Skip Navigation

Mark Schneider
Commissioner, National Center for Education Statistics

2003 National Assessment of Adult Literacy (NAAL) Results
December 15, 2005

Good morning. As the Commissioner of the National Center for Education Statistics, I am here today to release findings from the 2003 National Assessment of Adult Literacy, or NAAL.

Measuring literacy is a complicated task. One challenge is that we are testing adults, and you don't find them gathered in classrooms. Another challenge is that there are different types of literacy to measure. Yet another challenge is determining what it means to be literate. The report I am releasing today addresses these challenges and gives us solid data on the state of adult literacy in America.

Overview of the National Assessment of Adult Literacy (NAAL)
This assessment measures the English literacy of adults living in the United States. It's important to keep in mind that just because someone isn't literate in English doesn't mean he or she isn't literate in another language.

NAAL was last conducted in 1992. We are able to make comparisons to show how performance has changed since that time. We interviewed over 19,000 adults 16 or older in homes and prisons across the United States.

Our definition of literacy focuses on reading for a purpose. It emphasizes using printed and written information to function in society and to achieve one's goals and potential. This is somewhat different from a school-based definition of literacy, which focuses on learning to read and reading to learn.

NAAL measures three types of literacy using materials that are part of daily life:

  • Prose reading materials are arranged in sentences and paragraphs. Examples include the newspaper articles some of you will write on this report.
  • Documents are not organized in sentences. Documents include such things as bills, maps, bus schedules, and prescription labels.
  • Quantitative activities require simple calculations using numbers found in written materials such as tax forms or checking statements you receive from your bank.

Description of Literacy Levels
This assessment measures performance on a scale of 0 to 500. In addition, we use four main literacy levels based on work done by the National Research Council's Committee on Performance Levels for Adult Literacy.

Proficient means that someone can do complex activities such as comparing viewpoints in two editorials or interpreting a table about blood pressure and physical activity.

Intermediate means that a person can do moderately challenging tasks such as calculating the cost of an order from an office supply catalog or identifying a specific location on a map.

Basic means a person can perform simple and everyday tasks such as comparing the ticket price of two sporting events or understanding a pamphlet that describes how a person is selected for jury duty.

Below Basic indicates the lowest levels of performance such as signing a form or adding the amounts on a bank deposit slip.

There is also a fifth level for the population that is nonliterate in English, and I will talk about this later in more detail.

Before I present the results, it's important to remember that NAAL assessments are based on samples, and with samples there is a margin of error associated with each score. When we make comparisons between scores, or compare the percentages of adults at the various literacy levels, we test the differences to see if they are larger than the margin of error involved; if the differences are larger, they are statistically significant. These differences are indicated by asterisks in the tables and charts in the report.

Changes in Literacy: 1992-2003
I will first present the changes in literacy between 1992 and 2003. We reanalyzed the 1992 data for the prose, document and quantitative literacy scales using the 2003 procedures. Prose and document literacy did not change on average, while quantitative literacy increased.

Prose Literacy at Below Basic and Nonliterate in English Levels: 2003
Thirteen percent of all adults had Below Basic prose literacy. This translates into 30 million adults with Below Basic prose literacy-recall that these 30 million Americans cannot do much more than sign a form or search a simple document to find out what they are allowed to drink before a medical test.

The adults at the bottom of the Below Basic level did poorly on the easiest test questions. They represent three percent of the population.

This three percent of the population corresponds to seven million adults. These adults were considered to be nonliterate in English.

Another two percent of the population, or four million adults, couldn't take the test because of language barriers. We had interviewers who could speak English and spanish, but most of these adults spoke some other language. They also couldn't understand the interviewers when they tried to ask them about things such as their age and education. They are not included in the other results we're presenting today.

These four million adults, along with the seven million who did very poorly on simple test questions, are considered to be nonliterate in English.

A total of 11 million adults were nonliterate in English.

Next I'll present how many adults were in each of the literacy levels.

Percentage by Level: 1992-2003
Examining the percentage of adults in each of the four performance levels for each scale in both years, 14 percent of the population fell in the Below Basic category. Previously, I told you that only 13 percent were below basic. Why is that? Unlike the last set of data, these results are only for adults who could be tested. We haven't included the four million adults who couldn't be interviewed or tested because of language barriers. Thirteen percent of all adults had Below Basic literacy, but 14 percent of the adults who could be tested had Below Basic literacy skills.

One key finding I want to point out is that more adults scored in the Below Basic level on the quantitative scale (22 percent in 2003) than on the prose scale (14 percent) and on the document scale, at 12 percent.

Number of Adults in Each Prose Literacy Level: 1992-2003
The following results translate those percentages into actual numbers. Based only on adults who could be tested, we report 14 percent Below Basic.

As already mentioned, 30 million adults had Below Basic prose literacy. These adults can do no more than the most simple literacy activities.

Sixty-three million had Basic literacy. This means they are able to perform simple literacy activities such as understanding information in a pamphlet for prospective jurors.

Ninety-five million had Intermediate prose literacy. This means they can perform moderately difficult activities such as finding information in reference materials.

