1

Let $Y$ a real positive variable. I have a data frame with the measurements $X = mean(Y)$ of four distinct categories of subjects:

  • Healthy control during control condition
  • Healthy control during experimental condition
  • Diagnosed during control condition
  • Diagnosed control during experimental condition

The data frame then looks as follows:

Subject | SessionType | Subject Group | mean(Y)
   1          Control        Healthy       x1
   1          Exp            Healthy       x2
   2          Control        Depressed     x3
   2          Exp            Depressed     x4

and so on.

I want to compare the means of mean(Y) across SessionType and SubjectGroup using ANOVA (assume the conditions for this test are met). Since each subject is contributing two values, one for each sesion type, I must use a repeated measures ANOVA. But I was also suggested to fit a mixed linear-effects model on the data.

What would a mixed linear-effects model reveal in this context, and in contrast with what the ANOVA test would reveal?

lafinur
  • 235
  • 1
    There are quite many questions and answers here about this exact issue: https://stats.stackexchange.com/questions/24314/when-is-a-repeated-measures-anova-preferred-over-a-mixed-effects-model and https://stats.stackexchange.com/questions/397540/why-use-linear-mixed-models-instead-of-repeated-measures-anova for instance – Sointu Nov 24 '23 at 19:40

0 Answers0