Hi Everyone,
I have a multinomial regression model:
y=b0+b1x1+b2x2+b3x3+b4x4+error
I have 2 highly correlated independent variables, say x1 and x2, and want to orthogonalize them. I regress x1 on x2 and take the residuals of this regression. Then, I add these residuals and X2 variable in my regression.
I also encountered orthog function in Stata. As far as I understand, this function orthogonalizes two variables with respect to the order. When I use orthog and run the regression with newly created two orthogonal variables, I was thinking that I would get the same results as using residuals and x2 but that didn't happen.
Does anyone have an idea why I am getting different results when I use orthog? Thanks in advance.
Best,
Ulas
I have a multinomial regression model:
y=b0+b1x1+b2x2+b3x3+b4x4+error
I have 2 highly correlated independent variables, say x1 and x2, and want to orthogonalize them. I regress x1 on x2 and take the residuals of this regression. Then, I add these residuals and X2 variable in my regression.
I also encountered orthog function in Stata. As far as I understand, this function orthogonalizes two variables with respect to the order. When I use orthog and run the regression with newly created two orthogonal variables, I was thinking that I would get the same results as using residuals and x2 but that didn't happen.
Does anyone have an idea why I am getting different results when I use orthog? Thanks in advance.
Best,
Ulas
