SEARCH

SEARCH BY CITATION

Keywords:

  • race;
  • inequality;
  • health;
  • nutrition;
  • disease;
  • morbidity;
  • mortality

Abstract

  1. Top of page
  2. Abstract
  3. Introduction
  4. Background
  5. Hypothesis
  6. The study Data and measures
  7. Methods
  8. Analysis
  9. Discussion
  10. Conclusion
  11. Acknowledgements
  12. References

Black-White disparities in the incidence and prevalence of chronic disease and premature morbidity are persistent and well documented in the United States. Prevailing explanations for these disparities have focused upon socioeconomic inequality and related mechanisms as the causal factors. Yet, despite the explanatory power of socioeconomic status in models of health outcomes, an unexplained racial gap in health persists. This research contributes to the study of the Black-White health divergence by exploring a mechanism with the prospect of explaining a portion of the racial gap in health remaining after adjustment for socioeconomic status. Specifically, using random coefficient regression to analyse pooled data from the 1993–1999 California Dietary Practices Survey, I identify significant differences between Blacks and Whites, after adjustment for socioeconomic status and other controls, both in global nutritional healthfulness and across a range of nutritional behaviours with established links to the development of chronic disease. Given the compelling body of literature linking nutritional behaviour to health outcomes, these differences between Blacks and Whites constitute evidence for the potential explanatory value of nutrition in future studies seeking to explain the residual racial gap in health remaining after adjustment for socioeconomic status and correlates of socioeconomic status.


Introduction

  1. Top of page
  2. Abstract
  3. Introduction
  4. Background
  5. Hypothesis
  6. The study Data and measures
  7. Methods
  8. Analysis
  9. Discussion
  10. Conclusion
  11. Acknowledgements
  12. References

Racial differences in the incidence and prevalence of chronic disease and premature morbidity are subjects of long-standing interest in the social sciences, and the differences between Blacks and Whites in this regard are striking, disturbing, and, by some accounts, widening (Byrd and Clayton 2002). For example, the life expectancy gap between White and Black males in the US increased in favour of Whites from 6.4 years to 6.9 years between 1983 and 1998, such that White males now have expected life spans 10.2 per cent greater than Blacks (Murphy 2000). Similarly, the life expectancy gap between White and Black females, while relatively stable over the 15-year period, continues to favour Whites by 5.2 years (7.0%).

To characterise this underlying mortality gap from a different perspective, Blacks and Whites differ substantially in estimates of survival rates to middle age. Only 73.3 per cent of Black males born in 1970 in the United States are expected to survive to age 50 compared with 87.6 per cent of White males in that birth cohort (Hayward et al. 2000). The gap between Blacks and Whites widens with age, as the expected 60-year survival rates drop to 57.5 per cent and 76.0 per cent, respectively (Hayward et al. 2000). Consistent with these statistics, the age-adjusted mortality rate for Blacks is 153 per cent of that of Whites (Murphy 2000).

Given the racial differences in life expectancy, survival rates, and overall mortality, it is not surprising to find that Blacks are disadvantaged relative to Whites, after age adjustment, on 12 of the 15 major causes of death in the US (Murphy 2000).1 Among the top seven causes of death in 1998 (76.7% of all mortality) were five chronic diseases: heart disease, cancer, cerebrovascular disease (e.g. stroke), chronic obstructive pulmonary disease, and diabetes (the first, second, third, fourth, and seventh leading causes of death, respectively). Blacks were disadvantaged on all of these except chronic obstructive pulmonary disease. Blacks experienced a rate of mortality due to diabetes that was 240 per cent of the rate of Whites. Likewise, rates of mortality due to cerebrovascular disease, heart disease, and cancer were 178 per cent, 150 per cent, and 133 per cent of the rates of Whites. Taken together, these four chronic diseases constituted 61.2 per cent of all mortality in the US in 1998, and 56.8 per cent of excess deaths among Blacks.

In this paper, I test for differences between Blacks and Whites in nutritional behaviours with established associations to the development of heart disease, cancer, stroke, and diabetes. In addition to other variables, I control for socioeconomic status (SES) − the historically favoured explanation for racial disparities in health outcomes − in order to differentiate racial differences in nutritional behaviour from variation in nutritional behaviour associated with SES.2 The purpose of these tests is to investigate whether empirical evidence supports the finding of Black-White disparities in health-related nutritional quality, independent of SES. Any such identified disparities would provide a rationale for the inclusion of nutritional quality in future data collection efforts and analyses addressing race and health. However, the data used here do not include measures of health, so the mediating role of nutrition in the race-health relationship cannot be ascertained, and this study must rely upon the existing body of literature supporting inverse associations between nutritional healthfulness and the development of chronic disease.

Background

  1. Top of page
  2. Abstract
  3. Introduction
  4. Background
  5. Hypothesis
  6. The study Data and measures
  7. Methods
  8. Analysis
  9. Discussion
  10. Conclusion
  11. Acknowledgements
  12. References

Race and health

There is a growing debate in the literature concerning whether race and the emphasis on different racial ethnic groups should continue to play a prominent role in health research. The argument against it is premised upon the facts that race is essentially an ideological construction, that race is not the basis upon which differential health outcomes can be explained, and that racial identification is not necessarily stable over time (Nazroo 1998, Smaje 1995). However, from an epidemiological perspective, race remains a significant predictor of health outcomes and, therefore, is arguably a matter worthy of continued empirical attention. This perspective is consistent with the historical inclination of sociological research concerning matters of race on a wide variety of substantive topics (e.g. Emerson et al. 2001, Hughes and Thomas 1998, Hunt et al. 2000). Moreover, Byrd and Clayton (2002: 564) argue against the reinforcement of ‘the general perception in the scientific community that race and ethnicity research lacks rigor in conceptualization, terminology, and analysis’, as it obscures the need to study, understand, and rectify continuing disparities in health outcomes among a health underclass defined, in part, along racial lines. Thus, it would be reasonable to conclude that, so long as multiple social inequalities intersect in categories of a variable labelled ‘race’, sociological inquiry into matters of race will remain important.

Nevertheless, race is a social construction, not a biological fact (Smaje 2000), and explanations for racial differences in health premised upon genetic dissimilarity for the most part have been dismissed (Hayward et al. 2000, Kong et al. 1994). In fact, fewer than one per cent of excess deaths among Blacks can be attributed to differential propensity for hereditary conditions (Leigh 1995). Thus, observed racial patterns of health outcomes are primarily consequences of social forces, rather than genetic history. As Smaje (2000: 114) explains, ‘race in itself can never be invoked as a self-adequate explanation for patterns of health experience’. Nevertheless, as Schulz et al. (2000: 315) argue, ‘[r]ace, as a social construct, remains a powerful organizing feature of American social life, and racial categories both reflect and reinforce group differentials in power and access to social resources’. It follows that race is still a useful organising tool in efforts to understand and rectify inequity on many fronts, not the least of which is health (Gillum 2004).

Socioeconomic status and the racial gap in health

Historically, efforts to explain racial disparities in health in Western societies have focused on socioeconomic inequality and related correlates. In research dating back seventy years, Tibbits (1937) argued:

[i]t is well known that the Negro population is less fortunately situated than the white in terms of income, education, opportunities for obtaining medical care, etc. Hence it seems relatively safe to assume that where there is variation in the degree of health among different economic and social groups of the white population and where Negroes show a higher rate of ill health than the whites, the explanation lies partly, at least, in the low-income status of the Negro (1937: 417–18).

Much of the subsequent research on the topic has adhered to this theory, attempting to explain observed racial differences in chronic disease and premature morbidity through mechanisms rooted in socioeconomic inequality. This theory invariably leads to the hypothesis that, once inequality and related variables are controlled, racial differences in the incidence and prevalence of chronic disease and premature morbidity will disappear. A number of recent studies have continued the longstanding investigation of this hypothesis (e.g. Lantz et al. 1998, Mutchler and Burr 1991, Potter 1991, Rogers 1992), and the findings of these studies support the argument that SES is an important explanatory variable in Black-White differences in a variety of health outcomes.

Much attention also has been given to the proximal causes of this health differential as distinct from the distal cause (i.e. socioeconomic inequality). The objective underlying this class of papers is the enumeration of the mechanisms connecting SES to health. Generally speaking, these proximal causes can be divided into two broad categories: (1) health risk behaviours that are correlated with SES, such as smoking, excessive alcohol consumption, and physical inactivity (e.g. Wickrama et al. 1999, Williams and Collins 1995), and (2) mechanisms arising directly from unequal distribution and access to resources, such as inadequate medical care, dangerous working conditions, and exposure to environmental toxins (e.g. Smith and Kington 1997a, Williams and Collins 1995). While this distinction is important, an extensive review of this topic is outside the scope of this paper.3

The residual racial gap in health

While research on the relationship between socioeconomic inequality and health has explained a substantial segment of the racial disparities in health outcomes, most research still documents an unexplained racial gap in health after adjustment for socioeconomic differences. For example, Rogers (1992) found that adjusting for income and a set of demographic variables narrowed the health gap between Blacks and Whites, but did not fully eliminate it. Christenson and Johnson (1995) found that the protective effect of education on mortality benefited Blacks less than Whites. Huie et al. (2003) found a residual Black disadvantage in mortality net of racial differences in educational attainment, income, net worth, and demographic variables. Mutchler and Burr (1991), adjusting for four indicators of SES as well as various demographic measures, found significant differences between older Blacks and Whites remaining on one of six self-reported health measures and an increase in the Black-White differential on another of the six measures. Ferraro and Farmer (1996) found significant disadvantages for Blacks relative to Whites in the incidence of serious illness, in the declination of self-assessed health over time, and in survival rates, despite numerous controls, including education, income, access to medical care, and prominent health risk behaviours. Ferraro et al. (1997) found greater erosion of self-assessed health and greater increases in the incidence of chronic illness and disability among Blacks compared with Whites, net of prior health conditions, SES, and other relevant controls. Schnittker (2004) found poorer self-assessed health and lower scores on an index of physical health among Blacks relative to Whites after controlling for education, income, and the income-health gradient as it varied by level of educational attainment. In a particularly comprehensive study, Hayward et al. (2000) found a significant Black-White health gap remaining on seven health outcomes after controlling for educational attainment, household income, wealth, availability of health insurance, numerous health risk behaviours, a number of psychosocial characteristics, and other variables.

