Hello. I am a grad student anticipating the use of SEM for my thesis. Have been reading up on this method and I am stuck on figuring out exactly what constitutes data points vs parameter estimates as I need to ensure I have an overidentified model. Any simply explanation of exactly what this means in terms of what I am counting would be wildly appreciated
I think by data points you mean the H1 model. With no missing data, the sample statistics are the parameters in the H1 model. The H0 model is the analysis model. It should have fewer parameters than the H1 model for the model to be overidentified.