Description:
The purpose of this course is to develop an in-depth knowledge of bivariate and multiple regression/correlation techniques. This includes the theory underlying the methods (in particular, OLS, some WLS and logistic regression), computational procedures, and interpretation of results. The course will be an introduction to specific applications. These include: (1) the full range of correlation indices; (2) a range of regression strategies (e.g., reduced-form regression, path analysis, ordered and unordered step-wise inclusion); (3) statistical power; (4) regression diagnostics; (5) nonlinear regression and linearizing transformations; (6) testing interactions, and (7) conditions for causal analysis and analysis of change (primary focus is on manifest variables using OLS and logistic models). Spring (3) Cordray.