Suchergebnisse

AbstractDoes the internet facilitate selective exposure to politically congenial content? To answer this question, I introduce and validate large‐N behavioral data on Americans' online media consumption in both 2015 and 2016. I then construct a simple measure of media diet slant and use machine classification to identify individual articles related to news about politics. I find that most people across the political spectrum have relatively moderate media diets, about a quarter of which consist of mainstream news websites and portals. Quantifying the similarity of Democrats' and Republicans' media diets, I find nearly 65% overlap in the two groups' distributions in 2015 and roughly 50% in 2016. An exception to this picture is a small group of partisans who drive a disproportionate amount of traffic to ideologically slanted websites. If online "echo chambers" exist, they are a reality for relatively few people who may nonetheless exert disproportionate influence and visibility.

Self-reported measures of media exposure are plagued with error and questions about validity. Since they are essential to studying media effects, a substantial literature has explored the shortcomings of these measures, tested proxies, and proposed refinements. But lacking an objective baseline, such investigations can only make relative comparisons. By focusing specifically on recent Internet activity stored by Web browsers, this article's methodology captures individuals' actual consumption of political media. Using experiments embedded within an online survey, I test three different measures of media exposure and compare them to the actual exposure. I find that open-ended survey prompts reduce overreporting and generate an accurate picture of the overall audience for online news. I also show that they predict news recall at least as well as general knowledge. Together, these results demonstrate that some ways of asking questions about media use are better than others. I conclude with a discussion of survey-based exposure measures for online political information and the applicability of this article's direct method of exposure measurement for future studies.

AbstractDigital literacy is receiving increased scholarly attention as a potential explanatory factor in the spread of misinformation and other online pathologies. As a concept, however, it remains surprisingly elusive, with little consensus on definitions or measures. We provide a digital literacy framework for political scientists and test survey items to measure it with an application to online information retrieval tasks. There exists substantial variation in levels of digital literacy in the population, which we show is correlated with age and could confound observed relationships. However, this is obscured by researchers' reliance on online convenience samples that select for people with computer and internet skills. We discuss the implications of these measurement and sample selection considerations for effect heterogeneity in studies of online political behavior. We argue that there is no universally applicable formula for selecting a given non-probability sample or operationalization of the concept of digital literacy; instead, we conclude, researchers should make theoretically informed arguments about how they select both sample and measure.

AbstractDiscussions around declining trust in the US media can be vague about its effects. One classic answer comes from the persuasion literature, in which source credibility plays a key role. However, existing research almost universally takes credibility as a given. To overcome the potentially severe confounding that can result from this, we create a hypothetical news outlet and manipulate to what extent it is portrayed as credible. We then randomly assign subjects to read op-eds attributed to the source. Our credibility treatments are strong, increasing trust in our mock source until up to 10 days later. We find some evidence that the resulting higher perceived credibility boosts the persuasiveness of arguments about more partisan topics (but not for a less politicized issue). Though our findings are mixed, we argue that this experimental approach can fruitfully enhance our understanding of the interplay between source trust and opinion change over sustained periods.

We introduce the rationale for a new peer-reviewed scholarly journal, theJournal of Quantitative Description: Digital Media. The journal is intendedto create a new venue for research on digital media and address severaldeficiencies in the current social science publishing landscape. First,descriptive research is undersupplied and undervalued. Second, researchquestions too often only reflect dominant theories and received wisdom.Third, journals are constrained by unnecessary boundaries defined bydiscipline, geography, and length. Fourth, peer review is inefficient andunnecessarily burdensome for both referees and authors. We outline thejournal's scope and structure, which is open access, fee-free and relies on aLetter of Inquiry (LOI) model. Quantitative description can appeal to socialscientists of all stripes and is a crucial methodology for understanding thecontinuing evolution of digital media and its relationship to importantquestions of interest to social scientists.

Though commentators frequently warn about "echo chambers," little is known about the volume or slant of political misinformation people consume online, the effects of social media and fact-checking on exposure, or its effects on behavior. We evaluate these questions for the websites publishing factually dubious content often described as "fake news." Survey and web traffic data from the 2016 U.S. presidential campaign show that Trump supporters were most likely to visit these websites, which often spread via Facebook. However, these sites made up a small share of people's information diets on average and were largely consumed by a subset of Americans with strong preferences for pro-attitudinal information. These results suggest that widespread speculation about the prevalence of exposure to untrustworthy websites has been overstated.

AbstractHow well does public policy represent mass preferences in U.S. states? Current approaches provide an incomplete account of statehouse democracy because they fail to compare preferences and policies on meaningful scales. Here, we overcome this problem by generating estimates of Americans' preferences on the minimum wage and compare them to observed policies both within and across states. Because we measure both preferences and policies on the same scale (U.S. dollars), we can quantify both the association of policy outcomes with preferences across states (responsiveness) and their deviation within states (bias). We demonstrate that while minimum wages respond to corresponding preferences across states, policy outcomes are more conservative than preferences in each state, with the average policy bias amounting to about two dollars. We also show that policy bias is substantially smaller in states with access to direct democratic institutions.

