HomeLarge Type Edition
HOME ARCHIVE SEARCH TABLE OF CONTENTS
This Article
Right arrow Abstract Freely available
Right arrow Full Text (PDF)
Services
Right arrow Download to citation manager
Citing Articles
Right arrow Citing Articles via HighWire
PubMed
Right arrow PubMed Citation
The Journals of Gerontology Series A: Biological Sciences and Medical Sciences 59:M390-M395 (2004)
© 2004 The Gerontological Society of America

The Relationship Between Literacy and Cognition in Well-Educated Elders

Deborah E. Barnes1,, Ira B. Tager2, William A. Satariano2 and Kristine Yaffe3

1 Department of Medicine, University of California, San Francisco.
2 Department of Epidemiology, University of California, Berkeley.
3 Departments of Psychiatry, Neurology, and Epidemiology, University of California, San Francisco.

Address correspondence to Deborah E. Barnes, PhD, San Francisco VA Medical Center, 4150 Clement St. (MC 181G), San Francisco, CA 94121. E-mail: barnes{at}medicine.ucsf.edu


   Abstract
Top
 Abstract
Methods
 Results
 Discussion
 References
 
Background. Literacy is correlated with general intelligence and is often used to estimate premorbid intelligence in persons with dementia. However, little is known about the relationship between literacy and specific cognitive domains.

Methods. Study participants were 664 community-living adults aged 65 years or older who were participating in a community-based study of health and function in Sonoma, California. Literacy was measured using the North American Adult Reading Test, which evaluates the ability to pronounce words with irregular spellings (such as indict). Cognitive function was assessed using a neuropsychological test battery that included the Mini-Mental State Examination and measures of attention and executive function (Trails B, Stroop, Digit Symbol), verbal learning and memory (California Verbal Learning Test), and verbal fluency (letter "s," animals).

Results. The mean age of the participants was 76 years, 50% were women, 97% were white, and 92% had 12 or more years of education. A strong, linear association was observed between literacy and all cognitive measures (all p <.001). Results were similar after adjustment for age, sex, education, and health-related covariates and were consistent in subgroups of the study population (e.g., women vs men; English vs other native languages). Education was not associated with most cognitive measures after adjustment for literacy.

Conclusions. Literacy is strongly associated with cognitive function across all cognitive domains in well-educated, elderly white persons. Future studies should determine whether interventions to improve lifetime literacy may help prevent cognitive deterioration with age.

WORD-READING ability is strongly correlated with general intelligence in adults without dementia (1) and is typically well maintained in elderly persons with mild cognitive impairment or early dementia (2). Therefore, literacy is often used to estimate premorbid intelligence in elderly persons with dementia or other cognitive disorders.

However, less is known about the relationship between literacy and specific domains of cognitive function, particularly in elderly persons without dementia. It is possible that literacy is correlated more strongly with measures of language ability and less strongly with other aspects of cognition, such as attention or executive function (the ability to plan and carry out goal-directed behavior). Alternatively, literacy could be associated equally with all cognitive domains and, therefore, could be a general marker of intelligence or cognitive reserve (3,4).

Most previous studies performed in elderly people have found that literacy is associated with global measures of intelligence [e.g., Wechsler Adult Intelligence Test–Revised (5–8)] or cognition [e.g., Mini-Mental State Examination (MMSE) (9–12)]. One of the few studies that examined multiple cognitive domains (13) was restricted to elders with 0–3 years of higher education, and, therefore, the general application of its results is limited.

Most previous studies have also focused on the effect of low literacy on cognitive function. For example, two studies used a literacy measure in which the highest category was associated with an 8th Grade (9) or 9th Grade (10) reading level. However, literacy reflects a continuum that ranges from those who are illiterate to those who are highly literate (i.e., prolific readers with advanced vocabularies), and it is important to study both ends of the spectrum.

