Social inequalities, stressors and self reported health status among African American and white women in the Detroit metropolitan area

Social Science & Medicine 51 (2000) 1639±1653

www.elsevier.com/locate/socscimed

Social inequalities, stressors and self reported health status among African American and white women in the Detroit metropolitan area A. Schulz a,*, B. Israel a, D. Williams b, E. Parker a, A. Becker c, S. James d a

Health Behavior and Health Education, School of Public Health, University of Michigan, Ann Arbor, MI 48109-2029, USA b Sociology and the Institute for Social Research, University of Michigan, Ann Arbor, MI 48109-2029, USA c Community Health Sciences, School of Public Health and Tropical Medicine, Tulane University Medical Center, USA d Epidemiology & Health Behavior and Health Education, School of Public Health, University of Michigan, Ann Arbor, MI 481092029, USA

Abstract This article examines the cumulative e€ects of multiple stressors on women's health, by race and area of residence. Speci®cally, we examine socioeconomic status, experiences of unfair treatment and acute life events by race and residential location, and their cumulative e€ects on the health status of African American and white women living within the city of Detroit and in the surrounding metropolitan area. African American women, regardless of whether they live inside or outside the city, report more frequent encounters with everyday unfair treatment than white women. African American women who live in the city report a greater number of acute life events than white women who live outside the city. Regression analyses used to examine the cumulative e€ects of exposure to these stressors by race and area of residence show that: (1) socioeconomic status, everyday experiences with unfair treatment and acute life events each make a signi®cant contribution to di€erences in health status; and (2) the contribution of each of these variables to explaining variations in health status varies by area of residence. We suggest that di€erences in socioeconomic status, exposure to unfair treatment or discrimination and experiences of acute life events make signi®cant contributions to racial di€erences in women's health status. 7 2000 Elsevier Science Ltd. All rights reserved. Keywords: Self-reported health; Social inequality; Women; African American; Stress; USA

Introduction Di€erentials in morbidity and mortality between African American and European American women in the United States are persistent and well documented in the social science and epidemiological literatures.

* Corresponding author. Fax: +1-734-763-7379.

African American women experience greater morbidity and mortality at younger ages, and are more likely to experience the death of an infant, than are white women in the same age range (Collins & David 1990; Kreiger, Rowley, Herman, Avery & Phillips, 1993; Geronimus, 1992; Hummer, 1993). These di€erentials follow general and well-established relationships between socioeconomic position and morbidity and mortality, with lower socioeconomic status consistently

0277-9536/00/$ - see front matter 7 2000 Elsevier Science Ltd. All rights reserved. PII: S 0 2 7 7 - 9 5 3 6 ( 0 0 ) 0 0 0 8 4 - 8

1640

A. Schulz et al. / Social Science & Medicine 51 (2000) 1639±1653

related to higher levels of morbidity and mortality (Duleep, 1995; Kaplan, Pamuk, Lynch, Cohen & Balfour, 1996; LaVeist, 1996; Lillie-Blanton, Parsons, Gayle & Dievler, 1996; Pappas, Queen, Hadden & Fisher, 1993). However, some disparities remain between, as well as within, racial groups after adjusting for socioeconomic status with commonly used indicators such as individual level education or household income (Geronimus, Bound, Waidmann, Hillemeier & Burns, 1996; Lillie-Blanton et al., 1996; Polednak, 1993; Williams & Collins 1995). Public health researchers have increasingly questioned the utility of race as an explanatory research variable, urging that attention be shifted toward understanding the impact of social and economic divisions of populations based on systems of racial classi®cations on health (Fullilove, 1998; LaVeist, 1996; Williams, 1997). Women's experience of these racial classi®cations may di€er from those experienced by men, that is, the forms of discrimination experienced by African American men may di€er qualitatively from those experienced by African American women (Collins & David, 1990; Crenshaw, 1993, 1995; Kreiger et al., 1993). Understanding the ways that the particular stressors experienced by women may di€er by African American and white racial categories can contribute to an understanding of the pathways through which these di€erences contribute to racial disparities in health. In this paper we focus on understanding racial di€erences in subjective experiences of unfair treatment, household income and exposure to acute life events by African American and white women living in the Detroit metropolitan area. We examine the distribution of these experiences and their implications for health across four groups of women: African American women living in an economically marginalized area of Detroit, African American women living in the city of Detroit as a whole, and African American and white women living in Detroit metropolitan area but outside of the city of Detroit. Stressors and health The research presented here draws upon a conceptual model of the stress process (Avison & Gotlib, 1994; House, 1981; Israel, House, Schurman, Heaney & Mero, 1989; Israel & Schurman, 1990; Katz & Kahn, 1978) that identi®es the contributions of multiple factors to health outcomes. This model is a comprehensive framework that focuses attention on stressors, conceptualized as ``environmental demands (that) tax or exceed the adaptive capacity of an organism resulting in psychological and biological changes that may place persons at risk of disease'' (Cohen, Kessler & Gordon, 1997 p.3). The experience of stres-

