Physical inactivity is a major independent risk factor for chronic disease, and considerable evidence now exists that people of racial/ethnic minority groups share a disproportionate burden of a sedentary lifestyle compared with their white counterparts (26). In addition, data from three surveillance systems for leisure-time physical activity (6,9,19) identified Hispanic and non-Hispanic blacks as being more inactive than other racial/ethnic minority groups and reported that trends were often exacerbated among women (30). These findings led to the development of national objectives that prioritize population-wide increases in moderate- and vigorous-intensity physical activity and elimination of disparities between physical activity for whites and racial and ethnic minority groups (25). Evidence suggests progress toward these objectives. The prevalence of leisure-time inactivity among most racial/ethnic groups seems to be decreasing, but the rate of decrease has been slowest among Hispanics and non-Hispanic blacks (14).
The reasons for racial/ethnic differences in leisure-time inactivity remain unclear; in some cases, cultural attitudes about desirable and healthy body weight may influence physical activity behavior (10). However, focusing on racial/ethnic differences in the prevalence of physical activity downplays the importance of social class, which is correlated with physical activity (30) and is known to be an independent determinant of health status (12,13,15,31). Relationships among race/ethnicity, social class, physical activity, and disease are complex, but it has been hypothesized that social class may moderate the relationship between race/ethnicity and physical activity (9) and that physical activity may mediate the relationship between race/ethnicity and disease (16,27).
Measures of social class in studies of leisure-time physical activity are often based on single indicators of occupation, education, or income. Because social class indicators may be both independent and interdependent, epidemiologists have recommended the use of multiple indicators, especially if the conceptual relevance is unclear (17,21). Few studies in the literature on physical activity have measured social class by using multiple indicators, and only one (9) examined social class as a moderating factor in the relationships between race/ethnicity and leisure-time physical inactivity. Using data from the Third National Health and Nutrition Examination Survey, Crespo and colleagues (9) found that differences in leisure-time inactivity among racial/ethnic groups were not explained entirely by education, occupation, employment, poverty status, and marital status. However, a limitation of that study was that occupational physical activity was not measured. This is an important limitation because the amount of physical activity accrued during leisure time has been shown to be inversely related to the amount of physical activity accrued at work and because occupational physical activity is known to differ by social class and race/ethnicity (6).
To elucidate the role of social class in the relationship between race/ethnicity and leisure-time physical inactivity, researchers need more data from nationally representative population samples that contain multiple indicators of social class. Assessment of occupational physical activity is also important. The purposes of this study were to determine 1) the prevalence of leisure-time physical inactivity in a nationally representative sample of non-Hispanic white, non-Hispanic black, and Hispanic men and women; 2) the prevalence of inactivity between these racial/ethnic groups across different indicators of social class; and 3) the relationship between occupational physical activity and leisure-time physical inactivity among these racial/ethnic groups, independent of other social class indicators.
METHODS
The National Physical Activity and Weight Loss Survey (NPAWLS) was a nationwide telephone survey conducted between September and December 2002. All research procedures were approved by the institutional review board at the University of South Carolina. Survey questions included items pertaining to overall health status and quality of life, weight control measures, and participation in physical activity. The overall objective of the study was to obtain data on individual physical activity and nutrition behaviors that present health risks. The sample of respondents was drawn from the total noninstitutionalized U.S. adult population residing in telephone-equipped locations. Excluded from the referent population were institutionalized adults and those who lived in group quarters with 10 or more unrelated residents, had no telephone, or did not speak English or Spanish well enough to be interviewed.
Sample Design
The sample population was obtained by random digit dialing (RDD) of telephone-equipped households in the United States. A list of all operating telephone exchanges within each U.S. area code was used. These telephone exchanges, combined with all four-digit numbers from 0000 to 9999, were divided into blocks of 100 numbers; each block was examined to ensure that it contained at least one residential number. Qualifying blocks were combined to create the sampling frame; numbers were systematically sampled from this frame. Nonresidential, nonworking, fax, and modem numbers were filtered to increase the connection rate of telephone interviews with potential respondents.
