VIDEO‐BASED SITUATIONAL TESTING

In two separate studies, video‐based tests of situational judgment were developed and validated against measures of performance for hourly service workers. In the first study, 684 employees were used to develop a test of retail associate judgment and 787 newly hired employees were used to cross‐validate the instrument. In the second study, 412 current employees were used to develop a test of nursing home caregiver judgment and 148 newly hired caregivers were used to cross‐validate this video‐based test. In both studies, responses to video‐based situational vignettes were empirically keyed against supervisory ratings of performance. The resulting keys produced uncorrected cross‐validities in the low .20s. The video‐based test scores were also found to be related to measures of cognitive ability and, to a lesser extent, experience. These results suggest that cognitive ability and possibly experience account for some but not all of the predictiveness of video‐based situational tests. Video‐based situational tests demonstrated score differences between Whites and non‐Whites of roughly one‐half a standard deviation, indicating that the use of such tests could produce adverse impact against non‐Whites. In the second study, customer preferences regarding desired behavior were also used to develop a rational scoring key. This customer driven key was significantly related to performance ( r = .33) in the cross‐validation sample, demonstrating that customers' preferences can provide the basis for developing valid predictors. Potential benefits of video‐based situational tests in selection and directions for future research are discussed.