While just a few of the recent analyses of racial differences in health are documented here, the findings of numerous similar studies agree that SES and its correlates can explain much, but not all, of the differences in health outcomes between Blacks and Whites (Crimmins et al. 2004, Livingston et al. 2004, Smith and Kington 1997a). Authors’ conclusions often agree with the explanation offered by Williams and Collins (1995) in their review of the research addressing the White/Black health differential: ‘within each level of SES, blacks generally have worse health than whites’ (1995: 364). For example, Mutchler and Burr (1991) note that, ‘in terms of self-rated health and some associated health-related behaviours, older Blacks appear to have poorer health than Whites regardless of socioeconomic status’ (1991: 353). Likewise, Ferraro and Farmer (1996) observe that, even after accounting for the SES differential between Blacks and Whites, ‘there are important differences between the health status and health assessments of Black and White Americans’ (1996: 37). Thus, prior research repeatedly demonstrates a stubborn racial gap in health outcomes that persists despite adjustment for socioeconomic inequality and related mechanisms.

Consequently, we must ask, what are the missing factors that will explain the residual racial differences in health outcomes? Socioeconomic status and related mechanisms have been controlled statistically, yet health differences between Whites and Blacks remain. Why?

Explaining the unexplained racial gap in health

One possible explanation for the residual racial gap in health lies in the nature of socioeconomic controls. Measures of SES can only approximate the complex and multifaceted construct, leaving open the possibility of unmeasured covariation between SES and health outcomes (Krieger 2000, Williams and Collins 1995). In other words, the inability of SES to explain fully the race-health relationship may be a methodological artifact (Kaufman et al. 1997). However, the persistence of the residual race-health relationship in numerous studies, using a variety of data, and despite increasingly comprehensive measures of SES, suggests at least the possibility of an excluded explanatory variable operating independently of SES.

Among the potential excluded explanatory variables, chronic stress caused by experiences of racial discrimination is receiving increased attention in the literature (e.g. Brondolo et al. 2003, Clark 2003, Clark and Adams 2004, Karlsen and Nazroo 2002, Krieger 2000, Krieger et al. 1993, Leigh 1995, Livingston 1994, Myers et al. 2004). This line of research argues that the excess of stressors experienced by Blacks as a result of both intergroup and intragroup racism contribute to a chronic elevated physiological stress response (Clark et al. 1999). This elevated stress response is associated with a range of health problems (Berkman and Kawachi 2000, Fremont and Bird 2000, Krieger 2000), including at least two chronic diseases for which Blacks are severely disadvantaged: coronary heart disease and diabetes (Livingston and Carter 2004).

Another potential explanatory variable that is sometimes mentioned, but which has received comparatively little empirical attention, is nutritional healthfulness. Interestingly, the health-related consequences of stress are intertwined closely with nutritional deficiencies. Prolonged stress tends to deplete physiological stores of, and increase physiological demand for, essential nutrients (Semmes 1996). In addition, healthy nutritional behaviours serve as a barrier against the damaging effects of prolonged stress, and poor nutrition itself can induce a physiological stress response (Semmes 1996). Lastly, chronic stress contributes to feelings of helplessness, hopelessness, and loss of control, which, in turn, are associated with a greater propensity for unhealthy lifestyle behaviours, such as poor nutritional choices (Kristenson et al. 2004). Thus, as they pertain to health outcomes, nutritional quality and stress are interrelated.

In addition to the interrelationship between nutrition and stress, nutritional quality has a well-established, strong, and direct relationship to health outcomes.4 As Blocker (1994) observes, ‘[g]ood nutrition is crucial to the maintenance of health, and dietary factors contribute substantially to preventable chronic illness and premature death’ (2004: 267). The relationships between nutrition and cardiovascular disease, cancer, stroke, and diabetes − the first, second, third, and seventh leading causes of death in the United States − are particularly strong (Blocker 1994). An estimated 30 per cent of all deaths due to cancer in the United States can be attributed to diet (Harvard School of Public Health 1996), and dietary and physical inactivity patterns collectively cause 14 per cent of all deaths in the United States, second only to tobacco as the leading cause of death (McGinnis and Foege 1993). In addition to cancer prevention, it is estimated that dietary changes could reduce the risk of heart attack and stroke by 20 per cent to 30 per cent and the risk of preventable diabetes by 50 per cent to 75 per cent (Foerster et al. 1999).

The specific dietary practices associated with reduced risk of chronic diseases have been demonstrated repeatedly and, due in part to expanding public health campaigns, are rapidly becoming topics of common knowledge (although not necessarily common practice). Weisburger (2000) details a number of nutritional factors associated with lowered risk of chronic disease, including: reduced consumption of dietary fat, increased consumption of cereal bran fibre (e.g. wholegrains), increased consumption of fruits and vegetables, reduced consumption of fried and broiled foods, and increased consumption of dairy products. The consumption of saturated fats is correlated positively with the incidence of cancer, while the consumption of fruits, vegetables, and fibre is associated negatively with the incidence of cancer (Bal et al. 2001, Colditz et al. 2000, Willett 1994). The incidence of diabetes has been connected to dietary fat consumption (Blair et al. 1996). Consumption of fruits, vegetables, and whole grains is associated with reduced risk of stroke (Joshipura et al. 1999, Liu et al. 2000a). The incidence of cardiovascular disease (the broader category of chronic disease within which belong strokes) varies negatively with fruit, vegetable, and dietary fibre consumption, and varies positively with fat consumption and dietary cholesterol (Blair et al. 1996, Foerster et al. 1999, Liu et al. 2000b, Stampfer et al. 2000, Willett 1994). Moreover, these findings of specific associations between particular nutritional behaviours and the incidence of chronic disease are buttressed by research linking reduced risk of several major chronic diseases to comprehensive patterns of healthful eating (Hu et al. 2000, McCullough et al. 2000a, 2000b, Willett 1994). Also of note, foods that are high in fat have an interesting dual role in the development of chronic disease in that, in addition to being associated with increased incidence of chronic disease, they also tend to supplant healthier nutritional choices that have protective effects (Kant 2000). In sum, there is general agreement between medical and health researchers that increased consumption of fruits, vegetables, grains, fibre-rich foods, beans, and dairy products, and decreased consumption of dietary fat, substantially reduce the risk of many of the most common chronic diseases, including cardiovascular disease, cancer, stroke, and diabetes (Eyre et al. 2004, Foerster et al. 1999, Willett 1994).

Racial differences in nutritional behaviour

The omission of nutritional behaviour from prior analyses of the race-health relationship introduces one of two assumptions. The analyses must assume either that nutritional healthfulness does not differ meaningfully between racial groups or, alternatively, that nutritional healthfulness, as one expression of the general class of personal health behaviours, is perfectly correlated with (accounted for by) measures of SES. In point of fact, when variation in nutritional quality is mentioned as a potential mediating variable underlying Black-White differences in health outcomes, it is most often attributed to disparities in SES (e.g. Siewe 1999, Wickrama et al. 1999) or, more specifically, to the high rate of poverty among Blacks (e.g. Blocker 1994, Kittler and Sucher 1989, Leigh 1995). In other words, nutritional quality is theorised to mediate the SES-health relationship, to the exclusion of any mediating role it may play in the residual race-health relationship existing independently of the SES-health relationship.

Given, however, that controlling for SES has not eliminated the Black-White health disparities, and given the well-established relationship between nutrition and health, it is reasonable to conclude that other causal channels linking race to nutritional quality are worthy of exploration. In support of this argument, there is limited evidence that suggests that health-related nutritional behaviours do vary between Blacks and Whites independent of SES (Airhihenbuwa et al. 1996, Lindquist et al. 2000, Popkin et al. 1996). But each of the several previous studies that address the race-nutrition relationship, independent of SES, evidence significant weaknesses. For example, the Popkin et al. (1996) study was weakened by an oversimplified operationalisation of SES as a trichotomous indicator and the absence of controls for important demographic variables. The Lindquist et al. (2000) study was limited to a small convenience sample of young children. The Airhihenbuwa et al. (1996) qualitative study relied upon respondents’ own assessment of the effect of income on nutritional behaviour.

In addition to the limited empirical evidence, racial differences in health-related nutritional behaviours, independent of SES, would be expected for a number of reasons. Among these, Semmes (1996: 128) argues that nutritional deficiencies among Blacks are, in part, a result of ‘maladaptive dietary practices rooted in slave culture’, which is a perspective echoed in the interviews of Blacks conducted by Airhihenbuwa et al. (1996). Semmes (1996) argues that many present nutritional practices among Blacks are derived from ‘slave culture’, leading to an over-reliance on processed sugars and fatty meats and an underutilisation of fruits and vegetables. Kiple and King (1981) offer a contrasting, and much more detailed, assessment of Black nutrition under slavery, but the conclusion of significant nutritional deficiencies is the same. Other researchers have documented support for the influence of racial subcultures on dietary practices (e.g. Witt 1999) and, following from this, differences in nutritional quality (Kittler and Sucher 1989). In other words, cultural differences in preferred foods and cooking methods may contribute to differences in nutritional quality between Blacks and Whites, independent of SES. This causal explanation points to the importance of dietary patterns as components of social history and cultural memory, whereby knowledge and practice of nutritional behaviours are passed from generation to generation within families (Airhihenbuwa et al. 1996, Birch 1999).

One should note that, although this argument may be interpreted as ‘blaming the victim’, such an interpretation would be inaccurate. In particular, Semmes’ (1996) use of the word ‘maladaptive’ appears, on the surface, to imply that Black slaves had some degree of choice in the nutritional content of their diets. Of course, this was not the case. Instead, this argument points to the opposite situation: a people group who, by enslavement, were deprived of their culture, including their food culture, and forced to adapt to a severe state of scarcity, dislocation, and dispossession.

Racial differences in nutritional quality also may be rooted in residential segregation, which is, itself, another expression of racial discrimination (Krieger 2000). Blacks are subjected to the highest rate of residential segregation of any minority group in the US (Steinmetz and Iceland 2003), and Blacks in the US are 144 per cent more likely to live in ‘urban cores’ than are Whites (McKinnon 2003). The overrepresentation of Blacks in urban cores holds even for affluent Blacks (Bullard 1994), indicating that poverty is not the only contributing factor to racial residential segregation (Emerson et al. 2001, Erbe 1975, Farley 1977, Farley and Frey 1994, Iceland and Wilkes 2006, Massey et al. 1987). In fact, the evidence suggests that improvements in SES among Blacks generally contribute little to rectifying rates of residential segregation (Hwang et al. 1985, Villemez 1980).

