- Email: [email protected]
On the relationship between implicit and explicit self-esteem: The moderating role of dismissing attachment
Personality and Individual Differences 52 (2012) 173–177
Contents lists available at SciVerse ScienceDirect
Personality and Individual Differences journal homepage: www.elsevier.com/locate/paid
On the relationship between implicit and explicit self-esteem: The moderating role of dismissing attachment Francesco Dentale ⇑, Michele Vecchione, Alessandra De Coro, Claudio Barbaranelli ‘‘Sapienza’’ University of Rome, Rome, Italy
a r t i c l e
i n f o
Article history: Received 19 May 2011 Received in revised form 30 September 2011 Accepted 10 October 2011 Available online 4 November 2011 Keywords: Implicit measures Self-esteem Dismissing Attachment style
a b s t r a c t Recently, dual models of social cognition have distinguished the implicit (affective) and the explicit (propositional) levels of self-esteem. Many studies have found that implicit and explicit self-esteem are weakly or even not correlated. Concerning the moderating factors of such a relationship, other studies stressed the importance of emotional awareness in the translation of implicit self-evaluations into a propositional format. If so, the defensive strategy which characterizes dismissing attachment should determine a fictitious explicit self-image, by impairing the awareness of self-related emotional experiences. At the implicit level, instead, such defensive processes should not operate. We therefore hypothesized that dismissing attachment would increase the implicit–explicit self-esteem discordance. Participants were 145 university students with a mean age of 20.4 (SD = 2.75). Moderated multiple regression analyses were conducted to test for the interactive effects of the Self-Esteem Implicit Association Test and the Attachment Style Questionnaire scales on the Rosenberg Self-Esteem Scale. Findings revealed an interaction effect only for the relationships as secondary scale, confirming that dismissing style, but not other attachment dimensions, increases implicit–explicit self-esteem discordance. Ó 2011 Elsevier Ltd. All rights reserved.
1. Introduction 1.1. Implicit and explicit self-esteem In the last decade, an increasing attention has been paid to the use of indirect measures of self-esteem and self-concept (Greenwald & Farnham, 2000; Karpinski & Steinman, 2006; Koole, Dijksterhuis, & van Knippenberg, 2001). Among them, the most popular is the Implicit Association Test (IAT), first developed by Greenwald, McGhee, and Schwartz (1998) and successively used in many areas of psychological research. A growing body of evidence suggests that implicit and explicit measures of self-esteem are only weakly or not correlated (see Buhrmester, Blanton, & Swann, 2011, for a review). This is in accordance with findings of Nosek and Smyth (2007), who demonstrated that IAT and self-report measures refer to distinct constructs. In order to explain such results, recent models of social cognition distinguished two levels of information processing, that are the implicit (i.e. associative) and the explicit (i.e. propositional) ones (Gawronski & Bodenhausen, 2006; Strack & Deutsch, 2004). The associative level is intimately connected to the immediate affective reaction towards a target object, while the propositional level is more connected to deliberative thinking. For instance in the area of attitudes, according to Hofmann, Gschwender, Nosek, and Schmitt ⇑ Corresponding author. E-mail address: [email protected] (F. Dentale). 0191-8869/$ - see front matter Ó 2011 Elsevier Ltd. All rights reserved. doi:10.1016/j.paid.2011.10.009
(2005), implicit and explicit representations are two structurally distinct constructs that may be interrelated if certain conditions are satisfied. The implicit attitudes constitutes the associationbased foundation of attitudes, which may be conceived as an activation pattern in an associative network of concepts. This network can be activated fast, automatically, and without the use of cognitive resources. The implicit attitudes is indirectly assessed via implicit measures. In contrast, it is assumed that the explicit attitudes are represented in a propositional format and are formed by reasoned judgments that can be measured with self-report measures. In this vein, implicit self-esteem should be formed by intuitive and affective evaluations towards the self while explicit self-esteem should depends on deliberative and propositional judgments. On the basis of such a distinction we may assume that implicit and explicit measures of self-esteem should not correlate necessarily. Hofmann et al. (2005) suggesting that such a relationship, increases (or decreases) on the basis of many specific conditions, have classified five groups of moderating variables that influence implicit–explicit consistency: translation between implicit and explicit representations factors (e.g., representational strength, awareness), additional information integration for explicit representations factors (e.g., need for cognition), properties of explicit assessment factors (e.g., social desirability concerns), properties of implicit assessment factors (e.g., situational malleability), and research design factors (e.g., sampling bias, measurement correspondence).
174
F. Dentale et al. / Personality and Individual Differences 52 (2012) 173–177
1.2. The emotional awareness facilitates implicit to explicit translation Of particular interest here are a series of factors pertaining to awareness (e.g., private self-consciousness, mindfulness, faith in intuition, alexithymia) that are able to facilitate the translation from an implicit to an explicit format. According to Hofmann et al. (2005) such a translation process may depend upon the degree to which people are able to form accurate propositional representations of their underlying associative representations. In other terms, the association-based attitudes (e.g., me-good) are able to be integrated into the propositional attitudes (e.g., I am a good person) only if the subjects are able to achieve a sufficient awareness of the internal and external cues of the implicit activations (Hofmann, Gschwender, & Schmitt, 2005). Therefore, all factors that impair (or ameliorate) the ability to identify the internal (e.g., somatic sensations, feelings, mood changes) or the external cues (e.g., self-observation of spontaneous behaviors) of the implicit associative activations, should also impair (or ameliorate) the quality of the translation process and thus should decrease (or increase) implicit–explicit consistency. Concerning the ability to identify the internal cues of implicit associative activations, previous studies have examined the influence of private self-consciousness and mindfulness on implicit–explicit consistency (Brown & Ryan, 2003; Hofmann, Gschwender, & Schmitt, 2005). Private self-consciousness was defined as the extent to which individuals pay attention to their bodily and emotional experience. Mindfulness refers to the enhanced attention to and awareness of current experience or present reality, which may be reflected in more regular or sustained consciousness of ongoing events and experiences. Empirical findings revealed that high scorers in private self-consciousness and in mindfulness exhibited stronger implicit–explicit consistency than low scorers. With regard to the implicit–explicit self-esteem consistency, Jordan, Whitfield, and Zeigler-Hill (2007) argued that implicit self-esteem (ISE) is based on intuitive self-views that are connected to immediate affective reactions toward the self, whereas explicit self-esteem (ESE) is based on deliberative selfviews that are connected to evaluative judgements. They found that high scorers in faith in intuition, namely people’s confidence about their feelings as bases for decision making (Epstein, Pacini, Denes-Raj, & Heier, 1996), showed a greater consistency between ISE and ESE than low scorers. In a similar vein, Dentale, San Martini, De Coro, and Di Pomponio (2010) found that alexithymia moderates the relationship between implicit and explicit self-esteem. In particular, they demonstrated that both the difficulties to identify and describe feelings increase the discordance between ISE and ESE. These results indicate that emotional awareness plays a relevant role in the translation of implicit self-associations into propositional and explicit formats. Along this line of reasoning, it is reasonable to expect that personality factors that impair emotional awareness may increase the implicit–explicit self-esteem discordance. Among such personality factors we can find certain traits that defend subjects against too extreme emotions, such as the repressive coping style (Weinberger, Schwartz, & Davidson, 1979), the isolation of affect defense style (Bond, Gardner, Christian, & Sigal, 1983) or the dismissing style of attachment (Bowlby, 1973). 1.3. Dismissing attachment and emotional awareness of the self According to the literature on the attachment behavioral system (Bowlby, 1973; Main, Kaplan, & Cassidy, 1985), dismissing subjects are characterized by a defensive emotional regulation that induce an impoverished emotional awareness (e.g., Waller & Scheidt, 2006), especially in close relationships: ‘‘dismissing status of attachment is linked to the disability or to defensive forms of processing and expressing emotions, whereas attachment security
