Part B. Collections of Information Employing Statistical Methods
The CCD collections do not use statistical methods. The survey system is intended to provide directory and statistical information about the universe of public schools and education agencies. Although the respondents are experts in their states, there are opportunities for potential error, either through technical mishap or misinterpretation of the intent of a survey item. Over the life of the CCD survey system, staff members have developed a series of checks designed to flag these errors for review by the respondent and action by the CCD team. These edits rely on internal logic checks; consistency within specified tolerances over time; and consistency within a given state and type of unit. Items reported as missing or not applicable are confirmed with states and the designation may be changed if the rationale is not convincing. Comparisons with copyrighted proprietary lists such as those of Quality Education Data and Market Data Retrieval have consistently shown CCD school and agency coverage be as good or better than these lists in the coverage of regular schools an agencies
We anticipate that up to 47 states will report via the EDEN-CCD pilot for 2006-07. We anticipate that the other jurisdictions and up to four states will not. The universe includes the state education agencies, or their equivalent for non-states, for the 50 states, District of Columbia, Bureau of Indian Education, Department of Defense Dependents Schools (domestic and overseas), American Samoa, Commonwealth of the Northern Mariana Islands, Guam, Puerto Rico, and the U.S Virgin Islands. The survey response rate is 100 percent in almost every year of the CCD’s history. (Guam did not report in 2002–03 and the Bureau of Indian Affairs did not report in 2004-05).
None of the procedures described under this item (stratification, estimation, etc.) are used in the CCD collection.
Survey response is typically 100 percent. When there is item nonresponse, the item is identified as “missing” on the files and in reports for the school and local agency universes. Imputations are limited to the 50 states and the District of Columbia because the other reporting agencies may be qualitatively different in their structure. For example, for the 2004-05 school years prekindergarten teachers were imputed for 3 states; kindergarten teachers for 2 states; elementary teachers for 2 states; secondary teachers for 2 states; ungraded teachers for 2 states; instructional aides for 2 states; instructional coordinators for 2 states; librarians for 1 state; library support staff for 3 states; school district support staff for 3 states; school support staff for 3 states; other support staff for 2 states; and prekindergarten students for 1 state. A missing response on the state nonfiscal survey is imputed if the item was reported by the state in the past. The imputation method applies the rate of change among responding states to the prior year’s value for the reporting state. (That prior year’s value may have been imputed.)
The NPEFS imputes missing values by applying a current year ratio or rate of change, whichever is appropriate, among reporting states to the non-reporting states.
No such tests will be undertaken. Items are not added to the CCD unless a substantial majority (usually two-thirds percent or more) of respondents say that they can provide the item within 1 year of its introduction. This information is gathered during CCD Coordinator training and at EIMAC meetings. State CCD Coordinators also have the opportunity to respond on-line to proposals for new items.
There are no statistical aspects to the design of the CCD collections and thus no individuals were contacted.
This page intentionally left blank.
| File Type | application/msword |
| File Title | Part B |
| Author | barbara.spinner |
| Last Modified By | DoED |
| File Modified | 2006-11-22 |
| File Created | 2006-11-22 |