The Generalized Linear Model (GLM) and Maximum Likelihood Estimation (MLE) (Houston, TX)

Instructor(s):

  • Ling Zhu, University of Houston
  • Justin Kirkland, University of Houston

From survey responses to event occurrences and event counts, political and social science data are nearly always categorized by the limited distribution of the variables we are interested in modeling. In this class we will explore limited-dependent variable models, and how social scientists can make use of these models to test arguments. We will tackle both the theoretical under-pinnings of these models, and the application of these models to real-world political and social data. Students in this class will learn to tackle research questions with regression-style analyses regardless of the distribution of the variables-of-interest. With careful introductions to logistic regression, count models, ordinal and categorical models, simulation, and out-of-sample prediction, students will emerge with a well-rounded methodological tool kit and well-prepared for observational research of many types.

Prerequisites: A basic statistics course and familiarity with linear regression

Software: R Statistical Computing Environment

Fee: Members = $1700; Non-members = $3200

Tags: MLE, maximum likelihood estimation, GLM, Generalized Linear Model

Course Sections

Section 1

Location: University of Houston -- Houston, Texas

Date(s): May 21 - May 25

Time: 9:00 AM - 5:00 PM

Instructor(s):

  • Ling Zhu, University of Houston
  • Justin Kirkland, University of Houston