is associated with open, flexible affect expression and the ability to explore and process emotional experiences without employing defensive strategies (Waller & Scheidt, 2006, p. 19)’’. As is well known, recurrent experiences of the attachment figure’s availability and responsiveness would warrant for an evaluation of the self as a person toward whom anyone, and the attachment figure in particular, is likely to respond in a helpful way (Kobak, 1999). To preserve their explicit self-image, dismissing subjects tend to divert their attention from distressful attachment signals (Bowlby, 1973; Dixon, 1981) and to suppress the associated negative emotions (Dozier & Kobak, 1992), protecting themselves from attachment relationships that appear too painful. As a consequence, in spite of attachment experiences of refusal or even threat, dismissing subjects usually showed a fictitious explicit self-image that is independent by actual relationships. Such a defensive process permits them to stabilize the explicit self-esteem in respect to extreme fluctuations that may endanger their mental and physical well-being. As a side effect, in our view, it should emerge as an impaired explicit access to the spontaneous self-directed feelings that make up implicit self-esteem. But what about the implicit associative level? On the basis of dual models of social cognition (Gawronski & Bodenhausen, 2006; Strack & Deutsch, 2004), we have no reasons to assume that dismissing attachment operates by defensively distorting the implicit self-evaluations, as occur for the explicit self-judgments. On the contrary, it seems likely that automatic associations are closely connected to the affective experiences that determine the spontaneous and intuitive selfevaluations. As a consequence, similarly to what has been hypothesized for other moderating factors pertaining to awareness (Brown & Ryan, 2003; Dentale et al., 2010; Hofmann, Gschwender, & Schmitt, 2005; Jordan et al., 2007), dismissing style should impair implicit to explicit translation process, and thus decrease the implicit–explicit self-esteem consistency. Therefore, we expected that dismissing attachment style, but not secure, anxious and fearful/ avoidant styles, increase the discordance between implicit and explicit self-esteem. 1.4. Dismissing style decreases implicit–explicit self-esteem consis tency To test our hypothesis, we administered the Attachment Style Questionnaire (ASQ, Feeney, Noller, & Hanrahan, 1994) to measure five styles of attachment: confidence, discomfort with closeness, relationships as secondary, need for approval, and preoccupation with relationships. As Fossati et al. (2003) noted, of all ASQ dimensions, only relationships as secondary is fully consistent with Bartholomew’s concept of dismissing attachment (Bartholomew & Horowitz, 1991). To assess implicit–explicit consistency, we administered a self-report scale of self-esteem, namely the Rosenberg Self-Esteem Scale (RSES, Rosenberg, 1965), and two parallel forms of IAT for measuring implicit self-esteem (SE-IAT). Our hypothesis is that high scorers in relationships as secondary scale, but not in confidence, preoccupation with relationships, need for approval and discomfort with closeness, exhibit higher discrepancies between implicit self-esteem and explicit self-esteem than low scorers. 2. Methods 2.1. Participants and procedures Participants in the study were 145 undergraduate students (33 males and 112 females) enrolled in a psychology course at the ‘‘Sapienza’’ University of Rome. The age of the respondents ranged
175
F. Dentale et al. / Personality and Individual Differences 52 (2012) 173–177
from 18 to 20 years (M = 20.4, SD = 2.75). Participants came to the laboratory individually, where they first completed a booklet of questionnaires including measures of explicit self-esteem and attachment styles. Then, they were given two parallel forms of SE-IAT. All participants were informed and consented to participate in the study, for which they earned course credits. On completion of the study, all participants were given debriefing statements and thanked for participation. 2.2. Measures 2.2.1. Attachment styles Attachment styles were measured by the ASQ (Feeney et al., 1994), an inventory yielding five subscale scores representing different styles of attachment: Confidence, Preoccupation with Relationships, Need for Approval, Discomfort with Closeness and Relationships as Secondary. Confidence (Confid) (in self and others) reflects a secure attachment orientation. Preoccupation with Relationships (PreoRel), which involves an anxious and dependent approach to relationships, is a core feature of Hazan and Shaver’s (1987) original conceptualization of anxious/ambivalent attachment. Need for Approval (NeedApp) reflects respondents’ need of acceptance and confirmation from others, and characterizes Bartholomew’s fearful and preoccupied individuals. Discomfort with Closeness (Disclos) is a central theme of both Hazan and Shaver’s (1987) conceptualization of avoidant attachment and Bartholomew’s view of fearful subjects. Finally, Relationships as Secondary (RelSec) is the unique dimension that is consistent with Bartholomew’s concept of dismissing attachment. The Italian version of the instrument (Fossati et al., 2003) presented a good internal consistency for the RelSec subscale (a = .73), and moderate to low coefficients for Confid (a = .69), PreoRel (a = .64), NeedApp (a = .69), and Disclos (a = .68). 2.2.2. Explicit self-esteem Explicit self-esteem was measured by the RSES (Rosenberg, 1965). The Italian version of the scale (Prezza, Trombaccia, & Armento, 1997) has proved to be reliable: Cronbach’s alpha and test–retest reliability coefficients were .84 and .76, respectively. 2.2.3. Implicit self-esteem Two parallel forms of SE-IAT (Greenwald & Farnham, 2000) were used, in which the stimuli-words of the target-concept (me vs. not-me) were the same in both forms (e.g., me, self vs. it, that), whereas the stimuli-words of the attribute-dimension (pleasant vs. unpleasant) were emotionally loaded nouns (IATn) in the first form (e.g., luck, joy vs. vomit, agony) and emotionally loaded attributes (IATa) in the second form (e.g., positive, good vs. negative, bad)1. In both SE-IATs the participants performed two types of categorization tasks, with 5 stimuli-words for each category. The words were presented in random order within each block of trials. As described by Greenwald et al. (1998) the entire procedure consisted of seven blocks of trials: 1 (me vs. not-me), 2 (pleasant vs. unpleasant) and 5 (not-me vs. me) were single categorization blocks of 20 trials whereas 3–4 and 6–7 were combined blocks (me or pleasant vs. not-me or unpleasant) of 20 (3–6) and 40 (4–7) trials. Subjects were requested to respond as quickly and accurately as possible to the stimuli-words that appeared on the monitor. Following Greenwald, Nosek, and Banaji (2003), data from blocks 3–4 and 6–7 were used to compute SE-IAT difference scores (Dbiep), according to the built-in error penalty procedure and through the following formula:
Dbiep ¼ ðM IB MCB Þ=SD
1
ð1Þ
A complete description of the test is available from the first author upon request.
where MIB is the mean latency for incompatible blocks, MCB is the mean latency for compatible blocks; SD is the pooled standard deviation for compatible and incompatible blocks. Positive scores indicate high implicit self-esteem and negative scores indicate low implicit self-esteem. The internal consistencies of the two parallel IATs were estimated by a split-half index, based on two partial scores, Dbiep1 and Dbiep2, respectively computed from blocks 3 and 6 (20 + 20 trials), and from blocks 4 and 7 (40 + 40 trials). The correlation between the two parallel measures was r = .44 (p < .01). Since this result is affected by measurement error, we disattenuated the coefficient by taking into account the unreliability of the scales. After correction the correlation increased to .79. A global score of implicit self-esteem was computed by summing the standardized values of the Dbiep score of the two parallel forms. The composite score was used in the following analyses. 3. Results 3.1. Descriptive statistics Descriptive statistics and reliabilities of the variables included in the study are reported in Table 1. All measures were approximately normally distributed, since skewness and kurtosis values were close to |1| (Tabachnick & Fidell, 2001). As can be observed, all self-report measures have acceptable-to-high levels of internal consistency. Reliabilities of the two IAT measures, instead, were not fully adequate (rs-h < .60) but not too different from results commonly found (e.g., Krause, Back, Egloff, & Schmukle, 2011). Table 2 provides the intercorrelations among the study variables. Regarding the ASQ subscales, Confid was negatively and significantly correlated with all the other attachment scales. Anxious (PreoRel and NeedApp) and avoidant (Disclos and RelSec) scales were strongly and positively intercorrelated to each other, but weakly or not correlated between them (this follows the same pattern found in Fossati et al., 2003). The correlation between ISE and ESE was close to zero, in accordance with earlier findings (e.g., Greenwald & Farnham, 2000). Confid was strongly and positively correlated with ESE, while it was unrelated with ISE. The other ASQ subscales (with the exception of RelSec) correlated negatively and significantly with ESE but not with ISE. 3.2. Moderated regression analyses To investigate whether dismissing style moderates the relationship between implicit and explicit self-esteem, a hierarchical moderated regression analysis was conducted. In this analysis, ESE was considered as a criterion variable, in accordance with the theoretical model proposed by Hofmann, Gschwender, & Schmitt (2005) and
Table 1 Descriptive statistics.
