Dear All,
I have the following question. I am using the lme package to estimate the effect of a given covariate on the conditional mean of my dependent variable. my data is longitudinal (with respect to time). when I considered the random intercept model (i.e. only the intercept is considered as random effects), the effect of my covariate was negative and higly significant. However, when I used the randon slope model (i.e. the intercept and slope are considered as random effects) the effect of my covariate is no longer significant. In addition, the AIC of the random slope model is relatively small than that one of the random intercept model.
I would be grateful is someone can explain me this behaviour,
Thanks,
Abderrahim
Abderrahim Oulhaj, Phd in Statistics
Oxford University
Department of Pharmacology
Mansfield Road
Oxford OX1 3QT
Tel: +44 1865 224098
Fax: +44 1865 224099
Email: [log in to unmask]
|