• History of the CCD
• The Survey System
• Uses of the CCD
• Issues in Quality
• Recent, Ongoing, and Upcoming Developments
• References

The earliest responsibility of the CCD—and its predecessors—was to make available a list of public schools and the local education agencies that managed them. In the 1960s, publication of the Directory of Public School Systems began to follow a regular schedule as Part II of the Department of Education's Education Directory series. The corresponding universe of school names and addresses was maintained as an internal, unpublished listing. The next decade saw this system expand into the Elementary and Secondary Education General Information System. ELSEGIS was a more ambitious data system, collecting and publishing more extensive information on public schools and education agencies.

The current design of the CCD was introduced in 1986-87, with the first data of publishable quality collected in the following school year. In short order, the survey system acquired characteristics that distinguish it from its precursors and give the CCD its current identity.

Unlike earlier education directory systems, the CCD began in the late 1980s to edit the school and agency universe files with the same attention that had formerly been directed toward the state-level data collections. (These latter continued to stand as the official state statistics for the reported data.) At about the same time, the CCD discontinued the then-common practice among commercial listing services of maintaining "open" files: that is, allowing information to be updated as it became available, with the result that a file could include data from different years for different states. NCES decided that all CCD data reported by states would be those that were current as of October 1 of the school year. Old data could not be carried over from one year to the next, which made issues in the timeliness of state reporting more crucial than they had been in the past.

The CCD of the late 1980s expanded the content of the surveys as well as established new reporting and editing conventions. This expansion—and the process through which states agreed to report data items with common definitions—was accomplished in partnership with the Council of Chief State School Officers. CCSSO is a professional association representing the heads of state public education agencies, and it has a longstanding role in brokering the data interests between state data reporters and federal data collectors. In 1985, CCSSO began, under contract to NCES, to examine the completeness and comparability of data reported on the CCD as it was then structured. The project initiated the practice of negotiation, consensus, and verification between state agencies and NCES that continues today.

The CCD is a series of six separate annual data collections reporting information drawn from the administrative records of state education agencies. It encompasses the universe of public schools and local education agencies¹ and is limited to elementary and secondary school data.

Two surveys provide financial data. State-level information about revenues and expenditures is reported on the "National Public Education Financial Survey." The "Annual Survey of Government Finances: School Systems" collects comparable information about individual local education agencies. More commonly known by its form number (F33), this survey is supported jointly by the Bureau of the Census (Governments Division) and NCES.

Nonfiscal information about students, staff, and institutions is drawn from three CCD surveys. The "State Nonfiscal Survey of Public Elementary/Secondary Education" is the source of official state-level counts of public school students, teachers and other staff, and high school completers.

The "Local Education Agency Universe Survey" provides information about approximately 16,500² local agencies, in addition to the state education agencies, that are responsible for providing public education or services that support it. This survey includes information about institutional characteristics, numbers of education staff, numbers of students participating in selected education programs, dropouts, and high school completers.

The "Public Elementary/Secondary School Universe Survey" is similar. In 1997-98, it encompassed more than 90,000 public schools. Like its state and local agency counterparts, the school universe survey reports institutional characteristics and the numbers of teachers, student enrollment by grade (with detail on gender and racial/ethnic category), and students participating in selected education programs.

All of these collections report data that are at least 1 year out of date—the information collected by the state that reflects conditions on October 1 may not be reported to NCES until the following September. The "Early Estimates of Public Elementary/Secondary Education Survey" offers a sample of more current data. This survey reports, before the end of each school year, the estimated numbers of students and teachers, high school graduates, and revenues and expenditures for education.

The CCD is accessible through a number of print and electronic products that reflect increasing numbers of uses and users. Its historic function as a directory of public schools and education agencies remains, and the CCD serves as a mailing list, institution registry, and sampling frame.

