event/trials predictors and outcomes in glimmix

[telligent]

I am modeling a proportion in glimmix using event/trials format of the outcome variable.  If my predictors are also proportions with the same denominator as the the outcome, would I just enter the numerator as my predictor since the number of trials is already accounted for?  Or should I use the actual proportion or fraction as my predictor. 

[StatDave · Accepted solution]

You would use the actual proportions as your predictor values.

[telligent]

Curious if you think it is important to model proportions this way or if modeling as continuous is just as well.