RSES Con PWR NFA DWC RAS IATn Dbiep IATa Dbiep
No. of items
Reliability
Mean
SD
Skewness
Kurtosis
10 8 8 7 10 7 – –
.82a .67a .71a .71a .74a .65a .56rs–h .55rs–h
31.58 32.60 30.58 21.24 36.14 13.37 .67 .62
4.71 5.06 6.16 5.75 7.23 4.15 .32 .31
.63 .42 .07 .37 .03 1.08 .42 .17
.03 .11 .42 .38 .11 1.85 .57 .46
Note: a: Cronbach’s alpha; rtt: test–retest; rs-h: split-half; RSES: Rosenberg SelfEsteem Scale; Con: Confidence; PWr: Preoccupation With Relationships; NFA: Need For Approval; DWC: Discomfort With Closeness; RAS: Relationships As Secondary; IATn: Implicit Association Test with emotionally loaded nouns as stimuli; IATa: Implicit Association Test with emotionally loaded attributes as stimuli; Dbiep: SEIAT scores according to the built in error penalty method.
176
F. Dentale et al. / Personality and Individual Differences 52 (2012) 173–177
Table 2 Correlations among study variables. RSES RSES IAT-SE CON PWR NFA DWC RAS
IAT-SE
Con
PWR
1 .04 .10 .12 .01 .11
1 .35** .45** .47** .25**
NFA
DWC
RAS
1 .20* .12
1 .30**
1
1 .03 .61** .38** .63** .25** .14
1 .53** .08 .07
Note: RSES: Rosenberg Self-Esteem Scale; IAT-SE: IAT Self-Esteem; Con: Confidence; PWR: Preoccupation With Relationships; NFA: Need For Approval; DWC: Discomfort With Closeness; RAS: Relationships as Secondary. * p < .05 (two tails). ** p < .01 (two tails). Fig. 1. Explicit Self Esteem (ESE) as a function of Implicit Self-Esteem (ISE) and Relationships as Secondary (RelSec) facet of dismissing attachment style.
coherently with what has been made in other similar studies on the moderating role of factors pertaining to awareness (Brown & Ryan, 2003; Hofmann, Gschwender, & Schmitt, 2005; Jordan et al., 2007). In the first step, RelSec and ISE scores were entered as predictors (which represent main effects). In the second step, their cross product term (RelSecxISE) was entered. The predictors were centered before creating the cross-product vector (Aiken & West, 1991), which represents the interaction effect. Results (Table 3e) showed no significant main effects for ISE and RelSec, and a significant interaction term. The lack of main effects is not surprising but in line with the literature that showed null or weak correlations between ISE and ESE (e.g., Buhrmester, Blanton, & Swann, 2011) on the one hand, and between RelSec and ESE on the other hand (e.g., Feeney et al., 1994; Fossati et al., 2003). The interaction between RelSec and ISE explained a significant incremental amount of variance in ESE (DR2 = .03, p < .05), demonstrating the expected moderating role of dismissing attachment on the ISE-ESE consistency. The negative sign of the beta coefficient for the interaction term indicated that the concordance between ISE and ESE tended to increase as dismissing attachment decreases.
Table 3 Results of moderated regression analyses on the Rosenberg Self-Esteem Scale. Predictors a
b
c
d
e
Step 1 ISE CON Step 2 ISExCON Step 1 ISE PWR Step 2 ISExPWR Step 1 ISE NFA Step 2 ISExNFA Step 1 ISE DWC Step 2 ISExDWC Step 1 ISE RAS Step 2 ISExRAS
b (p)
R2 change
F Change (p)
Df
.38
42.35 (<.001)
2, 137
.00
0.14 (.75)
1, 136
.15
11.78 (<.001)
2, 137
.00
0.30 (.59)
1, 136
.40
47.20 (<.001)
2, 140
.00
0.08 (.78)
1, 139
.06
4.81 (.01)
2, 138
.02
2.49 (.12)
1, 137
.02
1.57 (.21)
2, 137
.03
4.14 (<.05)
1, 136
To interpret the interaction effect, the simple slopes for the relationship between ISE and ESE were tested at one standard deviation below (1 SD) and one standard deviation above (+1 SD) the RelSec mean, in accordance with Cohen and Cohen’s (1983) guidelines (see Fig. 1). At 1 SD, the relation between ISE and ESE was positive and significant (b = .217; t136 = 2.32, p < .01), indicating a concordance of medium size (Cohen & Cohen, 1983). On the other hand, at +1 SD the relation was negative and not significant (b = .121; t136 = .99, p = .16). To further examine the relationship between attachment styles and implicit–explicit self-esteem concordance, the regression analysis was repeated for each of the remaining attachment scales (Confid – see Table 3a; PreoRel – see Table 3b; NeedApp – see Table 3c; Disclos – see Table 3d). In all of the four regressions, the main effects of the attachment scales were significant at the first step. In line with the literature (e.g., Feeney et al., 1994; Fossati et al., 2003), PreoRel, NeedApp, and Disclos were found to be negatively related with ESE, whereas Confid had a positive relation. Differently, in accordance with previous results (Buhrmester et al., 2011), the main effect of ISE was not significant in each regression. At the second step, the interaction terms were not significant. Only the DisclosxISE cross product term exhibited a certain tendency to be significant (p = .12) with a small and negative beta coefficient. In sum, the five regressions indicated that only RelSec (and not Confid, PreoRel, NeedApp and Disclos) showed a significant moderating effect on the relationship between implicit and explicit self-esteem.
.06 (.34) .62 (<.01) -.03 (.70) .01 (.85) .38 (<.001) .04 (.59) .05 (.43) .64 (<.001) .02 (.78) .041 (.62) .252 (<.01) .13 (.12) .05 (.56) .14 (.14) .17 (<.05)
Note: ISE: Implicit Self-Esteem; CON: Confidence; PWR: Preoccupation with Relationships; NFA: Need for Approval; DWC: Discomfort with Closeness; RAS: Relationships as Secondary.
4. Conclusions The present study showed that dismissing/avoidant attachment (as measured by the ASQ relationships as secondary scale) increases implicit–explicit self-esteem discordance. As an illustration of our results, it was found that at one standard deviation below the mean of the RelSec scale (i.e. for individuals who present a low dismissing attachment tendency), there was a positive and significant, though moderate, correlation between implicit and explicit self-esteem. At one standard deviation above the mean, instead (i.e. for individuals who present a high dismissing attachment tendency), the correlation was not significant and weakly negative. As expected, of all ASQ scales, only relationships as secondary emerged as a moderator of the relationship between implicit and explicit self-esteem, although data showed that Disclos (i.e. fearful/avoidant dimension of ASQ) tends to increase the discordance between the two measures. The lack of interaction between Disclos and ISE may be explained by noting that this scale catches some avoidant behaviors that are accompanied by a certain degree of awareness of negative emotions in attachment interactions.
