Political Science A Politician Can Love

by John Sides on September 11, 2010 · 6 comments

in Political Science and Journalism

Ezra Klein’s new column is entitled “Poli Sci 101.” He pivots off last week’s APSA meetings to note that politicians don’t pay enough attention to political science, and that is a bad thing. He lists a few lessons they could learn, citing the work of George Edwards, David Canon, and Frank Baumgartner, Beth Leech, et al.

He quotes me in this passage::

But if politicians took these findings to heart, it would free them to do their jobs better. “The fact that much of what cable news is talking about on any given day is not important probably is empowering,” Sides says. Particularly combined with the finding that what does matter, both for elections and for people’s lives, is how well the country is doing. Worrying less about tomorrow’s polls and news releases and more about the effect of today’s policies could make for better bills—and happier, more successful politicians.

I want to emphasize this. Political scientists are sometimes accused of denying that politicians can affect anything. For example, we say that election outcomes depend a lot on the economy, and somehow this means that all campaigning by politicians is superfluous.

In fact, Ezra’s point and my quote suggests quite the opposite: political science really does empower politicians. It tells them to ignore a lot of gossip and trivia. It tells them not to sweat every rhetorical turn of phrase. It tells them that as useful as Mike Allen’s Playbook might be in some ways, it captures a conversation that the vast majority of American voters knows nothing about.

Freed from these concerns, politicians can, as Ezra suggests, focus on what they really can and do affect: the policy agenda and the content of legislation. Now that’s a simplification, because certainly members’ power to affect these things will depend on election outcomes—theirs and others’—as well as their status as a majority or minority party member. So we would not expect members to put campaigning entirely out of their mind. Political science is more implying a reweighting of priorities.

I assume this would be sweet, sweet music to politicians, many of whom complain routinely about things like the media and the time they must spend fundraising for their campaign, and profess their true passion for crafting public policies big and small.

A lesson of political science is that the stuff they hate is not as important as they fear, and the stuff thy love is what they can spend more time doing.


Seth Ackerman September 11, 2010 at 8:51 pm

But isn’t there sort of a logical fallacy embedded in this idea that analyzing historical data can tell you that the news cycle doesn’t matter? It’s like the old saw about how in 70 years there’s never been a robbery attempt at Fort Knox, therefore we should send home all the armed guards, dismantle the fortifications and leave it undefended. Of course, if you did that, there *would* be robbery attempts; the historical datum of zero robberies was premised on the actual existence of fortifications. (N.B.: I have no idea if there have actually been robbery attempts at Fort Knox or not.)

Likewise, your scatter charts plotting election outcomes against the economy might show all these election data points hugging the trend-line relatively closely. But those data points represent elections in which the candidates *did* take the news cycle very seriously. You have no idea if the statistical relationship would still hold if you started adding imaginary elections in which the candidates *didn’t* obsess about every rhetorical turn of phrase.

Brian in NOLA September 12, 2010 at 12:02 pm

While it might be appealing to the MC to turn away from “the stuff they hate” and do more policy crafting, don’t they surround themselves with a bunch of people who have the opposite incentives? Media liaisons, campaign consultants, office staff (many of whom come from the campaign) no doubt benefit from having the MC devote time (and money!) on driving the narrative and focusing on polls/campaigning. How can a PR firm or fundraising consultant justify his/her fees unless these things are hugely important?

John Sides September 12, 2010 at 2:25 pm

Seth: My point is not that the economy matters and the campaign does not, my point is that (a) the campaign does not matter as much as politicians think it does and (b) some of the aspects of the campaign (news cycle) that politicians think matter are in fact the aspects least likely to matter. Those conclusions are based on what data we have. There is no way to address the counterfactual you suggest, short of convincing politicians not to campaign. All I can say is that if you put up, say, the trial heat polls from a presidential campaign and then look to see if those polls change after various news eruptions over gossip and gaffes, you will usually see that they do not.

Justin September 12, 2010 at 7:45 pm

I have a lot of methodological worries about these claims in poli sci, but as an outsider, I don’t know where to go looking. Do you know of any good overviews of what kind of evidence y’all appeal to in making these claims?

Btw: I think you sell your discipline short if you say “well, since all politicians campaign, there’s no way to evaluate this stuff”. Unfortunately, I’m saying this in a very abstract way, but think of the variety of experimental designs, ways of looking for natural experiments, etc, that are used elsewhere in the social sciences.

John Sides September 12, 2010 at 9:29 pm

Justin: This blog has presented a lot of evidence for various claims over the past couple years — e.g., on the economy and elections, on presidential speeches, etc. Of course, there is scholarly work that informs the blogposts, but it is too extensive to summarize easily in a comment.

And of course political scientists leverage the variety of experiments — natural and otherwise — that you suggest. But as I understood Seth’s comment, he was suggesting a counterfactual — what would happen if presidential candidates didn’t campaign? — that no experiment could address.

Justin September 15, 2010 at 12:21 pm

Thanks for the answer: occasionally one gets lucky, and someone has written a paper that covers a really wide range of results with relevant references to other papers. I’m not a habitual monkey cage reader, so I don’t know where to start.

On Seth’s point, he was suggesting a counterfactual in which candidates don’t respond to heavily to the news cycle. I don’t have any idea how you’d evaluate it, but it’s far less hard to imagine or make judgments about than a circumstance in which candidates don’t campaign.

Comments on this entry are closed.

Previous post:

Next post: