PSYC 311 Chapter Notes - Chapter 16-17: Statistical Significance, Statistical Power, Research
Document Summary
Chapter 16: the politics, production, and ethics of research. Cost benefit analysis: make best use of limited resources. To become basis for policy or admin action - need statistical significance. Don"t want to use this too rigidly, depending on the context higher p values can be evidence to act. Also need look at practical significance: confidence intervals in addition to significance tests. Magnitude of the effect big enough to be worth implementing. Effect sizes a a first step cost - benefit analysis is useful to use here. Statistical and practical significance is not enough to cause causation. Before implement program: need establish that effect is causal. Real world cannot always wait for definitive evidence, sometimes will act even without definitive causal evidence. Replicate the same studies is sometimes not enough to provide evidence to justify action. 2 kinds uncertainty in applying research to make policy. Statistical question: is there a real difference of just fluke.