Possible Duplicate:
Interpretation of log transformed predictors in logistic regression
Should quantitative predictors be transformed to be normally distributed?
I was hoping for some general advice. I am using STATA to undertake a binary logistic regression with several independent variables. Most of these are categorical except for one which is a continuous variable. I have already found that this variable is not normally distributed and so I have log10 transformed it. In my regression analysis I have used the log transformed values of the continuous and get an odds ratio of 3. However, I am struggling how to interpret this value in terms of the outcome (disease or not disease).
Any thoughts or advice would be really appreciated.
Thanks
