) MD NS
72) Kapci and Cramer (1998) MD MS
73) Shrauger, et al. (1998) MD MS
Taken together, these studies seem to indicate that depressive realism may
not be an interminable concept that defies all boundaries and tasks. Rather, it seems
that depressive realism may be a bounded contextual concept, as depressed
individuals are more accurate in their perception of reality in some circumstances or
conditions and less accurate in their perception of it in others. Furthermore, in some
situations, depressed individuals actually display distinct negative biases in their
cognition that are different from both mildly depressed and non-depressed
individuals. This is consistent with the past reviews of depressive realism (Ackerman
and DeRubeis, 1991; Dobson & Franche, 1989; Alloy and Abramson, 1988).
These studies also provide strong evidence that depression may lead
individuals to engage in distinct cognitive processes from non-depressed individuals
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
46
that result in differential rates in accuracy depending on a number of conditions.
Some of the conditions that seem to change depressed individuals’ judgments away
from accuracy or bound depressive realism are: the presence of another person in the
room while conducting the experiment, the motivation of the participant, the location
of the task, the task difficulty, the amount of feedback the person receives, the
relevance of the task to the person’s identity, the valence of the information
(positive, negative, or neutral), the length of time in which the participant is making
the judgment, the level of the person’s depression, and the reference point of the
judgment.
Furthermore, differences did emerge in the findings based on the level of
depression (see table 2). Studies employing severely depressed individuals seem to
provide more support for a negative distortion bias in which depressed individuals
perceive the event or task as worse than objectively warranted. This lends support to
current depression theories that view depression as a faulty cognitive processing
system. However, even studies with clinically depressed inpatients found instances
of depressive accuracy. Studies that used mildly depressed individuals provide
stronger support for depressive realism. This pattern suggests that depressive realism
may be a phenomenon more strongly associated with moderate levels of depression
rather than with severe levels.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
47
Table 2.
Level o f Depression and Degree of Support for Depressive Realism
Level o f Depression
Number
and % o f
Full
Support
Number
and % o f
Partial
Support
Number
and % o f
No
Support
Total
Number
and % o f
Support
Mildly Depressed 20 (37.0%) 24
(44.4%)
10
(18.6%)
54 (100%)
Clinically Depressed 6 (31.6%) 6
(31.6%)
7
(36.8%)
19(100%)
Total 26 (35.6%) 30
(41.1%)
17
(23.3%)
73 (100%)
Lastly, differences did emerge based on the type of task (see Table 3). It
seems that self/other social comparison studies provided the most optimal conditions
for the depressive realism effect to emerge, with self-evaluation studies and
judgment of control studies following closely behind. Interestingly, recall of self
perceived performance and the prediction of future studies provided the least support
for depressive realism. This may be due to the increasing element of time with these
studies that allow for another process to undermine depressed individuals accuracy.
However, the explanation of these differences can only be speculated due to
the limitations with the current research that the present study will attempt to remedy
in that none of the studies demonstrate that depressed and non-depressed individuals
engage in different cognitive processes that lead to differential rates of accuracy.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
48
Table 3.
Type o f Study and Percentage of Support
Tvpe of Study
Number
and % o f
Full
Support
Number
and % o f
Partial
Support
Number
and % o f
No
Support
Total
Number
and % o f
Support
Judgment of Control Studies 8 (36.4%) 9 (40.9%) 5 (22.7%) 22 (100%)
Recall of Self-Perceived
Performance
2 (14.3%) 8(57.1%) 4 (28.6%) 14(100%
Self-Evaluation Studies 10(45.5%) 7(31.8%) 5 (22.7%) 22(100%)
Self/Other Social Comparison
Studies
6 (50.0%) 4 (33.3%) 2(16.7%) 12(100%)
Prediction of Future Studies 0 (n/a) 2 (66.7%) 1 (33.3%) 3 (100%)
Total 26 (35.6%) 30
(41.1%)
16
(21.9%)
73 (100%)
Overall Limitations in Current Research
As is true of all psychological research, each of the individual study’s
reported here have their own methodological faults or weaknesses, but the literature
as a whole suffers from four main problems that make any definitive conclusions
from the literature difficult.
The first problem with the current research is that the methods used to not
permit an understanding of the cognitive processes that lead to differential rates of
accuracy between depressed and non-depressed individuals. Instead, studies can
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
49
point to possible differences in personality characteristics, cues, or task demands of
the experiment but these are merely speculations. New studies need to be
implemented that either ask individuals the reasoning behind their choices,
predictions, or ratings that provide a process to examine the cognitive mechanisms
that lead to the differential responses. Until this is done, it will remain difficult to
explain the current variability among the findings of the studies of depressive
realism. The assumption that drove this present study was that depressed and non-
depressed individuals utilize different processes in the same situation and that these
lead to differential rates of accuracy.
Second, as described in the previous reviews (Ackerman and DeRubeis,
1991; Dobson & Franche, 1989; Alloy and Abramson, 1988) most studies continue
to examine the phenomenon within the confines o f a laboratory. Findings from an
experimental design in a laboratory may not generalize to depressive cognition in
general. Moreover, the experimental paradigm which necessitates the introduction of
a novel situation and different stimuli to the individual could be another factor which
bounds depressed individuals ability to make accurate judgments. It was possible to
locate only six studies that examined depressive realism more naturalistically. The
depressive accuracy effect was found in all but one (Kendall, Stark, and Adam
(1990, experiment 3).
Third, the two most common methodologies that provide the strongest
support for the phenomenon (self/other social comparison studies and self-evaluation
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
50
studies, Table 2) employ an “unbiased observer” or “objective rater.” Although these
observers have some objectivity in their responses, it cannot be ruled out that these
raters are biased as well (Ackermann & DeRubeis, 1991). Thus, these methods lack
the ability to be a strong test for the depressive realism effect. New methods need to
be employed that can use the same type of judgment task but eliminate the objective
rater.
Fourth, studies of depressive realism have continued to examine the
phenomenon with either four-year university students or clinically depressed
inpatients. Few other populations have been used in this research. A review of
studies demonstrated that traditional four-year college students manifest different
levels of symptomatology and rates of depression than the general population
(Nolen-Hoeksema, 1987). On the other hand, clinically depressed inpatients are at a
severe level of distress that can also debilitate their ability for accuracy and may
have partially led to the lower evidence of depressive realism in these populations.
Studies that have examined the phenomenon beyond these populations have found
strong evidence for the effect. Yet, only five different populations have been
examined: depressed mothers (Richters & Pellegrini, 1989), depressed women ages
40 to 60 (Keller, Lipkus & Rimer, 2002), depressed nurses (Glass, McKnight, &
Valdimarsdottir, 1993), children (Kendall, Stark, & Adams, 1990), and adolescents
(Kistner, Balthazor, Risi, & David, 2001). New populations need to be examined in
order to explore the validity of depressive realism.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
51
The Present Study
The present study was designed to address some of the methodological
problems found in past research on depressive realism. Specifically it employed a
community sample and a daily diary methodology.
A daily diary methodology allows people to serve as their own measure of
accuracy, basically creating an objective measure of a person’s subjective
experience. As people fill out the measure each day over eight consecutive days, the
average across the last seven days would be the objective account of the person’s
recall of the week and any deviation from the average upon recall of that week would
be inaccurate. Thus, this study creates the objectiveness lacking in the other self-
evaluation studies.
This methodology also allows for the exploration of the phenomenon in an
ecologically valid context; a major concern identified in past reviews (Ackerman
and DeRubeis, 1991; Dobson & Franche, 1989; Alloy and Abramson, 1988). People
answered the questions over the phone; and the questions related to the person’s own
experience during the day. No laboratory was used.
Another strength of this study is that it utilized a clinically depressed sample
of individuals rather than mildly depressed or dysthymic individuals. This too was a
major issue of concern identified by past reviews (Ackerman and DeRubeis, 1991;
Dobson & Franche, 1989; Alloy and Abramson, 1988).
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
52
A further strength of the daily diary format is that it permits a quantitative
examination of the problem solving strategies of the participants’ in their recall of
different areas. Because participants’ completed the measures each day, it was
possible to determine what amount or combination of daily judgments account for
the recall report. This is important as the present study made the predication that
depressed and non-depressed use different problem solving strategies that lead, in
turn, to differential rates of accuracy.
Two problem solving strategies have emerged as explanatory mechanisms for
how people recall emotional experiences (Kahneman, 1999; Thomas & Diener,
1990;). One of these problem solving strategies, which will be referred as the
average hypothesis, assumes temporal integration in the recall of events. The average
notion suggests that each person will equally weigh each aspect of an event into his
or her appraisal of the event. Thomas and Diener (1990) used the average hypothesis
and found that people tend to overestimate the intensity of their positive and negative
experiences.
Another set of studies found instead of temporal integration, that recall of any
experience can be best accounted for by knowledge of the peak emotional intensity
of the experience and the end experience (Almeida & Mroczek, 2004; Kahneman,
1999; Redelmeier & Kahneman, 1996; Fredrickson & Kahneman, 1993; Kahneman,
Fredrickson, Schreiber, & Redelemeier, 1993). Kahneman (1999) articulated this
theory after conducting a series of studies in which he and his colleagues compared
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
53
real time evaluations of events with recall of the events. For example, one series of
studies compared in the moment ratings of patients undergoing a colonoscopy and
found that retrospective ratings of the procedure were best predicted by the most
intense period and the end period of the procedure (Redelmeier & Kahneman, 1996).
In another study he manipulated the experienced pain at the end of the procedure and
found that patients recalled less amounts of pain when the end of the colonoscopy
was less painful even though the rest of the procedure was conducted in the same
manner (Kahneman, Fredrickson, Schreiber, & Redelemeier, 1993). Almeida,
Mroczek, and Neiss (2004) extended these findings to the retrospective recall of
negative emotions across the lifespan.
This study examined both integration patterns (average hypothesis and peak
end hypothesis) to determine if the different groups used different processes in their
recall of their psychological distress and positive mood.
The features of the present study that so far have been noted were to
overcome some o f the methodological weaknesses of past studies of depressive
realism and to attend to the areas of future research identified by previous reviewers.
But, this study also was designed to extend the depressive realism research in three
ways. First, it was designed to extend the ability to explore accuracy to the recall of
one’s mood. Second, it was designed to extend the findings to a population that has
never been studied in regard to depressive realism, a community sample of male and
female adults ages 25 to 60. Third, this methodology allows for an exploration of the
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
54
process which may lead to differences in accuracy, something that only study that
was found has begun to be able to accomplish (Msetfi, Murphy, Simpson, &
Kombrot, 2005 (experiments 1 and 2).
This study extended the depressive realism notion to a new task, recall of
one’s emotional mood over the course of the week. In the context of the depressive
realism literature, this study falls within a hybrid between two types of depressive
realism studies, self-evaluation (as participants are evaluating their own week) as
well as recall of perceived performance (as participants are asked to recall their
week). For this, it employed a new methodology, a daily diary format. As such, this
study makes the following hypotheses:
1) Depressed individuals will more accurately recall their average weekly
psychological distress than non-depressed individuals.
2) Depressed individuals will more accurately recall their average weekly
positive good mood than non-depressed individuals.
Based on the notion that differences in accuracy of recall are due to different
problem solving approaches by the two different groups, this study makes these
further hypotheses:
3) Depression status will moderate the relationship between recall of weekly
psychological distress and the valuing of peak psychological distress and
end day psychological distress, such that depressed individuals and non-
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
55
depressed individuals will use different processes in their valuing of peak
and end experience in their recall of their own psychological distress.
4) Depression status will moderate the relationship between recall of weekly
psychological distress and the averaging of their weekly experience, such
that depressed individuals and non-depressed individuals will use
different processes in the averaging of their weekly experience in their
recall of their own psychological distress.
5) Depression status will moderate the relationship between recall of weekly
positive affect and the valuing of peak positive affect and end day affect,
such that depressed individuals and non-depressed individuals will use
different processes in their valuing of peak and end experience in their
recall of their own positive affect.
6) Depression status will moderate the relationship between recall of weekly
positive affect and the averaging of their weekly positive affect, such that
depressed individuals and non-depressed individuals will use different
processes in the averaging o f their weekly experience in their recall of
their own positive affect.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
56
Chapter 3: Methodology
This chapter pertains to the methodology implemented in this study. It
contains detailed descriptions of the demographic participants in the study. It details
the measures employed in the study including how the measures were created
coupled with the psychometric properties of these measures. Lastly, it describes the
