rr-lavaan

lavaan - fixing standardized parameter values

In lavaan if we would want to test if a certain parameter is greater than 0.6 we would use:

fit.model.free <- '
F1 =~ V1 + V2 +V3
F1 ~~ F1
' %>% cfa(data)

fit.model.fix <- '
F1 =~ 0.6*V1 + V2 +V3
F1 ~~ F1
' %>% cfa(data)

anova(fit.model.free, fit.model.fix)

But what if I'm actually interested to see if the standardized parameter is greater than 0.6? What code should I use to specify such a model?

Solution

  • To standardize the loading, you need both the observed variable and the common factor to be standardized. You can do this with options on the cfa() call:

    std.ov=T # standardizes observed variables std.lv=T # standardizes the common factors

    Then the .60 in your second model will be a standardized loading.