Multiple weights

[emily taylor]

Hi there,

I am currently conducting a regression discontinuity using an analytical weight of county population. I now wish to also incorporate an inverse distance weighting (where results further from the discontinuity carry less weight).

I have tried adding [pw=inversetime] into my regression, but it states 'too many weights'.

Is there a way of adding more than one weight into a regression on stata?

Many thanks!

[David Radwin]

You can combine multiple weight variables into one variable, but that is a tricky business that depends in part on the underlying meaning and purpose of each set of weights. This approach is used, for example, in certain surveys to adjust for unequal probability of selection and nonresponse (and often other factors as well). You might consult sections 7.2 and 7.4 of Valliant, R., & Dever, J. A. (2018). Survey weights: a step-by-step guide to calculation. College Station, TX: Stata Press.

[Denat Ephrem Negatu]

Is there more info on combining survey weight and inverse weight variables into one variable? I don't have access to the book

[David Radwin]

This is not my area of expertise, but I skimmed the relevant sections and I believe the answer is no.