Urban neighbourhoods, poor neighbourhoods, and neighbourhoods in which the residents are predominantly Black are characterised disproportionately by the notable absence of major supermarkets and specialty food stores, requiring residents to rely on the similarly notable excess of small convenience stores, which generally have higher prices and limited nutritional options (Cheadle et al. 1991, Emmons 2000, Krebs-Smith and Kantor 2001, Leigh 1995, Macintyre and Ellaway 2000, Morland et al. 2002, Williams and Collins 2001). A similar problem, termed ‘food deserts’ to refer to urban environments that lack adequate nutritional infrastructure, has been identified in the UK (Wrigley 2002), and improvements in nutritional infrastructure have been found to be associated with improvements in dietary quality (Wrigley et al. 2002). Furthermore, the problem of limited nutritional infrastructure is exacerbated by the high rate of poverty among Blacks, which tends to limit transportation alternatives, placing access to supermarkets in suburban and predominantly White neighbourhoods out of reach (Krebs-Smith and Kantor 2001, Leigh 1995, Morland et al. 2002).

Interestingly, several recent studies have found that Black-White differences on certain health outcomes shrink substantially, or even to statistical insignificance, once variables that address overarching aspects of community context (e.g. neighbourhood affluence or poverty, neighbourhood population, residential stability, racial segregation) are controlled (e.g. Browning and Cagney 2003, Huie et al. 2002, Robert 1998, Robert and Lee 2002, Subramanian et al. 2005). While the mechanisms that mediate the relationship between community context and health outcomes are still unclear, and while these prior studies do not address nutritional infrastructure specifically, these findings represent an intriguing parallel to the literature that addresses racial segregation, nutritional infrastructure, and nutritional quality. In particular, the contextual variables found to be significant in prior studies would be expected (either intuitively or based on prior research) to be correlated with nutritional infrastructure. Thus, nutritional infrastructure may constitute a key mediating variable in the relationship between community context and racial health disparities.

Another potential explanation for racial differences in nutritional quality is disparities in the nutritional information provided to Blacks via race-targeted product advertising. Semmes (1996) notes that advertisers of alcohol and tobacco have a long history of targeting the Black segment of the market, which, as one would expect, is correlated with residential segregation (Williams and Collins 2001). Research has identified differences in nutritional advertising as well. For example, Pratt and Pratt (1996) found evidence of significant disparities in the healthfulness of nutritional product advertising in popular print media targeted at Blacks versus that targeted at Whites. Specifically, in a longitudinal content analysis, Pratt and Pratt found that one widely circulated magazine targeted at Whites (Ladies’ Home Journal) carried significantly more nutritional advertising in support of dairy products, breads, cereals, vegetables, and fruits, than did two widely circulated magazines targeted at Blacks (Ebony and Essence). It follows that, to the extent that consumers depend upon nutritional advertising for making healthy nutritional choices, the gap in healthful nutritional information provided to Blacks may contribute to racial disparities in nutritional practice. In this regard, there is limited evidence that Blacks depend more heavily on advertising in making decisions about food choices than do Whites (Bock et al. 1998), potentially amplifying the effects of any differences in racially-targeted product advertising.

Finally, differences in the household compositions of Blacks and Whites may contribute to average differences in nutritional quality. Fully 58 per cent of all Black family households with children in the US are single-parent households, compared to 23 per cent of White family households with children (Fields 2004). While the effect of single-parent versus dual-parent living arrangements on adult nutritional quality is unknown, one might reason that single-parent living arrangements allow less time for grocery shopping and meal preparation, on average, than do dual-parent living arrangements. Thus, the excess of single-parent families among Blacks may contribute to a depression in average nutritional quality among Black adults.

Hypothesis

  1. Top of page
  2. Abstract
  3. Introduction
  4. Background
  5. Hypothesis
  6. The study Data and measures
  7. Methods
  8. Analysis
  9. Discussion
  10. Conclusion
  11. Acknowledgements
  12. References

In this analysis, I expect to find relatively fewer healthy nutritional behaviours among Blacks as compared with Whites, after adjustment for SES and other variables. In other words, I hypothesise that a relationship exists between race and nutritional quality that is independent of SES. Given the documented significance of nutritional quality in the development of the chronic diseases for which Blacks are disadvantaged relative to Whites, it follows that identifying significant racial differences in nutritional quality, independent of SES, would provide evidence that favours the potential utility of nutritional behaviour in future studies that seek to explain the residual racial gap in health. Any other finding would suggest that nutritional behaviour lacks explanatory utility in modeling racial disparities in health outcomes.

The study Data and measures

  1. Top of page
  2. Abstract
  3. Introduction
  4. Background
  5. Hypothesis
  6. The study Data and measures
  7. Methods
  8. Analysis
  9. Discussion
  10. Conclusion
  11. Acknowledgements
  12. References

Data

I use data collected by the California Department of Health Services (CDHS) in the California Dietary Practices Survey (CDPS) to test the hypothesis that Blacks differ significantly and negatively from Whites in terms of health-related nutritional behaviour, net of SES and other controls. The CDPS is a computer-assisted telephone interview of randomly selected California residents, employing a prescreened random digit dialing procedure. The sampling procedure draws upon a sampling frame that includes approximately 97 per cent of all households in California. The survey was administered in six biennial waves over 11 years (1989 through 1999) with samples varying in size from 1,000 to 1,703 (Foerster et al. 1999).

The CDPS collects data on a wide variety of health-related practices and beliefs. The centrepiece of the survey is a 24-hour dietary recall (Krebs-Smith and Kantor 2001), consisting of a comprehensive set of questions designed to elicit from the respondent all foods and beverages consumed during the previous 24 hours. The dietary data are then collapsed into a series of measures of key nutritional behaviours. In addition, a set of demographic measures is integrated into the survey.

Dependent variable

I selected the CDHS Healthy Eating Practices Index (HEPI) as the primary outcome measure for exploring racial differences in health-related nutritional behaviour (Foerster et al. 1999). The HEPI offers a global summary of important nutritional health behaviours and is consistent with previous findings of clusters of healthy and unhealthy nutritional behaviours (Hu et al. 1999). The HEPI is constructed by tallying points for key nutritional behaviours. In accordance with the earlier discussion, it addresses five dimensions of nutritional behaviour related to the incidence and prevalence of chronic disease and premature morbidity: fruit and vegetable consumption, consumption of dairy products, consumption of fibre/grains, consumption of beans, and consumption of fat. Specifically, a single point is added to a respondent's index score if he/she consumed at least one serving of fruit and at least one serving of vegetables (including fruit and vegetable juices) during the previous 24 hours. An additional point is added if a respondent's combined total number of servings of fruits, vegetables, and fruit/vegetable juices is at least five. A point is added if a respondent consumed any milk, yogurt, or cheese. An additional point is added for the consumption of any low fat milk, nonfat milk, or yogurt. Finally, a point each is added for the consumption of any wholegrain breads or corn tortillas, the consumption of high fibre cereals, and the consumption of any beans (Foerster et al. 1999).

The HEPI is weakened by three problems. First, the data upon which it is based are dichotomous rather than continuous, reducing actual variation in the measure. In analyses such as this one, however, the dichotomous nature of the variables composing the index tends to attenuate differences between groups, leading to underestimation, rather than overestimation, of coefficients. Likewise, the dichotomous nature of the variables tends to minimise bias associated with reporting error by ignoring small differences in respondent recall of number of servings consumed.

Secondly, the Healthy Eating Practice Index lacks specific focus on substantial sources of dietary fat, which, as noted earlier, play a dual role in the development of chronic disease. While the HEPI includes a single index point for the consumption of low fat and nonfat milk products, it does not address nutritional choices that would increase dramatically a respondent's total dietary fat. To treat this problem, the HEPI is modified here to include an additional point added to a respondent's score if he/she did not consume any deep-fried foods and fried snacks. To distinguish between the HEPI as defined by the California Department of Health Services and the modified HEPI employed here, I refer to the index used in this analysis as the Expanded Healthy Eating Practices Index (EHEPI).

Finally, it should be noted that the HEPI and EHEPI do not attempt to account for culturally specific eating behaviours, which may lead to racial bias in the measurement of nutritional healthfulness. However, empirical work on the topic of racially specific nutritional measurement (e.g. Schlundt et al. 2003) suggests that the HEPI and EHEPI have only two deficiencies in this regard. In particular, they do not account for added fats in cooking and seasoning vegetables, and they do not account for variation in the healthfulness of consumed meats (e.g. fish and poultry versus beef and pork) (Airhihenbuwa et al. 1996, Kumanyika and Adams-Campbell 1991, Schlundt et al. 2003).

Data were not collected for all components of the index in early sample years, so use of the index is confined to sample years 1993 through 1999, which are pooled for this analysis. Scores on the index range from a minimum of zero points, indicating very poor nutritional behaviours, to a maximum of eight points, indicating compliance with recommended healthy-eating practices. The mean EHEPI score in the pooled sample is 3.68, indicating that respondents exhibited less than half of the recommended nutritional behaviours on average. The standard deviation of the index is 1.69, and the reliability is 0.48.5

Independent variables

The primary independent variable of interest in this analysis is respondent's race, derived from a multicategorical measure of race in the CDPS data. For the purposes of this analysis, I consider only the dichotomy of Blacks versus Whites, excluding other racial/ethnic groups from the analytical sample. It should however be noted that the gross Black-White dichotomy employed in this study, and common in many prior studies, ignores ethnic distinctions in the historical origins and cultural contexts of respondents, thereby potentially masking important differences (Livingston and Carter 2004, Livingston et al. 2004, Luke et al. 2001, McBarnette 1996, Nazroo 1998, Smaje 1995). While this is an unfortunate limitation of the data employed in this study and a matter to be considered in future research on the topic, it is consistent with the practice of prior research to make generalisations concerning Black-White differences in nutritional behaviour (Blocker and Forrester-Anderson 2004).

In order to distinguish the relationship between race and nutritional behaviour from the relationship between SES and nutritional behaviour, two measures of SES are included in the analysis: annual household income (measured on an eight-point ordinal scale) and educational attainment (measured on a five-point ordinal scale). In the interest of parsimoniousness, household income has been converted here into a continuous variable by calculating the midpoint of the income range indicated at each value of the ordinal scale. The large number of values of the income variable and the effective linearity of the relationship between average index score and income (as evidenced in Table 1) make this conversion feasible, and the net effect is improved model fit.

