Available online at www.sciencedirect.com
ScienceDirect Comprehensive Psychiatry 58 (2015) 57 – 67 www.elsevier.com/locate/comppsych
Changes of explicitly and implicitly measured self-esteem in the treatment of major depression: Evidence for implicit self-esteem compensation Ingo Wegener a,⁎, Franziska Geiser a , Susanne Alfter a , Jan Mierke b , Katrin Imbierowicz a , Alexandra Kleiman a , Anne Sarah Koch a , Rupert Conrad a a
Department of Psychosomatics, University Hospital Bonn, Sigmund-Freud-Str. 25, 53105 Bonn, Germany b (r)evolution GmbH, Heussallee 12, 53113 Bonn, Germany
Abstract Background and objectives: Self-esteem has been claimed to be an important factor in the development and maintenance of depression. Whereas explicit self-esteem is usually reduced in depressed individuals, studies on implicitly measured self-esteem in depression exhibit a more heterogeneous pattern of results, and the role of implicit self-esteem in depression is still ambiguous. Previous research on implicit selfesteem compensation (ISEC) revealed that implicit self-esteem can mirror processes of self-esteem compensation under conditions that threaten self-esteem. We assume that depressed individuals experience a permanent threat to their selves resulting in enduring processes of ISEC. We hypothesize that ISEC as measured by implicit self-esteem will decrease when individuals recover from depression. Methods: 45 patients with major depression received an integrative in-patient treatment in the Psychosomatic University Hospital Bonn, Germany. Depression was measured by the depression score of the Hospital Anxiety and Depression Scale (HADS-D). Self-esteem was assessed explicitly using the Rosenberg Self-Esteem Scale (RSES) and implicitly by the Implicit Association Test (IAT) and the Name Letter Test (NLT). Results: As expected for a successful treatment of depression, depression scores declined during the eight weeks of treatment and explicit self-esteem rose. In line with our hypothesis, both measures of implicit self-esteem decreased, indicating reduced processes of ISEC. Limitations: It still remains unclear, under which conditions there is an overlap of measures of implicit and explicit self-esteem. Conclusions: The results lend support to the concept of ISEC and demonstrate the relevance of implicit self-esteem and self-esteem compensation for the understanding of depression. © 2014 Elsevier Inc. All rights reserved.
1. Introduction Major depressive disorder is known to be the most prevalent mental disorder. About 5% of the population suffer from depression [1], while approximately one fifth develop one or more episodes of major depression in their life [2]. Self-esteem has been claimed to be an important factor in the development and maintenance of depression [3]. Studies show self-esteem scores to be reduced in depressed individuals [4,5]. The majority of these studies employ self-report questionnaires to measure self-esteem. However, although widely used, self-report data have longtime been
⁎ Corresponding author. Tel.: +49 228 287 16299; fax: +49 228 287 15382. E-mail address:
[email protected] (I. Wegener). http://dx.doi.org/10.1016/j.comppsych.2014.12.001 0010-440X/© 2014 Elsevier Inc. All rights reserved.
criticized for at least two reasons: First, participants have to be motivated to disclose their genuine attitudes, e.g., individuals may give biased answers due to reasons of social desirability. Second, they must also be able to access the construct that they are asked about, e.g., some may misattribute emotions and thus cannot answer questions aiming at the reasons of their emotions correctly. The implicit measurement of an attitude aims at overcoming these limitations by the following principles: (1) the participant's awareness of the fact that this specific attitude is measured may be reduced, (2) the participant may not have conscious access to the attitude in focus, and (3) the participant's control over the outcome of the measurement may be limited [6]. Thus, “implicit measures might be less biased by deliberate attempts to conceal the attitude and that they might even reflect attitudes of which the respondent is not aware” (p. 401) [7].
58
I. Wegener et al. / Comprehensive Psychiatry 58 (2015) 57–67
Although measures of implicit self-esteem were repeatedly criticized to have only low to modest reliability and insufficient convergent validity [8,9], several studies obtained results indicating that implicit self-esteem comprises information that goes beyond the information provided by explicit measures of self-esteem [10,11]. Bosson et al. [8] concluded that among measures of implicit self-esteem, reliability was best for the Implicit Association Test (IAT) [12] and the Name-Letter Test (NLT) [13,14]. Subsequently, the majority of empirical research relies on these measures [9] (see method section for a detailed description of both measures). Dual process accounts to cognitive vulnerability to depression [15,16] suggest that two processing systems determine how an event is interpreted. The automatic or implicit system operates by automatically activating memory concepts in an effortless and unintentional manner and without charging cognitive resources. The deliberate or explicit system is characterized by effortful, intentional resource consuming processing. Haeffel et al. [16] assume that life events first trigger a rapid, automatic, and unintentional response that may activate negative self-schemas. In a second step this interpretation may be reinterpreted by explicit processes. Whereas some researchers presume that the main cause of cognitive vulnerability lies within negative implicit self-schemas [17], other investigators assume that explicit cognition can be the source of vulnerability to depression [18]. Haeffel et al. [16] tested the predictions of dual process models and observed in their first study, that only participants with lower implicit self-worth (IAT) experienced immediate emotional distress after a failure feedback. In their second study, using a prospective design, they found that implicit self-worth (IAT) as well as explicit cognitive styles interacted with negative life events in predicting later depression. When entered simultaneously into a regression model, only explicit self-worth interacted significantly. The authors argue that implicit self-worth affects immediate distress, whereas explicit cognitions determine the long-term risk to depression. Also supporting the role of explicit processing, Steinberg et al. [19] found implicit self-esteem as measured by IAT to predict depressiveness only for individuals with depressogenic cognitive style. Taking a closer look at implicit self-esteem in depressed individuals reveals a mixed picture (see DeHart et al. [20] for a more detailed review): De Raedt et al. [21] observed implicit self esteem in currently depressed individuals to be as high as in non-depressed controls using the IAT (study 1) and the NLT (study 2). In study 3 they observed higher implicit self-esteem for depressed compared to nondepressed participants using the Extrinsic Affective Simon Task (EAST, cf. De Houwer [22]). Accounting for suicidal ideation in depressed individuals, Franck et al. [23] find lower implicit self-esteem (IAT) in a depressed sample without suicidal ideation than in a non-depressed group as well as in depressed individuals with suicidal ideation. Implicit self-esteem of depressed with suicidal ideation was as high as in the non-depressed group. Unfortunately,
suicidal ideation was not controlled in other studies on implicit self-esteem in depressed samples and as a consequence we do not know to what extent the results are affected by suicidal ideation. Gemar et al. [24] observed higher implicit self esteem (IAT) in formerly depressed than in never depressed and currently depressed participants, but after negative mood was induced, implicit self-esteem of the formerly depressed dropped to the level of the never depressed and the currently depressed. This pattern of results was replicated by Franck et al. [25]. However, Franck et al. [26] observed no differences in implicit self-esteem using the NLT in currently, formerly, and never depressed individuals. Nevertheless, implicit self-esteem predicted future depressive symptomatology. Risch et al. [27] accounted for the number of depressive episodes and observed that implicit self-esteem (IAT) was the same for never depressed and remitted patients with recurrent depressive episodes, which both had higher implicit self-esteem than first-onset depressive patients and currently depressive patients with recurrent depressive episodes. Taken together, the most frequent finding is that currently depressed and non-depressed samples do not differ with respect to implicit self-esteem [21,25,26], although implicit self-esteem has also been found to be reduced [27] or elevated [21] in currently depressed individuals compared to non-depressed. Furthermore, implicit self-esteem can be moderated by suicidal ideation [23] and history of depression in depressed [27] or remitted [27,24,25] patients, but see Franck et al. [26] for contrary results. Moreover, implicit self-esteem in remitted patients is affected by mood induction [24,25]. Despite the stable finding of reduced explicit self-esteem in depressed populations, reduced implicit self-esteem seems not reliably associated with clinical depression in cross-sectional studies. A possible explanation for these heterogeneous results may be compensatory responses to self-threatening situations, as will be outlined in the following paragraphs. Based on research demonstrating that implicit self-esteem improves after participants have been confronted with selfthreatening situations, several authors proposed that selfdefensive processes can be triggered under conditions of threat to the self and that these processes can compensate potential loss in self-esteem [10,28,29]. They argue that measures of implicit self-esteem are particularly sensitive to these compensatory processes. The studies that focused on the consequences of self-threatening situations in healthy individuals used different ways to manipulate self-threat. For example, Jones et al. [29] observed enhanced implicit self-esteem after participants with high explicit self-esteem wrote about an aspect of themselves they wish to change but not when they were asked to write about a positive aspect of themselves or a self-irrelevant topic. Other studies yielded analogous effects on the liking of others with similar names [30] or choosing brands with names resembling their own name [31] after participants had performed a self-threatening writing task. Improved implicit self-esteem was also observed after negative life events [28] and social rejection [10].
