Linear mixed model without random or repeated effects? GLMZ?

[jcs691]

Hi,

I'm working on a dissertation, of which many chapters have longitudinal data and I've been using linear mixed models to estimate growth. I also have some non-longitudinal data (i.e. cross-sectional) that I'd like to estimate growth with in a similar way. The cross-sectional data spans a good age-range, and I'd like to construct a model for the DV over this age period.

Can I conduct a linear 'mixed' model without random or repeated effects (is this a marginal model) to just get ML parameter estimates and build a model? Is this the same thing as a Generalized Linear Model? I get almost exactly the same parameters with each.

I'd ideally like to work within the LMM framework as I now know it pretty well, but more important is that my stats are sound!

Any help would be MUCH appreciated!
Thanks,
J

 

[Link]

If you don't have repeated measures data, then it'd be best to just set up a GLM model. It's very similar to setting up GLMM, except less complicated. I haven't tried it out yet, but I imagine that setting up a mixed model without repeated measures would give you the exact same result as the linear model.