Providing corrective information can reduce factual misperceptions among the public but it tends to have little effect on people’s underlying attitudes. Our study examines how the impact of misinformation corrections is moderated by media choice. While the news content is held constant, our treatment manipulates whether participants are allowed to freely choose a media outlet or are randomly assigned. Our results demonstrate the importance of people’s ability to choose: While factual misperceptions are easily corrected regardless of how people gained access to information, subsequent opinion change is conditional on people’s prior willingness to seek out alternative sources.
The following publications are some of the recent research by our members. More about the research activities of our members can be found on their individual websites.
Gender differences in political knowledge are a well-known empirical finding in public opinion research. Scholars working in this area have proposed various explanations for this phenomenon, often focusing on issues regarding the format and content of factual knowledge batteries. Yet, there are surprisingly few works that focus on how scholars might diversify the content of political knowledge measures to develop items that are less biased toward male areas of expertise. In this paper, we propose an inductive framework to develop more gender-balanced knowledge batteries by including political issues that are of particular relevance to women and women’s lives.
Grant and Lebo (2016) and Keele et al. (2016) clarify the conditions under which the popular general error correction model (GECM) can be used and interpreted easily: In a bivariate GECM the data must be integrated in order to rely on the error correction coefficient, \( \alpha^{*}_{1} \), to test cointegration and measure the rate of error correction between a single exogenous x and a dependent variable, y. Here we demonstrate that even if the data are all integrated, the test on \( \alpha^{*}_{1} \) is misunderstood when there is more than a single independent variable. The null hypothesis is that there is no cointegration between y and any x but the correct alternative hypothesis is that y is cointegrated with at least one—but not necessarily more than one—of the x's. A significant can occur when some \( I(1) \) regressors are not cointegrated and the equation is not balanced. Thus, the correct limiting distributions of the right-hand-side long-run coefficients may be unknown. We use simulations to demonstrate the problem and then discuss implications for applied examples.
The question of how descriptive representation might affect political behaviour and attitudes is important when considering the role political attitudes play in facilitating a functioning democracy. What role, if any, does co-racial descriptive representation play in the relationship between citizens and local government? What factors underlie attitudes toward local government, generally? Employing a unique set of survey data collected across several dozen cities combined with city-level contextual data, the analysis offers a comprehensive picture of trust toward local governments. Overall, the findings support the hypothesis that descriptive representation has a positive effect on feelings of trust in local government. However, these effects are limited to mayoral representation. Increased levels of descriptive representation in less-visible city councils do not have the same effect.
There is reason to believe that an increasing proportion of the news consumers receive is not from news producers directly but is recirculated through social network sites and email by ordinary citizens. This may produce some fundamental changes in the information environment, but the data to examine this possibility have thus far been relatively limited. In the current paper, we examine the changing information environment by leveraging a body of data on the frequency of (a) views, and recirculations through (b) Twitter, (c) Facebook, and (d) email of New York Times stories. We expect that the distribution of sentiment (positive-negative) in news stories will shift in a positive direction as we move from (a) to (d), based in large part on the literatures on self-presentation and imagined audiences. Our findings support this expectation and have important implications for the information contexts increasingly shaping public opinion.
Competitive, vigorous campaigns have been shown to increase participation across a variety of elections, including those at the state and local level. Building on previous work that examines the impact of money in judicial elections, this study explores the impact of campaign effort on participation in state court elections. Using data from 260 state supreme court elections occurring from 1990-2004 across 18 states, I find that competitive campaigns—not just expensive ones—are important for encouraging participation in these contests. Additionally, the study uncovers differential effects of challenger and incumbent spending. Ultimately, the findings contribute to our understanding of campaign effects in judicial elections while also providing an additional test of the idea that campaigns matter, especially in low-information contests.
We investigate the effect of personality on prosocial behavior in a Bayesian multilevel meta-analysis (MLMA) of 15 published, interdisciplinary experimental studies. With data from the 15 studies constituting nearly 2500 individual observations, we find that the Big Five traits of Agreeableness and Openness are significantly and positively associated with prosocial behavior, while none of the other three traits are. These results are robust to a number of different model specifications and operationalizations of prosociality, and they greatly clarify the contradictory findings in the literature on the relationship between personality and prosocial behavior. Though previous research has indicated that incentivized experiments result in reduced prosocial behavior, we find no evidence that monetary incentivization of participants affects prosocial tendencies. By leveraging individual observations from multiple studies and explicitly modeling the multilevel structure of the data, MLMA permits the simultaneous estimation of study- and individual-level effects. The Bayesian approach allows us to estimate study-level effects in an unbiased and efficient manner, even with a relatively small number of studies. We conclude by discussing the limitations of our study and the advantages and disadvantages of the MLMA method.
This study explores whether and how individuals evoke moral considerations when discussing their political beliefs. By analyzing open-ended responses in the 2012 American National Election Study using a previously validated dictionary, I find systematic ideological differences in moral reasoning, even when respondents are not explicitly asked about morality. The study proceeds to show that the reliance on moral considerations in attitude expression is amplified by the moral content of individual media environments.