I. Wegener et al. / Comprehensive Psychiatry 58 (2015) 57–67
Rudman et al. [10] propose an automatic self-esteem defense mechanism called implicit self-esteem compensation (ISEC) that automatically regulates self-esteem whenever the self is seriously threatened. They argue that given the high frequency of threats to the self in everyday life, the idea of a process that buffers self-esteem would be compelling. They propose that ISEC is an automatic self-regulatory mechanism that countervails self-decrementing situations in an effortless manner. This account may explain why implicit self-esteem is not reduced in depressed samples as reviewed above. Based on the findings that healthy individuals employ this mechanism when confronted with ego-threats, it makes sense to assume a similar mechanism in depressed individuals: depressed individuals reveal low explicit selfesteem [4,5], adverse self-schemas [32–34] and overestimate the frequency of future life events [35]. They are prone to disadvantageous attribution styles [36] as they attribute negative life events internally but positive life events externally (see Sweeney et al. [37] for a meta-analytic review of 104 studies) thereby experiencing more selfthreatening events [38–40]. Given this daily struggle with threatening events on the one hand, and the results of elevated implicit self-esteem after manipulations of threats to the self [10,29] on the other hand, we expect implicit self-esteem compensation to be a most relevant process in depression. ISEC may thus be a reason, why depressed individuals reveal consistently reduced levels of explicit but not implicit self-esteem as reviewed above. The heterogeneous results concerning the relative extent of implicit self-esteem of depressed in comparison to healthy individuals may be a consequence of ISEC protecting the depressed self from adverse threats. Implicit self-esteem compensation may often reach or sometimes even prevail the level of healthy controls. Within the dual process framework ISEC can be understood as a mechanism stabilizing implicit self-esteem. In this conception implicit self-esteem may fluctuate in response to self-threatening events and thus exhibit a state component [28]. In dual process accounts for the vulnerability to depression, implicit self-esteem is often understood as rapid and automatic processing mode that develops over a long period [15] and is based on stable memory constructs [16]. However, there is evidence that implicit self-esteem can be affected by classical conditioning [41], induction of negative mood [24,25], daily negative events [28], or after threats to the self [10,28–31]. DeHart and Pelham [28] argue that implicit self-esteem has trait as well as state aspects and propose that implicit self-esteem has a trait level around which the state level can fluctuate. If depressed individuals possess a low level of implicit self-esteem, ISEC may represent such a fluctuation that can elevate implicit self-esteem to a level observed in non-depressed controls. This may explain, for example, why depressed individuals with suicidal ideation, but not without suicidal ideation, revealed higher implicit self-esteem than non-depressed [23]. Self-threat is presumably extremely pronounced in individuals with suicidal ideation and, as a
59
consequence, ISEC may yield particularly high levels of implicit self-esteem. Furthermore ISEC may also explain the frequent finding that implicit self-esteem is generally positive, even in samples with major depression. For example the performance of currently depressed participants in the IAT reveals stronger associations for self-positive and other-negative than for self-negative and other-positive combined trials [21,27]. Also the NLT is positive for depressed individuals as they prefer letters from their own name over other letters [21,26]. Given the more negative selfschemas that are observed in depressed individuals, one may expect negative scores for these measures of implicit self-esteem. The frequently observed positive implicit self-esteem in depressed samples may also be a consequence of ISEC, because depressed individuals are in particular need of processes that compensate the omnipresent threats to their self. In the present study we examine the role of implicit self-esteem compensation for depressed individuals by focusing on the change of explicit and implicit self-esteem during the course of treatment of inpatients with major depression. Because many studies demonstrated that explicit self-esteem improved as a consequence of the successful therapy of depression [4,5] we accordingly expect to replicate this pattern in the present study (hypothesis 1). If, in contrast to explicit self-esteem, implicit self-esteem compensation is protecting the self of depressed individuals against daily threats as suggested above, one would suppose a particularly high implicit self-esteem at the beginning of treatment which, as protective processes decrease during recovery, decreases as patients recover from depression. We thus expect implicit self-esteem to decline during the course of treatment (hypothesis 2). 2. Method 2.1. Participants 51 consecutive inpatients of the Psychosomatic University Hospital Bonn, Germany, diagnosed with major depression participated in the present study. Main reason for hospitalization was limited functioning in everyday life that did not improve sufficiently during out-patient treatment. All had been diagnosed by an experienced psychotherapist of the Psychosomatic University Hospital Outpatient Unit in an approximately two hour diagnostic and anamnestic interview using DSM-IV-TR [42] diagnostic criteria. Included were participants aged from 20 to 65 years. Exclusion criteria were a comorbid psychotic episode or limited German language skills that were insufficient to fill in the questionnaires. However, no participants had to be excluded. Patients were asked to complete the Implicit Association Test [12] on a computer and afterwards to fill in several questionnaires for a first time during the first week and for a second time during the last week of their inpatient treatment (see below). During the treatment, six patients dropped out. In two cases, patients left hospital a few days
60
I. Wegener et al. / Comprehensive Psychiatry 58 (2015) 57–67
before the date of planned discharge due to organizational reasons. In one case the treatment was discontinued because the patient with comorbid anorexia nervosa was not fulfilling her weight gain contract for more than three weeks. One patient discontinued the treatment because of difficulties in emotion regulation when confronted with annoying behavior of some other patient. One patient had somatic complications and one patient discontinued her treatment when she realized that a former coworker of her had been admitted to the treatment program. At the beginning of the treatment, there were no differences between drop outs and completers with respect to age (t(50) = 0.02, p = .98), gender (χ 2 (1) = 0.27, p = .61), depression measured by HADS-D score (t(49) = 1.77, p = .08), explicit self-esteem (t(49) = 0.03, p = .98), or implicit self-esteem (t(50) = 0.28, p = .78 for the IAT score and t(48) = 1.60, p = .12 for the Name Letter Test). Among the 45 participants who completed questionnaires at the beginning and end of treatment there were 36 females (80%) and 9 males (20%). Average age was 35.24 years (SD = 11.73), 27 (60%) had completed high school, seven (16%) had a university degree. At the entry of the treatment 13 (29%) received no psychotropic medication, 23 (51%) received antidepressants, 10 (22%) sedatives, and 4 (9%) antipsychotics. Psychiatric comorbidities were personality disorders (n = 20), eating disorders (n = 15), anxiety disorders (n = 9), somatoform disorders (n = 4), and posttraumatic stress disorder (n = 1). 