Is it appropriate to do a multivariate linear regression where some of the explanatory variables are represented as percentages (continuous from 0 to 1) and the rest are represented in absolute terms (take on values such as 1300, 2000, 2312)? The dependent variable is expressed as a percentage as well. What is a potential workaround to this?
Asked
Active
Viewed 39 times
1
-
"The dependent variable is expressed as a percentage as well. What is a potential workaround to this?: - has this been addressed on this site before? http://stats.stackexchange.com/questions/103731/percentage-outcomes-in-linear-regressions – charles Apr 16 '16 at 08:46