28 November 2005
There was an article in The New York Times on Friday discussing the "demise" of the Conservative Party in New York State. New York is one of the few jurisdictions where candidates can receive votes on multiple ballot lines and the totals are simply added together. For example, George Pataki ran on the Republican and Conservative line in 1994, 1998, and 2002. Pataki averaged about 350,000 votes on the Conservative line in 1994 and 1998, but dropped to about 175,000 votes in 2002. The article mentions briefly that the Conservative Party was listed in the third spot on the ballot in 1994 and 1998, but dropped to fourth in 2002. Could the simple change in ballot position explain the drop-off in support?
Dan Ho and Kosuke Imai have a forthcoming article in the Journal of the American Statistical Association that shows that ballot position did make a difference for minor party candidates in the California Recall Election. California rotates the ballot position of the candidates in each electoral district, using a randomized alphabet lottery. This randomized feature allows Ho and Imai to make causal inferences about the effects of ballot position. In the recall election, minor parties did seem to gain more votes when listed on the first page of the ballot.
In the New York case, we are only dealing with a single-page ballot for governor. But in 1994 and 1998, the Republican Party was in the second slot on the ballot and the Conservative Party in the third, so Pataki was back-to-back on two lines. In 2002, however, the Republican Party was first and the Conservative Party fourth. Perhaps the greater degree of separation contributed to the lower voter totals for Conservatives, as maybe some voters didn't make it quite so far down the ballot.
The Times mentions two other possibilities. First, the Conservative Party has seen a modest decline in registration of about 20,000 voters recently. But the New York State Board of Elections reports that enrollment has been constant around 160,000-170,000 for several years, so it is unlikely that this is a big factor. The Conservatives were clearly doing much better than their party enrollment in 1994 and 1998. The second explanation is that Conservatives have become disenchanted with the party continuing to endorse a liberal Republican. Yet the same was true in 1998 (and 1994).
While the evidence is hardly conclusive, the article does raise some interesting questions about the role of "third" parties when they end up endorsing a major party candidate anyway.
This Friday the Political Psychology and Behavior Workshop features a paper by Dan Hopkins titled "The Diversity Discount: How Increasing Ethnic and Racial Diversity Dampens Support for Tax Increases." The paper shows that communities are less likely to be presented with public spending opportunities when they are diverse. This finding tweaks the simpler hypothesis that affluent majorities dislike diversity and vote against redistribution. Rather, elites appear to be driving (or at least co-piloting) the train by not offering proposals in diverse communities. Perhaps the abstract will entice you to read more:
Racial and ethnic diversity reduces U.S. municipalities' investment in public goods--or so recent research has argued. Yet this growing body of research suffers from key weaknesses: its theories are imprecise; its evidence only shows correlation, not causation; and it ignores the political processes that set levels of public investment. This paper addresses these deficiencies by studying the impact of changing racial and ethnic demographics on voter approval for property tax hikes in Massachusetts. Employing a variety of statistical approaches with time-series cross-sectional data, it contends that increasing diversity dampens municipalities' willingness to make long-term public investments. Departing from past literature, it concludes that increases in diversity, not its level, appear most influential. And the impact of those changes is marked: all else equal, if a
town diversifies, the probability that it holds a vote on a major capital project drops by 20%. Theoretically, these findings point us away from the dominant understanding of diversity as divergent preferences, and toward approaches that emphasize the disruption of existing political patterns and relationships.
Posted by Barry Burden at 10:26 AM
21 November 2005
When I was a graduate student my department interviewed a senior faculty member for a position in American politics. During a meeting with grad students the candidate asked us to describe our research projects. When I mentioned my interests, s/he responded that I should give up on political behavior since it is "dead field." I was shocked. That was a decade ago.
I suppose it depends on how one defines "political behavior." Recalling only the work done at Michigan in the 1960s is quite different from assessing the field as it stands today. I'm not sure how the story will end, but the last 10 years have been a great success for the field. The new work on heuristics, partisanship, campaign advertising, electoral systems, network effects, and information has kept the field vital. The growing acceptance of field and lab experimentation, and the rise of new survey methodologies like the Annenberg data, TESS, and Knowledge Networks have helped to keep the study of mass behavior fresh. The Elections, Public Opinion, and Voting Behavior (EPOVB) section is one of the largest APSA organized sections. The NES has new funding and purpose. There is always room for improvement. Dormant subfields such as socialization need to be revitalized. Survey technology will continue to evolve as land lines are replaced by cell phones and internet communication. But I am convinced that political behavior is far from a "dead field."
