Reliability estimated by scoring key equivalence across samples.

Abstract

Two studies estimated the reliability of a biographical inventory (BI) by correlating scores from preexistent BI scoring keys with scores from empirically constructed BI scoring keys. Study I demonstrated that a preexistent scoring key, developed on National Aeronautics and Space Administration scientists to predict creativity, was correlated .87 (n = 294) and .91 (n = 296) with scoring keys empirically constructed to predict creativity for two samples of industrially employed scientists and engineers. Study II demonstrated that preexistent scoring keys, constructed on university freshmen to predict grade point average (GPA), were correlated .88 (for both 5,349 males and 5,897 females) with scoring keys empirically constructed to predict GPA for high school students. Test-retest reliability estimates and validity generalization estimates supported the above results. TYPE OF RESULT TYPE OF ANALYSIS Cross-Xfnifi'li' Reliability estimated by validation Z ation" key equivalence