The Effect on Supermarket Food Purchases from Point of Sale Nudges with Community Reinforcement
A project was designed with a supermarket to provide point of sale signs or shelf talkers which identified foods in specific categories that were considered to be healthier compared to other items in the category and which were already commonly purchased. In the store, posters were also put up that encouraged healthy choices (and advertise the shelf talkers) and food tastings with recipes were provided. Within the community, a targeted information campaign to explain the effort and encourage people to buy the products identified by the shelf talkers was implemented in a rural community from June 2011 to December 2012 including local newspaper, television public service announcements, and presentations to community groups.
This paper will evaluate the effect of the program on changing food purchases at the store level. Four hundred and two items that were relatively healthier among alternatives within different food category had shelf talkers. The monthly sales data of these products with shelf talkers from intervention and control store across four years will be used. The data set also includes two years’ monthly sales data of the entire produce item being sold in the intervention and control store. All the monthly sales data were obtained from computerized cash registers which record individual purchases by Universal Product Codes. These data will be used in two separate analyses to see if the point of sale nudges and supermarket health promotion intervention could increase the consumption of relatively healthier items and fresh fruits and vegetables. The focus is on quantitative outcomes, on whether or not the supermarket intervention was effective in changing food behavior.
Statistical Analysis
The difference between the control and intervention store sales data on point of sale nudges items and produce purchases will be tested using independent t-tests. The effect of intervention on the purchase of point of sale nudges item and produce will be analyzed using multiple linear regressions. In this a cross-sectional time series regression with an error term with autocorrelation and cross sectional correlation structure will be used. The analysis will be done separately for the 402 shelf talker item and the produce sales. In the case of shelf talker items the analysis will be done with total sales and the percent sale as the dependent variable. Similar model with total sales as the dependent variable will be used to analyze the impact on sale of produce items. The independent variable in both the analysis will be price, quadratic term for price, control for location, year and monthly variations, the variable of interest is a location and intervention year dummy.
Implication
There are few contemporary evidence-based intervention strategies for reducing the risk of obesity that focus on the supermarket with community reinforcement. This research will contribute significantly towards an increasing effort in implementing health promotion programs through community-based approach including supermarket as a key component.