2.2. Measures 2.2.1. Rosenberg Self-Esteem Scale (RSES) Explicit self-esteem was measured using the German translation by von Collani and Herzberg [43] of the Rosenberg Self-Esteem Scale [44]. Von Collani and Herzberg criticized the psychometric properties of one item of an earlier German version of the RSES [45]. They translated this item more closely to Rosenberg's English version and obtained an improved corrected item-totalcorrelation for this item. The German RSES comprises 10 items that are 4-point Likert-scaled from 0 (does not apply at all) to 3 (applies completely). Von Collani and Herzberg observed Cronbach's α of .84 and .85 in two independent samples. In the present study, we observed a Cronbach's α of .87 at the beginning as well as at the end of treatment. 2.2.2. Hospital Anxiety and Depression Scale (HADS) Depressive and anxious symptomatology was assessed using the German version [46] of the Hospital Anxiety and Depression Scale [47]. Respondents are asked to indicate how they have been feeling over the past week on a 4-point Likert scale. The questionnaire consists of a 7-item subscale measuring anxious and a 7-item subscale measuring depressive symptoms. The latter scale was used in the analyses reported in the present study. Based on a review of 747 studies using the HADS, it has been stated that the two factor structure was supported by the data and that reliability,
sensitivity, and specificity were sufficient [48]. In the present study Cronbach's α of .79 for anxiety and .82 for depression were observed at the beginning of the treatment, whereas the Cronbach's α were .81 and .86 for anxiety and depression, respectively, at the end of the treatment. 2.2.3. Implicit Association Test (IAT) The Implicit Association Test for self-esteem was developed by Greenwald and Farnham [12] as an implicit measure of self-esteem on the basis of the procedure of the standard IAT [49]. In this computerized speed categorization task, the participants' task is to assign words of the categories self versus other and positive versus negative by pressing one of two response keys. Usually participants perform better in compatible trials, where self-related words (e.g., “me”, “my”) are mapped on the same key as positive words (e.g., “luck”, “pleasure”) and other-related words (e.g., “it”, “the”) on the same key as negative words (e.g., “poison”, “betrayal”). If self-related words are assigned to the same key as negative words and other-related words to the same key as positive words, participants perform worse (incompatible trials). It is assumed that the faster reaction times for compatible than for incompatible trials reflect stronger associations for the materials mapped on the same keys in the compatible trials. If the self is associated with positive rather than negative concepts, as one would expect in case of positive self-esteem, participants should answer faster if they are required to press the same key for “me” and “pleasure” (compatible trial) than for “me” and “betrayal” (incompatible trial). In order to reduce the influence of training effects, Greenwald [49] suggested using different kinds of blocks consisting of training trials. In the present study the trials were presented in blocks of 48 trials each. It was balanced between patients whether they had to work on compatible or incompatible blocks first. The IAT procedure consisted of 672 trials grouped in five phases, including training blocks: 1. Two blocks with self- and other-related words only, 2. Two blocks with positive and negative words only, 3. Four mixed blocks (these blocks were all compatible or incompatible, depending on balancing condition), 4. Two blocks with self- and other-related words only, but with switched key assignment, 5. Four mixed blocks using this switched assignment. The dependent variable was calculated from the two phases with mixed blocks (phases 3 and 5) according to the algorithm suggested by Greenwald et al. [50], whereas phases 1, 2, and 4 served as training blocks only. Stimulus words were presented in black letters on a light gray background in a rectangle of 20 mm × 120 mm on a 17 inch display. Response stimulus interval was set to 250 ms. The response keys were to be pressed by different hands, and matching of stimulus categories to the dominant versus non-dominant hand was counterbalanced across patients. Responses were accepted as soon as a stimulus was visible. The IAT is the most frequently used measure of implicitly measured self-esteem [9]. Although studies on the IAT were
I. Wegener et al. / Comprehensive Psychiatry 58 (2015) 57–67
quite promising at the beginning [12,51], more recent overviews compiled more mixed results [9,52]. 2.2.4. Name-Letter Test (NLT) The Name-Letter Test [13,14] is another frequently used task that aims at assessing self-esteem implicitly. The idea behind this measure is that a person's self-esteem affects the extent of liking the letters of one's own name. Indeed it is a stable finding that people like their initials more than other letters [53]. Although it has been questioned if this preference can serve as a measure of self-esteem [9,53], there is evidence that NLT comprises information that goes beyond explicit self-esteem [28,54,55]. In the present study participants were asked to evaluate all letters of the alphabet with regard to how much they like each letter of the alphabet on a 9-point Likert scale. Letters were presented in randomized order that was generated for every patient and for each time of measurement anew. To avoid artifacts because letters that are generally evaluated more favorable may be more frequently found as name letters, the evaluation of a participant's name letter is usually corrected by calculating the difference of this evaluation and the average evaluation of this letter by all other participants that do not have this letter in their name. However, Albers et al. [56] argued and demonstrated that this score is still confounded with the general liking of letters of the participant. They propose an alternative name-letter score that is calculated as the average evaluation of someone's names letters minus the weighted average evaluation of letters that are not part of the person's name and the weighted average evaluation of the same letters rated by participants not possessing these letters in their names. Albers et al. [56] found significant correlations of this score with the Rosenberg Self-Esteem Scale and with internal but not external narcissism. In the present study, we use this improved scoring algorithm of the Name-Letter Test. 2.3. Treatment conditions The mean duration of the inpatient treatment in the present study was 55.69 days (SD = 10.76). Patients participated four times a week in a 1.5 hours psychodynamic group therapy. In addition they received one individual psychodynamic therapy session and two 1.5 hours group sessions of concentrative movement therapy every week. Concentrative movement therapy aims at making biographical material topic by using experiences emerging from movement work. It is based on psychodynamic and Gestalt principles. Furthermore, patients joined a weekly cognitivebehavioral role-play group and a weekly psychodynamic art therapy session lasting 1.5 hours each. Finally, they participated in progressive muscle relaxation groups twice a week. Four patients with panic disorder and/or agoraphobia and one patient with a specific phobia in addition joined a manual based cognitive-behavioral therapy with ten individual sessions including exposure therapy focusing panic symptoms and avoidant behavior. The four patients with
