University of South Florida
Evaluation of Structural Equation Modeling (SEM) Fit Indices in Distinguishing between Circumplex and Other Factor Models
Abstract
dc:descriptionThe circumplex offers a useful paradigm for simultaneously modeling relationships between latent variables and visually representing an individual’s profile. Previous research has demonstrated the Circular Stochastic Process Model (CSPM) can establish the presence of the circumplex in sample data using Confirmatory Factor Analysis (CFA), allowing researchers to test the circumplex with familiar global fit indices such as the Confirmatory Fit Index (CFI), Tucker-Lewis Index (TLI), Root Mean Squared Error of Approximation (RMSEA), and Unbiased Standardized Root Mean Squared Residual (SRMRu). However, it is not yet clear if these fit indices are sensitive to detect circumplexes under a variety of conditions. The current Monte Carlo simulation study evaluates how the number of common scores, sample size, and position of the common scores on the circumplex affects the sensitivity of these global fit indices to detect and accept circumplexes and reject non-circumplexes. In general, the results suggest the CFI, TLI, and RMSEA are too likely to erroneously accept non-circumplexes, while the SRMR accurately distinguished between circumplexes and non-circumplexes. Researchers are advised to plot common score theta locations, cosine factor loadings, sine factor loadings, β0 values, and β1 values carefully before accepting the circumplex. Thresholds for global fit indices when evaluating the CSPM are also provided. Supplemental code, figures, and tables are available at thur.st/circum-diss.
Degree
thesis:*- Name thesis:degree_name
- Doctor of Philosophy (Ph.D.)
- Level thesis:degree_level
- Dissertation
- Discipline thesis:degree_discipline
- Psychology
- Grantor
- University of South Florida
- Year dc:date
- 2021
Author and committee
dc:creator, dc:contributor.*- Author dc:creator
-
- Thurston, Andrew J.
- Contributors dc:contributor
-
- Stephen E. Stark, Ph.D.
- Brenton M. Wiernik, Ph.D.
- Walter C. Borman, Ph.D.
- Michael Coovert, Ph.D.
- Marina Bornovalova, Ph.D.
- Logan M. Steele, Ph.D.
Subjects
dc:subject × 5Identifiers
dc:identifier.*- Repository record dc:identifier
- https://digitalcommons.usf.edu/etd/9618