PERSONNEL DECISION-MAKING - THE IMPACT OF MISSING INFORMATION

Introductory psychology students evaluated hypothetical employees for promotion to a supervisory position in a computer-systems department. Employees were described in terms of management and computer-programming test scores; some employees were missing one test score. The correlation between employees' test scores was varied across groups of subjects (-.85, .00, .85). In general, subjects penalized employees with missing test scores, though the pattern varied by correlation condition. The results were consistent with the hypothesis that subjects would predict the missing test score from the correlation between the test scores, make an evaluation, and then apply a penalty. The fact that subjects did not ignore the missing score invalidates a procedure commonly used to distinguish adding and averaging models: comparing the missing-data slope to complete-data slopes.