A/B test questions: split users into 2 groups or test twice on the same users If we are testing drug effects (A placebo vs. B), and here are the two options: