Measuring over time growth, but not o... PreviousNext
Mplus Discussion > Growth Modeling of Longitudinal Data >
 Katy Pearce posted on Wednesday, March 30, 2011 - 2:20 pm
I have large datasets (same variables) over many years (nationally representative). I've done SEM models for each year, but I'd like to test for statistically significant NATIONAL growth over the years.

The SEM model is fairly simple with a latent variable for SES (3 continuous variables) that predicts binary yes/no ownership of technologies as well as skill (continuous).

I think that doing this with latent growth modeling makes sense, but as the cases (individuals) are not constant over the years, maybe it doesn't.

Any thoughts? Thank you.
 Linda K. Muthen posted on Wednesday, March 30, 2011 - 3:00 pm
Growth modeling requires that the same individuals be measured at each time point.
 Katy Pearce posted on Wednesday, March 30, 2011 - 3:39 pm
Thank you for your prompt response. Do you know of any other test that I can use to measure growth?
 Bengt O. Muthen posted on Thursday, March 31, 2011 - 10:09 am
If you have independent samples of individuals for the different years, you can do a multiple-group analysis (year as group) and study mean differences in the factor over the groups.
 Bengt O. Muthen posted on Thursday, March 31, 2011 - 11:40 am
I should add that my suggestion does not give a measure of growth but year-to-year differences. The population (or sampling procedure) might have changed rather than the individuals you started with.
 Katy Pearce posted on Thursday, March 31, 2011 - 12:56 pm
Thank you. That is a good idea.
Back to top
Add Your Message Here
Username: Posting Information:
This is a private posting area. Only registered users and moderators may post messages here.
Options: Enable HTML code in message
Automatically activate URLs in message