GRAD 6102: Categorical Outcomes

University of North Carolina, Charlotte

While linear models like ordinary least squares regression are robust and useful, they are only applicable when your dependent variable is continuous and uncensored. �Unfortunately, social scientists often want to explain phenomena that are "limited" in that they are not continuous and/or uncensored. �Fortunately, there are alternative estimation techniques for analyzing most models with non-continuous (e.g., dichotomous, nominal, ordinal) and/or censored (e.g., event count or event history) dependent variables. �A variety of techniques are examined for estimating models with "limited" dependent variables.

Average GPA: 3.80

Grade distribution records: 26 students across 8 terms.

Grade distribution

Based on 26 student grade records across 8 terms and 2 professors.

Instructors

• Yang Cao 22 students, Average GPA 3.77
• John Szmer 4 students, Average GPA 4.00