About the Society
Papers, Posters, Syllabi
Submit an Item
Polmeth Mailing List
Below results based on the '2010' year search
Total number of records returned: 13
Inferring Strategic Voting
We estimate a model of strategic voting and quantify the impact it has on election outcomes. Because the model exhibits multiplicity of outcomes, we adopt a set estimator. Using Japanese general-election data, we find a large fraction [75.3%, 80.3%] of strategic voters, only a small fraction [2.4%, 5.5%] of whom voted for a candidate other than the one they most preferred (misaligned voting). Existing empirical literature has not distinguished between the two, estimating misaligned voting instead of strategic voting. Accordingly, while our estimate of strategic voting is high, our estimate of misaligned voting is comparable to previous studies.
A New Solution to Ecological Inference
Truncated Normal Distribution
This poster presents a new solution to ecological inference (the example of southern vote registration, King, 1997). It solves the ecological regression directly with the maximum likelihood method and a mathematical solution to the bounded Gaussian integral. Assuming the proportion of black voters Xi and the overall voter's turnout Ti normally distributed between (0,1), the error distribution can be mathematically deduced from tomography lines as normally distributed with a doubly truncation , where Bb and Bw (the percentage of black's and white's turnout) is fixed. Given the analytical form of the error distribution has been known, we can directly solve the regression through MLE and derive an analytical solution to the Ecological Inference problem.
Beyond Voting: A Generalized Model of Political Participation
latent class analysis
finite mixture modeling
random utility models
For decades political scientists have studied the motives underlying individual engagement in political activities, but the field still lacks a comprehensive study taking into account the commonalities between different forms of political engagement. It is likely that some of the factors that affect the decision to vote also affect the choice of working for a campaign or donating money, as well as engagement in other political endeavors. In this paper I construct a model that allows measuring how observed individual attributes influence decisions to participate specific activities, as well as identifying common patterns of behavior across different form of activism. Most importantly, since empirical models cannot possibly account for all factors affecting participation decisions, and since other factors are likely to follow heterogeneous distributions across the population, I develop a generalized specification where respondents are classified into classes with lower or greater propensities toward civic voluntarism depending on the distribution of unobserved attributes. If underlying heterogeneities have considerable effects on propensities toward participation, learning about these heterogeneities is important for gaining a better understanding of how well activists and elected candidates represent diverse constituencies, and is also useful for the design of mobilization campaigns. I applied this model to survey data collected during the 2008 electoral period and found that while a relatively small proportion of individuals exhibited high propensity toward participation and were fairly sensitive to variations in variables such as education attainment, most respondents were assigned to a group with low tendencies toward activism and low sensitivity to changes in observable attributes.
Do you feel Angry? Are you sure? Testing the Reliability of Overt Emotional Cues and the Effects of Semantic Self Reports in Experimental Research
In the experimental study of political affect two assumptions are implicit. First, scholars assume that the affective state intended by the treatment is actually invoked. Second, scholars assume that semantic prompts such as, “Has (Barack Obama/John McCain) -- because of the kind of person he is, or because of something he has done -- ever made you feel: (insert word from feeling scale),” provide an accurate reliability check on the former assumption. However, work in psychology demonstrates that the use of semantic self reports is unreliable because participants do poorly at accurately reporting experienced emotion (Breckler 1984; Shacter and Singer 1962; Weber et al. 2007). If the presence of the prompt introduces error into the model or participants do not reliably recall their affective state then the use of semantic affective prompts is problematic. I ask: Q1: Is the semantic affective prompt an effective check on the reliability of an emotional cue? Additionally, I examine the use of overt anger cues versus subliminal anger cues in eliciting anger. Though most scholars use semantic self-reports as a direct test that the emotion of interest was elicited, others use subliminal primes to elicit emotional states outside of awareness (Bargh 1997). I ask: Q2: Is there a significant difference between models that invoke emotion overtly versus subliminally? I utilize a unique research design to tease out the effects of interest. To do so I set up treatment conditions which vary in the way the affective state is invoked (overtly/subliminally) and in the presence or absence of a semantic affective prompt. If find that challenges to the use of the semantic affective prompt are warranted: there is a mean difference in the responses of participants assigned to the semantic affective prompt condition and participants assigned to the no affective prompt condition.
Economic Voting: Causal Mediation of Retrospective Evaluations
In this paper, we show that an increase in economic growth has a positive effect on the share of voters who support the party of the chief executive and that it does this through retrospective evaluations of the economy. In order to do this, we expand on the results of Duch and Stevenson (2005, 2008). Using causal mediation analysis, we show that an increase in economic growth leads to an increase in the number of survey respondents whose retrospective evaluations of the economy are positive. This, in turn, leads to an increase in the number of voters who support the party of the chief executive. A similar result holds using annual unemployment change as the treatment. In both cases, the effect is weaker when the chief executive is a member of a coalition. The evidence for existence of mediation effects is robust to the inclusion or exclusion of a number of control variables, including an interaction between individual ideology and government ideology.
