300294 Method Class quantitativ: Selection Bias in Applied Research (S) (SoSe 2023)

(Self) selection into samples of analysis and missing data are pervasive challenges of applied research that potentially bias inferences from the data. In fact, many (doctoral) students will encounter issues of selection bias in their research and will address their implications, either directly in their analysis or indirectly in the discussion of limitations of their conclusions. This methods class introduces conceptual debates on selection bias and outlines alternative solutions, such as weighting of data, multiple imputation of missing data as well as model-based techniques (e.g., Heckman Selection Models).

In hands-on sessions using own data as well as example data, we exercise the implementation of these methods using statistical software packages (Stata, but also R if required). Please indicate preferred methods and your interests in the topic in advance.