procedure used to attain the sample for the study.
Method
Partcipants
Participants were 342 men (42.9% of the sample) and 455 females (57.1% of
the sample). Almost all (90.1%) of the sample identified themselves as Caucasian,
4.3% as African American, 0.4% as Native American, 0.6% as Asian or Pacific
Islander, 1.3% as Other, and 0.9% as multiracial. The average age of the sample was
48.81 years and the average total household income $56,077.34 (1998 dollars). In
terms of educational status, 93.0% had obtained at least a high-school degree, 38.9%
had obtained either a college or community college degree, and 9.9% had obtained
an advanced degree of either a Masters or Doctorate. Approximately two-thirds
(66.4%) of the sample was married at the time of study, 2.0% separated, 15.2%
divorced, 6.4% widowed, and 10% never married. The average number of biological
children per person was 2.10. Eighty eight and two tenths percent of the sample (703
participants) received no diagnosis of depression and 11.8% (94 participants)
received a diagnosis for depression.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
57
Measures
Three measures were utilized in this study. The World Health Organization
Composite
International Diagnostic Interview Short-Form was used to diagnose depression.
Two measures were used to assess affect that varied by intensity and duration of the
Measure: The Non-Specific Psychological Distress scale and the Positive Affect
measure.
World Health Organization Composite International Diagnostic Interview
Short-Form (CIDI-SF). The CIDI-SF is a fully standardized diagnostic interview for
the assessment of mental disorders according to the definitions and outlined criteria
of the Diagnositic and Statistical Manual of Mental Disorders- revised (DSM-III-R;
APA 1987). The interview as a whole has shown exceptional test-retest and inter
rater reliabilities as well as strong validity between clinical checklists and CIDI
diagnosis (Wittchen, 1994, Blazer, Kessler, McGonagle, & Swartz, 1994). This was
especially evident in the area of depressive disorders (K =.84). The CIDI-SF has
been tested world-wide with adult samples and is one of the main diagnostic tools
researchers and clinicians use to diagnose psychological difficulties (Wittchen,
1994):
The CIDI-SF was used to categorize participants as meeting a diagnosis of a
major depressive episode by utilizing the criteria specified in the third edition-
revised of the American Psychiatric Association’s (APA) Diagnostic and Statistical
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
58
Manual of Mental Disorders (DSM-III-R; 1987). These criteria require a period of at
least two weeks of either depressed mood or anhedonia most of the day, nearly every
day, and a series of at least four other associated symptoms typically found to
accompany depression. These include problems with eating, sleeping, energy,
concentration, feelings of self-worth, and suicidal thoughts or actions. Participants
were asked a series of questions surrounding each criteria for depression and whether
the person experienced these symptoms in the past 12 months for most of the day,
nearly every day during the same two week period. If the person met the major
criterion for depression and at least four symptoms of depression, then they were
coded as having been depressed.
Non-Specific Psychological Distress Scale. The 10 item non-specific
psychological distress scale (Mroczek & Kolarz, 1998) was developed in previous
studies and utilized in this study. The non-specific psychological distress scale
comes from several well-known instruments: The Affect Balance Scale (Bradbum,
1969), the University o f Michigan’s Composite International Diagnostic Interview
(Kessler, McGonagle, Zhao, Nelson, Hughes, Eshleman, Wittchen, & Kendler
(1994), the Manifest Anxiety Scale (Taylor, 1953) and the Center for
Epidemiological Studies Depression Scale (Radloff, 1977). The scale assesses
intensity o f eight emotions: sadness, nervousness, worthlessness, hopelessness,
anxiety, irritability, depression, and fidgetiness. Respondents indicated how much of
the time they experienced each emotion on a 5-point scale from none of the time to
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
59
all of the time. Two emotions, sadness and nervousness, were assessed via two item
measures that used the same likert scales rather than a one item measure.
Psychological distress scores were summed such that higher scores were indicative
of more psychological distress. This measure was used as the basis for categorizing
five scores used in this study:
A Daily Distress score was assessed during the daily diary in which
respondents indicated how often they felt each of the eight emotions “during the past
24 hours.” On each day, mean scores across the ten items were computed for a daily
distress score (Cronbach’s alpha=.82).
A Weekly Aggregate Distress score was created by computing the mean of
daily distress scores across the last seven diary day.
A Peak Distress Day score was created by selecting the highest score among
the seven daily distress scores (Cronbach’s alpha=.82).
An End Distress Day score was created by selecting the daily distress score
on the last day of the interview (Cronbach’s alpha=.85).
A Recall Distress score was assessed at the conclusion of the final day of
interviewing when respondents were asked on the same likert scale as the daily
distress score how often they felt each of the 8 emotions “during the past week.” The
score was obtained by computing the mean score across the 10 items (Cronbach’s
alpha=.83).
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
6 0
A Within Person Distress Variation score was assessed by summing the
absolute value of the daily deviations around the average daily distress score
(Cronbach’s alpha = .85).
Positive Affect. Positive affect was assessed via a one item measure in which
participants described how often they felt “in good spirits during the past 24 hours.”
Scores ranged on a 5-point scale from none of the time to all of the time such that
higher scores indicate more positive affect. This item was selected from the an
inventory o f six positive affect emotions used in the MIDUS survey including being
cheerful, calm and peaceful, satisfied, and full of life (Mroczek & Kolarz, 1998) and
was chosen based on its significant contribution to the strength of its scale as well as
its high degree of association with the sum of the other four items (r = .79). This
single item was used as the basis for computing six scores used in this study:
A Daily Positive Affect score was created using the participants response to
the item on each of the diary days.
A Weekly Aggregate Positive Affect score was created by computing the
mean frequency of positive affect across the last seven diary day.
A Peak Positive Affect score was created by selecting the highest score
among the seven positive affect scores.
An End Positive Affect Day score was created by selecting the final day daily
positive affect score.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
61
A Recall Positive Affect score was created from a question on the final day
of interviewing when respondents were asked how often on the same likert scale as
the daily positive affect item how often they felt in good spirits “during the past
week.”
A Within Person Positive Affect Variation score was assessed by summing
the absolute value of the daily deviations around the average daily positive affect
score (Cronbach’s Alpha = .80).
Procedure
The hypotheses were investigated using participants who completed the
major depressive episode diagnosis portion of the Midlife in the United States Study
(MIDUS) and who subsequently completed at least six of the final seven days of the
eight day National Study of Daily Experiences (NSDE).
The MIDUS study consisted of a nationally representative telephone-mail
survey designed to explore successful aging and midlife (for a detailed review of the
MIDUS project see Keyes & Ryff, 1998; Mroczek & Kolarz, 1998). Participants
were recruited through random digit dialing which yielded a total of 3032 people in
the age range of 25 to 74 and occurred between 1995 through 1996. Participants
were diagnosed as having a major depressive episode in the last year through the
World Health Organization Composite International Diagnostic Interview Short-
Form (CIDI-SF). A description of the measure and its psychometric properties can
be found in the measures section.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
6 2
The NSDE started immediately following the culmination of the MIDUS
study and data collection lasted over the course of a year (March 1996 to March
1997). Participants in the NSDE were a randomly selected subsample of the MIDUS
study who were offered $20 to take part in this follow-up study designed to
understand the daily stressors, emotional experiences, and the impact that these
events had on the person’s day and productivity. This study consisted of a short
telephone interview administered over the course of eight consecutive evenings and
on the final evening of the administration, respondents answered an additional series
of questions related to their recall of the entire week. O f the 1242 MIDUS
respondents who were contacted, 1031 participated yielding a response rate of 83%.
Data collection consisted of 40 separate “flights” of interviews with each flight
representing the eight-day sequence of interviews. Approximately 38 respondents
took part in each interview flight and the starting day of each flight was staggered
across the days of the week to control for the possible confounding between day of
study and day of week. Respondents completed an average of 7 of the 8 interviews
yielding a total of 7221 daily interviews.
The present analyses resulted in the inclusion of 797 adults (77.3% of the
original NSDE sample) who completed the major depressive episode diagnosis
portion of the Midlife in the United States Study (MIDUS) and who subsequently
completed at least six of the final seven days o f the NSDE. Thus, the analysis
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
63
involved 5,327 days of data (252 participants who filled out 6 of the last 7 diary days
and 545 who filled out the last 7 days).
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
64
Chapter 4: Results
Descriptive Differences
A series of chi-square and analyses of variance were conducted to determine
demographic and appraisal differences between the depressed and non-depressed
groups who had completed at least six of the final seven days of the study. The
demographic differences that were tested included gender, ethnicity, marital status,
age, years of education, and household total income. Then a hierarchical linear
regression was conducted to determine if the demographic differences between the
groups accounted for the difference in recall reports above depression status. The
appraisal differences included a series of analysis of variance which explored
differences in average daily psychological distress, peak day distress, within person
distress variation, recall of psychological distress, average daily positive affect, peak
day positive affect, within person positive affect variation, and recall of positive
affect.
Assessing Differences in Accuracy Recall by Depression Status
Independent sample t-tests were conducted to explore hypotheses one and
two. These analyses explored both the difference and the absolute value o f the
difference between average weekly reports of either psychological distress or
positive affect and recall of the same period between the depressed and non-
depressed groups.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
65
Accounting for Differences in Recall
To answer hypotheses 3 and 5, linear hierarchical multiple regressions were
performed with step one being depression status, step two adding peak weekly
experience and end weekly experience, and step three adding the interaction of
depression status by peak weekly experience and end experience predicting recall of
the weekly experience (distress or positive affect). If interaction effects were found,
two steps were performed to fully understand the nature of the interactions. The first
step included subsequent regressions with each independent group (depressed and
non-depressed). These regressions followed the same format with peak experience
and end experience predicting recall of the weekly experience (distress or positive
affect). The second step included graphing the interaction effects of the hierarchical
linear regression in order to understand where the differences lay.
To answer hypotheses 4 and 6, hierarchical multiple regressions were
performed with step one being depression status, step two adding weekly average
affect, and step three adding the interaction of depression status by weekly average
affect (distress or positive affect). If interaction effects were found, two steps were
performed to fully understand the nature of the interactions. The first step included
subsequent regressions with each independent group (depressed and non-depressed).
These regressions followed the same format with weekly average affect predicting
recall of the weekly experience (distress or positive affect). The second step involved
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
66
graphing the interaction from the hierarchical linear regression in order to understand
where the differences lay.
O f note, all predictors were centered prior to computing the interactions as
suggested by Aiken and West, 1992. This was done in order to reduce non-essential
multi-colinearity.
Descriptive Results
Tables 4 and 5 report the comparisons of depressed and non-depressed
groups on demographic variables. Whereas the two groups were relatively similar in
terms of ethnicity (% 2 (3) = 3.76, p > .05), education in years (F (1, 795) = .01, p >
.05, number of children (F (1, 795) = .84, p > .05, and total household income (F (1,
795) = 3.60, p > .05), they were significantly different from each other in terms of
gender, marital status, and age. The depressed group had a significantly greater
proportion of women (and, therefore, a smaller proportion of males) than the non-
depressed group (68.1% female depressed compared to 55.6% female non-
depressed), (% 2(1) = 5.26, p < .05). The depressed group had a significantly greater
proportion of not married people (and, therefore, a smaller proportion of married
people) than the non-depressed group (48.9% not married depressed compared to
31.6% unmarried non-depressed), (% 2(1) = 11.19, p < .01). The depressed group was
also significantly younger in age (Ma g e = 45.25, SD = 12.36) than the non-depressed
group (Ma g e = 49.29, SD = 13.13), (F (1, 793) = 7.94, p < .01).
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
67
Table 4.
Chi-Square Demographic Differences between Depressed and Non-Depressed
Groups
Depressed
N and
Percentage
Non-Depressed
N and
Percentage
Pearson Chi-
Square
Gender 5.26*
Male 30(31.9%) 312 (44.4%)
Female 64 (68.1%) 391 (55.6%)
Ethnicity 3.76
Caucasian 82 (89.1%) 636 (93.0%)
African 4 (4.3%) 29 (4.2%)
American
Asian 1 (1.1%) 4 (0.6%)
Other 5 (5.4%) 15 (2.2%)
Marital 11.19**
Status
Married 48 (51.1%) 481 (68.4%)
Not Married 46 (48.9%) 222 (31.6%)
Note. N = 797. (Depressed = 94. Non-depressed =703.) *p < .05. **p < .01
Table 5.
One Way Anova Demographic Differences between Depressed and Non-Depressed
Groups
Demographic
Variable
Chapter 2 Depresssed
Mean (SD)
Chapter 3 Non-
Depressed
Mean (SD)
F
Age (In years) 45.26 (12.36) 49.29 (13.13) 7.94**
Education (In Years) 14.22 (2.26) 14.20 (2.32) 0.01
Number of Children 1.95 (1.70) 2.11 (1.72) 0.84
Total Household 47,159.34 (45,827.40) 57,269.79 3.60
Income (48,880.08)
Note. N = 797. (Depressed = 94. Non-depressed = 703.) * p < .05. **p < .01.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
68
Table 6 reports the results of the hierarchical linear regression that
determined if the demographic differences between the groups accounted for the