sors may produce short term responses, which may be manifest as behavioral (e.g. smoking), physiological (e.g. elevated blood pressure), or psychological (e.g. tenseness) outcomes. These short term responses, if repeated over time, may lead to irreversible negative health e€ects. The relationships among stressors, short term responses and long term health outcomes may be modi®ed by what are sometimes referred to as ``conditioning variables'' (Israel et al., 1989). These include individual factors, such as coping behaviors, as well as social factors, such as access to material resources or social support (Cohen et al., 1997; House, Landis & Umberson, 1988; Parker & Eng, 1997; Sarason et al., 1994; Turner & Roszell, 1994). Stressors may be acute life events, such as the death of a loved one, or they may be chronic, ongoing life conditions, such as ongoing experiences with discrimination (Israel et al., 1989; Wheaton, 1994). Pearlin (1989) has suggested that the distribution of stressful life events and life circumstances re¯ect structural contexts, and there is considerable evidence to support this thesis. Lower status groups, for example women and African Americans, appear to be both more likely to experience stressful life conditions and to be more negatively a€ected by these stressors than higher status groups (Aneshensel, 1992; Kessler & Neighbors, 1986; Thoits, 1995; Turner, Wheaton & Lloyd, 1995). Relatively little research has speci®cally examined the particular stressors experienced by women of di€erent racial groups, the relationships of these stressors to social contexts, and the ways that di€erences in these experiences may contribute to racial di€erentials in women's health. Socioeconomic status There is considerable evidence that links lower socioeconomic status with poorer health, and racial di€erentials in socioeconomic status are an important factor in accounting for racial di€erences in health status (Lillie-Blanton et al., 1996; Williams & Collins, 1995). However, racial disparities in health status often remain even after adjusting for household or individual income and education (Lillie-Blanton et al., 1996; O'Campo, Xue, Wang & O'Brien Coughy, 1997; Williams & Collins, 1995). There are many potential explanations for these remaining di€erentials in health status. They include: limitations of current measures of social class, particularly for women (Duncan, Brooks-Gunn & Klebanov, 1994; Krieger & Fee, 1994; Krieger et al., 1993; McDonough, Williams, House & Duncan, 1999; Sheak, 1988); contributions of income inequalities to health disparities, as distinguished from the e€ects of material deprivation (Kaplan et al., 1996; Pappas et al., 1993); long term e€ects of childhood poverty on adult health

