« New Poll - Americans on Wiretapping | Main | Krosnick on survey questions »

13 January 2006

Congress and corruption, midterm edition

Journalists and academics often offer contradictory explanations of mass political behavior, and different predictions for what's likely to happen next. As an example, pundits’ accounts of the 2004 election being decided by “moral values� voters have come under close scrutiny by political scientists at PPBW and elsewhere. Over the holidays, I noticed this CNN story about the 2006 midterms which also seemed to clash with what political scientists have long known about congressional elections.

The CNN poll suggests that half of the public believe most Members of Congress are corrupt, and around 55% claim that corruption will be "the most important" or a "very important" issue that they weigh when deciding who to vote for in November. The story uses these results to suggest that “Congress' image could emerge as an election topic� and impact the outcome of the midterms. I don't dispute the methodology of the poll or that respondents were likely to have answered the questions in these ways. So why am I skeptical of their conclusions?

The CNN story doesn’t link to the full poll results or questions asked, so we have little way of knowing how closely the public has been following the unfolding scandals in Washington (for a similar line of thought, see Casey's comment on Traci's post yesterday). We likewise don't know how respondents weighted corruption relative to other issues, or whether they were primed by a series of questions on the subject to say it would be an important factor in their vote choice.

But I think the real reason to question CNN's conclusion that these results can be used to predict voters' behavior in the midterms comes from two long-standing political science research findings.

First, the public does good job of distinguishing between their Member of Congress (who they generally like) and the Congress as a collective institution (which they generally dislike). Second, as Gary Jacobson has tirelessly pointed out, congressional elections pivot around local issues and local candidates: voters cast ballots for their Representative or her opponent on the basis of their appeals, not for or against the institution as a whole.

A similar USA Today poll to the CNN one shows that the first research finding continues to hold in 2006. 69% say that most Members of Congress are out of touch with average Americans, while 38% say the same about their Member of Congress; 54% think most Representatives are focused on the needs of special interests, while 40% say the same about their Representative. And unsurprisingly, the same pattern holds for corruption: only 22% think their Member of Congress is corrupt, compared to 38% who think most Members are corrupt.

Which is where the second research finding comes in. Voters' perceptions of corruption in Congress are likely to affect the midterm results only if more voters start to view their own Member of Congress as corrupt. And whether that happens is dependent on local challengers' abilities to link the incumbent to ongoing scandals in Washington. As much as I wish it could, examining the public's views towards Congress as a collective body can't tell us whether local candidates will effectively link voters' opinions of incumbent Representatives with their opinions towards Congress. I guess we'll just have to wait and see how this one plays out.

Posted by Phil Jones at January 13, 2006 2:54 PM

Comments

Political image does have a role to play come November, but a larger role is the economic situation. If people are in a good economical situation they will normally stay the course with the incumbent.

Posted by: Jack Foreman at January 18, 2006 4:09 PM

The 1992 elections resulted in a lot of turnover because the House Bank scandal broke in time for strong challengers to mount campaigns against tainted incumbents. The Abramoff scandal so far has only affected a few leaders publicly. Concerns about corruption will only have large effects on the House elections if the other tainted members are outed in time for challengers to emerge.

Posted by: Barry Burden at January 19, 2006 3:49 PM

Barry's point is a good one.

Recruitment is key. However, corruption is only one issue that depresses the optimism potential candidates see. Keep in mind that the Admin’s approval had been bogged down for multiple reasons (e.g., war, gas prices, hurricane response, corruption).

I think we are beginning to see that the Admin. is having difficulty recruiting top candidates.

Florida is a perfect case where Rove has apparently been desperately and unsuccessfully trying to find an alternative to Katherine Harris in her bid to run against a vulnerable Bill Nelson.

Another factor might be the fact that they have been recruiting candidates for 3 cycles now and the obviously strong candidates (e.g., Norm Coleman, Mel Martinez) have already run/won.

Posted by: Ben Bishin at January 20, 2006 1:43 PM