Secondary Statistical Modeling with the National Assessment of Adult Literacy

Book excerpt: This paper offers recommendations to the National Center for Education Statistics (NCES) on the development of the background questionnaire for the National Assessment of Adult Literacy (NAAL). The recommendations are from the viewpoint of a researcher interested in applying sophisticated statistical models to address important issues in adult literacy. The paper focuses on five issues, each of which is the subject of a section of the paper: sampling; selection bias; measurement; policy modeling; and gauging cohort effects. Each section considers the scope of the issue and then makes recommendations to NCES. These recommendations include providing all appropriate sampling weights in NAAL data; examining contextual effects on the distribution of literacy ability in the population; considering relevant auxiliary variables that would constitute the selection equation; considering the hypothesized number of factors and including at least four variables measuring each factor in the questionnaire; obtaining retrospective data on general and job-specific literacy-related activities; and exploring the possibility of linking NAAL with existing longitudinal surveys. (Contains 21 references.) (YLB)