SASS School Library Media Center Survey (SLS)

4. SURVEY DESIGN

TARGET POPULATION

The universe of library media centers/ libraries in elementary and secondary schools with any of grades 1–12 in the 50 states and the District of Columbia

SAMPLE DESIGN

In 1993–94, the library media center sample was a subsample of the SASS school sample. Drawn from the 13,000 schools in the SASS, the library sample consisted of 5,000 public schools, 2,500 private schools, and the 180 BIE schools in the United States.

The strata used for library sampling were the same as those used in the public school sampling of the Schools and Staffing Survey (SASS) (see SASS chapter for details) (state and grade level). All BIE schools were selected for the library survey, so no stratification or sorting was needed. Within strata, public schools in the 1993–94 sample were sorted on the following variables:

The sample schools were then systematically subsampled using a probability proportionate to size algorithm, where the measure of size was the square root of the number of teachers in the school as reported in the CCD (the public school sampling frame for SASS) multiplied by the inverse of the school’s probability of selection from the public school sample file. Any school with a measure of size larger than the sampling interval was excluded from the library sampling operation and included in the sample with certainty.

The private school library frame for 1993–94 was identical to the frame used for the SASS private school survey, except that it excluded schools with special program emphasis (special education, vocational, or alternative curriculum schools). Private schools were stratified by recoded affiliation (Catholic, other religious, nonsectarian); grade level (elementary, secondary, combined); and urbanicity (urban, suburban, rural). Within each stratum, sorting occurred on the following variables: (1) frame (list frame and area frame); and (2) school enrollment.

Within each stratum, private schools were systematically selected using a probability proportionate to size algorithm. The measure of size used the school’s measure of size (i.e., the square root of the number of teachers in the school as reported in the CCD) multiplied by the inverse of the school’s probability of selection. Any library with a measure of size larger than the sampling interval was excluded from the probability sampling process and included in the sample with certainty. In all, 2,500 private schools were selected for the library sample in the 1993–94 SASS. In 1999-2000, the Library Media Center questionnaire was administered to all school within the SASS sample.

In 2003–04 and 2007–08, the Library Media Center questionnaire was administered to public and BIE SASS school samples, excluding private schools. Each sampled school received a library media center questionnaire. The sampling design for Library Media Center Survey follows that of the public school sample and BIE school sample of SASS. The BIE schools were selected for the sample with certainty. A number of changes were made in the sample design (i.e., stratification, sample sizes, sample sort, and school definition) from the 1999–2000 SASS to the 2003–04 SASS to the 2007–08 SASS. See more information on the 2007–08 public and BIE school sampling in Schools and Staffing Survey (SASS) chapter. 

Top

Data Collection and Processing

The U.S. Bureau of the Census is the collection agent for the School Library Media Center Survey. Data collection and processing procedures are discussed below.

Reference Dates. Most data items refer to the most re-cent full week in the current school year. Questions on collections and expenditures refer to the previous school year.

Data Collection. The School Library Media Center Survey is delivered with other SASS components beginning in October of the survey year. The survey is delivered to the school librarian or another staff member familiar with the library. (The follow-up procedures are described in SASS chapter).

Editing. Once data collection is complete, data records are processed through a clerical edit, preliminary interview status recode (ISR) classification, computer pre-edit, range check, consistency edit, and blanking edit. (See SASS chapter for details). After the completion of these edits, records are processed through an edit to make a final determination of whether the case is eligible for the survey and, if so, whether sufficient data have been collected for the case to be classified as an interview. A final ISR value is assigned to each case as a result of the edit. 

Top

Estimation Methods

Weighting. In the SASS School Library Media Center component, data are used to estimate the characteristics of schools with library media centers as well as schools without library media centers. Whenever possible, sampled schools with library media centers and sampled schools without library media centers are adjusted separately. Thus, interviewed library media centers are weighted up to the weighted estimate of sampled schools known to have library media centers, as determined at the time school library media center questionnaires were distributed. Likewise, the number of interviewed schools without library media centers is weighted up to the weighted number of all schools without library media centers as determined from the questionnaire distribution. This is done to study the characteristics of each type of school.

When it is not possible to adjust the library weights by the type of school, all sampled school library media centers and schools without library media centers are adjusted as a whole. This is necessary to handle instances in which the existence of the library media center cannot be established during data collection. Due to reporting inconsistencies between the school library media center questionnaire and the school questionnaire, school library media center survey data are not adjusted directly to schools reporting to have library media centers on the school questionnaire.

Imputation. Items from the SASS School Library Media Center questionnaire that still had items that were “not answered” went through a first stage of imputation in which unanswered items were imputed from other items on the same library media center record or items on the corresponding school record. The library media center data then went through the second stage of imputation in which some of the remaining “not answered” items were filled using either the data record from a similar record, regression imputation, or random ratio imputation. The third stage of imputation filled in the remaining “not answered” items that were not resolved during the first two stages of imputation (i.e., imputed clerically). After all stages of imputation were completed and no more “not answered” items remained, the library media center data from BIE-funded schools were separated into a single dataset. Recent Changes The School Library Media Center Survey has not been administered to private schools since the 1999-2000 school year. 

Recent Changes

The School Library Media Center Survey has not been administered to private schools since the 1999–2000 school year.

Top