How ‘Primary Colors’ Helps Political Reporters

Share With Friends

I’ll have a longer post about this over at Primary Colors with examples of different types of news events our data will provide useful insight on for journalists, but here’s a quick example from my run-down of the 2014 primaries to watch.

In MI-14, you have the incumbent Congressman running for US Senate, and an open seat primary to replace him. Naturally, people are interested in the question of how liberal or conservative the next member will have to be to win the district.

Up until now, there hasn’t been a good basis for anchoring that discussion. But now the analysis can be contextualized in terms of the district, rather than the ideological preferences of the departing MOC:

The departing member’s information is included for reference, and their ideological data appears in the next three columns – their Primary Score (where 0 is a good liberal and 10 is too conservative), their Actual Score (how often they vote with progressives) and their Expected Score (how often we’d expect a MOC from a district like this to vote with progressives).

This is key because seeing the departing MOC’s scores can help anchor expectations for the new candidates running for the seat. Maybe the departing member was unusually conservative, as in the case of Gary Peters in MI-14. The candidates running to replace Peters might think they have to vote as conservatively as Peters did, but in reality, an MOC representing a D+29 district would be expected to vote with progressives about 92.2% of the time. Peters only voted with progressives 74.2% of the time:

This entry was posted in US House.

6 Responses to How ‘Primary Colors’ Helps Political Reporters

  1. YoungPhillyProgressive says:

    Didn’t Gary Peters’ district become a lot more blue in the latest round of redistricting? Is this compensated for?

    • Jon Geeting says:

      I’ll ask Ryan. Basically the recent scores from the current Congress are weighted a lot more heavily than lifetime record, so recent behavior will impact the scores more than older votes.

      • YoungPhillyProgressive says:

        OK I don’t know a ton about DW-NOMINATE (but I suppose knowing *anything* is more than most) but I do know they have them broken down by session.

        So if anything we should be holding longtime members accountable to their new districts. I wonder, though, if there should be some type of exception for old members in a new district like in the first cycle, if that makes sense. For example, if Gary Peters was still sitting in a D+29 in 2016 and still voting like a moderate.

        I dunno if that’s something to consider, but that’s my thought.

        • You’re definitely not wrong — it’s a fairly difficult question to answer, because the algorithm does take this into account, but not directly.

          The way the algorithm is set up is to heavily way this current session of congress far more than the previous sessions. So, while they are in there, the vast majority of each member’s score is based solely, in this case, on 2013-2014.

          The main reason for this is because, while lifetime voting is important, a member’s voting patterns this congress is a much more important factor to base a primary on. But it also helps take care of the sub-reason for districts partisan leans changing after redistricting.

  2. Pingback: 3/12 Morning Buzz | PoliticsPA

  3. Kingofsmoke says:

    Of course him trying to run for senate has to be considered for some of his rightward lean. Although he might just be this way naturally.