Self-esteem and depression revisited: Implicit positive self-esteem in depressed patients?

https://doi.org/10.1016/j.brat.2005.08.003Get rights and content

Abstract

The cognitive behavioural model of depression holds that negative cognitions related to the self have etiological importance for the maintenance and relapse of depression. This has been confirmed by research using questionnaires. Recent research using the Implicit Association Test, however, showed positive implicit self-esteem in formerly depressed participants, even after negative mood induction [Gemar, Segal, Sagrati, & Kennedy (2001). Mood-induced changes on the implicit association test in recovered depressed patients. Journal of Abnormal Psychology, 110, 282–289]. These results are not in line with cognitive theory of depression. Since this could be an artifact of the specific procedure that was used, we investigated implicit self-esteem of currently depressed participants and healthy controls using three different paradigms: The Implicit Association Test, the Name Letter Preference Task, and the Extrinsic Affective Simon Task. The results of the three experiments are unequivocally indicative of positive implicit self-esteem in currently depressed patients. However, it remains an intriguing question what exactly these indirect measures assess.

Introduction

Many researchers from different theoretical orientations such as Freud (1961), Maslow (1954), and more recently Baumeister and Leary (1995) consider the desire to view the self positively and the human need for esteem as an important universal aspect of well being. This general idea is supported by different kinds of research, showing that people tend to make self-serving attributions (Zuckerman, 1979), use self-enhancing and self-presentational strategies (Jones, 1991) and make self-serving assessments of ability (Dunning, Meyerowitz, & Holzberg, 1989).

In cognitive theories of depression, however, the self-concept is considered to be biased in a negative way (e.g., Beck, Rush, Shaw, & Emery, 1979; Clark, Beck, & Alford, 1999). In comparison with the idea of inaccurately positive illusions and esteem-enhancing biases in healthy people, depressed people would have inaccurately negative self-cognitions. Reviewing the literature on this topic, Haaga, Dyck, and Ernst (1991) concluded that the thinking of depressed patients is more negatively biased as compared to absolute standards of objective reality and that increased negative thinking about the self is a central feature of depression. Indeed, cognitive distortions such as enhanced negative thinking and increased accessibility to negative information have been reported frequently as a core cognitive style observed during depression (Williams, Watts, MacLeod, & Mathews, 1997). Within the cognitive research tradition, the existence of negative self-related dysfunctional schemata is emphasised to explain maintenance and vulnerability for future episodes of depression (Williams, 1997). Using questionnaires, it has already been thoroughly demonstrated that depressive people think negatively and report lower self-esteem than nondepressed controls (see: Ingram, Miranda, & Segal, 1998, for a review). However, this research uses explicit self-report measures and is therefore susceptible to bias and one could even doubt whether meaningful results concerning the underlying self-schemata can be achieved in this way. First, self-reporting may be influenced by demand characteristics, social desirability, and self-presentation. Second, within cognitive models of depression it is assumed that the crucial schemata are not always consciously accessible and thus cannot be reported (Beck et al., 1979; Young, 1994). They are considered primary cognitive processes that are automatically activated but which escape reflection and logical reasoning (Clark et al., 1999). For this reason, a distinction is drawn between the underlying schema processes that are not accessible and the products of such processes that are accessible within the conscious mind as opinions, inferences and interpretations (Ingram & Wisnicki, 1991).

Recently, new paradigms have been developed that could provide a satisfactory means of evaluating the underlying schema-processes related to self-esteem. These paradigms measure so-called “implicit self-esteem” which can be defined as an automatic evaluation of the self (Greenwald & Banaji, 1995). Since a basic feature of the concept of implicit self-esteem is its automatic nature, it must be measured indirectly and by means of a procedure that is relatively free of influence of self-representational processes (Bosson, Swann, & Pennebaker, 2000). Recent research suggests that automatic and controlled self-evaluations stem from different sources and should be seen as different constructs (Rudman, 2004).

The currently available experimental paradigms to evaluate implicit self-esteem are mainly based on the assumption that people assign value to objects that are associated with the self, and that attitudes can be activated automatically upon confrontation with attitude-objects. One of the earliest methods to evaluate implicit self-esteem that has been widely used is the Name Letter Preference Task (NLPT). This task is based on the assumption that the initials of one's own name are closely associated with the self and that an assessment of how well people like their initials relative to other letters reflects their implicit attitudes toward the self (Nuttin, 1985). Because people are not necessarily aware of the logic behind the task, their attitudes towards their own initials can be interpreted as an implicit index of self-esteem (see Greenwald & Banaji, 1995). In a recent study, Koole, Dijksterhuis, and van Knippenberg (2001) found that the positive bias for name letters became smaller when participants were asked to respond in a deliberative way. Moreover, Koole et al. could demonstrate that implicit self-evaluations corresponded with explicit self-evaluations only when participants made the explicit ratings very quickly or when under cognitive load. These findings support the idea that implicit self-esteem as measured by the NLPT is driven by automatic self-evaluations.

