Should we trust our first impressions? An attempt to reconcile the thin slices and interview validity literatures

How much should we trust our first impressions of other people? Thin slices researchers and interview validity researchers have traditionally reached different conclusions about the predictive validity of observers\u27 initial evaluations. Thin slices researchers argue that performance can be accurately predicted using the conclusions untrained judges reach after a few moments of observation, whereas interview validity researchers argue that interviewers cannot reliably predict a job applicant\u27s future performance after a single one-on-one interaction. Drawing from research on perception, cognition, and multilevel aggregation, I hypothesize that previous investigations may have reached different conclusions about the predictive validity of first impressions because of differences in the dyadic observation process and the levels of analysis that researchers have used in their investigations. Data from three laboratory studies provide mixed support for my hypotheses. The results of these studies suggest that the predictive validity of observers\u27 evaluations increases when observers\u27 evaluations are aggregated. The presence of idiosyncratic biases appears to make individual observers much less reliable predictors of performance than the aggregated evaluations of multiple observers. The data also provide tentative evidence that providing observers with biographical information influences the predictive validity of their performance evaluations. Biographical information has the potential to either increase or decrease the predictive validity of performance evaluations, but does not appear to have a consistent effect on predictive validity of these evaluations