Table 1.  Bivariate analysis of the Expanded Healthy Eating Practices Index (n = 3,350)
  Expanded Healthy Eating Practices Index
nmeanstd dev
  1. Notes: *p ≤ 0.05; ***p ≤ 0.001.

RaceWhite2,8593.821.67
Black  4912.841.55
t 12.83*** 
Educationless than H.S. grad.  2193.261.66
H.S. graduate  7523.301.64
some college1,2183.591.67
college graduate  7363.991.66
post graduate  4254.261.68
F 33.55*** 
Household income<$10,000  3303.281.73
$10,000–15,000  2863.381.69
$15,001–20,000  2573.531.79
$20,001–25,000  3753.471.66
$25,001–35,000  3593.701.76
$35,001–50,000  4893.911.70
$50,001–65,000  3273.901.57
>$65,000  6503.851.62
missing income  2773.771.67
F 7.20*** 
Sexmale1,2593.551.68
female2,0913.751.70
t 3.38*** 
Age18–24  2993.321.63
25–34  5843.431.60
35–44  8413.621.67
45–54  5593.641.70
55–65  4073.681.72
65+  6604.161.72
F 16.44*** 
Household size1  9333.731.76
21,0713.751.68
3  5533.611.66
4  4683.621.65
5  2033.361.68
6+  1223.631.61
F 2.30* 
Year of data collection1993  7883.871.72
1995  6883.781.72
19971,0033.511.64
1999  8713.611.69
F 7.90*** 

In addition to race and SES, three control variables are included in the models: sex, age (measured in years), and household size (total number of persons). The latter control is included to account for the divisionary effect of household size on household income (Kawachi 2000), which is important in this study because Black families tend to be larger than White families (Myers et al. 2004).

Measures of race (Black = 1; White = 0) and sex (female = 1; male = 0) are entered in the models as dichotomous variables. Household income, age, and household size are entered as continuous variables. Educational attainment is entered as a set of categorical variables with the comparison category of ‘high school graduation’.

Note that several characteristics of the CDPS make it appropriate for use in identifying racial differences in health-related nutritional behaviour while controlling for SES, including the large sample sizes, probabilistic sampling method, numerous measures of nutritional behaviour, and dual measures of SES. However, the data do not contain measures of health and, therefore, cannot be used to determine if racial differences in nutritional behaviour contribute to racial disparities in health outcomes. Therefore, this analysis relies upon the existing body of literature supporting relationships between nutritional patterns and health outcomes to supply evidence for the importance and relevance of identified dietary differences between Blacks and Whites.

Methods

  1. Top of page
  2. Abstract
  3. Introduction
  4. Background
  5. Hypothesis
  6. The study Data and measures
  7. Methods
  8. Analysis
  9. Discussion
  10. Conclusion
  11. Acknowledgements
  12. References

My analysis involves four phases. First, I examine the bivariate relationships between each of the independent variables and the EHEPI. Secondly, I examine the bivariate relationships between race and each of the nutritional behaviours composing the EHEPI. Thirdly, I use nested random coefficient linear regression (Raudenbush and Bryk 2002) to examine the relationship between race and global nutritional healthfulness, first without, and then with, adjustment for SES.6 Finally, I use random coefficient logistic regression to identify which of the eight health-related nutritional behaviours underlie the racial gap in overall nutritional practice.

Cases with missing data on any of the independent variables other than income (n = 68) or lacking sufficient information to assess the EHEPI (n = 213) are excluded, resulting in the loss of 281 cases (7.7% of the sample).7 Cases for which income is missing (n = 277) had the missing datum replaced with an imputed value based on respondent's race, sex, age, household size, education, and year of data collection. A dummy variable indicating this imputed condition is included in the models. The final analytical sample includes 3,350 respondents, of which 491 (14.7%) are Black.

Analysis

  1. Top of page
  2. Abstract
  3. Introduction
  4. Background
  5. Hypothesis
  6. The study Data and measures
  7. Methods
  8. Analysis
  9. Discussion
  10. Conclusion
  11. Acknowledgements
  12. References

Bivariate analysis

Bivariate analyses of the relationships between the EHEPI and each of the independent variables are presented in Table 1. The relationship between race and the EHEPI is of primary interest, and this relationship is consistent with the hypothesis that Blacks differ significantly and negatively from Whites in terms of nutritional behaviour. While the mean index scores of both Whites and Blacks are less than half the possible range of the index (indicating nutritional practices that differ substantially and negatively from recommended guidelines), the mean EHEPI score for Blacks is 25.7 per cent less than the mean EHEPI score for Whites. Statistically significant variation in EHEPI scores also is observed across categories educational attainment, household income, sex, age, household size, and survey year.

Bivariate analyses of Black-White differences in the unadjusted probability of exhibiting each of the nutritional behaviours composing the EHEPI are presented in Table 2. The results indicate that Whites are more likely, on average, to exhibit six of the eight healthy nutritional behaviours. These include the consumption of one fruit and one vegetable, at least five fruits and/or vegetables, dairy products, high-fibre cereals, and lowfat/nonfat dairy products, and avoiding the consumption of deep-fried foods and snacks. Significant differences are not observed between Blacks and Whites in the probability of consuming wholegrain products nor in the probability of consuming beans.

Table 2. Proportion of Whites versus Blacks exhibiting each of the nutritional behaviours composing the Expanded Healthy Eating Practices Index (nWhites= 2,859; nBlacks= 491)
 One fruit and one vegetableFive fruits and/or vegetablesDairy productsLow fat dairy productsWholegrain products High fibre cerealsBeansDeep-fried foods and snacks
  1. Notes: ***p ≤ 0.001; standard deviations in parentheses.

White0.6440.3540.8040.4020.4890.2110.2380.322
(0.479)(0.478)(0.397)(0.490)(0.500)(0.408)(0.426)(0.467)
Black0.5070.2550.6130.1730.4790.0710.2180.481
(0.500)(0.436)(0.488)(0.379)(0.500)(0.258)(0.413)(0.500)
t5.78***4.30***9.38***9.70***0.447.30***0.966.84***

Regression analysis of the EHEPI

The results of the nested random coefficient linear regressions of the EHEPI on race, SES, sex, age, and household size are presented in Table 3. In Model 1, I regress the EHEPI on race, sex, age, and household size in order to estimate the average nutritional gap between Blacks and Whites, net of controls. In Model 2, I add education and household income to the variables included in Model 1 in order to determine if a residual racial gap in nutritional behaviour persists after the additional adjustment for SES.

Table 3. Estimated coefficients and standard errors for two nested random coefficient linear regressions of the Expanded Healthy Eating Practices Index on selected variables (n = 3,350)
  Expanded Healthy Eating Practices Index
Model 1Model 2
  1. Notes: *p ≤ 0.05; **p ≤ 0.01; ***p ≤ 0.001; standard errors in parentheses.

RaceBlack (vs. White)−0.888***−0.771***
(0.098)(0.095)
Education<H.S. grad. (vs. H.S. grad) −0.049
 (0.146)
some college (vs. H.S. grad) 0.349***
 (0.089)
college grad. (vs. H.S. grad) 0.691***
 (0.127)
>college grad. (vs. H.S. grad) 0.882***
 (0.164)
Household income×103 dollars 0.001
 (0.002)
missing income 0.004
 (0.139)
Sexfemale (vs. male)0.226*0.307**
(0.092)(0.103)
Ageyears0.013***0.015***
(0.002)(0.002)
Household sizepersons0.0150.023
(0.023)(0.023)
Intercept 3.013***2.437***
(0.151)(0.148)

The results presented in Model 1 reveal a substantial average difference in EHEPI scores between Blacks and Whites, net of sex, age, and household size. The average EHEPI score of Blacks is estimated to be 0.89 points less than that of Whites, net of controls but prior to the introduction of SES. In Model 2, a net reduction in the race-nutrition relationship is observed following the introduction of SES. However, the average difference in health-related nutritional behaviour between Blacks and Whites continues to be significant and substantial, after adjustment for SES and the controls. The average EHEPI score of Blacks is estimated to be 0.77 points less than that of Whites, net of controls. In addition, the effect of educational attainment is strong and positive, with respondents who completed some college coursework, respondents who graduated from college, and respondents who participated in graduate school having average EHEPI scores that are 0.35, 0.69, and 0.88 points higher than that of respondents who attained only a high school diploma, respectively. Household income does not have a statistically significant relationship to the EHEPI, net of controls, nor is a significant difference in EHEPI scores observed for respondents who did not report income.

The consequential weight of the differences in nutritional behaviour between Blacks and Whites becomes evident when one accounts for the average index score across the sample. As mentioned earlier, the average EHEPI score in the sample is 3.68, indicating that respondents are accomplishing less than half of the recommended healthful nutritional recommendations. Given the globally low scores on the index, the estimated average discrepancy of 0.77 points between Blacks and Whites appears to be a substantively important nutritional difference.

Regression analysis of the EHEPI components

To explore which specific health-related nutritional behaviours underlie the significant difference in EHEPI scores between Blacks and Whites, I used random coefficient logistic regression to identify significant racial differences in each of the index component variables. All component variables were dichotomised to replicate the manner in which the variables are used to construct the EHEPI, and each variable was regressed on the set of independent variables employed in Model 2 (Table 3). The results are presented in Table 4.

Table 4. Estimated coefficients and standard errors for the random coefficient logistic regressions of each component of the Expanded Health Eating Practices Index on selected variables (n = 3,350)
  One fruit and one vegetableFive fruits and/or vegetablesDairy productsLow fat dairy productsWholegrain productsHigh fibre cerealsBeansDeep-fried foods and snacks
  1. Notes: *p ≤ 0.05; **p ≤ 0.01; ***p ≤ 0.001; standard errors in parentheses.