difference in recall reports above depression status. Step one had gender, age, and
marital status and step two had Depression Status predicting weekly recall of
psychological distress. The first step of the regression analysis revealed that gender
(P = .11, p < .05) was a significant predictor of weekly recall of
Table 6.
Hierarchical Multiple Regression of Gender. Age. Marital Status, and Diagnosis
predicting Weekly Recall of Psychological Distress
Step 1 Step 2
b (SE)
P
b (SE)
P
Gender .09 (.04) .11* .08 (.04) .10
Age .07 .05 .08 .05 (.05) .06
Marital Status -.05 .05 -.06 -.06 .05 -.07
Diagnosis .27 .06 .25**
R2 2.20 8.40**
Change in R 6.20**
Note. N = 799. (Depressed = 94. Non-depressed = 703.) **p < .01. *p < .05.
psychological distress while age ((3 = .07, p > .05) and marital status (P = -.06, p >
.05) were insignificant predictors. The total variance accounted for by these variables
was 2.20% of the variance in weekly recall of psychological distress. On the second
step, each demographic predictor became insignificant gender (P = . 10, p > .05), age
(P = .06, p > .05), and marital status (p = -.07, p > .05), while depression status (p =
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
69
.25, p < .01) was a significant predictor of weekly recall of psychological distress.
Depression status added an additional 6.20% to the variance in weekly recall of
psychological distress. These results indicate that the demographic differences found
between the two groups did not account for the differences found in recall of weekly
psychological distress whereas depression status does.
Table 7 reports the appraisal differences between depressed and non-
depressed individuals in their reports of their weekly average psychological distress,
peak day distress, end day distress, within person distress variation, recall of
psychological distress, weekly average positive affect, peak day positive affect, end
day positive affect, within person positive affect variation, and recall of positive
affect.
Levels of psychological distress across all time frames was significantly
higher for depressed individuals than for non-depressed individuals. Depressed
individuals experienced significantly higher rates of weekly average psychological
distress (M= 1-35, SD = 0.36) compared to non-depressed individuals (M = 1.13, SD
= 0.19), (F (1, 795) = 82.48, p < .01). Depressed individuals significantly recalled
higher rates of psychological distress (M= 1-56, SD = 0.48) compared to non-
depressed individuals (M= 1-26, SD = 0.31), (F (1, 762) = 82.48, p < .01). Depressed
individuals had higher peak day distress (M = 1.80, SD=0.69) and end day distress
(M = 1.38, SD=0.69) compared to non-depressed individuals peak day
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
70
Table 7
Description of Psychological Distress Variables by Depression Level and Reference
Period
Psychological Variable
Depressed
Mean (SD)
Non-Depressed
Mean (SD) F
Average Weekly 1.35 (0.36) 1.13 (0.19) 82.48**
Distress
Recall Distress 1.56(0.48) 1.26 (0.31) 61.10**
Peak Day Distress 1.80 (0.69) 1.35 (0.42) 79.75**
End Day Distress 1.38 (0.59) 1.12(0.25) 73.21**
Within Person Distress 1.78 (1.50) 0.87 (0.90) 69.88**
Variation
Average Weekly 3.79 (0.80) 4.21 (0.59) 40.05**
Positive Affect
Recall Positive Affect 3.73 (0.78) 4.14(0.60) 34.30**
Peak Day Positive 4.41 (0.71) 4.65 (0.52) 15.67**
Affect
End Day Positive Affect 3.77 (1.06) 4.25 (0.77) 30.08**
Within Person Positive 3.74 (2.44) 2.80 (2.31) 13.43**
Variation
Note. N = 797. (Depressed = 94. Non-depressed = 703.) *p < .05. **p < .01.
(M = 1-35, SD = 0.42) and end day (M = 1.12, SD = 0.25), (F (1, 795) = 79.75, p <
.01 for peak day and F (1, 795) = 73.21, p < .01 for end day). Lastly, depressed
individuals experienced more variation in their distress across the weak (M = 1-78,
SD = 1.50) than non-depressed individuals (M = 0.87, SD = 0.90), (F (1, 795) =
69.89, p < . 01).
Reporting of positive affect across all time frames was significantly lower for
depressed individuals than non-depressed individuals. Depressed individuals
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
71
experienced significantly lower rates of average weekly positive affect (M = 3.79,
SD = 0.80) compared to non-depressed individuals (M = 4.21, SD = 0.59), (F (1,
795) = 40.05, p < .01). Depressed individuals significantly recalled lower rates of
positive affect (M = 3.73, SD = 0.78) compared to non-depressed individuals (M =
4.14, SD = 0.60), (F (1, 795) = 34.30, p < .01). Depressed individuals had lower peak
positive affect (M = 4.41, SD=0.71) and end day positive affect (M = 3.77, SD=1.06)
compared to non-depressed individuals peak day (M = 4.65, SD = 0.52) and end day
(M = 4.25, SD = 0.77), (F (1, 795) = 15.67, p < .01 for peak day and F (1, 795) =
30.08, p < .01 for end day). Lastly, depressed individuals experienced more variation
in their positive affect across the weak (M = 3.74, SD = 2.44) than non-depressed
individuals (M = 2.80, SD = 2.32), (F (1, 795) = 13.43, p < .01).
Accuracy Tests
Hypothesis 1. Accuracy o f Recall o f Psychological Distress
Flypothesis 1 proposed in support of depressive realism that depressed
individuals will accurately recall the amount of average weekly psychological
distress whereas non-depressed individuals will inaccurately recall their average
weekly psychological distress. Two independent sampleJ-tests were performed to
test this hypothesis and the results are displayed in figures 1 and 2. The first test
(Figure 1) explored the difference between weekly average psychological distress
and recall of psychological distress in order to explore accuracy as well as determine
whether the difference between the two groups was due to an underestimation
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
72
(difference between daily psychological distress and recall of psychological distress
is positive) or overestimation (difference between daily psychological distress and
weekly psychological
Figure 1. Difference Between Weekly Psychological Distress and Recall of
Psychological Distress
3.05 i:? J . i l i r
Depressed
— ■
m o im a
,
■ ■ ■ ■ i l l
■ ■ ■ ■ ■ ■ I
°-1 f:?sl!: 7' ■ '
H H I
1 5 1
0.2 -
-0.13
- 0.21
Figure 2. Absolute Difference Between Weekly Psychological Distress and
Recall of Psychological Distress
0.25
Depressed
0.15
Non-Depressed
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
73
distress is negative) in their recall of psychological distress. Results revealed that
both groups tend to overestimate the amount of psychological distress with depressed
individuals (M = --21, SD = .26) having a greater discrepancy between average
weekly psychological distress and recall of psychological distress than non-
depressed individuals (M =-.13, SD = .18), (t (101.48) = 2.89, p < .01). As
depressed individuals display greater within person distress variation than non-
depressed individuals an additional independent samples t-test was conducted to
further insure the robustness of the findings and to insure that the significant results
were not due to the valence of the discrepancy between weekly average distress and
recall distress.
A similar test (Figure 2) was conducted, yet this time exploring the absolute
value of the difference between weekly average psychological distress and recall of
psychological distress. Results strengthened the findings from the first analyses, as
depressed individuals have a greater overall discrepancy (M = .25, SD = .22) than
non-depressed individuals (M = .15, SD = .17), (t (103.17) = -3.98, p < .01).
Hypotheses 2. Accuracy o f Recall o f Positive Affect
Hypothesis 2 proposed in support of depressive realism that depressed
individuals will accurately estimate the amount of positive affect they experienced
during the week whereas non-depressed individuals will overestimate their amount
of positive affect. Two independent sample t-tests were performed to test this
hypothesis and both are provided figures 3 and 4. The first test (Figure 3) explored
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
74
the difference between weekly average positive affect and recall of positive affect in
order to explore accuracy as well as determine whether the difference between the
two groups was due to an underestimation (difference between average weekly
positive affect and recall of positive affect is positive) or overestimation (difference
between average weekly positive affect and recall of positive affect is negative) in
their recall of positive affect. Results
Figure 3. Difference Between Weekly Positive Affect and Recall of Positive
Affect
0.07 0.07
Depressed Non-Depressed
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
75
Figure 4. Absolute Difference Between Weekly Positive Affect and Recall of
Positive Affect
0.42
0.33
Depressed Non-Depressed
revealed that depressed individuals (M= .07, SD = .62) and non-depressed
individuals (M - .07, SD = .50) tend to underestimate the amount of weekly positive
affect in their recall of their week. However, there was no difference in the level of
accuracy between the two groups, (t (104.94) = 0.04, p > .05).
Due to the lack of significant findings, an additional independent samples t-
test was conducted to determine if the varying valences in estimation are what lead
to the lack of significant results. Using the absolute value of the difference between
weekly aggregate positive affect and recall of positive affect (Figure 4), results found
depressed individuals (M = .42, SD = .46) are more inaccurate in their recall of their
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
76
positive affect during the week than non-depressed individuals (M = .33, SD =.37),
recall of positive affect (t (761) =-2.00, p < .05).
Hypothesis 3. Peak End Distress Predicting Recall o f Weekly Psychological
Distress by Depression Status.
Hypothesis 3 proposed in support of notion that depressed and non-depressed
individuals use different processes in their recall of their own psychological distress
that depression status will moderate the relationship between recall of weekly
psychological distress and the valuing of peak psychological distress and end day
psychological distress.
Hierarchical linear regression (see Table 8) was performed to test the interaction of
peak end hypothesis by diagnosis on weekly recall of psychological distress. Step
one had Depression Status, step two had Peak Day Psychological Distress and End
Day Psychological Distress, and step three had Peak Day Psychological Distress X
Depression Status and End Day Psychological Distress by Diagnosis predicting
weekly recall of psychological distress. The first step of the regression analysis
revealed that diagnosis was a significant predictor of weekly recall of psychological
distress ((3 = -.27, p < .01), accounting for 7.40% o f the variance in weekly
recall of psychological distress. On the second step, diagnosis became insignificant
([3 = .04, p > .05) and Peak Day Distress (|3 = .55, p < .01) and End Day Distress (|3 =
.27, p < .01) significantly added to the prediction of weekly recall of psychological
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
77
distress, accounting for an additional 50.90% of the variance beyond the effects of
depression status. On the third step, diagnosis achieved significance again ((3 = .33,
P < .01) while Peak Day Distress ((3 = .54, p < .01) and End Day Distress ([3 = .37, p
Table 8.
Hierarchical Multiple Regression of Diagnosis, Peak Day Distress, Final Day
Distress predicting Weekly Recall of Psychological Distress
Step 1 Step 2 Step 3
b (SE) 3 b (SE) p b (SE) 3
Diagnosis .29 (.04) - .04 (.03) .04 .35 (.08) .33
.27*
*
**
Peak Day .40 (.02) .55** .39 (.03) .54
Distress
**
End Day .30 (.03) .27** .41 (.04) .37
Distress
**
Diagnosis X .00 (.06) .00
Peak Day
Distress
Diagnosis X -.25 (.07) -
End Day .34
Distress
**
R2 7.40 58.30 59.40
* **
Change in 50.90 1.10
R2
Note. N = 799. (Depressed = 94. Non-depressed = 703.) **p < .01. *p < .05.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
78
Table 9.
Multiple Regression of Peak Day Distress, Final Day distress predicting Weekly
Recall o f Psychological Distress for Depressed and Not-Depressed Groups
Depressed Non-Depressed
b (SE) p b (SE) p
Peak Day Distress
End Day Distress
R2
.39 .08 .56**
.16 .09 .19
50.01**
.39 .02 .53**
.41 .04 .32**
58.20**
Note. N = 799. (Depressed = 94. Non-depressed = 703.) **p < .01. *p <.05.
< .01) maintained its significance. The interaction of End Day Distress X Diagnosis
was significant ([3 = -.34, p < .01) while the interaction of Peak Day Distress X
Diagnosis was insignificant ([3 = .00, p > .05) and the additional interactions
contributed 1.10 % of the variance beyond the effects of depression status and peak
psychological distress and end day psychological distress.
As interactions were found between Diagnosis by End Day Distress, separate
linear regressions were performed with depressed and non-depressed individuals to
understand the differences in recall of psychological distress by peak day distress
and end day distress (see Table 9). Recall of psychological distress for non-depressed
people was significantly influenced by both peak day distress ((3 = .53, p < .01) and
end day distress ((3 = .32, p < .01), accounting for 58.20% of the variance in recall of
psychological distress. However, only peak day distress ((3 = .56, p < .01) and not
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
79
end day distress ((3 = .19, £ > .05), was a significant contributor to recall of
psychological distress for depressed individuals, accounting for 50.01% of the
variance in recall of psychological distress.
Figure 5 further illustrates the nature of the interaction of end day distress by
depression status as a separate regression was performed solely with end day distress
and the interaction of end day distress and depression status.
Figure 5. Interaction of End Day Distress By Depression Status
3
( / )
if)
C D
2.5
-+-i
if)
Q
7 5 2
o
5)
o
O 1.5
. C
o
< / >
CL 1
H -
0
8 0 5
01
Low End Distress High End Distress
Level of End Day Distress
Regression lines were plotted representing the prediction of weekly recall of
distress for depressed and non-depressed individuals in low and high end day
conditions. Low end distress was calculated using 2 standard deviations below the
mean of end distress and high end distress was calculated using 2 standard deviations
above the mean of end day distress. As shown, the relationship between end day
Non-depressed
■*— Depressed
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
80
distress and recall of psychological distress changed as a function of depression
status. Non-depressed individuals who experienced low end day distress or high end
day distress recalled greater amounts of that psychological distress than non-
depressed individuals and the magnitude of the difference was greater in the high end
condition. Depressed individuals with high end day distress failed to take in this
information in their recall of psychological distress.
Hypothesis 4. Average Weekly Distress Predicting Recall o f Weekly Psychological
Distress by Depression Status.
Hypothesis 4 proposed in support of notion that depressed and non-depressed
individuals use different processes in their recall of their own psychological distress
that depression status will moderate the relationship between recall of weekly
psychological distress and average weekly psychological distress.
Hierarchical linear regression was performed to test the interaction of average
weekly distress by diagnosis on weekly recall of psychological distress (see Table
10). Step one had Depression Status, step two had Weekly Average Psychological
Distress, and step three had Weekly Average Psychological Distress X Depression
Status predicting weekly recall of psychological distress. The first step of the
regression analysis revealed that diagnosis was a significant predictor of weekly
recall of psychological distress ([3 = .27, p < .01), accounting for 7.40% of the
variance in weekly recall of psychological distress. On the second step, diagnosis
