Está en la página 1de 3

Personality and Individual Differences 95 (2016) 159161

Contents lists available at ScienceDirect

Personality and Individual Differences


journal homepage: www.elsevier.com/locate/paid

Short Communication

Relations between explicit and implicit self-esteem measures


and self-presentation
Maarit Johnson
Department of Psychology, Stockholm University, SE-106 91 Stockholm, Sweden

a r t i c l e

i n f o

Article history:
Received 25 November 2015
Received in revised form 14 February 2016
Accepted 17 February 2016
Available online 27 February 2016
Keywords:
Self-presentation
Explicit self-esteem
Implicit self-esteem
Validity

a b s t r a c t
Three explicit self-esteem measures with different conceptualization and item content were compared with regard to their association with implicit self-esteem (SE) and positive self-presentation. The results revealed a pattern where affectiveexperiential basic SE appears to have more similarity with implicit self-esteem than
cognitiveevaluative general SE measures. Basic SE was the only explicit SE measure that predicted signicantly
implicit self-esteem. Most of the self-presentational styles lacked association with implicit SE and basic SE but
played a substantial role for the general SE scales. The results suggest the importance of considering self-report
measures potential to tap unbiased self-esteem.
2016 Elsevier Ltd. All rights reserved.

1. Introduction

1.1. Self-esteem mediated via affective contra cognitive processes

In research on self-esteem it is crucial for validity of the results to


develop concepts and measures that capture a person's genuine selfregard rather than a favorable self-presentation (Farnham, Greenwald,
& Banaji, 1999; Forsman & Johnson, 1996). It is widely agreed that selfreport measures of self-esteem are plagued by various selfpresentational inuences, such as self-deception or impression management (Farnham et al., 1999; Paulhus, 1991). In contrast, implicit selfesteem, often measured via reaction times of semantic associations, is
considered a more unbiased indication of self-esteem than explicit controllable measures (Buhrmester, Blanton, & Swann, 2011; Dijksterhuis &
Nordgren, 2006). However, despite their appeal as objective predictors
of behavior, implicit measures are known to have certain impediments.
For instance, implicit and explicit measures have shown weak correlations and appear to tap related but distinct constructs (Buhrmester
et al., 2011; Hoffman, Gawronski, Gschwender, Le, & Schmitt, 2005).
Therefore, it is important to examine further the relative potential of unbiased responses in available self-report measures of self-esteem
(Johnson, 2013). The present study compared three explicit self-esteem
measures, with different conceptualization and item content, with regard
to their closeness to implicit self-esteem and their susceptibility to selfpresentational inuences.

Inconsistent support found for predictive validity of explicit measures


of general self-esteem scales is often explained by their susceptibility to
self-presentation bias (Farnham et al., 1999; Paulhus, 1991). As a consequence, indirect measures, such as Implicit Association Test (IAT;
Greenwald & Farnham, 2000), were developed assuming an existence
of a valenced association toward oneself. The distinction between implicit
and explicit measurement techniques is based on models of dual processing of information where peoples' attitudes, for instance to the self, are
considered a result of automatic affective processes versus rational cognitive processes (e.g., Dijksterhuis & Nordgren, 2006). Accordingly, implicit
measures of self-esteem refer to uncontrolled responding whereas selfreports allow control of responses (Farnham et al., 1999). However, instead of emphasizing distinctive measurement techniques the focus
should be laid on more or less controllable judgment processes (De
Houwer, 2005; Ranganath, Smith, & Nosek, 2008). Indeed, the conceptualisation and item content of self-report measures can bear on the
controllability and sensitivity for self-presentational alterations of responses (see e.g., Johnson, 2013).

E-mail address: mtjn@psychology.su.se.

http://dx.doi.org/10.1016/j.paid.2016.02.045
0191-8869/ 2016 Elsevier Ltd. All rights reserved.

1.1.1. Two kinds of explicit measures of self-esteem


The most used self-esteem measures refer to general or global selfesteem (e.g., Rosenberg, 1965). These mainly empirically derived scales
capture people's self-views based on perceived popularity, skills, or appearances by using items such as I feel I have a number of good qualities with low subtlety and high fakability (Burkhart, Christian, &
Gythner, 1978; Farnham et al., 1999). These kinds of direct statements
are prone to trigger self-deception tendencies that protect the selfimage for the moment, but obscure the presence of genuine self-

