Political Behavior Blog

« December 2005 | Main | February 2006 »

31 January 2006

Elections to the North

The Conservatives won the elections last week in Canada, although Stephen Harper will be leading a minority government. While the polls right before the campaign ended predicted a Conservative victory, there a number of interesting questions that the election raises, including the impact of the economy and implications for Duverger's Law.

First, on the economy. Canada has enjoyed a strong economy in recent years, and the Liberal Party naturally tried to take full credit for it. The Conservative victory may provide further evidence for the thesis that incumbents suffer more from bad economic times than they are rewarded in good times. This is probably especially true when there is a long period of prosperity (such as in the US in the 1990s), and voters may become somewhat complacent about the current state of affairs. On the other hand, apparently crime has been rising in Canada and of course the Liberal Party was mired in a corruption scandal. So if we broaden our conception of "retrospective voting" to cover issues besides the economy, the result may be somewhat less puzzling.

With respect to Duverger's Law, Canada has long had a multiparty system, despite having plurality rule. As Cox (1997) summarizing this literature points out, the common explanation is that the third party nationally (recently, the Bloc Quebecois) is concentrated in only one area. Hence, Duverger's Law by itself only implies local bi-partism, not necessarily national bi-partism, and in Quebec the race of late has usually been between the Bloc and the Liberals. Indeed, the Conservatives only mustered 8.8% of the vote in Quebec in 2004 (and didn't win a single seat), and only around 10% in 2000. But this time around, the Conservatives actually moved into second with 24%, and the Liberals with 20%, so it appears to have been a genuine 3-way race. No doubt the corruption scandals hurt the Liberals badly in Quebec, but it will be interesting to see if this is a one-time event or something more permanent. In addition, there seems to have been a number of 3-way contests out west in British Columbia between the NDP, the Liberals, and the Conservatives.

Posted by Ian Yohai at 10:03 PM

26 January 2006

PPBW Kickoff with McDermott

Also on Friday, the next semester of the Political Psychology and Behavior Workshop (PPBW) begins at Noon. Our first paper is "Cognitive Neuroscience and Politics" by Rose McDermott of UCSB. Here's an excerpt from the conclusion:

Earlier political science scholars may have ignored or neglected the impact of emotion on politics at least in part because it was so difficult to study and measure. As methods in the cognitive neuroscience improve, our understanding of human brain functioning improves in ways that make empirically grounded investigations more plausible and gratifying. Yet we are clearly at the very beginning of our understanding of the complex relationships between brain, behavior and emotion. Humility is an appropriate emotion in the face of this realization. However, recognition of such daunting knowledge should not detract or distract us from our goal of seeking greater understanding of the myriad ways in which emotion can influence political thought and action. Using more sophisticated and multiple measures in service of more specific theories can help advance our knowledge in important ways.

Posted by Barry Burden at 8:30 PM

Kansas Redux Redux

February 3 brings a CAPS talk by Larry Bartels of Princeton titled "What's the Matter with What's the Matter with Kansas?" Thomas Frank's WTMWK (whose subtitle is How Conservatives Won the Heart of America) was a hot selling book during the 2004 presidential campaign and resonated with people who thought the election turned on moral values. This view was dealt with in a couple of the papers presented at PPBW last semester with Ansolabehere et al. coming down on one side and Abramowitz on the other. It only got mild support at a conference I just attended at OSU on the 2004 election.

It would be helpful to clarify what the debate is. If the ongoing argument after the 2004 election is about whether social issues have trumped economic issues for many lower-income voters, then Bartels and Franks are on opposite sides. Whatever the truth, this is different from arguing that moral values played a role in the election. There is no doubt that religious conservatives contributed to Bush's victory. But even this is different from asserting that gay marriage initiatives were a key ingredient in the outcome.

