Secondary Statistical Modeling With the National Assessment of Adult Literacy: Implications for the Design of the Background Questionnaire

With the advent of simple-to-use advanced statistical software packages, it is becoming increasingly easy to specify and estimate complex statistical models for addressing substantive questions in adult literacy and other areas of education. Such problems as the role of literacy in voting behavior or the job market experiences of individuals with GEDs, can be addressed with relative ease. However, a thoughtful application of statistical models to educational data leads to the recognition that certain assumptions must be met for the model estimates to be useful for theoretical explanation and/or policy analysis. The purpose of this paper is to offer recommendations to the National Center for Education Statistics on the development of the background questionnaire for the National Assessment of Adult Literacy. The recommendations presented in this paper are from the viewpoint of a researcher interested in applying sophisticated statistical models to address important issues in adult literacy. This paper focuses on five issues: (1) sampling, (2) selection bias, (3) measurement, (4) policy analysis, and (5) cohort effects.

Online Availability

• Download, view and print the report as a pdf file.