Frequently Asked Questions -> Assessment Design

• Can I use NAAL to assess the literacy skills of individuals?
• Can I obtain NAAL test booklets for secondary analysis?
• How do you know the sample size is large enough?

Can I use NAAL to assess the literacy skills of individuals?

No. NAAL is designed to assess large groups (not individuals) and to produce national and, in the case of SAAL, state literacy estimates. The assessment is carefully designed to minimize the number of literacy tasks that each individual will need to answer to obtain reliable national estimates for each of the three literacy scales. Since assessment questions are apportioned among 26 booklets; no respondent takes a sufficient number of each task type to warrant a scale score for individuals. Rather, the scores are aggregated across participants to produce national estimates (or state estimates for SAAL states) for each of the literacy scales. Thus, NAAL is not a good tool for assessing literacy at the local or individual level.

Can I obtain NAAL test booklets for secondary analysis?

Because NCES will reuse large portions of the 2003 NAAL assessment questions in future assessments, very few questions will be released to the public. Certain assessment questions from the 2003 NAAL will be released once it is determined that they will not be reused in the next assessment to support trend integrity and analysis. Yet approximately 100 questions from 1992 NALS (that were not reused in 2003) are posted in the Sample Test Questions section of this website. These items bear a close resemblance to those used in the 2003 assessment.

How do you know the sample size is large enough?

Determining the appropriate sample size requires a balance between available resources and the benefits of increasing the sample size. Large samples provide greater accuracy than small samples, but they are more expensive to obtain. Also, their increase in accuracy is not as dramatic as many people expect. This is because the precision of a survey's estimates increases with the square root of the sample size (rather than the absolute difference in sample size). For example, a sample of 10,000 provides estimates that are only about twice as precise as a sample of 2,000 (the square roots being 100 and 45). The NAAL sample is not large enough to provide reliable estimates for individual states (except for states participating in the State Assessment of Adult Literacy (SAAL), which have large enough samples and are designed for state estimation). However, NAAL will provide valid estimates for the nation and major national population groups.

Top