FPC for IPW and Survey Weights

[Luis Mijares Castaneda]

Hello, I am trying to implement an IPW methodology using data from a survey. I followed the methodology in the Ridgway paper, "Ridgeway, Greg et al. “Propensity Score Analysis with Survey Weighted Data.” Journal of causal inference 3.2 (2015): 237–249. PMC. Web. 11 June 2018."
Propensity Score Analysis with Survey Weighted Data - PMC

"We find that using sampling weights in the propensity score estimation stage (as weights, not as a covariate), computing final weights as the product of the sampling weight and the propensity score weight, and using those final weights in an outcome model will be robust in these scenarios and be competitive in other considered scenarios." I make these final weights with the code below; however, when I then use the -fw- final weights in the logit model below, I get an inflated population size/FPC, is there a formula or method to adjust the FPC after creating weights from multiplying IPW and survey weights?

 

gen fw = IPW * SurveyWeights

svyset psu [pweight = fw], strata(strata)   

svy: logit dental shootings `controls', or

[Richard Williams]

My inclination is to do this. Suppose the inflated population size is 10 million and it is supposed to only be 5 million. So, I would divide fw by 2.

im not sure how much it matters though. So long as the relative weights for each case are correct, I think the coefficients and standard errors will be correct.

i haven’t read the paper though, so it may be more complicated than I realize.