An Examination of Gender-Moderated Test Bias on the Levenson Self-Report Psychopathy Scale

Abstract

In this investigation, we examined the validity of Levenson's Self-Report Psychopathy Scale (LSRP; Levenson, Kiehl, & Fitzpatrick, 1995) in a sample of undergraduate students (n = 403). Utilizing a step-down hierarchical multiple regression procedure (e.g., Arbisi, Ben-Porath, & McNulty, 2002), we used LSRP scores to predict psychopathic personality traits to determine the presence of gender-moderated test bias. Results indicate similar correlational magnitudes for men and women. However, systematic over- and underprediction of scores (i.e., differential intercepts) was found for external criteria measuring key affective (i.e., low empathy) and behavioral (i.e., aggression, antisocial behavior) components of the psychopathic personality, although these effects were generally small.

Acknowledgments

Parts of this article were presented at the 2010 Annual Meeting for the American Psychology-Law Society in Vancouver, BC, Canada.

Notes

However, some evidence indicates that the factor structure of the PCL–R differs between men and women, with a three-factor model exhibiting the best fit in women (Cooke & Michie, 2001; Warren et al., 2003). The most recent version of the PCL–R manual (Hare, 2003) also indicates a four-facet structure subsumed within the two factors; however, we focus on the two broader psychopathy factors in this article, as they are linked to the two factors of the Levenson Self-Report Psychopathy scale (LSRP; discussed later).

All correlations were in the same direction for both men and women.

We also conducted the same analyses to examine the Brinkley et al. (2008) three-factor model of the LSRP. Interestingly, the Callous and Egocentricity factors behaved exactly like the Primary scale, and the Antisocial factor as the Secondary scale. Because the general psychopathy literature has focused primarily on Primary and Secondary subscales, they remained in focus in this study.