Am I looking for a better behaved distribution for the independent variable in question, or to reduce the effect of outliers, or something else?