Premium
A META‐ANALYSIS OF CONDITIONAL REASONING TESTS OF AGGRESSION
Author(s) -
BERRY CHRISTOPHER M.,
SACKETT PAUL R.,
TOBARES VANESSA
Publication year - 2010
Publication title -
personnel psychology
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 6.076
H-Index - 142
eISSN - 1744-6570
pISSN - 0031-5826
DOI - 10.1111/j.1744-6570.2010.01173.x
Subject(s) - psychology , aggression , meta analysis , test validity , coding (social sciences) , test (biology) , incremental validity , criterion validity , predictive validity , external validity , social psychology , statistics , psychometrics , developmental psychology , construct validity , mathematics , medicine , paleontology , biology
James et al. (2005) reported an estimate of criterion‐related validity (corrected only for dichotomization of criteria) of r = .44 across 11 conditional reasoning test of aggression (CRT‐Aggression) validity studies. This meta‐analysis incorporated a total sample size more than twice that of James et al. Our comparable validity estimate for CRT‐Aggression scales predicting counterproductive work behaviors was r = .16. Validity for the current, commercially marketed test version (CRT‐A) was lower ( r = .10). These validity estimates increased somewhat (into the .24–.26 range) if studies using dichotomous criteria with low base rates were excluded from the meta‐analysis. CRT‐Aggression scales were correlated r = .14 with measures of job performance. As we differed with James et al. in some of our coding decisions, we reran all analyses using James et al.'s coding decisions and arrived at extremely similar results.