Twenty-eight million had Proficient literacy. This means they can perform complex and challenging literacy activities such as comparing viewpoints in two different editorials.

The number of adults with Below Basic literacy was similar on the document scale, but more adults had Below Basic quantitative literacy.

Characteristics of Adults Below Basic in Prose Literacy
The adults in Below Basic are different from adults in the population as a whole as reflected in the NAAL survey. In the general population, 15 percent of adults did not graduate from high school. Among adults with Below Basic prose literacy, 55 percent did not graduate from high school. To make this more concrete, adults who did not graduate from high school were almost four times more likely to be at the Below Basic level of performance as would be expected by chance.

In the population as a whole, 13 percent of adults did not speak any English before starting school. Among adults with Below Basic literacy, 44 percent did not speak any English before starting school. These adults were over three times more likely to have below basic literacy skills than we would expect by chance alone.

Results by Race/Ethnicity: 1992-2003
In 2003, White adults had the highest prose literacy, followed by Asian/Pacific Islanders, Blacks, and then Hispanics. White adults had higher quantitative literacy in 2003, while their prose and document literacy did not change significantly. Black adults had higher literacy on all three scales. Hispanics were the only group whose prose and document literacy scores decreased. This finding should be considered in light of demographic changes. The Hispanic population today is larger and different from the Hispanic population a decade ago. For example, there has been an increase in the number of Hispanic adults who are not native English speakers, as well as an increase in the percentage of Hispanics who were not born in the United States and who arrived here as teenagers or adults. Asian/Pacific Islanders had higher average prose literacy, with no changes for document and quantitative literacy.

Results by Gender: 1992-2003
Women are doing better and men are doing worse. Women have closed the gap with men in document literacy. Men are still scoring higher than women in quantitative literacy, but the gap between them is smaller. Women now do better than men in prose literacy.

Age Cohorts: 1992-2003
We separated adults into six age groups. Adults who were younger and adults who were older had lower quantitative literacy than adults in the middle. The oldest adults had the lowest literacy, but they are doing better than they were in 1992. This may be driven by the higher levels of formal education among the older cohorts. For example, in 1992 the over 65 cohort had 11 years of school; in 2003, the cohort of the same age had 12.3 years of schooling. The pattern of results was similar on the prose and document scales.

Prose Literacy Results by Educational Attainment: 2003
Overall, the more education adults had, the higher their prose literacy. Document and quantitative literacy showed the same pattern.

Now I will present the prose literacy results in terms of literacy levels. The average prose score of adults who did not graduate from high school was in the Below Basic level. Having Below Basic literacy means that they have no more than the most simple literacy skills. For example, they may be able to find a straightforward piece of information in a very simple pamphlet. The adults at the bottom of the Below Basic level can't even do that.

Now I'll talk about the Basic level for prose literacy. Having Basic literacy means that you can perform simple and everyday literacy activities, such as finding factual information in a newspaper article. The average prose score of adults who ended their education after obtaining a GED or graduating from high school was at the top of the Basic level.

Adults with Intermediate literacy can perform moderately challenging literacy activities. They can do things like read and understand articles about different political systems. The average prose score of adults with any postsecondary education fell into the Intermediate category.

Adults with Proficient literacy can perform more complex and challenging literacy activities. For example, they can critically evaluate information in legal documents. No group, not even those with the highest levels of formal education, had an average in the Proficient level on any of the scales.

It's important to remember that we are presenting averages here. There is a spread in scores among adults in all the education groups. For example, 41 percent of adults with graduate education had Proficient prose literacy.

Educational Attainment: 1992-2003
I will now present the results on change in scores between 1992 and 2003 for selected educational attainment levels. There were no increases in literacy in any of any of the educational attainment levels. Prose literacy decreased among adults at every level of education. This decrease calls out for more research. On the quantitative scale, there were no changes in literacy at any level of educational attainment. For document literacy, those with higher levels of education showed a decline while those with less education had no change. With scores dropping in prose literacy for every level of education, you might wonder why there was no overall decline in the average score for this type of literacy. This is because adults with higher educational levels tend to outperform those with lower educational levels, and the percentage of adults with high educational levels-those with "some college" or more-has been increasing, while the percentage with low levels of education has been declining. We have more higher-scoring adults with high levels of education, and fewer lower scoring adults with low levels of education, which offsets the fact that average scores for highly educated adults are declining.

Employment and Earnings
We also find an association between levels of literacy, employment, and earnings. At each increasing level of prose literacy, more adults were employed full time. Median weekly earnings also increased with each level of literacy.

Closing Remarks
This concludes my remarks on the 2003 National Assessment of Adult Literacy. There is much more information in the NAAL First Look report and the Key Concepts report, and even more on the NAAL website (

Powerpoint presentation for National Assessment of Adult Literacy - A First Look at the Literacy of America's Adults in the 21st Century. (PPT 2,365KB)

I'd like to thank the adults who participated in the 2003 assessment. I would also like to thank the members of the National Research Council's Committee on Performance Levels for Adult Literacy who developed the levels we are using for reporting these results, as well as everyone else who contributed to the NAAL.

Thank you.