F. Dentale et al. / Personality and Individual Differences 52 (2012) 173–177
Taken together, these results give further support to the role ascribed to emotional awareness in the propositional (explicit) translation of affective (implicit) evaluations (Hofmann, Gschwender, & Schmitt, 2005; Jordan et al., 2007), and are compatible with Hofmann et al.’s (2005) working model of implicit–explicit consistency, positing awareness in a general group of moderators. In another possible interpretation, the moderating effect of dismissing attachment may depend on adjustment processes (Hofmann, Gschwender, & Schmitt, 2005) that bias explicit self-esteem scores, similarly to what happens as a consequence of social-desirability, impression management or self-deception effects. If so, dismissing individuals may be a part of those people who defend their self-image using compensatory responses (e.g., Jordan, Spencer, Zanna, Hoshino-Browne, & Correll, 2003). In other words, the moderating effect we found might be ascribed to a proneness to self-enhancement in the explicit assessment, rather than to an impaired emotional awareness that decreases the quality of the implicit to explicit translation process. In this vein, perhaps not coincidentally, high scorers on the RelSec scale showed a weak, though not significant, negative correlation between ISE and ESE. That is, for individuals with a high dismissing attachment, lower scores on implicit self-esteem appear to be associated to higher levels of explicit self-esteem. Looking at the recent literature on implicit self-esteem, one cannot exclude that this tendency would be due to a kind of compensatory bias. Further studies are necessary to examine the extent to which such a negative correlation is meaningful and generalizable across different samples. A limitation of the present study is that the attachment styles were assessed only via a self-report inventory and exclusively in a dimensional perspective. If we maintain that dismissing subjects are impaired in the explicit processing of affective information, their ability to provide valid self-reports on the ASQ may also be impaired to some extent, as this measure requires a great deal of introspective ability. A relevant confirmation of our results may come from investigations using clinical assessment of attachment states of mind through the Adult Attachment Interview (Steele & Steele, 2008). A further limitation of the present study is that implicit self-esteem was measured only through classical IAT, in which two target categories are opposed to each other. To examine the robustness of our results, it may be appropriate to conduct further studies using other implicit techniques, such as the Go/No-Go Association Task (GNAT, Nosek & Banaji, 2001) or the Single-Category IAT (SC-IAT, Karpinski & Steinman, 2006). Notwithstanding its limits, this contribution supports the hypothesis – traditionally confined into clinical assumptions – that dimensions relevant for interpersonal regulation of affects, such as attachment working models or defensive strategies, are to be considered in the theory and assessment of self-esteem. References Aiken, L. S., & West, S. G. (1991). Multiple regression: testing and interpreting interactions. London: Sage. Bartholomew, K., & Horowitz, L. M. (1991). Attachment styles among young adults: a test of a four-category model. Journal of Personality and Social Psychology, 61, 226–244. Bond, M. P., Gardner, S. T., Christian, J., & Sigal, C. (1983). Empirical study of selfrated defence styles. Archives of General Psychiatry, 46, 455–460. Bowlby, J. (1973). Attachment and loss: Vol. 2. Separation. New York: Basic Books. Brown, K. W., & Ryan, R. M. (2003). The benefits of being present: Mindfulness and its role in psychological well-being. Journal of Personality and Social Psychology, 84, 822–848. Buhrmester, M. D., Blanton, H., & Swann, W. B. Jr., (2011). Implicit self-esteem: Nature, measurement and a new way forward. Journal of Personality and Social Psychology, 100, 365–385. Cohen, J., & Cohen, P. (1983). Applied multiple regression/correlation analysis for the behavioral sciences. Hillsdale, NJ: Erlbaum.
177
Dentale, F., San Martini, P., De Coro, A., & Di Pomponio, I. (2010). Alexithymia increases the discordance between implicit and explicit self-esteem. Personality and Individual Differences, 49, 762–767. Dixon, N. F. (1981). Preconscious processing. England, Wiley: Chichester. Dozier, M., & Kobak, R. R. (1992). Psychophysiology in attachment interview: converging evidence for deactivating strategies. Child Development, 63, 1473–1480. Epstein, S., Pacini, R., Denes-Raj, V., & Heier, H. (1996). Individual differences in intuitive-experiential and analytical-rational thinking styles. Journal of Personality and Social Psychology, 71, 390–405. Feeney, J. A., Noller, P., & Hanrahan, M. (1994). Assessing adult attachment. In M. B. Sperling & W. H. Berman (Eds.), Attachment in adults: Clinical and developmental perspectives (pp. 128–152). New York: Guilford Press. Fossati, A., Feeney, J. A., Donati, D., Donini, M., Novella, L., Bagnato, M., et al. (2003). On the dimensionality of the Attachment Style Questionnaire in Italian clinical and non-clinical participants. Journal of Social and Personal Relationships, 20, 55–79. Gawronski, B., & Bodenhausen, G. V. (2006). Associative and propositional processes in evaluation: An integrative review of implicit and explicit attitude change. Psychological Bulletin, 132, 692–731. Greenwald, A. G., & Farnham, S. D. (2000). Using the Implicit Association Test to measure self-esteem and self-concept. Journal of Personality and Social Psychology, 79, 1022–1038. Greenwald, A. G., McGhee, D. E., & Schwartz, J. K. L. (1998). Measuring individual differences in implicit cognition: The Implicit Association Test. Journal of Personality and Social Psychology, 74, 1464–1480. Greenwald, A. G., Nosek, B. A., & Banaji, M. R. (2003). Understanding and using the Implicit Association Test: An improved scoring algorithm. Journal of Personality and Social Psychology, 85, 197–216. Hazan, C., & Shaver, P. (1987). Conceptualizing romantic love as an attachment process. Journal of Personality and Social Psychology, 52, 511–524. Hofmann, W., Gschwender, T., Nosek, B. A., & Schmitt, M. (2005). What moderates implicit–explicit consistency? European Review of Social Psychology, 16, 335–390. Hofmann, W., Gschwender, T., & Schmitt, M. (2005). On implicit–explicit consistency: The moderating role of individual differences in awareness and adjustment. European Journal of Personality, 19, 25–49. Jordan, C. H., Spencer, S. J., Zanna, M. P., Hoshino-Browne, E., & Correll, J. (2003). Secure and defensive self-esteem. Journal of Personality & Social Psychology, 85, 969–978. Jordan, C. H., Whitfield, M., & Zeigler-Hill, V. (2007). Intuition and the correspondence between implicit and explicit self-esteem. Journal of Personality and Social Psychology, 93, 1067–1079. Karpinski, A., & Steinman, R. B. (2006). The single category implicit association test (SC-IAT) as a measure of implicit social cognition. Journal of Personality and Social Psychology, 91, 16–32. Kobak, R. (1999). The emotional dynamics of dysruptions in attachment relationships. In J. Cassidy & P. R. Shaver (Eds.), Handbook of attachment: Theory, research and clinical applications (pp. 21–43). New York-London: Guilford Press. Koole, S. L., Dijksterhuis, A., & van Knippenberg, A. (2001). What’s in a name: Implicit self-esteem and the automatic self. Journal of Personality and Social Psychology, 80, 669–685. Krause, S., Back, M. D., Egloff, B., & Schmukle, S. C. (2011). Reliability of implicit selfesteem measures revisited. European Journal of Personality, 25, 239–251. Main, M., Kaplan, N., & Cassidy, J. (1985). Security in infancy, childhood, and adulthood: A move to the level of representation. Growing points in attachment theory and researchI. Bretherton & E. Waters (Eds.). Monographs of the Society for Research in Child Development, 50, 66–106. Nosek, B. A., & Banaji, M. R. (2001). The go/no-go association task. Social Cognition, 19, 625–666. Nosek, B. A., & Smyth, F. L. (2007). A multitrait-multimethod validation of the Implicit Association Test: Implicit and explicit attitudes are related but distinct constructs. Experimental Psychology, 54, 14–29. Prezza, M., Trombaccia, F. R., & Armento, L. (1997). La scala dell’autostima di Rosenberg. Traduzione e validazione italiana [The Rosenberg Self-Esteem Scale: Italian translation and validation]. Bollettino di Psicologia applicata, 223, 35–44. Rosenberg, M. (1965). Society and the adolescent self-image. Princeton, NJ: Princeton University Press. Steele, H., & Steele, M. (2008). Clinical applications of the adult attachment interview. New York-London: Guilford Press. Strack, F., & Deutsch, R. (2004). Reflective and impulsive determinants of social behavior. Personality and Social Psychology Review, 8, 220–247. Tabachnick, B. G., & Fidell, L. S. (2001). Using multivariate statistics (4th ed.). Needham Heights, MA: Allyn & Bacon. Waller, E., & Scheidt, C. E. (2006). Somatoform disorders as disorders of affect regulation: A development perspective. International Review of Psychiatry, 18, 13–24. Weinberger, D. A., Schwartz, G. E., & Davidson, R. J. (1979). Low-anxious, high anxious and repressive coping styles: Psychometric patterns and behavioral responses to stress. Journal of Abnormal Psychology, 88, 369–380.
