Date of Award

Summer 8-2021

Document Type

Dissertation

Degree Name

Doctor of Philosophy (PhD)

First Advisor

Tilman Sheets

Abstract

Cognitive ability testing and cognitively loaded measures in employee selection have been utilized, developed, and improved upon for over a century; however, it is not without its faults. Two major problems facing cognitive ability tests are their tendency to produce adverse impact when used in selection systems and the costs associated with creating a well-constructed measure. This paper proposed that Automated Item Generation (AIG) may provide a solution to both of those problems. The first study focused on the construct validation of the Katyem Object Tracking Assessment (KOTA), a nonverbal AIG measure of fluid intelligence, that would allow test takers to practice as much as they want, comparing it to the emotionality portion of the HEXACO and to the short form of the Hagen progressive Matrices. After cleaning and removing careless responders from the sample of 458 participants, 89 remained, far below the 200-participant sample size needed to find a medium effect size. The data were analyzed using the Multitrait-multimethod matrix. Support for the hypotheses were not found. Afterward, the measure was used in a second study to determine if allowing participants to practice reduces adverse impact in a hypothetical employment situation. After cleaning and removing careless responders from the sample of 172 participants, 56 remained and were analyzed using two-way repeated measures ANOVA, Chi-squared goodness of fit test, Fisher's Exact test, and the four-fifths rule. The hypotheses concerning group differences and practice effects were unsupported, however, the hypothesis for the KOTA not having adverse impact was supported. Directions for future research are also provided.

Share

COinS