Every year, we see a glut of stories in the media debunking the myth of the holiday weight gain. The general narrative is one of disbelief that the average holiday weight gain is 5 to 10 pounds (see for example, this SF Gate story) and that the majority of weight gained by the average person in a year is gained during the period between Thanksgiving and Christmas (see for example, this article). This straw man of holiday weight gain is then torn down by scientific studies that show the actual weight gain is “less than a pound.” (including this 2007 blog from the NY Times, this 2013 story from Fox News, and this recent Yahoo Health story).
Almost every article I found cited this study and this study from the year 2000, which states that “On average, weight gain during the 6-week winter period from Thanksgiving through New Year averaged only 0.37 kg.” For those of you not up on your metric conversions, this is the equivalent of about .816 pounds.
I am a big fan of scientific studies, but I was curious about the fact that every story seemed to be citing the same “1 pound” statistic. Is it really possible that an entire narrative of the myth of holiday weight gain could continue to cite back to original studies from 15 years ago? Yes, it turns out. And, I decided to take a closer look at the original studies.
First, both rely on what may be the same “convenience sample” of 195 people. A convenience sample is a non-representative sample of people who are, as one website describes it, “readily available” to participate. This approach surely means the findings are potentially not representative of what we would expect in the broader population. In fact, the authors say that in the abstract: “Further studies using representative populations are needed to confirm these findings.”
Another key finding concerned the fact that people don’t lose their holiday weight gains. But, according to the second study, only 165 of the original 195 people were available to be measured a year later – making it difficult to know for sure exactly how many people dropped the extra pounds.
Finally, because of something called the “confidence interval” it’s possible that the supposed weight gain could actually be a weight loss. A confidence interval is a way to measure the level of statistical certainty about results. Typically expressed as a range of values, the confidence interval tells you the range of values within which you’re likely to see the estimate (assuming, of course, you have a random—and representative—sample). In this case, the confidence interval is approximately 3.35 pounds. What that means is that although the best estimate of weight gain is about .8 pounds, given the precision of the study, the actual weight gain could be between losing 2.55 pounds (which is .8 pounds minus 3.35 pounds) and gaining 4.15 pounds.
In other words, the original studies that debunked the myth 15 years ago are nowhere near as definitive as one might think.
Whether you’re trying to watch your weight – or deciding whether or not to leave cookies out for Santa – a good consumer of this type of every day data (what I call everydata) will dig a little deeper to look at the underlying research that underpins a decade long narrative.