261 — Eat & Tell: A Randomized Trial of Random-Loss Incentive to Increase Dietary Self-Tracking Compliance

245 individuals from Singapore volunteered to track their food intake for thirty days. Half of these individuals were told that of their starting S$35, one dollar would be deducted for every day that they failed to log their food.

But the other group was told something different: They would lose somewhere between S$0 and S$3 each time they failed to log their food for the day.

The expected values of loss for the two programs were — unbeknownst to the subjects, but beknownst to us — exactly the same. (Both groups knew they wouldn’t owe money at the end of the survey, of course.)

The random-loss group was noticeably more likely to log their food intake than the control group, which confirms the long-held economics hypothesis that random loss is more threatening than known-loss, and conserves meaning better across individual interceptions (i.e. a single penalty).

This is very interesting for any case involving compliance: If the penalty of noncompliance doesn’t change, then it is likely to fade in significance in the eyes of the participant, whereas a random and unpredictable penalty maintains efficacy (at least for 30 days).