15 November 2005
Look for a new publication outlet called The Journal of Spurious Correlations. JSpurC will be launching soon and is soliciting manuscripts. Unlike the Journal of Irreproducible Results, this is a serious endeavor. It looks like the journal is seeking to publish negative results more than spurious ones, but either way it is an interesting development. The editors are experimenting with "triple blind" submission in which authors of negative findings may remain temporarily anoymous. (I suppose "quadruple blind" submission would prevent authors from even knowing about their own results!) The "file drawer problem" has long been a concern that JSpurC might help remedy. Gerber, Green, and Nickerson find evidence of publication bias in field studies of voter turnout but Sigelman's essay is skeptical. Is JSpurC going to solve this problem?
Posted by Barry Burden at 10:00 AM
14 November 2005
This week's meeting of PPBW features a paper by Harvard grad student Samuel Abrams titled "Interests, Parties, and Social Embeddedness: Why Rational People Vote." The paper offers a provocative argument: "When people vote and read about politics it is because this is sometimes a way get respect from other people in their social networks. Politicians understand this and try to influence what groups discuss and find important." Here's a longer summary of the findings from the paper's conclusion.
Because it is difficult to understand voting in standard rational choice models as investments in desirable outcomes, participation is often seen as quasi-irrational. The same is true of political knowledge acquisition. Voting becomes an act that is poorly linked to the real interests of groups, and that is easily manipulated by elites. By contrast, we have argued in this paper that voting and political knowledge are in fact safely anchored in group interests and can indeed be understood as an investment in desirable outcomes. As implied by network theory, the objective for individuals is not to influence the result of an election, but to maintain and improve their standing in the networks and communities to which they belong. Being knowledgeable about group interests, and being prepared to act in the interest of the group, are key ingredients in establishing such standing.
10 November 2005
What Americans think of their leaders is at times simple and at other times complex. The story of the Bush presidency so far has been more the former than the latter. George W. Bush the person is viewed just about as positively as George W. Bush the president.
Contrast that with our last two Democratic presidents. Both Bill Clinton and Jimmy Carter showed substantial divergence between the job approval and personal favorability ratings. Clinton had high approval ratings in his second term but was not well-liked as an individual after the Lewinsky affair. Jimmy Carter was well-liked, but his performance in office was not. Surveys have shown that the candidate who more respondents report wanting to "have a beer with" tends to win, yet academic focus has been on job approval.
4 November 2005
Today, the Political Psychology and Behavior Workshop co-sponsored a talk by Prof. John Aldrich of Duke University. Prof. Aldrich discussed "Party and Constituency in the House" - a project that seeks to reconcile the research traditions focusing on parties in government and parties in the electorate by examining these two dimensions of party politics together. Aldrich and his co-authors have put together an impressive data set - running from 1983 to 2002 - that includes 86 variables measuring a variety of socio-demographic characteristics for each of the 435 congressional districts. In his broad overview of the data, Prof. Aldrich discussed how such district characteristics might help explain voter preferences, predict winners and losers in House races, and in turn illuminate the behavior of elected representatives.
Posted by Federico Ferrara at 3:55 PM
3 November 2005
In the PPBW workshop last Friday, Tammy Frisby presented her paper on how terms limits affect the strategic entry of challengers to state legislative races. Frisby analyzes state legislative races from 1996 – 2004 in California, Colorado, Florida, and Ohio, and finds that challengers are just as likely to run against a last term incumbent as a non-last term incumbent. Additionally, the quality of these challengers who face last-term incumbents does not seem to be lower. The discussion centered on a few themes; first, a discussion of the utility function used to describe the strategic choice of potential challengers to enter a race. The basic utility function strategic politicians are assumed to use, developed by Gordon Black and popularized by Jacobson & Kernell, states that the utility is the result of the product of the probability of winning the office (P) times the value of the office (B), less the risk (R) involved in waging a campaign. To improve on this model, the audience suggested indexing terms by i for each individual candidate and by j for the district. Additionally, there was discussion as to how the values of all the variables may be different for state legislative candidates (as opposed to federal office candidates).
There was debate over whether the effects observed were due to the interaction between term limits and strategic political behavior, or whether the effects could be attributed to one of these explanations more so than the other. Lastly, some time was spent making the distinction between strategic politicians and savvy politicians; the former being those candidates who employ cost benefit analysis to inform their decision regarding entry to the race, the latter group of savvy politicians describes those who have better estimates of the variables to plug into the utility function.
Posted by Sarah Sled at 7:43 PM