About the Data

The Civil Rights Data Collection in the U.S. Department of Education (prior to 2004, the Elementary and Secondary School Survey) is used by the Office for Civil Rights (OCR) to obtain data regarding access to elementary and secondary schools and to programs or services within those schools. This Web site contains Civil Rights Data Collection data for the 2006-2007 school year (2006 Civil Rights Data Collection). We encourage you to complete the Web site tutorial on how to select which data you want, customize its presentation, and save it in a file.

In order to understand the data clearly, it is important for you to be familiar with the definitions used. There are differences in definitions in some data items between the 2006 Civil Rights Data Collection and previous surveys. Most differences in definitions are minor; others are more substantive; and there are definitions of new items.

The Common Core of Data (CCD) database maintained by the National Center for Education Statistics (NCES), U.S. Department of Education is used as the universe for identifying public school districts for participation in the Survey. Sampling methodology is a statistical process that is used on a state by state basis to ensure that a representative sample of school districts is included in each state sample. Representative sampling ensures a balance between four different sizes of school districts; i.e., 1-300 students, 301-3,000 students, 3,001-5,000 students, and 5,001-25,000 students. For this reason, some districts in some size "strata" in some states may have a higher probability of being sampled frequently, in order to provide appropriate data in each size stratum to ensure a representative, accurate set of statewide projections. Additionally, the following types of districts are sampled with certainty in every Survey cycle: districts having more than 25,000 students; all districts in states having 25 or fewer public school districts; and districts subject to, or recently released from, a Federal Court Order and monitored by the U.S. Justice Department.

The statistical projections presented on this Web site are derived from the 2006 Civil Rights Data Collection (CRDC) conducted by the Department of Education Office for Civil Rights (OCR). The state and national projections were prepared for OCR in accordance with the sampling methodology for the Civil Rights Data Collection.

The 2006 CRDC was administered to a nationally representative stratified sample of public school districts and schools. A total of 6,028 school districts (i.e., approximately 40 percent of all eligible public school districts in the nation) was included in the 2006 CRDC. Of the 6,028 sampled districts, 99 (1.6 percent) were determined to be ineligible (out-of-scope) by OCR—e.g., districts that closed or merged. Of the remaining 5,929 eligible districts, all 5,929 participated (responded) in the survey, for an overall district response rate of 100 percent. Within the 5,929 responding districts, 62,484 eligible schools were identified. Of these, 62,211 provided usable data for the statistical projections, for an overall school-level response rate of 99.6 percent.

Item response rates, which refer to the percentage of cases without missing (unknown) values for a particular data item, were generally very high (usually no less than 96 percent and very often close to 100 percent) for most survey items in almost every state. To compensate for item nonresponse, imputation procedures were used to estimate values for the missing items.

To provide a measure of the precision of the estimates, standard errors have been computed for all of the statistical projections. Note that projections with relative standard errors (i.e., the standard error divided by the estimate) in excess of 0.25 are deemed to be subject to considerable uncertainty and are flagged as such in the statistical tables.

Documentation is available from OCR that describes the procedures used for the projections, including weighting of the sample, imputation for item non-response and for schools with unusable data, standard errors, and quality control procedures. In addition, documentation is available from OCR for projections that should be used with caution due to large statistical uncertainty in the estimate, including factors that contributed to the extent of this statistical uncertainty for the 2006 Civil Rights Data Collection. This hardcopy documentation, available upon request, is contained in the “2006 Civil Rights Data Collection (CRDC) Projections and Documentation.”