The effectiveness of performance appraisal training: Alpha, beta, and gamma congruence.
Item
-
Title
-
The effectiveness of performance appraisal training: Alpha, beta, and gamma congruence.
-
Identifier
-
AAI9130319
-
identifier
-
9130319
-
Creator
-
Gracin, Lynn.
-
Contributor
-
Adviser: Roger E. Millsap
-
Date
-
1991
-
Language
-
English
-
Publisher
-
City University of New York.
-
Subject
-
Psychology, Industrial
-
Abstract
-
The purpose of this study was to introduce the proposed constructs of beta and gamma congruence and to assess promising performance appraisal paradigms, Frame of Reference (FOR) and Rater Error Training (RET), in light of these proposed constructs. Historically, these training paradigms have been assessed at an alpha level, that is, the correspondence of observed performance ratings with true score estimates provided by expert raters. Alpha congruence however presumes a common expert--trainee metric (beta congruence) and conceptual domain (gamma congruence). Expert ratings were obtained from nine (9) Ph.D. I/O Psychologists and sixty-five (65) I/O Psychology doctoral students. Twelve undergraduate classes (236 students) were randomly assigned to one of three conditions: Frame of Reference Training, Frame of Reference and Rater Error Training (FRET), and no training (Comparison).;Gamma congruence was found for both the FOR and comparison groups. The FRET group did not exhibit gamma congruence. These findings demonstrate that FOR training may not result in more accurate frame utilization than is evidenced by untrained raters, and that RET training may decrease accuracy in conceptual domains.;Beta congruence was absent for the FOR and comparison group. High correlations among the factors signaled that trainees perceived less differentiation among the relevant constructs than experts. The effect of RET on scale calibration could not be examined since members of the FRET group were not evaluating the same constructs as the expert group.;Traditional accuracy assessments of elevation and distance accuracy, analogous to alpha level assessments, revealed that the trained groups (FOR, FRET) were significantly different from the comparison group. These findings are explained with regard to relative rather than absolute accuracy. It is concluded that traditional alpha congruence assessments are inappropriate and misleading in the absence of beta and gamma congruence.
-
Type
-
dissertation
-
Source
-
PQT Legacy CUNY.xlsx
-
degree
-
Ph.D.
