Hi all.
I have question regarding the best way of coding a glme with time point as a factor in nlme.
We have basically have individuals with Fl and FC scores at 3 time points before birth (FL_T1, FL_T2 etc.) and an outcome variable that is to do with their offspring (Outcome_offspring). Individuals have multiple offspring, so their FL and FC scores at each time point apply to multiple outcomes.
We are debating which code is better as the two give drastically different results.
lme(Outcome_offspring~Time_pointFIFC, data=data_setl,random=~1 | Facility/Litter, method= "REML")
or
lme(Outcome_offspring~Fl_T1+Fl_T2+Fl_T3+FC_T1+FC_T2+FC_T3,data=data_setl,random=~1 | Facility/Litter, method= "REML")
The first method is what I came up with and leads to use not only repeating mother variables for multiple offspring, but the outcome recorded for each offspring being repeated for the multiple time points the mother was recorded.
The second method is what the student came up with --> she is getting significant results here when the other method is getting nothing.
I didn't imagine they would be quite so different.
Any advice would be appreciated!