The goal of our study was to describe the association between literacy and cognitive function across multiple cognitive domains in a cohort of well-educated elders. We also considered the effects of education, age, sex, and health-related covariates on the relationship between literacy and cognitive function.


   METHODS
Top
 Abstract
 Methods
Results
 Discussion
 References
 
Study Population
Study participants were part of a larger, longitudinal study of health and function in 2092 community-living adults aged 55 years or older in Sonoma, California (14). Cohort members have been interviewed on an approximately biennial basis since 1993. In 1999, a detailed cognitive test battery was administered to a random, sex-stratified subsample of the cohort (15). Cohort members were considered eligible if they had completed a full interview at the third evaluation (approximately 1 year earlier), still resided in the area, were at least aged 65 years on July 1, 1999, could speak and read English, and did not require a proxy respondent. Of the 824 adults who were invited to participate, 664 (81%) agreed. Those who refused to participate in the cognitive substudy did not differ from those who agreed with respect to age, sex, or frequency of symptoms and medical conditions, but they had slightly fewer years of education (mean, 14.2 vs 14.7 years; p =.07). Our analytic sample is the 664 cohort members who participated in the cognitive substudy.

The Committee for the Protection of Human Subjects at the University of California, Berkeley, approved all study procedures, and all participants gave written informed consent.

Literacy
We measured literacy using the North American Adult Reading Test (NAART) (16), which consists of a list of 61 words with irregular spellings. Participants were asked to pronounce each word, and pronunciations were scored as correct or incorrect based on a standardized key that accounts for regional differences in pronunciations. Scores may range from 0 (all incorrect) to 61 (all correct). Words range in difficulty from debt to synecdoche. Previous studies have found that test–retest reliability is.98 and interrater reliability is.88 (17).

Performance on the NAART is relatively well maintained in elderly persons with mild cognitive impairment and early stage dementia (2). Therefore, the NAART is often used in clinical settings to estimate premorbid intelligence in persons with dementia or other cognitive disorders.

Cognitive Function
Four domains of cognitive function were evaluated: 1) global, 2) executive function and attention, 3) verbal memory, and 4) verbal fluency. Global cognitive status was measured using the MMSE (18), which is a 30-point test that evaluates (a) orientation to time and place, (b) registration and recall of 3 words, (c) attention and calculation, (d) language ability, and (e) visuospatial ability.

Executive function refers to the ability to plan and carry out goal-directed activities, and it is closely linked to the ability to focus on a given task (attention) (17). Measures of executive function and attention included (a) the Trail-Making Test, Part B (19), in which the participant must draw a continuous line alternating between letters and numbers arranged randomly on a page (i.e., 1-A-2-B); (b) the Stroop Interference Test (20), in which the participant must name the color of ink in which words are printed (e.g., the word green printed in red ink); and (c) the Digit Symbol test (21), in which the participant must fill in blank squares based on a key in which numbers are paired with nonsense symbols. Analyses were based on the number of seconds to completion for the Trails B and Stroop tests and on the number of squares filled in correctly within 2 minutes for the Digit Symbol test.

Verbal memory was evaluated using the California Verbal Learning Test (22). Participants were given 5 trials to learn a 16-word list and then asked to recall the words after a brief interference and a 20-minute delay. Analyses were based on the total number of words learned and the total number recalled after 20 minutes.

Verbal fluency was measured using the Controlled Oral Word Association Test (17). Participants were asked to name as many different words as possible beginning with the letter "s" in 1 minute and then to name as many different animals as possible in 1 minute. Analyses for both tests were based on the number of unique words named.

Other Measures
Other measures were selected because they have been associated with cognitive test performance in other studies. We measured educational attainment by asking participants how many grades they had completed in school, including grade school, high school, college, or technical-vocational training. We also asked participants whether they currently smoked cigarettes; the frequency and usual quantity consumed for beer, wine, and other alcoholic beverages; and their native language. Depressive symptoms were assessed using the Center for Epidemiologic Studies-Depression Scale (23), and the standard cutoff of >=16 points was used to classify participants as depressed. History of comorbid medical conditions was evaluated by asking participants whether they had ever been told by a physician that they had any condition included on a list of specific conditions (e.g., hypertension, diabetes, stroke, heart attack).

