I have weight data from 4 animal groups measured at multiple time points. Some data at 2 time points is moderately non-normal. Should I still use repeated measures ANOVA with Greenhouse–Geisser correction, or switch to a non-parametric test? Also, to check for baseline differences, would ANCOVA be more appropriate? I'm not very confident with stats, and my main aim is to assess treatment effects on weight over time. Any suggestions on the best analysis approach?
Humaira Shah
Jochen Wilhelm
Thank you for your suggestions. Body weight is the response variable, and the goal is to assess how it changes over time across different treatment groups, using Day 0 as the baseline. The data has no missing values, and time intervals are equally spaced. Could you kindly recommend a specific regression model that would be appropriate for this type of data, or would linear mixed models be suitable?
0 votes
0 thanks
Humaira Shah
Jochen Wilhelm
Thank you for the detailed explanation. That helps a lot. My dataset includes adult mice only, with weights measured at 6–8 equally spaced time points over a short treatment period. Based on your suggestion, it sounds like a linear mixed model might still be appropriate in this case.I’m considering a log transformation if needed. Would it be best to check residuals and variance first before deciding?
0 votes
0 thanks
Humaira Shah
Jochen Wilhelm
Thanks a lot for your valuable suggestions. They gave me a clearer idea.
0 votes
0 thanks
More Humaira Shah's questions See All
Similar questions and discussions