BigSurv18 program
Wednesday 24th October Thursday 25th October Friday 26th October Saturday 27th October
Fake News! Information Exposure in Complex Online Environments |
|
| Chair | Ms Colleen McClain (University of Michigan) |
| Time | Saturday 27th October, 11:00 - 12:30 |
| Room: | 40.006 |
When Does the Campaign Matter? Attention to Campaign Events in News, Twitter, and Public Opinion
Final candidate for the monograph
Dr Josh Pasek (University of Michigan) - Presenting Author
Dr Lisa Singh (Georgetown University)
Dr Stuart Soroka (University of Michigan)
Dr Jonathan Ladd (Georgetown University)
Dr Michael Traugott (University of Michigan)
Dr Ceren Budak (University of Michigan)
Dr Leticia Bode (Georgetown University)
Dr Frank Newport (Gallup)
Scholars of political communication have long been interested in the messages that people see and recall about candidates and elections. But because we typically do not know about salient events before they occur, it has typically not been feasible to ask targeted questions about these events in surveys as they evolve. This has rendered it difficult to determine how news media information filters down to what people are thinking about. Instead, tests of information flows have focused either on the provision of specific pieces of information or on the relative salience of various categories of issues across media.
The current study uses a novel approach to identify events in news media, tweets, and open-ended survey responses that allows for the examination of where different kinds of events garner different amounts of attention and how patterns of attention compare across these data streams. Specifically, more than 50,000 respondents to the Gallup daily poll (500 each day for nearly four months) were asked open-ended questions about what they had recently seen or heard about each of the candidates in the lead-up to the 2016 U.S. Presidential election. By treating these responses as a daily bag-of-words that can be compared to the words in election news and candidate-mentioning twitter streams, we could identify terms that experienced notable spikes in attention over short periods of time. These words could then be clustered by their association with events that had occurred in the news or that people were otherwise attending to.
In total, we identified more than 200 keywords that were associated with 39 distinct events that appeared across the news, Twitter, and survey data. These events varied in the extent to which they garnered attention from each data stream. For example, compared to other media, candidate comments at the Al Smith Charity Dinner were disproportionately salient in the open-ended responses whereas Hillary Clinton’s “Basket of Deplorables” comment about Trump supporters was a focus of much media attention, but was rarely mentioned in open-ended responses.
Importantly, we find that the prominence of identifiable events in the news, twitter, and survey data streams is distinct from the content that is commonly extracted from textual data using topic modeling techniques. Focusing on topics and events reveals distinct patterns of attention over time and notably different stories about how media messages influence public perceptions and recall.
Is Informal Flagging for Propaganda in User Comments Helpful to Identify Anti-Western Narratives? The Benefits and Risks of Relying on User-Based Labeling
Mr Vlad Achimescu (University of Mannheim, Germany)
Mr Dan Sultanescu (CPD SNSPA, Bucharest, Romania) - Presenting Author
Mrs Dana Sultanescu (CPD SNSPA, Bucharest, Romania)
The comments section of any online news source ideally functions as a virtual public sphere, where engaged citizens freely debate the topics contained in the article. However, the anonymity feature offered by some of these websites can encourage different types of distortions of public opinion. Recent concerns of academics and state actors refer to the potential emergence of a new form of propaganda through distributing anti-Western messages: attacking human rights, freedom of association, or spreading false information and conspiracy theories, by means of bots or recruited commentators operating from ‘troll factories.’ Online anti-Western propaganda allegedly contains characteristics of big data: high volume and velocity, inconsistency, and multichannel distribution. It is particularly difficult to identify and distinguish from freely expressed opinion.
Flagging comments as spam or as hate speech has been successfully used to identify non-flagged comments of the same nature and reduce their prevalence. While there is no ‘flag as propaganda’ button, we observed that informal flagging occurs quite frequently on news websites, when another user replies to a comment with a text accusing the original poster of spreading propaganda. This type of unstandardized signaling does not result in censoring the initial message. Since manual coding is expensive and time consuming, we are checking whether it is useful to use these unofficial flags to identify potential topics promoted by anti-Western propaganda. We discuss the potential dangers of externalizing the labeling process to website users, such as an increased false positive rate.