RaceBlack (vs. White)−0.324*−0.198 −0.988***−1.001***−0.015 −0.571***−0.154 0.557***
(0.138)(0.128)(0.134)(0.208)(0.162)(0.151)(0.156)(0.149)
Education< H.S. grad. (vs. H.S. grad)−0.0510.042−0.080−0.0340.009−0.1260.1940.128
(0.206)(0.245)(0.283)(0.198)(0.242)(0.347)(0.308)(0.192)
some college (vs. H.S. grad)0.410***0.397***0.0790.339*0.221−0.0110.1870.045
(0.107)(0.114)(0.197)(0.138)(0.129)(0.158)(0.144)(0.138)
college grad. (vs. H.S. grad)0.743***0.582**0.2540.597**0.378**0.0730.340*−0.176
(0.148)(0.189)(0.191)(0.232)(0.123)(0.186)(0.144)(0.132)
> college grad. (vs. H.S. grad)0.801**0.919***0.2160.663***0.562***0.1700.399*−0.271
(0.265)(0.209)(0.163)(0.175)(0.138)(0.162)(0.168)(0.173)
Household income×103 dollars0.0040.003−0.001 0.003−0.002−0.002−0.002 0.001
(0.003)(0.004)(0.003)(0.002)(0.002)(0.002)(0.003)(0.002)
missing income−0.010−0.1970.045−0.0450.0400.027−0.030−0.165
(0.154)(0.307)(0.277)(0.219)(0.187)(0.249)(0.282)(0.156)
Sexfemale (vs. male)0.438***0.370*−0.008 0.278**−0.049 0.065−0.050 −0.379***
(0.084)(0.148)(0.089)(0.097)(0.116)(0.155)(0.142)(0.106)
Ageyears0.017***0.015***−0.0050.0030.009***0.019***−0.006−0.016***
(0.003)(0.003)(0.006)(0.003)(0.003)(0.003)(0.003)(0.003)
Household sizepersons0.0380.0410.071−0.076*0.0540.0650.0300.079**
(0.033)(0.032)(0.058)(0.037)(0.029)(0.044)(0.036)(0.030)
Intercept −1.144***−2.140***1.423***−1.022***−0.760***−2.309***−1.059***0.070
(0.200)(0.200)(0.265)(0.282)(0.183)(0.358)(0.306)(0.203)

Consistent with the findings on the global index of nutritional behaviour, Blacks vary significantly and negatively from Whites on five of the eight index components, after adjustment for SES, age, sex, and household size. Specifically, the estimates indicate that Whites are 169 per cent, 172 per cent, 77 per cent, and 38 per cent more likely than are Blacks to consume dairy products, lowfat/nonfat dairy products, high fibre cereals, and at least one serving each of fruits and vegetables, respectively. Conversely, Blacks are 75 per cent more likely to consume deep-fried foods and snacks than are Whites. Significant differences are not observed in the likelihood of consuming wholegrain products, beans, and at least five servings of fruits and/or vegetables.

The finding of a significant difference in the likelihood of consuming one fruit and one vegetable is surprising given the finding of no significant difference in the likelihood of consuming at least five servings of fruits and/or vegetables. In part, this discrepancy may be due to the low probability generally of consuming five fruits and/or vegetables. On average, only 33.9 per cent of all respondents consumed at least five fruits and/or vegetables, with 35.4 per cent of Whites and 25.5 per cent of Blacks meeting this health-related objective. However, to explore further this incongruity, I separately regressed (using random coefficient linear regression) total fruit consumption and total vegetable consumption on the set of independent variables employed in Model 2 (Table 3). The results, which are presented in Table 5, indicate that Blacks differ significantly and negatively from Whites in total number of servings of vegetables consumed, but not in total number of servings of fruit consumed. On average, Blacks consume 0.35 fewer servings of vegetables and vegetable juices per day than do Whites, after adjustment for SES and controls. Given the low level of vegetable consumption in the sample (Total= 1.99; Blacks= 1.58; Whites = 2.06), it then is not surprising to find a significant difference between Blacks and Whites in the likelihood of consuming at least one serving of fruit and one serving of vegetables per day.

Table 5. Estimated coefficients and standard errors for the random coefficient linear regressions of total fruit consumption and total vegetable consumption on selected variables (n = 3,350)
  Total fruit consumptionTotal vegetable consumption
  1. Notes: *p ≤ 0.05; **p ≤ 0.01; ***p ≤ 0.001; standard errors in parentheses.

RaceBlack (vs. White)−0.005 −0.347**
(0.186)(0.114)
Education<H.S. grad. (vs. H.S. grad)−0.055−0.129
(0.247)(0.153)
some college (vs. H.S. grad)0.231*0.194*
(0.096)(0.083)
college grad. (vs. H.S. grad)0.463***0.345**
(0.124)(0.128)
>college grad. (vs. H.S. grad)0.554**0.591***
(0.170)(0.154)
Household income×103 dollars0.0030.002
(0.002)(0.003)
missing income0.044−0.177
(0.161)(0.154)
Sexfemale (vs. male)0.219*0.173*
(0.092)(0.071)
Ageyears0.016***0.004*
(0.003)(0.002)
Household sizepersons0.0240.049
(0.026)(0.028)
Intercept 0.567***1.357***
(0.171)(0.169)

Supplementary analysis

As discussed earlier, the differences between Blacks and Whites in nutritional healthfulness could be attributed to one or more of several mechanisms, including culturally-based dietary practices, residential segregation, differences in nutritional knowledge, and the disproportionate number of single-parent Black households. While it is not possible to test these mediating mechanisms in a comprehensive fashion with the CDPS data, additional supporting evidence for several of these explanations can be found in these data. Specifically, evidence for the roles of culture and nutritional knowledge in explaining racial differences in nutritional behaviour can be found in differences between Blacks’ and Whites’ explanations for food choices.

To elaborate, in the 1999 wave of the CDPS a question was added to the interview schedule addressing the one main reason why each respondent does not eat more fruits and vegetables. The responses, which are collapsed here into six categories, include: cost (respondent indicates that fruits and vegetables are too expensive), practice (respondent is unsure how to determine quality of fruits and vegetables or unsure how to prepare them), preference (respondent does not like the taste of fruits and vegetables or respondent's family does not like them), time (respondent indicates that fruits and vegetables take too much time to prepare), habit (respondent is not in the habit of, or not used to, eating fruits and vegetables), and knowledge (respondent believes that he/she already consumes enough fruits and vegetables). Given the literature addressing Black-White differences in nutritional behaviour, several preliminary hypotheses concerning this variable may be posited. On the one hand, if poverty is the primary reason for Black-White differences in nutritional behaviour (Blocker 1994, Kittler and Sucher 1989, Leigh 1995, Siewe 1999, Wickrama et al. 1999), and if the controls for SES addressed in this study are insufficient, one would expect to find that Blacks are more likely than are Whites to report that the cost of fruits and vegetables is prohibitive, even after accounting for household income and educational attainment. On the other hand, if culturally based food practices contribute to Black-White differences in nutritional behaviour (Semmes 1996), one would expect to find that Blacks are more likely to report preference and habit as the primary reasons for not eating more fruits and vegetables. In other words, to the extent that personal taste and food preparation habits are a product of culture, a finding of racial differences in the likelihood of reporting preference and habit as primary reasons for not consuming more fruits and vegetables represents evidence for culture as an obstacle to healthy nutritional behaviour. Lastly, if differences in nutritional knowledge, which may be, in part, a consequence of racially targeted advertising (Pratt and Pratt 1996), contribute to Black-White differences in nutritional behaviour, one would expect to find that Blacks are more likely than are Whites to report that they believe they eat enough fruits and vegetables, net of actual consumption.

To explore these preliminary hypotheses, I present in Table 6 a multinomial logistic regression of the primary reason for not eating more fruits and vegetables on race, education, household income, sex, age, household size, and total fruit and vegetable consumption. The results indicate that, relative to the likelihood of reporting that fruits and vegetables are ‘too expensive’, Blacks are significantly more likely than are Whites to report that they, or their families, do not like the taste (preference), that they are not in the habit of eating fruits and vegetables (habit), and that they believe they consume enough fruits and vegetables (knowledge). These findings indicate that cost is not the primary reason for Black-White differences in nutritional behaviour, at least as far as the consumption of fruits and vegetables is concerned. Rather, the evidence supports the argument that culture and knowledge are primary reasons for Black-White differences in consumption.8

Table 6. Estimated coefficients and standard errors for the multinomial logistic regression of respondent's one main reason for not eating more fruits and vegetables on selected variables (n = 726; comparison category = Cost)
  PracticePreferenceTime HabitKnowledge
  1. Notes: *p ≤ 0.05; **p ≤ 0.01; ***p ≤ 0.001; standard errors in parentheses.

RaceBlack (vs. White)−0.0751.216*0.0510.849*0.914*
(0.747)(0.480)(0.451)(0.423)(0.449)
Education< H.S. grad. (vs. H.S. grad)−0.8790.706−0.388−0.1680.228
(1.161)(0.613)(0.592)(0.546)(0.574)
some college (vs. H.S. grad)0.4360.2700.4090.4010.561
(0.571)(0.485)(0.401)(0.393)(0.424)
college grad. (vs. H.S. grad)0.1070.2670.4300.4550.459
(0.713)(0.610)(0.513)(0.509)(0.548)
> college grad. (vs. H.S. grad)0.1680.8280.8280.4111.310
(1.070)(0.925)(0.832)(0.843)(0.848)
Household income×103 dollars0.041**0.038**0.042***0.042***0.031**
(0.014)(0.012)(0.011)(0.011)(0.011)
missing income−0.359−0.490−0.352−0.111−0.757
(0.790)(0.681)(0.565)(0.547)(0.608)
Sexfemale (vs. male)−0.391−0.2770.042−0.426−0.466
(0.487)(0.415)(0.362)(0.354)(0.374)
Ageyears0.015−0.0050.0000.0000.029**
(0.013)(0.011)(0.009)(0.009)(0.010)
Household sizepersons−0.067−0.1340.017−0.0080.045
(0.172)(0.135)(0.113)(0.111)(0.118)
Consumptiontotal servings of f&v0.016−0.0010.051−0.0330.125*
(0.082)(0.070)(0.059)(0.060)(0.060)
Intercept −2.086*−0.528 −0.398 0.221−2.147**
(1.037)(0.811)(0.687)(0.668)(0.750)

Discussion

  1. Top of page
  2. Abstract
  3. Introduction
  4. Background
  5. Hypothesis
  6. The study Data and measures
  7. Methods
  8. Analysis
  9. Discussion
  10. Conclusion
  11. Acknowledgements
  12. References

In comparing Blacks and Whites in California, the findings presented here indicate that health-related nutritional behaviours differ significantly and in a direction consistent with the well-documented differential in health outcomes between the two groups, independent of SES. Blacks exhibit less healthy nutritional practice, relative to Whites, both globally and across a range of specific nutritional behaviours, after adjustment for the effects of educational attainment, household income, sex, age, and household size. Blacks do not exhibit nutritional behaviours that are healthier than those of Whites on any of the variables addressed in this study. In a supplementary analysis, I find that, at least as far as fruit and vegetable consumption is concerned, Blacks are more likely than are Whites to offer explanations for their nutritional choices that do not include the issue of cost, which adds to the credibility of a finding of a race-nutrition relationship that is independent of SES. Instead, the evidence suggests that culture and knowledge are the primary reasons Blacks eat fewer fruits and vegetables than do Whites, which is consistent with deductions drawn from the literature (e.g.Airhihenbuwa et al. 1996, Bock et al. 1998, Kiple and King 1981, Pratt and Pratt 1996, Semmes 1996, Witt 1999). Thus, considered globally, the findings presented here suggest that a relationship between race and nutritional healthfulness exists that is independent of SES.

