Dependent variables rarely cover the full line of real numbers. Ignoring the true characteristics of your data may lead to inefficient and inconsistent estimates and may even generate nonsensical predictions. This skills training therefore introduces students to:
different types of limited and/or nonmetric dependent variables and the inherent dangers of ignoring the data’s real nature;
models that take into account the peculiarities of the data; and
a particularly popular estimation technique that is flexible enough to estimate all studied models, namely Maximum Likelihood Estimation (MLE).

This skills training consists of four building blocks: i) Maximum Likelihood Estimation (the estimation technique that we will use throughout the course), ii) models for count data, iii) models for nominal and ordered data, and iv) models for censored data. Each of these four topics will be introduced in a lecture. Immediately after each lecture, students start working on an assignment that involves the estimation of one or more of the introduced models.

Selected chapters from textbooks, course slides, course book

Knowledge of regression analysis and Ordinary Least Squares, Knowledge of elementary and matrix algebra and basic calculus, Experience with a statistical package like SPSS
An advanced level of English