The data consists of all the comments posted over a 3.5 year period between 2015 and 2018 on two popular Romanian online newspapers, collected by web scraping. We complement it with information from telephone surveys collected in the same period. We use a two-step supervised machine learning approach in order to identify messages that represent potential anti-Western propaganda. In the first step, flaggers are identified by using specific keywords and separating between comments that contain the accusation and those that do not. In the second step, all informally flagged comments are used as the training set, with a test set containing all other comments from the same website and a different online news website. The predictors used in the second step include phrases from the identified comments and paradata, such as length, rating, and posting time. Topics resulting from the identified posts are classified and compared to the topics approached by news channels associated with promoting anti-Western narratives, as well as the Romanian public agenda as it results from surveys. Preliminary results show that comments identified as propaganda are posted by a limited number of users on a diversity of topics, have more replies and more negative reviews.
The data we have so far can only help in identifying potential propagandistic messages, channels and rhetoric devices, but not their effects on public opinion. For this second goal, we will create a panel of Romanian online media users in 2018, where survey responses can be linked to the user’s online activity.
Echo Chambers: Twitter Versus Online News Exposure
Professor Susan Banducci (University of Exeter) - Presenting Author
Dr Iulia Cioroianu (University of Exeter)
Dr Lorien Jasny (University of Exeter)
Dr Travis Coan (University of Exeter)
Dr Hywel Williams (University of Exeter)
Dr Iain Weaver (University of Exeter)
It has been argued that the Internet, and social media in particular, has increased selectivity in news and information exposure. However, it has also been argued that the Internet has removed the gatekeepers of the mainstream media and opened up new sources of information to users creating a market place of ideas. Individuals are faced with a wider range of options (from social and traditional media), new patterns of exposure (socially mediated and selective), and alternate modes of content production (e.g. user-generated content). These processes, many have argued, increase polarisation and segregation in news consumption. However, recent research has questioned whether all online news exposure is alike and finds the extent to which online exposure reflects polarisation or echo chambers depends on whether news stories are consumed through social media, search engines, news aggregators or direct clicks.
One of the issues that hampers a comprehensive study of the extent of selectivity in news exposure is the availability of data that allows an examination of online news consumption. In this paper we compare news exposure by comparing topics of news stories during the 2016 Brexit campaign in the UK to news stories viewed by a panel of respondents who answered a series of attitudinal measures across a 3 wave survey and also allowed their web browsing histories to be captured. Using these sources of data, our analysis makes two contributions to the debate on selectivity and news exposure. First, using browsing histories we can investigate actual news consumption online across traditional and social media. Second, in order to assess the attitudinal basis of selectivity, we combine the online browsing histories with a panel survey. Third, using topic models, we compare the news stories consumed by our panel to the Brexit stories circulating on Twitter. In this way, we can investigate exposure and examine the extent of selectivity across subgroups of the citizenry and compare to selectivity on Twitter.
Our approach allows comparison of new exposure patterns by domains versus news exposure to topics, to our knowledge, the first analysis to allow this comparison. We find, consistent with recent empirical work, the extent of segregation in exposure may be overstated but importantly segregation and selectivity are less evident amongst domains than amongst the topics of stories. Furthermore, the degree of segregation and selectivity varies across groups that are defined by holding shared political preferences. For example, in our case of Brexit, those who supported the ‘Leave’ side were more selective in their news exposure. We also find that topics on Twitter differ from the topics read by the respondents in our online panel. We conclude that studies that rely exclusively on Twitter are likely to overreport polarisation amongst all citizens and show topics that are different than those consumed by a more representative sample of individuals.
Boys on the Tweet Bus: Identifying Information Flows Between Political Journalists During the 2016 U.S. Presidential Campaign
Professor Jonathan Ladd (Georgetown University) - Presenting Author
As social media has become ubiquitous in American society, it has gained particular prominence among specific communities. Political journalists, for instance, are rampant Twitter users, and have been shown to use Twitter both to communicate and to gain information to use for story ideas and sources (Parmalee, 2013). What remains to be seen, however, is how this affects the influence of individual journalists. Specifically, we are interested in determining 1) how information flows through a network of identified political journalists, and 2) how this affects traditional understandings of influence.
To answer these questions, we collected tweets and follower networks from 936 political journalists during the 2016 U.S. presidential election, beginning in fall 2015. These data allow us to first visualize the follower network of all journalists, to see how information might potentially flow between these individuals. Based on network structure, who are the information disseminators? Who are the information brokers? We then use this follower network in conjunction with the tweets sent and retweeted by these journalists to better understand how information spreads through the network. We look at information spread in general, as well as by topic of the tweets. This allows us to identify who is most influential journalist with respect to flow within the journalist network and flow beyond the network.
While traditional understandings of political journalism suggest a hierarchy of sorts, with most established brands like the New York Times and Wall Street Journal garnering the greatest influence, these analyses allow us to determine whether that hierarchy is maintained or upended in the new era of political journalism that includes social media.