Twitter  This week in review, learn why Twitter's bringing data sales in house and who's killing their comments section nie.mn/1h9wNrB  
Nieman Journalism Lab
Pushing to the future of journalism — A project of the Nieman Foundation at Harvard
social media, facebook, twitter, pinterest, instagram, iphone apps

Interrogating the network: The year in social media research

Do comment sections build a bias against expertise? Do people remember Facebook posts? How much does news drive search, and vice versa? These are some of the most noteworthy findings in academic research in 2013.

Editor’s note: There’s a lot of interesting academic research going on in digital media — but who has time to sift through all those journals and papers?

Our friends at Journalist’s Resource, that’s who. JR is a project of the Shorenstein Center on Media, Politics and Public Policy at the Harvard Kennedy School, and they spend their time examining the new academic literature in media, social science, and other fields, summarizing the high points and giving you a point of entry.

This year, John Wihbey has been summing up the top papers in digital media and journalism each month. Now it’s time for the best of the year.

Over the course of 2013, this column has spotlighted academic papers and reports on a monthly basis; about 100 in total were highlighted. In searching for what to include, we reviewed about 800 articles this year. Most came from among a group of about 30 journals, as well as some think-tanks. We also consulted with a small network of about a dozen scholars on a regular basis to maintain field awareness. Below are a dozen highlights. (See last year’s bunch for a point of comparison.) Of course, this is only meant to be a sample. Thanks to all who contributed suggestions on Twitter (@JournoResource). See you again in 2014.

“Mapping the global Twitter heartbeat: The geography of Twitter”: Study from the University of Illinois Graduate School of Library and Information Science, University of Illinois at Urbana-Champaign, published in First Monday. By Kalev Leetaru, Shaowen Wang, Guofeng Cao, Anand Padmanabhan, and Eric Shook.

One of the most comprehensive assessments to date about the effect of new technologies on human communications worldwide, this study examined patterns among more than 1.5 billion tweets from 70 million users over a one-month period in late 2012. It provides empirical evidence that the world is indeed shrinking: “There appears to be only weak geographic affinity in communicative link formation in that users retweet and reference users far away nearly as often as they do those physically proximate to them.” Further, on average people tweet news that happens locally and news about far-away events with about equal frequency. Twitter is “not simply a mirror of mainstream media” and has its own distinct conversational dynamics. The data also show that significant portions of the “world’s most influential Twitter users” were in places such as Indonesia, Western Europe, Africa and Central America. The overall takeaway is that where we live is beginning to matter less in terms of our knowledge, interests and social networks: “Geographic proximity is found to play a minimal role both in who users communicate with and what they communicate about, providing evidence that social media is shifting the communicative landscape.”

“Private traits and attributes are predictable from digital records of human behavior”: Study from the University of Cambridge and Microsoft Research, published in the Proceedings of the National Academy of Sciences (PNAS). By Michal Kosinski, David Stillwell, and Thore Graepel.

The study examines the degree to which information available online can successfully predict an individual’s personal — and private — attributes. The researchers correlated public records of Facebook “Likes” from more than 58,000 users with results from personality and intelligence tests and information from public profiles. The researchers were able to accurately predict a male user’s sexual orientation 88 percent of the time. While less than 5 percent of users were explicitly linked to gay policy or advocacy groups, “predictions rely on less informative but more popular Likes, such as ‘Britney Spears’ or ‘Desperate Housewives’ (both moderately indicative of being gay).”

The model was able to predict a user’s ethnic origin (95 percent) and gender (93 percent) with a high degree of accuracy. “Patterns of online behavior as expressed by Likes,” the researchers write, “significantly differ between those groups, allowing for nearly perfect classification.” The model also predicted a user’s religion (82 percent), political views (85 percent), relationship status (67 percent) and substance use (between 65 percent and 75 percent for drugs, alcohol and cigarettes) with a high degree of accuracy. The researchers caution against the potential negative outcomes that ready access to this type of personal data might have: “Commercial companies, governmental institutions, or even one’s Facebook friends could use software to infer attributes such as intelligence, sexual orientation or political views [that] could pose a threat to an individual’s well-being, freedom or even life.”

“Major Memory for Microblogs”: Study from UC-San Diego, the University of Scranton, and the University of Warwick, published in Memory and Cognition. By Laura Mickes, Ryan S. Darby, Vivian Hwe, Daniel Bajic, Jill A. Warker, Christine R. Harris, and Nicholas J. S. Christenfeld.

