We’re in a boom time for analytical Web journalism that uses data to make politics and policy compelling to readers. Since parting ways with The New York Times, Nate Silver has been hiring staff in preparation for launching an expanded FiveThirtyEight site under the Disney corporate umbrella. At Silver’s former home, meanwhile, David Leonhardt is heading up a new data-focused project that will cover economics, politics, sports, and more. And just this morning, Wonkblog creator Ezra Klein announced that he is leaving The Washington Post for what is likely to be a large-scale independent venture.
Even though it doesn’t cover politics and policy, there’s another publication that belongs in this discussion: Grantland, the ESPN-owned sports and pop culture site. Like Wonkblog and FiveThirtyEight, Grantland is a largely independent satellite of a larger media brand; it was spun off from ESPN under the direction of editor Bill Simmons and given significant creative freedom to experiment with format and content. And though the site may be best known for its long, writerly features (including a spectacularly ill-conceived story on the inventor of a new golf putter last week), its most notable contribution to the visual grammar of the Internet may come from the young contributors to its sports coverage, who are developing creative ways to combine stats-based analysis, commentary, and illustrative visual examples. While this approach is a natural fit for sports coverage, it offers lessons for coverage of politics and policy as well.
Over the past few years, journalists and designers have experimented with how to combine visuals and text more effectively. Two models have been especially successful. At one extreme are splashy, expensive, heavily engineered multimedia extravaganzas like The New York Times‘ “Snow Fall” or its South China Sea story. We’re starting to see experimentation with similar approaches in special investigations or long features on politics and policy as well. For instance, the Detroit Free Press’s acclaimed investigation “How Detroit Went Broke” integrated charts, animated graphs, and video interviews into its main narrative. While often compelling, this type of story is expensive and difficult to replicate on a consistent basis. The other approach that has worked extremely well is the “chartblog,” which features quick, punchy explanatory Web items built around a chart or graph. Much of the content at Wonkblog or the earlier version of FiveThirtyEight follows this model.
Grantland’s best analytical work succeeds by blending the virtues of these different approaches while sidestepping their pitfalls–offering more narrative and multimedia variety than a typical chartblog, more voice and better integration of data than a traditional news article, and a less demanding and time-consuming user experience than “Snow Fall.” The result is a form of readable, smart, creative coverage that doesn’t take buckets of money, or weeks of time, to produce.
Grantland’s NBA coverage in particular stands out. Lead NBA writer Zach Lowe bases his analyses on statistics rather than anecdotes, but uses illustrative YouTube clips and animated GIFs as well as annotated stills from game footage to show how player and team tendencies help create the patterns we observe in the data. For instance, this recent column on the improvement of Indiana Pacers star Paul George, who has become one of the best players in the league, uses a series of still images from games (including one annotated with an arrow and text) to illustrate improvements in George’s ball-handling and decision-making. The images are integral to the data-based analysis, but they appear in the flow of the text. Another Grantland contributor, geography professor Kirk Goldsberry, uses shot charts to illustrate player shooting efficiency statistics by floor location and frequency–which often allows him to challenge conventional wisdom about players like Boston Celtics point guard Rajon Rondo (who is actually now a good midrange shooter).
Along similar lines, Grantland baseball contributor Ben Lindbergh has created memorable columns using animated GIFs to illustrate variation in the pitch-framing abilities of catchers and how Derek Jeter compares with a top defensive shortstop. In both columns, Lindbergh identified well-chosen examples to show how individual skills translate into meaningful statistical differences between players. (Grantland football contributor Chris B. Brown also uses animated GIFs effectively to illustrate team and individual tendencies using specific plays as examples.)
This approach won’t always extend easily to politics or policy. Not only are videos and data more easily accessible in sports than politics, but sports statistics represent the cumulation of directly observed actions, whereas the influences on, say, public opinion are not filmed and measured over time. Nonetheless, it’s worth thinking about how political coverage can go beyond chartblogging to present other forms of visual information that help readers understand patterns or trends.
How might this be done? Consider polling coverage. Write-ups of public opinion data sometimes include quotes from follow-up interviews with respondents as well as graphical analyses in sidebars. For instance, this New York Times story on disapproval for healthcare reform among uninsured people quotes a handful of respondents reached after participating in the poll and includes two sophisticated graphics online. Why not embed short audio or video clips from illustrative voter interviews, along with poll charts, within the write-ups? While those individual clips would not be representative of the national population, journalists who use care in selecting appropriate examples could effectively help to illustrate larger trends in the data.
You could imagine a similar approach to James Fallows’s quadrennial presidential debate previews for The Atlantic. While these articles often overstate the effects of debates on presidential election outcomes, Fallows is an incisive analyst of televised debates and the way they are perceived in the media. But his articles–which are primarily oriented toward print–lack visual information. The online version of his 2012 story includes a five-minute interview with Fallows speaking over historical clips, but the clips are not integrated within the story and require a reader to watch the video separately. The 2016 version of this story might combine Fallows’s analyses with shorter embedded clips and graphics showing post-debate changes (or the lack thereof) in candidates’ standing in the polls.
Similar approaches are possible in coverage of public policy. Reporters writing about, say, a change in positions or emphasis on an issue like gun control could examine patterns in bill topics or word frequencies in Congressional debate and use clips from C-SPAN and cable news to show how the parties are changing in their approach–a relatively easy way to dramatize what could be an otherwise dry or lifeless story.
Admittedly, none of these examples would be as revolutionary as “Snow Fall” or the original chartblog format. Maybe Silver, Klein, or Leonhardt will dream up something bigger–or maybe the next advance in analytical journalism will be incremental, building on and blending elements of what came before. Either way, in politics as in sports, the best analysis will combine statistics with visual information and narrative to present an account that is both more grounded in evidence and easier to understand than the text and anecdote-driven style of traditional reporting. Who will figure out the new model first?
Follow @USProjectCJR for more posts from this author and the rest of the United States Project team.
Brendan Nyhan is an assistant professor of government at Dartmouth College. He blogs at brendan-nyhan.com and tweets @BrendanNyhan.