POL 314 Lecture Notes - Lecture 1: Jerzy Neyman, Bayes Estimator, Counterfactual Conditional
Document Summary
But it really took off under the name behaviorism in the 1950s: regressions, once you have collected some data, you want to analyse them to infer some causal relationship b/w your independent (or exogenous ) variables the presumed causes and the. It is based on counterfactual reasoning and is an attempt to replicate direct experimentation. Then you treat group 1 (the treatment group) w/ drug a and group 2 w/ a placebo (a sugar pill). If there is a difference between the outcomes for the treatment group with those for group. To make sure that they have identified the proper causal linkages. Controlling for variable z, in this case) to measure the real effect of a on b. Too often researchers who come up with multivariate regressions feel the need to serially control for each one of their variables. Had you collected your data on another day or slightly differently, the results could have been totally different!