The study explores how social media content is read and remembered. The researchers conducted three experiments to assess how well Facebook posts are remembered as compared to other types of information — particularly news such as CNN articles or entertainment stories — and the extent to which remembering is enhanced by a perceived social connection or post content. The researchers also compared reader comments and the content of news articles.

The findings defy expectations that social media posts are ephemeral and fleeting; in fact, in many cases they are more memorable than professionally produced content. “Especially memorable Facebook posts and reader comments, generated by ordinary people,” the researchers write, “may be far closer than professionally crafted sentences to tapping into the basic language capacities of our minds…Some sentences — and, most likely, those without careful editing, polishing, and perfecting — are naturally more ‘mind-ready.’” The study proposes that the language of social-networking sites and microblogs has shifted contemporary expectations for writing from formal conventions toward increased spontaneity.

“The ‘Nasty Effect’: Online Incivility and Risk Perceptions of Emerging Technologies”: From George Mason University and University of Wisconsin-Madison, published in the Journal of Computer-Mediated Communication. By Ashley A. Anderson, Dominique Brossard, Dietram A. Scheufele, Michael A. Xenos, and Peter Ladwig.

The study, published in February, has the distinction of being one of the few to actually help change editorial policy this year at a publication. Popular Science cited the study in its announcement that it was shutting down its comments section to push back against a perceived “war on expertise.” There remained some controversy, however, about whether the study’s conclusions were broad enough to justify that editorial decision. The researchers used online surveys with embedded experiments to test how people responded to articles about nanotechnology; some were accompanied by nasty comments, others not. The study’s findings suggest that “impolite and incensed blog comments can polarize online users based on value predispositions utilized as heuristics when processing the blog’s information.” Further, the researchers note, “The effects of online, user-to-user incivility on perceptions towards emerging technologies may prove especially troublesome for science experts and communicators that rely on public acceptance of their information. The effects of online incivility may be even stronger for more well-known and contentious science issues such as the evolution vs. intelligent design debate or climate change.”

“Did Twitter Kill the Boys on the Bus? Searching for a better way to cover a campaign”: From Harvard’s Shorenstein Center, by Peter Hamby.

CNN’s Hamby shows how social media shaped news narratives and campaign-press dynamics during the 2012 presidential campaign. Filled with up-close anecdotes and opinions from prominent actors, it’s a 95-page white paper that is part fast-paced narrative, part field ethnography and part zoological study of the political and chattering classes. Hamby interviewed 70 journalists and political operatives involved in the race. He uses the Romney presidential campaign as a case study to look at the digital disruption among journalists and decision-makers inside campaigns. “Reporters are not exactly a humble bunch,” Hamby concludes. “But most of the journalists interviewed for this piece expressed some form of regret about how they used Twitter during the campaign. It was, by far, the biggest source of dismay and angst in discussions with reporters about the current state of political journalism. No one is complaining about the revolutionary gateway to news and information that Twitter provides. But plenty of people in politics are anxious about the way the Twitter conversation thrives on incrementalism, self-involvement and snark.”

“Exploring News Apps and Location-Based Services on the Smartphone”: Study from San Diego State University, published in Journalism & Mass Communication Quarterly. By Amy Schmitz Weiss.

Are news media missing business and growth opportunities by not offering and utilizing more geolocation functionality in their mobile apps? Weiss analyzes more than 100 native apps from top TV network affiliates and radio stations, as well as other news apps in Apple’s App Store. She combines that content analysis with results from an online survey of young news consumers, who are increasingly likely to employ geolocation “check-ins” and location-based services as part of their mobile experience.

She finds that the “adoption of geo-located news stories is nonexistent among the traditional media examined. Six apps that did offer geo-located news were mainly user-generated apps.” The verdict on news organizations is damning, and the implications are clear: “Legacy news organizations analyzed in this study show that they are failing to keep up with the demand based on what news consumers, particularly young adults, are doing and using on their smartphones. This is supported by the proven hypothesis in this study that found younger adults who use location based services are also likely to consume news on their smartphone.”

“Twitter as a Reporting Tool for Breaking News”: Study from the University of Sheffield, published in Digital Journalism. By Farida Vis.