Statistical Analysis
Scatter diagrams were used to examine the linearity of relationships among literacy, education, and measures of cognition. Both linear models and smoothed models (smoothing spline with 4 degrees of freedom) were fitted to the data. Smoothing allows visualization of relationships between variables without making assumptions about the shape of the relationship.

We examined raw cognitive test scores by literacy tertiles. Chi-square tests for trend were performed to determine whether cognitive test scores changed in an incremental manner over literacy tertiles. We also converted cognitive test scores to z-scores ([observed value for participant i – mean for all participants]/standard deviation [SD]) so that units would be comparable across cognitive tests. A study participant with a z-score of 1 on a given cognitive test would have scored 1 SD above the mean, whereas a study participant with a z-score of –1 would have scored 1 SD below the mean.

We performed a series of linear regression analyses to evaluate relationships among literacy, education, and cognitive function: (a) cognitive z-scores as a function of literacy only; (b) cognitive z-scores as a function of years of education only; and (c) cognitive z-scores as a function of literacy and education together. Results are reported as the increase in z-score per 10 words pronounced and per 5 years of education. We performed additional analyses to determine whether our results were affected by age, sex, native language, or health-related covariates such as alcohol consumption, smoking, depression, or medical conditions.

We also performed a series of sensitivity analyses to determine whether our results differed among subgroups of the study population. Specifically, we repeated our analyses after stratifying the cohort by sex, age (<75 years vs >=75 years), educational attainment (<16 years vs >=16 years), and native language (English vs other languages).


   RESULTS
Top
 Abstract
 Methods
 Results
Discussion
 References
 
Table 1 shows the characteristics of the 664 study participants. The mean age of the participants was 76 years, 50% were women, and 97% were white. Ninety-two percent had 12 or more years of education. Although all study participants could read and write English, English was not their native language in 13%.


View this table:
[in this window]
[in a new window]
  		Table 1. Characteristics of 664 Study Participants.

 
NAART literacy scores ranged from 7 to 60 (mean, 38; SD, 11). Study participants in the lowest, middle, and highest literacy tertiles pronounced a mean of 26 words (range, 7–33), 39 words (range, 34–44), and 50 words (range, 45–60) correctly, respectively. Literacy and education were strongly correlated (r =.51). In addition, participants with lower literacy scores were more likely to have a native language that was not English, to have a history of diabetes mellitus, and to be currently depressed based on the Center for Epidemiologic Studies-Depression Scale (all p <=.05).

Cognitive test scores increased progressively with literacy tertile for all measures (Table 2). For example, study participants in the lowest, middle, and highest literacy tertiles scored 27.9, 28.7, and 29.1 points on the MMSE and remembered 6.9, 7.5, and 8.3 words on the California Verbal Learning Test, respectively (Table 2). In addition, the strength of the relationship between literacy and cognitive function was remarkably consistent across cognitive domains (Figure 1). Smoothed graphs revealed that the association between literacy and cognitive function was linear for all cognitive measures (Figure 2).


View this table:
[in this window]
[in a new window]
  		Table 2. Association Between Cognitive Function and Literacy Tertile.