61
anorexia nervosa additionally approved a weight gain contract and had to maintain a food diary. Similar integrative treatment conditions are common in German psychosomatic hospitals and have repeatedly been shown to provide good effectiveness [57,58]. Moreover, the specific treatment conditions at the Psychosomatic University Hospital Bonn were demonstrated to be effective [59,60]. 2.4. Procedure All patients participated during the first and the last week of their in-patient treatment at the Psychosomatic University Hospital, Bonn, Germany, on a voluntary basis. Patients were first administered an Implicit Association Test for self esteem [12] on a computer and afterwards they were asked to fill in several questionnaires including Rosenberg SelfEsteem Scale [43] and the Name-Letter-Test [13,14]. The sessions took about two hours at beginning as well as at end of treatment. 2.5. Statistical analysis For the statistical analyses IBM SPSS version 20 for Windows was used. Alpha error level was set to .05. Power analyses with g*power [61,62] for the matched t-tests used below with a sample size of 45 to test for medium effects resulted in a power of (1 − β) = .95. Because one patient did not fill in the HADS properly at the end of treatment and one other patient did not complete the ratings for the NLT at the end of treatment, the sample size is accordingly diminished for the statistical tests on these variables. 2.6. Ethical statement All patients participated on a voluntary basis and signed informed consent forms. Permission to conduct the study was granted by the ethics committee of the University of Bonn, Germany (Lfd.Nr. 175/09).
3. Results Comparisons of explicitly and implicitly measured selfesteem as well as depression at beginning and completion of treatment are presented in Table 1. Differences were tested by matched pairs t-tests revealing significantly increasing explicit self-esteem, while implicit self-esteem and depression decrease during the course of treatment. Spearman's Rho correlations between the two points of measurement indicate comparably high stability of implicitly measured self-esteem, implying some extent of reliability inherent in these measures. Note that the means of IAT and NLT scores are significantly larger than zero at both times of measurement (all p's b .001), indicating a positive implicit self-esteem. This means, e.g. for the Name Letter Test, that patients evaluated the letters of their own name generally more positive than did patients that did not have these letters
62
I. Wegener et al. / Comprehensive Psychiatry 58 (2015) 57–67
Table 1 Means, standard deviations, matched pairs t-statistics, effect sizes, and correlations of depression, explicit, and implicit self-esteem. Beginning of treatment
Explicitly measured depression HADS-D Explicitly measured self-esteem RSES Implicitly measured self-esteem IAT NLT
End of treatment
Comparisons of beginning vs. end of treatment
Mean
SD
Mean
SD
T-value
Cohen's d
10.73
4.39
5.95
4.17
7.11⁎⁎⁎
1.07
13.47
6.97
18.78
6.29
−4.64⁎⁎⁎
−0.69
0.64 2.84
0.38 0.98
0.55 1.95
0.32 0.93
2.15⁎ 6.04⁎⁎⁎
0.32 0.91
Spearman's Rho (t1-t2) .42⁎⁎ .30⁎ .62⁎⁎⁎ .59⁎⁎⁎
RSES = Rosenberg Self-Esteem Scale; IAT = Implicit Association Test; NLT = Name Letter Test; HADS-D = Hospital Anxiety and Depression Scale – depression score; t-tests were conducted with df = 43 for NLT and HADS-D and df = 44 for all other variables. ⁎ p b .05. ⁎⁎ p b .01. ⁎⁎⁎ p b .001.
in their names, but this positive evaluation decreased across the course of treatment. The Spearman's Rho correlations of the three measures of self-esteem are presented in Table 2 for both points of measurement separately. Neither the correlations of the two implicitly measured self-esteem scores nor the correlations of implicitly and explicitly measured self-esteem are statistically significant. However, correlations of explicit self-esteem and depression scores reveal a strong negative relationship mirroring that the more depressed individuals reported lower explicit self-esteem, whereas no significant correlations were observed for depression scores and implicitly measured self-esteem.
4. Discussion Current psychological theory defines self-esteem as a person's overall appraisal of his or her own worth. “The self-concept is what we think about the self; self-esteem, the positive or negative evaluation of the self, is how we feel about it” (p. 107) [63]. Since psychological theory distinguishes between automatic and unconscious self-evaluation on the one hand and a conscious and more reflective self-evaluation on the other it is useful to distinguish between two different constituents of self-esteem [64]: firstly, implicit self-esteem involving mainly automatic processes [65–67] and secondly,
explicit self-esteem, involving higher cognitive processes [68,69]. Thus, self-esteem is built on (at least) two modes of information processing [70,71]. Implicit associations with the self are thought to be more primitive and develop earlier compared to their explicit counterparts, being closely connected to early social interactions [8,54,67]. Obviously, cognitive restructuring during psychotherapy primarily addresses reflective self-evaluation, which involves higher cognitive processes [72–74] and is closely associated with explicit self-esteem [68,69]. To shed more light on factors that affect explicit and implicit self-esteem, the present study examined the changes of explicitly and implicitly measured self-esteem across the course of an inpatient treatment in individuals with major depression. This specific treatment has been demonstrated to reduce symptom distress in depressed individuals [59,60]. In line with these previous results, we observed significantly decreasing HADS-D scores across the course of treatment in the present study. Consequently, we were able to study changes in self-esteem over an eight week time interval with significantly decreasing depressiveness. We used the Rosenberg Self-Esteem Scale as an explicit measure of selfesteem and, both, the Implicit Association Test for selfesteem and the Name Letter Test as implicit measures of self-esteem. All three measures are most frequently used in self-esteem research [9]. As self-esteem is an important factor in the development and maintenance of depression [3], our first hypothesis stated
Table 2 Correlations of explicit and implicit self-esteem and depression scores. Correlations at the beginning of treatment
RSES IAT NLT HADS-D
Correlations at the end of treatment
RSES
IAT
NLT
HADS-D
RSES
IAT
NLT
HADS-D
1.00
.03 1.00
.18 .09 1.00
−.55⁎⁎⁎ .13 .05 1.00
1.00
.03 1.00
.21 .00 1.00
−.63⁎⁎⁎ −.10 −.19 1.00
RSES = Rosenberg Self-Esteem Scale; IAT = Implicit Association Test; NLT = Name Letter Test; HADS-D = Hospital Anxiety and Depression Scale – Depression Score. ⁎⁎⁎ p b .001.