A. Schulz et al. / Social Science & Medicine 51 (2000) 1639±1653

status (Lynch, Kaplan & Cohen, 1994; Polednak, 1996); neighborhood e€ects which may in¯uence community level stressors or access to health care services (Krieger, Williams & Moss, 1997; O'Campo et al., 1997; Parker, Lichtenstein, Schulz, Steinman, Israel & James, 2000; Robert, 1998); and non-economic e€ects, such as physiological responses to subjective experiences of unfair treatment (Williams, 1997). In this paper, we focus our attention on the last two of these possibilities. That is, we examine the distribution of acute life experiences and subjective experiences of unfair treatment by race and by area of residence (which, we argue, in the Detroit area is also in¯uenced by systems of racial classi®cation). Furthermore, we examine the contribution of these factors to understanding racial di€erences in women's health status. Subjective discrimination and health Experiences with unfair treatment or discrimination have been linked to multiple health outcomes. For example, individuals who experience unfair treatment or hate crimes, whether on the basis of race, gender or sexual orientation, experience feelings of powerlessness and fear (Barnes & Ephross, 1994; Essed, 1990; Feagin & Sikes, 1994). A substantial body of research has established links between subjective experiences with unfair treatment with a broad range of mental and physical health outcomes (Dion, 1975; Dion, Dion & Pak, 1992; James, 1993a, 1993b; Kessler, Mickelson & Williams, 1999; Krieger, 1990, 1999; Krieger & Sidney, 1996; Noh, Belser, Kaspar, Hou & Rummens, 1999; Ren, Amick & Williams, 1999; Williams, Yu, & Jackson, 1997a; Williams, Yu, Jackson & Anderson, 1997b). To the extent that women are more likely than men to experience unfair treatment on the basis of gender, and that African Americans are more likely than whites to experience unfair treatment on the basis of race, both groups are more vulnerable to the health e€ects of such unfair treatment. African American women may experience unfair treatment on the basis of both race and gender, with subsequent implications for mental and physical health outcomes. Systemic discrimination In addition to subjective experiences with unfair treatment, social systems organized around racial and gender inequalities also in¯uence health outcomes through pathways that may or may not be subjectively perceived by the individual. For example, social processes that contribute to the creation of racially segregated, economically marginalized communities with insucient social, structural and economic resources (e.g. racially based housing discrimination), contribute to poorer health outcomes among residents of those

1641

communities (Collins, 1997; Massey & Denton, 1993; Polednak, 1993; Williams & Collins, 1995). African Americans, who are disproportionately represented in racially segregated urban communities, encounter the burden of decreased access to employment, municipal services and other institutional supports for community life (Massey & Denton, 1993; Massey & Fischer, 1998; Sugrue, 1996). Residential segregation primarily occurs on the basis of race and class, and thus its in¯uences on access to employment a€ect both women and men. However, African American women's employment experiences and economic status are in¯uenced by employer's stereotypic perceptions of African American women Ð that is, by the combined categories of race and gender (Kennely, 1999; Krieger et al., 1993). Furthermore, women who are raising children as single heads of households encounter the dual strain of being the sole economic provider for their families at the same time that they struggle to meet the social and emotional needs of their children (Edin & Lein, 1996). On the east side of Detroit, approximately two-thirds of women who are raising children report that they are doing so as single heads of households, and the incomes of those households are substantially lower than those in which two parents are present (Schulz et al., 1999). Recent research further underscores the contribution of community or neighborhood e€ects to health status (Geronimus et al., 1996; Haan, Kaplan & Camacho, 1987; Robert, 1998; Schulz et al., 2000), linking economic disinvestment to neighborhood conditions, which in turn may shape health outcomes (James, Schulz & vanOlphen, 1999). The public health impact of these processes is seen in the disproportionate burden of morbidity and mortality borne by residents of highpoverty urban communities, increasingly recognized as a concern by public health practitioners and researchers (Freudenberg, 1998; Fullilove, 1998; Fullilove, Green & Fullilove, 1999). Cumulative e€ects The cumulative e€ects of exposure to multiple stressors over time has been hypothesized to contribute to more rapid deterioration in the health of African American women when compared to white women, an e€ect that has been termed the ``weathering hypothesis'' (Geronimus, 1992). Turner and colleagues (1995) have argued for more comprehensive measures of the variety of ways that these di€erential life experiences may contribute to di€erential exposure to stressors and ultimately to di€erentials in health status. We consider both race and gender to be social constructs that shape, and are shaped by, patterns of social interaction that substantially in¯uence the conditions of peoples' lives (Auerbach & Figert, 1995; Essed, 1990; Feagin,

1642

A. Schulz et al. / Social Science & Medicine 51 (2000) 1639±1653

1991; James, 1993a; LaVeist, 1996; Massey & Denton, 1993; Polednak, 1996). In this paper, we examine the distribution of a number of stressors that may contribute to racial di€erences in women's health status, and the ways that these stressors are distributed by race and by area of residence1.

