1
$\begingroup$

The title of this question says it all. I know that all methods have limitations, and while I know some of the strengths of raking weights (e.g., often, only marginal distributions for auxiliary variables are necessary), I'm not sure where raking weights wouldn't be appropriate or what their limitations might be.

Can anyone help me out?

Improve this question
$\endgroup$
1
  • 1
    $\begingroup$ Rim aka raking weighting assumes that the "forces" which distorted the structure in the sample (relative the true, population structure), operated on the variables independently, aka marginally. The distortion between genders (say) occured independently from the distortion between age groups. And that is logically why we "cure" the distortion by the proportional fitting algo which operates on marginals. $\endgroup$
    – ttnphns
    Commented Oct 23, 2021 at 19:02

0

Reset to default

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Browse other questions tagged or ask your own question.