FNS RESEARCH CORNER

Please note that this study was published before the implementation of Healthy, Hunger-Free Kids Act of 2010, which went into effect during the 2012-13 school year, and its provision for Smart Snacks Nutrition Standards for Competitive Food in Schools, implemented during the 2014-15 school year. As such, certain research may not be relevant today.

Summary Of Research

 The FNS Research Corner provides a continuing series of summaries of recently completed and current research conducted by the U.S. Department of Agriculture’s (USDA) Food and Nutrition Service (FNS) in the area of child nutrition. For further information, contact the Office of Analysis, Nutrition, and Evaluation (OANE) at (703) 305-2117. Links to published studies and reports, as well as descriptions of ongoing studies conducted by OANE, are available from the FNS Internet web site at http://www.fns.usda.gov/oane/.

Recently Completed Research

Data Matching in the National School Lunch Program: 2005 Background

This study examined the feasibility of expanding the use of computer matching for certification and verification of children eligible for free and reduced-price meals funded under the National School Lunch Program (NSLP). It provides a detailed description of how computer matching is used currently and how it could be used. It addressed the feasibility and effectiveness of different practices, from the point of view of both state and local agencies. Computer matching for the NSLP has the potential to improve the efficiency and integrity of the certification and verification processes without deterring eligible households from applying for free or reduced- price meals. The Child Nutrition and WIC Reauthorization Act of 2004 (P.L. 108-265) mandated direct certification of children in Food Stamp households, to be phased in over three years beginning in 2006.

Study Methods

The study collected data through four activities: 1) A panel consisting of five experts in the fields of information technology, computer matching, and data processing prepared papers to address sources of data for matching with student information to determine or verify eligibility, computer matching processes, data acquisition methods, matching algorithms, and privacy issues; 2) site visits were conducted in New Jersey and Nebraska in September 2004 to interview school food authorities (SFA) and certain state agencies, including those administering Child Nutrition, Education, Food Stamps, Labor, and Medicaid programs; 3) surveys in all states were conducted in late Summer 2005 with state Child Nutrition, Education, and Medicaid agencies, to collect information about current computer matching practices and capabilities for the NSLP and other education programs; and 4) in-depth telephone interviews were conducted during Winter 2005 with state and local agencies in six states that used a variety of approaches to computer matching.

Prevalence of Matching for Certification

The percentage of public school districts using direct certification appears to have been nearly constant over the past ten years: 63% in 1996-97, 61%in 2001-02, and 63% in 2004-05.

However, the percentage of free certified students who were directly certified rather than certified through application rose from 17.9% in 2001-02 to 28.1% in 2004-05.

There are three main methods used for direct certification: matches performed at the state level, matches performed at the school district level, and the letter method. State-level and district-level matches involve linking two databases: children in the Food Stamp Program (FS) or Temporary Assistance for Needy Families (TANF) households and students enrolled in school. With the letter method, letters are mailed to FS/TANF households, which parents return to the child’s school, rather than completing an application. Over time, the number of states providing state- level computer matches has increased. From 1996 to 2004, the number of states providing state- level computer match results increased from 13 to 18 and the number of states providing data to school districts for matching increased from 18 to 22.

State-level matching, implemented state-wide, is the most effective method of direct certification, reaching 75% of all FS/TANF children receiving free meals. States implementing state-level matching, but not requiring school districts to participate, only certified 51% of all FS/TANF children receiving free meals. District-level matching matched 63% of FS/TANF children receiving free meals, and the letter method reached 52%. The best computer matching systems for certification were those using timely records from FS/TANF and student information systems, obtaining accurate matches, distributing match results to the correct districts, and providing a mechanism for directly certifying unmatched children. School districts are required to verify a sample of NSLP applications by obtaining documentation to confirm the income or program participation reported by the sampled households. Verification is typically conducted by requiring households to provide documentation. School districts also use collateral contacts.

Direct verification is the process of verifying approved applications using income and program information from a public agency administering a means-tested program. Many school districts contact the local FS/TANF office to verify program participation. However, in 2004-05, only four states had an automated system for school districts to verify case numbers. Another 11 states were investigating options for direct verification using computer matching.