Study setting As one of the most racially segregated cities in the United States (Massey & Denton, 1993; Sugrue, 1996), Detroit o€ers an important opportunity to examine the combined e€ects of race-based residential segregation and subjective experiences of discrimination on women's health. Over the past four decades, white residents and employers in the Detroit area have increasingly relocated from aging urban industrial areas to the everexpanding suburban areas. There is evidence that the proportion of white employees to African American employees increased in corporations that relocated to the suburbs (Zax & Kain, 1996), while employment opportunities declined in increasingly segregated urban communities (Jaynes & Williams, 1989; Massey & Denton, 1993; Sugrue, 1996; Thomas, 1997). Some urban employers responded to this situation by reducing rates of pay, bolstered by the knowledge that the available labor pool far exceeded the number of positions available in the area (Turner, 1997). The Detroit metropolitan area re¯ects processes of economic disinvestment and increasing race-based residential segregation over the past several decades that have been experienced by urban industrial areas throughout the northeast and midwest (Schulz et al., 2000; Sugrue, 1996). Little research has speci®cally examined the implications of these patterns for the lives of women living in urban communities and surrounding areas, nor the relationship of those patterns to racial di€erences in women's health status. The Detroit metropolitan area o€ers an opportunity to examine the relationship of these economic and social 1

Despite considerable ethnic variation and intermixing, individuals of African and European ancestry in the United States have generally been identi®ed as either black or white. In this paper, we use self-identi®ed race as an indicator of group membership, but with the assumption that this identity has been constructed through interactions with others who identify the individual as within a particular social category, such as black or African American. Through the paper, we refer to women who self-identify as ``black'' or ``African± American'' as African American and to women who selfidentify as ``white'', ``caucasian'' or ``European American'' as white.

patterns to the stressors experienced by women, as these may vary by race and area of residence, and their implications for health. Hypotheses for this study This study begins with the premise that race is a social category that shapes patterns of social interaction and the distribution of social resources di€erentially for African American and white women. The stress process framework is used to examine the di€erential distribution of stressors by area of residence and the cumulative e€ects of these di€erent life experiences on women's health status. Speci®cally, we test the hypotheses listed below. 1. African American women will report lower levels of household income at any given level of education than white women living in the same metropolitan area. Furthermore, African American women living in more economically marginalized urban areas will report lower household incomes at each level of education than women who live in less economically marginalized areas. 2. African American women will experience more frequent unfair treatment or discrimination than white women living in the same metropolitan area. 3. African American women who live in economically marginalized urban communities will experience a greater number of acute life events than their counterparts in more economically stable communities. 4. The cumulative e€ects of di€erentials in socioeconomic status, experiences with unfair treatment or discrimination and acute life events will be associated with racial di€erences in women's health status.

Methods Sample Data for this study are drawn from two surveys conducted in the Detroit metropolitan area in 1995 and 1996. In 1996 a survey was conducted with 700 women aged 18 and older living in a geographically de®ned area on the east side of Detroit who care for children under 18 years of age at least ®ve hours a week. This community survey was conducted by the East Side Village Health Worker Partnership under the auspices of the Detroit Community-Academic Urban Research Center (URC) funded by the Centers for Disease Control and Prevention. The East Side Village Health Worker Partnership is a community-based participa-

