Logging data is almost as dangerous as logging trees; if you don’t do either the right way, the consequences could be undesirable. Below I will explain why we log data in OLS. Then I will discuss how to interpret the coefficient of a logged independent variable in an OLS model. The interpretation is not as straight forward as it normally is in a linear OLS model. If it is not done carefully, the results of the model may be difficult to understand.
More following the jump….
Julie VanDusky-Allen is at Boise State University and received her PhD in Political Science from Binghamton University in 2011. Her research focuses on institutional choice and development, political parties, the legislative process, and Latin American politics.