Learning about Our Vices from Devices: A Model of Individual Learning with an Application to Consumer Food Waste

Abstract

The proliferation of personal, household and workplace sensors and devices has created individual environments rich with purposeful and incidental feedback capable of altering behavior. We formulate an empirical learning model suitable for understanding individual behavioral responses in such environments. We estimate this model using data collected about the joint personal decisions of food selection, intake, and waste during a study in which users photographed their meal selections and plate waste over the course of a week with a cell phone. Despite neutral recruitment language and no expectation that participants would alter food intake in response to the assessment procedures, we found a substantial learning-by-doing effect in plate waste reduction as those who document greater plate waste in their captured photographs waste less on subsequent days. Further we identified that participants reduced plate waste by learning to eat more rather than by learning to reduce the amount of food selected.

Keywords: Learning by doing; behavior tracking; dietary intake; food; food waste; obesity.

Figure A1.

Google search trend in Food Waste and Obesity in the U.S. from 2008 – 2020

Figure 1.. Average grams of plate waste per person per day

Notes: 95% confidence intervals presented on bar figures.

Figure 2.

Association between cumulative plate waste experience and current plate waste generation