became insignificant ((3 = .02, p > .05) and Weekly Average Distress ((3 = .84, p <
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
81
.01) significantly added to the prediction of weekly recall of psychological distress,
accounting for an additional 64.30% of the variance beyond the effects of depression
status. On the third step, diagnosis achieved significance again ((3 = .22, p < .01) and
Weekly Average Distress ((3 = .87, p < .01) maintained its significance. The
interaction of Weekly Average Distress X Diagnosis was significant ([3 = -.21, p <
.05) and the additional interaction contributed 0.30 % of the variance beyond the
effects of depression status and weekly average distress.
Table 10.
Hierarchical Multiple Regression of Diagnosis. Weekly Average Distress predicting
Weekly Recall of Psychological Distress
Step 1 Step 2 Step 3
b (SE) ( 3 b (SE) [ 3 b (SE) ( 3
Diagnosis .29 (.04) - .02 (.02) .02 .35 (.09) .22**
.27
**
Weekly 1.27 (.03) .84 .39 (.04) .87**
Average
**
Distress
Diagnosis X -.16 (.07) -
Weekly .21**
Average
Distress
R2 7.40 71.70 72.0
** ** o**
Change in 64.30 0.30
R2
Note. N = 799. (Depressed = 94. Non-depressed = 703.) **p < .01. *p < .05.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
82
Table 11.
Multiple Regression of Weekly Average Distress predicting Weekly Recall of
Psychological Distress for Depressed and Not-Depressed Groups
Depressed Non-Depressed
b (SE) ( 3 b (SE) ( 3
Weekly Average
Distress
R2
1.16 .07 .86**
73.20**
1.31 .03 .83**
68.50**
Note. N = 799. (Depressed = 94. Non-depressed = 703.) **p < .01. *p < .05.
As interactions were found between Diagnosis by Weekly Average Distress,
separate linear regressions were performed with depressed and non-depressed
individuals to understand the differences in recall of psychological distress by
weekly average distress (see Table 11). Recall of psychological distress for non-
depressed people were significantly influenced by weekly average distress ((3 = .83,
P < .01), accounting for 68.50% of the variance in recall of psychological distress.
Recall of psychological distress for depressed people were also significantly
influenced by weekly average distress ((3 = .86, p < .01), accounting for 73.20% of
the variance in recall of psychological distress.
Figure 6 further illustrates the nature of the interaction of weekly average
distress by depression status on recall o f psychological distress. Regression lines
were plotted representing the prediction of weekly recall of distress for depressed
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
83
and non-depressed individuals who experienced low and high average daily distress.
Low end distress was calculated using 2 standard deviations below the mean of end
distress and high end distress was calculated using 2 standard deviations above the
mean of end day distress. As shown, the relationship between weekly average day
distress and recall of psychological distress changed as a function of depression
status. Non-depressed individuals in the low average range of weekly distress
accurately recalled less amounts of distress than depressed individuals and similar
amounts of distress in the high daily average range. Depressed individuals in the low
average range of daily distress recalled greater amounts psychological distress than
actually occurred. This may account for why depressed individuals were inaccurate
as they overestimated the amount of psychological distress they experienced in when
they were not experiencing significant amounts o f distress.
Figure 6. Interaction of Weekly Average Distress by Depression Status
Recall of
Psychological
Distress „
Depressed
• —Non-depressed
0.4
0.2
Low Weekly Average High Weekly Average
Distress Distress
Degree of Average Distress
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
84
Hypothesis 5. Peak End Positive Affect Predicting Recall o f Positive Affect
by Depression Status.
Hypothesis 5 proposed in support of notion that depressed and non-depressed
individuals use different processes in their recall of their own positive affect that
depression status will moderate the relationship between recall of weekly positive
affect and the valuing of peak positive affect and end day positive affect.
A hierarchical linear regression was performed to test the interaction of peak
end hypothesis by diagnosis on weekly recall of positive affect (see Table 12). Step
one had Depression Status, step two had Peak Day Positive Affect and End Day
Positive Affect, and step three had Peak Day Positive Affect X Diagnosis and End
Day Positive Affect by Diagnosis predicting weekly recall of positive affect. The
first step of the regression analysis revealed that diagnosis was a significant predictor
of weekly recall of positive affect ((3 = -.21 .01), significantly accounting for
4.30% of the variance in weekly recall of positive affect. On the second step,
diagnosis was still significant ((3 = -.10, p < .01) and Peak Day Positive Affect ((3 =
.17,p < .01) and End Day Positive Affect ((3 = .47, p < .01) significantly added to the
prediction of weekly recall of positive affect, accounting for an additional 33.30% of
the variance beyond the effects of depression status. On the third step, diagnosis was
no longer significant ((3 = -.07, p > .05) while Peak Day Positive Affect ((3 = . 16, p <
.01) and End Day Positive Affect ((3 = .47, p < .01) maintained its significance.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
85
Table 12.
Hierarchical Multiple Regression of Diagnosis, Peak Day Positive Affect. Final Day
Positive Affect predicting Weekly Recall of Positive Affect
Step 1 Step 2 Step 3
b (SE) p b (SE) p b (SE) p
Diagnosis - (-07) - -.20 (.06) -.10** -.13 (.40) -.07
.4 .21
Peak Day .21 (.04) .17** .19 (.05) .16**
Distress
End Day .37 (.03) .27** .40 (.03) .51**
Distress
Diagnosis .12 (.11) .27
X Peak
Day
Distress
Diagnosis
I
L /i
0
1
*
*
X End Day
Distress
R2 4.30 37.60 38.00*
** ** *
Change in 33.30 0.40
R2
Note. N = 799. (Depressed = 94. Non-depressed = 703.) **p < .01. *p < .05.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
8 6
Table 13.
Multiple Regression of Peak Day Positive Affect. Final Day Positive Affect
predicting Weekly Recall of Positive Affect for Depressed and Not-Depressed
Groups
Depressed Non-Depressed
b (SE) 3 b (SE) 3
Peak Day Distress
End Day Distress
R2
.31 .13 .26**
.24 .09 .33*
28.30**
.19 .04 .16**
.40 .03 .50**
36.70**
Note. N = 799. (Depressed = 94. Non-depressed = 703.) **p < .01. *p < .05.
The interaction of End Day Positive Affect X Diagnosis was significant ((3 = -.31, p
< .05) while the interaction of Peak Day Positive Affect X Diagnosis was
insignificant ([3 = .27, p > .05) and the additional interactions contributed .40 % of
the variance beyond the effects of depression status and peak positive affect and end
day positive affect.
As interactions were found between Diagnosis by End Day Positive Affect,
separate linear regressions were performed with depressed and non-depressed
individuals to understand the differences in recall of positive affect by peak day
affect and end day affect (see Table 13). Recall of positive affect for non-depressed
people was significantly influenced by both peak day positive affect ((3 = .16, p <
.01) and end day positive affect ((3 = .50, p < .01), accounting for 36.70% of the
variance in recall of positive affect. Recall of positive affect for depressed
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
87
individuals was also influenced by both the peak day positive affect ((3 = .26, g < .01)
and end day positive affect ((3 = .33, g < .05), however to a different degree. The
differences in degree lead to peak end only accounting for 28.30% of variance for
depressed individuals in recall of positive affect.
Figure 7 further illustrates the nature of the interaction of end day positive
affect by depression status as a separate regression was performed solely with end
day positive affect and the interaction of end day positive affect and depression
status.
Regression lines were plotted representing the prediction of weekly recall of
positive affect for depressed and non-depressed individuals with low and high end
day positive affect. Low end day positive affect was calculated using 2 standard
Figure 7. Interaction of End Day Positive Affect by Level of Depression
Recall of
Positive
Affect
■Depressed
■ Non-Depressed
Low End Day High End Day
Positve Affect Positive Affect
Level of End Day Positive Affect
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
deviations below the mean of end positive affect and high end positive affect was
calculated using 2 standard deviations above the mean of end day positive affect. As
shown, the relationship between end day positive affect and recall of positive affect
changed as a function of depression status. Non-depressed and depressed individuals
who experienced low end day positive affect recalled their low end positive affect to
a similar degree while non-depressed individuals who experienced high positive
affect recalled a greater amount of that positive affect than depressed individuals.
The depressed individuals who experienced high end day positive affect and failed to
recall their high end positive affect may be what lead to non-depressed individuals
being more accurate in their recall of positive affect.
Hypothesis 6. Average Weekly Positive Affect Predicting Recall o f Positive Affect
by Depression Status.
Hypothesis 6 proposed, in support of notion that depressed and non-
depressed individuals use different processes in their recall of their own positive
affect, that depression status will moderate the relationship between recall of weekly
positive affect and the average weekly positive affect.
Hierarchical linear regression was performed to test the interaction of weekly
average positive affect by diagnosis on weekly recall of positive affect (see Table
14). Step one had Depression Status, step two had Weekly Average Positive Affect,
and step three had Weekly Average Positive Affect X Depression Status predicting
weekly recall of positive affect. The first step o f the regression analysis revealed that
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
89
diagnosis was a significant predictor of weekly recall of positive affect ((3 = -.21
.01), accounting for 4.30% of the variance in weekly recall of psychological distress.
On the second step, diagnosis remained significant ((3 = -.07, p < .05) and Weekly
Average Positive Affect ((3 = .66, p < .01) significantly added to the prediction of
weekly recall of positive affect, accounting for an additional 41.40% of the variance
beyond the effects of depression status. On the third step, diagnosis became
insignificant ((3 = -.09, p > .05) and Weekly Average Positive Affect ((3 = .66, p <
.01) maintained its significance. The interaction of Weekly Average Positive Affect
X Diagnosis was insignificant ((3 = .02, p < .05) and there was no additional
contribution to the variance beyond the effects of depression status and weekly
positive affect.
As no interactions were found between depression status and weekly positive
affect, no subsequent regressions needed to be performed.
Table 14.
Hierarchical Multiple Regression o f Diagnosis, Weekly Average Positive Affect
Predicting Weekly Recall of Positive Affect
Step 1 Step 2 Step 3
b (SE) ( 3 b (SE) p b (SE) p
Diagnosis -.41 (.07) -
.21
**
-.13 (.05) -.07* -.16 (.28) -.09
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
90
Table 14 - continued.
Hierarchical Multiple Regression of Diagnosis. Weekly Average Positive Affect
Predicting Weekly Recall of Positive Affect
Step 1 Step 2 Step 3
b (SE) p b (SE) 3 b (SE) 3
Weekly .67 (.03) .66** .67 (.03) .66
Average
**
Positive
Affect
Diagnosis X .01 (.07) .02
Weekly
Average
Positive
Affect
R2 4.30 45.70 45.70*
** **
Change in 41.40 0.00
R2
*
Note. N = 799. (Depressed = 94. Non-depressed = 703.) **p < .01. *p < .05.
Conclusions
Demographic differences were found between depressed and non-depressed
individuals in terms of gender, marital status, and age. In terms of these differences,
the depressed group had a significantly more proportionate amount of women and
less proportionate amount of males than the non-depressed group, a significantly
more proportionate amount of unmarried people and less proportionate amount of
married people than the non-depressed group, and the depressed group was also
significantly younger in age. There were no differences found between the groups in
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
91
terms o f ethnicity, education in years, number of children, and total household
income. Furthermore, the demographic differences found did not significantly
influence recall of psychological distress beyond depression status.
Differences were also found in the reporting of distressing experiences and
weekly recall of those experiences between depressed and non-depressed
individuals. Depressed individuals experienced significantly higher rates of weekly
average psychological distress, experienced a more distressing peak day during the
week as well as ended the week reporting more psychological distress than non-
depressed individuals. Depressed individuals also experienced more variation within
their distress across the week and recalled having experienced higher levels of
psychological distress at the end.
Differences also emerged between depressed and non-depressed individuals
in the reporting of positive affective experiences and weekly recall of those
experiences. Depressed individuals experienced significantly lesser rates of weekly
average positive affect, experienced a lower level of peak positive affect during as
well as ended the week experiencing less positive affect. Depressed individuals did
experience more variation within their positive affect across the week and recalled
having experienced lower levels of positive affect at the end of the week.
Differences did emerge in depressed and non-depressed groups accuracy in
their appraisal o f their own psychological distress. Both depressed and non-
depressed individuals significantly overestimated the amount of psychological
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
92
distress they experienced during the week, with the depressed group having a
significantly higher overestimation of their own distress. Further tests revealed that
the absolute difference between weekly psychological distress and recall of
psychological distress was also significantly higher for depressed individuals than
non-depressed individuals.
Mixed results emerged with depressed and non-depressed groups accuracy in
their appraisal of their own positive affect. Both depressed and non-depressed
individuals tended to underestimate the amount of positive affect they experienced
during the week. However, the first set of analysis failed to demonstrate differences
between the two groups in terms of the amount of discrepancy in the recall of
positive affect. Subsequent analysis that explored the absolute value of the difference
between weekly average positive affect and recall of positive affect did reveal
depressed individuals are more inaccurate in their recall of their positive affect
during the week than non-depressed individuals.
Depressed and non-depressed individuals did use different processes in their
recall of their own psychological distress. Results revealed that in the peak end
appraisal process non-depressed individuals factored in both their peak distressing
period and their current mood in their recall of their own psychological distress that
lead them to be more accurate. However, depressed individuals weighed their peak
experiences more heavily into their recall of their own distress rather than also
valuing their current experience. The lack of utilization o f the end experience in their
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
93
recall may have been what led depressed individuals to be more inaccurate in their
appraisal of their own psychological distress. Results also demonstrated differences
