In PA-18, Everything Mattered

After any special election victory, various people are keen to claim that their one strategy worked–or would have worked better had the candidate not followed someone else’s strategy. We’re seeing the same thing happen with Conor Lamb’s victory in Pennsylvania, were a district that went 58-38 for Trump in 2016, was a toss-up win for Democrat Lamb (BAND NAME!), where there are arguments over how this massive swing occurred. Was it a Democratic rally, Republican discouragement, 2016 non-voters returning to the fold, and so on. Of course, as is usually the case (and annoying as hell), these are proxy arguments for policies and the direction of the Democratic Party.

The problem is, as best as I can tell, based on exit polling data, the answer is yes. In a trivial sense, this is always true in a close election. If you win a basketball game by one point, everything mattered: making that extra free throw, not throwing up that bad shot, that lucky call (or no-call). But it really does seem that there was no One Neat Trick that let Lamb win. It’s also clear that 2016 Trump supporters really didn’t show up, while, relatively speaking, 2016 Clinton supporters surged (though turnouts were lower overall).

The best way to see this is to use specific vote estimates based on an exit poll (pdf), and the 2016 and 2018 election results. Here are the figures on the 2016 and 2018 turnouts, as well as the estimated numbers of Clinton and Trump voters (note: don’t take the precise numbers too seriously; estimates from the exit polls and the vote totals differ by around 1,000, which is pretty good, as polls have a margin of error):

(“Trump 2018” and “Clinton 2018” refer to Trump and Clinton supporters who voted in the PA-18 election).

So Trump voters, in a relative sense, stayed home compared to Clinton supporters, who also turned out at very high rates for an off-year election. So people saying that strong Clinton voter turnout doesn’t explain much of the massive swing aren’t making much sense: most of Lamb’s votes came from Clinton 2016 voters. At the same time, Lamb’s ability to convince eleven percent of Trump 2016 voters to vote for him put him over the top. And Trump supporters didn’t show up–and that, too, put Lamb over the top. But so did Lamb’s ability to keep 94 percent of Clinton 2016 voters in line, as well as reach out third party/2016 non-voters (who went 70-26 Lamb). All of these things mattered in determining victory, if not the massive shift from 2016.

So what does this mean, in the sense of claiming victory for one strategy or another? First, rallying the base does matter: if one percent of Clinton voters who voted for Lamb stayed home, Lamb loses. Without base turnout, you can’t win. Shitting on your base isn’t helpful–which Lamb did not do. At the same time, if Democrats can pick off a few Trump voters, attract some previous non-voters, and convince a few Republicans to stay home, they can pick up seats in all but the most conservative areas; as we’ve discussed many times, elections are often won at the margins, so these voters matter too.

Realistically, there isn’t going to be one model or political strategy to do that in different districts. And Saccone did worse than expected in turnout–was that a general Trump effect, or a ‘I don’t like Saccone’ effect? Since non-voters weren’t polled, we don’t know.

Anyway, your favorite just-so story is probably correct in explaining why Lamb won, but so are many of the others.

This entry was posted in Democrats, Voting. Bookmark the permalink.

3 Responses to In PA-18, Everything Mattered

  1. daved says:

    I’ve seen almost nothing about the presence of Drew Grey Miller in the race — had most of his voters thrown support to Saccone, Saccone would have won (the difference between Lamb and Saccone was around 700 votes, and Miller got 1300).

  2. Pingback: Mike's Blog Round Up – NEWSFUZZ

  3. Pingback: Mike's Blog Round Up – Liberal View News

Comments are closed.