Administrative data is often used to establish eligibility for means-tested programs. For example students who appear in administrative records of Supplemental Nutrition Assistance Program (SNAP) data may be directly certified for school meal benefits. For programs such as these, improving data matching effectiveness can expand access to benefits and improve program equity. We used State SNAP data and National School Lunch Program (NSLP) application data from seven States to explore factors associated with effective administrative data matching for NSLP direct certification. Findings from the analysis have implications for improving access to school meal benefits, as well as to other programs using administrative data for eligibility determination.

In the first stage of our study, we analyzed statewide SNAP participant data from two participating States, comparing characteristics of children who were directly certified for school meal benefits and those who were not. These comparisons identified patterns in age, name commonality, local area school characteristics, and economic characteristics associated with more successful or more challenging direct certification.

In the second part of the study, we collected data on children certified for school meal benefits based on program participation through an application (rather than direct certification). These data, drawn from randomly sampled districts in all seven participating States, represent students who could have been directly certified for school meal benefits but were not. We matched these students with their respective State SNAP participant lists using a two-stage matching process that simulated data matching processes that may be used in NSLP direct certification. In the first stage, we conducted a deterministic match, requiring exact matches on key variables such as name and date of birth. This process mirrored the deterministic processes many States use for direct certification. In the second stage, we used a probabilistic match that incorporated more flexible algorithms and allowed inexact matches between data fields—a more sophisticated process that is not typically used in NSLP direct certification.

The results of this independent matching process indicate the extent to which students categorically eligible for NSLP benefits can be identified in SNAP participation data. By comparing the characteristics of students matched deterministically, those matched probabilistically, and those not matched in either process, we identified characteristics associated with more challenging matches to administrative records. These findings allow us to evaluate the value of data completeness, the extent to which certain types of data—such as uncommon or misspelled names—present challenges to matching, and the potential for probabilistic matching to improve program access.