PSYB04H3 Lecture Notes - Lecture 7: Regression Analysis, Sampling Frame, Discriminant Validity
Document Summary
No sampling frame or bad sampling frame. Both variables are measured (but none are manipulated) The degree to which measure variables correlate: examine relationships to make predictions, no indication of cause and effect, bivariate correlation = 2 variables, multivariate correlation = 3+ variables. Variables should be selected logically: construct validity: Correlation: expresses the strength and direction of the relationship between two variables. Estimates how well a set of predictors forecasts a criterion. There are 3 things needed for causality: temporal precedence (the establishment of which variable came first, covariance (positive correlation, negative correlation) Unrelated correlation are correlated due to unseen factors. Ways to meet some of the requirements for causal claims: longitudinal designs, sample people over time, help establish causation by establishing covariance and temporal precedence. Including more variables: deals with some problem of internal validity, rules out some 3rd variables by controlling them. Answe(cid:396)s the (cid:862)why(cid:863) (cid:395)uestion (cid:894)gets at the unde(cid:396)lying p(cid:396)ocess(cid:895)