Additionally, my findings provide indirect evidence of the causal role of residential segregation in racial differences in nutritional behaviour, specifically in the types of foods for which Blacks and Whites are found to differ in consumption. As discussed earlier in this paper, Blacks tend to be subjected to intense residential segregation into neighbourhoods often characterised by limited nutritional infrastructure (e.g. Bullard 1994, Cheadle et al. 1991, Emmons 2000, Hwang et al. 1985, Krebs-Smith and Kantor 2001, Leigh 1995, McKinnon 2003, Morland et al. 2002, Steinmetz and Iceland 2003). Consistent with this, significant differences were not identified between Blacks and Whites in the consumption of wholegrain products and beans, which are relatively inexpensive foods that require little special handling or storage, that have extended shelf lives, and that are subject to wide distribution across a diverse range of retail establishments. Blacks were found to consume significantly more deep-fried foods and snacks, which also require little special handling, are relatively inexpensive, are widely distributed, and represent a substantial proportion of the available nutritional choices at fast food restaurants and convenience stores. Conversely, the foods Blacks were found to be less likely to consume all require either special handling or special storage arrangements (e.g. dairy products and vegetables) and, therefore, are less widely distributed across a range of retail establishments, or are somewhat specialised items (e.g. high fibre cereals) confined primarily to supermarkets and large grocery stores. The absence of significant differences in fruit/fruit juice consumption may be explained by the wider availability of fruits and fruit juices relative to vegetables and vegetable juices. Thus, the findings of this analysis are consistent with the argument that racial residential segregation and neighbourhood nutritional infrastructure play a role in Black-White differences in health-related nutritional behaviour.

This study has, however, a number of important limitations. First, the data do not contain measures of SES other than educational attainment and household income. While these are well-established and useful measures of SES (Williams and Collins 1995), they do not allow for the ‘fine grain’ distinctions in social class discussed by Bartley and colleagues (1998) or account for other potentially important aspects of SES, such as accumulated wealth, income source, labour market position, or material deprivation (e.g. Cooper 2002, Huie et al. 2003, Macintyre et al. 2003, Smaje 1995, Smith and Kington 1997b). Thus, this study faces the same methodological problem that arises in all studies involving race and SES as primary independent variables, and perhaps to an even greater degree due to the limitations of the measures of SES available in these data. Namely, one cannot be certain if the observed residual racial differences (in this case, in nutritional behaviour) are a product of forces other than SES (as I argue) or simply a consequence of the operationalisation of SES (Cooper and Kaufman 1998, Kaufman et al. 1997, Nazroo 1998). In other words, the documented residual gap in nutritional healthfulness between Blacks and Whites yet may be a consequence of differences in SES between Blacks and Whites, albeit aspects of SES that are not addressed in this study.

Secondly, while my findings indicate what appears to be a sizeable residual racial gap in nutritional healthfulness, and while prior evidence suggests that nutritional healthfulness is strongly correlated with health outcomes, the absence of direct measures of health in the data does not allow for a determination concerning whether the nutritional differences identified here are of sufficient magnitude to explain the residual health differences between Blacks and Whites documented in previous studies. Moreover, the threshold at which poor nutritional behaviour begins to affect health outcomes is unclear. Thus, it is uncertain if the Black-White gap in nutrition documented here is large enough to make it a plausible explanation for the health gap.

Thirdly, while the results of this study are generalisable to California, caution must be exercised in generalising these findings to other populations. In particular, the wide availability and comparably low prices of fruits and vegetables in California generally would be expected to attenuate differences in EHEPI scores across strata of SES, leading to smaller coefficients than would be expected in geographic locales with higher prices. The other nutritional components of the EHEPI would be expected to vary less in availability and price, and therefore should have associated coefficients that are more stable vis-à-vis locale. Perhaps more importantly, the intervening mechanisms posited here to explain racial differences in nutrition, particularly culture and residential segregation, do vary regionally. Dietary practices associated with culture are not homogenous throughout the US, nor is the degree of residential segregation. Thus, considered globally, the generalisability of the findings is uncertain.

Conclusion

  1. Top of page
  2. Abstract
  3. Introduction
  4. Background
  5. Hypothesis
  6. The study Data and measures
  7. Methods
  8. Analysis
  9. Discussion
  10. Conclusion
  11. Acknowledgements
  12. References

The findings I present here suggest that previous studies, which have depended largely upon SES and correlates of SES as explanations for health differences between Blacks and Whites, may have excluded a potentially important variable that may aid in explaining racial disparities in health outcomes. However, these results cannot verify conclusively that nutritional behaviour explains the ‘unexplained’ racial gap in health outcomes left after adjustment for socioeconomic status. Such a conclusion would require a comparative longitudinal study of health outcomes that monitors nutritional behaviour, as well as other health risk behaviours, socioeconomic status, and demographic characteristics. Nevertheless, with due consideration to the limitations of this study, my findings support a conclusion in favour of the potential scientific fruitfulness of nutritional behaviour, independent of socioeconomic status, for explaining racial disparities in health outcomes. Given the well-established relationship between nutritional quality and chronic disease, the significant racial differences in nutritional behaviour identified here indicate that nutritional behaviour should be considered in future empirical work focused on explaining racial differences in health outcomes. Moreover, these findings constitute a step forward in the study of nutrition and inequalities, as recommended in prior work (Murcott 2002).

Notes
  • 1

    Chronic obstructive pulmonary disease, suicide, and Alzheimer's are the three major causes of mortality in the US for which Whites are disadvantaged relative to Blacks.

  • 2

    As noted by one referee, socioeconomic status is a complex and sometimes contested concept. In the interest of clarity, the phrase socioeconomic status, used at numerous points in this paper, refers to the concept marked by the intersection of education, occupation, income, and prestige (Hauser and Warren 1997).

  • 3

    For a recent and particularly comprehensive discussion of the pathways connecting SES to health, please see Crimmins et al. (2004).

  • 4

    Kumanyika (1993) argues that the nutrition-health relationship may be defined along three differing dimensions: undernutrition, the treatment of conditions that respond to dietary adjustments, and the increased risk of developing chronic diseases associated with poor nutrition. The primary theoretical focus of this paper as it pertains to the nutrition-health relationship is the last of the three.

  • 5

    This reliability score indicates a moderate intercorrelation among the items. However, the theoretical foundation of the items selected for the index is of substantially greater importance for this analysis than the interrelationship of the items, so the moderate reliability is of minor concern.

  • 6

    Random coefficient regression is appropriate in this case to account for variation in nutritional behaviour across the several waves of the biennial survey.

  • 7

    The mean EHEPI scores of respondents who were included and those who were excluded due to missing data on the independent variables were found to not differ significantly (excluded= 3.53; included= 3.68; t = 0.655).

  • 8

    The finding that knowledge is a primary reason for deficient fruit and vegetable consumption among Blacks is opposed to some extent by prior findings of no difference between Blacks and Whites in the accuracy of perceptions of dietary healthfulness (Airhihenbuwa et al. 1996, Variyam et al. 2001). Although the finding presented here does not contradict precisely the findings of prior studies, the discrepancy is worthy of further consideration in future research.

Acknowledgements

  1. Top of page
  2. Abstract
  3. Introduction
  4. Background
  5. Hypothesis
  6. The study Data and measures
  7. Methods
  8. Analysis
  9. Discussion
  10. Conclusion
  11. Acknowledgements
  12. References

I am grateful for the assistance of the following individuals: Diana Cassidy, Rodney Clark, Diane Felmlee, Susan Foerster, Christopher Gardner, Jennifer Gregson, Janet Hankin, Willard Hom, Renato Littaua, Monica Martin, Michelle Oppen, Mark Regnerus, Xiaoling Shu, and Sharon Sugerman. I thank the Public Health Institute and Freeman, Sullivan & Company for access to the California Dietary Practices Survey data, and Mark Hudes for his assistance in preparing the data. Finally, I thank the editors and anonymous referees of Sociology of Health & Illness for their insightful recommendations concerning improving this work. This research was made possible by funds received from the Cancer Research Fund under grant agreement No. 98-16026 with the Department of Health Services, Cancer Research Program.

