Feasibility and reliability of digital imaging for estimating food selection and consumption from students’ packed lunches

Publication date: 1 January 2018
Source:Appetite, Volume 120
Author(s): Jennifer C. Taylor, Carolyn Sutter, Lenna L. Ontai, Adrienne Nishina, Sheri Zidenberg-Cherr
Although increasing attention is placed on the quality of foods in children's packed lunches, few studies have examined the capacity of observational methods to reliably determine both what is selected and consumed from these lunches. The objective of this project was to assess the feasibility and inter-rater reliability of digital imaging for determining selection and consumption from students' packed lunches, by adapting approaches previously applied to school lunches. Study 1 assessed feasibility and reliability of data collection among a sample of packed lunches (n = 155), while Study 2 further examined reliability in a larger sample of packed (n = 386) as well as school (n = 583) lunches. Based on the results from Study 1, it was feasible to collect and code most items in packed lunch images; missing data were most commonly attributed to packaging that limited visibility of contents. Across both studies, there was satisfactory reliability for determining food types selected, quantities selected, and quantities consumed in the eight food categories examined (weighted kappa coefficients 0.68–0.97 for packed lunches, 0.74–0.97 for school lunches), with lowest reliability for estimating condiments and meats/meat alternatives in packed lunches. In extending methods predominately applied to school lunches, these findings demonstrate the capacity of digital imaging for the objective estimation of selection and consumption from both school and packed lunches.



http://ift.tt/2xXyCcn