Skip Navigation
small header image
National Assessment of Adult Literacy (NAAL)
Publications & Products   |   Data Files   |   Staff   |   Search   |   Site Index

Data Files from the 2003 National Assessment of Adult Literacy

Information on Data Files

This U.S. Department of Education’s National Center for Education Statistics has released the household and prison public-use data files for the 2003 National Assessment of Adult Literacy and the 1992 National Adult Literacy Survey. The accompanying 2003 National Assessment of Adult Literacy Public-Use Data File User's Guide PDF File (1.7 MB) explains how the data was collected and how it can be analyzed and includes codebooks for the datasets as well as instructions for using AM Software to analyze the data.

Restricted use files are available with a complete set of the variables collected on the NAAL. Users must apply for an NCES license to access restricted data.

The data derive from one-on-one interviews with adults in homes and prisons across the United States. Household interviews were conducted in 2003 and prison interviews were conducted in 2004. The household sample was multi-stage and selected on the basis of Census-defined geographical areas, while the prison sample was institution-based.

To keep participant burden to a minimum, NAAL administers only a fraction of the assessment items on each scale to each participant—too few items to produce accurate scale scores for each adult. To account for this, marginal maximum likelihood (MML) models are used to estimate proficiencies for the population and for groups within the population based upon responses to questions that were answered. The AM Software provides these estimates.

Top

Data Files

The 2003 combined household-prison public use file Zip File (5.3 MB) includes background data and responses to test items that can be used to estimate proficiency on the prose, document, and quantitative scales using a marginal maximum likelihood model. There are 19,258 completed interviews in this file. An additional 456 adults could not be interviewed and tested because of language barriers or for cognitive reasons (such as Alzheimer’s or dementia). These adults are not included in the data file. A separate 2003 public-use file Zip File (5.3 MB) permits estimates of proficiency on the health literacy scale.

The 1992 combined household-prison public use file Zip File (8.5 MB) also includes background data as well as test items that can be used to estimate proficiency. Background variables that are the same in both years were given the same name.

Top