A Tutorial on Confirmatory Factor Analysis Using lavaan: Evaluating Model Fit using dynamic
Listed in
This article is not in any list yet, why not save it to one of your lists.Abstract
Measurement is at the core of research in many fields such as psychology. One way thatresearchers can test the factorial structure of an instrument is to use confirmatory factor analysis(CFA). When using CFA, researchers evaluate how well the hypothesized structure of a measurefits the data. Fit indices, which quantify the degree of misfit, are then used to evaluate thefactorial model. Traditionally, researchers have used fixed index cutoff values to judge theappropriateness of their factorial models. However, many have discussed the limitations of usingfixed cutoffs, as fixed values do not generalize to all kinds of models. Recently, McNeish &Wolf (2023) have developed the Dynamic Fit Index approach (DFI) which enables thegeneration of fit index values that are tailored to the characteristics of the model being tested. Inthe following tutorial, we conduct a CFA on the Attainment of School Achievement Goal Scale(A-SAGS) using the lavaan package in R. We then generate fit indices using the dynamicpackage. We conclude that the DFI approach, although very promising and interesting,necessitates further testing. Dynamic fit indices can be overly strict and do not always agree onthe appropriateness of factorial models, which stresses the need to further test its algorithm to seewhether it is of practical value when evaluating model fit.Keywords: Confirmatory Factor Analysis, Fit evaluation, Fit indices, Dynamic fit. Tools:R