Does the supplemental nutrition assistance program really increase obesity? The importance of accounting for misclassification errors

Achilleas Vassilopoulos, Andreas C. Drichoutis, Rodolfo M. Nayga, Jr, Panagiotis Lazaridis

Research output: Contribution to journalArticlepeer-review

1 Citation (Scopus)

Abstract

The prevalence of obesity among US citizens has grown rapidly over the last few decades, especially among low-income individuals. This has led to questions about the effectiveness of nutritional assistance programs such as the Supplemental Nutrition Assistance Program (SNAP). Previous results on the effect of SNAP participation on obesity are mixed. These findings are however based on the assumption that participation status can be accurately observed, despite significant misclassification errors reported in the literature. Using propensity score matching, we conclude that there seems to be a positive effect of SNAP participation on obesity rates for female participants and no such effect for males, a result that is consistent with several previous studies. However, an extensive sensitivity analysis reveals that the positive effect for females is sensitive to misclassification errors and to the conditional independence assumption. Thus analogous findings should also be used with caution unless examined under the prism of classification errors and of other assumptions used for the identification of causal parameters.

Original languageEnglish
Pages (from-to)1-10
Number of pages10
JournalJournal of Applied Statistics
DOIs
Publication statusAccepted/In press - 2017 Dec 20
Externally publishedYes

Keywords

  • misclassification
  • obesity
  • propensity score matching
  • sensitivity analysis
  • SNAP

ASJC Scopus subject areas

  • Statistics and Probability
  • Statistics, Probability and Uncertainty

Fingerprint Dive into the research topics of 'Does the supplemental nutrition assistance program really increase obesity? The importance of accounting for misclassification errors'. Together they form a unique fingerprint.

Cite this