Lack of efficacy of a salience nudge for substituting selection of lower-calorie for higher-calorie milk in the work place

Abstract

Obesity is a major burden on healthcare systems. Simple, cost effective interventions that encourage healthier behaviours are required. The present study evaluated the efficacy of a salience nudge for promoting a change in milk selection from full-cream to low-fat (lower calorie) in the kitchen of a university-based research institute that provided full-cream and low-fat milk free of charge. Milk selection was recorded for 12 weeks (baseline). A sign with the message "Pick me! I am low calorie" was then placed on the low-fat milk and consumption was recorded for a further 12 weeks. During baseline, selection of low-fat milk was greater than selection of full-cream milk (p = 0.001) with no significant milk-type × time interaction (p = 0.12). During the intervention period overall milk selection was not different from baseline (p = 0.22), with low-fat milk consumption remaining greater than full-cream milk selection (p < 0.001) and no significant milk-type × time interaction (p = 0.41). However, sub-analysis of the first two weeks of the intervention period indicated an increase in selection of both milk types (p = 0.03), but with a greater increase in low-fat milk selection (p = 0.01, milk-type × time interaction). However, milk selection then returned towards baseline during the rest of the intervention period. Thus, in the present setting, salience nudging promoted a transient increase in low-fat milk consumption, but also increased selection of full-cream milk, indicating that nudging was not effective in promoting healthier milk choices.

Keywords: behavioural economics; milk selection; nudging; obesity; salience nudge; workplace.

Figure 1

Selection of low-fat and full-cream milk prior to and during a 12 week nudging intervention. Data analysis was based on weekly selection but has been presented as the mean of 12 weeks of baseline data and means of two- and four-weekly periods during the intervention period for simplicity of graphical representation. Data are mean ± standard error. *Significantly different from baseline (p = 0.03) and significantly greater increase in low-fat milk selection compared with full-cream milk selection (p = 0.01; first two weeks of the intervention period only).