Contents lists available at SciVerse ScienceDirect
Personality and Individual Differences journal homepage: www.elsevier.com/locate/paid
On the relationship between implicit and explicit self-esteem: The moderating role of dismissing attachment Francesco Dentale ⇑, Michele Vecchione, Alessandra De Coro, Claudio Barbaranelli ‘‘Sapienza’’ University of Rome, Rome, Italy
a r t i c l e
i n f o
Article history: Received 19 May 2011 Received in revised form 30 September 2011 Accepted 10 October 2011 Available online 4 November 2011 Keywords: Implicit measures Self-esteem Dismissing Attachment style
a b s t r a c t Recently, dual models of social cognition have distinguished the implicit (affective) and the explicit (propositional) levels of self-esteem. Many studies have found that implicit and explicit self-esteem are weakly or even not correlated. Concerning the moderating factors of such a relationship, other studies stressed the importance of emotional awareness in the translation of implicit self-evaluations into a propositional format. If so, the defensive strategy which characterizes dismissing attachment should determine a fictitious explicit self-image, by impairing the awareness of self-related emotional experiences. At the implicit level, instead, such defensive processes should not operate. We therefore hypothesized that dismissing attachment would increase the implicit–explicit self-esteem discordance. Participants were 145 university students with a mean age of 20.4 (SD = 2.75). Moderated multiple regression analyses were conducted to test for the interactive effects of the Self-Esteem Implicit Association Test and the Attachment Style Questionnaire scales on the Rosenberg Self-Esteem Scale. Findings revealed an interaction effect only for the relationships as secondary scale, confirming that dismissing style, but not other attachment dimensions, increases implicit–explicit self-esteem discordance. Ó 2011 Elsevier Ltd. All rights reserved.
1. Introduction 1.1. Implicit and explicit self-esteem In the last decade, an increasing attention has been paid to the use of indirect measures of self-esteem and self-concept (Greenwald & Farnham, 2000; Karpinski & Steinman, 2006; Koole, Dijksterhuis, & van Knippenberg, 2001). Among them, the most popular is the Implicit Association Test (IAT), first developed by Greenwald, McGhee, and Schwartz (1998) and successively used in many areas of psychological research. A growing body of evidence suggests that implicit and explicit measures of self-esteem are only weakly or not correlated (see Buhrmester, Blanton, & Swann, 2011, for a review). This is in accordance with findings of Nosek and Smyth (2007), who demonstrated that IAT and self-report measures refer to distinct constructs. In order to explain such results, recent models of social cognition distinguished two levels of information processing, that are the implicit (i.e. associative) and the explicit (i.e. propositional) ones (Gawronski & Bodenhausen, 2006; Strack & Deutsch, 2004). The associative level is intimately connected to the immediate affective reaction towards a target object, while the propositional level is more connected to deliberative thinking. For instance in the area of attitudes, according to Hofmann, Gschwender, Nosek, and Schmitt ⇑ Corresponding author. E-mail address: [email protected] (F. Dentale). 0191-8869/$ - see front matter Ó 2011 Elsevier Ltd. All rights reserved. doi:10.1016/j.paid.2011.10.009
(2005), implicit and explicit representations are two structurally distinct constructs that may be interrelated if certain conditions are satisfied. The implicit attitudes constitutes the associationbased foundation of attitudes, which may be conceived as an activation pattern in an associative network of concepts. This network can be activated fast, automatically, and without the use of cognitive resources. The implicit attitudes is indirectly assessed via implicit measures. In contrast, it is assumed that the explicit attitudes are represented in a propositional format and are formed by reasoned judgments that can be measured with self-report measures. In this vein, implicit self-esteem should be formed by intuitive and affective evaluations towards the self while explicit self-esteem should depends on deliberative and propositional judgments. On the basis of such a distinction we may assume that implicit and explicit measures of self-esteem should not correlate necessarily. Hofmann et al. (2005) suggesting that such a relationship, increases (or decreases) on the basis of many specific conditions, have classified five groups of moderating variables that influence implicit–explicit consistency: translation between implicit and explicit representations factors (e.g., representational strength, awareness), additional information integration for explicit representations factors (e.g., need for cognition), properties of explicit assessment factors (e.g., social desirability concerns), properties of implicit assessment factors (e.g., situational malleability), and research design factors (e.g., sampling bias, measurement correspondence).
174
F. Dentale et al. / Personality and Individual Differences 52 (2012) 173–177
1.2. The emotional awareness facilitates implicit to explicit translation Of particular interest here are a series of factors pertaining to awareness (e.g., private self-consciousness, mindfulness, faith in intuition, alexithymia) that are able to facilitate the translation from an implicit to an explicit format. According to Hofmann et al. (2005) such a translation process may depend upon the degree to which people are able to form accurate propositional representations of their underlying associative representations. In other terms, the association-based attitudes (e.g., me-good) are able to be integrated into the propositional attitudes (e.g., I am a good person) only if the subjects are able to achieve a sufficient awareness of the internal and external cues of the implicit activations (Hofmann, Gschwender, & Schmitt, 2005). Therefore, all factors that impair (or ameliorate) the ability to identify the internal (e.g., somatic sensations, feelings, mood changes) or the external cues (e.g., self-observation of spontaneous behaviors) of the implicit associative activations, should also impair (or ameliorate) the quality of the translation process and thus should decrease (or increase) implicit–explicit consistency. Concerning the ability to identify the internal cues of implicit associative activations, previous studies have examined the influence of private self-consciousness and mindfulness on implicit–explicit consistency (Brown & Ryan, 2003; Hofmann, Gschwender, & Schmitt, 2005). Private self-consciousness was defined as the extent to which individuals pay attention to their bodily and emotional experience. Mindfulness refers to the enhanced attention to and awareness of current experience or present reality, which may be reflected in more regular or sustained consciousness of ongoing events and experiences. Empirical findings revealed that high scorers in private self-consciousness and in mindfulness exhibited stronger implicit–explicit consistency than low scorers. With regard to the implicit–explicit self-esteem consistency, Jordan, Whitfield, and Zeigler-Hill (2007) argued that implicit self-esteem (ISE) is based on intuitive self-views that are connected to immediate affective reactions toward the self, whereas explicit self-esteem (ESE) is based on deliberative selfviews that are connected to evaluative judgements. They found that high scorers in faith in intuition, namely people’s confidence about their feelings as bases for decision making (Epstein, Pacini, Denes-Raj, & Heier, 1996), showed a greater consistency between ISE and ESE than low scorers. In a similar vein, Dentale, San Martini, De Coro, and Di Pomponio (2010) found that alexithymia moderates the relationship between implicit and explicit self-esteem. In particular, they demonstrated that both the difficulties to identify and describe feelings increase the discordance between ISE and ESE. These results indicate that emotional awareness plays a relevant role in the translation of implicit self-associations into propositional and explicit formats. Along this line of reasoning, it is reasonable to expect that personality factors that impair emotional awareness may increase the implicit–explicit self-esteem discordance. Among such personality factors we can find certain traits that defend subjects against too extreme emotions, such as the repressive coping style (Weinberger, Schwartz, & Davidson, 1979), the isolation of affect defense style (Bond, Gardner, Christian, & Sigal, 1983) or the dismissing style of attachment (Bowlby, 1973). 1.3. Dismissing attachment and emotional awareness of the self According to the literature on the attachment behavioral system (Bowlby, 1973; Main, Kaplan, & Cassidy, 1985), dismissing subjects are characterized by a defensive emotional regulation that induce an impoverished emotional awareness (e.g., Waller & Scheidt, 2006), especially in close relationships: ‘‘dismissing status of attachment is linked to the disability or to defensive forms of processing and expressing emotions, whereas attachment security
