Fact check: How reliable are studies estimating immigrant use of SNAP and what methodologies produce the widest variation?

1. Why results diverge: measurement and definition drive the headlines

One cluster of analyses shows that policy environment measures — for example, classifying counties or states as “sanctuary” or “immigrant‑friendly” — are associated with higher SNAP participation among Latino and low‑income immigrant households, with reported effects described as a 1.1 percentage‑point increase in county‑level participation ^[5] and a 21% higher odds of enrollment in jurisdictions with sanctuary policies ^{[6] [1] [2]}. These two summaries are reconcilable only by noting that one reports an absolute percentage‑point difference while the other reports an odds ratio; researchers who report odds ratios tend to produce larger‑sounding effects when applied to low baseline participation rates. The apparent variation therefore reflects choice of statistical scale and baseline prevalence, not necessarily conflicting empirical realities, and underscores the importance of standardizing effect metrics when comparing studies.

2. Household complexity and mixed eligibility amplify uncertainty

Several analyses highlight that mixed‑eligibility households — poor immigrant households containing both eligible and ineligible members — complicate measurement and interpretation. A policy brief and related summaries report that millions live in such households (5.2 million reported in one brief) and that about 40% of individuals in poor immigrant households reside in households where some members are eligible and some are not ^[3]. Those facts mean administrative records that count household receipt do not map cleanly onto individual eligibility, and survey self‑reports can misclassify eligibility. Researchers choosing household versus individual units, or using administrative caseloads versus survey data, will therefore reach different conclusions about immigrant take‑up and its drivers. The presence of mixed eligibility also makes policy effects heterogeneous within households, so aggregate estimates obscure important distributional differences.

3. Administrative frictions and case processing shape observed take‑up

Analyses of SNAP application processing and experimental evidence on outreach show that paperwork, interviews, and state discretion materially affect who ends up on the rolls. A 2024 Indiana study documents denied and incomplete cases linked to procedural barriers, indicating that application processes can deter eligible immigrants and skew observed participation ^[4]. Experimental interventions from 2018 and 2019 demonstrate that information and assistance raise enrollment but may change program targeting because respondents who sign up after outreach can be less needy than those who sign up without prompting ^{[7] [8]}. This dual effect — raising participation while altering caseload composition — creates wide variation in measured take‑up depending on whether studies account for application denials, administrative churn, or the impact of outreach programs.

4. Policy design and state discretion create genuine geographic heterogeneity

Multiple summaries link state and county policy discretion to differences in access, implying that part of the variation across studies is substantive rather than methodological. Researchers find higher SNAP participation in immigrant‑friendly jurisdictions ^{[2] [1]}, and other work emphasizes how state rules shape eligibility pathways for mixed households ^[3]. These findings mean that even perfectly harmonized methods would still detect geographic variation driven by real policy differences. At the same time, the focus on Latino households in several studies signals a potential scope limitation: results may not generalize across immigrant populations with different legal statuses, ethnic composition, or geographic concentration, and some research points to differential food‑security effects by race and ethnicity during shocks like the COVID‑19 pandemic ^[9].

5. What produces the widest variation: scale, unit, and treatment of eligibility

The greatest divergence in estimates arises when studies differ on three axes: the statistical scale (odds ratios versus percentage points), the unit of analysis (individual versus household versus county), and the treatment of eligibility (counting administrative receipt, self‑reported receipt, or modeled eligibility in mixed households). Comparative summaries in the provided analyses show precisely these trade‑offs: policy‑environment studies report both percentage‑point and odds‑ratio effects ^{[1] [2]}, household‑level briefs emphasize mixed eligibility counts ^[3], and administrative‑processing studies highlight denials and case status dynamics ^[4]. Any single study’s headline number must therefore be read alongside its methodological choices to understand why it differs from others.

6. Bottom line for policymakers and analysts: triangulate, standardize, and report

Given the documented sources of variation, reliable conclusions require triangulating across administrative data, surveys, and natural‑experiment policy comparisons while reporting both absolute and relative effects and explicitly modeling mixed‑household eligibility. The provided literature collectively shows that methodological transparency — including metrics, units, and administrative adjustments — explains most of the inter‑study dispersion ^{[2] [1] [3] [4]}. Analysts should therefore treat single studies’ point estimates as conditional on design choices, use multiple effect metrics, and prioritize replication across contexts to produce robust estimates of immigrant SNAP use.