Does selectively reversing variable codes (to achieve uniform sign of coefficients) improve multiple regression in any way?

I was once advised to reverse the coding of any independent variables that have a negative (say) correlation with the dependent variable, so as to achieve the same sign for the coefficients of all...

24 August 2021 3,569 7 View