As you may have heard, the Democratic Congress just passed a big health care bill on a close-to-party-line vote, the Democratic President is about to sign it, and—oh yeah—the Democrats are expected to get slaughtered in the upcoming November elections.
But, as the timing of the above links reveals—the 2010 election prediction was made in September, 2009—it would be a mistake, when November 2010 rolls around, to attribute a Republican sweep to the events of March, 2010. During all those months from September through Feburary, a period when passage of the heath care bill was far from certain, the Republicans maintained their lead in the polls and thus their anticipated midterm election gains.
And I think that, when November rolls around, the pundits will realize this. Sure, one or two might say something about how the voters were punishing Obama and the Democrats for going too far in March, but then the more clueful pundits will get on their case and say: “No! Remember, the Republicans were expected to do well in the midterms, with authoritative predictions being made as early as September, 2009.”
One thing that interests me here is the role of the political science profession in this story. By laying out our 2010 predictions early, we short-circuited what could otherwise have been a popular narrative about the election. This is something I’ve been thinking about for over 20 years—ever since Michael Dukakis’s election loss was attributed (inappropriately, according to our research) to campaign strategies rather than to general economic and political conditions.
It feels good for once to be ahead of the story. And I think we as quantitative researchers should be proud of this, whether we’re happy or sad about the new health bill, and whether we’re happy or sad about the possibility of a Republican takeover in November.
Quantitative research is not just about making predictions; it’s also about changing the storyline.
P.S. Here’s the key graph (from Bafumi, Erikson, and Wlezien):
Follow the first link above for more discussion of the research.
{ 5 comments }
Andrew,
I think one of the under-appreciated mechanisms by which this occurs is that blogger/pundits such as Matt Yglesias or Ezra Klein take political science research much more seriously than, say, Matt Bai, and thus help transmit PS results (ala the research posted in this blog) into the larger media sphere.
I hope you’re right. But it seems to me that this lesson is never learned, because powerful political interests stand in the way of its ever being learned.
Political scientists have been writing about surge and decline for a long time, but those who want to claim that the public rejects the president’s agenda will have no hesitation in doing so if the midterm results give them anything to work with. There’s always enough uncertainty in politics that such claims get a fair amount of respect.
How many pundits claimed that the Mass. special election result was a repudiation of health care reform, even though Mass. opinion polls showed solid support for health care reform even among Brown voters? How was the discourse constrained by data in that case?
Did it matter that polls in 1994 showed that most voters had never heard of the Contract with America and, of the voters who had heard of it, most claimed that it did not make them more likely to vote for Republicans? Did that affect the claims of a mandate?
If/when the Democrats take a beating in 2010, Republicans will trumpet that it means that the public wants to stop the president’s agenda. And mainstream pundit opinion will generally endorse that view.
You should get your shoulder checked out. Atleast the Republican’s waited until they lost before they started their victory lap. So by predicting failure early, you get to claim the spin on why it happened? Just because you are right about the outcome doesn’t mean you are right about the reasons.
You have a higher opinion of political scientists than the general public does. They’ll believe the common sense view of the pundits who will say the democrats lost because of the choices they made during the previous 2 years rather than the professors who interpret the soft polling data. Count me among them. It seems every MonkeyCage post points to the economy and discounts anything else as a factor. The economy may be the big kahuna but these other issues must certainly be nudging public opinion in one direction or the other.
Maybe an ostrich would be a more appropriate mascot than a monkey.
Anonymous:
Please read the above post more carefully.
I did not point to the economy in regard to 2010, nor did I “discount anything else as a predictor.” Rather, I showed a graph and linked to work that shows that you can predict the Congressional vote given early polling. Those graphs are based on polling data in the year before the election, which is, in your words “during the previous 2 years.”
to me, the interesting question is, though – if Democrats don’t get slaughtered, does that mean the model is more broadly wrong or does that mean that HC changed the game?
Comments on this entry are closed.