A Directory of Public Elementary and Secondary Education Agencies is printed each year.³ The publication is also available on the NCES Web Site. The Internet has also enabled NCES to maintain a Public School and School District Locator, through which users can search for individual schools or districts and secure basic information such as address, telephone number, state and NCES identification codes, and some student and teacher data. The NCES identification code is important because it is used by applicants for Schools and Libraries Corporation e-rate (universal rate) telecommunications discount grants and for grants under the Safe Schools/Healthy Students program.

A Web product, the School District (Agency) and Public School Name and Address Files, is available. Vendors and marketers are particularly likely to rely on this product to prepare customized lists for their mailings.

The school and agency universe files are used in drawing samples for national studies. The CCD has been used recently in drawing samples for the NCES Schools and Staffing Survey and the National Assessment of Educational Progress. Its component school districts serve as the framework to which demographers are mapping extensive demographic data from the 2000 Decennial Census.

The CCD is a source of descriptive statistics about public schools and districts. This can be an important function for some education programs—for example, the state per pupil expenditure drawn from the "National Public Education Financial Survey" is used to calculate allocations for Title I, Impact Aid, and other federal programs. Descriptive statistics are reported for general use through short Statistics in Brief publications (three of which are featured in this section) and through more detailed reports, such as the Characteristics of the 100 Largest School Districts in the United States, as well as other publications. Each edition of the NCES Digest of Education Statistics incorporates CCD data in 40 or so of its tables. The CCD serves as the only annual universe report of the numbers of students enrolled in public schools, the numbers of high school diplomas awarded, and the numbers of public school teachers. For the more than 40 states that participate in the dropout data collection (the number is growing over time), the CCD provides comparable statistics about how many students drop out of school each year.

One relatively new use of the CCD is that of research database. In the 1999 Condition of Education, the CCD was used to analyze changes in the racial/ethnic isolation of students in public schools over a 10-year period. This application of the CCD has been aided by work to reconstruct lost documentation and by making archived files available through the International Archive of Education Data. It has also relied on the development of a longitudinal education agency research file that matches school districts over time and includes imputed data for a number of missing responses.

Administrative records data have had limited use in research and policymaking because they may not be considered trustworthy. The nonsampling errors to which administrative records are vulnerable include incomplete coverage, discrepancies in data definitions, varying periodicity, and inability to verify data quality.

A coverage evaluation of the elementary/secondary education agency universe concluded that the CCD did a good job of representing traditional types of education agencies (Owens 1997); a companion study of school coverage is currently in review. Nontraditional agencies, particularly those not administered by the state education agency, were more likely to be missed. These evaluations set the context for discussions with state data reporters about how to improve coverage and have led to changes that will ostensibly improve the surveys' representation. As one example, schools and districts under the jurisdiction of the Bureau of Indian Affairs (BIA) had been inconsistently reported by states. BIA schools and districts are now excluded from state reports and published as a separate reporting entity. Again, there was an explicit effort in the 1998-99 collection to improve the coverage of charter schools, whose governance structure varies widely across states. This was partially successful in capturing more schools. However, as long as the CCD continues to draw upon the administrative records of state education agencies, there will be some schools and agencies that fall outside its net.

Respondents may not follow definitions uniformly. For example, classification evaluations of the CCD, and subsequent technical review panels, found that states disagreed about when to classify an education agency as a supervisory union, a regional service agency, or a state operated agency. The survey continues to struggle with crafting a feasible and professionally acceptable definition of a vocational school. Definitional agreement in reporting has been approached in several ways. Topical technical review panels uncover possible sources of confusion (or inapplicability) in current item definitions, with resulting guidance to state data reporters through the CCD Home Page, direct mailings, and annual training. For the number of dropouts, a particularly high-stakes statistic, each CCD coordinator is questioned annually about the state's adherence to the CCD definition and reporting procedures.

Periodicity has been addressed by arbitrarily setting October 1 as the "as of" date for CCD counts. Current-year statistics are to be reported as they were observed on October 1; past-year statistics, such as the number of dropouts or the number of students receiving migrant education services, are reported as they were known to be on that day. Information collected from state CCD coordinators in the 1991-92 introduction of the dropout statistic found that the data from most states had been collected within a week of October 1, but that a few states reported data that had been collected 4 or 5 weeks earlier or later. This issue appears to be one that is improving due to developments in information technology. In past years, for example, the sum of students in the reported racial/ethnic categories often differed from the reported total of students—purportedly because the counts were taken at different times. Today this problem does not arise on the state-level survey.

