It’s difficult to choose which research articles to spotlight here as the most interesting or compelling — because scholars are doing so much interesting and compelling work. They’re continually asking tough questions to try to understand problems and trends within the digital news/social media space.
Recently, we enlisted help from faculty at a variety of institutions to select half a dozen studies published or released in recent months that we think you should know about. These studies investigate a range of topics — from native videos and Twitter usage among political journalists to inaccuracies in health reporting and the online harassment of female journalists.
Remember, this is just a sampling of new studies from the second quarter of 2018. If you’d like to suggest research for our next roundup, message us at @JournoResource.
This article was first published at Nieman Lab. It has been slightly edited to reflect Journalist’s Resource’s editorial style.
“Twitter Makes It Worse: Political Journalists, Gendered Echo Chambers, and the Amplification of Gender Bias”: From University of Illinois at Urbana-Champaign, Calvin College, and George Washington University, published in The International Journal of Press/Politics. By Nikki Usher, Jesse Holcomb, and Justin Littman.
Research on gender dynamics on social media generally focuses on political candidates. This study focuses on political journalists. And the findings, according to the researchers, are “deeply concerning.”
Usher, Holcomb and Littman looked at how elite political journalists in the Washington, D.C. area — those credentialed to cover Congress — interact on Twitter “to legitimate, amplify, and engage each other.” The three scholars analyzed the activity and profiles of 2,292 journalist Twitter accounts between June 1, 2017 and Aug. 1, 2017.
Here’s what they discovered: 22 of the 25 most prolific tweeters were men and the male political journalists had nearly twice as many followers as the female political journalists. The researchers also found that, on average, the entire sample of journalists was more likely to follow male journalists than female journalists. When political journalists retweeted other journalists, they retweeted men about 69 percent of the time and women about 31 percent of the time.
Male political journalists replied to other male political journalists 91.5 percent of the time and to female journalists 8.5 percent of the time, the study shows.
“Retweet patterns suggest that men live in a gendered echo chamber that promotes other male journalists at the expense of female ones: Men retweet other men almost three times more often than they retweet women,” the authors explain. “In contrast, women still retweet male journalists more than they retweet female journalists.”
“‘You Really Have to Have a Thick Skin’: A Cross-Cultural Perspective on How Online Harassment Influences Female Journalists”: From the University of Texas at Austin, National Chung Cheng University, El Paso Times, and Ludwig Maximilian University of Munich, published in Journalism. By Gina Masullo Chen, Paromita Pain, Victoria Y. Chen, Madlin Mekelburg, Nina Springer, and Franziska Troger.
This second study also looks at gender in journalism — how online harassment affects female journalists. Journalists are encouraged (or required) to interact with their audiences online, especially on social media. But when female reporters do it, the researchers found, they face “rampant” harassment, including sexist comments and threats of violence. And the strategies women use to deal with the abuse can interfere with their work.
Chen and her colleagues interviewed 75 female journalists who work or have worked in the United States, Germany, India, Taiwan, and the United Kingdom. They sought out women of different ages and races who worked at a variety of media outlets. While this small sample isn’t representative of female journalists as a whole, the researchers say their goal was to “find meaning through the female journalists’ words, not make generalizable inferences.”
Of the 75 journalists interviewed, 73 said they had been harassed online. The 23 TV journalists experienced abuse that often was most pronounced and frequent, the authors note. The study highlights specific examples of harassment and women’s efforts to handle them, described in their own words. For example, an online editor from Germany said: “The feedback [on this article] was not criticism, it was threats, it was death threats, it was calls for rape.” A broadcast journalist in the U.S. said people leave misogynistic comments so often on her professional Facebook page that she blocks certain words from appearing. “I have moderation on my page for the words ‘sexy,’ ‘hot,’ or ‘boobs,’” she said.
The researchers say journalism schools and media organizations must pay more attention to the issue. “However, women in our sample stressed that what is most important is that newsrooms change the culture that allows this abuse to continue,” they write.
“All Forest, No Trees? Data Journalism and the Construction of Abstract Categories”: From the University of Alabama, published in Journalism. By Wilson Lowrey and Jue Hou.
Newsrooms have begun to focus more enthusiastically on the use of data to tell important stories and highlight societal ills. Large, complex data sets and analyses are the centerpiece of some of the most powerful journalism projects published in recent years. But Wilson Lowrey, a professor in the Department of Journalism and Creative Media at the University of Alabama, has found that as newsrooms have emphasized numbers, their work has lost some important details — the colorful descriptions and personal anecdotes that bring news stories to life.
Lowrey led a review of 194 data journalism projects published in the United States and the United Kingdom between 2011 and 2016. Most — 58.2 percent — were produced by legacy news organizations while the remaining 41.8 percent came from non-legacy organizations, primarily digital-only outlets.
Here’s the gist of what Lowrey and his colleague found: Data projects increasingly relied on “abstract constructs” created from data sets — for example, an index of income inequality — to help audiences understand complicated issues. Over time, they also became less likely to incorporate anecdotal reporting. “Anecdotal reporting was linked with more than a third of all data projects, and this percentage decreased from 44.0 percent in 2011 to 26.0 percent in 2016,” the authors explain.
