Faking in Personality Assessment: A “Multisaturation” Perspective on Faking as Performance

Abstract

Concerns about socially desirable responding on self-report personality tests are heightened in employment settings where motivation to fake is elevated. Building on prior faking models and the classical X = T + e measurement model, we offer a unique performance-based perspective, in which opportunity, ability, and motivation to fake are jointly critical (P = O × A × M). Trait activation theory is used to show how impression management and self-deception can express multiple abilities and traits beyond those targeted and how response biases might be reduced. Three sets of testable hypotheses are offered. That nontargeted traits (e.g., ambition) serving faking might contribute to the prediction of job performance supports the view that faking could benefit selection decisions. Several arguments against this perspective are proposed.

ACKNOWLEDGMENTS

An earlier version of this article was presented at the 24th Annual Conference of the Society for Industrial and Organizational Psychology, April 4, 2009, New Orleans, LA.

Notes

¹As used throughout this article, “self-deception” and “impression management” are conceptual terms and are not intended as an endorsement of any particular measure of related constructs.

²If “putting one's best foot forward” is a crime of omission, faking is more seriously a crime of commission. Faking also bears distinction from prosocial deception, such as when a person avoids conflict by restraining from criticizing someone judged deserving of criticism. There is nothing prosocial about faking on a personality test.

³Intention to fake, a key ingredient in the theory of planned behavior, is to some extent redundant with faking, defined as the deliberate (i.e., intentional) attempt to present an overly desirable impression. Along similar lines, Snell et al. (1999) equate intention to fake with motivation to fake.

⁴In some versions of classical theory, T = the true score on whatever the test, in fact, is measuring. We adopt the more restrictive interpretation here as it allows us to articulate the effects of motivated distortion as distinct from T.

⁵This is simplest to discuss in the context of a positively desired trait, for example, methodicalness for the job of accountant. For traits on which it is desirable to fall at the low end (e.g., neuroticism in many jobs), faking is expected to occur in that direction. The same principles hold, nonetheless.

⁶One must include mental expressions because many traits, including cognitive styles (e.g., cognitive complexity) and psychopathy constructs (e.g., anxiety), are identified by their unique patterns of thought and affect.

⁷The strength of a selection situation depends largely on job demand, which, in turn, is affected by economic conditions (e.g., high unemployment), job characteristics (e.g., status, pay), and personal reasons (e.g., debt). Job demands are unlikely to completely overwhelm individual differences in the propensity to fake, but such propensities will have weakened influence when job demand is especially strong.

⁸Hiring someone who fakes on job-irrelevant traits or in the undesirable direction on relevant traits still carries risk, specifically that of hiring someone who may be disposed to dishonesty.

⁹Perhaps this is one way that lie scales work. Also, there is a notable irony and ethical duplicity here to the degree the impression that applicants are unlikely to fake is contrived.

¹⁰It might be expected, accordingly, that, on hearing that he or she missed the cut, the risky faker would be especially eager to know why.

¹¹Of interest, the median (absolute value) correlation between the Personality Research Form's content scales and the Crowne-Marlow scale has been reported to be .25 (Stricker, 1974), which is notably weaker than the corresponding median of .39 reported for the NEO–PI scales (Kurtz, Tarquini, & Iobst, 2008), developed without reference to item desirability.

¹²Not all confirmatory strategies are conceptually driven; they can also include empirically driven replications.

¹³Clever fakers may be less likely to fake on irrelevant scales to reduce the likelihood of presenting a “too-good-to-be-true” profile.