Hi all,
I have an independent variable in my regression that lies between .0000869 and .0610705. Its coefficient shows up rather large. So I rescaled the variable by using the following:
Code:
gen newvar=(oldvar-.0000869)*(1/(.0610705-.0000869))
And it seems that the magnitude of the coefficient lies in the same range as others do. The coefficients of other variables and their p-values remain the same. Only change is in the coefficient of the rescaled variable. Is this a correct way to deal with a large coefficient? It's not my main variable of interest but it's just a control variable so I thought it would be alright.
Many thanks.
0 Response to is this a correct way to rescale a variable?
Post a Comment