2010-2011 Fall

Rachel Stein, Graduate Student, Political Science

September 20, 2010

War and the Culture of Honor: Southern Support for the Use of Military Force

ABSTRACT: Anecdotal evidence has long been used to suggest that Southerners have a greater enthusiasm for war than people from other parts of the country. Numerous studies have documented a Southern proclivity for violence as evidenced by higher homicide rates, higher rates of execution of death row prisoners, less strict laws regarding domestic violence and killing in self-defense, and greater public support for the death penalty, gun ownership, corporal punishment, and the use of violence by police to quell social unrest. However, the question of whether this Southern “culture of violence” extends into the realm of foreign policy has remained largely unexplored. In this paper, I develop the argument that individuals born and raised in the South tend to hold a particular set of core values regarding the use of force that dispose them to approve of violence in both the domestic and international contexts, and test it using available public opinion data, such as the General Social Survey. This study will not only bring empirical evidence to bear on an old conventional wisdom about Southern enthusiasm for war, but it will also demonstrate the value of an approach to the study of mass public support for war that breaks down the barrier between the study of domestic policy opinion and the study of foreign policy opinion. The core values that inform citizens’ approval of the use of force in domestic life do not cease to matter at the water’s edge.

Rachel Stein
Josh Pasek, Professor of Communication Studies, University of Michigan

September 27, 2010

Determinants of turnout and candidate choice in the 2008 U.S. Presidential election

ABSTRACT: The presence of an African-American candidate on the ballot running for President in 2008 raises the possibility that the election outcome might have been influenced by anti-African-American racism among voters. This paper uses data from the Associated Press-Yahoo! News-Stanford University survey to explore this possibility, using measures of both explicit racism (symbolic racism) and implicit racism (the Affect Misattribution Procedure). The parameters of multinomial logistic regression equations were estimated to test the hypotheses that racism might have behaved differently on election day than they would have had racism been eliminated. The findings suggest that racism’s impact on the election outcome could have been substantial, by causing (1) people who would otherwise have voted for Obama to vote for McCain, for a nonmajor party candidate, or not to vote at all, (2) people who would not have voted to vote for McCain instead, and (3) people who would have voted for a nonmajor party candidate to vote for McCain instead.

Josh Pasek
Rebecca Weiss, PhD Candidate, Communication

October 4, 2010

Optimally Aggregating Elicited Expertise: Applications of the Bayesian Truth Serum

ABSTRACT: Subjective decision-making is essential to creating good policies; when a quantitative analysis cannot be performed in a timely fashion, we rely upon the judgments of experts. These judgments tend to lack external criteria for objective truth. Thus, it is not easy to evaluate these judgments in a way that accurately reflects both true and truthful opinion. Simple measures like the averaged judgment, or majority rule, can be inadequate since such information does not account for variations in the quality of opinions. This may occur when some judgments are unpopular or unusual. The “Bayesian Truth Serum” or BTS (Prelec, 2004; Prelec and Weaver, 2007; Prelec and Seung, 2007) is a scoring method that identifies judgments that possess the highest probability of being objectively true, provided that the participants are rational Bayesian decision-makers and that a single right answer exists. Essentially, BTS is a method that optimally aggregates opinion, even when the correct answers are anonymously dispersed in a crowd. Additionally, BTS may also overcome cognitive biases, where respondents ignore personal opinions in favor of a socially desirable answer (Prelec and Weaver, 2007). In this talk, I will introduce BTS as a scoring algorithm for survey instruments. I will review a series of studies that were performed to test the ability of BTS in highlighting true and truthful answers, using chess as the domain of expertise. Additionally, I will briefly introduce an interactive survey employing BTS scoring, since BTS has not been thoroughly tested in iterative domains where respondents can view previous scores of the same respondents.”

