Dear all,

I have two dependent variables (one is a percentage of a total and one is measured in cm). Both of the dependent variables are negatively skewed. The independent variables are categorical variables. I am thinking in using GLM but not sure which link and family to use. Also do I need to transform the dependent variable (e.g. square root transformation to improve normality??) before carrying out the regression analyses??

Many thanks for your help,

Regards,
Sonia