Message-ID: <20120920133014.40010@gmx.com>
Date: 2012-09-20T13:30:14Z
From: knallgrau at gmx.com
Subject: stats q: median regression and using ln(y+1)
Hi there,
I hope you don't mind me asking a general stats rather than a R-specific question.
I'm estimating a median regression model and would like to interpret my (mainly categorical) regression coefficients in terms of percentage changes in the dependent variable. There are about 10-15%(naturally occuring) zeros in my datasets. The dependent variable (y) varies from zero to about 100,000.
Would using ln(y+1) as the dependent variable be more defensible with quantile regression compared to the OLS context?
Best,
Irene