The answer to this question is rife with what Brendan Nyhan “calls”:http://www.brendan-nyhan.com/blog/2010/11/beware-context-free-election-analysis.html “context-free analysis.” Here’s his example regarding the Tea Party, from “MSNBC”:http://firstread.msnbc.msn.com/_news/2010/11/03/5403120-just-32-of-tea-party-candidates-win:
bq. For all the talk of the Tea Party’s strength – and there will certainly be a significant number of their candidates in Congress – just 32% of all Tea Party candidates who ran for Congress won and 61.4% lost this election.
He correctly notes that we can’t draw any conclusions about the Tea Party’s apparent effect if we don’t account for other factors.
Here’s an analysis that does. I used the New York Times’ “data”:http://www.nytimes.com/interactive/2010/10/15/us/politics/tea-party-graphic.html?ref=politics on which GOP candidates were aligned with the Tea Party. Brendan Nyhan had already compiled these data (see his earlier “post”:http://www.brendan-nyhan.com/blog/2010/10/did-the-tea-party-weaken-gop-candidate-quality.html) and kindly shared them.
My model estimates whether the Democratic vote in a contested House districts was higher or lower if the Republicans was affiliated with the Tea Party. I control for the “underlying partisanship of the district”:https://themonkeycage.org/2010/11/back_to_basics_districts_and_d.html, the Democrat’s share of the vote in 2008, the partisan balance of fundraising, and if there was a Democratic or Republican incumbent running.
The results: on average, the Democratic vote in districts with Tea Party Republicans was 1.3 point _lower_, other things equal. The 95% confidence interval ranges from -2.3 to -0.3. Thus, affiliation with the Tea Party helped Republicans.
We can refine the analysis. First, the vast majority of Tea Party Republicans (111 of 127) ran in races with Democratic incumbents. Isolating only those races, the same effect emerges, unsurprisingly: a drop of 1.3 points.
Second, in districts with open-seat races, Tea Party affiliation had no effect. The 15 Democrats running against Tea Party Republicans for these open seats did not better or worse than the 25 running against other Republicans.
Interestingly, I cannot find any evidence that the performance of Tea Party candidates depends on the partisanship of the district. So it’s not the case, say, that Tea Party Republicans do better in Republican-leaning or swing districts than in solidly Democratic districts.
This analysis cannot address three important questions. The first is simply whether there is some other factor at would account for the small Tea Party effect, but is not included in my model. If so, then even this 1.3-point effect is spurious.
Second, assuming the effect is real, it does not tell us_why_ the Tea Party candidate did a little bit better. I am controlling for fundraising, so it is not that Tea Party candidates do better raising money than other Republicans.
Third, it assumes that Tea Party affiliation is a cause of outcomes, when in fact the opposite could be true. Perhaps GOP candidates, anticipating that the outcome will be somewhat in their favor, believe that affiliating with the Tea Party is a good strategy.
With those important caveats in place, the provisional conclusion is this: in 2010, affiliating with the Tea Party helped GOP candidates win an additional 1% of the vote.