I'm trying to look for significant effects on "similarity" (of "isinpair" and controlling for time effects) using repeated measures with an in group sample. The intervention "isinpair" occurs after a short period of time (time intervals aren't equivalent, but there are many occuring with in a relative short time).
The question seems a variation of R's lmer cheat-sheet
Unfortunately, I am quite hazy about the syntax (and underlying statistical differences) of lme
vs. lmer
. I've seen both recommended for similar tasks.
Are these two examples accomplishing what I intend? How equivalent / different are they? Which is better illustrating what I'm hoping to examine?
lmer(similarity ~ 1 +time*isinpair +(time*isinpair|user1), data=nozeros) lme(similarity ~ isinpair*time, random=list(user1=pdBlocked(list(~1, pdIdent(~isinpair-1), pdIdent(~time-1)))), data=nozeros)
How does this compare to:
aov(simscore ~ isinpair * time + Error(user1), data = nozerosdemo))