1. Depressive Realism: A Meta-Analytic Review Michael T. Moore and David M. Fresco, Kent State University Depressive Realism: A Meta-Analytic Review Michael T. Moore and David M. Fresco, Kent State University ABSTRACT The current investigation represents the first attempt to quantitatively summarize and examine a body of work of substantial importance to cognitive theory of depression (Beck, Rush, Shaw, & Emery, 1979), the depressive realism literature. A search of this literature revealed 118 relevant studies representing 7013 participants from across the US and Canada, as well as from England, Spain, and Israel. Results generally indicated a small overall depressive realism effect, owing perhaps to substantial variability in various study outcomes. Overall, however, both dysphoric/depressed individuals and nondysphoric/nondepressed individuals evidenced a substantial positive bias, with this bias being larger in nondysphoric/nondepressed individuals. Examination of potential moderator variables indicated that studies that lack an objective standard of reality, that utilize self-report to measure symptoms of depression, whose measures are more generalizable, and that use student versus clinical samples were more likely to find depressive realism effects. Methodological paradigm was also found to influence whether results consistent with depressive realism were found. INTRODUCTION Clinicians have often noted anecdotally how depressed individuals seem to view events in their lives more negatively than nondepressed individuals. This observation has been incorporated into theories of depression and plays a prominent role in Beck’s (1987) theory of the etiology of this disorder. However, Beck (1987) characterized depressed individuals very differently from how they are portrayed in the extensive literature on what is referred to as the “depressive realism phenomenon” (see Alloy & Abramson, 1988, for a review). This descriptive and theoretical difference between these two well-validated theories is the subject of the current study. Beck’s theory (1987), which led to the development of cognitive therapy (Beck et al., 1979), posits that depressed affect is heavily influenced by recurrent thoughts with negative content, or automatic thoughts. These thoughts arise from deeply-held dysfunctional beliefs, or schemas, relating to the self, world, and future (e.g., “If I fail, no one will love me). Beck posited that schemas and automatic thoughts, and the depressed affect that results from them, tend to be self-perpetuating as the depressed person both attends more to negative events in their lives and interprets events that occur after the onset of the depressed mood in light of their own dysfunctional cognitions. The “depressive realism hypothesis“ (Alloy & Abramson, 1979) posits not only that depressed individuals can make realistic inferences, but that they do so to a greater extent as compared to nondepressed individuals. Evidence for this phenomenon comes in the form of studies utilizing the “judgment of contingency task”. In this task, participants are asked to press a button, which results in the illumination of a green light a percentage of the time that is predetermined by the experimenter. The dependent variable is the participant-rated perceived contingency between pressing the button and the illumination of the light. Consistent with the depressive realism effect, depressed individuals more accurately make these kinds of judgments as compared to nondepressed individuals (Alloy, Abramson, & Kossman, 1985; Musson & Alloy, 1987; Vazquez, 1987). Other paradigms have also demonstrated the relative accuracy of depressed or dysphoric participants. In tasks designed to measure how evenhandedly participants attend to positively- versus negatively-valenced stimuli, participants recall of task performance feedback, and evaluation of their own performance, depressed or dysphoric individuals have evidenced less bias in their perceptions than nondepressed or nondysphoric individuals (see Alloy & Abramson, 1988 for a review). Despite the research in favor of a depressive realism effect, the literature has been cogently critiqued. Ackermann and DeRubeis (1991) perceptively noted that much of the aforementioned research cannot be said to support depressive realism unequivocally as no objective standard of reality exists with which to compare many of the participants’ rating. Without a “gold-standard” measure of reality, it is theoretically impossible to state that one group or another’s ratings are more or less “realistic”. INTRODUCTION Clinicians have often noted anecdotally how depressed individuals seem to view events in their lives more negatively than nondepressed individuals. This observation has been incorporated into theories of depression and plays a prominent role in Beck’s (1987) theory of the etiology of this disorder. However, Beck (1987) characterized depressed individuals very differently from how they are portrayed in the extensive literature on what is referred to as the “depressive realism phenomenon” (see Alloy & Abramson, 1988, for a review). This descriptive and theoretical difference between these two well-validated theories is the subject of the current study. Beck’s theory (1987), which led to the development of cognitive therapy (Beck et al., 1979), posits that depressed affect is heavily influenced by recurrent thoughts with negative content, or automatic thoughts. These thoughts arise from deeply-held dysfunctional beliefs, or schemas, relating to the self, world, and future (e.g., “If I fail, no one will love me). Beck posited that schemas and automatic thoughts, and the depressed affect that results from them, tend to be self-perpetuating as the depressed person both attends more to negative events in their lives and interprets events that occur after the onset of the depressed mood in light of their own dysfunctional cognitions. The “depressive realism hypothesis“ (Alloy & Abramson, 1979) posits not only that depressed individuals can make realistic inferences, but that they do so to a greater extent as compared to nondepressed individuals. Evidence for this phenomenon comes in the form of studies utilizing the “judgment of contingency task”. In this task, participants are asked to press a button, which results in the illumination of a green light a percentage of the time that is predetermined by the experimenter. The dependent variable is the participant-rated perceived contingency between pressing the button and the illumination of the light. Consistent with the depressive realism effect, depressed individuals more accurately make these kinds of judgments as compared to nondepressed individuals (Alloy, Abramson, & Kossman, 