[R-sig-ME] testing for a significant correlation between random slopes (Jaime Ashander)

[Tove Jansson]

HI,

Correct me if I am wrong, but shouldn’t REML be used when comparing random effects (ML will supposedly produce overly conservative estimates and should only be used when comparing models with differences in fixed effects).

best,

DT


On Dec 19, 2014, at 12:00 PM, r-sig-mixed-models-request at r-project.org wrote:

> Re: testing for a significant correlation between random
>      slopes (Jaime Ashander)


	[[alternative HTML version deleted]]