Here are two headlines that ran half a year apart in The New York Times:
Obesity Rates Hit Plateau in U.S., Data Suggest (January 14, 2010)
Obesity Rates Keep Rising, Troubling Health Officials (August 3, 2010)
Did things really turn around that quickly? Probably not. The first story, which I noted here and here, was based on data from the National Health and Nutrition Examination Survey (NHANES), which uses actual measurements of weight and height. As of 2008, the NHANES numbers indicated that obesity rates had been essentially unchanged for at least five years among men and for nearly 10 years among women and children. By contrast, data from the Behavioral Risk Factor Surveillance System (BRFSS), the basis for the second story, show obesity continuing to rise during the same period.
Unlike NHANES, the BRFSS relies on self-reported weights and heights. As the Times notes, these data, for understandable reasons, tend to understate the percentage of Americans who have body mass indexes of 30 or more (the official cutoff for obesity). As the Times does not note, the shrinking gap between the NHANES and BRFSS numbers suggests that Americans are becoming increasingly precise and/or candid about such matters. "The data show a 1.1 percentage point increase—an additional 2.4 million people—in the self-reported prevalence of obesity between 2007 and 2009 among adults aged 18 and over," the CDC reports. "The number of states with an obesity prevalence of 30 percent or more has tripled in two years to nine states in 2009." But unless the next set of NHANES data (which, again, are based on actual measurements rather than self-reports) indicates the the BMI plateau was just a brief rest during a continuing climb, the recent upward trend in the BRFSS data will prove to be illusory.