PS295 Lecture Notes - Lecture 7: Confounding, Cognitive Dissonance, Statistical Hypothesis Testing
Document Summary
An extended example to compare correlation and experimental research. Steps of research followed with either method: formulate hypothesis, often involves causal relation, collect and summarize data, interpret the results, statistical tests, judgement about cause, explanation. Correlational study to test the initiation hypothesis: ask fraternity members two questions, how severe was the initiation rite you went through? (1-11, how much do you like the other fraternity members? (1-11) Alternative method for summarizing the data: classify people into 2 groups based on their severity ratings: severe initiation group, mild initiation group, do statistical test comparing the mean liking rating across the two groups. Interpretation: is the relationship causal: can"t be confident that it is - it might be, without manipulation and random assignment, can"t be sure that iv cause changes on. Explanation: discuss various causal patterns that might be producing the relationship, and offer a rationale for each one.
