Dubious Data Comes Home
I just returned from the Discovery 2009 and Innovator’s Summit in Chicago. The conference is all about analytics, data mining, visualization and the people who make their living understanding numbers.
One the speakers, Joel Best, really caught my attention. He is the author of several books including Stat-Spotting: A Field Guide to Identifying Dubious Data and his talk centered on how statistics, even when they’re wrong, can be cited and quoted and accepted into popular culture.
He describes a case in which a reporter covered the death of a celebrity due to complications of anorexia. The article stated that over 150,000 women in the US also suffer from the disease.
That was picked up by another reporter who mutated the statistic into a claim that 150,000 women die from the disease each year! This alarming stat (completely invented) was picked up by other news agencies, authors, and blogs as each author cited the previous article.
When Best looked into the matter, he discovered that the death rate for women between 15 and 40 (the age group most affected by anorexia) is only 55,000 per year – total. He quipped “I was surprised to learn that out of the 55,000 young women that die each year, 150,000 of them die due to anorexia” and finished with “ A bad statistic is harder to kill than a werewolf.”
With this lesson in mind, I returned home to catch an article in the Daily Tar Heel, our campus paper, about some protesters who want to see us become a coal free university. Their timing was good, because we had just released the climate action plan describing how we would do just that.
What struck me in the article was this line: “Outside the building, they held posters opposing the plant and claimed that the facility’s daily carbon emissions equaled that of 50,000 cars.”
The average person on the street has no experience of how much carbon is “a lot” or “a little.” Without context, how could you? It’s like asking your neighbor whether 21 degrees Celsius is hot or cold when they use the Fahrenheit system. Who knows?
But a little visit to the EPA’s carbon equivalency website reveals the error. Plant emissions of 320,000 tons of carbon ARE roughly equivalent to 50,000 cars (with certain assumptions about mileage and efficiency) but that represents the entire year of emissions, not a single day. From what I can see, the protestor’s signs make no distinction of time, but the article certainly does.
As they say in Rome, “caveat lector”: let the reader beware.
For more hilarious examples, see the Live Blog from Joel Best’s keynote speech here.
Related posts:
