lavaan longitudinal invariance CFA with a 2-factor model in R

Question

I use the longInvariance function to asssess longitudinal invariance between two time points for interval data; and I am wondering how the correct lavaan/semTools R-code should look like.

It works when I look at one factor at a time; but it generates an error when examining a two factor model.

Example code for one factor below:

  model.oneFactor <- '
  Factor1T1 =~ Item1 + Item2 + Item3
  Factor1T2 =~ Item1t2 + Item2t2 + Item3t2
  '

  # Create list of variables
  var1 <- c("Item1", "Item2", "Item3")
  var2 <- c("Item1t2", "Item2t2", "Item3t2")
  constrainedVar <- list(var1, var2)

  # Invariance of the same factor across timepoints
  longInvariance(model.oneFactor, auto=1, constrainAuto=TRUE, varList=constrainedVar, data=data, estimator="MLM", strict=TRUE)

However, when adding a second factor in the model generates an error. Example code for a two-factor model below:

  model.twoFactor <- '
  Factor1T1 =~ Item1 + Item2 + Item3
  Factor2T1 =~ Item4 + Item5 + Item6 + Item7
  Factor1T2 =~ Item1t2 + Item2t2 + Item3t2
  Factor2T2 =~ Item4t2 + Item5t2 + Item6t2 + Item7t2
  '

  # Create list of variables
  var1 <- c("Item1", "Item2", "Item3", "Item4", "Item5", "Item6", "Item7")
  var2 <- c("Item1t2", "Item2t2", "Item3t2", "Item4t2", "Item5t2", "Item6t2", "Item7t2")
  constrainedVar <- list(var1, var2)

  # Invariance of the same factor across timepoints
  longInvariance(model.twoFactor, auto=1, constrainAuto=TRUE, varList=constrainedVar, data=data, estimator="MLM", strict=TRUE)

The generated error is:

   Error in longInvariance(model.twoFactor, auto = 1, constrainAuto = TRUE,  : 
    The factor names of the same element of the 'varList' are not the same.

Answer 1

I think you also asked your question on the lavaan Google group.

As it was answered there and written in the semTools documentation, semTools only provides longInvariance() for one single scale. If you have two scales you need to specify the models by hand.