I'm currently running a time series analysis which requires me to fit lmer models to each of my data points. Here is my code :
# import data dt <- h5read('file.h5', 'd1/table') dt$cond <- as.factor(dt$cond) dt$sub <- as.factor(dt$sub) # close file H5close() # apply model to each time point fitted <- by(dt, dt$tp, function(x) lmer(size ~ cond + (1 + cond | sub), data = x))
In this model,
cond is a 2 levels fixed effect factor and
sub is a 14 levels random effect factor. For some reason, the model doesn't yield a t-value in its ouput :
> fitted $`0` Linear mixed model fit by REML ['lmerMod'] Formula: size ~ cond + (1 + cond | sub) Data: x REML criterion at convergence: -7694.797 Random effects: Groups Name Std.Dev. Corr sub (Intercept) 1.388e-02 cond1 1.169e-05 -1.00 Residual 7.922e-02 Number of obs: 3467, groups: sub, 14 Fixed Effects: (Intercept) cond1 0.023743 -0.001154
Is it the output I should have expected? It seems that R doesn't take the difference between both levels of
cond into account.