- Surveys & Programs
- Data & Tools
- Fast Facts
- News & Events
- Publications & Products
- About Us
Twitter
Table of Contents | Search Technical Documentation | References
NAEP Technical DocumentationWeighted Distributions of Students Before and After Student Absenteeism
 |
Weighted Student Percentages Derived from Sampled Schools Grade 4 Mathematics Weighted Student Percentages Derived from Sampled Schools Grade 4 Mathematics |
||
|
|||
To check the difference between the full sample and the assessed samples, weighted distributions of students before and after student absenteeism were studied. Weighted sampled percentages of students by jurisdiction, subject, and grade were calculated by gender (male) and race/ethnicity (White, not Hispanic; Black, not Hispanic; Hispanic), and by students-with-disabilities or limited-English-proficient (SD/LEP) status for the full sample of students (after student exclusion), and for the assessed sample. The mean student age in months is also presented on each basis.
Reporting populations differ by whether accommodations were offered to disabled or limited English proficiency (SD/LEP) students. The non-accommodated reporting population, also known as the R2 reporting population, includes all non SD/LEP students plus SD/LEP students from non-accommodated sessions. The accommodated reporting population, also known as the R3 reporting population, includes all non SD/LEP students plus SD/LEP students from accommodated sessions.
The weight used for the full sample was the student base weight, defined in Calculation of Student Base Weights. The weight for the assessed students was the final student weight, defined in Adjustments for Nonresponse at the Student Level. The difference between the estimates of the population subgroups is an estimate of the bias in estimating the size of the subgroup, resulting from student absenteeism.
Care must be taken in interpreting these results. First, note that there is generally little difference in the proportions estimated from the full sample and those estimated from the assessed students. While this is encouraging, it does not eliminate the possibility that bias exists within the state as a whole, within the results for gender and race/ethnicity subgroups, or within other subgroups. Second, when differences do exist, they cannot be used to indicate the likely magnitude or direction of the bias with any reliability. For example, in Wisconsin for grade 8 science (for the sample in which accommodations were permitted), the percentages of White and Black students in the full sample are respectively 80 and 8 percent. For assessed students, these percentages are 82 for White students and 7 for Black students. This indicates that White students are slightly over-represented and Black students are slightly underrepresented in the sample of assessed students. While these differences raise the possibility that some bias exists, it is not appropriate to speculate on the magnitude of this bias by considering the assessment results for White or Black students in comparison to other students in the state. The reason is that the over-represented White students or underrepresented Black students may not be typical of students that were included in the sample. Similarly, White students who are disproportionately underrepresented or Black students who are disproportionately over-represented may not be typical either, because not all students within the same race/ethnicity group receive the same student nonresponse adjustment.
One other feature to note is that, for assessed students, information about the student's gender and race/ethnicity is provided by the student, whereas for absent students, it is provided by the school. Evidence from past NAEP assessments indicates that there can be substantial discrepancies between those two sources, particularly for grade 4 Hispanic students.
Last updated 13 August 2008 (KL)