A. Schulz et al. / Social Science & Medicine 51 (2000) 1639±1653

tory research partnership that uses a lay health advisor approach to address and understand stressful life conditions and health protective factors for women and children on the east side of Detroit (Parker, Schulz, Israel & Hollis, 1998; Schulz et al., 1999). This survey (hereafter referred to as the VHW survey) was conducted in a geographically de®ned area on the east side of Detroit which is highly segregated by race (96% African American) and where 37% of all families and 65% of female headed families with children live below the poverty line. The VHW survey used a two-stage random sampling process. Households were randomly selected from a listing of all households in the de®ned area and, if more than one woman in a selected household met the eligibility criteria, respondents were randomly selected from the eligible members within the household. The response rate for this survey was 80%, with 97% of respondents self-reporting their race/ethnicity as African American. The analyses reported in this paper use only African American respondents from the VHW survey. Data are also analyzed from the 1995 Detroit Area Study (DAS), an annual survey of the Detroit metropolitan area encompassing Wayne (which includes the city of Detroit), Oakland and Macomb counties, which focused on social inequality and health. The DAS is a multistage area probability sample of 1139 adult respondents 18 years of age and older living in the tricounty area. The response rate for this survey was 70%, and the ®nal sample included 520 respondents who self-identi®ed as white, 586 who identi®ed as African American and 33 respondents who identi®ed as either Asian, Hispanic or American Indian (Williams et al. 1997b). The analyses reported in this paper use only African American and white female respondents from this survey (397 African American and 296 white women; 333 of the African American female respondents lived within the Detroit city limits and 64 lived outside the city; 15 of the white female respondents lived in the city limits and 281 lived in the metropolitan area). The VHW survey area is highly segregated on the basis of race (96% African American) and 37% of families living in that area of the city have incomes below the poverty line. Within the city of Detroit as a whole, 77% of residents are African American and the poverty rate is somewhat lower, with 31% of families below the poverty line. In the areas surrounding the city of Detroit, just 7% of residents are African American and 6% of households fall below poverty. The use of both the VHW and the DAS survey data allows comparisons of the life experiences of African Ameri2 See Williams et al. (1997a) for a discussion of the development of this scale.

1643

can and white women living in residential areas that di€er in terms of the extent of race-based residential segregation and the concentration of household poverty. This analysis divides respondents into four subgroups: African American women living on Detroit's east side (VHW survey area); African American women living within the city of Detroit as a whole; African American women living in the Detroit metropolitan area but outside of the Detroit city limits; and white women living in the Detroit metropolitan area but outside of the Detroit city limits. Ideally, we would have included a ®fth group, white women living within the city of Detroit. However, the very low numbers in this category (n = 15), precluded meaningful analysis for this group. Measures Both the DAS and the VHW surveys used a single item measure of general health status in which respondents were asked ``in general, would you say your health is: excellent, very good, good, fair or poor''. This item, which has been shown to be a reliable indicator of future population mortality (Idler & Benyamini, 1997), is used as the dependent variable in the analyses reported here. The independent variables included age in years as a sociodemographic control variable. Education was measured using an ordinal scale, (1=less than high school graduation; 2=high school graduation; 3=some college; 4=college graduate), and income (measured as total family income in the past year, with ten response categories, ranging from less than $5000 to more than $50,000). To translate these ordinal categories into more meaningful income ®gures in the descriptive data reported in this paper, we used the midpoint income ®gure for each category (e.g. for category 1= < $5000, we used the midpoint of $2500) when calculating mean incomes. For the uppermost category (>$50,000) we used $50,000 in our calculations. This ¯attens the range of income distributions somewhat, so that the results in Table 2 represent a conservative estimate of group di€erences in income. Race was measured by respondent self-identi®cation (0=white, 1=African American). Two measures of exposure to stressors were used as independent variables. The ®rst was a measure of daily encounters with discrimination. This scale was the mean of ®ve items that assessed the frequency of experiences of unfair treatment in the 12 months prior to the interview. Two representative scale items are: how often have you been treated with less courtesy than others, and how often have other people acted as if they were better than you2. Response categories ranged from 1=never to 5=very often. Cronbach's alpha for

1644

A. Schulz et al. / Social Science & Medicine 51 (2000) 1639±1653

the ®ve item scale was 0.82 and 0.81 for the VHW and DAS samples respectively3. The second measure of stressors was an index representing the sum of ®ve acute life events reported by respondents in the 12 months preceding the administration of the survey. Representative items included: whether the respondents had retired when they didn't want to, had someone close to them die, or had been the victim of a serious physical attack or assault.

