PSY201H1 Lecture Notes - Lecture 5: Mhow, Type I And Type Ii Errors, Null Hypothesis
Document Summary
Lecture 5 (october 18, 2016): hypothesis testing i. Proposed result of the conducted experiment: statistical hypothesis, null hypothesis. Null hypothesis significance testing (nhst: most used method of data analysis, often misunderstood. We want to offer up an explanation for a difference that exists: ex: drug group vs. placebo group. Three possible explanations for a difference between the groups. The drug may actually be working (systematic) Could be chance, random factors and errors. Could be a mix of systematic and chance factors. There"s always going to be chance factors at work: people vary, the explanation of the difference between groups is going to be entirely due to systematic factors. First we test the entirely chance explanation and if this doesn"t seem to offer a good enough explanation, then there"s enough evidence there to say that the drug is actually doing something. Hypothesis testing: using sample data to evaluate a particular hypothesis about the population, we make predictions about the unknown population.