I have a simple pretest-intervention-posttest that I need to analyze. I want to determine if there is an improvement in performance between pretest and posttest, and what that difference is. This is all ratio data (percentages) and the pretest is severely left-skewed. I have an ANOVA (maybe not valid with skewness) that shows significance, and a simple gain score analysis that shows a 36.1% increase on learning. n=37
I'm looking for someone to take my data, advise (and document) on what tests to run based on my design, and run those tests.