The effect of job information and validity generalization information on expert judgments of employment test validities.
Item
-
Title
-
The effect of job information and validity generalization information on expert judgments of employment test validities.
-
Identifier
-
AAI8914803
-
identifier
-
8914803
-
Creator
-
Weinberg, Karen.
-
Contributor
-
Adviser: Joel M. Lefkowitz
-
Date
-
1988
-
Language
-
English
-
Publisher
-
City University of New York.
-
Subject
-
Psychology, Industrial | Psychology, Psychometrics
-
Abstract
-
A sample of Industrial/Organizational Psychologists (N = 114) participated in a judgment task which required them to estimate the true criterion-related validity of 28 different predictor/criterion combinations for jobs with "clerical" and "sales" job titles. The judges were also required to estimate the 90% confidence interval around each of their validity estimates. Two types of task-related information were manipulated to study the effects on estimated validities: (1) the congruency/incongruency of job title and job SKAPs; and (2) the presence/absence of validity generalization (VG) information.;It was hypothesized that (1) validity judgments for jobs with identical SKAPs would be significantly different when different job titles were provided; (2) validity judgments for jobs with identical job titles (irrespective of SKAP information) would be significantly different when VG information was provided versus not provided; (3) confidence intervals would be smaller (i.e., confidence greater) when job information was congruent; and (4) confidence intervals would be smaller when VG information was provided.;MANOVA analyses were conducted to determine if the sets of judged validities and confidence judgments were significantly affected by these two types of manipulated information. Then, ANOVA analyses were used to identify the effects of job and VG information on individual validity or confidence judgments.;When sales SKAP information was provided, judges' validity estimates were significantly higher when the "clerical" versus "sales" job title was provided. This test of Hypothesis #1 was not supported for clerical SKAP judgments, however. Validity generalization information had a significant effect on the "clerical" job title validity judgments such that average judgments were higher in the presence of VG information. A smaller, non-significant effect on "sales" job title judgments was observed. Therefore, Hypothesis #2 also received only partial support. Neither of the manipulated factors significantly affected the experts' confidence judgments.;The results were discussed with respect to (a) why some types of task-related information may be more salient when experts estimate criterion-related validities; (b) how information-processing biases may affect the expert judgment process; and (c) why rational validity estimates are not as meaningful as empirically-derived validity estimates. It was suggested that validity estimates derived from this expert judgment process should be interpreted cautiously because (a) rational judgments are not comparable to empirical estimates of criterion-related validity; and (b) experts' validity estimates may not be free of unknown information-processing biases.
-
Type
-
dissertation
-
Source
-
PQT Legacy CUNY.xlsx
-
degree
-
Ph.D.
