Hi All,
I have seen likelihood ratio tests used to compare the fit of two nested multiple logistic regression models. By nested I mean the more complicated model includes all of the covariates included in the less complex model, plus additional covariates. However, I am wondering if it is appropriate to use lrtest to conduct a likelihood ratio test to compare the fit of two glmer models that are the same except for one covariate (i.e., the number of covariates in each model is the same, but each model contains one covariate the other model does not).
In simplified equations, to compare two models that differ in the same way these:
a) (Y ~ X1 + X2 +X3+ (1|X4)) vs b) (Y ~ X1 + X2 +X5+ (1|X4))
Thanks so much for any assistance.
Best,
Lucas