Gluten-Free January Survey Data, Part I: Demographics and Limitations

Thursday, March 10, 2011
Thanks to Matt Lentzner for organizing Gluten-Free January, and everyone who participated and completed the survey, we have a nice data set illustrating what happens when a group of people stop eating gluten for a month. Janine Jagger, Matt and I have been busy analyzing the data, and I'm ready to begin sharing our findings.

GFJ had over 500 participants, 527 of which received the survey and 279 of which completed the survey at the end of the month. Of those who received the survey, 53 percent completed it. I think these are respectable numbers for a survey of this nature, and it reflects the conscientious nature of the people who participated in GFJ.

Demographics

Although respondents were primarily from the United States, I'm happy to say that the data represent 18 different nationalities:

Respondents represented a diversity of ages, the largest group being 30-39 years old, with similar numbers in the 20-29 and 40-49 year groups.
Respondents were just under 2/3 women.

Respondents represented a variety of weights, but the sample was biased toward lean people, in comparison with the general population. There were not many obese participants.
Overall, I was pleased to see that the demographics were quite diverse, particularly in the age and gender categories.

Limitations

There are a few caveats to keep in mind when interpreting the survey results:
  1. GFJ participants do not represent a random cross-section of the population at large. They represent primarily health-conscious individuals who were motivated enough to make a substantial dietary change. In addition, many of the people who participated probably did so because they already suspected they had a problem with gluten.
  2. The survey response rate was 53%. Although I think that's a reasonable number considering the circumstances, it leaves open the possibility that survey responders differ from non-responders. It's conceivable that participants with better adherence and better outcomes were more likely to complete the survey than those who did not adhere to the diet or had neutral or unfavorable outcomes, despite our efforts to encourage everyone to complete the survey regardless of adherence or outcome. So the results could be biased toward positive outcomes, meaning that we will need to see a strong effect for it to be believable.
  3. This was a non-blinded diet trial without a control group. There's no way to know how much of the effect was due to avoiding gluten per se, how much was due to overall changes in diet patterns, and how much was a placebo effect.
With that in mind, what can we take from the survey data? I feel that we can use it to answer the following question: "what is likely to happen when a motivated, health-conscious person decides to avoid gluten for a month?" And I think we can also use it to generate (but not test) hypotheses about the effects of eating gluten on the general population.