Return to search

Latent trait, factor, and number endorsed scoring of polychotomous and dichotomous responses to the Common Metric Questionnaire

Although job analysis is basic to almost all human resource functions, little attention has been given to the response format and scoring strategy of job analysis instruments. This study investigated three approaches to scoring polychotomous and dichotomous responses from the frequency and importance scales of the Common Metric Questionnaire (CMQ). Factor, latent trait, and number endorsed scores were estimated from the responses of 2684 job incumbents in six organizations. Scores from four of the CMQ scales were used in linear and nonlinear multiple regression equations to predict pay. The results demonstrated that: (a) simple number endorsed scoring of dichotomous responses was superior to the other scoring strategies; (b) Scoring of dichotomous responses was superior to scoring of polychotomous responses for each scoring technique; (c) scores estimated from the importance scale were better predictors of pay then scores from the frequency scale; (d) the relationship between latent trait and factor scores is nonlinear; (e) latent trait scores estimated with the two-parameter logistic model were superior to latent trait scores from the three parameter model; (f) test information functions for each scale demonstrated that the CMQ scales accurately measured a relatively narrow range of theta; (g) the reliability of factor scores estimated from dichotomous data is superior to factor scores from polychotomous data. Issues regarding the construction of job analysis instruments and the use of item response theory are discussed. / Ph. D.

Identiferoai:union.ndltd.org:VTETD/oai:vtechworks.lib.vt.edu:10919/38981
Date28 July 2008
CreatorsBecker, R. Lance
ContributorsPsychology, Franchina, Joseph J., Frary, Robert B., Foti, Roseanne J., Sgro, Joseph A., Clum, George A. Jr.
PublisherVirginia Tech
Source SetsVirginia Tech Theses and Dissertation
LanguageEnglish
Detected LanguageEnglish
TypeDissertation, Text
Formatxi, 184 leaves, BTD, application/pdf, application/pdf
RightsIn Copyright, http://rightsstatements.org/vocab/InC/1.0/
RelationOCLC# 25611469, LD5655.V856_1991.B435.pdf

Page generated in 0.0021 seconds