I am using SAS to fit SEM to cross-sectional data from 2015. My theoretical model fits well with the data. Because I also have data from 2005 onward, I was thinking to fit the model with 2005 and 2010 data (all cross-sectional) and see how the goodness-of-fit indices, factor loadings and path coefficients change. I would also like to see whether those changes are (statistically) significant. I looked for literature on this, but could not find anything. Does this makes sense, and are there ways to do this? Or, is this something that is not feasible and/or not interpretable? Thank you!