« Should Law Students Take a Stats Course? | Main | Who Wins Patent Infringement Cases? »

22 May 2006


William Henderson

Jeff, that abstract seems to address what I am talking about. Thanks for posting it.

James presents a good way to frame the issue. Imagine a world where this equation holds true:

Votes/popularity = C + [Beta for substantive issues] + [Beta composition of district] + [Beta for negative campaigning] + [Beta for economy as applied to incumbents] + [Beta for leadership/sincerity/first principles]

We can't measure the last item, so political consultants use this equation:

Votes/popularity = C + [Beta for substantive issues] + [Beta composition of district] + [Beta for negative campaigning] + [Beta for economy as applied to incumbents] + E

I am arguing that under some circumstances, [Beta for leadership/sincerity/first principles], which is in the error term, can carry an election. But consultants overlook this fact because other factors (unrelated to sincerity) have a proven relationship to election outcomes.

Re Bill's point, I sweep new ideas, sincerity, and first principles together into a single leadership/sincerity construct. If a candidate thinks a single-payer health system is the best for the country, he or she makes that case to voters. If a candidate thinks torture is abhorrent and counterproductive, he or she says so, etc. In short, leaders sometimes ignore focus groups because of first principles; and voters like leaders. This may have been Kerry's biggest mistake.

So on the margins, Alan Keyes had some additional voter appeal because of his sincerity and first principles. Of course, Obama might have had similar appeal + substantive positions more in line with IL voters.


I was under the impression that you were trying to say that consultants were mis-specifying regressions in the following way...


where, if leadership/sincerity is in the error term, both belief1 and belief2 may be corellated with E. Therefore the consultants get biased/inconsistent results and scrap the idea of running a "belief's" based campaign all together, and favor a negative campaign instead. Of course, this problem could be fixed if there was (and Im not saying that there is) an instrument available. Let me just say that I have no idea if this is the case, it is just the impression I got from your post. I apologize if I misunderstood.

William Ford


There seem to be three distinct variables in your original post and subsequent comment, (1) new policy ideas, (2) sincerity, and (3) clearly articulated first principles. Do you see these as three different constructs? And if so, do you think all three are associated with additional votes? Why would a voter be more likely to vote for a candidate just because he or she has new ideas or articulates first principles? Sincerity, by contrast, seems more plausible to me. Since we cannot monitor everything a politician does, we might prefer candidates who sincerely hold their beliefs rather than candidates who pander. We might prefer a candidate who is disagreeable on one issue because we worry less that he or she will “sell us out” on other issues when no one is watching.

As a specific Illinois example, Alan Keyes should score well on at least two of these three items, i.e., sincerity and first principles. Would you predict that even though he lost by a wide margin, he still picked up some votes from people who otherwise disagreed with him, just because he was sincere or articulating first principles?


Jeff Yates

I'm not sure if the first principles test has been dealt with directly in pol sci. It opens up a number of related questions such as how does one discern if such principles are "clearly articulated" or "consistent". Also, is there a disconnect between rhetoric and policy making?

One way to gain some leverage on this issue is by assessing voter perceptions of party candidates' positions. The article (abstract) pasted below by Larry Bartels is a good example of this type of research.

Issue Voting Under Uncertainty: An Empirical Test

Larry M. Bartels

American Journal of Political Science, Vol. 30, No. 4. (Nov., 1986), pp. 709-728.

Enelow and Hinich (1981) proposed a simple model of issue voting for the case in which voters are uncertain about candidates' issue positions. This paper adds an equally simple model of survey responses for issue placement questions that makes it possible to estimate respondents' uncertainty indirectly from existing National Election Study (NES) data. An empirical test of Enelow and Hinich's model based on voting behavior in the 1980 presidential election indicates that voters dislike uncertainty, as suggested by the model. Indeed, uncertainty about candidates' issue positions appears to have been sufficiently pervasive and important to rival issue distances as a determinant of electoral choices.

Jeff Yates
Associate Professor
Department of Political Science
University of Georgia
SSRN page: http://ssrn.com/author=454290

William Henderson

Interesting comments. Re a proxy variable for sincerity/leadership, I am not sure how it would be constructed. It is a technical question with potentially valuable applications for modern politics.

But I am not sure what the endogeniety problem is and hence the need for a instrument. I am suggesting that it is possible to improve your political fortunes by clearly articulating first principles, regardless of focus group or polling results. There would be little correlation between short term political winds and first principles. Is the endogeniety problem the possibility that leadership would affect public opinion? Presumably that would be at time t + 1.

Re Jeff's point on negative campaigning, that was not the point I was trying make--admittedly, my facts are very stylized and simplistic.

Let me try again: Between the Republicans and the Democrats, the Republican may have more identifiable first principles; and presumably, some (or even most) Republican candidates sincerely hold those beliefs. Assuming similar Republican and Democratic dynamics and incentives for negative campaigning and the passing through the primary/election stages of the campaign, I am positing that clearly articulated first principles are associated with more votes.

Jeff Yates

I see part of this question being - does negative campaigning work? This is not my subfield, but it seems that this question isn't asked by pol scientists nearly as much as why presidential candidates decide to engage in negative campaigning (see Damore 2002; Haynes and Rhine 1998).

There's more work on the impact of negative campaigning in Senate races. For example, Lau and Pomper (2002) suggest that while it's a complicated calculation it seems to work for Senate challenger candidates, but that incumbants may be better with a positive campaign.

Re presidential campaigns we have to keep in mind that it's a step process - they have to win the nomination first and so, negative campaigning may unfold on two fronts, the nomination contest and the general election.

Jeff Yates
Associate Professor
Department of Political Science
University of Georgia
SSRN page: http://ssrn.com/author=454290


I haven't put much thought into it (though i am sure someone has) but why couldn't you use an instrumental variables approach to create a proxy for leadership/sincerty and get rid of the endogenaety problem?

The comments to this entry are closed.


June 2015

Sun Mon Tue Wed Thu Fri Sat
  1 2 3 4 5 6
7 8 9 10 11 12 13
14 15 16 17 18 19 20
21 22 23 24 25 26 27
28 29 30        

Site Meter

Creative Commons License

  • Creative Commons License
Blog powered by Typepad