Measurement invariance testing in multigroup CFA – scaling methods

Question

I am doing a multigroup CFA and want to test whether the latent factors are invariant across three groups (three timepoints). The CFA model consists of two latent factors and 7 items, where latent factor one has 4 items, and latent factor two has 3 items. In total, there are 92 observations. I got two questions relating to measurement invariance testing and the configural model:

Q1. When I estimate the configural method using the “marker variable method” (where loadings of a variable is fixed at 1, and the intercept at 0 across all groups), I am having convergence issues - the model won’t converge. Does anyone have any idea of why this might happen?

Q2. Due to the convergence issue mentioned in Q1, I also tried estimating the configural model using the “in every group method”, with the latent variables mean equal 0 and the variance equal 1 across all groups. This model was estimated. I am therefore wondering whether this approach is a good alternative to the first one, or perhaps Item response theory (IRT) might be a better way to assess the invariance of the latent factors across the groups?

Here’s the Stata command for the two approaches:

*Approach 1 marker variable method

sem (SMW -> pol1 pol2 pol3 pol4) ///
            (IWB -> pol5 pol6 pol7) ///
            (pol1 <- SMW _cons@0) (pol5 <- IWB _cons@0 ), ///
            group(timepoint) ginvariant(none) mean(SMW IWB) ///
            vce(sbentler) cov( SMW*IWB e.pol2*e.pol1)

*Approach 2 in every group method

sem (SMW -> pol1 pol2 pol3 pol4) ///
            (IWB -> pol5 pol6 pol7), ///
            group(timepoint) ginvariant(none) variance(SMW@1 IWB@1) means(SMW@0 IWB@0) ///
            vce(sbentler) cov( SMW*IWB e.pol2*e.pol1)

Any help or thoughts are much appreciated!