November 2017
Beginner
286 pages
8h 13m
English
You may be tempted to set up a single large regression to answer several causal questions that exist in a model or data; however, in observational settings (including experiments in which certain conditions of interest are observational), this approach risks bias. The bottom line here is, don't assume anything about any perceived relationships (or coefficients), especially don't assume that a coefficient can be interpreted causally. However, a casual inference can be effective (where appropriate) as a method used to improve a statistical model.
Read now
Unlock full access