PSYCH 241 Lecture Notes - Lecture 4: Causal Inference, Convergent Validity, Internal Validity
Document Summary
Internal validity the degree to which differences in performance (the dependent variable) can be attributed unambiguously to an effect of the independent variable. The conditions necessary for internal validity: time-order relationship, co-variation, elimination of other plausible alternatives. * to test hypotheses about the cause of phenomena : hypothesis may be based on descriptive or predictive research. True experiment is best available method for establishing cause and effect (drawing causal inference: has to be done correctly assign, manipulate, control. Multimethod approach very powerful convergent validity. If numerous experiments lead to same conclusions strengthened: replicate, replicate and extent, within & across labs confidence. Three key steps: random assignment of subjects, manipulation of iv dv, control of external factors. Control is essential, and is achieved through: manipulation of iv, holding other variables constant, & balancing. Manipulate iv to observe effect on dv. Appropriate control allows researcher to eliminate plausible alternatives and make a causal inference the iv caused the change in the dv.