Data analysis Diagnostic tests for heteroskedasticity and multicollinearity of the independent and dependent variables were conducted. Plots of the dependent variables were examined against the independent variables to assess whether assumptions for homoskedasticity were upheld. The variance in¯ation factors (VIF) and tolerances for the predictor variables were also examined. The assumption of homoskedasticity was met, and the scales used in the analyses were within prescribed tolerances (Neter, Wasserman & Kutner, 1985). Simple descriptive statistics using the data from the DAS and the VHW Survey compared the responses of African American women living within the VHW study area, African American women in the city of Detroit, and African American and white women in the surrounding metropolitan area by self reported health status, income, education, frequency of experiences with everyday discrimination or unfair treatment and acute life events. A series of regression analyses were conducted to examine the independent and cumulative e€ects of socioeconomic status, daily discrimination and acute stressors on general health status within each of the four subgroups. Finally, the samples were combined and a series of regression analyses was conducted to examine relationships among age, race, income, everyday discrimination or unfair treatment, acute life events and residence, with self-reported health as the dependent variable. 3

The alpha for the daily discrimination scale in the DAS sample reported here was calculated using only female African American and white respondents from that study and may di€er slightly from results reported elsewhere using the full sample. 4 Although not shown in Table 1, age speci®c self-reported general health status for women living on Detroit's east side were not signi®cantly di€erent than those reported by African American women city wide. 5 Comparisons between African American and white respondents in the Detroit metropolitan area are reported in more detail in Williams et al. (1997b).

Results Health status Overall, white women in the sample reported more favorable general health status than did African American women (Table 1). When examined by race and area of residence, the results indicate that African American women, regardless of where they lived, reported signi®cantly poorer health status than white women living outside the city limits ( p < 0.01). There were not signi®cant di€erences in self-reported health status between African American women living on the east side, in the city as a whole, or outside the city4. Thus the di€erences in self-reported health status that appear between Detroit city and the surrounding metropolitan area appear to be primarily a function of poorer health status among African American women, who are disproportionately represented within the city limits. Education and household income Table 1 also shows mean education and income for women within each of the four subsamples. No signi®cant di€erences were found in mean level of education between the four groups, i.e. African American women in the east side of Detroit or Detroit as a whole or between African American or white women living outside the city but in the metropolitan area. However, as shown in Table 2, there were di€erences in mean household income by both race and area of residence at each level of education. White women living outside the city reported higher incomes than African American women living in the same metropolitan area at each level of education ( p < 0.05). Furthermore, mean income at each educational level was higher for African American women city-wide than for African American women living on Detroit's east side. African American women living on the east side with some college education reported household incomes roughly similar to those for African American women living elsewhere in the city and outside the city who were high school graduates, and comparable to household incomes for white women living outside the city who had not completed high school. Daily discrimination Table 1 also compares reported experiences of unfair treatment or discrimination for each of the four groups5. Regardless of where they lived, African American women reported signi®cantly more frequent experiences with everyday unfair treatment than white women living outside the city ( p < 0.001).

A. Schulz et al. / Social Science & Medicine 51 (2000) 1639±1653

Acute life events Table 1 shows mean number of acute life events experienced in the preceding 12 months for women in each of the four subgroups. White women living outside Detroit reported the fewest acute life events, followed by African American women living outside the city. African American women living in the city reported the greatest number of life events. Di€erences in the number of acute life events experienced by African American and white women who lived outside the city were not statistically signi®cant. However, African American women who lived in the city experienced signi®cantly more acute life events than white women who lived outside the city ( p < 0.01). Table 3 shows that there were di€erences by race and by area of residence in the type of acute life events experienced. While women in each of the four groups were equally likely to report that they had retired when they did not want to in the past year, African American women, regardless of area of residence, were more likely to report that they had been the victim of a physical attack than were white women who lived outside the city ( p < 0.05). African American women who lived in the city were more likely to report that someone close to them had died in the past 12 months than were women of either race who lived outside the city ( p < 0.05). These patterns suggest that risk of

