We Need More Data Journalism to Reveal Biased Journalism
The Silent Author lurks.
Like many cultural reporters I know, I got into journalism for selfish reasons: I like to write. There’s a fair share of us, I must assume, who were actually drawn to reporting because they are genuinely fascinated by the art of collecting information and delivering it to the masses. But many of us are here because we like the sound of our own words read aloud.
I didn’t realize the fault in this approach to journalism until my first real newsroom internship, when my editor pushed me to write and publish a story that had style—not content—that made me squirm with unease. The story, a delicate subject with a main source that had brushed with death, was ripe for stylistic exploitation and my editor encouraged me to juice it for all it was worth. The result? An overly subjective, dramatized recounting of events that, while factual and approved by my source, felt to me to be a gross injustice to the story at hand. Since this story, I’ve never looked at narrative journalism, once my dream medium, in the same way.
The fact of the matter is, while narrative storytelling touches readers and can stay with them for years (I think of this article often, even five years after first reading it), it is a tool that should only be deployed when appropriate, yet is often abused by view-hounding editors like mine and literary-minded writers like myself. When inappropriately applied, narrative reporting can warp a story beyond recognition, obscuring necessary facts and magnifying others until the final product transforms into literature with little journalistic purpose.
In the worst cases, the hidden voice of the narrative journalist can steer readers to a predetermined conclusion. The reporting surrounding the “Who Is the Bad Art Friend?” controversy comes to mind; readers felt manipulated by reporters, specifically because they could barely parse the story facts from the writer’s opinions. While reporting can never be unbiased, no matter the medium, I believe readers deserve news that gives them the chance to form their own opinions—or at the very least, identify the journalist’s personal logic and disagree.
This, I believe, is why readers desperately need data journalism.
What Is Data Journalism and Data Visualization?
Although data journalism technically describes any form of reporting that uses data as a key source, its definition has changed dramatically over the years and varies depending on who you ask. The data journalism to which I’m referring is more accurately called “data visualization” and takes the form of interactive graphics that allow readers to easily explore and manipulate a data set. You likely became familiar with data visualization during COVID while perusing the interactive maps depicting infection rates around the world.
I personally became acquainted with data visualization when, a year or so before the pandemic, I stumbled upon The Pudding, a cultural publication whose data essays changed the way I view the format. As you might have noticed, the two articles I have criticized in this essay—my click-fueled drama narrative and “Who Is the Bad Art Friend?”—don’t seem to lend themselves to data visualization, mostly because they are human interest stories. That’s how I used to view data visualization: as an inventive but narrowly applicable tool. If you peruse The Pudding’s stories, however, you might be surprised at how apparently effortlessly these journalists apply data visualization to human interest stories—from colorism in high fashion to women’s minuscule pockets.
It was this discovery, of the infinite applications of data visualization in any given sector of reporting, that ultimately convinced me to go back to school for a master’s degree in data science. Pursuing a career in data visualization was a no-brainer for me, not only because I’m a visual learner but because I struggle with the knowledge that my inherent biases will always impact my work and, by extension, my readers.
How Data Journalism Can Kill the (Silent) Author
Click through a map of COVID infections and you’ll start to get an idea of how data visualization can allow readers to hold more power over their own learning by doing away with what I like to call the Silent Author. The Silent Author is the quiet bias journalists bring to their reporting, the assumptions, personal opinions and blind spots we (intentionally or not) sprinkle throughout our articles and which often go unnoticed by the author as well as the reader.
Say a reader is investigating infection rates around the world to plan a trip. A data-driven article outlining the safest countries to visit during COVID might be informed by the journalist’s interpretation of infection data, but it is limited to just that: their personal interpretation, which likely includes unconscious biases, like which countries are worth visiting in the first place. For a multitude of reasons, these biases can go unnoticed by the reader; for example, it is difficult to perceive when the scope of an article is warped or when an author makes an interpretive leap that they wouldn’t agree with. Many readers simply don’t think to question a journalist’s authority, especially when they publish with a major news outlet.
In contrast, an interactive map of the same data set allows readers to peruse infection data free of the Silent Author’s guiding voice. Of course, data journalists usually accompany interactive graphics with their own analysis, but their biases are no longer silent; readers can approach the author’s interpretations with a critical eye by exploring the data themselves. Data visualization takes your middle school English teacher’s instruction to “show, not tell” seriously, with the added bonus of allowing you to do both and giving the reader the opportunity to determine whether your “show” and “tell” match up.
We Can Never Get Rid of the Author Entirely
You’ll notice that even this essay is biased.
Unfortunately for you, I have only just started my data science degree and am not yet capable of producing a graphic that allows you to compare the bias in traditional print journalism and data visualization. Instead, I find myself, a self-professed data visualization zealot, writing an article that focuses on the positive potential of this reporting style. But, in an attempt to be a Not-So-Silent Author, I will tell you that data visualization is not, and could never be, truly unbiased.
One of the greatest flaws in data visualization projects is that we often mistakenly consider data objective when in reality, data sets are inherently skewed. Take our interactive COVID infection maps, for example. Not all countries are equally efficient or honest ind their reports of infections, China being a noted one. This alone will warp a graphic’s accuracy. In other cases, journalists might use data sets that are not large or diverse enough to be representative of the population or subject at hand. Worse yet, since journalists rarely collect their own data sets, there is always the possibility that a reporter might knowingly or unknowingly use one that has been produced by a disreputable source. And then, of course, data visualization does not solve the issue of story choice bias, wherein the stories that make it to publication always reflect an editor’s opinion of what (or who) is important enough to merit news coverage.
All of these factors and more prevent data visualization from being some sort of perfect solution to media bias, but such a solution will never exist. Data journalism as a whole has long been heralded as the solution to any number of media woes: gendered harassment in the workplace, bias in reporting, the existential threat of how to provide reporting of increasingly higher quality, among others. While I don’t believe data visualization will solve any of these puzzles, the way I see it, it is the next step that journalism can take toward owning up to the natural biases that reporters bring to their stories.