Date of Award

6-1-2013

Document Type

Dissertation

Degree Name

Ph.D.

Organizational Unit

Morgridge College of Education

First Advisor

Kathy E. Green, Ph.D.

Second Advisor

Antonio Olmos-Gallo

Third Advisor

Donald Bacon

Keywords

Character development, Confirmatory factor analysis, Factorial validity, Item response theory, Latent class analysis, Leadership development

Abstract

This study leveraged the complementary nature of confirmatory factor (CFA), item response theory (IRT), and latent class (LCA) analyses to strengthen the rigor and sophistication of evaluation of two new measures of the Air Force Academy’s “leader of character” definition—the Character Mosaic Virtues (CMV) and the Leadership Mosaic Inventory (LMI). Special CFA methods involving robust weighted least squares estimation were implemented to analyze the rated responses through linear structural equation modeling. Most informative at the subscale level, the CFA techniques provided evidence of factorial validity and other desirable psychometric properties in support of previous exploratory results for a nine-factor CMV and a unidimensional LMI. As an alternative to CFA, IRT’s nonlinear approach and capability of estimating the probability of a response based on the amount of a latent trait was applied to the CMV and LMI data. Most informative at the item level, individual item difficulty parameters were estimated as the amount of the latent trait required for a cadet to give a particular response to an item and mapped against person ability estimates. The IRT analyses were extended to explore the underlying dimensions of the CMV and LMI through multidimensional IRT (MIRT)—results supporting the theoretical dimensional structure of the CMV were substantiated while new evidence of a multidimensional LMI structure was concluded. LCA techniques permitted the inference of classifying mixtures of unobserved cadet subpopulations based on their responses to the CMV and LMI. The analyses uncovered the meaning and number of underlying subpopulations not evident through the other two traditional factor structure analysis techniques. The creation of the study’s latent class models complemented the more traditional dimensional approaches of structure assessment with an understanding of the unobserved cadet subpopulations which could lead to future targeted cadet developmental opportunities being applied at organizational levels or groups deficient in certain latent traits. By exposing more researchers, decisionmakers, and other stakeholders to these three advanced psychometric evaluation methods, this study benefited the fields of moral development and leadership development, especially at the nation’s service academies.

Publication Statement

Copyright is held by the author. User is responsible for all copyright compliance.

Rights Holder

David L. Higginbotham

Provenance

Received from ProQuest

File Format

application/pdf

Language

en

File Size

298 p.

Discipline

Quantitative psychology and psychometrics, Educational tests & measurements, Statistics

Download



Share

COinS