What the Illinois Evidence Actually Measures
The subject here is the evidence architecture behind the Illinois CACFP study: which records identify reimbursement status, which survey answers describe a home daycare's operation, and how the two are compared over time. The specific task is to decide what an estimate about “more childcare” actually means when the data record offered services, reported rates, and quality proxies rather than every family transaction. Understanding that distinction lets readers judge both the study's causal confidence and the reach of its conclusions.
Two records, two kinds of evidence
The study joins two sources that answer different questions. Annual recipient lists from the Illinois State Board of Education cover 2003 through 2017. They identify CACFP daycare-home providers, record addresses and participation modes, and link providers over time with unique identifiers. These records establish the reimbursement state: whether a provider was in school, census, categorical, all-high, mixed, or all-low mode.
The second source is the Illinois Market Rate Survey, or MRS, collected in the years used for the analysis from a broader sequence of survey waves. The federal government requires states to conduct a biennial survey for the childcare subsidy system. In Illinois, regional Child Care Resource and Referral agencies combined administrative records with telephone interviews. The MRS asks providers about the operation they offer, including food service, age categories, shifts, opening and closing times, prices, and selected quality information.
The distinction is crucial. The recipient list can tell the researchers that a provider's reimbursement status changed. It cannot tell them whether that provider began offering dinner or closed later. The MRS can describe those services. It cannot by itself determine whether the provider qualified for the high schedule, whether a reported meal was actually served to an enrolled child, or whether a listed price was paid by a particular family. Linking the sources creates the policy-to-service connection, but it does not make either source more precise than it is.
The researchers then geocode the providers' addresses and place them within National Center for Education Statistics school-district boundaries. That step supplies local context: child-poverty rates and a locale classification such as city, suburb, town, or rural area. It also connects the reimbursement transition to the geographic unit used for the comparison. The policy is tied to school boundaries, while the available boundary information is organized at the school-district level; that mismatch will matter when the design's limits are considered.
Who enters the comparison
The study does not begin with every licensed provider in Illinois. In 2009, the recipient records contained 9,865 CACFP daycare-home providers, and in 2011 they contained 9,515. Matching those records to active providers observed in the Market Rate Survey produced smaller year-specific groups. Further losses came from missing school districts, locale codes, or outcome variables. This narrowing is not a flaw by itself. A causal comparison often needs a well-defined population. But every restriction changes the population to which the estimate most directly applies.