1

Is there a way to create a synthetic control using pooled cross-sectional data, rather than a panel?

I'm analyzing a staggered reform and, so far, the methods I've found require having a panel: Ben-Michael, Feller & Rothstein 2020 (augsynth), Callaway & Sant'Anna 2019 (did), Xu 2017 (gsynth), and potentially also Robbins, Saunders, & Kilmer 2017 (microsynth).

luisdiego24
  • 13
  • 5
  • What's the unit of observations in your data? – dimitriy Aug 11 '20 at 01:31
  • Individuals within a household. My data set is made up of 22 pooled cross-sections, where each subsequent year a quarter of it is resampled, so I don't have the same households every year. – luisdiego24 Aug 11 '20 at 02:02

0 Answers0