SASS School Library Media Center Survey (SLS)

5. Data Quality and Comparability

Although data are imputed for nonrespondents, caution should be exercised when analyzing data by state, sector, or affiliation. Since nonresponse varies by state, the reliability of state estimates and comparisons are affected. Users should be especially cautious about using data at a level of detail where the nonresponse rate is 30 percent or greater. See below for more information on the types of errors affecting data quality and comparability.

Sampling Error

The estimators of sampling variances for SASS statistics take the SASS complex sample design into account. (See SASS chapter).

Nonsampling Error

Nonresponse Error.
Unit nonresponse. The weighted unit response rates for the 2007–08 School Library Media Center Survey were 76.9 percent for public schools and 82.1 percent for BIE schools.

Item Nonresponse. Some 95 percent of the items in the public school version of the 2007–08 School Library Media Center Survey had response rates above 85 percent and 93 percent of the items in the BIE version had response rates above 85 percent. All items in both versions had response rates above 70 percent, and there was no substantial evidence of bias in either case.

Measurement Error. A reinterview was conducted for the 2003–04 SASS, but it did not include questions from the School Library Media Center Survey.