Medicaid data is a potential source of income information for low-income children not participating in FS/TANF. Medicaid and state Child Health Insurance Program (SCHIP) income eligibility limits are higher than FS/TANF limits, and thus data from Medicaid and SCHIP can potentially verify a larger percentage of NSLP income application than FS/TANF data. Medicaid and SCHIP have good potential for matching because:

• Most states have maximum income eligibility limits that meet or exceed the limits for reduced-priced meals;
• Social Security Numbers (SSN) are required for Medicaid and most states have SSNs for the majority of SCHIP children; and
• 34 states have statewide Medicaid/SCHIP databases that can be used to attempt to verify all free and reduced-price applications, and five states have databases that can be used for all free and some reduced-price applications.

However, substantial challenges remain, including Medicaid’s definitions and recordkeeping on income, as well as identification of the assistance unit. Another challenge to using Medicaid/SCHIP data for direct verification is securing the active participation of the state Medicaid/SCHIP agency.

Conclusions

Computer matching systems for NSLP direct certification and verification are currently in place in a number of states. This study found considerable variation in the methods and effectiveness of direct certification across states, suggesting that it may be possible to increase effectiveness in some states and thus expand direct certification to more food stamp or TANF participants. The study also found that, as of 2005-06, direct verification was used almost exclusively for categorical applications, although a few states were beginning to develop or use systems for direct verification of income applications.

Accuracy of SFA Processing of School Lunch Applications – Regional Office Review of Applications (RORA) 2006

This is the second in a series of annual reports that examined the administrative accuracy of school food authority (SFA) approval and benefit issuance for free and reduced-price meals based on household applications. It does not examine the accuracy of household reporting of information on the applications or errors made in counting and claiming. (These issues will be examined in a larger, nationally representative study, The NSLP/SBP Access, Participation, Eligibility and Certification Study.) Similar to last year, a stratified two-stage cluster sample design was used in which school districts were stratified into 28 strata defined by seven FNS regions and four size categories within each region. A total of 2,751 applications were selected for review. An SFA’s determination of household size, total gross income, and certification status (free, reduced-price, paid) that it assigned to the selected student was recorded. FNS Headquarters staff reviewed each application and made an independent assessment of household size, total gross monthly income, and certification status, based on the information on the applications. SFA determinations were compared with FNS’ independent assessments.

Key Findings:

• Administrative errors are rare on applications that were approved based on categorical eligibility. Less than one percent of categorically eligible applications were processed
• Applications that are approved based on household size and income are more prone to administrative errors. SFAs are more accurate in determining household size (97%) than they are in determining monthly income (92%). Common errors in computing household size included not counting or double counting the student in the total household size. Common errors in the calculation of gross monthly household income include using the wrong monthly income conversion factor, incorrectly determining the frequency of receipt of household income, or incorrect addition or multiplication.
• SFA eligibility determinations were incorrect for 3% of students approved or denied on the basis of an The percentage of errors is slightly higher (3.4%) for students approved or denied on the basis of income-based applications.
• Accuracy of meal benefit status was slightly lower than the accuracy of eligibility determination at the time of Meal benefit status was correct for 96.2% of the students.
• The percentage of students incorrectly approved or denied for NSLP free or reduced- price meal benefits has remained relatively In SY 2005-06, the percent of students incorrectly certified was 3% compared to 3.5% the previous year. The percentages of incorrectly certified students that were over-certified and under-certified remained relatively constant across the two years at about 83% and 17%, respectively.

Research In Progress

The following section provides a brief description of some ongoing FNS research and the current status of these studies.

The School Nutrition Dietary Assessment Study III (SNDA-III)

As part of FNS’ periodic assessment of the nutritional effects of school meals, this study is designed to provide policy makers with updated information on four domains: 1) characteristics of the school food environment and school foodservice operations; 2) student participation, participant characteristics, satisfaction, and related attitudes toward the school lunch and breakfast program; 3) nutritional quality of meals offered and served in the school meal programs; and 4) student dietary intakes and the contribution of school meals to these dietary intakes. The last SNDA study, published in 2001, only examined the nutritional quality of meals offered not consumed.

Data were collected from school officials, students, and parents of students in a nationally representative sample of 398 schools in 129 school districts during the spring of the 2004-05 school year. School food authority (SFA) directors and school food managers were asked to provide data on foodservice characteristics and meals offered. Descriptive information about district-level operations and meal service characteristics were collected through telephone interviews with SFA directors. Foodservice managers in sampled schools reported information on foods offered at all breakfasts and lunches during a specified five-day target week. In addition, data on the types of food offered on an a la carte basis for one day of the target week were collected. A random sample of about 2,400 students enrolled in 300 of the participating schools was selected and interviewed about their food consumption over a 24-hour period.