A second measure that has gained an enormous amount of interest during the last few years is the Implicit Association Test (IAT) (Greenwald & Farnham, 2000; Greenwald, McGhee, & Schwartz, 1998). During a prototypical self-esteem IAT, participants are asked to categorize words that appear in the middle of a computer screen as referring to “me” (e.g., own name) “not-me” (e.g., other name), “negative” (e.g., evil) or “positive” (e.g., happy) by pressing one of two keys. During one block of trials, the same key is pressed for “me” and “negative” words and the other key for “not-me” and “positive” words. The reaction times (RTs) during this block are compared to those of another block of trials during which the same key is used for “me” and “positive” words and the other key for “not-me” and “negative” words. Previous research demonstrated that healthy people categorize words significantly faster when the same response is required for self and positive words as compared to the task during which self and negative words require the same response, which is indicative of a positive implicit self-esteem (e.g. Greenwald & Farnham, 2000).

Recently, De Houwer (2003) developed the Extrinsic Affective Simon Task (EAST), a modified version of the IAT based on a comparison of performance on trials within a single task rather than on a comparison of performance on two different tasks. During the first phase of a self-esteem EAST, positive and negative white words appear in the middle of a computer screen and participants have to use one key for the positive words and the other key for negative words. In this way, the keys become extrinsically associated with a negative or a positive valence. During the second phase, coloured words are displayed that are either self-relevant or self-irrelevant. All these words can either be blue or green. The subject is asked to react only on the basis of the colour of the word (e.g. blue=press left, green=press right). After these two preparatory phases, both white and coloured words are displayed in random order. Albeit irrelevant for the execution of the task, it has been found in an experiment with undergraduate students that the meaning of the coloured words does have an influence on reaction speed and accuracy (De Houwer, 2003). On trials where the coloured word was self-relevant, performance was superior when the correct response was assigned to the positive key. The reverse was true on trials where the coloured words were self-irrelevant.

Although (implicit) self-esteem is thought to play a crucial role in depression, at this moment, research that evaluates implicit self-esteem in depressed individuals is extremely limited. The only study we are aware of was conducted by Gemar, Segal, Sagrati, and Kennedy (2001). They used the self-esteem IAT to examine mood related changes in implicit and explicit self-esteem in formerly depressed people and never depressed controls. After negative mood induction, the formerly depressed group showed both a larger drop in implicit and explicit self-esteem relative to controls, but the two measures (explicit and implicit) were not associated. However, a closer inspection of the data reveals that both the formerly depressed group and the never depressed controls showed a positive implicit self-esteem, before as well as after mood induction. Moreover, the pre–post mood induction difference scores observed by Gemar and colleagues were, in fact, due to higher implicit self-esteem in the formerly depressed participants before mood induction: the magnitude of the remaining positive bias score after mood induction was apparently not different between the formerly depressed population and the controls. Interestingly, the authors mention that the implicit self-esteem scores shown by the formerly depressed after mood induction was not significantly different from that found in a group of currently depressed patients.

These results suggest that implicit self-esteem remains positively biased in depressed participants, which is not in line with the existence of negative self-schema processes as postulated by the cognitive theory of depression. However, since these findings might be an artifact of the specific procedure that was used, the question remains whether or not implicit self-esteem remains positive in depressed patients. Therefore, we tested implicit self-esteem using a different version of the self-esteem IAT and two other implicit measures. In our IAT version, we used labels that are more specific for the cognitive model of depression. In two subsequent studies, we used the NLPT and the EAST as measures of implicit self-esteem. Moreover, a matched nondepressed control group was used to investigate if the magnitude of implicit self-esteem differs between depressed and nondepressed participants.

Section snippets

Participants

The healthy control sample consisted of 15 staff members (11 women and 4 men) of a general hospital. They were screened for the absence of any depressive disorder using a structured clinical interview for DSM-IV axis 1 disorders (Mini International Neuropsychiatric Interview, Dutch translation by Overbeek, Schruers, & Griez, 1999). They were between 21 and 59 years old (M=42.6, SD=12.1), had Beck Depression Inventory (BDI) scores between 0 and 11 (M=2.7, SD=2.7) and their Hamilton Rating Scale

Results

Only RTs on data collection trials with correct responses were analysed (blocks 3, 4, 6, 7), discarding RTs of the first trial of each block. RTs below 300 ms and above 3000 ms were recoded to 300 and 3000 ms, respectively (see Greenwald et al., 1998). To reduce the skew associated with response latencies, RTs were log-transformed.