Posted by Barry Burden at 8:00 PM | Comments (1)

24 January 2006

Sons, Daughters, and Ideology

A recent working paper by Oswald and Powdthavee argues that having daughters makes parents more left-wing. Using a decade worth of panel data, it seems that support for Labour in Britain is positively correlated with having daughters versus sons. This effect holds even after controlling for family size. Our political economy discussion group was fascinated by the idea that children could affect their parents' attitudes rather than the other way around. One of the concerns raised was that the "fertility stopping rules" might be different for liberals and conservatives if the latter values boys more. See Andy Gelman's discussion of this issue in his December 27th post (scroll down). An earlier paper by Dahl and Moretti discussed in our lunch argued that father prefer boys. As a result, there are more single mothers with girls than should be expected. We were also concerned that the analysis tosses out respondents without a party affiliation. In the U.S. most of the short-run change in party affiliation is in shift between independents and partisans rather than the more dramatic shift from Conservative to Labour.

Posted by Barry Burden at 1:32 PM | Comments (2)

20 January 2006

Krosnick on survey questions

The past two days have seen Jon Krosnick present his new book, The Handbook of Questionnaire Design: Insights from Social and Cognitive Psychology to a packed conference at Harvard. Krosnick's aim is an extraordinary one - to distil the thousands of studies that look at how individuals respond to survey questions and derive basic rules about how to design reliable and valid questions.

Much of the past two days has focused on the "mechanics" of survey questions - how many response options should we give respondents and how should we order them; should we use rating scales or ranking methods; should we offer a "don't know" response and so on. One of today's sessions stood out, however, for its lack of precise rules.

Jon discussed how to word questions - and basically offered two pieces of guidance. First, use clear, unambiguous language to ask directly about the item of interest. This seems straightforward to me. This fall, I wrote some questions for a poll with Steve Ansolabehere at MIT, and largely relied on my own instincts and the advice of colleagues to make sure the questions were unproblematic.

But Krosnick's second piece of advice was basically not to trust your own instincts, since survey respondents often interpret questions in ways that are not predictable when sitting in a university office! His answer to the problem is to pre-test questions, asking respondents to explain what they understood the question to be asking. The poll I worked on - and probably any poll that I will work on whilst a grad student - didn't have the resources to run this kind of pre-test. Presumably other researchers also have this problem: does anyone have a good solution to cheaply and conveniently check how ordinary respondents will react to the questions we ask?

Posted by Phil Jones at 8:31 PM | Comments (1)

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 2:54 PM | Comments (3)

11 January 2006

New Poll - Americans on Wiretapping

A new Washington Post poll found that 51 percent of Americans consider the interception of telephone and email communications without court approval an acceptable way for the government to investigate terrorism. Of course, when broken down by party lines, 75 percent of Republicans but only 37 percent of Democrats find the wiretapping acceptable.

A link to the poll results is available on the Washington Post homepage (www.washingtonpost.com).

Posted by Traci Burch at 5:09 PM | Comments (2)

5 January 2006

Abortion Politics Continue

In anticipation of the Senate confirmation hearings for judge Samuel Alito next week, yesterday All Things Considered devoted almost half an hour to the politics of abortion. It is remarkable to recall that it was initially Republicans such as then California Governor Ronald Reagan who pushed for more "liberal" abortion policies in the pre-Roe 1960s.

Even when Reagan was elected president, party identification was not a strong correlate of citizens' abortion attitudes. The (imperfect) 4-point NES abortion question shows that in 1980 30% of Republicans and 34% of Democrats were staunchly pro-choice ("always a personal choice"). In fact, there were more staunchly pro-life ("never permitted") Democrats (13%) than Republicans (8%). By 2000 29% of Republicans and a whopping 50% of Democrats were clearly pro-choice. Adams (1997 AJPS) has documented this "issue evolution" around abortion attitudes, as does Sanbonmatsu's 2002 book. Abramowitz (1995 JOP) argues that the 1992 presidential election was about abortion rights as much as the economy. Toying with the 2004 NES suggests that abortion attitudes affected vote choices even after accounting for party ID, views about the Iraq war, and demographics.

Posted by Barry Burden at 1:11 PM