Our Students are Hungry: Food Insecurity Impacts More Than We Knew, Part 1 — Coming to College Hungry
Influenced by the emergent work of multiple colleagues and the #RealCollege work coming from the Hope Center, I ensured that a core objective of a large-scale research agenda that I led, focused on better understanding how food insecurity and other non-cognitive attributes impact students’ first-year engagement, performance, and persistence. This three-part series highlights the progression from descriptive to causal research that my team (particularly Dr. Fitzpatrick, who is jointly writing these entries) and I conducted.
The current post draws attention to our findings of the descriptive study we titled Coming to College Hungry: Correlating Food Insecurity to Amotivation, Stress, and Engagement, and First-Semester Performance and Persistence in a 4-Year University. The next two posts respectively focus on the inclusion of food insecurity into an already-developed structural equation model (SEM), and whether our nudging random assignment protocol impacted first-year college student performance and persistence — as well as any changes in first-year students’ food insecurity.
Coming to College Hungry: Correlating Food Insecurity to Amotivation, Stress, Engagement, and First-Semester Performance and Persistence in a 4-Year University (Descriptive Study)
Limited research considers first-year college students’ incoming food insecurity. Although research is expanding our understanding on the characteristics of students who report food insecurity, a distinct gap remains both in how food insecurity relates to students’ incoming non-cognitive attributes and also in how being food insecure correlates with first-semester performance and with persistence into a second semester — this study explored such. 
We found that 42% of students reported any level of food insecurity, which is similar to prior research employing a national survey (see graph below) — although, a noticeably lower percentage of students at this institution reported experiencing “very low” (10% v. 22%) food security. This finding may be in part due to the fact that our participants flowed into the site from neighborhoods with median family incomes around $70,000 — higher than national ($60,336) and state ($54,909) averages. Even with the average student here coming from a family who lives in a neighborhood with elevated income, the level of food insecurity remains a concern at this institution.
Next, we wanted to know who comes to the institution with elevated food insecurity — the following factors correlated with reporting higher food insecurity: (1) reporting being multi-racial (B=.57), (2) being an international student (B=.53), (3) being a transfer student (B=.39), (4) self-identifying as LGBTQ (B=.26), and (5) first-generation status (B=.22).
These correlations were unsurprising as each has a foundation in already established research, in other food insecurity and higher education studies more widely examining students’ finances. However, the findings remain important because they contribute to an extremely limited understanding of which factors contribute to or predict elevated food insecurity; an understanding that must be strengthened in order to develop targeted services. 
We were also able to examine linkages between food insecurity and eight other non-cognitive attributes, as captured by previously-tested scales: (1) financial stress, (2) psychological distress, (3) amotivation, (4) conscientiousness, (5) peer-group engagement, (6) faculty interactions, (7) staff interactions, and (8) cognitive engagement. Few research agendas, let alone individual studies, can capture such a robust range of non-cognitive attributes.
Here, food insecurity was associated with increased financial stress (B=.31), psychological distress (B=.15), and amotivation (B=.08), and decreased peer-group interaction (B= -.05). But held no significant connections with staff and faculty engagement or cognitive (academic) engagement.
Food insecurity also connected to lower Fall (first-semester) GPA (B=-.07) and credits earned (B=-.52), and had a marginally-significant relationship with fall-to-spring retention (B=-.02) — when controlling for the other non-cognitive attributes
When food insecurity is included in models predicting first-semester performance and persistence into the spring semester, relationships between other non-cognitive attributes and the outcomes are nearly all non-significant. Our descriptive findings on first-semester performance and persistence into the spring semester seem to suggest that most of these non-cognitive measurements may not matter more than food insecurity to first-semester performance and persistence outcomes. Indicating that interventions addressing students’ non-cognitive attributes — like moderating amotivation or encouraging peer-group interactions — without also easing food insecurity may produce weak or no results for first semester GPA (and probably persistence).
So what? We must do a better job of identifying who comes to college with food insecurity so that we can immediately provide assistance to ease it. As suggested by our data, easing food insecurity should reduce students’ overall anxiety and amotivation, encourage increased peer engagement, and first-semester performance.
Ultimately, our suggestion is simple, low cost, and could be easily rolled out with financial aid packages or via other institutionalized processes, such as when students pay for their deposits. We promote that campuses deploy a survey consisting of the USDA food security scale (6 items) to identify which students come to college with food insecurity — with the intention that stakeholders will flow additional resources (financial, human capital, social) towards easing students’ unmet basic needs and employing already tested psycho-social interventions that illustrate research-supported outcomes.
In most cases, structures aimed at easing food insecurity already exist on campus — including Invisible Need, need-based micro-grants, and accessing SNAP. What tends to be missing is the mechanism to immediately gauge students’ food insecurity and a strong research-based understanding of who comes to college hungry. Due to current limitations in understanding who comes to college hungry, we are missing out on opportunities to more quickly curtail incongruence and provide to our incoming students an opportunity to have an overall better college experience and support their early college performance (and persistence).
Next week we expand upon our findings on food insecurity by highlighting the direct, indirect, and total influences of food insecurity uncovered by Structural Equation Modeling.
Dr. Daniel A. Collier is a Sr. Post-Doctoral Research Associate and Director of Research of SWMU Initiatives at Western Michigan University. Have any questions feel free to email Dan Collier at Daniel.Collier@wmich.edu and follow me on Twitter at Dcollier74. Dan is currently seeking a tenure track role.
Dr. Dan Fitzpatrick is a Post-Doctoral Research Associate at Western Michigan University. Email Dan Fitzpatrick at firstname.lastname@example.org and follow him on Twitter at FitzEdPolicy. Dan is currently seeking research-related roles within Michigan or remotely.
*Email Dan Collier for a working version of the paper.
 Sample size is N=700 (15% of incoming students at to the institution) that mimics the population in neighborhood income ($70,000), urbanicity percentage (82%), and high school free-and-reduced lunch percentage (30%).
 We will also note that according to the adjusted R-squared, our food insecurity model remains relatively weak. The low R-squared illustrates that despite otherwise robust data — as we were able to control for race, high school performance, and school and neighborhood characteristics — modeling correlates to food insecurity remains difficult without student-level access to financially related variables (like EFC). These low R-squared are not uncommon in studies examining student non-cognitive attributes and behavior.