Rebecca Weiss
Ali Valenzuela, Professor of American Politics, Princeton University

October 11, 2010

Integrating Institutions: Political Determinants of American Latino Identity

ABSTRACT: Over successive generations, most white ethnics shed their group identity and became politically indistinguishable from other whites. In contrast, among African Americans, increasing diversity has not led to a decline in racial group behavior. These divergent processes of social identification suggest distinct determinants of identity among American subgroups, yet research rarely inquires about the institutional and political sources of group identification. This paper explores the determinants of linked fate, pan-ethnic, American and partisan identification among U.S. Latinos and tests the influence of institutions—electoral context, religious observance, media exposure and demographic patterns—on such identification. The point of departure, that social identities are malleable, is consistent with findings in social psychology and comparative politics that show identities can be manipulated to vary in their salience under different conditions. Whether and how this process occurs for Latinos in American politics has not been studied empirically. I review the literatures in American and comparative politics on ethnic identity formation and social context, and generate a social transmission theory of Latino identity choice. I test these ideas using data from the 2006 Latino National Survey (LNS), a representative study of more than 8,600 Latinos, and exploit respondent county information to create a new dataset with survey responses, Census population estimates, county electoral data and measures of news media coverage of the 2006 immigrant protest marches. Using these comprehensive data, I estimate models of social identification among Latinos in the U.S. I find that electoral competition is a key variable in processes of Latino identification, dramatically mediating the influence of demographic context, religious observance and media exposure on Latino identification. This supports a view of identity choice processes contingent on elite mobilization and political information flow to individual Latinos. A focus on electoral context helps clarify the role of elite actors and social influence on processes of politicized Latino identification with pan-ethnic, American or partisan categories. These and other findings situate Latinos within the larger panoply of racial and ethnic groups in American politics.

Ali Valenzuela
Tess Heintz, Researcher, GfK

October 18, 2010

The Time Taken to Earn a Bachelor’s Degree: Are Students Taking Longer Now Than They Have in the Past?

ABSTRACT: We use data from the Survey of Income and Program Participation (SIPP) to explore whether the time it took students to earn their bachelor’s degrees increased over the time period 1925 to 2004. We found that the time to degree did increase over this time period with students taking about one and a half weeks to two and a half weeks longer per year, depending on the model used. African American students had a time to degree that was loger than that of White students and which increased over time at a higher rate. We included various robustness tests to account for definitions of time to degree, the overlapping data structure of the SIPP and for measurement error that may occur because of our reliance on respondents having to recall their past event of graduation.

David Yeager, Assistant Professor of Psychology, University of Texas, Austin

November 1, 2010

Psychological interventions in education: Inquiry at the intersection of psychological theory and educational policy

ABSTRACT: In the past decade, the lion’s share of attention and resources in education research have been shifted toward large-scale evaluations of structural changes to the educational environment using randomized controlled trials. Yet increasingly these studies have yielded disappointing or null effects, even when programs were well-implemented. When social scientists interpret such findings, it is tempting to conclude either that interventions are too small in scope (even when they are large in sample size) to have an effect beyond the term of the intervention, or that the behavior under investigation is not changeable. In this talk I discuss psychological interventions as a supplementary approach to traditional educational reform. I propose that sustained changes in educationally-relevant behavior can be made, and that relatively brief and time-limited psychological messages can be used to do so. I examine this in the context of interventions to reduce aggression and depression in high schools. In various laboratory studies I develop a new theory of the causes of reactions to victimization in school, and using field experiments I test the effectiveness of this theory. Implications for basic theories of aggression and of educational research will be discussed.

Josh Pasek, Professor of Communication Studies, University of Michigan

November 29, 2010

Bias in the Voting Booth: The Case of Ballot Order

ABSTRACT: It’s not enough to get voters to the polls. Democracy also requires that citizens make “informed” decisions in the voting booth. Despite millions of dollars spent on campaigns, a sizable number of people remain truly undecided when they must choose among candidates. And their decisions influence elections. The heuristics undecided voters use range from picking candidates based on gender, to choosing whichever candidate has the coolest last name. One particularly pernicious cue that voters use is the order of candidates’ names on the ballot. Not all heuristics are created equal. Ballot order effects introduce a systematic and arbitrary bias into election results. This talk presents the most comprehensive analysis of ballot order effects to date – showing that the order of candidate names on the ballot induces a persistent bias that holds for all types of candidates, across varying contests in different states, and using different statistical methods. After exploring variation in ballot order effects across electoral contexts, I nest this research into a broader framework for examining how sources of information feed into the political decision-making process.

Josh Pasek