I. Wegener et al. / Comprehensive Psychiatry 58 (2015) 57–67
that explicit self-esteem will rise during the successful treatment of depression. In support of this hypothesis, we observed a significant increase of explicit self-esteem over time. Furthermore, we argued that (1) increased implicit self-esteem has been observed after manipulations of threat [10,29], which has been proposed to be a result of implicit self-esteem compensation [10] and (2) depressed individuals are prone to dysfunctional self-schemas [32,33] and attribution styles [36,37] which is why everyday life bears frequent threats to their self. Based on this reasoning we hypothesized that a reduction of depression during treatment is accompanied by declining processes of implicit selfesteem compensation. In line with this second hypothesis, both implicit measures of self-esteem decreased significantly across the course of treatment. These results support the notion that ISEC may buffer depressed individuals' self-esteem. As reviewed above, depressed individuals reveal lower explicit, but equal implicit self-esteem in comparison to non-depressed controls [21,25,26], although also higher [21] or lower [27] implicit self-esteem has been observed. ISEC can account for these unexpected and heterogeneous results as it may boost the state component of implicit self-esteem in depressed to a level comparable to non-depressed. Depending on the intensity of the threat as well as the level of depression, it may not reach or even pass by the level of implicit self-esteem in non-depressed. For these studies demonstrating a dissociation of implicit and explicit self-esteem in depressed but not in non-depressed using a cross-sectional design, one cannot rule out that sample characteristics not directly linked to depression may be responsible for differences in depressed and non-depressed. Because the present study is first to demonstrate a dissociation of explicit and implicit self-esteem depressed patients in one single sample over time, we can exclude that differences in sample characteristics are mainly responsible for the observed dissociation. Note that both implicit scores were positive at the beginning and end of treatment, indicating a generally positive implicit self-esteem. In case of the IAT these positive scores denote that patients made fewer mistakes when positive and self-related words as well as negative and other-related words were combined than in conversed assignments. For the Name Letter Test the positive values indicate that patients evaluated letters that are part of their name more positive than do other patients that do not have these letters in their name. These results are in line with previous studies that found positive values of implicit self-esteem in IAT [21–27] and NLT [21,26] and may also be a consequence of ISEC. However, in the context of contradicting results of previous studies it is necessary to draw a more elaborate picture of the possible mechanisms of ISEC, which can be located within the framework of dual process account. ISEC is conceptualized as an automatic and unconscious reaction to severe challenges to self-esteem [10]. That means, this
63
reaction is not activated in any case, but only in cases, when a perception of a serious self-threat makes it necessary. To put it in other words there is an individually specific threshold, above which the initiation of ISEC takes place. The extent of subjectively experienced self-threat is determined by several factors: First, the external situation, which can be perceived as secure and supporting or challenging and hostile [75]. Thus, for example, at the beginning of a psychiatric inpatient treatment patients often experience therapeutic circumstances as far more challenging and threatening as in the further course of treatment. Second, the global cognitive vulnerability of an individual, which is determined by the mental representation of explicit and implicit attitudes [76]. The implicit attitudes are mainly automatic and affect driven, the explicit attitudes belong to the rational system being deliberative and duty driven [70,77,78]. Both systems interact and influence behavior to maintain the adequate level of self-esteem. This highly individual level depends on the stability of explicit self-esteem, the awareness of current, past or future threats to self-esteem, the management of previous self-threats and the associated implicit affective experiences. The likelihood of feeling seriously challenged will be greater if explicit self-esteem is fragile [10,79,80], more threatening situations have been experienced or are anticipated, and management has been unsuccessful. Third, the type of self-threat, which may be relevant or irrelevant for subjective self-schemas. For example, an individual defining him- or herself mainly by academic performance will be more vulnerable to poor grades compared to an individual, defining him- or herself by physical appearance [81]. The complex interplay between these factors determines, whether the critical threshold is reached and ISEC is initiated. The results of Franck et al. [23] corroborate this model as depressed individuals with suicidal ideation, that we would expect to exhibit stronger ISEC, revealed higher implicit self-esteem than those without suicidal ideation. Furthermore, in remitted depressed patients [25], the past experience of low self esteem, poor self esteem management, and the anticipation of possible recurrence can enhance the likelihood of ISEC. On the other hand, the experience of three or more depressive episodes [27] may lead to a change of implicit appraisal, in the sense that even though several threatening situations have been experienced, repeated successful management makes it less threatening and activation of ISEC less likely. The decrease of implicit self-esteem after mood induction in formerly depressed patients [25] can be explained by the fact that the recall of sad experiences is not synonymous with the recall of self-threatening experiences. Distraction by other experiences may disengage ISEC.