The study design targeted Hispanic and non-Hispanic black respondents. A replicate design mixing telephone numbers from three independent samples was used to achieve target percentages of racial/ethnic groups in the final sample. The first independent sample was drawn from a national RDD list. The second and third samples were drawn from frames constructed to yield higher percentages of persons from targeted minority groups. These two samples, one targeting Hispanic households and the other targeting non-Hispanic black households, were constructed by using subsets of telephone exchanges in the national frame. Exchanges were included in each subset only if incidence of the targeted ra ce/ethnicity of the households was greater than 20%. Ethnic household incidence was determined by area code exchange and census-based demographic information.
Selection of Respondents
A minimum of 15 attempts were made during at least 5 d to reach each telephone number. Interviewers asked the number of adults 18 yr of age or older in the contacted household; then they asked for a listing of all men and women in the household. The survey respondent was randomly selected from this roster. Survey interviews were completed with 11,211 respondents, 4695 men, and 6516 women. The response rate was 30.9%, and the cooperation rate (completed interviews/[refusals + terminations + completed interviews]) was 54.3%, by calculations using CASRO (7).
Weighting
The weights for the NPAWLS data were calculated as the product of three components: 1) a sampling weight accounting for differential probabilities of selection, defined as the inverse of the product of the probability of household selection and the probability of respondent selection; 2) a poststratification factor adjusting weight totals to population figures from the 2000 U.S. Census (24), by age, race/ethnicity, and sex, computed for the three independent samples, so that the sum of the adjusted weights for each sample would produce an estimate of the eligible U.S. population; and (3) a factor scaling the weights for each sample to allow for estimates based on the combined data that would not overestimate population totals.
Assessment and Definition of Physical Inactivity
Self-reported leisure-time physical activity was assessed by using the questions from the 2001 Behavioral Risk Factor Surveillance System (BRFSS) survey (4). This approach incorporates multiple domains (e.g., leisure, household, transportation) in the definition of leisure-time physical activity. The question about moderate physical activity was: "In a usual week, do you do moderate activities for at least 10 min at a time, such as brisk walking, bicycling, vacuuming, gardening, or anything else that causes small increases in breathing or heart rate?" The question about vigorous physical activity was: "In a usual week, do you do vigorous activities for at least 10 min at a time, such as running, aerobics, heavy yard work, or anything else that causes large increases in breathing or heart rate?" Participants who responded "no" to both questions were classified as physically inactive. This approach differs from the method used in BRFSS to track physical inactivity (14).
Occupational physical activity was assessed with three questions about the time spent 1) sitting or standing, 2) walking, or 3) performing heavy labor at work:
- "In a usual day, do you do any sitting or standing while working, such as desk work, using hand tools, light assembly, laboratory technician, or driving a car or truck for work?"
- "In a usual day, do you do any walking at work, such as walking in the halls, postal carrier, waiter, or roving salesperson?"
- "In a usual day, do you do any heavy labor or use power tools at work, such as moving furniture, carpentry, jackhammers, or using a shovel or pick?"
For each type of activity, respondents reported the number of hours per day they participated in the activity. Respondents who reported that they were retired, were unable to work, or worked less than 20 h·wk−1 were excluded from the analyses involving occupational physical activity.
The number of hours spent sitting, walking, or performing heavy labor at work were multiplied by the metabolic equivalents (MET) of 1.5, 3.0, and 7.0, respectively. One MET is roughly equivalent to the energy cost of sitting quietly and is defined as the activity metabolic rate divided by the resting metabolic rate (1), expressed as kilocalories per kilogram per hour. The MET hours per week spent sitting, walking, and performing heavy labor were summed to compute total MET hours per week at work. This variable was then transformed into quartiles specific to each racial/ethnic group.
Assessment of Social Class
Indicators of social class used in this study were education level, family income, employment status, and marital status. Education level was based on the number of years of schooling, in four categories: less than high school, high school graduate, some college, and college graduate. Family income was based on the annual household income from all sources in four categories: <$25,000; $25,000 to $49,999; $50,000 to $74,999; and ≥$75,000. Employment status was classified into three categories: employed for wages, not employed for wages, and not in the labor force. Marital status was classified in two categories: married or partnered, and not married and not partnered.
