The Washington Post’s Wonkblog backpedalled slightly after being accused of making questionable claims about alcohol consumption patterns in the United States. Wonkblog published a second piece Oct. 3, clarifying the statistics from the Sept. 25, article [1].
Sharing statistics calculated by Philip J. Cook, a Duke University professor, the original article included a graphic, which asserted the top 10 percent of drinkers in America consume 73.85 drinks per week on average.
According to Wonkblog reporter Christopher Ingraham, “That works out to a little more than four-and-a-half 750 ml bottles of Jack Daniels, 18 bottles of wine, or three 24-can cases of beer. In one week. Or, if you prefer, 10 drinks per day.”
Find this statistic hard to believe? So did Trevor Butterworth, a contributor at Forbes, who challenged Cook on his methodology. In an article entitled “When Data Journalism Goes Wrong” [2] on Oct. 2, Butterworth pointed out several critical mistakes in the Wonkblog post, including how Cook’s 2007 study used old data published by the National Institute on Alcohol Abuse and Alcoholism (NIAAA) from 2001 and 2002.
Butterworth also criticized Cook for the way he adjusted the data. “[W]e find that Cook corrected these data for under-reporting by multiplying the number of drinks each respondent claimed they had drunk by 1.97 in order to comport with the previous year’s sales data for alcohol in the US,” Butterworth said.”
This strategy was flawed according to Butterworth who pointed out, how could “everyone who drank, say, four or eight drinks per week systematically forget that they actually had eight or sixteen? That seems like a stretch.”
In its follow-up piece [3], Ingraham was forced to acknowledge, “Almost everyone wanted to know how Cook arrived at his figures.”
Ingraham went back to Cook for more information and Cook admitted, “I have gone back to the original data … and re-estimated everything.” But Cook ultimately defended his calculations as being in the right “ballpark,” concluding that the top ten percent of drinkers consume closer to 8.7 drinks per day on average.
But as Butterworth noted, the Wonkblog only went back to the original source for clarification rather than turning to Cook’s critics or even to other estimates of alcohol consumption. And there were estimates to be found. Butterworth wrote about several sets of data on drinking that were far more recent and said “none come remotely close to Cook’s estimate.”