64
I. Wegener et al. / Comprehensive Psychiatry 58 (2015) 57–67
Our findings can also be seen within the framework of dual-process account without the assumption of implicit self-esteem compensation. Dual-process accounts usually assume that explicit processing is resource consuming, but that the processing rules can be learned relatively fast, whereas implicit processing is effortless but has to be acquired via repetitive experiences [15,16,77]. While research has focused on the implicit and explicit vulnerability to depression, little is known about the contributions of these processes in the treatment of and recovery from depression. If explicit strategies are mainly responsible for recovery from depression, explicit self-esteem should rise during the treatment, while implicit self-esteem should rise if a change in the implicit self-schema is responsible for improvements. In this line of reasoning, the present results can be interpreted as supporting the role of explicit processing in the recovery from depression, because explicit, but not implicit self-esteem is rising across the course of therapy. If changes in implicit processing indeed require repetitive experiences, the present results suggest that much more experiences have to be made in order to improve implicit self-schemata than our eight week treatment could provide. Still, in the present study, implicit self-esteem did not only fail to increase, but it decreased significantly. Such impairments in implicit processing may be a consequence of changes in well-learned implicit memory structures. Indeed it has been demonstrated that learning changed implicit representations usually leads to reduced output [82,83]. As an example, imagine changes to a well learned mostly automatic and effortless process like driving a car. If you are used to drive a car with manual gear shift, changing to a car with automatic gearbox may temporarily cause difficulties, deceleration, and mistakes, although an automatic gear system may be regarded as easier to learn for a beginner. Likewise, a depressed individual that makes experiences of self-efficacy due to therapeutic interventions may have to integrate this new information in a less favorable implicit self-schema. This may cause a weakening of old representations in the beginning yielding impaired performance in tasks like the IAT or NLT. Thus one may expect that positive experiences can cause changes to a less positive implicit self-concept thereby interfering with the well learned more automatic processing until the changes have been consolidated. In the light of this interpretation, it is primarily explicit processing that is responsible for the immediate recovery during the treatment of depression, while implicit processing may possess relevance for later phases of stabilization. This alternative explanation may be regarded as more parsimonious because it does not require the assumption of self-defensive needs or processes. On the other hand, this account has difficulties explaining positive implicit self esteem in depressed individuals, which has been found in several studies [21,24–26]. With regard to our findings, some methodological issues should be discussed. Neither at the beginning nor at the end of treatment any of the three measures of self-esteem correlated significantly with each other. This complies with
the results of meta-analyses that find very small to nonexistent correlations between implicit and explicit measures of self-esteem as well as between different measures of implicit self-esteem [8,9]. After all, the strong negative correlation of explicit self-esteem with depression corroborated the validity of the Rosenberg Self-Esteem Scale. Although we did not find significant correlations of implicit and explicit measures, we observed strong correlations across time for each of the implicit measures which lends support to a reasonable extent of reliability and stability. An alternative explanation for the decreasing implicit self-esteem may be that reduced depression in our study did influence processing speed which may have influenced the reaction times in the IAT. However, this explanation cannot account for the results of the Name-Letter Task. In conclusion it is rather unlikely that implicit and explicit measures mirror the same aspects of self-esteem, because our data showed a clear dissociation of both measures. If indeed implicit measures of self-esteem are affected by processes of self-esteem compensation, one cannot expect a strong correlation with explicit self-esteem. An individual with high self-esteem in a situation that does not threaten the self will reveal high explicit but low implicit self-esteem. If this individual is exposed to a self-threatening situation, the threat to self-esteem will automatically be compensated for, and high implicit and high explicit self-esteem will be observed. Despite the dissociation of implicit and explicit measures of self-esteem, we do not assume that both procedures are affected by completely disjunct processes, because some studies find significant correlations between both approaches. Moreover, explicit measures of self-esteem have been demonstrated to increase under conditions of threat [84–86], but also the reversed pattern was observed [87,88]. There seems to be a considerable overlap of, but also significant differences between implicit and explicit measures of self-esteem. More research is needed to specify, which processes affect each of the two measures under which conditions. A major limitation of the present work is the absence of a non-depressed control group. Hence we do not know whether implicit and explicit self-esteem is higher, lower or on an equal level compared to a healthy control group. Nevertheless, the dissociation of implicit and explicit measures of self-esteem remains a significant and valid result, even without control group. A second limitation is that we did not use a measure of self-threat that would allow to corroborate our assumption of elevated self-threat in our depressed sample. However, it is very difficult to measure self-threat in all its significant facets, as it embraces relevant aspects not consciously retrievable. A third limitation is that we failed to observe a significant correlation of the two measures of implicit self-esteem. Also other studies using more than one measure of implicit self-esteem found that these different measures of implicit self-esteem were affected similarly, although they have repeatedly been demonstrated to be unrelated [42,89,90]. One may speculate that different
I. Wegener et al. / Comprehensive Psychiatry 58 (2015) 57–67
implicit measures of self-esteem tap different aspects of implicit self-esteem or that methodological differences of the tasks (e.g., fast vs. slow evaluations in IAT vs. NLT) hinder correlations to reach significance, especially in smaller samples.
5. Conclusions The findings underline the importance of implicit selfesteem compensation in depressed patients as a relevant implicit mechanism for stabilization that may undergo specific changes during treatment. Within the dual process account for cognitive vulnerability in depression, ISEC can explain previously not well understood findings such as a higher implicit self-esteem in depressed patients. Measures of implicit self-esteem may help to gain more insight into these processes of protecting self-esteem, thereby hopefully paving the way to optimize the treatment of depression.
References [1] Murphy JM, Laird NM, Monson RR, Sobol AM, Leighton AH. A 40year perspective on the prevalence of depression: The Stirling County Study. Arch Gen Psychiatry 2000;57:209-15. [2] Kessler RC, Berglund P, Demler O, Jin R, Merikangas KR, Walters EE. Lifetime prevalence and age-of-onset distributions of DSM-IV disorders in the National Comorbidity Survey Replication. Arch Gen Psychiatry 2005;62:593-602. [3] Sowislo JF, Orth U. Does low self-esteem predict depression and anxiety? A meta-analysis of longitudinal studies. Psychol Bull 2013;139:213-40. [4] Battle J. Relationship between self-esteem and depression. Psychol Rep 1978;42:745-6. [5] Orth U, Robins RW. Understanding the link between low self-esteem and depression. Curr Dir Psychol Sci 2013;22:455-60. [6] De Houwer J. What are implicit measures and why are we using them? In: Wiers RW, & Stacy AW, editors. Handbook of implicit cognition and addiction. Thousand Oaks: Sage; 2006. p. 11-28. [7] De Houwer J, De Bruycker E. The implicit association test outperforms the extrinsic affective Simon task as an implicit measure of interindividual differences in attitudes. Br J Soc Psychol 2010;46:401-21. [8] Bosson JK, Swann WB, Pennebaker JW. Stalking the perfect measure of implicit self-esteem: the blind men and the elephant revisited? J Pers Soc Psychol 2000;79:631-49. [9] Buhrmester MD, Blanton H, Swann Jr WB. Implicit self-esteem: nature, measurement, and a new way forward. J Pers Soc Psychol 2011;100:365-85. [10] Rudman LA, Dohn MC, Fairchild K. Implicit self-esteem compensation: automatic threat defense. J Pers Soc Psychol 2007;93:798-813. [11] Spalding LR, Hardin CD. Unconscious unease and self-handicapping: behavioral consequences of individual differences in implicit and explicit self-esteem. Psychol Sci 1999;10:535-9. [12] Greenwald AG, Farnham SD. Using the Implicit Association Test to measure self-esteem and self-concept. J Pers Soc Psychol 2000;79:1022-38. [13] Nuttin JM. Narcissism beyond Gestalt and awareness: the name letter effect. Eur J Soc Psychol 1985;15:353-61. [14] Nuttin JM. Affective consequences of mere ownership: the name letter effect in twelve European languages. Eur J Soc Psychol 1987;17:381-402. [15] Beevers CG. Cognitive vulnerability to depression: a dual process model. Clin Psychol Rev 2005;25:975-1002.
