This paper assesses the quality of survey-based reporting of SNAP participation by using linked administrative records from three states, Illinois, Maryland, and Virginia, covering the years 2009-2012. We compare reports of SNAP participation from the 2008 panel of the Survey of Income and Program Participation (SIPP) with SNAP administrative records, evaluating the consistency of reported SNAP participation across the two data sources and identifying characteristics of individuals associated with conflicting reports.

We find evidence of a slight over reporting of SNAP participation in the SIPP for the combined states and years evaluated. Though the vast majority of responses in the survey align with the administrative data, approximately 4.5 percent of the observations can be classified as either a false positive or a false negative. False positives, defined as occurring when a survey respondent reports that they receive benefits, but do not appear in the administrative data, are more likely to occur for the foreign born relative to those born in the U.S, for non-Hispanic Blacks and Hispanics than for non-Hispanic Whites, and for unmarried adults and married adults without children. Conversely, rural residents, those with less than a high school education, women, unmarried parents, and the unemployed are more likely to have false negative responses (appearing as a beneficiary in the administrative data, but not reporting receipt in the survey). Both false positives and false negatives are more likely to occur when survey data on SNAP participation are imputed, raising the possibility that current imputation methods for this survey could be improved by incorporating administrative data.