The study looks at the Twitter usage patterns of The Guardian’s Paul Lewis and The New York Times’ Ravi Somaiya as they covered the London riots in 2011. Vis also draws on millions of tweets from hundreds of thousands of users employing specific hashtags during that same four-day period in August 2011. Lewis’ account was the second-most mentioned (he tweeted 441 times and was mentioned more than 30,000 times by others), while Somaiya was 34th (290 tweets and about 3,500 mentions.) They employed different strategies: “it is clear that Paul Lewis’ tweets are most often original tweets (312 tweets, 71 per cent) compared with Ravi Somaiya’s (133 tweets, 46 per cent).” Other differences include: “Paul Lewis uses 54 @ replies…whilst Ravi Somaiya dedicates more than a third of his tweets (89 tweets…) to @ replies”; Lewis shared 111 links, while Somaiya shared 70; and Lewis sent out 82 tweets with crowdsourcing requests for information, compared to three such requests from Somaiya.

Vis concludes that “studying breaking news on Twitter and early adopters in these situations is important as it can highlight the emergence of new journalistic conventions, which a focus on established journalistic norms alone may fail to identify.” For example, Somaiya voiced more opinion than old-school norms might allow, suggesting an evolving “hybrid norm,” the study notes. In many ways, the study is a natural follow-up to an earlier paper, by Alfred Hermida, Seth Lewis and Rodrigo Zamith, titled “Sourcing the Arab Spring: A Case Study of Andy Carvin’s Sources During the Tunisian and Egyptian Revolutions.”

“State of the News Media 2013″: Report from the Pew Research Center’s Project on Excellence in Journalism. By Amy Mitchell et al.

As always, Pew’s annual reports provide a foundation for informed conversation about the media (the Lab also published a valuable interview with the report’s lead author). Key findings for this year include: As news outlets have slashed staff and reduced the quantity and quality of coverage, the report suggests, many consumers have responded negatively: “Nearly one-third — 31% — of people say they have deserted a particular news outlet because it no longer provides the news and information they had grown accustomed to, according to [a] survey of more than 2,000 U.S. adults in early 2013.” About half of all people surveyed said news stories are not as thorough as they were previously. Of the consumers who reported abandoning certain news outlets, 61 percent said the decision was based on issues of quality, while 24 percent said there were not enough stories. Newspaper ad revenue is now down 60 percent compared to a decade ago. The number of U.S. news jobs is likely now below 40,000, compared to the historic high of 56,900 in 1989, a 30 percent decrease overall.

Amid the gloom, bright spots include: “In 2012, total [online] traffic to the top 25 news sites increased 7.2%, according to comScore. And according to Pew Research data, 39% of respondents got news online or from a mobile device ‘yesterday,’ up from 34% in 2010, when the survey was last conducted.” Further, the emerging mobile market offers another opportunity, as many people appear to be consuming more news because of Internet-enabled devices. This offers opportunities for the news business: “One piece of [the mobile] market that news can exploit is sponsorship advertising, and in 2012, so-called native advertising (a type of sponsorship ad) made headlines. Though it remains small in dollars, the category’s growth rate is second only to that of video: sponsorship ads rose 38.9%, to $1.56 billion; that followed a jump of 56.1% in 2011. Traditional publications such as The Atlantic and Forbes, as well as digital publications BuzzFeed and Gawker, have relied heavily on native ads to quickly build digital ad revenues, and their use is expected to spread.”

“Reuters Institute Digital News Report 2013: Tracking the Future of News”: Paper from University of Oxford Reuters Institute for the Study of Journalism, edited by Nic Newman and David A. L. Levy.

This report provides tremendous comparative perspective on how different countries and news ecosystems are developing both in symmetrical and divergent ways (see the Lab’s write-up of the national differences/similarities highlighted). But it also provides some interesting hard numbers relating to the U.S. media landscape; it surveys news habits of a sample of more than 2,000 Americans.

Key U.S. data points include: the number of Americans reporting accessing news by tablet over a given week rose, from 11 percent in 2012 to 16 percent in 2013; 28 percent said they had accessed news on a smartphone in the past week; 75 percent of Americans reported accessing news online in the past week, while 72 percent said they got news through television and 47 percent reported having read a print publication; TV (43 percent) and online (39 percent) were Americans’ preferred platforms for accessing news. Further, a yawning divide exists between the preferences of those ages 18 to 24 and those over 55: among the younger cohort, 64 percent say the Web is their main source for news, versus only 25 percent among the older group; as for TV, however, 54 percent of older Americans report it as their main source, versus only 20 percent among those 18 to 24. Finally, 12 percent of American respondents overall reported paying for digital news in 2013, compared to 9 percent in 2012.