 


View larger version (18K):
[in this window]
[in a new window]
  		Figure 1. Z-scores on cognitive tests are shown as a function of literacy tertile. Lowest, middle, and highest literacy tertiles reflect mean of 26 words (range, 7–33), 39 words (range, 34–44), and 50 words (range, 45–60), respectively, pronounced correctly on the North American Adult Reading Test. Elders in the lowest literacy tertile scored 0.2–0.4 standard deviations (SD) below the mean on all cognitive tests, while elders in the highest literacy tertile scored 0.2–0.4 SD above the mean (all p <.001 for trend over tertiles). "Missing" bars reflect z-score = 0 (i.e., performance at the mean)

 


View larger version (15K):
[in this window]
[in a new window]
  		Figure 2. A linear association was observed between the North American Adult Reading Test literacy scores and all cognitive measures data for Digit Symbol scores are shown for illustrative purposes. The solid line reflects a linear relationship estimated using least squares regression. The dotted line reflects a smoothed relationship estimated using smoothing splines with 4 degrees of freedom

 
Higher literacy and more years of education were both associated with better cognitive function in unadjusted linear regression analyses (all p <=.05, Table 3). For example, MMSE scores increased by approximately 0.3 SD for every 10 words correctly pronounced on the NAART and for every 5 years of education (Table 3). However, when we considered literacy and education together in the same linear regression models, literacy was strongly associated with all 8 cognitive measures, whereas education was weakly associated with 3 measures and not associated with 5 measures. These findings were unchanged by further adjustment for age, sex, native language, depression, and medical diagnoses (Table 3).


View this table:
[in this window]
[in a new window]
  		Table 3. Cognitive Performance as a Function of Literacy and Education.

 
Our findings were similar in different subgroups of the study population (e.g., men vs women; age <75 years vs >=75 years; education <16 years vs >=16 years; English vs other native languages) and when we excluded participants who had low MMSE scores (<=23). In all subgroups, literacy was strongly associated with all measures of cognition in both unadjusted and adjusted analyses, whereas education was no longer associated with most cognitive measures after adjustment for literacy.


   DISCUSSION
Top
 Abstract
 Methods
 Results
 Discussion
References
 
We found that elderly persons with higher literacy scores performed better on all measures of cognition. The association between literacy and cognition was linear, was largely unchanged by adjustment for years of education and other covariates, and was consistent across different subgroups of the cohort. Our elderly participants with higher levels of education also performed better on most cognitive measures. However, when we included both literacy and education in the same statistical models, cognitive scores were associated much more strongly and consistently with literacy than with education.

Our results are consistent with those of several previous studies that evaluated the relationship between literacy levels and measures of global cognitive function (9–12,24) or intelligence (6,7). For example, Weiss and colleagues (9) found that MMSE scores were correlated more strongly with reading level (0 to >= 8th grade) than with years of education, age, or ethnicity in 214 seniors living in public assistance housing. Our results extend the findings of previous research to show that the relationship between literacy and cognitive function is present at higher literacy levels and is consistent across multiple cognitive domains.

There are several possible explanations for the association between literacy and cognitive function. The association could be a test bias, in which elderly persons with higher literacy levels perform better on cognitive tests because they are more adept at taking tests in general. However, we believe this is unlikely because our study population did not include any persons who were illiterate, and most study participants had graduated from high school. Furthermore, our results were similar when we restricted our analyses to persons with 16 or more years of education (i.e., had received a college degree). Presumably, a test bias would be minimized in elders with high levels of education.

Literacy also could reflect a person's innate level of cognitive ability or the effect of early life exposures and educational experiences. Previous studies have found that better mental ability at age 11 years (25) and more complex sentence structure at age 22 years (26) are associated with a reduced risk for dementia.

Alternatively, literacy could be causally associated with better cognitive function or could be a marker of other causal factors. For example, reading and writing could lead directly to enrichment of neural networks, which could enable persons with higher levels of literacy to have larger cognitive reserves and to process cognitive information more efficiently (3).

Literacy also could be a marker of crystallized intelligence, which reflects, in part, the effects of formal and informal educational experiences during a person's lifetime and the degree to which a person participates in mentally stimulating activities. One study found that elderly persons with higher levels of intelligence, as estimated by a measure of literacy, were less likely to develop incident dementia (27). Similarly, high levels of educational attainment (28) and cognitive stimulation (29) have been associated with a reduced risk for dementia in previous studies.