in the average valuing of weekly psychological distress as well. Depressed
individuals seemed to recall more psychological distress in conditions in which they
did not experience intense psychological distress. It is this overestimation of
depressed individuals own psychological distress in these low intensity conditions
that contributed significantly to depressed individuals lack of accuracy.
Mixed results were found in the use of different processes in the recall of
depressed and non-depressed individuals positive affect. Results revealed that in the
peak end appraisal process both depressed and non-depressed individuals factored in
the peak positive affective experience and their current mood in their weekly recall
of their own positive affect. However, each group valued the contribution of the end
affective experience differently. Non-depressed individuals especially those who
experienced high end day positive affect more accurately recalled their high end
positive affect while depressed individuals who experienced high end day positive
affect underestimated the amount of positive affect they experienced. Results failed
to demonstrate that depressed and non-depressed individuals utilized different
processes in the average valuing o f their daily experience in their recall of their
positive affect as their was no interaction of depression status by weekly average
positive affect.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
94
Chapter 5: Discussion
This study explored whether depressive realism exists as a result of depressed
and non-depressed individuals use of different problem solving approaches. It is
proposed that the different problem solving approaches is what leads one group to be
more accurate than the other group in certain situations and less accurate in other
conditions. In order to test this notion, the current study investigated whether
depressed individuals were more accurate in their appraisal of their own
psychological distress and positive affect than were non-depressed individuals.
Whether the effect was found or not, the study then investigated two cognitive
appraisal processes, peak end evaluation and average evaluation, to explore whether
depressed and non-depressed valued different aspects within these processes that
would lead to differential rates in accuracy.
This chapter discusses the results of this study including an analyses of each
hypothesis in the study and its connection with previous literature, a summary of the
findings, a discussion of the validity of recall of average daily ratings as an accuracy
measure, implications of this current study, limitations, and future research stemming
from this study.
Tests o f Hypothesis
Hypothesis 1
Data did not support the prediction that depressed individuals would be more
accurate than non-depressed individuals in their recall of their own psychological
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
95
distress. In fact, data supported the opposite view: both depressed and non-depressed
individuals overestimated the amount of psychological distress that they actually
experienced. Furthermore, depressed individuals overestimated their psychological
distress to a greater degree than their non-depressed counterparts. These findings
support previous depressive realism studies especially in the area o f recall of self
perceived performance (Kendall, Stark, Adam (1990, experiments 1, 2, and 3);
Gotlib, 1983; Gotlib, 1981; Roth & Rehm, 1980). Each of these studies with
clinically depressed samples found that depressed individuals overestimated the
amount of negative feedback or reinforcement they were provided. Regardless of
whether the results support the hypothesis, they provide evidence that clinical
depression may provide one of the conditions that causes individuals to utilize
processes that lead to inaccurate recall.
Hypothesis 2
Data did not support the prediction that depressed individuals are more
accurate than non-depressed individuals in their recall o f their own positive affect.
The first set of analysis showed no differences in accuracy o f positive affect between
depressed and non-depressed individuals. Both groups underestimated the amount of
positive affect. The second set of analysis revealed that the absolute value, that is the
total amount of discrepancy between depressed and non-depressed individuals in
their daily average distress and the recall of their distress, of the difference between
weekly positive affect and recall of positive affect was significantly greater for
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
96
depressed individuals than non-depressed individuals. This means that depressed
individuals had an overall greater amount of deviation from accuracy than non-
depressed individuals.
These findings support previous studies in the self-perceived performance as
well as self-evaluation ranges that found depressed individuals overall underestimate
the amount of positive feedback they received compared to non-depressed
individuals (Albright, Alloy, Barch, & Dykman, 1993; Dykman, Horowitz,
Abramson, & Usher, 1991; Kennedy & Craighead, 1988; Dykman, Abramson,
Alloy, & Hartlage, 1989; Cane & Gotlib, 1985; Dobson & Shaw, 1981, (first 40
trials); Buchwald, 1977; Nelson & Craighead, 1977, (experiment 1)). These studies
corroborated these results across different experimental tasks within these ranges
(Albright, Alloy, Barch, & Dykman, 1993, experiment #2; Dykman, Horowitz,
Abramson, & Usher, 1991; Abramson, Alloy, & Hartlage, 1989; Dykman, Kennedy
& Craighead, 1988; Dobson & Shaw, 1981, (first 40 trials); Buchwald, 1977; Nelson
& Craighead, 1977, (experiment 1)) as well as in social evaluations (Cane & Gotlib,
1985).
Other studies also revealed that the lack of recall of positively valenced
information may be related to depression status, such that more clinically depressed
individuals are more prone to underestimating positive affect than are mildy
depressed individuals (McKendree-Smith and Scogin, 2000; Kendall, Stark, and
Adam, 1990, experiment 3; Dow & Craighead, 1987; Gotlib, 1981). As the
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
97
individuals in this study were in the clinical range of depression, it provides further
supports this notion and the overall premise that depression status can lead to
differential cognitive processes that lead to differential rates of accuracy across
different situations.
Hypothesis 3
Data fully supported the hypothesis that depression status would moderate
the relationship between recall of weekly psychological distress and the valuing of
peak psychological distress and end day psychological distress; that depressed
individuals and non-depressed individuals would use different processes in their
valuing of peak and end experience in their recall of their own psychological
distress. Results revealed that in the peak end appraisal process non-depressed
individuals factored in both their peak distressing period and their current mood in
their recall of their own psychological distress. However, depressed individuals
weighed their peak experiences more heavily into their recall of their own distress
rather than also valuing their current experience. The process that non-depressed
individuals used led to more accurate recall of their psychological distress and led
depressed individuals to overestimate its amount.
These findings shed further light on the previous studies that were able to
demonstrate that depressed individuals recall greater amounts of negatively valenced
information than non-depressed individuals (Kendall, Stark, Adam (1990,
experiments 1, 2, and 3); Gotlib, 1983; Gotlib, 1981; Roth & Rehm, 1980). By
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
98
failing to take into account their current mood in their recall of the experience,
depressed individuals consistently overestimate the amount of negativity they
experience.
Hypothesis 4
Data fully supported the hypothesis that depression status moderates the
relationship between recall of weekly psychological distress and the averaging of
their weekly experience. Results showed that whereas depressed and non-depressed
individuals appraised and recalled highly distressing events in the same manner,
depressed individuals seemed to recall more psychological distress in conditions in
which they did not experience intense psychological distress. Results revealed that it
is this overestimation of depressed individuals own psychological distress in these
low intensity conditions that contributed significantly to depressed individuals lack
of accuracy.
This analysis reveals another possibility as to why there is mixed evidence in
the depressive realism literature surrounding recall of negatively valenced
information. It may be that some of the studies that found depressed individuals to be
more accurate had relied on highly distressing conditions. The evidence from this
study seems to point that this condition is part of the process that would lead
depressed individuals to be more accurate. This notion seems to be supported in the
literature as tasks that had more difficulty and more distress associated with them
produced results of greater or equal amounts of accuracy for depressed individuals
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
99
compared to non-depressed individuals (Dykman & Volpicelli, 1983; Craighead,
Hickey, & DeMonbreun, 1979).
Hypothesis 5
Results fully supported the notion that depression status moderates the
relationship between recall of weekly positive affect and the valuing of peak positive
affect and end day affect. Data demonstrated that in the peak end appraisal process
both depressed and non-depressed individuals had been affected by their the peak
positive affective experience and their current mood in their weekly recall of their
own positive affect. However, each group valued the contribution of the end
affective experience differently which contributed to different rates of accuracy. Data
supported the notion that the difference in accuracy seem to result in recall of end
day positive affect such that non-depressed individuals especially those who
experienced high end day positive affect more accurately recalling their high end
positive affect while depressed individuals who experienced high end day positive
affect underestimating the amount of positive affect they experienced.
These findings are also consistent with one study that explored the effects of
differential rates of positive feedback depending on the rate of positive feedback at
the end of the experiment (DeMonbreum & Craighead, 1977). Demonbreum and
Craighead’s (1977) slide task in which individuals had to perceive the out-of-focus
syllables found that the second 40 trials in which there was a high rate of positive
feedback that depressed individuals significantly underestimated their rate of positive
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
100
feedback. However, depressed individuals that were in end conditions with low rates
of positive feedback, were accurate in their perception of positive feedback they
received.
Taken together, this finding can help explain some of the variability in results
surrounding studies that explored accuracy in recall of positive feedback. It may be
that the studies that found depressive accuracy explored it in conditions in which the
amount of positive feedback at the end of the experiment was low. This seems to
trigger the optimal conditions or cognitive processes for depressed individuals to
accurately perceive their amount of positive feedback. When conditions at the end of
the experiment call for recall of high amounts of positive recall, depressed
individuals seem to underestimate the amount of positive affect.
Hypothesis 6
Data failed to support the hypothesis that depression status moderates the
relationship between recall of weekly positive affect and the averaging of weekly
positive affect. Results seem to indicate that both groups average the amount of
weekly positive affect in the same way which leads to similar rates in recall of
positive affect.
The lack of significant differences in accuracy in recall may partly be due to
lack of differences in accuracy recall between depressed and non-depressed
individuals. Both groups had similar rates of accuracy of recall of positive affect and
differences only emerged using the absolute amount of discrepancy between weekly
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
101
positive affect and daily positive affect. Secondly, it may be that the average valuing
across the seven days is able to mask the differences found in the effects of the end
day on positive affect. As there exists more data points to create the average, the
differences may disappear using the longer time frame.
Summary o f Findings
Overall, the data failed to support the notion of depressive realism, for
depressed individuals were more inaccurate in their recall of their own psychological
distress and positive affect than non-depressed individuals. Although this is not the
first study to reach such conclusions, it is unique in that it demonstrates that
depressed and non-depressed individuals engage in different processes in their
appraisal that led to differential rates of accuracy. In this case, the differential
cognitive processes depressed individuals utilized led them to more inaccurate recall.
The one exception was in the area of average valuing with positive affect.
In the area of recall of psychological distress, depressed individuals in the
peak end analysis did not take into account their end day distress in recalling their
own psychological distress. Rather, they focused on their peak experience in the
week and used this as the basis for recall. This was different than non-depressed
individuals who were affected by both their peak experience and their end
experience. The different processes in recall led to non-depressed individuals more
accurately recalling their psychological distress than non-depressed individuals.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
102
In the area of positive affect, non-depressed and depressed individuals did
weigh in their end day positive affect and their peak day positive affect in their recall
of their positive affect. However, results revealed that each group valued the
contribution of the end affective experience differently. Non-depressed individuals,
especially those who experienced high end day positive affect, more accurately
recalled their high end positive affect whereas depressed individuals who
experienced high end day positive affect underestimated the amount of positive
affect they experienced. This difference could account for the differences in accuracy
evaluations.
The average cognitive process analysis revealed differences between
depressed and non-depressed individuals by type of emotional recall. In the area of
psychological distress, results revealed that whereas depressed and non-depressed
individuals appraised and recalled highly distressing events in the same manner,
depressed individuals seemed to recall more psychological distress in conditions in
which they did not experience intense psychological distress. Results revealed that it
is this overestimation by depressed individuals own psychological distress in these
low intensity conditions that contributed significantly to depressed individuals lack
of accuracy. In the area of positive affect, results revealed that depressed and non-
depressed individuals engaged in similar cognitive processes. This may explain why
depressed and non-depressed individuals demonstrated no significant differences in
accuracy of recall of positive affect without the exploration of absolute accuracy.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
103
Validity o f the Weekly Average as a Measure o f Accuracy
Accuracy or the objective standard was defined as the average of the daily
positive or distressing experiences across the week. This study acknowledges that
there may be other ways to explore accuracy in terms of emotional experiences