References

  1. Top of page
  2. Abstract
  3. Introduction
  4. Background
  5. Hypothesis
  6. The study Data and measures
  7. Methods
  8. Analysis
  9. Discussion
  10. Conclusion
  11. Acknowledgements
  12. References
  • Airhihenbuwa, C.O., Kumanyika, S., Agurs, T.D., Lowe, A., Saunders, D. and Morssink, C.B. (1996) Cultural aspects of African American eating patterns, Ethnicity and Health, 1, 24560.
  • Bal, D.G., Foerster, S.B., Backman, D.R. and Lyman, D.O. (2001) Dietary change and cancer: challenges and future direction, Journal of Nutrition, 131, 1815.
  • Bartley, M., Blane, D. and Smith, G.D. (1998) Introduction: beyond the Black Report, Sociology of Health and Illness, 20, 5, 56377.
  • Berkman, L.F. and Kawachi, I. (2000) A historical framework for social epidemiology. In Berkman, L.F. and Kawachi, I. (eds) Social Epidemiology. Oxford: Oxford University Press.
  • Birch, L.L. (1999) Development of food preferences, Annual Review of Nutrition, 19, 4162.
  • Blair, S.N., Horton, E., Leon, A.S., Lee, I.M., Drinkwater, B.L., Dishman, R.K., Mackey, M. and Kienholz, M.L. (1996) Physical activity, nutrition, and chronic disease, Medicine and Science in Sports and Exercise, 28, 33549.
  • Blocker, D.E. (1994) Nutrition concerns of Black Americans. In Livingston, I.L. (ed.) Handbook of Black American Health. Westport: Greenwood Press.
  • Blocker, D. and Forrester-Anderson, I.T. (2004) Nutrition concerns of Black Americans. In Livingston, I.L. (ed.) Praeger Handbook of Black American Health, Volume 1. Westport: Praeger.
  • Bock, M.A., Read, M., Bruhn, C., Auld, G., Gabel, K., Lauritzen, G., Lee, Y., McNulty, J., Medeiros, D., Newman, R., Nitzke, S., Ortiz, M., Schutz, H. and Sheehan, E. (1998) Gender and ethnic differences in factors that influence food intake, Journal of Consumer Studies and Home Economics, 22, 2537.
  • Brondolo, E., Rieppi, R., Kelly, K.P. and Gerin, W. (2003) Perceived racism and blood pressure: a review of the literature and conceptual and methodological critique, Annals of Behavioral Medicine, 25, 5565.
  • Browning, C.R. and Cagney, K.A. (2003) Moving beyond poverty: neighborhood structure, social processes, and health, Journal of Health and Social Behavior, 44, 55271.
  • Bullard, R.D. (1994) Urban infrastructure: social, environmental, and health risks to African Americans. In Livingston, I.L. (ed.) Handbook of Black American Health. Westport: Greenwood Press.
  • Byrd, W.M. and Clayton, L.A. (2002) An American Health Dilemma: Race, Medicine, and Health Care in the United States 1900–2000, Volume 2. New York: Routledge.
  • Cheadle, A., Psaty, B.M., Curry, S., Wagner, E., Diehr, P., Koepsell, T. and Kirstal, A. (1991) Community-level comparisons between the grocery store environment and individual dietary practices, Preventive Medicine, 20, 25061.
  • Christenson, B.A. and Johnson, N.E. (1995) Educational inequality in adult mortality: an assessment with death certificate data from Michigan, Demography, 32, 21529.
  • Clark, R. (2003) Self-reported racism and social support predict blood pressure reactivity in Blacks, Annals of Behavioral Medicine, 25, 12736.
  • Clark, R. and Adams, J.H. (2004) Moderating effects of perceived racism on John Henryism and blood pressure reactivity in Black female college students, Annals of Behavioral Medicine, 28, 12631.
  • Clark, R., Anderson, N.B., Clark, V.R. and Williams, D.R. (1999) Racism as a stressor for African Americans: a biopsychological model, American Psychologist, 54, 80516.
  • Colditz, G.A., Atwood, K.A., Emmons, K., Monson, R.R., Willett, W.C., Trichopoulos, D. and Hunter, D.J. (2000) Harvard report on cancer prevention, Volume 4: Harvard cancer risk index, Cancer Causes and Control, 11, 47788.
  • Cooper, H. (2002) Investigating socio-economic explanations for gender and ethnic inequalities in health, Social Science & Medicine, 54, 693706.
  • Cooper, R.S. and Kaufman, J.S. (1998) Race and hypertension: science and nescience, Hypertension, 32, 81316.
  • Crimmins, E.M., Hayward, M.D. and Seeman, T.E. (2004) Race/ethnicity, socioeconomic status, and health. In Anderson, N.B., Bulatao, R.A. and Cohen, B. (eds) Critical Perspectives on Racial and Ethnic Differences in Health in Late Life. Washington, D.C.: The National Academics Press.
  • Emerson, M.O., Chai, K.J. and Yancey, G. (2001) Does race matter in residential segregation?: Exploring the preferences of White Americans, American Sociological Review, 66, 92235.
  • Emmons, K.M. (2000) Health behaviors in a social context. In Berkman, L.F. and Kawachi, I. (eds) Social Epidemiology. Oxford: Oxford University Press.
  • Erbe, B.M. (1975) Race and socioeconomic segregation, American Sociological Review, 40, 80112.
  • Eyre, H., Kahn, R. and Robertson, R.M. (2004) Preventing cancer, cardiovascular disease, and diabetes: a common agenda for the American Cancer Society, the American Diabetes Association, and the American Heart Association, CA: a Cancer Journal for Clinicians, 54, 190207.
  • Farley, R. (1977) Residential segregation in urbanized areas in the United States in 1970: an analysis of social class and racial differences, Demography, 14, 497518.
  • Farley, R. and Frey, W.H. (1994) Changes in the segregation of Whites and Blacks during the 1980s: small steps towards a more integrated society, American Sociological Review, 59, 2345.
  • Ferraro, K.F. and Farmer, M.M. (1996) Double jeopardy to health hypothesis for African Americans: analysis and critique, Journal of Health and Social Behavior, 37, 2743.
  • Ferraro, K.F., Farmer, M.M. and Wybraniec, J.A. (1997) Health trajectories: long-term dynamics among Black and White adults, Journal of Health and Social Behavior, 38, 3854.
  • Fields, J. (2004) America's families and living arrangements: 2003, Current Population Reports, P20-553. Washington, D.C.: U.S. Census Bureau.
  • Foerster, S.B., Wu, S., Gregson, J., Hudes, M. and Fierro, M.P. (1999) California Dietary Practices Survey: Overall Trends in Healthy Eating among Adults, 1989–1997, a Call to Action, Part 2. Sacramento: California Department of Health Services.
  • Fremont, A.M. and Bird, C.E. (2000) Social and psychological factors, physiological processes, and physical health. In Bird, C.E., Conrad, P. and Fremont, A.M. (eds) Handbook of Medical Sociology. New Jersey: Prentice Hall.
  • Gillum, R.F. (2004) Trends in cardiovascular diseases: an overview of evolving disparities. In Livingston, I.L. (ed.) Praeger Handbook of Black American Health, Volume 1. Westport: Praeger.
  • Harvard School of Public Health (1996) Harvard report on cancer prevention, Volume I: Causes of human cancer, Cancer Causes and Control, 7.
  • Hauser, R.M and Warren, J.R. (1997) Socioeconomic indexes for occupations: a review, update, and critique, Sociological Methodology, 27, 177298.
  • Hayward, M.D., Crimmins, E.M., Miles, T.P. and Yang, Y. (2000) The significance of socioeconomic status in explaining the racial gap in chronic health conditions, American Sociological Review, 65, 91030.
  • Hu, F.B., Rimm, R., Smith-Warner, S.A., Feskanich, D., Stampfer, M.J., Ascherio, A., Sampson, L. and Willett, W.C. (1999) Reproducibility and validity of dietary patterns assessed with a food-frequency questionnaire, American Journal of Clinical Nutrition, 69, 2439.
  • Hu, F.B., Rimm, E.B., Stampfer, M.J., Ascherio, A., Spiegelman, D. and Willett, W.C. (2000) Prospective study of major dietary patterns and risk of coronary heart disease in men, American Journal of Clinical Nutrition, 72, 91221.
  • Hughes, M. and Thomas, M.E. (1998) The continuing significance of race revisited: a study of race, class, and quality of life in America, 1972 to 1996, American Sociological Review, 63, 78595.
  • Huie, S.A.B., Hummer, R.A. and Rogers, R.G. (2002) Individual and contextual risks of death among race and ethnic groups in the United States, Journal of Health and Social Behavior, 43, 35981.
  • Huie, S.A.B., Krueger, P.M., Rogers, R.G. and Hummer, R.A. (2003) Wealth, race, and mortality, Social Science Quarterly, 84, 66784.
  • Hunt, M.O., Jackson, P.B., Powell, B. and Steelman, L.C. (2000) Color-blind: the treatment of race and ethnicity in social psychology, Social Psychology Quarterly, 63, 35264.
  • Hwang, S.S., Murdock, S.H., Parpia, B. and Hamm, R.R. (1985) The effects of race and socioeconomic status on residential segregation, 1970–80, Social Forces, 63, 73247.
  • Iceland, J. and Wilkes, R. (2006) Does socioeconomic status matter? Race, class, and residential segregation, Social Problems, 53, 24873.
  • Joshipura, K.J., Ascherio, A., Manson, J.E., Stampfer, M.J., Rimm, E.B., Speizer, F.E., Hennekens, C.H., Spiegelman, D. and Willett, W.C. (1999) Fruit and vegetable intake in relation to risk of ischemic stroke, Journal of the American Medical Association, 282, 12339.
  • Kant, A.K. (2000) Consumption Of energy-dense, nutrient-poor foods by adult Americans: nutritional and health implications. The third national health and nutrition examination survey, 1988–1994, American Journal of Clinical Nutrition, 72, 92936.
  • Karlsen, S. and Nazroo, J.Y. (2002) Agency and structure: The impact of ethnic identity and racism on the health of ethnic minority people, Sociology of Health and Illness, 24, 1, 120.
  • Kaufman, J.S., Cooper, R.S. and McGee, D.L. (1997) Socioeconomic status and health in Blacks and Whites: the problem of residual confounding and the resiliency of race, Epidemiology, 8, 6218.
  • Kawachi, I. (2000) Income inequality and health. In Berkman, L.F. and Kawachi, I. (eds) Social Epidemiology. Oxford: Oxford University Press.
  • Kiple, K.F. and King, V.H. (1981) Another Dimension to the Black Diaspora: Diet, Disease, and Racism. Cambridge: Cambridge University Press.
  • Kittler, P.G. and Sucher, K. (1989) Food and Culture in America. New York: Van Nostrand Reinhold.
  • Kong, S.B., Kong, W. and McAllister, S.B. (1994) Politics and health: the coming agenda for a multiracial and multicultural society. In Livingston, I.L. (ed.) Handbook of Black American Health. Westport: Greenwood Press.
  • Krebs-Smith, S.M. and Kantor, L.S. (2001) Choose a variety of fruits and vegetables daily: understanding the complexities, Journal of Nutrition, 131, 487501.
