Here is a strange statistical result. My colleagues and I have been teaching
high school students to investigate the causes of obesity in their community—as
a form of civic education. This fall, they are going to conduct and tape interviews
and create a radio show to publicize their results. To give them some data to
work fro
m,
we surveyed all the students in the school’s health classes. The response rate
was poor, because students had to bring in parental permission slips before they could complete
the survey; and there was no penalty for failing to participate. Nevertheless, we
received enough surveys to draw tentative statistical conclusions. Here is the
one that surprises me. None of the 17 kids who said that they ate fast food
every day are overweight (according to their self-reported combination of height and weight). However, 43% of those who said they eat "hardly
any" fast food are considered clinically overweight.
What’s going on? Maybe a lot of kids are mistaken or dishonest, but it’s strange
that the relationship between fast food and body weight would be so linear and negative. The sample is too small for serious statistical analysis,
but we noticed that immigrant kids are more likely to eat fast food, yet less
likely to be overweight. So maybe immigrants eat good food at home but go out a lot
to McDonalds.
There are more possible explanations. For instance, the Washington Post’s “Kid’s Post” section reported last Wednesday that young people order less healthy food at restaurants like Outback Steakhouse and Red Lobster than they do at fast-food places. So maybe it’s good to go to McDonalds if it keeps you from ordering the “surf and turf” at a sit-down restaurant. But most of the kids we surveyed cannot afford regular visits to real restaurants.
In any case, the students’ research task is a lot harder because of this result.