“Effects of gender and tie strength on Twitter interactions”: Study from Rutgers, Google, and Cornell NYC Tech, published in First Monday. By Funda Kivran-Swaine, Samuel Brody, and Mor Naaman.

The researchers set out to study gender differences in behavior on Twitter by analyzing some 78,000 messages among more than 1,700 pairs of persons. They conclude: “Gender differences revealed in our analysis have mostly confirmed observations in traditional settings; women use higher levels of [first person plural, or "we,"] [first person singular, or "I"], intensifiers, and emoticons in their speech, with levels escalating even more when they converse with other women, hinting at accommodation.”

The study catalogues the words that most distinctively characterize — that are most predictive of — female-to-female messages (“love”) and male-to-male interactions (“dude” or “man”). Many of the old Venus and Mars clichés are at work, with some nuances: “These results suggest that, in their Twitter interactions, women tend to reference both themselves and others, more than men do…In general, the female linguistic style that was manifested in our study is more socially aware than linguistic style exhibited by men. This may be due to the fact that even when conversing with those they feel close to, in Twitter, women’s interactions are more about people and social happenings, whereas men prefer a style that is less personal.”

In related research, also see “News sourcing and gender on Twitter,” from Claudette G. Artwick at Washington and Lee University. The study, published in Journalism, analyzes 2,731 tweets from journalists (26 men, 25 women) at 51 different newspapers during 2011. The problems in this area are persistent and well-documented, and Artwick reviews the prior literature on gender imbalances in news stories. In her sample, she finds sources named in about 19 percent of tweets (507 sources quoted overall). Just 11 percent of those quoted were women, thus “women’s voices were relatively silent in the quotes on these reporters’ Twitter streams.” Further, at larger papers, “less than 8 percent of female reporters’ quotes featured women, and male reporters quoted no women at all.” Through the use of “@” mentions, however, reporters were “engaging with a more diverse community”: Nearly four of every 10 “@” mentions were women.

“Beyond Cognitions: A Longitudinal Study of Online Search Salience and Media Coverage of the President”: Study from DePaul University, published in Journalism & Mass Communication Quarterly. By Matthew W. Ragas and Hai Tran.

The paper takes an empirical look at the evolving two-way street of how news coverage can drive online search — and how online search can also drive media coverage. Ragas and Tran use the Associated Press and Reuters as their representative indicators of news coverage and analyze data from the U.S. Search Intelligence database of Experian Hitwise. The study looks at coverage of President Obama during 2009-2010. Predictably, more AP and Reuters coverage — particularly negative coverage — was associated with more online search around Obama. But, interestingly, Ragas and Tran found that “coverage volume was also influenced by search trends, demonstrating an instance of reverse agenda setting with the media seemingly monitoring and taking cues from Internet users. Moreover, the impact of search salience on media salience occurred relatively quickly (starting within a week), while the media-led influence appeared to accumulate over a five-week span.”

The findings validate greater media investment in monitoring of the digital space — they “lend empirical support to recent observations of journalists monitoring, influencing, and reacting to search trends and the rise of the active audience in web environments.” For communications and journalism scholars, the study is particularly interesting because it shows that the traditional dynamics of media “agenda setting” — telling the public what to think about, and how to think about it — is changing and becoming a more dynamic process.

“The Internet and American Political Campaigns”: From George Washington University, published in The Forum. By David Karpf. (Pre-print open version here.)

Part of a growing cohort of academics pioneering the subfield of online politics, Karpf provides a short, useful summary of the state of research in this area. For journalists, the works cited page alone is a valuable who’s who — fill up that contact list for campaigns 2014 and 2016 — but the narrative also underscores some basic truths: The web has not changed many forms of participatory inequality; polarizing candidates frequently win the small donations race; the “culture of testing” and analytics are changing how campaigns allocate resources; liberals and conservatives typically use technology differently for campaigns.

One striking insight: “We are potentially moving from swing states to swing individuals, employing savvy marketing professionals to attract these persuadables and mobilize these supporters with little semblance of the slow, messy deliberative practices enshrined in our democratic theories.” But definitive answers remain elusive on many other fronts. “There is still, frankly, a lot that we do not know,” Karpf writes. For more insights in this area, see Kathleen Hall Jamieson’s response, “Messaging, Micro-Targeting and New Media Technologies.”

                                   
What to read next
los-angeles-register
Ken Doctor    April 16, 2014
The first edition of the Orange County Register’s expansion into the Times’ turf is hot off the presses. Is this about selling papers or positioning for a further shakeout of the newspaper market?