160

M. Johnson / Personality and Individual Differences 95 (2016) 159161

regard and elicit spuriously high self-esteem scores (Forsman &


Johnson, 1996; Johnson, 2013; Paulhus, 1991). It has also been shown
that a global component extracted in factor analyses of self-esteem
items to a certain extent may be accounted for by various responsesets (e. g., Tomas & Oliver, 1999).
To capture a person's self-regard without references to skills and
competencies, Forsman and Johnson (1996) developed a construct
and measure of basic self-esteem. Theoretically anchored in early selfdevelopment basic self-esteem captures an individual's affectiveexperiential perception of oneself, mediated via autonomic rather than
cognitive-evaluative processes (see Forsman & Johnson, 1996). So,
reecting stable non-evaluative representations arising from early
socialisation experiences (Johnson, 1998; Johnson & Patching, 2013)
basic self-esteem appears to have a common ground with implicit
self-esteem (DeHart, Pelham, & Tennen, 2006). Moreover, statements
such as I feel safe or I can freely express what I feel are, due to their indirect and non-evaluative character, considered less likely to trigger
motivational distortions in responses than direct evaluative items
(Farnham et al., 1999; Johnson, 2013). Indeed, Johnson (2013) showed
that people seem more inclined to reject negative information about
themselves when responding to a global self-esteem scale than when
responding to the Basic self-esteem scale. On these grounds, basic selfesteem appears less sensitive for self-presentational biases than global
self-esteem and may reect some aspects of self-regard intended to be
captured by implicit measures (De Houwer, 2005; Koole & DeHart,
2007).
The present study focuses on explicit measures' relative potential to
tap unbiased self-esteem (SE) by examining: (i) implicit SE's association
with two widely used general SE measures and with the affectiveexperiential concept and measure of basic SE; (ii) the different explicit SE
measures' association with self-presentational styles. It was hypothesized that basic SE, as compared to general SE has stronger association
with implicit SE and weaker association with self-presentational variables (see e. g., Hoffman et al., 2005).
2. Methods
2.1. Participants
In total 168 social science undergraduates participated in the study.
They were 70 males and 98 females aged from 20 to 59 years (M = 27.2;
SD = 4.6).
2.2. Measures
2.2.1. Explicit SE and self-presentation measures
Basic self-esteem was measured with a short form of the Basic Selfesteem (BSE) Scale (Forsman & Johnson, 1996), e.g., I feel relaxed when
I'm with other people (Cronbach's alpha = .82). Global general selfesteem was measured with a 25-item version of Coopersmith's (1981)
Self-Esteem (CSE) Inventory, e.g., I'm not as nice and good looking as
most people, reversed coding, (alpha = .79), and with Rosenberg's
Self-Esteem (RSE) Scale (1965) e.g., I feel that I have a lot to be proud
of, (alpha = .84). All SE scales had equal number of positive and negative wordings of items. The Balanced Index of Desirable Responding
(Paulhus, 1991) was used to measure impression management
(alpha = 0.74) and self-deception, (alpha = .76). A short version of
Crowne and Marlowe's Social Desirability Scale (Strahan & Gerbasi,
1972) was used to measure socially correct responding (alpha = .71).
The items of these scales were randomly mixed in the questionnaire.
A 5-point Likert scale was used ranging from strongly disagree (1) to
strongly agree (5). Finally, NPI-16 (Ames, Rose, & Anderson, 2006)
was used to measure narcissism, (alpha = .77). The score is the mean
score of the pairs of items referring to narcissistic and non-narcissistic
responses. There was no item overlap between the different selfreport measures used in the present study.

2.2.2. Implicit SE measure


Implicit Association Test (IAT) for self-esteem (Greenwald &
Farnham, 2000) was used to assess implicit self-esteem. In the present
study a free version of IAT (Meade, 2009) adapted for self-esteem was
used. The stimuli-words used referred to Self (e.g., me, mine) vs. Others
(e.g., they, them) and to emotional attributes Pleasant (e.g., love, happy)
vs. Unpleasant (e.g., terrible, pain). Each participant performed two
kinds of categorization tasks, with ve stimuli for each category. The
words were presented in random order within each block of trials.
The procedure comprised seven blocks of trials in total: 1 (Self vs.
Others), 2 (Pleasant vs. Unpleasant) and 5 (Others vs. Self) were single
categorization blocks (practicing) of 20 trials, whereas 34 and 67
were combined blocks (Self or Pleasant vs. Others or Unpleasant) of
20 (34) and 40 (47) trials. Participants were instructed to respond
(by key pressing) as fast and accurately as possible to the stimuluswords on the computer screen. Following the algorithm of Greenwald,
Nosek, and Banaji (2003), data from blocks 34 and 67 were used to
obtain IAT difference scores, according to the built-in error penalty
method. Positive scores indicate high implicit SE whereas negative
scores indicate low implicit SE. The internal consistency of the scores
(see Greenwald & Farnham, 2000), in the present study was .62.
2.3. Procedure
After being individually seated in a small room with a computer, the
participants were briey informed about the study and that their answers would be treated anonymously. Half of the participants completed the IAT rst and half completed the questionnaire rst.
3. Results
Table 1 shows inter correlations (Pearson, two-tailed) of major interest in the study. The explicit SE measures were highly correlated,
however, the global scales correlated stronger with each other than
with BSE; RSE showed signicantly stronger relation to CSE than to
BSE. Further, of the explicit measures, only BSE correlated signicantly
with implicit self-esteem (IAT). This correlation was signicantly stronger than the correlation between IAT and RSE.
All explicit SE measures correlated signicantly with social desirability but both BSE and IAT showed signicantly weaker correlation with
social desirability than both global scales. BSE displayed higher degree
of socially desirable responses than IAT but the difference was not signicant. Both BSE and IAT showed weaker associations with selfdeception than the global scales, however, the difference in correlations
was signicant only regarding RSEs association with self-deception.
Only CSE correlated signicantly with impression management whereas all SE measures correlated signicantly with narcissism, except IAT.

