Publication Date

2020

Document Type

Thesis

Committee Members

David LaHuis, Ph.D. (Committee Chair); Nathan Bowling, Ph.D. (Committee Member); Joseph Houpt, Ph.D. (Committee Member)

Degree Name

Master of Science (MS)

Abstract

Research on the optimal number of response options on graphic rating scales has yielded mixed results such as that more scale points are better; there is an optimal range; or that it does not matter. The present study compared the psychometric properties of dichotomous and polytomous personality items using several methods of scoring including summed scores, item response theory (IRT), and item response trees. It was found that regression models based on dichotomous items explained similar amounts of variance in careless responding as models based on polytomous items. In addition, scores from dichotomous models were more closely related to the trait-level variance from the IR tree model. Results suggests that a 2- or 3-point graphic rating scale can achieve comparable trait measurement as what is offered by longer alternatives while reducing the cognitive burden on the respondent.

Page Count

59

Year Degree Awarded

2020

ORCID ID

0000-0003-1966-1645


Share

COinS