Hello,

first of all I would like to thank you all for all the help I have been getting through this medium the last few weeks. I am in the middle of my data analysis, and I am struggling with some stata related questions, and some statistical questions.

1. Before performing OLS, I winsorize all my independent variables and my dependent variable. Do all variables have to be winsorized at the same % level? for example, I have very big outliers in my dependent variable, but these outliers are a lot smaller in my independent variables. Also, I have taken the natural log of my firm size variable, does it make sense to winsorize this variable, given that it is already expressed as a log?
2. I have performed the hettest for heteroskedasticity. If this test indicates that there is heteroskedasticity, is it sufficient to regress making use of ,robust? and what is this robust standard error exactly? is it White's?
3. I have expressed most of my variables as a ratio (decimal number), and I am strugglign on how to interpret these results. For example, my regression shows a coefficient for independent variable leverage of -0,165119. My dependent variable is the relative change in carbon intensity between 2013 and 2018. Is it correct if I interpret the coefficient as follows: if the debt-to-equity ratio of a firm (= my measure for leverage) is 10 percentage point higher, the carbon intensity will be reduced with 16,51 percentage point more. Or should I multiply these coefficients with 100?
4. if some of my control variables are statistically significant, what does this tell me exactly? Or do they just serve for a more accurate estimation of the coefficients of the independent variables and do not need to be interpret? Same with a constant term that is statiscally significant?

Kind regards,
Timea De Wispelaere