Factor Analysis and Structural Equation Modeling
At present, I use the Lavaan package to perform CFA and SEM (https://lavaan.ugent.be/tutorial/index.html). I also use semoutput to extract loadings, paths, and correlations (https://dr-jt.github.io/semoutput/).
​
I will not directly explain these packages, since the associated tutorials are some of the best available. Instead, I will use this space to validate these packages against LISREL results from studies that I have published. I will, however, provide my Lavaan code for anyone interested in certain techniques.
​
​
​
​
​
This study explored the relation between working memory capacity, sensory discrimination, and attention. I found that working memory capacity and sensory discrimination could be combined into one latent factor that was predicted by both attention control and sustained attention.
​
​
1. Full CFA
1.1 Fit Statistics
1.2 Lavaan Code
CFA1 <- ' VSWM =~ VAcircSq + VAarrows + SplitSpan
AUDdisc =~ Pitch + Loud
AC =~ AntiSacc + Flanker + Stroop
Sust =~ PVT + CRT + SART'
CFA1fit <- cfa(CFA1, data = Stabledata)
summary(CFA1fit, fit.measures = TRUE)
fitMeasures(CFA1fit)
​