I am working on longitudinal data (from a panel survey) that were analyzed using mixed effects modelling to estimate risk differences, the outcome is contact with any health provider, and predictors being both fixed (like socio demographic factors) and random (like having symptoms). I am finding it difficult to interprete the results of the adjusted models, certainly because I don't fully understand the concept of this modelling technique. Could you advise on any reference for (non statistician) clinicians? Thank you
I recommend you consult the course notes by Prof Marie Davidian at the Stat department of NCSU FOR longitudinal models. They should be of great help to you. They include useful R and SAS Programs that may be of help to you. google should help you find all of this. Best wishes, David Booth
The course suggested above by David Eugene Booth is really useful.
Btw, how do you measure the outcome?
Thank you very much David, I have checked the course and what a (generous) asset it is.
Andrey, the outcome is binary (Y/N) and was collected by self-report on daily questionnaires.
Monica, because your outcome variable is binary you want to build your model with the methods of Marie Davidian's course notes but do the regression not by ols but by logistic regression. I have attached a second set of course notes by Frank Harrell that cover this topic with examples. . for more details and examples you can Google logistic regression. Best wishes, David Booth
- Badges
- Science topic
- Similar topics
- Engineering
- Modeling