Evaluating the outcomes of food assistance programs comes with high costs, especially if program settings are geographically dispersed, as they are in home settings. If program administrative data are of sufficient scope and quality, they will support some types of evaluations. In other cases, low-cost ways of supplementing administrative data may exist. The latter better describes the case of evaluating the USDA Food and Nutrition Service’s Child and Adult Care Food Program (CACFP). This paper examines the
efficacy and limitations of using administrative data to evaluate CACFP as it applies to family child care homes.
CACFP reimburses child care providers for serving meals and snacks to a daily average of 2.9 million children. About 850,000 of these children receive care in the homes of child care providers, often called family child care homes (FCCs). CACFP payments to FCCs can be substantial. In Illinois, they can equal as much as 39 percent of the State’s reimbursement for full-time child care under Illinois’s child care subsidy program.
FCCs participate in CACFP through local sponsor organizations that have an agreement with the State authority responsible for administering CACFP. Sponsors recruit, train, monitor, and authorize payments to FCCs. Monitoring by sponsors consists of reviewing menus and making three visits to each FCC annually, including unannounced visits.
CACFP has the potential to grow substantially. Some 6.5 million children under the age of 6 receive care in FCCs. Of these, 39 percent have incomes below 200 percent of the Federal poverty level. Among school-age children, at least 1.5 million (and as many as 10 times that number) receive regular care in FCCs.
CACFP home provider participation largely consists of licensed FCCs regulated by State agencies. Yet the majority of FCCs in the United States are home providers who are not licensed and of whom, as a group, little is known. Families with lower incomes are more likely to use this license-exempt child care. Only some States reimburse license-exempt FCCs under CACFP. Little is known about the outcomes.
After describing FCCs in the United States, reviewing basic features of CACFP, and discussing the role of CACFP in FCCs, the study presents the administrative data that an evaluator might use for various outcome evaluations. It summarizes several evaluation questions and discusses possible research designs appropriate to measuring program outcomes, including nutrition, nutrition knowledge, food behaviors, FCC and parent food expenditures, and the cost of complying with monitoring. It discusses to what extent evaluators can rely on administrative data for this research and identifies what additional data evaluators would need to collect.
Data sources include administrative forms used by CACFP sponsors and interviews with program staff and specialists.
Basic administrative data in CACFP, including menus, allow evaluators to conduct descriptive studies, such as geographic coverage of FCCs by CACFP or basic foods served. The study discusses administrative data storage and data quality differences across CACFP sponsors and other limitations of administrative data.
Several lower cost ways of supplementing administrative data exist, including two in particular. First, supplementing CACFP administrative data with other data sets, such as census data and administrative data from other programs that subsidize, license, or train FCC providers, make more sophisticated evaluations possible. For example, they make possible the analysis of differences in takeup rates by types of providers across geographies, including those with identifiable income, racial, or ethnic characteristics.
Second, it might be possible to train the FCC home-visit staff of CACFP sponsors to collect observational or survey data during their normal visits extended by a few minutes. The cost of this form of data collection with the case of a Chicago sponsor paying for extending each of three annual FCC visits by 15 minutes was estimated at $32.64.
Collecting pretreatment data or data for a control group of FCCs will typically cost an evaluator substantially higher amounts. (Exceptions include less typical evaluations where both treatment and control groups can be drawn from among CACFP participants or where administrative data contain
sufficient pretreatment data.) Continuing with the example of using sponsors’ FCC home-visit monitoring staff to collect these data, the study concludes that collecting pretreatment data from the treatment group will typically require at least making an additional home visit and possibly several visits. Similarly, collecting comparable data from a control group will generally require additional home visits. The study estimates that collecting data during each additional visit will cost 2.5 times the cost of extending the duration of three annual home visits (approximately $80 in Chicago).
Additional costs include the costs of using specialists or special materials and equipment. Other substantial costs of making additional home visits include inconveniencing the child care provider, an inconvenience that might need to be compensated to avoid nonparticipation or attrition.
Direct inquiries about this study to the Project Contact listed above.