Table 1
Correlations between explicit and implicit SE and self-presentation measures.

1. BSE
2. RSE
3. CSE
4. IAT
5. Social desirability
6. Self-deception
7. Impression management
8. Narcissism

.68***#
.79***
.28**#
.24*##
.16##
.12
.27**

.93***##
.05##
.49***#
.40***#
.19*
.36***#

.15
.52***#
.31**
.26**
.42***#

.05##
.12##
.09
.14##

Note. Coefcients marked with # are signicantly different from those marked with ##. Zs
were signicant, p b .05; Steiger, 1980, formula 15. Intercorrelations between the self-presentation scales were all signicant, p b .01, but are not presented in the table, for clarity.
BSE = basic self-esteem; CSE = Coopershmith's self-esteem, RSE = Rosenberg's selfesteem.
p b .05.
p b .01.
p b .001.

M. Johnson / Personality and Individual Differences 95 (2016) 159161

IAT's weak association with narcissism differed signicantly from the


correlations obtained with CSE and RSE, but not from that obtained
with BSE.
Three multiple regression analyses (stepwise) were performed on
the scores of each explicit SE measure with social desirability, selfpresentation scales, and narcissism as predictors. The model with BSE
as criterion (R = .29; R2 = .09) revealed that the only signicant predictor of BSE was narcissism ( = .26; t = 1.99, p b .05). The model with
RSE as criterion (R = .47; R2 = .23) revealed three signicant predictors
of which the strongest was self-deception ( = .37; t = 2.84, p b .01)
followed by social desirability ( = .28; t = 2.10, p b .01), and narcissism ( = .23; t = 2.03, p b .05). Finally, the model with CSE as criterion
(R = 50; R2 = .25) revealed three signicant predictors, of which the
strongest was social desirability ( = .42; t = 2.87, p b .01) followed
by narcissism ( = .34; t = 2.72, p b .01) and self-deception ( = .21;
t = 1.98, p b .05). Impression management was not found to be a significant predictor of any of the self-report measures of self-esteem.
Multicollinearity statistics of these analyses showed that all VIF values
were b2.
4. Discussion
The present study compared three explicit measures of self-esteem
(SE) with regard to their association with implicit SE and selfpresentation variables. The results indicate that affective-experiential
basic SE appears distinguished from the two self-evaluative general SE
measures, by its stronger association with implicit SE (IAT) and by its
weaker association with self-presentational styles.
Overall, the results suggest a pattern where basic SE appears to have
more similarity with implicit SE than the global measures. Most of the
self-presentational styles lacked association with both IAT and basic
self-esteem but played a signicant role for the general SE scales. That
narcissism was positively associated with all SE scales, may bear on
the narcissism scale used in the present study. Whether the established
scales tap an adjusted or maladjusted narcissism has been a controversy
in the literature (Rose, 2002). Further, that only Coopersmith's SE was
associated with impression management may reect this SE scale's relative emphasis on skills and competencies (Burkhart et al., 1978).
In main, the present ndings support the hypotheses and resonate
with previous research (Hoffman et al., 2005; Johnson, 2013). The ndings suggest that the basic SE measure appears to evade some sources of
bias which are more readily triggered by direct evaluative global scales
(see also Johnson, 2013). This robustness can pertain to different but
intertwined determinants of psychometric relevance. First, referring to
emotional candor, integrity, and trust, considered the essence of selfesteem (see Forsman & Johnson, 1996), this scale's fakability and face
validity is low, and the respondent might not be aware what the test
is measuring (Burkhart et al., 1978). Second, though the conceptual content of implicit SE is still unclear (Buhrmester et al., 2011), basic SE, arising from early reactions to environmental inputs, may share some latent
construct with implicit SE (DeHart et al., 2006; Ranganath et al., 2008).
The present results suggest closeness between implicit and basic SE
and disparity between these two measures and global measures. However, it should be kept in mind that IAT measures are known to be noisy
and have not shown very consistent and stable results (Buhrmester
et al., 2011). Moreover, while the mutual association between implicit
and basic SE suggests a conceptual correspondence, the implicitexplicit links may nonetheless reect various other factors such as motivation
and spontaneity during self-report (e.g., Hoffman et al., 2005). Forthcoming studies, including additional implicit and explicit SE measures
might elucidate further the implications of the patterns found in this
study.

