r - How to model fixed effects in lme? -

i have longitudinal measurements , basic demographic variables age , gender , ı'd model measurements lme.

what things must take account of when modeling fixed effects part in lme? i've read many questions , answers on topic, i'm not quite sure 1 apply.

for analysis, model fixed effects part, first of used graphics examine relationship betweeen response , explanatory variables(one one). used possible options modeling fixed effects , utilized information criterias (aic, bic) decide model use among these options. utilized both graphics , information cirtria values , tried univariate analyses (such t-test, chi squared tests) variable selection find potential risk factors response, i'm not sure true apply univariate analyses.

after applying these methods, decided use main effects in fixed effects part because model gave smallest aic , bic did not find trend in graphics shows interaction between candidate variables. possible include main effects or not logic? not know answer. therefore appreciated.

thanks in advance!