They do, actually, according to this methodology. It's a broken methodology. The study asks only about the previous day's diet. Assume a population where ever

They do, actually, according to this methodology. It's a broken methodology. The study asks only about the previous day's diet. Assume a population where everyone has the same diet of a milkshake randomly about once a week. If you ask the question asked by this study you will find similar results. About 12% of the people will report having a milkshake "yesterday" if everyone is uniformly consuming around one milkshake every 8.3 days.