Dear all,

I have a question on rescaling/transforming a variable, particularly a weight variable. In some papers, I have seen that authors rescale/transform the sample weight variable before running regression models, for example:
Code
Code:
gen new_weight = original_weight*1000
or 
gen new_weight1 = original_weight/1000
Can anyone help me to understand why they need to do such a transformation?

Thank you.