Evaluating food environment assessment methodologies: a multi-level examination of associations between food environments and individual outcomes

  • Author / Creator
    Minaker, Leia M.
  • Background: The food environment (FE) is being increasingly recognized as an important and modifiable determinant of diet quality and weight status. Hundreds of FE measures exist, resulting in a lack of comparability between studies. This is problematic given that evaluating FEs’ impact on population health requires valid and reliable FE measures. Methods: A population-based, stratified random sample was recruited from southern Ontario (N=4902 individuals within 2228 households). Socio-demographic data, and self-reported weight, height and waist circumference (WC) were collected from participants. Diet quality was assessed in a subset of participants (n=1170 individuals within 690 households). Neighbourhood FE perceptions were gauged. Seven objective measures characterized the FE of 421 food stores and 912 restaurants in the study region. Euclidean-distance buffers around each household were created at 250m, 500m, 1000m, and 1500m; FE scores from each measure were aggregated within each buffer. Datasets were used to investigate three different research issues. Construct validity of four of the measures was examined using a multitrait-multimethod matrix (MTMM). Multilevel multiple regression analyses determined the extent to which perceptions and objective measures predicted individuals’ body mass index (BMI), WC and diet quality. Mediation analyses were conducted to determine whether residents’ perceptions explained associations between objective measures and outcomes. Results: MTMM results revealed that common FE measures purportedly assessing the same constructs may in fact be measuring different constructs. Perceptions were not highly correlated with objective FE measures. Objective measures (notably food access and food affordability measures) predicted BMI and WC while perceptual variables did not. Mediation analyses findings revealed that perceptions do not mediate associations between objective measures and diet quality, BMI, or WC. Conclusion: MTMM results suggest a method effect, in that what is actually being measured seems to differ by assessment method employed, which has implications for research and practice. Findings may support FE policies or programs focused on objective (rather than perceived) FE features, since objective features better predict weight outcomes and perceptions do not mediate these associations. Specifically, strategies to restrict convenience store access and improve the affordability of nutritious foods relative to non-nutritious foods seem to be supported by these findings.

  • Subjects / Keywords
  • Graduation date
  • Type of Item
  • Degree
    Doctor of Philosophy
  • DOI
  • License
    This thesis is made available by the University of Alberta Libraries with permission of the copyright owner solely for non-commercial purposes. This thesis, or any portion thereof, may not otherwise be copied or reproduced without the written consent of the copyright owner, except to the extent permitted by Canadian copyright law.
  • Language
  • Institution
    University of Alberta
  • Degree level
  • Department
  • Supervisor / co-supervisor and their department(s)
  • Examining committee members and their departments
    • Nykiforuk, Candace (School of Public Health)
    • Farmer, Anna (School of Public Health)
    • Raine, Kim (School of Public Health)
    • Lytle, Leslie (School of Global Public Health, University of North Carolina Chapel Hill)
    • Wild, T. Cameron (School of Public Health)