65
[16] Haeffel GJ, Abramson LY, Brazy P, Shah J, Teachman B, Nosek B. Explicit and implicit cognition: a preliminary test of a dual-process theory of cognitive vulnerability. Behav Res Ther 2007;45:1155-67. [17] Abramson LY, Alloy LB, Hankin BL, Haeffel GJ, MacCoon DG, Gibb BE. Cognitive vulnerability-stress models of depression in a self-regulatory and psychobiological context. In: Gotlib IH, & Hammen CL, editors. Handbook of depression. New York: Guilford; 2002. p. 268-94. [18] Scher CD, Ingram RE, Segal ZV. Cognitive reactivity and vulnerability: Empirical evaluation of construct activation and cognitive diatheses in unipolar depression. Clin Psychol Rev 2005;25:487-510. [19] Steinberg JA, Karpinski A, Alloy LB. The exploration of implicit aspect of self-esteem in vulnerability-stress models of depression. Self Identity 2007;6:101-17. [20] DeHart T, Peña R, Tennen H. The development of explicit and implicit self-esteem and their role in psychological adjustment. In: Zeigler-Hill V, editor. Self-esteem: current issues in social psychology. New York: Psychology Press; 2013. p. 99-123. [21] De Raedt R, Schacht R, Franck E, De Houwer J. Self-esteem and depression revisited: implicit positive self-esteem in depressed patients? Behav Res Ther 2006;44:1017-28. [22] De Houwer J. The extrinsic affective Simon task. Exp Psychol 2003;50:77-85. [23] Franck E, De Raedt R, Dereu M, Van den Abbeele D. Implicit and explicit self-esteem in currently depressed individuals with and without suicidal ideation. J Behav Ther Exp Psychiatry 2007;38:75-85. [24] Gemar MC, Segal ZV, Sagrati S, Kennedy SJ. Mood-induced changes on the Implicit Association Test in recovered depressed patients. J Abnorm Psychol 2001;110:282-9. [25] Franck E, De Raedt R, De Houwer J. Activation of latent self-schemas as a cognitive vulnerability factor for depression: the potential role of implicit self-esteem. Cogn Emot 2008;22:1588-99. [26] Franck E, De Raedt R, De Houwer J. Implicit but not explicit selfesteem predicts future depressive symptomatology. Behav Res Ther 2007;45:2448-55. [27] Risch AK, Buba A, Birk U, Morina N, Steffens MC, Stangier U. Implicit self-esteem in recurrently depressed patients. J Behav Ther Exp Psychiatry 2010;41:100-206. [28] DeHart T, Pelham BW. Fluctuations in state implicit self-esteem in response to daily negative events. J Exp Soc Psychol 2007;43:157-65. [29] Jones JT, Pelham BW, Mirenberg MC, Hetts JJ. Name letter preferences are not merely mere exposure: implicit egotism as selfregulation. J Exp Soc Psychol 2002;38:170-7. [30] Jones JT, Pelham BW, Carvallo M, Mirenberg MC. How do I love thee? Let me count the Js: implicit egotism and interpersonal attraction. J Pers Soc Psychol 2004;87:665-83. [31] Brendl CM, Chattopadhyay A, Pelham BW, Carvallo M. Name letter branding: valence transfers when product specific needs are active. J Consum Res 2005;32:405-15. [32] Beck AT. Depression: clinical, experimental, and theoretical aspects. New York: Harper & Row; 1967. [33] Beck AT. Cognitive models of depression. J Cogn Psychother 1987;1:5-37. [34] Halvorsen M, Wang CE, Richter J, Myrland I, Pedersen SK, Eisemann M, et al. Early maladaptive schemas, temperament and character traits in clinically depressed and previously depressed subjects. Clin Psychol Psychother 2009;16:394-407. [35] Strunk DR, Lopez HL, DeRubeis RJ. Depressive symptoms are associated with unrealistic negative predictions of future life events. Behav Res Ther 2006;44:875-96. [36] Abramson LY, Metalsky FI, Alloy LB. Hopelessness depression: a theory based subtype of depression. Psychol Rev 1989;96:358-72. [37] Sweeney PD, Anderson K, Bailey S. Attributional style in depression a meta-analytic review. J Pers Soc Psychol 1986;50:974-91. [38] Crocker J, Karpinski A, Quinn DM, Chase S. When grades determine self-worth: consequences of contingent self-worth for male and female engineering and psychology majors. J Pers Soc Psychol 2003;85:507-16.
66
I. Wegener et al. / Comprehensive Psychiatry 58 (2015) 57–67
[39] Crocker J, Sommers S, Luhtanen R. Hopes dashed and dreams fulfilled: contingencies of self-worth in the graduate school admissions process. Pers Soc Psychol Bull 2002;28:1275-86. [40] Ingram RE, Ritter J. Vulnerability to depression: cognitive reactivity and parental bonding in high-risk individuals. J Abnorm Psychol 2000;109:588-96. [41] Dijksterhuis A. I like myself but I don't know why: enhancing implicit self-esteem by subliminal evaluative conditioning. J Pers Soc Psychol 2004;86:345-55. [42] American Psychiatric Association. Diagnostic and Statistical Manual of Mental Disorders – DSM-IV-TR. Washington: American Psychiatric Association; 2000. [43] von Collani G, Herzberg PY. Eine revidierte Fassung der deutschsprachigen Skala zum Selbstwertgefühl von Rosenberg (A revised version of the German adaptation of Rosenberg's Self-Esteem Scale). Z Diff Diagn Psychol 2003;24:3-7. [44] Rosenberg M. Society and the adolescent self-image. Princeton: Princeton University Press; 1965. [45] Ferring D, Filipp SH. Messung des Selbstwertgefühls: Befunde zu Reliabiliät, Validität und Stabiliät der Rosenberg- Skala. Diagnostica 1996;42:284-92. [46] Herrmann C, Buss U, Snaith RP. HADS-D – Hospital Anxiety and Depression Scale – Deutsche Version: Ein Fragebogen zur Erfassung von Angst und Depressivität in der somatischen Medizin. Bern, Switzerland: Huber; 1995. [47] Zigmond SA, Snaith RP. The hospital anxiety and depression scale. Acta Psychiatr Scand 1983;67:361-70. [48] Bjelland I, Dahl AA, Haug TT, Neckelmann D. The validity of the Hospital Anxiety and Depression Scale: an updated literature review. J Psychosom Res 2002;52:69-77. [49] Greenwald AG, McGhee DE, Schwartz JLK. Measuring individual differences in implicit cognition: the Implicit Association Test. J Pers Soc Psychol 1998;74:1464-80. [50] Greenwald AG, Nosek BA, Banaji MR. Understanding and using the implicit association test: 1. An improved scoring algorithm. J Pers Soc Psychol 2003;85:197-216. [51] Greenwald AG, Banaji MR, Rudman LA, Farnham SD, Nosek BA, Mellott DS. A unified theory of implicit attitudes, stereotypes, selfesteem, and self-concept. Psychol Rev 2002;109:3-25. [52] Hofmann W, Gawronski B, Gschwendner T, Le H, Schmitt M. A metaanalysis on the correlation between the Implicit Association Test and explicit self-report measures. Pers Soc Psychol Bull 2005;31:1369-85. [53] Krizan Z, Suls J. Are implicit and explicit measures of self-esteem related? A meta-analysis for