Data Analysis
All statistical analyses were performed with the use of SAS (22) and SUDAAN version 8.0 (23) to incorporate the survey's complex sampling design and the sampling weights described here. Age-adjusted prevalence estimates were computed by using the direct method based on the 2000 census population as the standard. Multivariate logistic regression was used to calculate odds ratios (OR) and 95% confidence intervals (CI) adjusted for age, sex, and social class indicators. Parameter estimates were obtained by using maximum likelihood techniques. A significant difference (nonequivalence) between two prevalence estimates was determined by nonoverlapping 95% CI. Using the CI as an indicator of effect provides more information than the P value in that it gives a range of possible values within which the value falls (22). An OR was considered indicative of no effect if its 95% CI did not include 1.0.
RESULTS
The NPAWLS sample included 9806 adults (4140 men and 5666 women), of which approximately 72% were white. Demographic characteristics of the sample are presented in Table 1.
;)
TABLE 1: Demographic characteristics of participants in the National Physical Activity and Weight Loss Survey, 2002-2003.
Across all racial/ethnic groups, the age-adjusted prevalence of leisure-time physical inactivity was 12.4% ± 0.6 (standard error, SE) for men and 15.1% ± 0.6 for women The age-adjusted prevalence of leisure-time inactivity was 9.9% ± 0.6 and 12.0% ± 0.6 for white men and women, respectively; 19.0% ± 2.5 and 25.2% ± 2.1 for non-Hispanic black men and women, and 20.9% ± 2.1 and 27.3% ± 2.5 for Hispanic men and women (Table 2).
TABLE 2: Prevalence of physical inactivity during leisure time among white, non-Hispanic black, and Hispanic men and women, aged 18 yr or older, National Physical Activity and Weight Loss Survey, 2002-2003.
The age-adjusted prevalence of leisure-time physical inactivity was higher in women than in men. For both men and women, non-Hispanic blacks and Hispanics had similar levels of inactivity, but both racial/ethnic groups were more inactive than their white counterparts. The lowest prevalence of leisure-time inactivity was among white men (10%), and the highest prevalence was among Hispanic women (27%). For all racial/ethnic groups, the highest prevalence of inactivity was among those 65 yr and older.
The prevalence of leisure-time inactivity among men and women across different indicators of social class is shown in Tables 3 and 4. Men with lower levels of education and household income tended toward a higher prevalence of inactivity (Table 3). However, within social class strata, overlapping CI among racial/ethnic groups in the prevalence of inactivity were more evident. Hispanic men who had less than a high school education or earned less than $25,000 per year were more inactive than white men of the same status. Non-Hispanic black men who were college graduates had the lowest prevalence of inactivity (4%) of all the racial/ethnic and social class groups, including whites or Hispanics with the same level of education. Hispanic men who were married or partnered or employed for wages were more inactive than white men of the same status. Within the other strata of social class, differences among racial/ethnic groups in the prevalence of leisure-time physical inactivity were not considered significant because of overlapping CI.
TABLE 3: Age-adjusted prevalence of physical inactivity during leisure time among white, non-Hispanic black, and Hispanic men, by indicators of social class, National Physical Activity and Weight Loss Survey, 2002-2003.
TABLE 4: Age-adjusted prevalence of physical inactivity during leisure time among white, non-Hispanic black, and Hispanic women, by indicators of social class, National Physical Activity and Weight Loss Survey, 2002-2003.
Regardless of race/ethnicity, women who were not in the workforce, had less than a high school education, or had a yearly household income less than $25,000 had a higher prevalence of leisure-time inactivity. Whereas Hispanic and non-Hispanic black women had a higher prevalence of inactivity than white women overall, few differences (95% CI overlap) were evident among women within the same social class strata. However, non-Hispanic black women with some college education had a higher prevalence of inactivity (24.3%; 95% CI 16.7, 31.9%) than white women (13.9%; 95% CI 11.5, 16.3%) with the same level of education. The prevalence of leisure-time inactivity was higher among non-Hispanic black and Hispanic women who were not employed for wages than that for white women with the same employment status. Finally, non-Hispanic black women who were married or partnered had a higher prevalence of inactivity than did whites of the same status.