Thereafter, mean log-transformed RTs were calculated separately for both tasks: (a) the task during which one response was required for the self-items and adjectives

Participants

The control group consisted of 16 healthy persons (13 women and 3 men) (for the screening and selection procedure, see Experiment 1) ranging in age between 22 and 53 years (M=37.2, SD=11.6). They had BDI scores between 0 and 9 ((M=2.9, SD=2.5) and HRSD scores between 0 and 7 (M=1.2, SD=2.6).

The clinical sample consisted of 16 inpatients (13 women and 3 men) who were referred to the psychiatry department of a general hospital (specialized reference centre for depression) with depressive

Results

In preparing the data, we followed the guidelines of Koole et al. (2001). For each participant, all 26 ratings were first Z-transformed based on the mean and the standard deviation of all ratings of that participant. In this way, the data were corrected for inter-individual differences in rating tendencies. Based on these Z-transformed scores, a global baseline evaluation for each group (depressed and controls) was calculated for each letter. This was the mean evaluation for each letter of the

Participants

Thirteen control participants (9 women and 4 men) volunteered to take part (for the selection and screening procedure, see Experiment 1). They were between 26 and 68 years old (M=39.6, SD=14.5), their BDI scores ranged between 0 and 5 (M=2.3, SD=1.8), and their HRSD scores between 0 and 3 (M=0.7, SD=1.2).

Thirteen inpatients (9 women and 4 men) referred to the psychiatry department of a general hospital (specialized reference centre for depression) with depressive symptoms participated in the

Results

Only RTs of correct responses on coloured me and not-me data collection trials were analysed. In line with De Houwer (2003), RTs below 300 ms and above 3000 ms were recoded to 300 and 3000 ms, respectively. RTs were log-transformed to reduce the skew associated with response latencies but untransformed RTs can be found in Table 2. Thereafter, mean log-transformed RTs were calculated separately for (a) trials on which a self-item was presented and an extrinsically positive response was required

General discussion

We investigated implicit self-esteem in currently depressed patients and nondepressed controls. When the IAT was used to measure implicit self-esteem (Experiment 1), RTs were shorter when the same response was required for the self-words and adjectives referring to worth than when the same response was required for the self-words and worthlessness words. Depressed patients and nondepressed controls matched for age and gender showed the same pattern. In the NLPT (Experiment 2), depressed

References (39)

  • J. De Houwer

    A structural and process analysis of the implicit association test

    Journal of Experimental and Social Psychology

    (2001)
  • J.E. Roberts et al.

    Level and stability of self-esteem as predictors of depressive symptoms

    Personality and Individual Differences

    (1995)
  • J.E. Roberts et al.

    A multidimensional model of self-esteem in depression

    Clinical Psychology Review

    (1994)
  • F.A. Albersnagel et al.

    Depressie

    (1989)
  • R.F. Baumeister et al.

    The need to belong: Desire for interpersonal attachments as a fundamental human motivation

    Psychological Bulletin

    (1995)
  • A.T. Beck et al.

    Cognitive therapy of depression

    (1979)
  • D.C. Berry et al.

    Implicit learning: Theoretical and empirical issues

    (1993)
  • J.K. Bosson et al.

    Stalking the perfect measure of implicit self-esteem: The blind men and the elephant revisited?

    Journal of Personality and Social Psychology

    (2000)
  • T.K. Bouman et al.

    Enige ervaringen met de Beck depression inventory (BDI)

    Gedrag

    (1985)
  • D.A. Clark et al.

    Scientific foundations of cognitive theory and therapy for depression

    (1999)
  • J. De Houwer

    The extrinsic affective Simon task

    Experimental Psychology

    (2003)
  • D.J.A. Dozois et al.

    Information processing and cognitive organisation in unipolar depression: Specificity and comorbidity issues

    Journal of Abnormal Psychology

    (2001)
  • D. Dunning et al.

    Ambiguity and self-evaluation: The role of idiosyncratic trait definitions in self-serving assessments of ability

    Journal of Personality and Social Psychology

    (1989)
  • A. Fernald

    Approval and disapproval: Infant responsiveness to vocal affect in familiar and unfamiliar languages

    Child Development

    (1993)
  • Freud, S. (1961). The ego and the ID. In: J. Strachley (Ed. and Translation), The standard edition of the complete...
  • M.C. Gemar et al.

    Mood-induced changes on the implicit association test in recovered depressed patients

    Journal of Abnormal Psychology

    (2001)
  • M.S. Greenberg et al.

    Depression versus anxiety: Processing of self—and other—referent information

    Cognition and Emotion

    (1989)
  • A.G. Greenwald et al.

    Implicit social cognition: Attitudes, self-esteem, and stereotypes

    Psychological Review

    (1995)
  • A.G. Greenwald et al.

    An unified theory of implicit attitudes, stereotypes, self-esteem, and self-concept

    Psychological Review

    (2002)
  • Cited by (0)

    View full text