Nate Cohn

Nate Cohn @Nate_Cohn

@cwarshaw @baseballot @lxeagle17 @DrewLinzer @Civiqs and i think there's a distinct chance that we'd transition to MRP in future cycles -- not because of small area estimation but because it's a vastly better framework for adjusting on multiple variables — PolitiTweet.org

Nate Cohn @Nate_Cohn

@cwarshaw @baseballot @lxeagle17 @DrewLinzer @Civiqs @kabir_here @doug_rivers i guess i'll add one caveat -- i do think there's a good case for 538 to exclude Civiqs modeling, not on the groudns that it's MRP but because they appear to pool data over time (but @DrewLinzer can correct me on that). that is the kind of thing that 538 wants to handle itself — PolitiTweet.org