Racial discrimination persists despite established antidiscrimination laws. A common government strategy to deter discrimination is to publicize the law and communicate potential penalties for violations. We study this strategy by coupling an audit experiment with a randomized intervention involving nearly 700 landlords in New York City and report the first causal estimates of the effect on rental discrimination against blacks and Hispanics of a targeted government messaging campaign. We uncover discrimination levels higher than prior estimates indicate, especially against Hispanics, who are approximately 6 percentage points less likely to receive callbacks and offers than whites. We find suggestive evidence that government messaging can reduce discrimination against Hispanics but not against blacks. The findings confirm discrimination's persistence and suggest that government messaging can address it in some settings, but more work is needed to understand the conditions under which such appeals are most effective.

The precise mechanisms by which the information ecosystem polarizes society remain elusive. Focusing on political sorting in networks, we develop a computational model that examines how social network structure changes when individuals participate in information cascades, evaluate their behavior, and potentially rewire their connections to others as a result. Individuals follow proattitudinal information sources but are more likely to first hear and react to news shared by their social ties and only later evaluate these reactions by direct reference to the coverage of their preferred source. Reactions to news spread through the network via a complex contagion. Following a cascade, individuals who determine that their participation was driven by a subjectively "unimportant" story adjust their social ties to avoid being misled in the future. In our model, this dynamic leads social networks to politically sort when news outlets differentially report on the same topic, even when individuals do not know others' political identities. Observational follow network data collected on Twitter support this prediction: We find that individuals in more polarized information ecosystems lose cross-ideology social ties at a rate that is higher than predicted by chance. Importantly, our model reveals that these emergent polarized networks are less efficient at diffusing information: Individuals avoid what they believe to be "unimportant" news at the expense of missing out on subjectively "important" news far more frequently. This suggests that "echo chambers"—to the extent that they exist—may not echo so much as silence.

AbstractIn this note, we provide direct evidence of cheating in online assessments of political knowledge. We combine survey responses with web tracking data of a German and a US online panel to assess whether people turn to external sources for answers. We observe item-level prevalence rates of cheating that range from 0 to 12 percent depending on question type and difficulty, and find that 23 percent of respondents engage in cheating at least once across waves. In the US panel, which employed a commitment pledge, we observe cheating behavior among less than 1 percent of respondents. We find robust respondent- and item-level characteristics associated with cheating. However, item-level instances of cheating are rare events; as such, they are difficult to predict and correct for without tracking data. Even so, our analyses comparing naive and cheating-corrected measures of political knowledge provide evidence that cheating does not substantially distort inferences.

We examine the role of overconfidence in news judgment using two large nationally representative survey samples. First, we show that three in four Americans overestimate their relative ability to distinguish between legitimate and false news headlines; respondents place themselves 22 percentiles higher than warranted on average. This overconfidence is, in turn, correlated with consequential differences in real-world beliefs and behavior. We show that overconfident individuals are more likely to visit untrustworthy websites in behavioral data; to fail to successfully distinguish between true and false claims about current events in survey questions; and to report greater willingness to like or share false content on social media, especially when it is politically congenial. In all, these results paint a worrying picture: The individuals who are least equipped to identify false news content are also the least aware of their own limitations and, therefore, more susceptible to believing it and spreading it further.

The use of individual-level browsing data, that is, the records of a person's visits to online content through a desktop or mobile browser, is of increasing importance for social scientists. Browsing data have characteristics that raise many questions for statistical analysis, yet to date, little hands-on guidance on how to handle them exists. Reviewing extant research, and exploring data sets collected by our four research teams spanning seven countries and several years, with over 14,000 participants and 360 million web visits, we derive recommendations along four steps: preprocessing the raw data; filtering out observations; classifying web visits; and modelling browsing behavior. The recommendations we formulate aim to foster best practices in the field, which so far has paid little attention to justifying the many decisions researchers need to take when analyzing web browsing data.

Widespread misperceptions about COVID-19 and the novel coronavirus threaten to exacerbate the severity of the pandemic. We conducted preregistered survey experiments in the United States, Great Britain, and Canada examining the effectiveness of fact-checks that seek to correct these false or unsupported misperceptions. Across three countries with differing levels of political conflict over the COVID-19 response, we demonstrate that fact-checks reduce targeted misperceptions, especially among the groups who are most vulnerable to these claims, and have minimal spillover effects on the accuracy of other beliefs about COVID-19. However, the positive effects of fact-checks on the accuracy of respondents' beliefs fail to persist over time in panel data even after repeated exposure. These results suggest that fact-checks can successfully change the beliefs of the people who would benefit from them most but that their effects are disappointingly ephemeral.