The amount of occupational physical activity was related to the race/ethnicity of the respondent; Hispanics expended more energy at work than did other racial/ethnic groups (Pearson chi-square = 73.5, P < 0.01). For each racial/ethnic group, the crude and adjusted odds of being inactive during leisure time across quartiles of occupational physical activity are presented in Table 5. The analysis was based on data from 5333 participants who reported doing paid work for at least 20 h·wk−1. This group constituted 54.4% of the 9806 men and women in the study. To preserve cell sizes, data are presented as a combined sample of men and women.
TABLE 5: Odds of leisure-time physical inactivity based on levels of occupational physical activity, by racial/ethnic group, National Physical Activity and Weight Loss Survey, 2002-2003.
The likelihood of being physically inactive during leisure time was not significantly different across quartiles of occupational physical activity. This finding was unchanged after adjustment for age, sex, and indicators of social class in the logistic regression model. One exception was that Hispanics in the highest quartile of occupational physical activity (> 160.7 MET·h·wk−1) were more likely to be inactive during leisure time than Hispanics in the lowest quartile of occupational physical activity (< 60 MET·h·wk−1; odds ratio 2.18, 95% CI 1.07, 4.45). For this analysis, values were adjusted only for age and sex. However, this finding was no longer evident after adjustment for social class variables.
DISCUSSION
Our data show that the age-adjusted prevalence of leisure-time inactivity was lower in white men and women than in non-Hispanic black and Hispanic men and women. Our estimates are consistent with contemporary surveillance data on physical activity that include household and transportation behaviors in calculations of physical activity (19). However, estimates of physical inactivity from our study are lower than estimates derived from historical surveillance systems (6,9) that exclude "lifestyle" physical activities in calculations of physical activity. Within racial/ethnic groups in our study, respondents of lower social class tended to be more physically inactive during leisure time.
Education
In this sample, racial/ethnic groups within the same strata of education generally had similar levels of inactivity. This finding supports the hypothesis that education moderates the relationship between race/ethnicity and leisure-time inactivity. Educational experience has been hypothesized to be the single most important social influence on health (20), largely because it is "upstream" from other social status factors and consistently predicts more of the variance in health and mortality than occupation and income do (29). Also, education level is an especially good measure of social class in studies of health behavior because it is more stable than occupation and income (17).
Education seemed to moderate relationships between race/ethnicity and leisure-time inactivity, particularly among women. One exception was that white women with some college education had a lower prevalence of leisure-time inactivity compared with non-Hispanic black women with the same education status. In general, these results differ from those of previous studies (9,23). Those studies reported that differences in leisure-time inactivity among racial/ethnic groups of women remained after adjusting for educational attainment.
Our data suggest that education may play an important role in reducing some health disparities among minority women. It is unclear why Hispanic men with less than a high school education had a higher prevalence of inactivity than whites with the same level of education. This difference was not evident for Hispanic men compared with non-Hispanic blacks. Non-Hispanic black men who were college graduates had the lowest prevalence of leisure-time inactivity compared with that of Hispanics and whites of the same status.
The relationship between educational experience and health is complex, but a possible mechanism involves education as a builder of "human capital"-the productive capacity to use resources and skills to solve problems (20). Formal education is thought to instill the knowledge and values that are important for seeking, understanding, evaluating, and acting on health information (29). Because education likely improves the ability to gather and interpret information about health, it may also increase a sense of personal control over health (20), which is an important predictor of change in exercise behavior (2). Educational attainment may increase cognitive awareness of the relationship between behavior and health, and it is also related to economic and social advancement because people with more education are likely to get better jobs and earn more money. This finding may explain why people with more education report fewer barriers to participation in physical activity (3,32).
Occupation
In our sample, the age-adjusted prevalence of leisure-time inactivity was highest among those who were not in the workforce. These trends were consistent for men and women and for all racial/ethnic groups. The highest prevalence of inactivity (52.2%) was among Hispanic men not in the workforce. However, this finding was based on a small sample size (N = 66) and should be interpreted with caution. However, in general, our findings suggest that being employed may offer some protective effect against inactivity during leisure time.
