Appendix B—Methodology and Technical Notes—Response Rates

Unit response rates. The unit response rate indicates the percentage of sampled cases that met the definition of a complete interview. The weighted SASS unit response rate was produced by dividing the base-weighted number of respondents who completed questionnaires by the base-weighted number of eligible sampled cases.⁷ Table B-1 summarizes the base-weighted unit response rates for each survey type. Because response rates vary between surveys, it is possible not to have information for all the components related to a particular school. For example, it is possible that a district does not have a corresponding school record or there may not be a principal record for every school.

Overall response rates. The overall response rate represents the response rate to the survey taking into consideration each stage of the survey. For teachers, the overall response rate is calculated as the product of the response rate to two stages: the Teacher Listing Form and the teacher questionnaire.⁸ The base-weighted overall response rate for public school teachers was 72.4 percent; for BIE school teachers, 71.4 percent; and for private school teachers, 65.9 percent. For the other surveys, the overall and unit response rates are the same since they have only one data collection stage.

Unit nonresponse bias analysis. Because the NCES Statistical Standards (4-4) require analysis of nonresponse bias for any survey stage with a base-weighted response rate less than 85 percent, all SASS files were evaluated for potential bias. First, the base-weighted unit response rate was calculated by state, region, or affiliation depending on the sector (public, BIE, and private respectively). If the base-weighted response rate for any state, region, or affiliation was below 85 percent, a detailed comparison of respondents to the frame population was conducted by examining the following characteristics: grade level, urbanicity, enrollment, and state/affiliation. A difference between the frame and respondent population was considered noteworthy if the difference was statistically significant and the following four conditions were met:

• The relative difference between the frame and respondent population was greater than 10 percent;
• The absolute difference was greater than one percentage point;
• The coefficient of variation was less than 15 percent; and
• The cell for each subpopulation contained at least 30 interviews.

As shown in table B-1, the base-weighted response rate was 81.7 percent for public school library media centers and 78.9 percent for BIE school library media centers. When response rates were calculated further by state for public school library media centers and region for BIE school library media centers, 31 states and 3 regions had rates below 85 percent. Table B-2 contains a list of the comparisons between the frame and the weighted distribution that were analyzed for potential bias, with an indication of the comparisons with evidence of potential bias.

Nonresponse adjustments were designed to reduce or eliminate nonresponse bias. The following variables were included in the nonresponse adjustments: grade level, urbanicity, enrollment, and state/affiliation. The final-weighted comparison to the frame reflects the nonresponse adjustment. Table B-2 shows those comparisons that have evidence of potential bias after the nonresponse adjustments were included. For further information on unit response rates and nonresponse bias analysis, see the Documentation for the 2007–08 Schools and Staffing Survey (Tourkin et al. forthcoming).

Item response rates. The item response rate indicates the percentage of respondents who answered a given survey question or item. The weighted SASS item response rate is calculated by dividing the base-weighted number of respondents who provided an answer to an item by the base-weighted number of respondents who were eligible to answer that item.⁹ Table B-3 provides a brief summary of the base-weighted item response rates for each survey. The nonresponse bias analysis conducted at the item level revealed no substantial evidence of bias in the school files. For further information on nonresponse bias analysis and item response rates, see the Documentation for the 2007–08 Schools and Staffing Survey (Tourkin et al. forthcoming).

Top

---

⁷ For the formula used to calculate the unit response rate, see NCES Statistical Standards (U.S. Department of Education 2003).
⁸ For the formula used to calculate the overall response rate, see NCES Statistical Standards (U.S. Department of Education 2003).
⁹ For the formula to calculate the item response rate, see NCES Statistical Standards (U.S. Department of Education 2003).