How scientists can think about Prism's mixed model analysis

Most scientists can think of the mixed model in Prism as repeated measures ANOVA that allows missing values.

What to say in Methods sections or Figure legends

Here is some suggested wording for explaining how repeated measures ANOVA with Prism are done (using the Mixed-model approach):

Repeated measures ANOVA cannot handle missing values. We analyzed the data instead by fitting a mixed model  as implemented in GraphPad Prism 11.2. This mixed model uses  a compound symmetry covariance matrix, and is fit using Restricted Maximum Likelihood (REML). In the absence of missing values, this method gives the same P values and multiple comparisons tests as repeated measures ANOVA that assumes sphericity. In the presence of missing values (missing completely at random), the results can be interpreted like repeated measures ANOVA. Denominator degrees of freedom for tests of the fixed effects were computed by the Kenward-Roger method. 

Name whichever denominator degrees of freedom method you actually used, because it changes the P values and the confidence intervals. If you chose Satterthwaite or the Legacy method, say so in place of Kenward-Roger. Reviewers who work with mixed models will look for this, and it is what makes your analysis reproducible in R or SAS. See denominator degrees of freedom in the mixed model.

Prism does not apply the Geisser-Greenhouse correction when it fits a mixed model. If you also analyzed some data with repeated measures ANOVA (for example, by choosing It depends, which uses repeated measures ANOVA when there are no missing values), state whether that analysis used the Geisser-Greenhouse correction.