PSYCH 209 Lecture Notes - Lecture 5: External Validity, Internal Validity, Covariance
Document Summary
Is this difference statistically significant: external validity, to whom or what can we generalize this effect? (external validity is rarely prioritized in an experiment) Three criteria for causation: covariance, the two variables need to be associated, as a changes, b changes, temporal precedence, one of the variables always occurs before the other in time, a comes first in time, before b. Internal validity: the only thing that could"ve caused changes in the outcome, random assignment, there are no possible alternative explanations for the change in b, a is the only thing that changed. Designing an experiment: random assignment to conditions, at least one variable that is manipulated. Its variation is independent of, not influenced by, the other variables: at least one variable that is measured, dependent variable. It is left free to vary on its own, dependent only on any changes in the manipulated variable.