Range shrinkage of cognitive ability test scores in job-specific German applicant pools: Implications for range restriction corrections

Range restriction corrections require the predictor standard deviation in the applicant pool of interest. Unfortunately, this information is frequently not available in applied contexts. The common strategy in this type of situations is to use national-norm standard deviation estimates. This study used data from 8,276 applicants applying to nine jobs in German governmental organizations to compare applicant pool standard deviations for two cognitive ability tests with national-norm standard deviation estimates, and standard deviations for the total group of governmental applicants. Results revealed that job- and organizational context-specific applicant pool standard deviations were on average about 10-12% smaller than estimates from national norms, and about 4-6% smaller than standard deviations for the total group of governmental applicants.