MCFA using Lavaan

[meng han]

Hi there,

I came across the conversation with the subject "Warning message: In cov2cor(w.cov) : diag(.) had 0 or NA entries; non-finite result is doubtful", wherein Terence mentioned that we can use Lavaan to test MCFA model fit.

I managed to run the model with the following syntax. May I ask you to just confirm that my syntax is correct?

Specify the model

model <- '

fw1 =~ FCT1 + FCT2 + FCT3 + FCT4 + FCT5 + FCT6

fw2 =~ ICT1 + ICT2 + ICT3 + ICT4 + ICT5 + ICT6

fb1 =~ FCT1 + FCT2 + FCT3 + FCT4 + FCT5 + FCT6

fb2 =~ ICT1 + ICT2 + ICT3 + ICT4 + ICT5 + ICT6'

Model fit

fit <- sem(model = model, data = raw, cluster = "cluster")

summary(fit, fit.measures = TRUE, standardized = TRUE) 

Thanks,

Meng

[Ed Merkle]

In the model specification, I think you need lines that say "level: " to distinguish within from between. See here:

https://lavaan.ugent.be/tutorial/multilevel.html

Ed