1985; Musson & Alloy, 1987; Vazquez, 1987). Other paradigms have also demonstrated the relative accuracy of depressed or dysphoric participants. In tasks designed to measure how evenhandedly participants attend to positively- versus negatively-valenced stimuli, participants recall of task performance feedback, and evaluation of their own performance, depressed or dysphoric individuals have evidenced less bias in their perceptions than nondepressed or nondysphoric individuals (see Alloy & Abramson, 1988 for a review). Despite the research in favor of a depressive realism effect, the literature has been cogently critiqued. Ackermann and DeRubeis (1991) perceptively noted that much of the aforementioned research cannot be said to support depressive realism unequivocally as no objective standard of reality exists with which to compare many of the participants’ rating. Without a “gold-standard” measure of reality, it is theoretically impossible to state that one group or another’s ratings are more or less “realistic”. HYPOTHESES The current investigation seeks to both quantitatively evaluate the depressive realism literature for the first time and examine several proposed moderators of the depressive realism effect. 1.Consistent with expectations from the depressive realism hypothesis, but counter to expectations from Beck’s theory (1987), effects averaged across studies will show less perceptual/ attentional bias in dysphoric/depressed versus nondysphoric/ nondepressed participants. 2.Study quality will moderate the depressive realism effect. Specifically, studies that utilize an objective standard of reality and address the critique cited by Ackermann and DeRubeis (1991) will evidence larger depressive realism effects than studies that do not. 3.Method of assessment will moderate the depressive realism effect. Specifically, studies that utilize structured clinical interview will produce larger depressive realism effects than studies that utilize self-report. 4.The external validity of the study will moderate the depressive realism effect. Consistent with observations made by Dobson and Franche (1989), studies that are more readily generalizable to settings other than the laboratory will produce smaller differences between dysphoric/depressed and nondysphoric/nondepressed individuals and smaller depressive realism effects. 5.Sample characteristics will serve as a moderator of the depressive realism effect. Specifically, studies that use clinical populations will produce larger depressive realism effects then studies that use university students. RESULTS (CONTINUED) Sample characteristics significantly moderated the depressive realism effect. However, counter to Hypothesis 5, studies using student samples were found to produce mild depressive realism effects whereas studies using clinical samples found results contrary to depressive realism. The methodological paradigm used significantly moderated the depressive realism effect. However we had no specific hypotheses about which method would produce the most robust effects. Studies of attentional bias most strongly illustrated depressive realism, followed by judgment of contingency studies, whereas both recall of feedback and evaluation of performance studies produced results mildly counter to depressive realism. DISCUSSION The results of this first quantitative synthesis of the depressive realism literature were supportive of such an effect when averaging across all studies for which data could be obtained. However, this general effect was moderated by several variables. Depressive realism effects were found most strongly when: the methodological quality of the study was poor the design of the study was more readily generalized outside of the laboratory symptoms of depression were assessed via self-report non-clinical samples were used attentional bias or judgment of contingency paradigms were used Limitations There was a substantial, but randomly distributed, number of studies for which data could not be obtained (45 studies, 38% of the total). Some of the moderator group sizes were small (i.e. clinical interview [n = 6], clinical sample [n = 10], attentional bias studies [n = 4]), owing to a lack of research using particular methods/populations. Future Studies Research in depressive realism should attempt to address the aforementioned lack of studies using structured clinical interview and clinical populations. While there has been a substantial amount of research using particular paradigms (with the exception of attentional bias), there is little consistency in the particular tasks that are used. This operational heterogeneity is likely a contributing factor to the large variability in the effects obtained and lack of consistency in the literature. RESULTS A small depressive realism effect was found when averaged across all studies, (Cohen’s d [1988] = -.24, SD =.72). However, counter to predictions from depressive realism, both dysphoric/depressed and nondysphoric/ nondepressed participants were found to be strongly positively biased, although this bias was larger in the latter group (d =.79, SD = 2.33 and 1.16, SD = 2.43, respectively). Study quality significantly moderated the depressive realism effect (see Table 1 for test statistics), however, counter to Hypothesis 2, studies lower in quality (see Table 1 for means, SD’s, and ranges) found depressive realism effects more strongly. Method of assessment significantly moderated the depressive realism effect, however, counter to Hypothesis 3, studies that utilized self-report found depressive realism effects more strongly than studies utilizing clinical interview. Consistent with Hypothesis 4, the generalizability of the study significantly moderated the depressive realism effect, with more externally valid studies being more likely to produce depressive realism effects. Moderator Test StatisticModerator Group 1Moderator Group 2 Study QualityQ [df = 1] = 39.21, p <.0001High: d = -.46, SD =.92, Range = -.80 to.80 Low: d = -.04, SD =.39, Range = -3.59 to.99 Method of AssessmentQ [df = 1] = 9.99, p =.0016Clinical Interview: d =.01, SD =.52, Range = -.60 to.81 Self-Report: d = -.05, SD =.37 Range = -.80 to.80 External ValidityQ [df = 1] = 39.17, p <.0001High: d = -.08, SD =.38, Range = -.80 to.43 Low: d = -.02, SD =.40, Range = -.70 to.88 Sample CharacteristicsQ [df = 1] = 5.11, p =.0234Clinical: d =.04, SD =.50, Range = -.60 to.88 Student: d = -.07, SD =.35, Range = -.80 to.49 Methodological ParadigmQ [df = 3] = 53.97, p <.0001Attentional Bias: d = -.43, SD =.15, Range = -.60 to -.28 Judgment of Cont: d = -.04, SD =.28, Range = -.52 to.49 Eval. Of Perf: d =.03, SD =.50, Range = -.80 to.88 Recall of Feed: d =.08, SD =.40, Range = -.70 to.81 Table 1