A cash-back rebate program for healthy food purchases in South Africa: Selection and program effects in self-reported diet patterns

Ruopeng An, Roland Sturm

Research output: Contribution to journalArticlepeer-review

Abstract

Objectives: A South African insurer launched a rebate program for healthy food purchases for its members, but only available in program-designated supermarkets. To eliminate selection bias in program enrollment, we estimated the impact of subsidies in nudging the population towards a healthier diet using an instrumental variable approach. Methods: Data came from a health behavior questionnaire administered among members in the health promotion program. Individual and supermarket addresses were geocoded and differential distances from home to program-designated supermarkets versus competing supermarkets were calculated. Bivariate probit and linear instrumental variable models were performed to control for likely unobserved selection biases, employing differential distances as a predictor of program enrollment. Results: For regular fast-food, processed meat, and salty food consumption, approximately two-thirds of the difference between participants and nonparticipants was attributable to the intervention and one-third to selection effects. For fruit/ vegetable and fried food consumption, merely one-eighth of the difference was selection. The rebate reduced regular consumption of fast food by 15% and foods high in salt/sugar and fried foods by 22%-26%, and increased fruit/vegetable consumption by 21% (0.66 serving/day). Conclusions: Large population interventions are an essential complement to laboratory experiments, but selection biases require explicit attention in evaluation studies conducted in naturalistic settings.

Original languageEnglish (US)
Pages (from-to)152-162
Number of pages11
JournalAmerican journal of health behavior
Volume41
Issue number2
DOIs
StatePublished - Mar 2017

Keywords

  • Diet
  • Economic incentive
  • Instrumental variable
  • Rebate
  • Selection bias

ASJC Scopus subject areas

  • Health(social science)
  • Social Psychology
  • Public Health, Environmental and Occupational Health

Fingerprint Dive into the research topics of 'A cash-back rebate program for healthy food purchases in South Africa: Selection and program effects in self-reported diet patterns'. Together they form a unique fingerprint.

Cite this