I would like to ask you whether I can run two-way repeated measures ANOVA (GLM) to see the effect of scoring method on pass rate when my variables are categorical, and whether I define well my variables. Thank you for your help.
My data contain one categorical dependent variable with two categories (pass = 1 and fail = 0) and four levels: pass/fail count for test A at two cut scores 44% and 65 %, and pass/fail count for test B and cut scores 44% and 65 %. The categorical independent variable has two levels (groups X and Y). Each subject in the groups X and Y was measured twice: with one test (A or B) scored with the old method, and on second time, with another test (B or A) scored with the new method (see Table XXX). I assume there is no test order effect and both tests are equivalent.
Martina
Best Answer
If you have a categorical dependent variable with two levels, you do not want ANOVA or GLM (in SAS terminology), you want logistic regression (which, confusingly, is part of glm in R terminology.
(that confusion comes from the difference between general linear model and generalized linear model, which unfortunately have the same acronym).