161

To conclude, the concept and measure of basic self-esteem appears


less susceptible for self-presentation bias than general self-esteem measures (see also Johnson, 2013) and this affective experiential selfesteem may reect aspects of implicit self-esteem. Despite the further
development of implicit measures, self-report scales of self-esteem
will remain central in research practice. Therefore, for improved validity
of self-esteem studies the present results encourage a further consideration of explicit measures' potential to capture a person's genuine selfregard rather than favorable self-presentation. Self-esteem referring to
cognitive evaluations of one's qualities might be increasingly replaced
by theoretically based concepts and assessments capturing genuine
self-acceptance.
References
Ames, D. R., Rose, P., & Anderson, C. P. (2006). The NPI-16 as a short measure of narcissism. Journal of Research in Personality, 40, 440450.
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, 365385.
Burkhart, B., Christian, W., & Gythner, M. (1978). Item subtlety and faking on the MMPI: A
paradoxical relationship. Journal of Personality Assessment, 42, 7680.
Coopersmith, S. (1981). The antecedents of self-esteem. Palo Alto, CA: Consulting Psychologists Press.
De Houwer, J. (2005). What are implicit measures and direct measures of attitudes? Social
Psychological Review, 7, 1820.
DeHart, T., Pelham, B. W., & Tennen, H. (2006). What lies beneath: Parenting style and implicit self-esteem. Journal of Experimental Social Psychology, 42, 117.
Dijksterhuis, A., & Nordgren, L. F. (2006). A theory of unconscious thought. Perspectives on
Psychological Science, 1, 95109.
Farnham, S. D., Greenwald, A. G., & Banaji, M. R. (1999). Implicit self-esteem. In D. Abrams,
& M. Hoss (Eds.), Social identity and social cognition (pp. 230248). Oxford, UK:
Blackwell.
Forsman, L., & Johnson, M. (1996). Dimensionality and validity of two scales measuring
different aspects of self-esteem. Scandinavian Journal of Psychology, 37, 115.
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,
10221038.
Greenwald, A. G., Nosek, B. A., & Banaji, M. R. (2003). Understanding and using the implicit association test: I. An improved scoring algorithm. Journal of Personality and Social
Psychology, 85, 197216.
Hoffman, W., Gawronski, B., Gschwender, T., Le, H., & Schmitt, M. (2005). A meta-analysis
on the correlation between the Implicit Association Test and explicit self-report measures. Personality and Social Psychology Bulletin, 10, 13691385.
Johnson, M. (1998). Self-esteem stability: The importance of basic self-esteem and competence strivings for the stability of global self-esteem. European Journal of Personality, 12, 103116.
Johnson, M. (2013). Patterns of extreme responses to items in self-esteem scales: Does
conceptualisation and item content matter? Personality and Individual Differences,
55, 622625.
Johnson, M., & Patching, G. R. (2013). Self-esteem dynamics regulate the effects of feedback on ambition. Individual Differences Research, 11, 4458.
Koole, S. L., & DeHart, T. (2007). Self-affection without self-reection: Origins, mechanisms, and consequences of implicit self-esteem. In C. Sedikides, & S. Spencer
(Eds.), The self in social psychology (pp. 3686). New York: Psychology Press.
Meade, A. W. (2009). FreeIAT: An open-source program to administer the implicit association test. Applied Psychology Measurement, 33, 643.
Paulhus, D. L. (1991). Measurement and control of response bias. In J. P. Robinson, & P. R.
Shaver (Eds.), Measures of personality and social psychological attitudes. 1. (pp. 1759).
San Diego, CA: Academic Press.
Ranganath, K. A., Smith, C. T., & Nosek, B. (2008). Distinguishing automatic and controlled
components of attitudes from direct and indirect measurement methods. Journal of
Experimental Social Psychology, 44, 386396.
Rose, P. (2002). The happy and unhappy faces of narcissism. Personality and Individual
Differences, 33, 379391.
Rosenberg, M. (1965). Society and the adolescent self-image. Princeton, NJ: Princeton University Press.
Steiger, J. H. (1980). Tests of comparing elements of a correlation matrix. Psychological
Bulletin, 87, 245251.
Strahan, R., & Gerbasi, C. K. (1972). Short homogenous version of the MarloweCrowne
social desirability scale. Journal of Clinical Psychology, 28, 191193.
Tomas, J., & Oliver, A. (1999). Rosenberg's self-esteem scale: Two factors or method effects. Structural Equation Modeling: A Multidisciplinary Journal, 6, 8498.

También podría gustarte