1645

experiencing di€erent acute life events di€ers by both race and by area of residence. Stressors and health status Regression analyses were carried out to examine the relationships between socioeconomic status, experiences with everyday unfair treatment and acute life events, by race and area of residence, to indicators of general health. Table 4 shows results of regression analyses for each of the four subgroups. Self-reported general health status is the dependent variable, regressed on age, education, income, everyday experiences of unfair treatment and acute life events. Everyday experiences with unfair treatment ( p < 0.01) and acute life events ( p < 0.05) were signi®cantly and negatively associated with self reported general health for African American women living in the city of Detroit, above and beyond the e€ects of the sociodemographic variables. Neither everyday experiences with unfair treatment nor acute life events were signi®cant correlates of general health status for women who lived outside the city of Detroit, regardless of race. Finally, we examined the extent to which race and area of residence contributed to racial di€erentials in health status, above and beyond the contributions made by experiences of unfair treatment and acute life events. Table 5 shows results from a series of re-

Table 1 Mean self-reported general health status, age, education, income, everyday unfair treatment, and number of acute life events for African American women on Detroit's east side, Detroit City, and African American and white women living outside the city of Detroit in the Detroit metropolitan area African American women, east side Detroit (n = 679)

Self reported general 3.3 (1.0) health status (1=poor, 5=excellent) Age 38.9 (16.1) Education (1=less 2.1 (0.9) than high school, 4=college graduate) Income 14,183 (2.3)b Everyday unfair 2.3 (0.8) treatment (1=never, 5=very often) Acute life events 0.79 (0.9) a

African American women, Detroit city (n = 333)

African American women in Detroit metropolitan area (n = 64)a

White women in Detroit metropolitan area (n = 276)a

3.2 (1.1)

3.2 (1.2)

3.6 (1.0)c

45.2 (16.1) 2.5 (0.9)

44.8 (20.8) 2.6 (1.0)

47.7 (18.1) 2.7 (1.0)

23,116 (2.8) 2.2 (0.8)

27,227 (2.7) 2.3 (0.9)

35,045 (2.4) 1.8 (0.8)d

0.73 (0.8)

0.55 (0.8)

0.39 (0.6)e

Excluding women living in the city of Detroit. One SD=$5,000. c White women reported signi®cantly better health than African American women, regardless of area of residence ( p R 0.01). d White women reported signi®cantly fewer experiences with unfair treatment than African American women ( p R 0.001). e White women reported signi®cantly fewer acute life events than African American women living in Detroit ( p R 0.01). Di€erences between African American and white women living outside the city were not statistically signi®cant. b

14,417 (2.4) 19,080 (2.5) 25,420 (2.6) 37,188 (2.3) 23,116 (2.8)

$8493 (1.6)b 12,771 (2.1) 19,118 (2.4) 27,394 (3.0) 14,183 (2.3)c

Less than high school graduation (1) High school graduate or GED (2) Some college (3) College graduate (4) Mean 30,417 (2.1) 41,923 (1.5) 27,227 (2.7)

21,912 (3.0)

9500 (1.0)

African American women in Detroit metropolitan area (excluding Detroit city residents) (n = 64)

36,134 (2.4) 42,029 (1.9) 30,045 (2.4)d

33,315 (2.4)

20,690 (2.0)

White women in Detroit metropolitan area (excluding Detroit city residents) (n = 276)

b

Results shown are not age-adjusted. While age was signi®cantly associated with education, the di€erences reported here remain signi®cant after controlling for age. One SD=$5000. c The mean income for African American women on Detroit's east side was signi®cantly lower than the mean income for African American women city wide, and for African American and white women in the metropolitan area outside the city ( p R 0.001). d White women in the metropolitan area reported household incomes signi®cantly higher that African American women living in the metropolitan area ( p R 0.05).