(Thomas & Diener, 1990). These authors described how experiences could be
measured both by intensity of emotion or frequency of emotions. Thus, which type
of measure would be more accurate? Further, some philosophers would argue that
there is no objective way to accurately value a person’s subjective experience
(Heidegger, 1967) and fields of philosophy are dedicated to the subject,
phenomenology.
However, the chosen method seems to be the most plausible way to
accurately value subjective experience. Furthermore, using this method, this study
found differences between depressed and non-depressed individuals as well as found
that depressed and non-depressed individuals utilized different cognitive processes
that resulted in differential rates of accuracy of recall.
Implications
Assuming that the criterion task employed in this study is valid, the results of
this study have implications on various theoretical levels. These concern the core
tenets of psychopathology, cognitive theories of depression, the theory o f depressive
realism, and treatment of depression.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
104
One of the core tenets of psychopathology is that any psychological disorder
stems from a distortion of reality (Taylor & Brown, 1988). This study demonstrated
that contrary to the core tenets of psychopathology, both depressed and non-
depressed individuals distorted reality. But it did find that depressed individuals
distorted their reality to a greater degree than non-depressed individuals in both their
recall of positive and negative emotions. Furthermore, this study found that the type
of distortion varied as a function of the valence of the emotion. Depression was not
associated with a unilateral negative distortion. However, depression was associated
with an overestimation of the amount of psychological distress and an
underestimation of the amount of positive affect. Overall, these different types of
distortions would lead to depressed individuals feeling more negatively valenced
emotions.
This study also provides evidence to support Beck’s Cognitive Theory of
Depression that proposes that depression is caused or maintained by negative views
of self, other, and the world (Beck 2000; Beck, Rush, Shaw, & Emory, 1979; Beck,
1967). In this study, depressed individuals displayed negative views of their own
emotional experience such that they recalled greater amounts of distress than they
actually experienced and underestimated the amount of positive affect they had
during the week.
This study also provides evidence to support the notion that rumination may
maintain depression (Nolen-Hoeksema, 2000). Evidence for this comes from the
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
105
peak end analysis in which in each condition depressed individuals focused more on
their peak experience in their recall and failed to take in their end experience. Non-
depressed individuals that focused on the peak and end experience more accurately
recalled their distress and positive affect.
This study provided a new methodology to test depressive realism. As well, it
extended the findings to a more ecologically valid context, it explored the depressive
realism phenomenon with depressed individuals rather than dysphoric or mildly
depressed individuals, and extended the study of depressive realism to a new sample,
adults ages 25 to 76.
This study further extended the depressive realism literature as it provides
evidence that depressive realism may be due to differential cognitive processes
triggered by different conditions. Although this study did not find that depressed
individuals are more accurate than non-depressed individuals in their recall of their
own emotions, it did provide strong evidence that each group engaged in different
cognitive processes that lead to the differential rates of accuracy. This is also
consistent with recent research on depressive realism that found that depressed
individuals do engage in different processes or pay attention to different cues that
lead to differential rates of accuracy (Msetfi, Murphy, Simpson, & Kombrot, 2005).
Lastly, this study has implications for cognitive therapies for depression, as it
provides evidence of the locations of the cognitive distortions that lead depressed
individuals to be more inaccurate than non-depressed individuals. Based on these
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
106
results, cognitive therapists may want to focus on helping depressed individuals take
into account their current experiences in their appraisal and to stop focusing on their
peak experience. Another area on which cognitive therapists may want to focus is
helping depressed individuals recall more precisely their true distress or experience,
especially in cases where there exists minimal amounts of distress.
Limitations
Four limitations were identified in this study and efforts were made to
overcome these limitations when possible. The first limitation had to do with timing
of the diagnosis of depression. As the larger study from which this was drawn was
not specifically designed to be a test of depressive realism, diagnosis of individuals
took place up to one year prior to actual partaking in the diary portion of the study.
As such, there was no precise way to ensure that the individuals diagnosed in the
depressed group were still depressed at the time of the diary portion of the study.
Further, it is impossible to know whether some individuals in the non-depressed
group had become depressed. However, in support of the grouping, results did reveal
substantial differences between the depressed and non-depressed groups in ways that
would be consistent with depression. The depressed group reported higher
psychological distress across all measures and lower amounts of positive affect
across all measures.
Another issue with the diagnosis of depression in this study was that it was a
dichotomous measure that differentiated clinical depression from no depression.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
107
Thus, there was no way to explore depression along a continuum and explore
whether there is a relationship between degree o f depression and degree of accuracy.
Studies seem to indicate that mild depression may be one of conditions that lead to
cognitive processes that promote accuracy. Yet, this study could not explore this
notion and could only substantiate past studies that seem to indicate that major
depression results in cognitive processes that lead to greater psychological
negativity.
Another limitation of the current study had to do with the ability to explore
positive affect. Positive affect was assessed via a one item measure. Although this
one item had been shown to have the greatest item weight in a longer measure of
positive affect from the MIDUS study, there is no way to test the internal consistency
of a one item measure and therefore the validity of positive affect can be in question.
However, even with the one item measure, differences were found between the
depressed and non-depressed groups that were consistent with previous findings
within the depressive realism literature.
A last limitation that was discussed previously surrounds the operational
definition of accuracy. For a discussion of this limitation, please review the section
in the results entitled “Validity of the Weekly Average as a Measure of Accuracy.”
Future Studies
Future studies of depressive realism need to continue the trend of exploring
the process that leads depressed and non-depressed individuals to differential rates of
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
108
accuracy. The exploration, either quantitatively through measures or methodologies
used in this study, or qualitatively, by asking these individuals their reasoning behind
their choices will allow for a reconciliation of differences in the depressive realism
literature. As more studies explore the process, it is likely that they will support this
study’s notion that both depressed and non-depressed individuals can be accurate in
their appraisal in different situations, even though each group utilizes different
processes within the same task that lead to different rates of accuracy.
Additional studies utilizing the daily diary format need to explore different
types of tasks and accuracy. Emotional recall is only one way to explore accuracy yet
daily diary studies can explore other factors including daily time use and recall of
that time use, future time use and whether those predictions come true, as well as
frequency of daily stressors and recall of those stressors. An additional exciting
component of the daily diary format with depressive realism is that it can explore
different lengths of recall and the effect that time has on accuracy of recall. Future
studies should take advantage of the daily diary methodology in the exploration of
depressive realism.
Future studies of depressive realism need to explore whether depressive
realism is a state based phenomenon or trait like phenomenon. No study to date has
followed up with depressed individuals to see if the same accuracy process occurs
when depression lifts. This is an important point to explore as current depression
research is trying to differentiate between state aspects to depression and trait aspects
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
109
of depression in order to understand possible etiology or markers for people who
become depressed as well as factors that could predispose depressed individuals to
relapse Teasdale, Williams, Soulsby, Segal, Ridgeway, & Lau, 2000; Illardi,
Craighead, & Evans 1997; Hedlund & Rude, 1995; Rohde, Lewinsohn and Seely,
1990).
Future studies of depressive realism also should include a mildly depressed or
dysthymic group along with a depressed group and non-depressed group. Moreover,
they should determine depression status just before undergoing the diary portion and
then confirming the same diagnosis after the diary is completed, and use a measure
of positive affect rather than one item. Adding these additional components will
insure the accuracy of these results as well as explore the possibility that depressive
realism is more associated with mild dysphoria as opposed to major depression.
Conclusion
This study employed a novel approach to explore the validity of the
depressive realism theory. As well, it provided a means to explain the current
variability in the depressive realism literature. Using a daily dairy design, this study
failed to support the depressive realism theory as depressed individuals were more
inaccurate in their appraisal of their psychological distress and positive affect than
non-depressed individuals.
This study further showed that part of the inaccuracy was due to depressed
individuals overestimating the amount of psychological distress they experienced in
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
110
their recall of psychological distress to a greater degree than non-depressed
individuals. Depressed individuals also underestimated the amount of positive
feedback they experienced to a greater degree than non-depressed individuals.
This study further provided quantitative cognitive mechanisms to explain the
differential rates of accuracy of depressed and non-depressed individuals. In this
case, the differential valuing within the explored cognitive processes by depressed
individuals led them to more inaccurate recall, except this difference was unable to
be shown in the area of average valuing of positive affect.
Future research needs to continue to explore qualitatively and quantitatively
the processes that lead to differential rates of accuracy. These processes may hold the
key to understanding the phenomenon of depressive realism as well as depression in
general. Furthermore, a clear understanding o f the processes can provide more
insight into relapse prevention and cognitive therapies for depression. This study
provides clues that the effectiveness of cognitive therapy may be in that it helps
depressed people factor in their end experience into their recall of their own
experience which would minimize the amount of distress described by the depressed
individual.
Overall, much more research is needed. Yet this study provides a blueprint
for where future research needs to go to fully understand depressive realism.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Ill
References
Abramson, L. Y., Alloy, L. B. & Rosoff, R. (1981). Depression and the
Generation of Complex Hypotheses in the Judgment of Contingency. Behavioral
Research & Therapy, 19, 35-45.
Ackermann, R. & DeRubeis, R.J. (1991). Is Depressive Realism Real?
Clinicial Psychology Review. 11. 565-584.
Ahrens, A. H., Zeiss, A. M. & Kanfer, R. (1988). Dysphoric Defecits in
Interpersonal Standards, Self-Efficacy, and Social Comparison. Cognitive Therapy
and Research, 12(1), 53-67.
Aiken, L.S., & West, S.G. (1991). Multiple Regression: Testing and
Interpreting Interactions. Thousand Oaks, Ca: Sage Publications.
Albright, J. S., Alloy, L. B., Barch, D. & Dykman, B. M. (1993). Social
Comparison by Dysphoric and Nondysphoric College Students: The Grass Isn’t
Always Greener on the Other Side. Cognitive Therapy and Research, 17(6), 485-509.
Albright, J.S. & Henderson, M.C. (1995) How Real is Depressive Realism?
A Question of Scales and Standards. Cognitive Therapy and Research. 19(5). 589-
609.
Alloy, L. B. & Abramson, L. Y. (1979). Judgment of Contingency in
Depressed and Nondepressed Students: Sadder but Wiser? Journal of Experimental
Psychology: General. 108(4), 441-485.
Alloy, L. B. & Abramson, L. Y. (1982). Learned Helplessness, Depression,
and the Illusion of Control. Journal of Personality and Social Psychology, 42(6),
1114-1126.
Alloy, L. B. & Abramson, L. Y. (1988). Depressive realism: Four theoretical
perspectives. In L. B. Alloy (Ed.), Cognitive Processes in Depression, 223-265. New
York: Guilford Press.
Alloy, L.B., Abramson, L.Y., & Viscusi, D. (1981). Induced mood and the
Illusion of Control. Journal of Personality and Social Psychology, 41(6), 1129-1140.
Alloy, L. B. & Ahrens, A. H. (1987). Depression and Pessimism for the
Future: Biased Use of Statistically Relevant Information in Predictions for Self
Versus Others. Journal of Personality and Social Psychology, 52(2). 366-378.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
112
Alloy, L. B. & Clements, C. M. (1992). Illusion of Control: Invulnerability to
Negative Affect and Depressive Symptoms After Laboratory and Natural Stressors.
Journal of Abnormal Psychology. 101(2). 234-245.
Almeida, D.M. & Mroczek, D.K. (2004). The Effect of Daily Stress,
Personality, and Age on Daily Negative Affect. Journal of Personality. 72(2). 355-
378.
Altmann, E. O. & Gottlib, I. H. (1988). The Social Behavior of Depressed
Children: An Observational Study. Journal of Abnormal Child Psychology, 16(1).
29-44.
American Psychiatric Association. 1987. Diagnostic and Statistical Manual
for Mental Disorders Ill-Revised. American Psychological Association, Washington,
DC.
Beck,A.T. (1967). Depression: Clinical, experimental, and theoretical
aspects. New York: Hoeber.
Beck, A.T., Rush, A.J., Shaw, B.F., & Emery, G. (1979). Cognitive Therapy
of Depression. New York, New York: The Guilford Press.
Beck, A.T. & Weisher, M.E. (2000) Cognitive Therapy. In R.J. Corsini & D.
Wedding (Eds.), Current Psvchotherapies Sixth Edition (pp. 241-272). Itasca, 1 1 1 . :
F.E. Peacock Publishers, Inc..
Benassi, V. A. & Mahler, H. I. M. (1985). Contingency Judgments By
Depressed College Students: Sadder But Not Always Wiser. Journal of Personality
and Social Psychology. 49(5). 1323-1329.
Beyer, S. (2002). The effects of gender, dysphoria, and performance feedback
on the accuracy o f self-evaluations. Sex Roles. 47 (9/10). 453-464.
Bradbum, N.M. (1969). The structure o f psychological well-being. Chicago,
IL: Aldine.
Blazer, D.G., Kessler, R.C., McGonagle, K.A., & Swartz, M.S. (1994). The
prevalence and distribution of major depression in a national community sample:
