ASSESSMENT-CENTER CONSTRUCT-VALIDITY AND BEHAVIORAL CHECKLISTS - SOMEADDITIONAL FINDINGS

A quasi-experimental design was used to investigate the effects of an ''untranslated'' behavioral checklist (i.e., developed without the use of a retranslation procedure) on the construct validity of assessment center dimension ratings for four exercises. Multitrait-multimethod a nd confirmatory factor analysis approaches were used to compare the ch ecklist to a traditional graphic rating scale. Results indicated (a) g reater discriminant validity the behavioral checklist, (b) greater con vergent validity for the graphic rating scale, and (c) greater evidenc e of exercise factors rather than dimension factors for both methods. It is suggested that the cognitive reduction benefits of behavioral ch ecklists may improve the discriminant validity of dimension ratings an d should be investigated in conjunction with approaches associated wit h gains in convergent validity. Issues related to the content validity of behavioral checklists are discussed.