dropped rapidly. Adults 55 -64 and those 65 or older performed well below the levels of younger adults, even though their average years of education was not much different from the 16-18 year olds. Summarizing across the three literacy scales, about 4448 percent of those adults categorized in Level 1 were aged 55 or older, and 32-35 percent were 65 years old or older. Some 28-32 percent of those in Level 2 were 55 years old or older, and 16-18 percent were 65 or older.

From the NALS data it is not possible to say whether adults' literacy skills rise and then decline or whether the various age groups have performed at the levels indicated throughout adulthood. This would require longitudinal studies. However, referring to the human cognitive system of page 4 of this report, the NALS tasks do impose heavier burdens on working memory as they increase in difficulty. In fact, this may be one of the major reasons the tasks increase in difficulty. The authors of the NALS report note that, of several variables that might make tasks more difficult, two of the variables for prose and document tasks are the number of categories or features of information that the reader has to process or match, and the number of categories or features of information in the document that can serve to distract the reader or that may seem plausible but are incorrect. In the quantitative tasks, the number of operations needed to perform the task is given as a factor that may influence the difficulty of the task (Kirsch, et. al, 1993, pp. 74, 85, & 94).

Generally, holding features or categories of information in short term or working memory and then searching through other information places greater demands upon working memory, and there is considerable evidence that working memory performance declines with increasing age (Bernstein, Roy, Srull, & Wickens, 1988, p.401). This may explain, at least in part, the decrease in literacy skills as age increases.

As indicated in the human cognitive system model, one of the factors that is important for literacy is one's organized bodies of knowledge. The bodies of knowledge are what makes it possible to comprehend printed displays, to reason analogically (i.e., from one body of knowledge to another), and to make inferences (i.e., going from the information given in the display to another body of knowledge in one's mental knowledge base to create yet a third domain of knowledge needed to correctly perform an inference- type task). Generally, these organized bodies of knowledge continue to develop across adulthood and tend to resist deterioration in older age (Bernstein, et. al, 1988, p. 401). While the NALS includes tasks that include knowledge content from health, consumer economics, and others, it does not systematically assess people's organized bodies of knowledge in any domain (e.g., health, science, government, etc.). It is not possible to know whether poorly performing people's primary problems may be their lack of knowledge (e.g., vocabulary, concepts, etc.) or of working memory control, or both. But the rapid decline in performance with ages above 55 suggests a strong component of working memory control in the NALS tasks.

Health Conditions

A major contribution of the NALS was the sampling of adults with various forms of physical, mental or other health conditions. The survey reported that 12 percent of the adult population reported some type of health problem. Significantly, as a type of epidemiological indicator of the self-perceived extent of adult learning difficulties in the U. S. population, some 3 percent (7.5 million) of adults reported that they suffered from learning disabilities. Around 60 percent of these adults scored in Level 1, and some 22 percent scored in Level 2. Overall, the average scores for those self- reporting that they had a learning disability were: prose- 207; document- 203; and quantitative- 200.

Less than one-half of one-percent reported that they were mentally retarded. Eighty-six to 89 percent of these adults were placed in Level 1, with average scores of: prose-143; document-147; and quantitative 117.