Our study has several limitations. First, these results describe the cross-sectional relationship between literacy and cognition. Longitudinal studies should be conducted to determine whether literacy is associated with preservation of cognitive abilities over time. Second, our study population was primarily white and had a relatively high socioeconomic status. Although this was a strength in terms of the internal validity of the study, it is unclear whether our results can be generalized to other groups. However, 1 study found that reading level accounted for most of the difference in cognitive test scores between African Americans and white persons matched by years of education (24), which suggests that literacy also is an important predictor of cognitive function in African Americans.

Conclusion
We found that literacy was associated with cognitive test performance in a linear manner across all cognitive domains. In addition, literacy was stronger than education as a predictor of cognitive function. These results support the use of literacy as a marker of general intelligence in clinical and research settings. Future studies should determine whether interventions to improve literacy are associated with prevention of cognitive decline in elderly people.


   Acknowledgments
 
Supported by grants from the National Institute on Aging (RO1-AG09389, T32-AG00212) and a predoctoral National Research Service Award from the National Institute of Mental Health (F31-MH12665). Dr. Yaffe was supported by a Patient-Oriented Research Development Award from the National Institute on Aging (K23-AG00888), and she is a Paul Beeson Faculty Scholar in Aging Research.

Preliminary results were presented at the Gerontological Society of America meeting on November 24, 2002, in Boston, Massachusetts, and at the American Academy of Neurology meeting on April 3, 2003, in Honolulu, Hawaii.