In addition to employment status (employed, not employed), the intensity of physical activity while a person is at work is also important (6). One hypothesis is that because of a compensation mechanism, persons who are particularly active during working hours are more inclined to be sedentary during leisure time, because of fatigue, lack of motivation, or other factors. Conversely, persons who are particularly sedentary at work may be more inclined to be physically active during leisure time to compensate for a sedentary occupation. There is little evidence to support these hypotheses, but descriptive data from nationally representative samples (3,32) consistently show that lack of energy and lack of motivation are reported as major barriers to leisure-time physical activity.
Our data suggest that physical exertion at work is independent of physical inactivity during leisure time, a conclusion that seems to be consistent across racial/ethnic groups. This finding argues against a compensation mechanism and is consistent with 2001 BRFSS data (18). Those data indicate that the prevalence of leisure-time inactivity was similar for persons who were employed and engaged in mostly heavy labor at work (12.9%; 95% CI 10.3, 15.5%); engaged in mostly walking at work (15.7%; 95% CI 13.6, 17.8%); or engaged in mostly sitting or standing at work (12.8%; 95% CI 12.0, 13.6%). Moreover, our data are particularly informative because we made adjustments for age, sex, and other social class variables-factors known to be related to occupational physical activity. For some comparisons, the findings changed after adjustment for these factors. For example, among Hispanics, the most occupationally active were 2.2 times more likely to be inactive during leisure time than the least occupationally active. However, after accounting for age, sex, and other social class indicators, the odds of being inactive during leisure time were only 1.3 times those of the least occupationally active, and the 95% CI included 1.0. Thus, we recommend that future studies of occupational physical activity routinely measure and adjust for age, sex, and other social class factors.
Income
Physical activity may provide one mechanism to explain why economic indicators of social status are consistently related to health outcomes (11). In our sample, respondents with lower levels of annual household income tended to have higher levels of age-adjusted leisure-time inactivity. Although some cell sizes were relatively small, differences in the prevalence of leisure-time inactivity between racial/ethnic groups were less evident within income strata. These findings suggest that household income may moderate the relationship between race/ethnicity and leisure-time inactivity. However, Hispanic men with less than $25,000 annual income were more inactive than white men of the same status. This finding is important because it suggests that household income has less influence on leisure-time inactivity between racial/ethnic groups who earn the least. Because the largest differences in leisure-time inactivity were evident at lower income levels, possible relationships are likely to be complex. Income can have a direct impact on physical activity because some activities (e.g., cycling) require the purchase of equipment or access to fee-based social amenities (e.g., recreational facilities and sports clubs). However, relationships between household income and physical inactivity may also be mediated by neighborhood-level environmental factors that are known to correlate with physical activity behavior. For example, people with more money are more likely to live in better neighborhoods that are characterized by factors positively associated with physical activity, such as neighborhood attractiveness, safety, and access to recreation facilities (5,28).
Marital Status
In our sample, the prevalence of leisure-time inactivity did not seem to differ across marital status groups for men or women. However, Hispanic men and women who were married or partnered were more likely to be inactive than white men or women of the same marital status. Overall, our findings suggest that marital status does not moderate the relationship between race/ethnicity and physical inactivity.
A limitation of our data is that measured variables were based on self-reports that may not accurately reflect physical activity level or social class. Also, sample sizes for some racial/ethnic groups within social class strata were small; 15 of 78 subgroups had sample sizes less than 100. Consequently, the 95% CI for some point estimates of prevalence were inordinately large. In addition, because nonoverlapping confidence intervals were used as evidence of nonequivalent groups, some statistically significant differences between groups may have been missed. Future research should attempt to oversample racial/ethnic groups within specific social class strata to enable more precise estimates of prevalence.
Although social class was measured by multiple indicators, other variables that are correlated with both social class and physical activity were not examined. For example, perceptions of neighborhood safety and access to facilities seem to be important correlates of physical activity, and these perceptions are known to differ in communities of different socioeconomic status (28). The extent to which these factors influenced the prevalence of physical inactivity in our sample is not clear. Racial/ethnic differences in leisure-time physical inactivity may also be confounded by the level of acculturation to living in the United States (10). Acculturation may influence both actual behavior and the reporting of behavior because less acculturated racial/ethnic groups and those who have recently arrived in the United States are more likely to face social and economic hardship (8) and may have difficulty responding to questions that lack sensitivity to culture and language. Interviews with respondents in our sample were conducted in both English and Spanish, but the level of acculturation remains unknown.
