Publication Bias and the Validity of Conditional Reasoning Tests

Analyses were conducted to evaluate the possible presence and influence of publication bias on the validity of conditional reasoning tests for aggression. Multiple publication bias methods yielded results consistent with a conclusion of publication bias suggesting that the reported validity of the conditional reasoning tests may be overestimated. PUBLICATION BIAS 2 The development of a new personnel selection measure is typically prompted by a desire to address limitations of current measures. Over the last decade, promising work has been conducted by James and colleagues (e.g., James, 1998; James & Mazerolle, 2002; James et al., 2005) in the development of a new personnel selection measure, known as conditional reasoning tests for aggression (CRT-A). This line of research is an attempt to develop a measure that makes it more difficult for applicants to fake (i.e., to distort responses due to the fact that the “socially desirable ” answer may be apparent to test takers) and to explain construct variance not captured by conscious self-reports. Recently, a meta-analysis published by Berry, Sackett, and Tobares (2010) has offered some evidence of validity for these measures, but not as optimistic as offere