Self-deception and discrepancies in self-evaluation

Highlights

• Tested the validity of the Self-Deception Enhancement (SDE) scale.

• Participant (N = 150) completed explicit and implicit measures of self-evaluation.

• Higher SDE was associated with a more positive explicit self-evaluation.

• Higher SDE was associated with a more negative implicit self-evaluation.

• Overall, evidence indicates that the SDE measures discrepancies in self-evaluation.

Abstract

Self-deception is an intricate psychological construct, grounded in relatively little empirical evidence as an individual difference variable. The present research offers a direct test of one of the central scales of self-deception—the Self-Deception Enhancement (SDE) scale—by comparing explicit and implicit self-evaluations. Using the Affect Misattribution Procedure (AMP), participants (N = 150) reported about their implicit self-evaluation. These ratings were compared to explicit self-reports about self-evaluation. Results showed that higher self-deception was associated with a larger discrepancy between explicit and implicit self-evaluations. Specifically, higher self-deception was associated with a more positive explicit self-evaluation but with a more negative implicit self-evaluation. These findings offer one of the first direct tests of the SDE, supporting its viability.

Introduction

One of the strongest predictors of positive self-evaluation is self-deception (Paulhus, 1998, Taylor and Brown, 1988). Notwithstanding, unlike other constructs associated with self-evaluation, self-deception entails a paradox. Traditional views to this construct depict self-deceivers as explicitly believing that they own desirable attributes, while simultaneously implicitly holding opposing beliefs about their undesirability, which they actively suppress (Gur & Sackeim, 1979). This duality is at the core of what the construct of self-deception represents, yet not much research addressed the validity of the scales intended to measure it as a personality construct (e.g., Paulhus & Buckels, 2012). The present research aims to partly fill this void.

Self-deception is controversial theoretically and difficult to support empirically (Mele, 1997). Notwithstanding, experimental evidence has documented instances of self-deceptive behavior (e.g., Fernbach et al., 2014, Mijović-Prelec and Prelec, 2010, Quattrone and Tversky, 1984). In a classic experiment, Gur and Sackeim (1979) compared participants' verbal and physiological responses to the sound of their own voice (building on the assertion that people find their recorded voice aversive). Results showed that some participants misidentified their voice in their (explicit) verbal reports but were responsive in their (implicit) physiological responses, implying that at some level, correct identification was made. That is, those individuals held simultaneously contrasting beliefs, one of which was denied conscious awareness. Recently, Chance, Norton, Gino, and Ariely (2011) showed that people who cheated in an ability test self-deceived themselves into believing that they have high abilities. Their biased conviction brought them to erroneously predict their future performance, even at a cost of losing money for making inaccurate predictions. In both experiments, individual differences in self-deception exacerbated the effects, implying that individuals vary in the proneness to self-deceive.

In recent years, the most frequently used self-deception scale has been the Self-Deception Enhancement scale (SDE; Paulhus, 1984, Paulhus, 1991), which taps ego-enhancement expressed in exaggerated claims of positive attributes. Within the broad domain of self-enhancement, individual differences in self-deception represent a stable, inward directed, non-conscious mechanism to preserve a desirable self-evaluation (Paulhus, 1991). The self-deceivers are said to hold positive self-views that they believe to be true, but do not reflect their less conscious beliefs (Paulhus, 1984).

Research on the SDE scale has shown that high scorers experience (self-favoring) discrepancies between their positive self-evaluation and the way they are perceived by others (Paulhus, 1998, Uziel, 2014). More so, they show strong biases in their perceptions of their abilities on performance tests (Chance et al., 2011, Paulhus et al., 2003). That is, a basis exists to suggest that the SDE scale measures a discrepancy in one's self- perception. However, direct evidence is lacking for the SDE scale concerning this core aspect of self-deception—the co-existence of contrasting beliefs about oneself (Chance & Norton, 2015).

To address this issue there is a need to explore the discrepancy between explicit and implicit self-views. The few studies that have done so yielded inconclusive results. In Greenwald and Farnham (2000), the Implicit Association Test (IAT) served as an implicit measure of self-esteem. IAT performance was compared to explicit reports of self-esteem. The findings showed that SDE was positively correlated with explicit self-esteem, and also, less strongly, with implicit self-esteem, implying that SDE was not associated with a notable discrepancy. However, using a similar methodology, Jordan, Spencer, Zanna, Hoshino-Browne, and Correll (2003) reached the opposite conclusion, whereas Riketta (2005) reported mixed findings.

The inconclusive findings from the IAT-based studies could be attributed to general problems with the IAT. Notwithstanding, in the present context, the main limitation of the IAT is that it does not allow independent direct assessment of implicit beliefs. In the IAT, attitudes toward oneself are inferred relative to attitudes toward anyone that is “not me”. Thus, IAT findings could reflect more strongly the attitude toward the “not me” group than the attitude about oneself (Karpinski & Steinman, 2006).

The present research sought to explore whether individual differences in self-deception (measured with the SDE) form a basis for a discrepancy between explicit and implicit self-evaluation, using a methodology that allows independent assessment of one's self-views. Specifically, we used the Affect Misattribution Procedure (AMP; Payne, Cheng, Govorun, & Stewart, 2005) to measure implicit self-evaluation. The AMP is a reliable measure of implicit attitudes (Payne & Lundberg, 2014), which has proven effective as a measure of implicit self-evaluation (Schreiber, Bohn, Aderka, Stangier, & Steil, 2012). To the extent that self-deception reflects the existence of two incompatible self-related beliefs, we expected self-deception to be associated with a positive explicit self-evaluation alongside a negative implicit self-evaluation.

The focus of the present research is on self-deception. Notwithstanding, to gain a broader perspective, we have also measured narcissism. Theoretically, narcissism too embodies a discrepancy between explicit and implicit self-evaluation (Campbell et al., 2007, Morf and Rhodewalt, 2001). Narcissistic individuals are described as being overly self-centered, dominant, and as having a sense of entitlement, grandiosity, and vanity (Raskin & Terry, 1988). Narcissists’ positive self-view is often attributed to self-deceptive processes (Zeigler-Hill, 2006). If this is the case, narcissists—like self-deceiving individuals—should exemplify a discrepancy between their explicit and implicit self-view. However, existing research, which has mostly utilized the IAT for implicit self-evaluation, has been inconclusive (Bosson et al., 2008). The present research thus aims to inform the literature on narcissism by using a different methodology and by considering SDE, expecting the discrepancy for narcissism to be subsidiary to the effect found for self-deception (the core process).

The main dependent variable of the present investigation is the discrepancy between explicit and implicit self-evaluations. Difference scores are an intuitive approach to operationalize discrepancy and are often applied in studies involving explicit and implicit self-concepts (Briñol et al., 2006, Kehr, 2004). However, they are also criticized for not adequately addressing the source of the difference. To address this criticism, analyses of difference score were supplemented by multiple regressions involving implicit and explicit scores separately. Materials, data, and code are available at the Open Science Framework (OSF): https://osf.io/cw8t5/.