PSY201H1 Lecture Notes - Lecture 12: Statistical Hypothesis Testing, Sampling Distribution, Sample Size Determination
Document Summary
Should be no effect, but our study suggests that there is. Should be an effect, but our study suggest there isn"t. We never know the truth of the universe or the true status. We try to minimize the probability of mistake. Probability of making the correct decision (to retain h0) when h0 is true. So, mistakes will be rare when h0 true! Up to the researcher - we can control error by choosing alpha. If the consequence of t1e are serious, then we should choose a more stringent (smaller) value. 1- = power = probability of making a correct decision when h0 is false. How to t2e (and power? ) E. g. difference between groups, strength between variables. When an effect size is small, you need more power to detect the difference !! Bigger size is needed for us to detect a small.