The National Comorbidity Survey. American Journal of Pyschiatry. 151. 979-986.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
113
Bryson, S. E., Doan, B. D. & Pasquali, P. (1984). Sadder but wiser: A failure
to demonstrate that mood influences judgments of control. Canadian Journal of
Behavioral Science, 16(2), 107-119.
Buchwald, A. M. (1977). Depressive Mood and Estimates of Reinforcement
Frequency. Journal of Abnormal Psychology, 86(4), 443-446.
Campbell, J. D. (1986). Similarity and Uniqueness: The Effects of Attribute
Type, Relevance, and Individual Differences in Self-Esteem and Depression. Journal
of Personality and Social Psychology. 50(2). 281-294.
Cane, B.C. & Gotlib, I.H. (1985). Depression and the effects of positive and
negative feedback on expectations, evaluations, and performance. Cognitive Therapy
and Research, 9(2), 145-160.
Coyne, J.,C. & Gotlib, I., H. (1983). The Role of Cognition in Depression: A
Critical Appraisal. Psychological Bulletin. 94(3). 472-505.
Craighead,W. E., Hickey, K. S., & DeMonbreun, B. (1979) Distortion of
Perception and Recall of Neutral Feedback in Depression. Cognitive Therapy and
Research, 3(3). 291-298.
Crocker, J., Alloy, L., & Kayne, N. (1988). Attributional style, depression,
and perception of consensus of event. Journal of Personality and Social Psychology,
54 15), 840-846.
DeMonbreun, B. G. & Craighead, W. E. (1977). Distortion of Perception and
Recall of Positive and Neutral Feedback in Depression, 1(4), 311-329.
Dobson, K.S. (1989a).Real and Perceived Interpersonal Responses to
Subclinically Anxious and Depressed Targets. Cognitive Therapy and Research,
13(1). 37-47.
Dobson, K.S. (1989b). A meta-analysis of the efficacy of cognitive therapy
for depression. Journal of Consulting and Clinical Psychology. 57(3), 414-419.
Dobson, K.S. & Franche, R.L. (1989). A Conceptual and Empirical Review
of the Depressive Realism Hypothesis. Canadian Journal of Behavioural Science, 21
(4). 419-433.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
114
Dobson, K.S. & Pusch, D. (1995).A Test of the Depressive Realism
Hypthesis in Clinically Depressed Subjects. Cognitive Therapy and Research. 19(2).
179-194.
Dobson, K. S. & Shaw, B. F. (1981). The Effects of Self-Correction on
Cognitive Distortions in Depression. Cognitive Therapy and Research. 514). 391-
403.
Dow, M. & Craighead, W.E. (1987). Social Inadequacy and Depression:
Overt Behavior and Self-Evaluation Process. Journal of Social and Clinical
Psychology. 5(1). 99-113.
Dunning, D. & Story, A. (1991). Depression, Realism, and the
Overconfidence Effect: Are the Sadder Wiser When Predicting Future Actions and
Events? Journal of Personality and Social Psychology. 61(4). 521-532.
Ducharme, J. & Bachelor, A. (1993). Perception of Social Functioning in
Dysphoria. Cognitive Therapy and Research. 17(1). 53-70.
Dykman, B. M. & Volpicelli, J. R. (1983). Depression and Negative
Processing of Evaluative Feedback. Cognitive Therapy and Research, 7(6). 485-498.
Dykman, B. M., Abramson, L. Y., Alloy, L. B., & Hartlage, S. (1989).
Processing of Ambiguous and Unambiguous Feedback by Depressed and
Nondepressed College Students: Schematic Biases and Their Implications for
Depressive Realism. Journal of Personality and Social Psychology, 56(3), 431-445.
Dykman, B. M., Horowitz, L.M., Abramson, L. Y., & Usher, M. (1991).
Schematic and Situational Determinants of Depressed and Nondepressed Students’
Inerpretation of Feedback. Journal of Abnormal Psychology, 100(1). 45-55.
Edison, J. D. & Adams, H. E. (1992). Depression, Self-Focus, and Social
Interaction. Journal of Psychopathology and Behavioral Assessment. 14(1). 1-19.
Elkin, I.S., Watkins, M.T., Imber, J.T., Sotsky, S.M., Collins, J.F., Glass,
D R ., Pilkonis, P.A., Leber, W.R., Docherty, J.P., Fiester, S.J., & Perloff, M.B.
(1989). National Institute of Mental Health treatment of depression collaborative
research program: General effectiveness of treatments. Archives of General
Psvchiatrv,46, 971-983.
Ford, C. E. & Neale, J. M. (1985). Learned Helplessness and Judgments of
Control. Journal of Personality and Social Psychology, 49(5), 1330-1336.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
115
Fredrickson, B.L., & Kahneman, D. (1993). Duration neglect in retrospective
evaluations of affective episodes. Journal of Personality and Social Psychology. 65.
45-55.
Glass, D., McKnight, D., & Valdimarsdottir (1993). Depression, Burnout,
and Perceptions of Control in Hospital Nurses. Journal of Consulting and Clinical
Psychology, 61(1). 147-155.
Golin, S. & Terrell, F. (1977). Motivational and Associative Aspects of Mild
Depression in Skill and Chance Tasks. Journal of Abnormal Psychology. 86(4). 389-
401.
Golin, S. &Terrell, F. & Johnson, B. (1977). Depression and the Illusion of
Control. Journal of Abnormal Psychology. 86(4). 440-442.
Golin, S., Terrell, F, Weitz, J. & Drost, P. L. (1979). The Illusion of Control
Among Depressed Patients. Journal of Abnormal Psychology, 88(4). 454-457.
Gotlib, I. H. (1981). Self-Reinforcement and Recall: Differential Defecits in
Depressed and Nondepressed Psychiatric Inpatients. Journal of Abnormal
Psychology, 90(6). 521-530.
Gotlib, I. H. (1982). Self-Reinforcement and Depression in Interpersonal
Interaction: The Role of Performance Level. Journal of Abnormal Psychology. 91(1).
3-13.
Gotlib, I.H. & Robinson, L.A. (1982). Responses to Depressed Individuals:
Discrepancies Between Self-Report and Observer-Rated Behavior. Journal of
Abnormal Psychology, 91(4), 231-240.
Gotlib, I. H. (1983). Perception and Recall of Interpersonal Feedback:
Negative Bias in Depression. Cognitive Therapy and Research, 7(5), 399-412.
Gotlib, 1 . H. & Meltzer, S. J. (1987). Depression and the Perception of Social
Skill in Dyadic Interaction. Cognitive Therapy and Research, 11(1). 41-54.
Groth-Mamat, G. (2003). Handbook of Psychological Assessment (4th
edition). Hoboken, N.J.: John Wiley & Sons.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
116
Haaga, D.A. & Beck, A.T. (1995). Perspectives on Depressive Realism:
Implications for Cognitive Theory of Depression. Behaviour Research and Therapy.
3311). 41-48.
Haaga, D.A., Dyck, M.J., Ernst, D. (1991). Empirical Status of Cognitive
Therapy of Depression. Psychological Bulletin. 110(2), 215-236.
Hedlund, S., & Rude, S. (1995). Evidence of Latent Depressive Symptoms in
Formerly Depressed Individuals. Journal of Abnormal Psychology. 104(3). 517-525.
Heiby, E. M. (1981). Depression and Frequency of Self-Reinforcement.
Behavior Therapy. 12. 549-555.
Heidegger, M. (1967). On the Essence of Truth. In D. Krell (Ed.), Basic
Writings (pp 11-138). San Francisco: Harper Collins Publishers.
Hoehn-Hyde, D., Schlottmann, R. S. & Rush, A. J. (1982). Perception of
Social Interactions in Depressed Psychiatric Patients. Journal o f Consulting and
Clinical Psychology. 50(2). 209-212.
Hollon, S.D. & Beck, A.T. (2000).Cognitive Therapy. Washington, D.C.:
Oxford University Press.
Illardi, S., Craighead, E., & Evans, D. (1997). Modeling Relapse in Unipolar
Depression: The Effects of Dysfunctional Cognitions and Personality Disorders.
Journal of Consulting and Clinical Psychology. 65(3). 381-391.
Kahneman, D. (1999). Objective happiness. In D. Kahneman & E. Diener,
(eds.) Well-being: The foundations of hedonic psychology (pp 3-25). New York,
NY: Russell Sage Foundation
Kahneman, D., Fredrickson, D.L., Schreiber, C.A., & Redelemeier,
D.A. (1993). When more pain is preferred to less: Adding a better
end. Psychological Science. 4. 401-405.
Kapci, E., Cramer, D. (1998). The Accuracy of Dysphoric and Nondepressed
Groups’ Predictions of Life Events. The Journal of Psychology. 132(6). 659-670.
Kendall, P.C., Hollon, S.D., Beck, A.T., Hammen, C.L., & Ingram, R.E.
(1987). Issues and recommendations regarding use of the Beck Depression
Inventory. Cognitive Therapy and Research. 11. 289-299.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
117
Kendall, P. C., Stark, K. D. & Adam, T. (1990). Cognitive Defecit or
Cognitive Distortion in Childhood Depression. Journal of Abnormal Child
Psychology. 18(3). 255-270.
Kennedy, R. E. & Craighead, W. E. (1988). Differential Effects of
Depression and Anxiety on Recall of Feedback in a Learning Task. Behavior
Therapy. 19, 455-470.
Keller, P.A., Lipkus, I.M., & Rimer, B.K. (2002). Depressive realism and
health risk accuracy: the negative consequences of positive mood. Journal of
Consumer Research, 29. 57-69.
Keyes, C.L.M. & Ryff, C.D. (1998). Generativity and Adult Lives: Social
Structural Countours and Quality of Life Consequences. In D.P. McAdams & E. de
St. Aubin (Eds.) Generativity and Adult Development: How and Why We Care for
the Next Generation.(pp-227-263). Washington, DC: American Psychological
Association.
Kistner, J., Balthazor, M., Risi, S., & David, C. (2001) Adolescents’
perceptions of peer acceptance: is dysphoria associated with greater realism? Journal
of Social and Clinical Psychology, 20 (1), 66-81.
Koenig, L. J., Clements, C. M. & Alloy, L. B. (1992). Depression and the
Illusion of Control. Canadian Journal of Behavioural Science, 24(2), 233-252.
Kuiper, N. A. (1978). Depression and Causal Attributions for Success and
Failure. Journal of Personality and Social Psychology. 3, 236-246.
Lennox, S., Bedell, J.R., Abramson, L.Y., & Foley, F.W. (1990). Judgment
of contingency: A replication with hospitalized depressed, schizophrenic, and normal
samples. Journal of Social Behavior and Personality, 5, 189-204.
Lewinsohn, P. M., Mischel, W., Chaplin, W., & Barton, R. (1980). Social
Competence and Depression: The Role of Illusory Self-Perceptions. Journal of
Abnormal Psychology, 89(2), 203-212.
Lovejoy, M.C. (1991). Maternal depression: effects on social cognition and
behavior in parent-child interactions. Abnormal Child Psychology, 19, 693-706.
Martin, D. J., Abramson, L. Y., & Alloy, L.B. (1984). Illusion of Control for
Self and Others in Depressed and Nondepressed College Students. Journal of
Personality and Social Psychology, 46(1), 125-136.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
118
McNamara, K. & Hackett, G. (1986). Gender, sex-type, and cognitive
distortion: self-perceptions of social competence among mild depressives. Social
Behavior & Personality, 14. 113-121.
McKendree-Smith, N. & Scogin, F. (2000). Depressive Realism: Effects of
Depressive Severity and Interpretation Time. Journal of Clinical Psychology. 56(12).
1601-1608.
McKnight, J.D. & Glass, D. (1995). Perceptions of Control, Bumout, and
Depressive Symptomatology: A Replication and Extension. Journal of Consulting
and Clinical Psychology. 63(3). 490-494.
Mikuliner, M., Gerber, H., & Weisenberg, M. (1990). Judgment of control
and depression: the role of self-esteem threat and self-focused attention. Cognitive
Therapy and Research. 14(6). 589-608.
Mroczek, D. & Kolarz, C.M. (1998). The Effect of Age on Positive and
Negative Affect: A Developmental Perspective on Happiness. Journal of Personality
and Social Psychology. 75(5). 1333-1349.
Msetfi, R.M., Murphy, R.A., Simpson, J., & Kombrot, D.E. (2005).
Depressive Realism and Outcome Density Bias in Contingency Judgements: The
Effect of the Context and the Intertrial Interval. Journal of Experimental
Psychology: General. 134(1). 10-22.
Nelson, R. E. & Craighead, W. E. (1977). Selective Recall of Positive and
Negative Feedback. Journal of Abnormal Psychology. 86. 379-388.
Nolen-Hoeksema, S. (1987) Sex Differences in Unipolar Depression:
Evidence and Theory. Psychological Bulletin. 101(2). 259-282.
Nolen-Hoeksema, S. (2000) The role of rumination in depressive disorders
and mixed anxiety/depressive symptoms. Journal of Abnormal Psychology. 109(3).
504-511.
Pacini, R., Muir, F., & Epstein, S. (1998). Depressive Realism From the
Perspective of Cognitive Experiential Self-Theory. Journal of Personality and Social
Psychology. 74(4). 1056-1068.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
119
Pry, G. & Roth, D. (1977). Depression and Self-reinforcement Behavior in
Hospitalized Patients. Journal of Behavioral Therapy and Experimental Psychology.
8, 31-34.
Pusch, D., Dobson, K.S., Ardo, K., & Murphy, T. (1998). The Relationships
Between Sociotropic and Autonomous Personality Styles and Depressive Realism in
Dysphoric and Nondysphoric University Students. Canadian Journal of Behavioral
Science, 30(4), 253-265.
Radloff, L.S. (1977). The CES-D scale: A self-report depression scale for
research in the general population. Applied Psychological Measurement. 1, 385-401.
Redelmeier, D.A., Kahneman, D. (1996). Patients’ Memories of Painful
Medical Treatments: Real-time and Retrospective Evaluations of Two Minimally
Invasive Procedures. Pain. 66(1). 3-8.
Richters, J. & Pellegrini, D. (1989). Depressed mothers judgments about their
children: an examination of the depression-distortion hypothesis. Child
Development. 60. 1068-1075.
Rohde, P., Lewinsohn, P., & Seeley, J. (1990). Are People Changed by the
Experience of Having an Episode of Depression? A Further Test of the Scar
Hypothesis. Journal of Abnormal Psychology. 99(3). 264-271.
Rosenfarb, I.S., Burker, E.J., Morris, S.A., & Cush, D.T. (1993). Effects of
Changing Contingencies on the Behavior of Depressed and Nondepressed
Individuals. Journal of Abnormal Psychology. 102(4). 642-646.
Roth, D. & Rehm, L. P. (1980). Relationships Among Self-Monitoring
Processes, Memory, and Depression. Cognitive Therapy and Research. 4(2). 149-
157.
Rozensky, R.H., Rehm, L.P., Pry, G., & Roth, D. (1977). Depression and
Self-Reinforcement Behavior in Hospitalized Patients. Journal of Behavior Therapy
& Experimental Psvciatry. 8(1). 35-38.
Shrauger, J.S., Mariano, E., & Walter, T.J. (1998). Depressive Symptoms and
Accuracy in the Prediction of Future Events. Personality and Social Psychology
Bulletin. 24(8). 880-892.
Segrin, C. (2000). Social Skills Deficits Associated with Depression. Clinical
Psychology Review. 20(3). 379-403.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
120
Stehouwer, S. (1985) Beck Depression Inventory. In D. Keyer & R.
Sweetland (Eds.), Test Critiques Volume II (pp. 83-87). Kansas City, Missouri: Test
Corporation of America.
Stone, E.R., Dodrill, C.L., & Johnson, N. (2001). Depressive Cognition: A
Test o f Depressive Realism versus Negativity Using General Knowledge Questions.
Journal of Psychology: Interdisciplinary & Applied. 135(6), 583-602.
Strack, S. & Coyne, J. C. (1983). Social Confirmation of Dysphoria: Shared
and Private Reactions to Depression. Journal of Personality and Social Psychology.
44(4). 798-806.
Tabachnik, N., Crocker, J. & Alloy, L. B. (1983). Depression, Social
Comparison, and the False-Consensus Effect. Journal of Personality and Social
Psychology. 43(3). 688-699.
Taylor, J.A. (1953). A personality scale of manifest anxiety. Journal of
Abnormal and Social Psychology. 48. 285-290.
Taylor, S. E. & Brown, J. D. (1988). Illusion and Well Being: A Social
Psychological Perspective on Mental Health. Psychological Bulletin. 103(2). 193-
210 .
Teasdale, J., Williams, M., Soulsby, J. Segal, Z, Ridgeway, V, & Lau, M
(2000). Prevention of Relapse/Recurrence in Major Depression by Mindfulness-
Based Cognitive Therapy. Journal of Consulting and Clinical Psychology. 68(4).
615-623.
Thomas, D.L. & Diener, E. (1990). Memory Accurcay in the Recall of
Emotions. Journal of Personality and Social Psychology. 59(2). 291-297.
Vasquez, C. (1987). Judgment of Contingency: Cognitive Biases in
Depressed and Nondepressed Subjects. Journal of Personality and Social
Psychology. 52(2). 419-431.
Vestre, N. D. & Caufield, B. P. (1986). Perception of Neutral Personality
Descriptions by Depressed and Nondepressed Subjects. Cognitive Therapy and
Research. 10. 31-36.
Williams, J., Watts, F.,MacLeod, C., & Mathews, A. (1988). Cognitive
Psychology and Emotional Disorders. Oxford, England: John Wiley & Sons.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
121
Wittchen, H.U. (1994). Reliability and validity studies of the WHO
Composite International Diagnostic Interview (CIDI): A critical review. Psychiatric
Research. 28. 57-84.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.