In conclusion, our data show that the age-adjusted prevalence of leisure-time inactivity is lower in white men and women than in non-Hispanic black and Hispanic men and women. Among men, differences in leisure-time inactivity between racial/ethnic groups were less evident within the same strata of education, household income, employment status, and marital status. Among women, few differences in the prevalence of inactivity in racial/ethnic groups were evident within social class strata. This finding suggests that social class moderates the relationship between race/ethnicity and leisure-time inactivity, particularly among women. The odds of being inactive during leisure time did not differ significantly across quartiles of occupational physical activity. This finding was consistent for each racial/ethnic group and after adjusting for age, sex, and social class. Overall these data suggest that social class, but not occupational physical activity, moderates the relationship between race/ethnicity and leisure-time inactivity.
Additional data are needed to explore mechanisms that may contribute to moderation of leisure-time inactivity by social class factors known to differ by race/ethnicity. Improving understanding of how race/ethnicity and leisure-time physical inactivity are related can lead to better targeting and tailoring of interventions to focus on modifiable determinants of inactivity. The primary goal is to reduce health disparities in minority populations.
This work was supported by a SIP 20-01 Cooperative Agreement (U48/CCU409664) between the Centers for Disease Control and Prevention and the University of South Carolina Prevention Research Center.
REFERENCES
1. Ainsworth, B. E., W. L. Haskell, A. S. Leon, et al. Compendium of physical activities: classification of energy costs of human physical activities.
Med. Sci. Sports Exerc. 25:71-80, 1993.
2. Biddle, S. J. H., and C. Nigg. Theories of exercise behavior.
Int. J. Sports Psych. 31:290-304, 2000.
3. Canada Fitness & Lifestyle Research Institute. Barriers to physical activity.
Prog. Preven. Bulletin 4:1-6, 1996.
4. Centers for Disease Control and Prevention. Behavior Risk Factor Surveillance System. Available at:
http://www.cdc.gov/brfss. Accessed November 17, 2004.
5. Centers for Disease Control and Prevention. Neighborhood safety and the prevalence of physical inactivity-selected states, 1996.
Morb. Mort. Wkly. Rep. 48:143-146, 1999.
6. Centers for Disease Control and Prevention. Prevalence of leisure-time and occupational physical activity among employed adults in the United States, 1990.
Morb. Mort. Wkly. Rep. 49:420-424, 2000.
7. Council of American Survey Research Organizations.
On the Definition of Response Rates: A Special Report of the CASRO Task Force on Completion Rates. Port Jefferson, NY: CASRO, p.11, 1982.
8. Council on Economic Advisors.
Changing America: Indicators of Social and Economic Well-being by Race and Hispanic Origin. Washington, DC: Council on Economic Advisors. Publication Document PR: 42.8:C 36 0851-J, 1998.
9. Crespo, C., E. Smit, R. Andersen, O. Carter-Pokras, and B. E. Ainsworth. Race/ethnicity, social class and their relation to physical inactivity during leisure time: results from the third National Health and Nutrition Examination Survey, 1988-1994.
Am. J. Prev. Med. 18:46-53, 2000.
10. Crespo, C., E. Smit, O. Carter-Pokras, and R. Andersen. Acculturation and leisure-time physical inactivity in Mexican-American adults: results from NHANES III, 1988-1994.
Am. J. Public Health 91:1254-1257, 2001.
11. Daly, M. C., G. J. Duncan, G. A. Kaplan, and J. W. Lynch. Macro-to-micro links in the relation between income inequality and mortality.
Milbank Q. 76:315-339, 1998.
12. Farmer, M. M., and K. F. Ferraro. Are racial disparities in health conditional on socioeconmic status?
Soc. Sci. Med. 60:191-204, 2005.
13. Frank, J. W., R. Cohen, I. Yen, J. Balfour, and M. Smith. Socioeconomic gradients in health status over 29 years of follow-up after midlife: the Alameda county study.
Soc. Sci. Med. 57:2305-2323, 2003.