is associated with open, flexible affect expression and the ability to explore and process emotional experiences without employing defensive strategies (Waller & Scheidt, 2006, p. 19)’’. As is well known, recurrent experiences of the attachment figure’s availability and responsiveness would warrant for an evaluation of the self as a person toward whom anyone, and the attachment figure in particular, is likely to respond in a helpful way (Kobak, 1999). To preserve their explicit self-image, dismissing subjects tend to divert their attention from distressful attachment signals (Bowlby, 1973; Dixon, 1981) and to suppress the associated negative emotions (Dozier & Kobak, 1992), protecting themselves from attachment relationships that appear too painful. As a consequence, in spite of attachment experiences of refusal or even threat, dismissing subjects usually showed a fictitious explicit self-image that is independent by actual relationships. Such a defensive process permits them to stabilize the explicit self-esteem in respect to extreme fluctuations that may endanger their mental and physical well-being. As a side effect, in our view, it should emerge as an impaired explicit access to the spontaneous self-directed feelings that make up implicit self-esteem. But what about the implicit associative level? On the basis of dual models of social cognition (Gawronski & Bodenhausen, 2006; Strack & Deutsch, 2004), we have no reasons to assume that dismissing attachment operates by defensively distorting the implicit self-evaluations, as occur for the explicit self-judgments. On the contrary, it seems likely that automatic associations are closely connected to the affective experiences that determine the spontaneous and intuitive selfevaluations. As a consequence, similarly to what has been hypothesized for other moderating factors pertaining to awareness (Brown & Ryan, 2003; Dentale et al., 2010; Hofmann, Gschwender, & Schmitt, 2005; Jordan et al., 2007), dismissing style should impair implicit to explicit translation process, and thus decrease the implicit–explicit self-esteem consistency. Therefore, we expected that dismissing attachment style, but not secure, anxious and fearful/ avoidant styles, increase the discordance between implicit and explicit self-esteem. 1.4. Dismissing style decreases implicit–explicit self-esteem consis tency To test our hypothesis, we administered the Attachment Style Questionnaire (ASQ, Feeney, Noller, & Hanrahan, 1994) to measure five styles of attachment: confidence, discomfort with closeness, relationships as secondary, need for approval, and preoccupation with relationships. As Fossati et al. (2003) noted, of all ASQ dimensions, only relationships as secondary is fully consistent with Bartholomew’s concept of dismissing attachment (Bartholomew & Horowitz, 1991). To assess implicit–explicit consistency, we administered a self-report scale of self-esteem, namely the Rosenberg Self-Esteem Scale (RSES, Rosenberg, 1965), and two parallel forms of IAT for measuring implicit self-esteem (SE-IAT). Our hypothesis is that high scorers in relationships as secondary scale, but not in confidence, preoccupation with relationships, need for approval and discomfort with closeness, exhibit higher discrepancies between implicit self-esteem and explicit self-esteem than low scorers. 2. Methods 2.1. Participants and procedures Participants in the study were 145 undergraduate students (33 males and 112 females) enrolled in a psychology course at the ‘‘Sapienza’’ University of Rome. The age of the respondents ranged
175
F. Dentale et al. / Personality and Individual Differences 52 (2012) 173–177
from 18 to 20 years (M = 20.4, SD = 2.75). Participants came to the laboratory individually, where they first completed a booklet of questionnaires including measures of explicit self-esteem and attachment styles. Then, they were given two parallel forms of SE-IAT. All participants were informed and consented to participate in the study, for which they earned course credits. On completion of the study, all participants were given debriefing statements and thanked for participation. 2.2. Measures 2.2.1. Attachment styles Attachment styles were measured by the ASQ (Feeney et al., 1994), an inventory yielding five subscale scores representing different styles of attachment: Confidence, Preoccupation with Relationships, Need for Approval, Discomfort with Closeness and Relationships as Secondary. Confidence (Confid) (in self and others) reflects a secure attachment orientation. Preoccupation with Relationships (PreoRel), which involves an anxious and dependent approach to relationships, is a core feature of Hazan and Shaver’s (1987) original conceptualization of anxious/ambivalent attachment. Need for Approval (NeedApp) reflects respondents’ need of acceptance and confirmation from others, and characterizes Bartholomew’s fearful and preoccupied individuals. Discomfort with Closeness (Disclos) is a central theme of both Hazan and Shaver’s (1987) conceptualization of avoidant attachment and Bartholomew’s view of fearful subjects. Finally, Relationships as Secondary (RelSec) is the unique dimension that is consistent with Bartholomew’s concept of dismissing attachment. The Italian version of the instrument (Fossati et al., 2003) presented a good internal consistency for the RelSec subscale (a = .73), and moderate to low coefficients for Confid (a = .69), PreoRel (a = .64), NeedApp (a = .69), and Disclos (a = .68). 2.2.2. Explicit self-esteem Explicit self-esteem was measured by the RSES (Rosenberg, 1965). The Italian version of the scale (Prezza, Trombaccia, & Armento, 1997) has proved to be reliable: Cronbach’s alpha and test–retest reliability coefficients were .84 and .76, respectively. 2.2.3. Implicit self-esteem Two parallel forms of SE-IAT (Greenwald & Farnham, 2000) were used, in which the stimuli-words of the target-concept (me vs. not-me) were the same in both forms (e.g., me, self vs. it, that), whereas the stimuli-words of the attribute-dimension (pleasant vs. unpleasant) were emotionally loaded nouns (IATn) in the first form (e.g., luck, joy vs. vomit, agony) and emotionally loaded attributes (IATa) in the second form (e.g., positive, good vs. negative, bad)1. In both SE-IATs the participants performed two types of categorization tasks, with 5 stimuli-words for each category. The words were presented in random order within each block of trials. As described by Greenwald et al. (1998) the entire procedure consisted of seven blocks of trials: 1 (me vs. not-me), 2 (pleasant vs. unpleasant) and 5 (not-me vs. me) were single categorization blocks of 20 trials whereas 3–4 and 6–7 were combined blocks (me or pleasant vs. not-me or unpleasant) of 20 (3–6) and 40 (4–7) trials. Subjects were requested to respond as quickly and accurately as possible to the stimuli-words that appeared on the monitor. Following Greenwald, Nosek, and Banaji (2003), data from blocks 3–4 and 6–7 were used to compute SE-IAT difference scores (Dbiep), according to the built-in error penalty procedure and through the following formula:
Dbiep ¼ ðM IB MCB Þ=SD
1
ð1Þ
A complete description of the test is available from the first author upon request.
where MIB is the mean latency for incompatible blocks, MCB is the mean latency for compatible blocks; SD is the pooled standard deviation for compatible and incompatible blocks. Positive scores indicate high implicit self-esteem and negative scores indicate low implicit self-esteem. The internal consistencies of the two parallel IATs were estimated by a split-half index, based on two partial scores, Dbiep1 and Dbiep2, respectively computed from blocks 3 and 6 (20 + 20 trials), and from blocks 4 and 7 (40 + 40 trials). The correlation between the two parallel measures was r = .44 (p < .01). Since this result is affected by measurement error, we disattenuated the coefficient by taking into account the unreliability of the scales. After correction the correlation increased to .79. A global score of implicit self-esteem was computed by summing the standardized values of the Dbiep score of the two parallel forms. The composite score was used in the following analyses. 3. Results 3.1. Descriptive statistics Descriptive statistics and reliabilities of the variables included in the study are reported in Table 1. All measures were approximately normally distributed, since skewness and kurtosis values were close to |1| (Tabachnick & Fidell, 2001). As can be observed, all self-report measures have acceptable-to-high levels of internal consistency. Reliabilities of the two IAT measures, instead, were not fully adequate (rs-h < .60) but not too different from results commonly found (e.g., Krause, Back, Egloff, & Schmukle, 2011). Table 2 provides the intercorrelations among the study variables. Regarding the ASQ subscales, Confid was negatively and significantly correlated with all the other attachment scales. Anxious (PreoRel and NeedApp) and avoidant (Disclos and RelSec) scales were strongly and positively intercorrelated to each other, but weakly or not correlated between them (this follows the same pattern found in Fossati et al., 2003). The correlation between ISE and ESE was close to zero, in accordance with earlier findings (e.g., Greenwald & Farnham, 2000). Confid was strongly and positively correlated with ESE, while it was unrelated with ISE. The other ASQ subscales (with the exception of RelSec) correlated negatively and significantly with ESE but not with ISE. 3.2. Moderated regression analyses To investigate whether dismissing style moderates the relationship between implicit and explicit self-esteem, a hierarchical moderated regression analysis was conducted. In this analysis, ESE was considered as a criterion variable, in accordance with the theoretical model proposed by Hofmann, Gschwender, & Schmitt (2005) and
Table 1 Descriptive statistics.
