Dear Listers,

Using -contrast- after the -mixed- command allows you to test the main and interaction effects:
mixed y a##b || id:, contrast a##b The -contrast- command estimates a chi-square test - is this a Wald test? I have seen online references suggesting the chi-square test can also be converted into ANOVA- style F-ratio tests (chi-square/df) and in this resource it shows how for repeated measures, F-ratio tests can be directly obtained by specifying repeated methods and small samples: https://www.stata.com/stata-news/new...ing-contrasts/ mixed y a##b || id:, dfmethods(repeated) contrast a##b, small How small would the sample really need to be for one to be able to use such an approach? Most readers are familiar with the F-ratio so it may be helpful to report it over the chi-square. I would also welcome some more information on the relationship between the chi-square and F-ratio tests calculated by the -contrast- command following -mixed- Any resource welcome. Thanks in advance,
Laura