National School Lunch Program Eligibility and Its Relationship With NAEP Performance on the National and State Levels

• In Event: 25.065 - Community Eligibility and National School Lunch Program Status: A View From NAEP

Abstract

There has been a considerable concern in the research and policy community about the potential for a sudden decrease in the validity of NSLP as an SES proxy for NAEP in 2015 because of the implementation of the Community Eligibility Provision (CEP) in 2014-15. Regarding this question, two hypotheses were investigated regarding NAEP NSLP data. The data sources were the NAEP mathematics assessment data between 2009 and 2017 at grades 4, 8 and 12.
Hypothesis 1: Because of the implementation of the CEP in 2014-2015 where all students in a school or district can be identified as eligible for NSLP if certain criteria are met, there should be a significant increase in the percentage of students identified as NSLP eligible. Moreover, the share of NSLP eligible students whose eligibility status was determined school wide (rather than individually) should also see an increase in 2015 as compared to previous years.
This hypothesis was confirmed. At the national level, a steady growth in the percentage of students identified as NSLP eligible over time at all three assessed grades can be identified (e.g., from 45% in 2009 to 51% in 2017 at grade 4). The percentage of school-wide NSLP eligible students also increased (e.g., from 10% in 2011 to 33% in 2017 at grade 4).
Hypothesis 2: Given that a certain proportion of students who are not poor are reported as NSLP-eligible due to CEP, a diminished validity of the NSLP variable as an SES proxy would express itself in: first, a decrease in the performance gap between NSLP eligible and non-eligible students, and second, a decrease in the association between NSLP status and NAEP mathematics performance in terms of variance explained by the former.
The second hypothesis was not confirmed. Despite the changes caused by CEP, neither of the two above-mentioned trends is present at the moment. Over five NAEP administrations (2009, 2011, 2013, 2015, and 2017) at grade 4, for example, the performance gap between eligible and non-eligible students has remained about the same, ranging from 22.6 to 24.4 NAEP scale score points, and the association between NSLP status and NAEP mathematics performance has also remained virtually unchanged with a R-squared value between 0.14 and .15 across the four NAEP administrations.
Therefore, at the national level, NAEP data do not support the concern of a sudden decrease in the validity of reporting NAEP results by NSLP eligibility. At the state level, only few states showed a decreased association between NSLP status and NAEP scores, while in most states the associated remained the same or even saw an increase. Of course it is without dispute that in certain counties or school districts that have adopted CEP-ALL throughout their jurisdiction, no differentiation can be made any longer between students using this SES proxy since 100% of students in these jurisdiction are considered NSLP eligible. The importance of this paper is to present a data based counterargument to a wide-spread concern about the continued relevance of NSLP as an SES proxy.

Authors

• Markus Broer, American Institutes for Research

• Jing Chen, National Center for Education Statistics

• Qingshu Xie, MacroSys LLC