RSES Con PWR NFA DWC RAS IATn Dbiep IATa Dbiep
No. of items
Reliability
Mean
SD
Skewness
Kurtosis
10 8 8 7 10 7 – –
.82a .67a .71a .71a .74a .65a .56rs–h .55rs–h
31.58 32.60 30.58 21.24 36.14 13.37 .67 .62
4.71 5.06 6.16 5.75 7.23 4.15 .32 .31
.63 .42 .07 .37 .03 1.08 .42 .17
.03 .11 .42 .38 .11 1.85 .57 .46
Note: a: Cronbach’s alpha; rtt: test–retest; rs-h: split-half; RSES: Rosenberg SelfEsteem Scale; Con: Confidence; PWr: Preoccupation With Relationships; NFA: Need For Approval; DWC: Discomfort With Closeness; RAS: Relationships As Secondary; IATn: Implicit Association Test with emotionally loaded nouns as stimuli; IATa: Implicit Association Test with emotionally loaded attributes as stimuli; Dbiep: SEIAT scores according to the built in error penalty method.
176
F. Dentale et al. / Personality and Individual Differences 52 (2012) 173–177
Table 2 Correlations among study variables. RSES RSES IAT-SE CON PWR NFA DWC RAS
IAT-SE
Con
PWR
1 .04 .10 .12 .01 .11
1 .35** .45** .47** .25**
NFA
DWC
RAS
1 .20* .12
1 .30**
1
1 .03 .61** .38** .63** .25** .14
1 .53** .08 .07
Note: RSES: Rosenberg Self-Esteem Scale; IAT-SE: IAT Self-Esteem; Con: Confidence; PWR: Preoccupation With Relationships; NFA: Need For Approval; DWC: Discomfort With Closeness; RAS: Relationships as Secondary. * p < .05 (two tails). ** p < .01 (two tails). Fig. 1. Explicit Self Esteem (ESE) as a function of Implicit Self-Esteem (ISE) and Relationships as Secondary (RelSec) facet of dismissing attachment style.
coherently with what has been made in other similar studies on the moderating role of factors pertaining to awareness (Brown & Ryan, 2003; Hofmann, Gschwender, & Schmitt, 2005; Jordan et al., 2007). In the first step, RelSec and ISE scores were entered as predictors (which represent main effects). In the second step, their cross product term (RelSecxISE) was entered. The predictors were centered before creating the cross-product vector (Aiken & West, 1991), which represents the interaction effect. Results (Table 3e) showed no significant main effects for ISE and RelSec, and a significant interaction term. The lack of main effects is not surprising but in line with the literature that showed null or weak correlations between ISE and ESE (e.g., Buhrmester, Blanton, & Swann, 2011) on the one hand, and between RelSec and ESE on the other hand (e.g., Feeney et al., 1994; Fossati et al., 2003). The interaction between RelSec and ISE explained a significant incremental amount of variance in ESE (DR2 = .03, p < .05), demonstrating the expected moderating role of dismissing attachment on the ISE-ESE consistency. The negative sign of the beta coefficient for the interaction term indicated that the concordance between ISE and ESE tended to increase as dismissing attachment decreases.
Table 3 Results of moderated regression analyses on the Rosenberg Self-Esteem Scale. Predictors a
b
c
d
e
Step 1 ISE CON Step 2 ISExCON Step 1 ISE PWR Step 2 ISExPWR Step 1 ISE NFA Step 2 ISExNFA Step 1 ISE DWC Step 2 ISExDWC Step 1 ISE RAS Step 2 ISExRAS
b (p)
R2 change
F Change (p)
Df
.38
42.35 (<.001)
2, 137
.00
0.14 (.75)
1, 136
.15
11.78 (<.001)
2, 137
.00
0.30 (.59)
1, 136
.40
47.20 (<.001)
2, 140
.00
0.08 (.78)
1, 139
.06
4.81 (.01)
2, 138
.02
2.49 (.12)
1, 137
.02
1.57 (.21)
2, 137
.03
4.14 (<.05)
1, 136
To interpret the interaction effect, the simple slopes for the relationship between ISE and ESE were tested at one standard deviation below (1 SD) and one standard deviation above (+1 SD) the RelSec mean, in accordance with Cohen and Cohen’s (1983) guidelines (see Fig. 1). At 1 SD, the relation between ISE and ESE was positive and significant (b = .217; t136 = 2.32, p < .01), indicating a concordance of medium size (Cohen & Cohen, 1983). On the other hand, at +1 SD the relation was negative and not significant (b = .121; t136 = .99, p = .16). To further examine the relationship between attachment styles and implicit–explicit self-esteem concordance, the regression analysis was repeated for each of the remaining attachment scales (Confid – see Table 3a; PreoRel – see Table 3b; NeedApp – see Table 3c; Disclos – see Table 3d). In all of the four regressions, the main effects of the attachment scales were significant at the first step. In line with the literature (e.g., Feeney et al., 1994; Fossati et al., 2003), PreoRel, NeedApp, and Disclos were found to be negatively related with ESE, whereas Confid had a positive relation. Differently, in accordance with previous results (Buhrmester et al., 2011), the main effect of ISE was not significant in each regression. At the second step, the interaction terms were not significant. Only the DisclosxISE cross product term exhibited a certain tendency to be significant (p = .12) with a small and negative beta coefficient. In sum, the five regressions indicated that only RelSec (and not Confid, PreoRel, NeedApp and Disclos) showed a significant moderating effect on the relationship between implicit and explicit self-esteem.
.06 (.34) .62 (<.01) -.03 (.70) .01 (.85) .38 (<.001) .04 (.59) .05 (.43) .64 (<.001) .02 (.78) .041 (.62) .252 (<.01) .13 (.12) .05 (.56) .14 (.14) .17 (<.05)
Note: ISE: Implicit Self-Esteem; CON: Confidence; PWR: Preoccupation with Relationships; NFA: Need for Approval; DWC: Discomfort with Closeness; RAS: Relationships as Secondary.
4. Conclusions The present study showed that dismissing/avoidant attachment (as measured by the ASQ relationships as secondary scale) increases implicit–explicit self-esteem discordance. As an illustration of our results, it was found that at one standard deviation below the mean of the RelSec scale (i.e. for individuals who present a low dismissing attachment tendency), there was a positive and significant, though moderate, correlation between implicit and explicit self-esteem. At one standard deviation above the mean, instead (i.e. for individuals who present a high dismissing attachment tendency), the correlation was not significant and weakly negative. As expected, of all ASQ scales, only relationships as secondary emerged as a moderator of the relationship between implicit and explicit self-esteem, although data showed that Disclos (i.e. fearful/avoidant dimension of ASQ) tends to increase the discordance between the two measures. The lack of interaction between Disclos and ISE may be explained by noting that this scale catches some avoidant behaviors that are accompanied by a certain degree of awareness of negative emotions in attachment interactions.
