Statistically Significant Pile of Rubbish

Just in time for the new year when most people are thinking about losing weight, eating healthy etc, a new study is out that touts that being overweight may actually help. In case you did not catch this weight study done by “researchers” here is a link. Blogs and news outlets wrote articles with catchy titles like

“Research: A little extra fat may help you live longer” (NPR)

“Study: People who are overweight are less likely to die than people of normal weight” (WSJ)

“Study: Being overweight might have health benefits” (WCNC)

I should point out that Times picked a more reasonable and guarded title than any other outlet (“Study suggests lower mortality risk for people overweight“).

The said overweight study was conducted by a well qualified scientist working for none other than Center for Disease Control and Prevention and will publish the paper in prestigious JAMA. Name brand organization and someone with big degrees presenting a study that has done some level of data collection and analysis seems good enough for most. To good measure the author of the study adds,

“overweight is actually associated with a lower risk of death. It’s certainly not dramatic, but about a 6 percent decreased risk. It’s statistically significant,

Ahhh, the magic words, the difference is statistically significant.

The NPR report despite its oversimplified title digs deeper into the research and brings in others into the conversation to discuss the merits of the study. One such researcher, Walter Willett, from Harvard School of Public Health said,

“This study is really a pile of rubbish, and no one should waste their time reading it,”

Willett points out many flaws with the study starting with its use of BMI (which I wrote about as well). The said study overly relies on a single irrelevant metric to make its claim without considering context, people’s health and fitness, quality of life and other lurking variables (like people who are ill and hence thin).

But media and most readers do not stop to ask questions, especially question a study from CDC scientist that is accepted for JAMA publication and when the results are statistically significant.

Did you consider questioning this study like Willett did?

How many times have you seen similar articles in the business realm?  How many tweets have you seen that say, “McKinsey study finds”, “Name-brand-analyst says”, “Seth Godin says (Ridiculous is the new Remarkable)” etc. ?

Do you stop to look for what these popular gurus may have missed and whether or not you are simply reading a pile of rubbish?