The purpose of this experiment is to examine one previously untested premise of the validity generalization model proposed by Schmidt and Hunter on test validity. The sample consisted of 307 subjects who were tested on measures constructed to be examples of contaminated, deficient, relevant and ultimate criteria. These criteria were correlated to a predictor, student GPA. The predictor-criteria relationship was observed over various sample sizes ranging from n = 25 to n = 300. The validity coefficients were hypothesized to change over increasing sample sizes in the same pattern as the other artifacts from Schmidt and Hunter's validity generalization model. Student GPA was obtained from a biographical data sheet and this raw score was correlated through the use of Pearson Product-Moment correlations to raw scores on each of four quizzes, representing the four criteria. These correlations were then analyzed to determine if they follow four hypothesized patterns supporting the patterns of Schmidt and Hunter's controllable artifacts. The results of the study indicated that the four hypotheses of various patterns resulting among the four criteria generally supported Schmidt and Hunter's validity generalization model. Further investigation to determine more completely the extent of the effects of the uncontrollable artifacts is suggested as this exploratory study represents the first laboratory investigation of the effect of sample size on criterion contamination and criterion deficiency in the validity generalization model.
| Identifer | oai:union.ndltd.org:ucf.edu/oai:stars.library.ucf.edu:rtd-5752 |
| Date | 01 January 1985 |
| Creators | Pavlick, Cheryl L. |
| Publisher | University of Central Florida |
| Source Sets | University of Central Florida |
| Language | English |
| Detected Language | English |
| Type | text |
| Format | application/pdf |
| Source | Retrospective Theses and Dissertations |
| Rights | Public Domain |
Page generated in 0.0672 seconds