F. Dentale et al. / Personality and Individual Differences 52 (2012) 173–177
Taken together, these results give further support to the role ascribed to emotional awareness in the propositional (explicit) translation of affective (implicit) evaluations (Hofmann, Gschwender, & Schmitt, 2005; Jordan et al., 2007), and are compatible with Hofmann et al.’s (2005) working model of implicit–explicit consistency, positing awareness in a general group of moderators. In another possible interpretation, the moderating effect of dismissing attachment may depend on adjustment processes (Hofmann, Gschwender, & Schmitt, 2005) that bias explicit self-esteem scores, similarly to what happens as a consequence of social-desirability, impression management or self-deception effects. If so, dismissing individuals may be a part of those people who defend their self-image using compensatory responses (e.g., Jordan, Spencer, Zanna, Hoshino-Browne, & Correll, 2003). In other words, the moderating effect we found might be ascribed to a proneness to self-enhancement in the explicit assessment, rather than to an impaired emotional awareness that decreases the quality of the implicit to explicit translation process. In this vein, perhaps not coincidentally, high scorers on the RelSec scale showed a weak, though not significant, negative correlation between ISE and ESE. That is, for individuals with a high dismissing attachment, lower scores on implicit self-esteem appear to be associated to higher levels of explicit self-esteem. Looking at the recent literature on implicit self-esteem, one cannot exclude that this tendency would be due to a kind of compensatory bias. Further studies are necessary to examine the extent to which such a negative correlation is meaningful and generalizable across different samples. A limitation of the present study is that the attachment styles were assessed only via a self-report inventory and exclusively in a dimensional perspective. If we maintain that dismissing subjects are impaired in the explicit processing of affective information, their ability to provide valid self-reports on the ASQ may also be impaired to some extent, as this measure requires a great deal of introspective ability. A relevant confirmation of our results may come from investigations using clinical assessment of attachment states of mind through the Adult Attachment Interview (Steele & Steele, 2008). A further limitation of the present study is that implicit self-esteem was measured only through classical IAT, in which two target categories are opposed to each other. To examine the robustness of our results, it may be appropriate to conduct further studies using other implicit techniques, such as the Go/No-Go Association Task (GNAT, Nosek & Banaji, 2001) or the Single-Category IAT (SC-IAT, Karpinski & Steinman, 2006). Notwithstanding its limits, this contribution supports the hypothesis – traditionally confined into clinical assumptions – that dimensions relevant for interpersonal regulation of affects, such as attachment working models or defensive strategies, are to be considered in the theory and assessment of self-esteem. References Aiken, L. S., & West, S. G. (1991). Multiple regression: testing and interpreting interactions. London: Sage. Bartholomew, K., & Horowitz, L. M. (1991). Attachment styles among young adults: a test of a four-category model. Journal of Personality and Social Psychology, 61, 226–244. Bond, M. P., Gardner, S. T., Christian, J., & Sigal, C. (1983). Empirical study of selfrated defence styles. Archives of General Psychiatry, 46, 455–460. Bowlby, J. (1973). Attachment and loss: Vol. 2. Separation. New York: Basic Books. Brown, K. W., & Ryan, R. M. (2003). The benefits of being present: Mindfulness and its role in psychological well-being. Journal of Personality and Social Psychology, 84, 822–848. Buhrmester, M. D., Blanton, H., & Swann, W. B. Jr., (2011). Implicit self-esteem: Nature, measurement and a new way forward. Journal of Personality and Social Psychology, 100, 365–385. Cohen, J., & Cohen, P. (1983). Applied multiple regression/correlation analysis for the behavioral sciences. Hillsdale, NJ: Erlbaum.
177
Dentale, F., San Martini, P., De Coro, A., & Di Pomponio, I. (2010). Alexithymia increases the discordance between implicit and explicit self-esteem. Personality and Individual Differences, 49, 762–767. Dixon, N. F. (1981). Preconscious processing. England, Wiley: Chichester. Dozier, M., & Kobak, R. R. (1992). Psychophysiology in attachment interview: converging evidence for deactivating strategies. Child Development, 63, 1473–1480. Epstein, S., Pacini, R., Denes-Raj, V., & Heier, H. (1996). Individual differences in intuitive-experiential and analytical-rational thinking styles. Journal of Personality and Social Psychology, 71, 390–405. Feeney, J. A., Noller, P., & Hanrahan, M. (1994). Assessing adult attachment. In M. B. Sperling & W. H. Berman (Eds.), Attachment in adults: Clinical and developmental perspectives (pp. 128–152). New York: Guilford Press. Fossati, A., Feeney, J. A., Donati, D., Donini, M., Novella, L., Bagnato, M., et al. (2003). On the dimensionality of the Attachment Style Questionnaire in Italian clinical and non-clinical participants. Journal of Social and Personal Relationships, 20, 55–79. Gawronski, B., & Bodenhausen, G. V. (2006). Associative and propositional processes in evaluation: An integrative review of implicit and explicit attitude change. Psychological Bulletin, 132, 692–731. Greenwald, A. G., & Farnham, S. D. (2000). Using the Implicit Association Test to measure self-esteem and self-concept. Journal of Personality and Social Psychology, 79, 1022–1038. Greenwald, A. G., McGhee, D. E., & Schwartz, J. K. L. (1998). Measuring individual differences in implicit cognition: The Implicit Association Test. Journal of Personality and Social Psychology, 74, 1464–1480. Greenwald, A. G., Nosek, B. A., & Banaji, M. R. (2003). Understanding and using the Implicit Association Test: An improved scoring algorithm. Journal of Personality and Social Psychology, 85, 197–216. Hazan, C., & Shaver, P. (1987). Conceptualizing romantic love as an attachment process. Journal of Personality and Social Psychology, 52, 511–524. Hofmann, W., Gschwender, T., Nosek, B. A., & Schmitt, M. (2005). What moderates implicit–explicit consistency? European Review of Social Psychology, 16, 335–390. Hofmann, W., Gschwender, T., & Schmitt, M. (2005). On implicit–explicit consistency: The moderating role of individual differences in awareness and adjustment. European Journal of Personality, 19, 25–49. Jordan, C. H., Spencer, S. J., Zanna, M. P., Hoshino-Browne, E., & Correll, J. (2003). Secure and defensive self-esteem. Journal of Personality & Social Psychology, 85, 969–978. Jordan, C. H., Whitfield, M., & Zeigler-Hill, V. (2007). Intuition and the correspondence between implicit and explicit self-esteem. Journal of Personality and Social Psychology, 93, 1067–1079. Karpinski, A., & Steinman, R. B. (2006). The single category implicit association test (SC-IAT) as a measure of implicit social cognition. Journal of Personality and Social Psychology, 91, 16–32. Kobak, R. (1999). The emotional dynamics of dysruptions in attachment relationships. In J. Cassidy & P. R. Shaver (Eds.), Handbook of attachment: Theory, research and clinical applications (pp. 21–43). New York-London: Guilford Press. Koole, S. L., Dijksterhuis, A., & van Knippenberg, A. (2001). What’s in a name: Implicit self-esteem and the automatic self. Journal of Personality and Social Psychology, 80, 669–685. Krause, S., Back, M. D., Egloff, B., & Schmukle, S. C. (2011). Reliability of implicit selfesteem measures revisited. European Journal of Personality, 25, 239–251. Main, M., Kaplan, N., & Cassidy, J. (1985). Security in infancy, childhood, and adulthood: A move to the level of representation. Growing points in attachment theory and researchI. Bretherton & E. Waters (Eds.). Monographs of the Society for Research in Child Development, 50, 66–106. Nosek, B. A., & Banaji, M. R. (2001). The go/no-go association task. Social Cognition, 19, 625–666. Nosek, B. A., & Smyth, F. L. (2007). A multitrait-multimethod validation of the Implicit Association Test: Implicit and explicit attitudes are related but distinct constructs. Experimental Psychology, 54, 14–29. Prezza, M., Trombaccia, F. R., & Armento, L. (1997). La scala dell’autostima di Rosenberg. Traduzione e validazione italiana [The Rosenberg Self-Esteem Scale: Italian translation and validation]. Bollettino di Psicologia applicata, 223, 35–44. Rosenberg, M. (1965). Society and the adolescent self-image. Princeton, NJ: Princeton University Press. Steele, H., & Steele, M. (2008). Clinical applications of the adult attachment interview. New York-London: Guilford Press. Strack, F., & Deutsch, R. (2004). Reflective and impulsive determinants of social behavior. Personality and Social Psychology Review, 8, 220–247. Tabachnick, B. G., & Fidell, L. S. (2001). Using multivariate statistics (4th ed.). Needham Heights, MA: Allyn & Bacon. Waller, E., & Scheidt, C. E. (2006). Somatoform disorders as disorders of affect regulation: A development perspective. International Review of Psychiatry, 18, 13–24. Weinberger, D. A., Schwartz, G. E., & Davidson, R. J. (1979). Low-anxious, high anxious and repressive coping styles: Psychometric patterns and behavioral responses to stress. Journal of Abnormal Psychology, 88, 369–380.