The study points out a key shortcoming of the data projects they reviewed: Most offered limited evidence that journalists checked the validity and reliability of the data they used. Most projects failed to include an explanation of their data’s limitations.
“These findings and the concerns they raise become more poignant when we consider that many of the data projects examined here have been held up as model efforts,” Lowrey and his colleague write.
“Causal Language and Strength of Inference in Academic and Media Articles Shared in Social Media (CLAIMS): A Systematic Review”: From Harvard University, the University of North Carolina at Chapel Hill, Boston Children’s Hospital, University of Pennsylvania, Tufts University, Boston University, Johns Hopkins University, published in PLOS ONE. By Noah Haber, Emily R. Smith, Ellen Moscoe, Kathryn Andrews, Robin Audy, Winnie Bell, Alana T. Brennan, Alexander Breskin, Jeremy C. Kane, Mahesh Karra, Elizabeth S. McClure, Elizabeth A. Suarez, on behalf of the CLAIMS research team.
Over the last year and a half, numerous studies have focused on “fake news” and how bad information spreads online. This study offers insights about how journalists themselves contribute to the problem. It examines how accurately reporters and academics describe the findings of health research.
A team of researchers analyzed 50 health research studies and the 64 news articles written about them that were most shared on Facebook and Twitter in 2015.
Haber and his colleagues found that journalists often overstated the findings of the studies and made other errors. Almost 60 percent of news articles inaccurately reported the question, results, intervention or population of the study. Also, the authors of the research studies sometimes used language that did not accurately describe what they had found. “Thirty-four percent of academic studies and 48 percent of media articles used language that reviewers considered too strong for their strength of causal inference.”
The findings of this study don’t apply to all news coverage of health research or all health-related academic work — just the articles and studies that were examined. However, the findings indicate the need for a closer examination of the issue, Haber and his colleagues write. “Review of the strength of causal inference and language in work published by authors at the top medical institutions as well as reviews of media articles across a range of popularity levels is needed to better understand where and why issues occur in the research production and dissemination pathway,” they write.
“News Organizations’ Use of Native Videos on Facebook: Tweaking the Journalistic Field One Algorithm Change at a Time”: From Nanyang Technological University and the University of St. Gallen, published in New Media & Society. By Edson C. Tandoc Jr. and Julian Maitra.
This study looks at how newsrooms have responded to Facebook’s decision in June 2014 to change its News Feed algorithm to prioritize native videos — that is, videos directly uploaded to the platform, not linked to or embedded from elsewhere — so that they appear more prominently in news streams. The findings underscore the significant role Facebook and its more than 1 billion daily active users play in modern newsroom routines.
Tandoc and Maitra analyzed data collected from 232 Facebook pages operated by major U.S. news organizations in the year before and the year after Facebook announced it tweaked its algorithm. The scholars examined data on more than 10 billion interactions and more than 2 million posts, including 41,373 native videos.
They learned that print, broadcast and digital-native news organizations all began posting more native videos after the change. Broadcast newsrooms led the pack. Native videos accounted for 1.9 percent of broadcast organizations’ posts before the algorithm change. Afterward, the proportion jumped to 9.3 percent. “Digital natives increased the proportion of their Native Videos from 1.4 percent to 4 percent, while print-based publishers increased theirs from 0.2 percent to 1.5 percent,” according to the study.
The authors say their findings “provide yet another demonstration of the journalistic field’s heteronomy that persists, if not increases, as journalism increasingly gets digitized.” The findings, they write, “show how an agent external to the field, whose founder has specifically declared it is not a media company, extends its influence to the journalistic field.”
“Trust, Misinformation, and the Declining Use of Social Media for News: Digital News Report 2018”: From the Reuters Institute for the Study of Journalism at the University of Oxford. By Nic Newman, Richard Fletcher, Antonis Kalogeropoulos, David A. L. Levy, and Rasmus Kleis Nielsen.
This year’s annual report from the Reuters Institute covers a lot of ground in its 144 pages, looking at data collected via survey in 37 countries to offer insights on a range of digital news issues, including consumer distrust, messaging apps, podcasts and crowdfunded startups. Some of the main takeaways: Fewer people are turning to social media for news in the U.S. and a number of other markets. The proportion of Americans who said they used social media as a source of news within the past week fell 6 percentage points to 45 percent since last year. Most of that change is related to Facebook — news consumption via Facebook dropped 9 percentage points nationwide and 20 points among younger groups.
The report dedicates an entire section to “fake news” and takes a global look at concerns about online misinformation and people’s exposure to it. For example, people’s views about government intervention varied substantially by country. In the U.S., 41 percent of the individuals surveyed said the government should do more to make it easier to detect fake information online, compared to 61 percent in the United Kingdom and France and more than 70 percent in Spain and South Korea.
The report discusses the growing popularity of voice-activated digital assistants such as the Amazon Echo and their potential for distributing podcasts and other audio news products. “Usage has more than doubled in the United States, Germany, and the UK with around half of those who have such devices using them for news and information,” according to the report.
Interested in other roundups of digital media research? Check out our Top 10 list for 2017 and our roundup of the most interesting research of the first quarter of 2018. We also have write-ups on the evolving relationship between social media and news and the best length for online news videos.