Multiplicity adjustments for respecification searches in structural equation models

John Appiah Kubi, University of Texas at El Paso

Abstract

Structural Equation Modeling (SEM), as a statistical modeling technique, is one of the most comprehensive and flexible approaches to data analysis currently available. Its use has been increasing steadily over the past few decades. Generally, it refers to a family of techniques that employs the analysis of covariance to establish relationships among a set of variables. It allows researchers (or users) to assess the adequacy of their hypothesized models with their sample data. Often times, in assessing their models, researchers are not only interested in the overall fit of their model but they are also interested in knowing which proposed relationships (parameters) are significant. With respect to the evaluation of the significance of parameters, researchers risk capitalizing on chance and including unnecessary parameters (that is, producing a less parsimonious model) in their models when no form of controlling type I error rate is adopted. ^ In this thesis, the Functional Independence Measure (FIM) data was used to demonstrate the effectiveness of using a Scheffe-like procedure for controlling the rate of type I errors when multiple parameters are evaluated for significance.^

Subject Area

Statistics

Recommended Citation

Appiah Kubi, John, "Multiplicity adjustments for respecification searches in structural equation models" (2013). ETD Collection for University of Texas, El Paso. AAI1545148.
http://digitalcommons.utep.edu/dissertations/AAI1545148

Share

COinS