A theme that I have visited many times in the last six years is the use of data in enriching journalism. In an earlier column, “ New skills for new challenges ” (May 18, 2015), I had argued that for data journalism to gain gravitas, journalists must develop skills not only to understand quantitative methods but also to grasp the qualitative inputs that go into the process of data collection. In the last three years, we have seen data being presented in an engaging manner, with even a daily graphic, ‘Data Point’, being featured on the Op-Ed page.
An incomplete infographic
With context, data are illuminating; without it, data become a confusing set of numbers. Last week, The Hindu data team managed to illuminate and confuse. The team produced one of its finest data stories following the Karnataka election results in “To the spoiler belongs the victory?” (May 16, 2018). The one that failed the readers’ test was “Ministry report card”, published on May 15.
Professor Vinod Pavarala, chair holder for the UNESCO Chair on Community Media at the University of Hyderabad and one of the earliest campaigners for community radio in India, wrote about “Ministry report card”: “I am really puzzled by how the newspaper’s editors have allowed this in print. An interesting and aesthetically presented graphic in itself cannot be the reason, when you do not give readers any useful information to assess the credibility of this rating. We don’t know what this LocalCircles, described simply as ‘a community social media platform’, is. How can its registered users constitute any kind of a scientific sample for the credible ranking of ministries, however crude the ranking scale itself maybe? This is a nothing story/graphic really, one where more information is hidden than it is revealed.”
Professor Pavarala is right. The infographic was incomplete. The data team wanted to give some background about LocalCircles, the sample size as well as the methodology. They wanted to explain how LocalCircles has 10,000 registered users across cities and about its regular polling on many issues. They also wanted to explain how LocalCircles votes on Ministries based on a Ministry’s work compared to the party manifesto. But they were hit by a technical glitch. Instead of holding back the story for the next day, they went ahead and published the evaluation of various Ministries with the hope that it would make sense to the readers. It did not. Data stories dealing with policy issues are not breaking news and holding them for a day will not cause any harm.
The New York Times has four golden rules for data journalism: provide context, describe processes, reveal patterns and explain the geography. These rules apply to Indian data journalists too. The Assistant Managing Editor of The New York Times , Matthew Ericson, who was earlier Deputy Graphics Editor, observed: “Graphics should bring something new to the story, not just repeat the information in the lead. A graphics team that simply illustrates what the reporter has already told the audience is not doing its job properly. A graphic can bring together a variety of stories and provide context.” In the case of “Ministry report card”, we don’t know whether the participants are from urban areas or rural areas, we do not have data relating to gender, and we do not know about the online polling methodology.
Data that provide context
On the other hand, the analysis of the Karnataka polls results was rich. The graphic had a detailed regional analysis; it looked for some pattern by comparing numbers from three elections — 2013, 2014 and 2018; it looked at the urban-rural split in votes, it looked at the comparison between vote share and number of seats, and it even looked at the hypothetical situation of a pre-poll alliance between the Congress and the Janata Dal (Secular).
When dealing with the performance of more than 50 Ministries, there is a need to be extra cautious. Rebekah E.D. McBride, in her book The Ethics of Data Journalism, has an important note of caution: “Though structural constraints on time and resources exist for journalists, making this type of reporting difficult, the fact remains that the proper contextualization of data through systemic reporting is essential to avoid ethical missteps and the oversimplification of complex systemic issues”. Data journalism handbooks remind journalists that they should start with data and finish with a story.