Navigation auf uzh.ch
Causal Inference 101: Causal Graph, Potential Outcome and Identification Strategies
We are very happy to announce that Xi Chen will be visiting our chair 21st October. Xi plans to hold a workshop on causal inference and identification.
Long have we been told that “correlation is not causation”, but rarely are we explained exactly how the two differ. Nevertheless, it is essential to understand their differences in marketing, given the basic tenet of marketing is that marketers should be able to justify that intended marketing mix actions lead to suffcient customer responses which surpass the costs of the actions. In this workshop, we will discuss the issue of causal inference and program evaluation in marketing, which surprisingly receive rather limited attention. During the session, we will cover four major topics: first, we will discuss a series of conceptual questions about causality and try to solve the “mysteries” around causality through the discussion; second, we will learn the “new math” of causality - causal graph, and how to use it to “speak” about causality; third, we will learn about the potential outcome framework (i.e. Rubin model) and its link with Pearlian model; fourth, we will learn about the concept of “identification” and discuss four general identification strategies, which can potentially unify all empirical endeavors towards the establishment of causal claims.