Measuring food insecurity among college students

Nikolaus, Cassandra J

Permalink

https://hdl.handle.net/2142/105018 Copy

Description

Title

Measuring food insecurity among college students

Author(s)

Nikolaus, Cassandra J

Issue Date

2019-04-19

Director of Research (if dissertation) or Advisor (if thesis)

• Nickols-Richardson, Sharon M.
• Ellison, Brenna

Doctoral Committee Chair(s)

Engeseth, Nicki

Committee Member(s)

• An, Ruopeng
• Arthur, Anna

Department of Study

Food Science & Human Nutrition

Discipline

Food Science & Human Nutrition

Degree Granting Institution

University of Illinois at Urbana-Champaign

Degree Name

Ph.D.

Degree Level

Dissertation

Date of Ingest

2019-08-23T20:35:53Z

Keyword(s)

• food security
• food insecurity
• college students
• university students
• psychometric
• qualitative
• quantitative

Abstract

Background. Mounting evidence suggests that college (i.e., postsecondary) students are at a greater risk of experiencing food insecurity (FI) – the lack of sufficient access to adequate food for a healthy and active life – than the general U.S. population. However, there is a paucity of psychometric evaluations of common FI assessment surveys for this population. Objective. To understand if current FI surveying methods are appropriately capturing and classifying U.S. undergraduate college students. Methods. The existing literature was systematically reviewed and 51 study samples were analyzed using meta-analysis techniques. Specifically, an overall weighted estimate of college FI was produced and the impact of study characteristics (e.g., sampling strategy, survey type, and reference period) on FI estimates were investigated. In addition, an online survey was sent to a randomized sample of 4,000 undergraduate students to characterize response patterns on FI questionnaire items. Psychometric analyses were conducted on the 462 student responses to compare varying survey protocols and evaluate item response statistics. Among students who responded to the online survey and agreed to be contacted for future studies, a random subset was invited to participate in the in-person phase. An exploratory analysis was conducted with students (n=66) who participated in the in-person phase and provided responses on FI questions using paper-and-pencil surveys. This repeated-measures study sample was used to test the impact of survey modality (online or paper-and-pencil) on FI responses, and resulting FI prevalence rates. A subset of this sample (n=33) students participated in in-person qualitative cognitive interviews to understand how they interpreted questionnaire items and how students responded to experiences of FI. The transcripts of these interviews were analyzed using a basic interpretative approach. Results. The meta-analyses indicated that an estimated 38% (95% CI: 33%, 43%, I2: 99.7%) of U.S. college students experience FI. However, the type of survey and the reference period used in surveys were significantly related to the resulting FI rate, producing estimates ranging from 17% to 46% across the study sub-groups. Psychometric analyses of the online survey revealed students’ response patterns on FI questions deviated from expectations. Item infit and outfit statistics reflected poor discrimination for many of the questions. Each surveying protocol tested had low model fit, but the 10-item Food Security Survey Module (FSSM) paired with a 2-item screener had the best fit (McFadden’s R2=0.15 and Bayesian Information Criterion=-2049.72) compared to shorter surveys or those that did not include a screener. When students answered items on the 10-item FSSM and the 2-item screener, they were less likely to affirm each item on a paper-and-pencil survey, when compared with the online modality. Therefore, paper-and-pencil surveys produced a 25.8% FI prevalence compared to 40.9% FI estimated from the online survey. The cognitive interviews provided rich data that may explain the psychometric issues with the questionnaire items. Specifically, students had unique interpretations of key terms – such as “money for more” and “balanced meals” – used in FI surveys. Conclusion. Overall, this body of work illustrates several shortcomings of current FI surveying methods used with college students. Further psychometric tests on modified FSSMs, or new FI surveys, will be needed for this population to ensure students are not misclassified and that students experiencing FI are best supported.

Graduation Semester

2019-05

Type of Resource

text

Permalink

http://hdl.handle.net/2142/105018

Copyright and License Information

Copyright 2019 Cassandra Nikolaus

Owning Collections