[R-sig-ME] Predict response variable Lmer

[John Haart]

Dear List,

I have arrived at my most parsimonious set of models (AIC & BIC) using lmer, however i am unsure how i go from this to predicting the response variable when i only have the fixed effects.

I.e 

Best model = 

H~ a+b+d+f 

Where H is a binary response variable

How do i predict if H will be true or false if i have a,b,c,d,e,f etc?

My aim is to take the fixed effects and predict if the response will be true or false.


Thanks very much

John