In the context of the article.... self control correlates with an increase in obesity.
Wasn't the metric for calculating self control was how long you were willing to wait for an additional marshmallow? Seems like a proxy for measuring self control that would be biased towards people with a propensity for obesity.