Erroneous Payments In Childcare Centers Study (EPICCS)

<b>Title</b>Erroneous Payments In Childcare Centers Study (EPICCS)<br><b>URL</b>These data are not published elsewhere. The public webpage that provides information about the study is https://www.fns.usda.gov/research/cacfp/epiccs.<br><b>Description of the experiment setting: location, influential climatic conditions, controlled conditions (e.g. temperature, light cycle)</b><br>These data are from the Erroneous Payments In Childcare Centers Study (EPICCS) study and EPICCS Adjustment study. The study methods are summarized at https://www.fns.usda.gov/research/cacfp/epiccs.<br>These studies examined payment errors and compliance with USDA meal pattern requirements for breakfast and lunch meals in a nationally representative sample of CACFP child care centers serving 2- to 5-year-old children in 2017. The studies included sponsored centers, independent centers, and Head Start centers. The studies did not include infant care centers, family day care homes, at-risk afterschool centers, outside school hours centers, adult care centers, or emergency and homeless shelters.<br><b>Processing methods and equipment used</b>N/A<br><b>Study date(s) and duration</b>EPICCS data were collected between March 2017 and March 2018. EPICCS Adjustment data were collected between April and June 2023 but are about 2017.<br><b>Study spatial scale (size of replicates and spatial scale of study area)</b>The intended study universe was CACFP child care centers serving 2- to 5-year-old children. The study data (for the study sample) are accurate. However, due to a weighting error, the national estimates are inaccurate. Please read our disclaimer about the accuracy of some of these data, and a summary of the study methods, at https://www.fns.usda.gov/research/cacfp/epiccs.<br><b>Level of true replication</b>N/A<b>Sampling precision (within-replicate sampling or pseudoreplication)</b>EPICCS used a multistage-clustered sample design as follows: (1) A representative sample of 25 States in the contiguous 48 States and the District of Columbia; (2) a representative sample of 450 childcare centers (and their sponsors) within Primary Sampling Units (PSUs); and (3) a random sample of 5,400 households (with children ages 2-5 years) enrolled in the sampled childcare centers who applied for free and reduced-price meals or were categorically eligible for free meals.<br><b>Level of subsampling (number and repeat or within-replicate sampling)</b><br><b>Study design (before–after, control–impacts, time series, before–after-control–impacts)</b>Cross-sectional<br><b>Description of any data manipulation, modeling, or statistical analysis undertaken</b><br>The EPICCS data were analyzed to produce sample-level descriptive statistics about sources of payment error, sources of noncompliance with USDA meal pattern requirements, and error rates. EPICCS Adjustment data (about the sampled centers’ meal counting methods) were used to re-analyze certification error (a source of payment error). Please read a summary of the study methods at https://www.fns.usda.gov/research/cacfp/epiccs.<br>The sample-level data were weighted to produce national-level estimates, but due to a weighting error these estimates are inaccurate.<br>To preserve the confidentiality of respondents, variables that could be used to identify sponsors, centers, or households were suppressed in the public-use files.<br><b>Description of any gaps in the data or other limiting factors</b>Please read our disclaimer about the accuracy of some of these data at https://www.fns.usda.gov/research/cacfp/epiccs.<br>EPICCS was sampled on accurate criteria. However, the EPICCS and EPICCS Adjustment study final post-stratification weights were raked to inaccurate population parameters using FNS keydata. FNS keydata for CACFP child care centers include emergency shelters and at-risk afterschool programs in addition to child care centers (see the instructions of form FNS-44 https://www.fns.usda.gov/cacfp/fns-44). As a result, the population parameters that the weights were raked to were significantly different than the actual universe being analyzed in EPICCS. Because of the way CACFP data are reported, at-risk afterschool programs were all included in the population as “non-Head Start, not-for-profit outlets” leading to a severe over-weighting of these types of outlets (and children attending them) and under-weighting of Head Start and for-profit outlets. This is particularly notable because certification error is not a source of payment error in Head Start centers; children are categorically eligible for free meals, so FNS reimburses at the free rate without the need for Head Start centers to certify applications. As a result, the calculations using the incorrect weights very likely overestimated certification error and resulting payment errors. FNS chose not to publish these estimates in keeping with USDA scientific integrity guidelines. These weights are left in these data files for transparency.<br>In the EPICCS Adjustment data, significant imputation of claiming methods was required due to non-response to the survey. More information is provided in the technical manual accompanying the data files.<br><b>Outcome measurement methods and equipment used</b>Data collection instruments included surveys of sampled sponsors and centers, meal observation booklets, abstraction of applications and meal counts, and household interviews.