Algorithms and Agenda-Setting in Wikileaks’ #PodestaEmails Release
I’m pleased to share that Information, Communication & Society has recently accepted a paper by myself and Ed Summers (@edsu) titled “Algorithms and Agenda-Setting in Wikileaks’ #PodestaEmails Release.” You can read a pre-print copy of the full paper right here.
In the event that you don’t have the time to indulge the full paper right now, here’s a few relevant highlights:
In the month before the 2016 U.S. Presidential election, Wikileaks released a collection of e-mails authored by former Clinton campaign manager John Podesta. Wikileaks’ release methodology for this content was novel relative to its previous leaks. Rather than a bulk dump, Wikileaks serialized the release. Installments were posted once (and sometimes twice) a day, nearly every day in the month before the November 8th public vote. They also used a unique PodestaEmail related hashtag for each release (#PodestaEmails2, #PodestaEmails3, etc.). This use of sequential, but still unique hashtags for each batched release made it possible for Podesta-related content hit the Trending Topics lists again and again.
As shown below, Podesta e-mail related hashtags hit town-wide, country-wide, or worldwide Trending topics lists a total of 1,917 times in the 30 days before election day. Podesta e-mail related hashtags remained on Trending Topic lists nearly everyday in the 30 days before election day somewhere within the U.S.
As media organizations covered the content of the e-mails, some additionally made note of the fact that the topic was trending. For example, an article in The Washington Times stated:
“Twitter’s trending news feed was dominated by #PodestaEmails25 after a new batch of stolen documents belonging to John Podesta were posted online…”
The Washington Times story was subsequently shared by then candidate Donald Trump on his own Twitter page, extending attention both to the content, and the importance of the content vis-à-vis its trending. This movement of trends from Twitter to the news also broadened the scope of the potential audience to people who did not use the platform.
The release methodology was well-considered. In the full article, we describe how Wikileaks’ tweets carefully spoke to two audiences: Twitter users and Twitter algorithms, and how this ultimately furthered a dynamic flow of content between Twitter and traditional news outlets. We argue that as journalists rely on Trending Topics to identify breaking news, and report trending as newsworthy in itself or as part of a larger story, agenda-setting is now a potential by-product of getting content to trend.
In their original article defining the theory of agenda-setting, McCombs and Shaw (1972) write:
“In choosing and displaying news, editors, newsroom staff, and broadcasters play an important part in shaping political reality. Readers learn not only about a given issue, but also how much importance to attach to that issue from the amount of information in a news story and its position.”
On social media, news, editors, and newsroom staff are often not the ones making choices about what content appears prominently on a platform. Instead, algorithms are often making these decisions. When traditional news outlets cover content that is trending as newsworthy, the effect is that the algorithms play a part in shaping political reality. Users who have some degree of technical knowledge about the ways in which algorithms on social media sites work may now find they are more likely (but not guaranteed) to be able to act in an agenda-setting capacity.
Wikileaks use of serial hashtags only created the possibility for agenda-setting to occur. One does not get to the Trending Topics list merely by using lots of different hashtags. Instead, content must be shared repeatedly. There is an outstanding question of how broader user-action during the Podesta Email release made agenda-setting outcomes possible. It is beyond the scope of our data to comment on this aspect. However, as researchers move towards creating a more robust account of agenda-setting through social media, it will be critical for scholars to consider questions about the kinds of organic or coordinated power and resources necessary to achieve such outcomes.