I am using R to run random effect models and fixed effects models. I have data collected over 4 times points in which each of the 24 subjects took part in answering them. Now, I wanted to know how factors A and B affect C overtime (1) and over each subject (2) score in factor C? To do so, I ran fixed effect model to understand the effect of A and B on each individual's score C. I did this and I know that they are alternatives to this type of analysis but I really appreciate if someone out there could direct me whether what I did is logic and does make sense. My logic tells me that why not, why not to understand effects of factors on each subject on small scale studies (experimental for example). Maybe some visualizations will help, attached is a pic of results table I made
I would need some more information on your data. What are 'A' 'B' and 'C'? And what kind of research question do you want to answer?
Regardless of that, I would always advise you to run an mixed effects analysis, given that you have 4 measurement occasions within a subject. Most appropriately you take this dependency between scores into account (a.k.a. the problem of sphericity)
Thank Sven for your answer. It was something out of curiosity to run fixed effects model. I have collected data while a go for a project on developing chidlren reading practices.
I assumed that A and B (likert-scale tools) have unique effect on each student score of C ( reading test). So, I wanted to run analysis to observe each effect of A and B factors on each student score of C.
Again, thanks for knowledge sharing.
To all knowledge seekers out there,
I have had the chance to read the following document from Bodo Winter, University of California. It answers a lot of inquires I have had about R for random and fixed effects models, esp. in my field. I am posting the link to the document below.
http://www.bodowinter.com/tutorial/bw_LME_tutorial2.pdf
In case you are not able to get it, drop a message here.
- Badges
- Science topic