Linked assets

University of Southern California Dissertations and Theses

Conceptually similar

PDF

Gender role conflict, personality, and help -seeking in adult men

PDF

Integration of science and practice: A collective case study of scientist -practitioner programs in counseling psychology

PDF

Development and validation of the Cooper Quality of Imagery Scale: A measure of vividness of sporting mental imagery

PDF

An investigation of a new diagnostic sub-type: Post traumatic stress disorder with psychotic features

PDF

Gender differences in symptom presentation of depression in primary care settings

PDF

Gender differences in motivation for sexual intercourse: Implications for risky sexual behavior and substance use in a university and community sample

PDF

Dementia caregiving and ethnicity: African American caregivers and the sociocultural stress and coping model

PDF

A phenomenological inquiry into the essential meanings of the most intense experience of religiosity and spirituality

PDF

A cluster analysis of the differences between expert and novice counselors based on real time training

PDF

Family functioning, intergenerational conflict, and psychological symptomology of Asian American college students

PDF

Effects of test interpretation style and need for cognition on elaboration, favorability, and recall

PDF

Ideological threat, social identity, and consensus estimation

PDF

A qualitative examination of the nature and impact of three California minority engineering programs

PDF

Course completion and older adult students: The effects of self-efficacy, self -regulation, and motivation

PDF

Collective efficacy, anxiety, creativity/innovation and work performance at the team level

PDF

Border crossing: the experience of international students from Latin America in a California community college

PDF

Assessment of racial identity and self -esteem in an Armenian American population

PDF

Cross -cultural examination of mental health measures in dementia caregivers: Assessment of the Center for Epidemiologic Studies -Depression Scale and the Zarit Burden Inventory

PDF

Heterosexuals' attitudes toward lesbians and gay men and the attitude functions they serve: Correlates, stability and gender differences

PDF

Effects of cultural values and behaviors in sociocultural stress and coping among Korean-American and Korean caregivers

Asset Metadata

Creator Cypers, Scott Jared (author)

Core Title A daily diary approach to compare the accuracy of depressed and nondepressed participants' estimation of positive and negative mood: A test of the depressive realism hypothesis

School Graduate School

Degree Doctor of Philosophy

Degree Program Education (Counseling Psychology)

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

Tag OAI-PMH Harvest,psychology, clinical

Language English

Contributor Digitized by ProQuest (provenance)

Advisor Goodyear (committee chair), Hagedorn, Linda Serra (committee member), Silverstein, Merril (committee member)

Permanent Link (DOI) https://doi.org/10.25549/usctheses-c16-398457

Unique identifier UC11341011

Identifier 3196797.pdf (filename),usctheses-c16-398457 (legacy record id)

Legacy Identifier 3196797.pdf

Dmrecord 398457

Document Type Dissertation

Rights Cypers, Scott Jared

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