Falsification may be difficult to reach in your experiment. The design must be fit to prove the contrary of what you are trying to demonstrate. If you prove that the result can’t be explained by normal distribution, you will have an acceptable answer. Your experiment must replicable and replicated in many conditions.
It won’t be easy. First, asking someone to criticize a paper is a big clue to what you hope to get as a result. Second, having people patient enough to read the lessons is auto-selecting a category of the population. You won’t be able to know whether those people read your lessons because they have been training for critical thinking for a long time ; they may already be looking for such material and others may be just bored.
I think you need a pre-test and a post-test, and this test must very light or pretending something completely remote to what you really want to prove.