  • Krieger, N. (2000) Discrimination and health. In Berkman, L.F. and Kawachi, I. (eds) Social Epidemiology. Oxford: Oxford University Press.
  • Krieger, N., Rowley, D.L., Herman, A.A., Avery, B. and Phillips, M.T. (1993) Racism, sexism, and social class: implications for studies of health, disease, and well-being, American Journal of Preventive Medicine, 9, 82122.
  • Kristenson, M., Eriksen, H.R., Sluiter, J.K., Starke, D. and Ursin, H. (2004) Psychobiological mechanisms of socioeconomic differences in health, Social Science and Medicine, 58, 151122.
  • Kumanyika, S.K. (1993) Diet and nutrition as influences on the morbidity/mortality gap, Annals of Epidemiology, 3, 1548.
  • Kumanyika, S. and Adams-Campbell, L.L. (1991) Obesity, diet, and psychosocial factors contributing to cardiovascular disease in Blacks, Cardiovascular Clinics, 21, 4773.
  • Lantz, P.M., House, J.S., Lepkowski, J.M., Williams, D.R., Mero, R.P. and Chen, J. (1998) Socioeconomic factors, health behaviors, and mortality: results from a nationally representative prospective study of US adults, Journal of the American Medical Association, 279, 170308.
  • Leigh, W.A. (1995) The health of African American women, Health Issues for Women of Color: a Cultural Diversity Perspective. Thousand Oaks: Sage.
  • Lindquist, C.H., Gower, B.A. and Goran, M.I. (2000) Role of dietary factors in ethnic differences in early risk of cardiovascular disease and type 2 diabetes, American Journal of Clinical Nutrition, 71, 72532.
  • Liu, S., Manson, J.E., Stampfer, M.J., Rexrode, K.M., Hu, F.B., Rimm, E.B. and Willett, W.C. (2000a) Whole grain consumption and risk of ischemic stroke in women: a prospective study, Journal of the American Medical Association, 284, 153440.
  • Liu, S., Manson, J.E., Lee, I.M., Cole, S.R., Hennekens, C.H., Willett, W.C. and Buring, J.E. (2000b) Fruit and vegetable intake and risk of cardiovascular disease: the women's health study, American Journal of Clinical Nutrition, 72, 9228.
  • Livingston, I.L. (1994) Social status, stress, and health: Black Americans at risk. In Livingston, I.L. (ed.) Handbook of Black American Health. Westport: Greenwood Press.
  • Livingston, I.L. and Carter, J.J. (2004) Eliminating racial and ethnic disparities in health: a framework for action. In Livingston, I.L. (ed.) Praeger Handbook of Black American Health, Volume 2. Westport: Praeger.
  • Livingston, I.L., Brown, R.M. and Livingston, S. (2004) Vulnerability of African Americans to adverse health: the importance of social status and stress. In Livingston, I.L. (ed.) Praeger Handbook of Black American Health, Volume 1. Westport: Praeger.
  • Luke, A., Cooper, R.S., Prewitt, T.E., Adeyemo, A.A. and Forrester, T.E. (2001) Nutritional consequences of the African diaspora, Annual Review of Nutrition, 21, 4771.
  • Macintyre, S. and Ellaway, A. (2000) Ecological approaches: rediscovering the role of the physical and social environment. In Berkman, L.F. and Kawachi, I. (eds) Social Epidemiology. Oxford: Oxford University Press.
  • Macintyre, S., McKay, L., Der, G. and Hiscock, R. (2003) Socio-economic position and health: what you observe depends on how you measure it, Journal of Public Health Medicine, 25, 28894.
  • Massey, D.S., Condran, G.A. and Denton, N.A. (1987) The effect of residential segregation on Black social and economic well-being, Social Forces, 66, 2956.
  • McBarnette, L.S. (1996) African American women. In Bayne-Smith, M. (ed.) Race, Gender, and Health. Thousand Oaks: Sage.
  • McCullough, M.L., Feskanich, D., Rimm, E.B., Giovannucci, E.L., Ascherio, A., Variyam, J.N., Spiegelman, D., Stampfer, M.J. and Willett, W.C. (2000a) Adherence to the dietary guidelines for Americans and risk of major chronic disease in men, American Journal of Clinical Nutrition, 72, 122331.
  • McCullough, M.L., Feskanich, D., Stampfer, M.J., Rosner, B.A., Hu, F.B., Hunter, D.J., Variyam, J.N., Colditz, G.A. and Willett, W.C. (2000b) Adherence to the dietary guidelines for Americans and risk of major chronic disease in women, American Journal of Clinical Nutrition, 72, 121422.
  • McGinnis, J.M. and Foege, W.H. (1993) Actual causes of death in the United States, Journal of the American Medical Association, 270, 220712.
  • McKinnon, J. (2003) The Black population in the United States: March 2002. U.S. Census Bureau, Current Population Reports, Series P20-541. Washington, D.C.
  • Morland, K., Wing, S., Roux, A.D. and Poole, C. (2002) Neighborhood characteristics associated with the location of food stores and food service places, American Journal of Preventive Medicine, 22, 239.
  • Murcott, A. (2002) Nutrition and inequalities: a note on sociological approaches, European Journal of Public Health, 12, 20307.
  • Murphy, S.L. (2000) Deaths: final data for 1998, National Vital Statistics Reports, 48.
  • Mutchler, J.E. and Burr, J.A. (1991) Racial differences in health and health care service utilization in later life: the effect of socioeconomic status, Journal of Health and Social Behavior, 32, 34256.
  • Myers, H.F., Echiverri, A.T. and Odom, B.N. (2004) The role of the family in African American health. In Livingston, I.L. (ed.) Praeger Handbook of Black American Health, Volume 2. Westport: Praeger.
  • Nazroo, J.Y. (1998) Genetic, cultural or socio-economic vulnerability? Explaining ethnic inequalities in health, Sociology of Health and Illness, 20, 5, 71030.
  • Popkin, B.M., Siega-Riz, A.M. and Haines, P.S. (1996) A comparison of dietary trends among racial and socioeconomic groups in the United States, New England Journal of Medicine, 335, 71620.
  • Potter, L.B. (1991) Socioeconomic determinants of White and Black males’ life expectancy differentials, 1980, Demography, 28, 30321.
  • Pratt, C.A. and Pratt, C.B. (1996) Nutritional advertisements in consumer magazines: health implications for African Americans, Journal of Black Studies, 26, 50423.
  • Raudenbush, S.W. and Bryk, A.S. (2002) Hierarchical Linear Models: Applications and Data Analysis Methods. Thousand Oaks: Sage.
  • Robert, S.A. (1998) Community-level socioeconomic status effects on adult health, Journal of Health and Social Behavior, 39, 1837.
  • Robert, S.A. and Lee, K.Y. (2002) Explaining race differences in health among older adults, Research on Aging, 24, 65483.
  • Rogers, R.G. (1992) Living and dying in the U.S.A.: Sociodemographic determinants of death among Blacks and Whites, Demography, 29, 287303.
  • Schlundt, D.G., Hargreaves, M.K. and Buchowski, M.S. (2003) The eating behavior patterns questionnaire predicts dietary fat intake in African American women, Journal of the American Dietetic Association, 103, 33845.
  • Schnittker, J. (2004) Education and the changing shape of the income gradient in health, Journal of Health and Social Behavior, 45, 286305.
  • Schulz, A., Williams, D., Israel, B., Becker, A., Parker, E., James, S.A. and Jackson, J. (2000) Unfair treatment, neighborhood effects, and mental health in the Detroit metropolitan area, Journal of Health and Social Behavior, 41, 31432.
  • Semmes, C.E. (1996) Racism, Health, and Post-Industrialism: a Theory of African-American Health. Westport: Praeger.
  • Siewe, Y.J. (1999) Stages of limiting dietary-fat intake: implications for African Americans, Journal of Black Studies, 29, 73146.
  • Smaje, C. (1995) Health, ‘Race’ and Ethnicity: Making Sense of the Evidence. London: King's Fund Institute.
  • Smaje, C. (2000) Race, ethnicity, and health. In Bird, C.E., Conrad, P. and Fremont, A.M. (eds) Handbook of Medical Sociology. New Jersey: Prentice Hall.
  • Smith, J.P. and Kington, R.S. (1997a) Race, socioeconomic status, and health in late life. In Martin, L.G. and Soldo, B.J. (eds) Racial and Ethnic Differences in the Health of Older Americans. Washington, D.C.: National Academy Press.
  • Smith, J.P. and Kington, R. (1997b) Demographic and economic correlates of health in old age, Demography, 34, 15970.
  • Stampfer, M.J., Hu, F.B., Manson, J.E., Rimm, E.B. and Willett, W.C. (2000) Primary prevention of coronary heart disease in women through diet and lifestyle, New England Journal of Medicine, 343, 1622.
  • Steinmetz, E. and Iceland, J. (2003) Racial and ethnic residential housing patterns in places: 2000. Presented at the annual meeting of the Population Association of America. Minneapolis.
  • Subramanian, S.V., Acevedo-Garcia, D. and Osypuk, T.L. (2005) Racial residential segregation and geographic heterogeneity in Black/White disparity in poor self-rated health in the US: a multilevel statistical analysis, Social Science and Medicine, 60, 166779.
  • Tibbits, C. (1937) The socio-economic background of negro health status, Journal of Negro Education, 6, 41328.
  • Variyam, J.N., Shim, Y. and Blaylock, J. (2001) Consumer misperceptions of diet quality, Journal of Nutrition Education, 33, 31421.
  • Villemez, W.J. (1980) Race, class, and neighborhood: differences in the residential return on individual resources, Social Forces, 59, 41430.
  • Weisburger, J.H. (2000) Approaches for chronic disease prevention based on current understanding of underlying mechanisms, American Journal of Clinical Nutrition, 71, 171014.
  • Wickrama, K.A.S., Conger, R.D., Wallace, L.E. and Elder, Jr., G.H. (1999) The intergenerational transmission of health-risk behaviors: adolescent lifestyles and gender moderating effects, Journal of Health and Social Behavior, 40, 25872.
  • Willett, W.C. (1994) Diet and health: what should we eat? Science, 264, 5327.
  • Williams, D.R. and Collins, C. (1995) US socioeconomic and racial differences in health: patterns and explanations, Annual Review of Sociology, 21, 34986.
  • Williams, D.R. and Collins, C. (2001) Racial residential segregation: a fundamental cause of racial disparities in health, Public Health Reports, 116, 40416.
  • Witt, D. (1999) Black Hunger: Food and the Politics of U.S. Identity. New York: Oxford University Press.
  • Wrigley, N. (2002) ‘Food deserts’ in British cities: policy context and research priorities. Urban Studies, 39, 202940.
  • Wrigley, N., Warm, D., Margetts, B. and Whelan, A. (2002) Assessing the impact of improved retail access on diet in a ‘food desert’: a preliminary report. Urban Studies, 39, 206182.