Development of moderator variables to enhance the prediction of managerial effectiveness

Abstract

ITEM ANALYSES WERE USED TO DEVELOP 2 MODERATOR VARIABLES WHICH, ON CROSS VALIDATION, SUCCESSFULLY IDENTIFIED MANAGERS WHO WERE OVER- AND UNDERPREDICTED BY REGRESSION EQUATIONS DEVELOPED EARLIER. THESE MODERATOR TESTS WERE USED TO IDENTIFY AND ELIMINATE FROM THE CROSS-VALIDATION SAMPLE 55 MANAGERS (25% OF THE TOTAL) CLASSED AS "UNPREDICTABLE." THE POINT-BISERIAL CORRELATION BETWEEN PREDICTED CRITERION STATUS (ABOVE OR BELOW THE CRITERION MEDIAN) AND ACTUAL CRITERION STATUS FOR THE REMAINING MANAGERS WAS .73 AS COMPARED WITH A POINT-BISERIAL CORRELATION OF ONLY .65 FOR ALL MANAGERS IN THE SAMPLE. MOREOVER, THE DEGREE OF OVERLAP ON THE ACTUAL CRITERION SCALE BETWEEN MANAGERS PREDICTED TO BE "HIGH" OR "LOW" WAS REDUCED FROM 38%-28% BY USING THE 2 MODERATOR TESTS. IT IS CONCLUDED THAT THESE RESULTS PROVIDE FURTHER CONFIRMING EVIDENCE OF THE USEFULNESS OF MODERATOR VARIABLES FOR ENHANCING THE MAGNITUDE OF RELATIONSHIPS IN TEST VALIDATION AND SELECTION RESEARCH. (35 REF.) (PsycInfo Database Record (c) 2025 APA, all rights reserved)