Course Objective
AEM is the second course in the empirical methods sequence in the MPAff core curriculum. This section is an extension of the statistical inference portion of IEM, structured around the logic and the application of advanced regression modeling in social policy analysis.
No background in social policy is needed. We will use empirical research on social policy as the context to structure our learning process. That objective is to develop conceptual logic as well as practical skills in statistical analysis of policy issues. The logic and skills learned in this course should be applicable to other policy areas as well.
The substantive content of the course is about 50 percent on understanding advanced regression modeling, including logit, tobit, and poisson models, and dealing with data biases and violations of Gauss-Markov assumptions; and 50 percent on practical knowledge about research design, national social data files, and hands-on data-analytic skills using Stata as statistical software.
Learning Experience
Limited “textbook” type reading suggestions will be available in the first half of the semester, but course materials are organized mostly around actual empirical research articles. It is through the meticulous reading of and reflection on analytic articles, on average one per week, that we develop our analytic skills as well as substantive understanding of regression logic. Several problem sets will be offered to practice such analytic understanding. The course will also include two individual-based Friday Learning Experiences, one around the fifth or sixth week and the other around the tenth or eleventh week of the semester. This is the “substantive knowledge” part of the learning experience.
The second major component of this course is an independent research project—team-based or individual-based depending on class size—which each student will start working on from the beginning of the semester. This research project will require (1) the formulation of an empirical research question, (2) the use of a data set to analyze the problem, (3) the completion of two research papers, and (4) a teaching session on the analytic issues involved. This is the “hands-on” part of the learning experience.
Expectations
Successful completion of IEM at the LBJ School or its equivalent is a prerequisite. In particular, class members need to be proficient in calculus as analytic language and multivariate regression as background knowledge. The latter should include clear conceptual understanding of the logic of least square errors as well as associated Gauss-Markov assumptions.
Class members are expected to do preparatory work during winter break by (1) deciding on a researchable policy question and identify a useable data set for addressing that question, and (2) reviewing, if needed, mathematically oriented notes on statistical inferential logic. This course will begin its Blackboard-based learning process in early December. Interested students who fail to register for this class in October should contact the instructor as soon as possible to ensure access to Blackboard.
Class members can vote, on a majoritarian basis, whether to take notes in class or to refrain from note-taking in exchange for weekly notes from the instructor.