Effects of Interviewer Gender and Hijab on Gender-Related Survey Responses: Findings from a Nationally-Representative Field Experiment in Morocco
Gender of Interviewer Effects
Interviewer Religious Dress
Despite the recent expansion of surveying in the Muslim world, few published studies have addressed methodological questions, including how observable interviewer characteristics affect responses and data quality. Although there are a limited number of studies on interviewer dress effects, none examine interviewer gender. This study asks whether and why gender and religious dress affect responses to gender-related questions. Drawing upon original data from a nationally-representative, partially-randomized survey of 800 Moroccans conducted in 2007, the study finds strong evidence that gender and dress affect responses and item non-response. The paper argues that because hijab implies multiple personal, religious, and political dimensions of identity nested within gender identity, interviewer gender and dress must be considered as intersecting categories. For questions pertaining to women’s role in the public sphere, responses were affected by interviewer dress; respondents reported more progressive attitudes and were more likely to refuse to respond to female interviewers not wearing hijab than to veiled female interviewers and male interviewers. For support for gender equality in family law, results were affected by interviewer gender; respondents reported more liberal views and were more likely to fail to respond to female interviewers with both dress styles than male interviewers. Interviewer characteristics affected responses to more than half of the 174 questions included in the survey, including support for democracy and religiosity. Researchers conducting surveys should code and control for interviewer characteristics in order to reduce total survey error and better understand the social processes which generate public opinion in this crucial region.
A Comparison of Instrumental Variable Estimators in Models of Discrete Choice
Quiroz Flores, Alejandro
continuous endogenous regressors
Comparison of three instrumental variable estimators applicable to probit models. The first estimator uses conditional probabilities and MLE. The second estimator uses Newey’s two-step minimum chi-squared estimator. A new estimator presented here uses GMM to approach probit models as non-linear regression. These models are compared in a simulation experiment. Results show that conditional probability MLE model has superior performance both in terms of bias and efficiency, although the GMM estimator follows closely.
Robust Estimation of the Cox Proportional Hazards Model
Event History Modeling
Cox Proportional Hazards Model
Partial Likelihood Maximization
Iteratively-Reweighted Robust Estimation
The Cox proportional hazards model is often used with time-to-event data in political science. However, misspecification issues such as measurement error or omitted covariates can cause substantial coefficient bias when it is estimated via the conventional Partial Likelihood Maximization (PLM). Here we review an iteratively-reweighted robust (IRR) estimator of the Cox model that is proven to reduce this bias under such conditions and propose a cross-validated median fit (CVMF) test to select between PLM and IRR. Then we apply the test to data in political science. We consider several typologies of replications with respect to (1) the test's selection (PLM or IRR) and (2) the implications of IRR for the original hypotheses (less support, more support, or mixed results). Overall, we demonstrate that PLM and IRR can each be optimal, that substantive conclusions can depend on which one is used, and that the CVMF test is effective in choosing between them.
Non-Parametric Treatment Effect Estimation Strategy for Missing Treatment Data
Classic Treatment Effect
What should scholars do when faced with missing treatment data in randomized experiments or observational studies? Rather than, for example, assuming the treatment data is missing at random, Molinari (2010) introduces a non-parametric approach for computing bounds on treatment effects when there are missing treatment data. I review the Molinari approach and then use it to address an important question in international relations: is it true that ``issue linkage'' (the simultaneous negotiation of multiple issues for joint settlement) helps states conclude otherwise unattainable negotiated agreements?
A Bottom-Up Approach to Linguistic Persuasion in Advertising
This paper presents a new, bottom-up approach to understanding how the linguistic contents of television advertisements determine their persuasive effects. Rather than categorizing by topic or style the hundreds of ads run during a presidential election campaign, and then working with a few ad categories as independent variables, this approach instead estimates an effect for each unique ad and then draws inferences about the relationship between the contents of those ads and their effects using automated text analysis. Specifically, each unique ad run in 2004 is assigned its own variable (counting broadcasts per region per time), and since regressing (survey-measured) vote intention on all the ad variables at once would result in severe over-fitting, instead each ad variable is regressed separately, and vote intention is predicted by averaging the individual predictions. This approach is validated through extensive out-of-sample testing of predicted versus measured vote intention, and the collective effect of all ads is shown to be significant and largely pro-Democrat in 2004. To understand what in the content of those ads determines their varying effects, a variety of automated text analyses (eg, k-nearest-neighbor, distance weighting, and Bayesian) are adapted to predict ad effects based on the textual similarity between new ads and previously measured ads, and the technique is again validated through out-of-sample testing. The most effective pro-Democrat and pro-Republican words are established using a one-dimensional scaling of words in the ad-effect space, with suggestive differences between the two word sets. This new approach is of substantive interest for the insights the text analysis provides about why ads differ in their effects, and it is of methodological interest in grounding automated text analysis in a framework that allows real-world prediction and out-of-sample testing. The tools developed could also be of practical use in shaping future advertising campaigns.
Optimally Selecting Matched Samples
We apply a new and simple graphical method (the ``space graph"; Iacus, King, and Porro, 2010) for evaluating many matched samples and selecting the best one(s). We then use this technique to reveal patterns in the relative performance of matching methods across data sets. We also identify an important and previously unnoticed problem that causes propensity score matching with calipers to fail in precisely the applications for which it was designed.
Weighted Estimation for Analyses with Missing Data
inverse probability weighting
Missing data plague data analyses in political science. The recent applied statistics literature reflects renewed interest in weighting methods for missing data problems. Three properties are stressed in this literature: (i) robustness, (ii) the ability to use post-treatment information in causal analysis, and (iii) methods to gain efficiency. I present these results, hoping to show the potential in using refashioned weighting methods for political science research.
Parties, Pivots, and Policy: The Status Quo Test
This study applies a novel technique that measures policy status quo locations in relation to legislators’ preferences. The resulting status quo estimates allow for the first direct test of the policy consequences predicted by the pivotal politics and party cartel theories of legislative politics. The empirical tests indicate that parties interact with pivotal politics to contribute to policy gridlock and shape policy change. By bringing pressure to bear upon pivotal politics pivots and by blocking policy changes that would ‘roll’ the party, parties increase the range of policies subject to gridlock in the American political system.