the Name–Letter Test. Pers Individ Differ 2008;44:521-31. [54] Koole SL, Pelham BW. On the nature of implicit self-esteem: the case of the name letter effect. In: Spencer JS, Fein S, Zanna MP, & Olson JM, editors. Motivated social perception: the Ontario symposium, vol. 9. Mahwah: Lawrence Erlbaum; 2003. p. 93-116. [55] Shimizu M, Pelham BW. The unconscious cost of good fortune: implicit and explicit self-esteem, positive life events, and health. Health Psychol 2004;23:101-5. [56] Albers LW, Rotteveel M, Dijksterhuis AJ. Towards optimizing the Name Letter Test as a measure of implicit self-esteem. Self Identity 2009;8:69-77. [57] Haase M, Frommer J, Franke GH, Hoffmann T, Schulze-Muetzel J, Jäger S. From symptom relief to interpersonal change: treatment outcome and effectiveness in inpatient psychotherapy. Psychother Res 2008;18:615-24. [58] Salzer S, Streeck U, Jaeger U, Masuhr O, Warwas J, Leichsenring F, et al. Der Zusammenhang von interpersonalen Subtypen und der Veränderung interpersonaler Probleme am Beispiel stationärer Psychotherapiepatienten (The relationship between interpersonal subtypes and the modification of interpersonal problems in psychotherapy patients). Z Psychosom Med Psychother 2010;56:191-206. [59] Geiser F, Imbierowicz K, Conrad R, Liedtke R. “Turning against self” and its relation with symptom distress, interpersonal problems and
[60]
[61]
[62]
[63] [64]
[65] [66]
[67]
[68]
[69] [70]
[71] [72]
[73] [74]
[75] [76]
[77]
[78]
[79] [80]
[81]
[82] [83]
therapy outcome. A replicated and enhanced study. Psychother Res 2005;15:357-65. Wegener I, Alfter S, Geiser F, Liedtke R, Conrad R. Schema change without schema therapy: the role of early maladaptive schemata for a successful treatment of major depression. Psychiatry 2013;76:1-17. Faul F, Erdfelder E, Buchner A, Lang A-G. Statistical power analyses using G*Power 3.1: tests for correlation and regression analyses. Behav Res Methods 2009;41:1149-60. Faul F, Erdfelder E, Lang A-G, Buchner A. G*Power 3: a flexible statistical power analysis program for the social, behavioral, and biomedical sciences. Behav Res Methods 2007;39:175-91. Smith ER, Mackie DM. Social psychology. Philadelphia: Psychology Press; 2007. Zeigler-Hill V. Discrepancies between implicit and explicit selfesteem: implications for narcissism and self-esteem instability. J Pers 2006;74:119-44. Greenwald AG, Banaji MR. Implicit social cognition: attitudes, selfesteem, and stereotypes. Psychol Rev 1995;102:4-27. Koole SL, Dijksterhuis A, van Knippenberg A. What's in a name: implicit self-esteem and the automatic self. J Pers Soc Psychol 2001;80:669-85. Pelham BW, Hetts JJ. Implicit and explicit personal and social identity: toward a more complete understanding of the social self. In: Tyler T, & Kramer R, editors. The psychology of the social self. Mahwah: Lawrence Erlbaum; 1999. p. 115-43. Brown JD. Self-esteem and self-evaluation: feeling is believing. In: Suls J, editor. Psychological perspectives on the self, vol. 4. Hillsdale: Erlbaum; 1993. p. 27-58. Kernis MH. Toward a conceptualization of optimal self-esteem. Psychol Inq 2003;14:1-26. Smith ER, DeCoster J. Dual-process models in social and cognitive psychology: conceptual integration and links to underlying memory systems. Pers Soc Psychol Rev 2000;4:108-31. Wilson TD, Lindsey S, Schooler TY. A model of dual attitudes. Psychol Rev 2000;107:101-26. Garratt G, Ingram RE, Rand KL, Sawalani G. Cognitive processes in cognitive therapy: evaluation of the mechanisms of change in the treatment of depression. Clin Psychol Sci Pract 2007;14:224-39. Linden DE. How psychotherapy changes the brain—the contribution of functional neuroimaging. Mol Psychiatry 2006;11:528-38. Linden DE. Brain imaging and psychotherapy: methodological considerations and practical implications. Eur Arch Psychiatry Clin Neurosci 2008;258:71-5. Rudman LA. Sources of implicit attitudes. Curr Dir Psychol Sci 2004;13:80-3. Rudman LA, Fairchild K. Reactions to counterstereotypic behavior: The role of backlash in cultural stereotype maintenance. J Pers Soc Psychol 2004;87:157-76. DeCoster J, Banner MJ, Smith ER, Semin GR. On the inexplicability of the implicit. Differences in the information provided by implicit and explicit tests. Soc Cogn 2006;24:5-21. Smith ER, DeCoster J. Associative and rule-based processing: a connectionist interpretation of dual-process models. In: Chaiken S, & Trope Y, editors. Dual process theories in social psychology. New York: Guilford; 1999. p. 323-36. Roberts JE, Monroe SM. A multidimensional model of self-esteem in depression. Clin Psychol Rev 1994;14:161-81. Roberts JE, Monroe SM. Vulnerable self-esteem and depressive symptoms: prospective findings comparing three alternative conceptualizations. J Pers Soc Psychol 1992;62:804-12. Rudman LA, Goodwin SA. Gender differences in automatic ingroup bias: why do women like woman more than men like men. J Pers Soc Psychol 2004;87:494-509. Ikier S, Yang L, Hasher L. Implicit proactive interference, age, and automatic versus controlled retrieval strategies. Psychol Sci 2008;19:456-61. Kushner M, Cleeremans A, Reber A. Implicit detection of event interdependencies and a PDP model of the process. Proceedings of the
I. Wegener et al. / Comprehensive Psychiatry 58 (2015) 57–67 13th Annual Conference of the Cognitive Science Society. Hillsdale: Erlbaum; 1991. p. 215-20. [84] Baumeister RE, Jones EE. When self-presentation is constrained by the target's knowledge: consistency and compensation. J Pers Soc Psychol 1978;36:608-18. [85] Greenberg J, Pyszczynski T. Compensatory self-inflation: a response to the threat to self-regard of public failure. J Pers Soc Psychol 1985;49:273-80. [86] Pyszczynski T, Greenberg J, Solomon S, Arndt J, Schimel J. Why do people need self-esteem? A theoretical and empirical review. Psychol Bull 2004;130:435-68.
67
[87] Leary MR, Baumeister RF. The nature and function of self-esteem: sociometer theory. In: Zanna MP, editor. Advances in experimental social psychology, vol. 32. San Diego: Academic Press; 2000. p. 1-62. [88] Williams KD, Cheung CKT, Choi W. Cyberostracism: effects of being ignored over the Internet. J Pers Soc Psychol 2000;79:748-62. [89] Baccus JR, Baldwin MW, Packer DJ. Increasing implicit self esteem through classical conditioning. Psychol Sci 2004;15:498-502. [90] Pelham BW, Koole SL, Hardin CD, Hetts JJ, Seah E, DeHart T. Gender moderates the relation between implicit and explicit selfesteem. J Exp Soc Psychol 2005;41:84-9.
