Hi everyone!
I used linear mixed model for an analysis in my thesis and need to report the p-value of the model. Why is it not given? I am not comparing models and I need to be sure if I could report the simple effects, if the main model is statistically significant. AIC or BIC is not working for me since I don't compare models. Is there any way to observe the p value of the main model?
Thanks!
P value of the mixed model
-
mcfanda@gmail.com
- Posts: 549
- Joined: Thu Mar 23, 2017 9:24 pm
Re: P value of the mixed model
Hi
not in the current version. However, version 3 is in preview here (https://github.com/gamlj/gamlj#version-3) and does the overall model test. You just download the package from the links there and then sideload into jamovi (see instructions in the linked page). You need to re-run the models, because GAMLj3 installs itself in parallel to the previous version.
As regards testing "the overall model", you find two tests together with R-squared and p-values. The Conditional refers to the whole model, random and fixed effects together. This basically tests that some variance is explained by your model as a whole. the Marginal refers to the whole set of fixed effects, so it basically tests whether your independent variables together explain some variance of the dependent variable. The marginal is more important in terms of strength of associations.
let us know it works
not in the current version. However, version 3 is in preview here (https://github.com/gamlj/gamlj#version-3) and does the overall model test. You just download the package from the links there and then sideload into jamovi (see instructions in the linked page). You need to re-run the models, because GAMLj3 installs itself in parallel to the previous version.
As regards testing "the overall model", you find two tests together with R-squared and p-values. The Conditional refers to the whole model, random and fixed effects together. This basically tests that some variance is explained by your model as a whole. the Marginal refers to the whole set of fixed effects, so it basically tests whether your independent variables together explain some variance of the dependent variable. The marginal is more important in terms of strength of associations.
let us know it works
Re: P value of the mixed model
It worked! Thanks a lot!