275 participants (216 female) to examine how different groups score on a test of EI. Differences were compared for gender, ethnicity and age. Results indicated that females scored slightly higher than males and EI scores tended to increase with age. Group differences existed for ethnicity but favored minority groups, mitigating potential adverse impact concerns. Full implications for test development and organizational use are discussed.
2004 Elsevier Ltd. All rights reserved.
Keywords: Emotion; Emotional; Intelligence; Adverse impact; Group differences; Minority; Gender
1. Introduction
Many predictors have been proposed for use in personnel selection (cf. Guion, 1998; Schmidt &
Hunter, 1998) and the criterion-related validity of these predictors has been investigated extensively.
Further, researchers and practitioners alike have considered the potential adverse impact on minorities and protected demographic groups due to the use of different predictors (cf. Ones &
Viswesvaran, 1998; Reilly & Chao, 1982). Group difference in mean predictor scores is a likely cause of adverse impact (Sackett & Roth, 1996). Of course, adverse impact is determined by a
* Corresponding author.
E-mail address: dvanro01@fiu.edu (D.L. Van Rooy).
0191-8869/$ - see front matter 2004 Elsevier Ltd. All rights reserved. doi:10.1016/j.paid.2004.05.023 Personality and Individual Differences 38 (2005) 689–700 www.elsevier.com/locate/paid combination of several factors (e.g., selection ratio) of which group differences is only one (Sackett
& Roth, 1996). Nevertheless, it is important to investigate the presence of group differences in predictor scores as these differences have important implications for personnel selection procedures such as banding of test scores (cf. Schmidt & Hunter, 1998).
Of the many predictors used in personnel selection, general mental ability has been shown to have the best criterion-related validity across several jobs, but its use also results in adverse impact concerns (Guion, 1998; Schmitt & Chan, 1998). This combination of findings has prompted calls to search for alternative predictors that could supplement the use of general mental ability as valid predictors of job performance (e.g., Bobko, Roth, & Potosky, 1999; Hunter & Hunter, 1984;
Schmitt, Rogers, Chan, Sheppard, & Jennings, 1997). The recent use of increased teamwork where individuals depend on others to get the job done, as well as the resurgence in use of personality variables (cf. Guion, 1998), has resulted in the introduction of several new predictors. One such predictor is emotional intelligence (EI).
Moreover, the use of EI in personnel selection contexts is increasing (cf. Kostman & Bedwell,
2003; Miller & Zeng, 2003; Wong & Law, 2002). Given prevalent and steadily increasing use, it is imperative that industrial-organizational psychologists evaluate the potential applicability of EI as a predictor in personnel selection. Research on criterion-related validity and other considerations for use as a predictor in selection (e.g., adverse impact) is needed.
The predictive validity of EI measures has been replicated in different samples and a metaanalytic accumulation of this literature exists (Van Rooy & Viswesvaran, 2004). Anecdotally, many proponents for the use of EI claim that adverse impact concerns are less prevalent with the use of EI than with many other selection methods. Considering that EI has been shown to be moderately correlated with general mental ability (Van Rooy & Viswesvaran, 2004), the idea of a test of EI resulting in little or no group differences is questionable. However, few empirical studies to date have reported the scoring results of different groups (i.e., gender and ethnicity) and