14. Ham, S. A., M. M. Yore, J. E. Fulton, and H. W. Kohl. Trends in physical inactivity during leisure time, 35 states and District of Columbia, United States-1988-2002.
Morb. Mort. Wkly. Rep. 53:76-81, 2004.
15. Issacs, S. L., and S. A. Schroeder. Class-the ignored determinant of the nation's health.
N. Engl. J. Med. 351:1137-1142, 2004.
16. Kington, R. S., and J. P. Smith. Socioeconomic status and racial and ethnic differences in functional status associated with chronic diseases.
Am. J. Public Health 87:805-810, 1997.
17. Liberatos, P., B. G. Link, and J. L. Kelsey. The measurement of social class in epidemiology.
Epidemiol. Rev. 10:87-121, 1988.
18. Macera, C. A., S. A. Ham, M. M. Yore, et al. Prevalence of physical activity in the United States: behavioral risk factor surveillance system, 2001.
Preventing Chron. Dis. 2:A17, 2005.
19. Macera, C. A., D. A. Jones, M. M. Yore, et al. Prevalence of physical activity, including lifestyle activities among adults-United States, 2000-2001.
Morb. Mort. Wkly. Rep. 52:764-769, 2003.
20. Mirowsky, J., and C. E. Ross.
Education, Social Status and Health. New York, NY: Aldine de Gruyter, p. 94, 2003.
21. National Heart Lung and Blood Institute (NHLBI). Socioeconomic factors: working group summary. In:
Measuring Psychosocial Variables in Epidemiologic Studies of Cardiovascular Disease: Proceedings of Workshop., A. M. Ostfeld and E. D. Eaker (Eds.). Bethesda, MD: National Institutes of Health (NIH Publication No. 85-2270), pp. 41-52, 1985.
22. Rothman, K. J., and S. Greenland. Approaches to statistical analysis. In:
Modern Epidemiology. 2nd ed., K. J. Rothman and S. Greenland (Eds.). Philadelphia: Lippincott-Raven, pp. 183-199, 1998.
23. Shea, S., A. D. Stein, C. E. Basch, et al. Independent associations of educational attainment and ethnicity with behavioral risk factors for cardiovascular disease.
Am. J. Epidemiol. 134:567-582, 1991.
24. U.S. Census Bureau. U.S. Summary: 2000. Census 2000 Profile. Washington, DC: U.S. Department of Commerce Economics and Statistics Administration. U.S. Census Bureau, Government Printing Office, 2002.
25. U.S. Department of Health and Human Services.
Healthy People 2010: Understanding and Improving Health, Washington, DC: Government Printing Office, 2000.
26. U.S. Department of Health and Human Services. Physical activity and health: a report of the Surgeon General. Atlanta, GA: U.S. Department of Health and Human Services, Centers for Disease Control and Prevention, National Center for Chronic Disease Prevention and Health Promotion, p. 177, 1996.
27. U.S. Department of Health and Human Services.
Report of the Conference on Socioeconomic Status and Cardiovascular Health and Disease. Bethesda, MD: National Institutes of Health, pp. 22-30, 1995.
28. Wilson, D. K., K. A. Kirtland, B. E. Ainsworth, and C. L. Addy. Socioeconomic status and perceptions of access and safety for physical activity.
Ann. Behav. Med. 28:20-28, 2004.
29. Winkleby, M. A., D. E. Jatulis, E. Frank, and S. P. Fortmann. Socioeconomic status and health: how education, income, and occupation contribute to risk factors for cardiovascular disease.
Am. J. Public Health 82:816-820, 1992.
30. Winkleby, M. A., H. C. Kraemer, D. K. Ahn, and A. N. Varady. Ethnic and socioeconomic differences in cardiovascular disease risk factors.
JAMA. 280:356-362, 1998.
31. Wong, M. D., M. F. Shapiro, W. J. Boscardin, and S. L. Etner. Contribution of major diseases to disparities in mortality.
N. Engl. J. Med. 347:1585-1592, 2002.
32. Zunft, H. F., D. Friebe, B. Seppelt, et al. Perceived benefits and barriers to physical activity in a nationally representative sample in the European Union.
Pub. Health Nutr. 2:153-160, 1999.