Younger children were assisted by their parents in these interviews. Parents also supplied important background information on household characteristics. Analysis of the data is currently underway. A full report presenting the results of this study is expected in 2007.

The National School Lunch and Breakfast Programs: Access, Participation, Eligibility, and Certification Study (APEC)

FNS is undertaking a nationally representative study to provide information about children’s access, participation, eligibility, and certification in school nutrition programs to help the U.S. Congress and USDA improve program integrity and ensure that intended recipients have access to appropriate meal benefits. This study will look at the application, certification, and verification process to identify reasons that some families do not participate, difficulties they experience in applying, and sources of erroneous reimbursements (both under- and over-payments) due to an administrative error, household misreporting or errors in meal counting and claiming. FNS will use the study results to provide guidance to school districts and schools on how to enhance program administration and target benefits effectively.

Data were gathered from a nationally representative sample of elementary, middle, and high schools during the 2005-06 school year. A total of 266 schools in 87 school districts participated in the study. In-person interviews were administered to more than 3,000 households. More than 7,800 applications were reviewed. To meet the requirements of the Improper Payments Information Act of 2002, a final report will provide national estimates of erroneous payments due to certification errors and meal counting and claiming errors. This report is expected to be completed in 2007.

School Lunch and Breakfast Cost Study II

Since the last meal cost study was completed in 1994, the landscape of school foodservice has changed. School food authorities (SFAs) contend that they are now spending more to serve meals that meet the nutritional standards established by the School Meals Initiative for Healthy Children. Rising labor costs and other factors have made it difficult for foodservice managers to meet program requirements within the limits of existing reimbursement rates. Demands for educational improvement and other mandates have led school districts to minimize subsidies to SFAs, which increasingly have relied on other sources of revenues, including a la carte sales, vending machines, and other food sales outside mealtimes. In light of all these changes, there is a critical need for updated information on the average cost of producing reimbursable school lunches and breakfasts, including indirect and local administrative costs, using a direct measurement approach comparable to the methodology used in the last study of meal costs conducted over ten years ago. The methodology relies on the direct measurement of costs attributable to various SFA activities, rather than the use of indirect allocation rules.

A nationally representative sample of 122 school districts has been recruited to participate in this study. Onsite data collection occurred in about three schools per district during the 2005-06 school year and will be used to allocate the SFAs’ reported costs between lunch and breakfast production and between reimbursable and non-reimbursable meals. Review of the SFAs’ final SY 2005-06 annual financial statement and the school district’s indirect cost allocation, as well as identification of unreported costs attributable to school foodservice operations were conducted in Fall 2006. A final report is expected in Fall 2007.

Evaluation of Direct Verification

Currently, local education agencies must pull a sample of household applications for free or reduced-price meals, contact the households, and verify the facts of the application. This process, known as household verification, is burdensome on both schools and households. Often, households do not respond. When this happens, the child and the school lose the free or reduced- price subsidies, even though the household could be eligible. Direct verification, in which

National School Lunch Program (NSLP) applications are computer-matched to other means- tested programs, has the potential to avoid this consequence, increasing household access to school meals and reducing burden on parents and schools. Indiana, Oregon, South Carolina, Tennessee, and Washington are conducting demonstrations of direct verification using Medicaid data. The researchers conducting the Evaluation of Direct Verificationstudy are working with these states to assess the feasibility of these demonstrations to verify school lunch applications and document the lessons learned. The evaluation is due to the U.S. Congress by July 2007.

Evaluation of the Simplified Summer Food Program Pilot 2001-06

Twenty-seven states with lower-than-average participation rates in the Summer Food Service Program currently participate in the Simplified Summer Food Program Pilot. Under the pilot, meals served by eligible sponsors are reimbursed at the maximum allowable rate. In addition, administrative record keeping is reduced and sponsors no longer have to report costs to state agencies. The U.S. Congress has mandated an evaluation of the expanded Simplified Summer Food Program Pilot that examines the effects of the pilot on the participation of both sponsoring organizations and children. At press time, an in-house evaluation report that uses administrative data and information from telephone surveys of state agencies was being prepared for submission to Congress by April 30, 2007.