The quality of data is promoted by attention to the conditions under which data are collected and reported—conditions that were discussed in the preceding paragraphs—and by editing data once they are received. A recent evaluation concluded that the CCD survey processing included more edits than necessary, based on the number of data changes in response to edit challenges (Hamann 1999). The number of edit challenges reported to state CCD coordinators has been reduced. The "National Public Education Financial Survey" continues to send prior-year comparisons and add checks on all subtotals to the respondents, in part because sizable federal allocations are affected by the numbers. The state nonfiscal survey, which is Web based, incorporates soft edits that respondents can override with an explanatory note. Similar editing software for the school and education agency universe surveys is available to states. More comprehensive Web-based versions of this software are expected to be available by August 2000.

The revised CCD school and education agency universe surveys that were introduced in the 1998-99 school year added a number of features that were intended to make the CCD more useful for sampling and program uses. The idea of "flags" was introduced. In addition to its traditional type code (e.g., regular, vocational), a school is now flagged as magnet, charter, or Title I, as appropriate. Additional detail provides school membership by gender and racial/ethnic category within each grade, and there are now reports of the numbers of students receiving migrant and limited English proficiency services.

The school locale code, originally developed by CCD staff at NCES, has been refined in the last year. The variable identifies the degree to which a school is located in an urban setting, with codes ranging from "large city" to "rural." The addition of a "location address" (if this differs from the mailing address) has improved the assignment of locale codes on the basis of Census place. For example, if a rural school receives its mail at the post office in a neighboring town, the location address will ensure that the school is coded as "rural" rather than "small town." And, the existing rural category has been broken into "rural (outside a metropolitan statistical area)" and "rural, urban fringe (within a metropolitan statistical area)."

CCD staff, in cooperation with state education agencies, have also reached consensus on a high school completion rate that is being introduced with 1998-99 data. This rate is the proportion of students leaving school who leave as completers. It is the number of high school completers in a given year divided by the number of completers plus the number of dropouts from grades 12, 11, 10, and 9 in the current and 3 preceding years, respectively.

Because of the importance of the environment in which CCD data are first produced, NCES supports technical assistance activities that are peripheral to the CCD. These include comprehensive handbooks for elementary/secondary student and staff data. The Student Data Handbook was revised within the last year and will be updated annually.⁴ The Staff Data Handbook (Malitz 1995) will be revised by January 2001, with the same provision for annual review and revision. Financial Accounting for Local and State School Systems: 1990 (Fowler 1990) is undergoing substantial review in response to developments in program interests (e.g., better ability to reflect expenditures for technology or school safety) and new reporting recommendations from the Government Accounting Standards Board.

In the area of information technology, NCES participates actively in the development of national data standards. The X-12 subcommittee of the American National Standards Institute approves standards for the electronic exchange of information about people. The X-12 subcommittee has approved the format for electronic student record exchange developed by NCES and the members of the Center's elementary and secondary education and postsecondary education data cooperatives. At present, NCES participates in X-12, continues a project to develop electronic data interchange (EDI) standards for the CCD, and has joined an education and vendor group that is creating extensible markup language (XML) standards for the information typically collected and used by schools and school districts.

In the near future, NCES will be studying ways to better exploit the use of electronic data exchange in collecting and reporting administrative records data. The challenge will be to increase the timeliness of CCD data without threatening quality or adding to response burden. This effort can potentially challenge several of the CCD's basic operating principles—for example, one suggestion has been to improve timeliness by publishing directory information before the statistical information is available—but it will result in a better and more responsive survey system.

⁴ The current edition of the Student Data Handbook (Cheung and Young 2000) is excerpted later in this section, beginning on p. 66.