Received March 6, 2003

Accepted April 23, 2003


   References
Top
 Abstract
 Methods
 Results
 Discussion
 References
 

  1. Nelson HE, McKenna P. The use of current reading ability in the assessment of dementia. Br J Soc Clin Psychol.. 1975;14:259-267.[Medline]
  2. Fromm D, Holland AL, Nebes RD, Oakley MA. A longitudinal study of word-reading ability in Alzheimer' disease: evidence from the National Adult Reading Test. Cortex.. 1991;27:367-376.[Medline]
  3. Stern Y. What is cognitive reserve? Theory and research application of the reserve concept. J Int Neuropsychol Soc.. 2002;8:448-460.[Medline]
  4. Katzman R. Education and the prevalence of dementia and Alzheimer's disease. Neurology.. 1993;43:13-20.[Free Full Text]
  5. Nelson HE, O'Connell A. Dementia: the estimation of premorbid intelligence levels using the New Adult Reading Test. Cortex.. 1978;14:234-244.[Medline]
  6. Boekamp JR, Strauss ME, Adams N. Estimating premorbid intelligence in African-American and white elderly veterans using the American version of the National Adult Reading Test. J Clin Exp Neuropsychol.. 1995;17:645-653.[Medline]
  7. Crawford JR, Deary IJ, Starr J, Whalley LJ. The NART as an index of prior intellectual functioning: a retrospective validity study covering a 66-year interval. Psychol Med.. 2001;31:451-458.[Medline]
  8. Bright P, Jaldow E, Kopelman MD. The National Adult Reading Test as a measure of premorbid intelligence: a comparison with estimates derived from demographic variables. J Int Neuropsychol Soc.. 2002;8:847-854.[Medline]
  9. Weiss BD, Reed R, Kligman EW, Abyad A. Literacy and performance on the Mini-Mental State Examination. J Am Geriatr Soc.. 1995;43:807-810.[Medline]
  10. Mayeaux EJ, Jr, Davis TC, Jackson RH, et al. Literacy and self-reported educational levels in relation to Mini-Mental State Examination scores. Fam Med.. 1995;27:658-662.[Medline]
  11. Albert SM, Teresi JA. Reading ability, education, and cognitive status assessment among older adults in Harlem, New York City. Am J Public Health.. 1999;89:95-97.[Abstract/Free Full Text]
  12. Brayne C, Beardsall L. Estimation of verbal intelligence in an elderly community: an epidemiological study using NART. Br J Clin Psychol.. 1990;29:(Pt 2): M217-223.
  13. Manly JJ, Jacobs DM, Sano M, et al. Effect of literacy on neuropsychological test performance in nondemented, education-matched elders. J Int Neuropsychol Soc.. 1999;5:191-202.[Medline]
  14. Satariano WA, Smith J, Swanson A, Tager IB. A census-based design for the recruitment of a community sample of older adults: efficacy and costs. Ann Epidemiol.. 1998;8:278-282.[Medline]
  15. Barnes DE. Does physical activity protect against cognitive decline in older adults? [Doctoral Dissertation]. University of California, Berkeley: Department of Epidemiology; 2001.
  16. Blair J, Spreen O. Predicting premorbid IQ: a revision of the National Adult Reading Test. Clin Neuropsychologist.. 1989;3:129-136.
  17. Spreen O, Strauss E. A Compendium of Neuropsychological Tests: Administration, Norms, and Commentary. New York: Oxford University Press; 1998.
  18. Folstein MF, Folstein SE, McHugh PR. Mini-Mental State: a practical method for grading the cognitive state of patients for the clinician. J Psychiat.. 1975;12:189-198.
  19. Reitan R, Wolfson D. The Halstead-Reitan Neuropsychological Test Battery. Tucson, AZ: Neuropsychology Press; 1985.
  20. Stroop J. Studies of interference in serial verbal reaction. J Exp Psychol.. 1935;18:643-662.
  21. Wechsler D. Wechsler Adult Intelligence Scale-Revised. New York: The Psychological Corporation; 1981.
  22. Delis D, Kramer J, Kaplan E, Ober B. California Verbal Learning Test: Adult Version Manual. San Antonio, TX: The Psychological Corporation; 1987.
  23. Radloff LS. A self-report depression scale for research in the general population. Appl Psychol Meas.. 1977;1:385-401.
  24. Manly JJ, Jacobs DM, Touradji P, Small SA, Stern Y. Reading level attenuates differences in neuropsychological test performance between African American and white elders. J Int Neuropsychol Soc.. 2002;8:341-348.[Medline]
  25. Whalley LJ, Starr JM, Athawes R, Hunter D, Pattie A, Deary IJ. Childhood mental ability and dementia. Neurology.. 2000;55:1455-1459.[Abstract/Free Full Text]
  26. Snowdon DA, Kemper SJ, Mortimer JA, Greiner LH, Wekstein DR, Markesbery WR. Linguistic ability in early life and cognitive function and Alzheimer's disease in late life. Findings from the Nun Study. JAMA.. 1996;275:528-532.[Abstract/Free Full Text]
  27. Schmand B, Smit JH, Geerlings MI, Lindeboom J. The effects of intelligence and education on the development of dementia. A test of the brain reserve hypothesis. Psychol Med.. 1997;27:1337-1344.[Medline]
  28. Stern Y, Gurland B, Tatemichi TK, Tang MX, Wilder D, Mayeux R. Influence of education and occupation on the incidence of Alzheimer's disease. JAMA.. 1994;271:1004-1010.[Abstract/Free Full Text]
  29. Wilson RS, Mendes De Leon CF, et al. Participation in cognitively stimulating activities and risk of incident Alzheimer disease. JAMA.. 2002;287:742-748.[Abstract/Free Full Text]



This article has been cited by other articles:


Home page
 GerontologistHome page
D. Morrow, D. Clark, W. Tu, J. Wu, M. Weiner, D. Steinley, and M. D. Murray
Correlates of health literacy in patients with chronic heart failure.
Gerontologist, October 1, 2006; 46(5): 669 - 676.
[Abstract] [Full Text] [PDF]

This Article
Right arrow Abstract Freely available
Right arrow Full Text (PDF)
Services
Right arrow Download to citation manager
Citing Articles
Right arrow Citing Articles via HighWire
PubMed
Right arrow PubMed Citation

HOME ARCHIVE SEARCH TABLE OF CONTENTS