Search
Program Calendar
Browse By Day
Browse By Time
Browse By Person
Browse By Room
Browse By Unit
Browse By Session Type
Search Tips
Visiting Washington, D.C.
Personal Schedule
Sign In
X (Twitter)
The purpose of this presentation is to validate the NAEP teacher sampling weights using two other nationally representative datasets of NCES: the Schools and Staffing Survey (SASS) and the Common Core of Data (CCD). Such validation is important because it provides support for our assertion that the NAEP fourth grade teacher sample is nationally representative, and that the proposed method of creating the teacher sampling weights (i.e., applying the school sampling weight adjusted for teacher nonresponse) is adequate and appropriate.
Table 5 presents the weighted distribution of NAEP and SASS teachers by the following variables: race/ethnicity, certification, master’s degree, experience, percent non-white enrollment, and percent NSLP-eligible enrollment. There were no significant differences between the two data sets for any categories of any variables at the .05 alpha level, based on independent t-tests of weighted percentages for each sample. We conclude from these analyses that the weighted distribution of NAEP teachers is not significantly different from the weighted distribution of SASS teachers, for the variables analyzed.
While a comparison of weighted distributions is helpful, the comparison of weighted counts provides further validation of the proposed NAEP sampling weights. To carry out this validation, we compared the weighted sums of NAEP teachers to national estimates available through CCD. Results are provided in Table 6. Using CDD data, we estimated there are about 173,000 fourth grade teachers in U.S. public schools overall. We also estimated the number of fourth grade teachers by categories of school locale, size, percent non-white enrollment, and percent NSLP-eligible enrollment, shown in the second column. We then compared these estimates to the sum of the teacher sampling weights for the 2013 NAEP mathematics assessment (about 163,000). This sum was about 6% lower than the number of teachers estimated using CCD data. Across four different school characteristics, NAEP estimates tended to be lower than CCD estimates, with the exception of rural schools, small schools, and schools with up to 25% non-white enrollment. While estimates for some school categories varied substantially, others were quite close (such as schools with an enrollment of up to 300 students). Across the 17 school categories displayed in Table 6, the average discrepancy between NAEP and CCD was about 8.5%. For the purpose of conducted teacher-weighted analyses, this level of divergence between NAEP and CCD estimates is reasonable.
In summary, based on the comparisons between NAEP, SASS, and CCD, we conclude that the proposed NAEP teacher sampling weights, based on the NAEP school sampling weights, will allow NAEP analysts to generate estimates that are representative of the nation’s population of fourth grade teachers.