Visualizing swine flu versus tuberculosis

Hans Rosling, the brilliant Swedish statistician and doctor, has been on a tear lately, publishing a lovely series of short videos on Gapminder.org. The videos are short Rosling lectures, centered on data visualization, designed to dismiss your preconceptions regarding important questions about health and development.

The most recent video in the series focuses on one of my favorite statistics: the death to news ratio. Rosling looks at the 31 deaths from swine flu and the quarter million news stories he was able to find on Google News and compares this to the estimated 63,000 deaths from tuberculosis around the world in the same time period, and the 6 thousand stories about those deaths. There are over 8100 news stories for each death from swine flu, and less than a tenth of a story for each TB death, by Rosling’s calculations.

Rosling points out that, if swine flu does become a genuine pandemic, more media attention will be appropriate. In the meantime, he issues a media hype warning regarding the disproportionate attention paid to the disease.

The death to news stories ratio is a favorite topic of international news geeks – discussions usually centers on the question, “How many African kids have to die to make the front page of the New York Times?” The answers tend to be pretty depressing, particularly when cynical reporter friends introduce hierarchies that run from African children through to blond American kids at the bottom of wells, or kidnapped in Mexico.

There’s a slightly more affirmative way to look at these situations, offered by Johan Galtung and Mari Ruge, in their 1965 paper “The Structure of Foreign News“. Galtung and Ruge offer a dozen factors that help explain how a particular story becomes or does not become news. We could throw several of Galtung and Ruge’s factors at Rosling’s implicit question: why are we paying so much more attention to swine flu than to tuberculosis:

F6: The more unexpected the signal, the more probable it will be recorded as worth listening to.
…
F9: The more the event concerns elite nations, the more probable that it will become a news item.
F10: The more the event concerns elite people, the more probable that it will become a news item.

Rosling’s map makes it very clear that cases of TB are concentrated in Africa, especially in southern Africa, and largely affect people who appear to be, culturally and physically, very far away from the readers of many of the northern news sources tracked by Google News. TB has been a crisis for a long, long time, and therefore isn’t a surprise in the way that a new pandemic – potentially affecting the President of the United States, if his aide shook his hand on his visit to Mexico! – is.

And then there’s factor 7: “If one signal has been tuned into the more likely it will continue to be tuned into as worth listening to.” In other words, once you’ve run 100,000 swine flu stories, the next 150,000 may be inevitable.

Unfortunately, Galtung and Ruge weren’t as straightforward about offering solutions to problems of media attention as they were at identifying them. But visualizations like Rosling’s, showing just how big events we manage to ignore are in comparison to those we obsess over, have to be a good start.