DSAIL Black Twitter Project
DSAIL: New Approaches to Audience Research (February 2014)
Data Science at the Annenberg Innovation Lab (DSAIL)
Lead researchers: Prof. François Bar, Dayna Chatman, Kevin Driscoll, Alex Leavitt
http://datascience.annenberglab.com
This research project is conducted in partnership with the Media Impact Project, a global hub for collecting, developing and distributing the smartest approaches and best tools for measuring the impact of media.
http://mediaimpactproject.org
With generous support from IBM:
http://ibm.com
Summary
As television viewership expands across multiple channels, platforms and services, audience researchers are turning to social media services such as Twitter to complement traditional media metrics. Conventionally, social media analytics focus on population-level characteristics such as the total number of tweets and retweets containing a particular keyword or hashtag. We believe that this approach vastly understates the range of practices and thematic diversity among Twitter audiences. Furthermore, such macro-scale views are easily manipulated and difficult to compare across different types of programming.
DSAIL is currently developing a multi-method approach to studying public discourse on Twitter that explores both macro and micro-scale activity simultaneously in order to draw out particularly active, engaged “neighborhoods” within the larger population. Among the many different ways that audiences incorporate Twitter into their media ecologies, “live-tweeting” is one of the most promising for researchers. “Live-tweeting” refers to an open-ended discussion among casual viewers, producers, critics, fans, and anti-fans alike that unfolds in response to television programming, in connection with real-time viewing. From sports events to awards shows to original content, this sort of real-time activity offers a unique opportunity for researchers to listen in on live commentary from thousands of viewers at once.
In pursuit of a generalizable set of best practices for audience research on social media, DSAIL is working toward an understanding of highly active sub-community of Twitter users often self-identified as "Black Twitter.” This case study turns on a tricky classification problem: not everyone who identifies as black is a part of Black Twitter, nor does everyone participating in Black Twitter identify as black. In spite of this ambiguity, coverage of Black Twitter in the popular and trade press has exploded during the last year. The community is credited with mobilizing attention around a variety of news and entertainment events, including attracting millions of new viewers to the network drama Scandal and drawing national news coverage to the killing of Trayvon Martin. The data we are collecting will allow us to map specific user connections, explore how information is spread and by whom, and identify the types of communication practices that are unique to Black Twitter. As part of this research, we are engaged in highlighting and evaluating both the immediate and long-term social implications of these online interactions.
What is Black Twitter?
Black Twitter is a dynamic, open-ended, socio-technical phenomenon. Black Twitter exists as a fluctuating collective rather than a clearly defined “community,” although some Twitter users that identify as black racially (or in other ways) also identify as a member. Black Twitter manifests through discussion and interactions among individuals and groups as well as the political, cultural, and media events that resonate with these users’ collective interests and experiences. Users contributing to the – notably lively – Black Twitter discourse are among the most deeply engaged with the affordances of Twitter’s platform and employ technical features such as the hashtag, search function, and Trending Topics to raise the widespread visibility of content and themes that they value. As a result, newcomers to Twitter are occasionally surprised to find that the system can have a profoundly Black feeling to it.
Data Collection & Analysis
The dynamism of Black Twitter presents a methodological challenge because it defies the conventional keyword-based approach that researchers rely on to track communities of interest on social media. Unlike similarly open-ended groups who identify themselves using a shared hashtag (e.g., #tcot, or “top conservatives on Twitter”), the tweets by those users that make up Black Twitter tend to exhibit more complex rhetorical strategies – such as references to black culture or ways of calling out similarly-identified users – that evade simple mechanical classification. The terminology and humor of Black Twitter evolves rapidly over time, complicating any effort to generate a definitive list of key terms or phrases.
In order to observe the unpredictable flow of Black Twitter activity, we turn instead to a structured set of events around which a significant percentage of the Black Twitter community has gathered. Our case study focuses on the popular television show, Scandal (of which the protagonist is notable actress, Kerry Washington). From October 3 to December 12, 2013, we tracked the activity of any user tweeting about Scandal, and logged their Twitter conversations and user metadata. With this collection as a starting point, we have begun to map out the relationships among users who “live-tweet” Scandal in an effort to identify sub-groups of users that interact with one another outside of their shared interest in the TV show. From there, we hope to identify particular cliques of individuals that identify as part of Black Twitter to further explore their engagement with the show and each other on Twitter as well as their offline participation more generally in black culture and politics.
Next Steps
Currently, we are in the preliminary stages of analyzing the data collected during the fall season. Our initial, exploratory observations suggest that there are numerous overlapping sub-groups within the Scandal fandom (see figure1, right), but there are also a number of distinct smaller groups of individuals that are highly connected separate from the main group. We have also identified a large number of co-occurring hashtags in addition to #Scandal that we will use to triangulate practices amongst these users. Further, we observed that for hard-core fans of Scandal, live-tweeting while watching the show when it first airs is an important part of the experience -- to the extent that many West Coast viewers will log off twitter during the the East Coast airing of the show to avoid spoiling the experience. The immediate next steps for investigating this dataset include examining how users identify with Black Twitter and mapping those behaviors to other viewers of Scandal.
In the next stage of our research, we plan to dig more deeply into the full range of interests shared within and among these sub-groups to see how and in what ways their participation on Twitter differs. In other words, what do these users talk about in addition to Scandal, and what does this tell us about the social and political implications of Black Twitter? Further, we hope to uncover how Twitter as a communication platform enables and impacts individuals of this evolving collective.
Generalizability
Beyond the specific case of Black Twitter, we believe the techniques and research practices we are developing have broad applicability to the study of audience engagement. Similarly, we expect our research results to be relevant to other social media communities. In particular, our current research is providing key insights into “live tweeting” practices, which suggest powerful ways to zoom in on audience characteristics such as engagement, and spill-over between different topics. In addition, we believe that our observations about community belonging, inclusion and exclusion will have broad relevance beyond Black Twitter.
Updates
DSAIL is currently exploring the cultural practices of a highly active population of Twitter users who self-identify and are identified by others as "Black Twitter." During the last year, coverage in the trade press has exploded as "Black Twitter" is credited with mobilizing attention around political issues and popular culture events, including attracting millions of new viewers to the network drama Scandal and drawing national news coverage to the killing of Trayvon Martin.
The emergence of the term "Black Twitter" reflects the central role that young black Americans have played in shaping Twitter. According to the Pew Research Center's Internet & American Life Project, 40% of African Americans between 18-29 who are on the internet use Twitter. But not everyone who identifies as black (racially or in other ways) also identifies as a part of Black Twitter, nor does everyone participating in Black Twitter identify as black. Black Twitter exists as a dynamic, open-ended collective that manifests through discussion of particular political, cultural, and media events rather than a stable, clearly defined "community."
Through complementary efforts at micro-scale participation observation and macro-scale data collection, DSAIL is developing collective expertise that allows us to investigate and map specific user connections, explore how information is spread and by whom, and identify the types of communication practices that are unique to Black Twitter.
______________
1) A visualization of follower relationships among users who live-tweeted the October 3rd Scandal premiere. The algorithms used for this visualization will color the dots (users) more-strongly connected to one another.