Adults recruited on social media were asked to provide their current height and weight. They were then asked a question about their own childhood environment or behavior that could have led to them to have their current body mass index (BMI). For example, “When you were a child, did you drink juice or soda more often than water?” They were then given the opportunity to suggest new questions, such as “When you were a child, did you have many friends?” The new questions were included in the ones asked of subsequent participants and each person’s predicted BMI was continuously updated based on his or her answers.
All together, 532 people participated over a two week period. Thirty-seven of them generated 56 new questions that they thought could help predict whether children grow into slim or obese adults. Interestingly, the question most highly correlated to future BMI was “When you were a child, did someone consistently pack a lunch for you to take to school?”
You can see the top nineteen questions below.
Interestingly, some of the questions with the most predictive power were not ones typically considered by health care professionals.
To be clear, the data does not show that if you pack your child’s school lunch everyday, he’ll never become obese. For one thing, it’s just a correlation. There could and probably are underlying factors that go along with bag lunches, like how nutritious that lunch is likely to be, or even whether that child has a caretaker with the time to pack a lunch. This is true for all the questions. Also, the adults participating in the study were relying on their memories of their childhoods to generate and answer questions.
The bigger point is that crowdsourcing can be a way of finding correlations that experts might miss. By asking people what life was like for them as children, health care professionals can gather new insights into how to keep the next generation of kids from becoming obese.