a

African American women, Detroit city (n = 333)

African American women on Detroit's east side (n = 679)

Educational level

Table 2 Mean income by level of education among African American women on Detroit's east side, African American women in Detroit, and African American and white women living outside the city of Detroit in the Detroit metropolitan areaa,b

1646 A. Schulz et al. / Social Science & Medicine 51 (2000) 1639±1653

6.6 (22) 3.3 (11) 44.4 (147) 5.1 (17) 14.0 (47) 0.73 (331)

5.0 (33) 5.0 (33) 43.6 (290) 14.3 (95) 11.4 (77) 0.79 (662)

African American women, Detroit city (n = 331)

0.55 (62)

7.8 (5)

7.9 (5)

29.0 (18)

4.8 (3)

4.8 (3)

African American women Detroit metropolitan area (excluding Detroit city) (n = 63)

0.39 (271)

4.0 (11)

2.6 (7)

27.9 (76)

3.7 (10)

0.7 (2)

White women Detroit metropolitan area (excluding Detroit city) (n = 271)

b

Results shown are not age-adjusted. Di€erences in mean number of acute life events reported by African American women on the east side and city wide, and between African American women city wide and those living outside the city were not statistically signi®cant. White women living outside of the city reported signi®cantly fewer acute life events ( p R 0.01) than African American women living outside of the city.

a

Percent reporting they were victim of physical attack or assault Percent reporting they retired when they didn't want to Percent reporting anyone close to them died Percent reporting they moved to a worse residence Percent reporting their home had been robbed or burglarized Mean number of acute life events reportedb

African American women on Detroit's east side (n = 664)

Table 3 Percent of respondents who report each of ®ve acute life events, and mean number of acute life events among African American women on Detroit's east side, African American women in Detroit, and African American and white women living outside the city of Detroit in the Detroit metropolitan areaa

A. Schulz et al. / Social Science & Medicine 51 (2000) 1639±1653 1647

1648

A. Schulz et al. / Social Science & Medicine 51 (2000) 1639±1653

Table 4 Self-reported general health status regressed on age, household income, education, daily discrimination and acute life events by race and area of residence

Independent variables

African American women Detroit east side (n = 664)

African American women, Detroit city (n = 331)

African American women Detroit metropolitan area (excluding Detroit city) (n = 63)

White women Detroit metropolitan area (excluding Detroit city) (n = 271)

b (SE)

B (SE)

B

b (SE)

B (SE)

(0.04) ÿ0.33 (0.07) 0.12 0.02 0.11 (0.07) ÿ0.15

ÿ0.20 0.35 0.00 0.02

B

Age ÿ0.24 (0.03) ÿ0.35a Education 0.06 (0.05) 0.06 Income 0.09 (0.02) 0.20 Everyday unfair ÿ0.13 (0.05) ÿ0.10 treatment Acute life ÿ0.10 (0.04) ÿ0.09 events Constant 3.85 (0.17) R2 0.17 F 26.87 Signi®cant 0.000 a

p R 0.05.



p R 0.01.



ÿ0.23 0.14 0.04 ÿ0.20

ÿ0.19 (0.07) ÿ0.15 4.00 (0.31) 0.17 13.36 0.000

(0.10) ÿ0.30 (0.20) 0.28 (0.07) 0.00 (0.18) 0.01

ÿ0.21 (0.23) ÿ0.13 3.02 0.16 2.17 0.071

ÿ0.15 0.13 0.09 ÿ0.12

B

(0.04) ÿ0.25 (0.06) 0.12 (0.03) 0.21 (0.09) ÿ0.08

ÿ0.15 (0.09) ÿ0.09 3.40 (0.34) 0.16 10.34 0.000

p R 0.001.

gression analyses conducted with African American and white respondents living in Detroit and the surrounding metropolitan area, using the DAS dataset6. General self-reported health status is regressed, in the ®rst model, on age, education, income and race (African American=1, white=0), and in the second model, on age, education, income and Detroit city residence. Models 3 and 4 examine the contributions of race and Detroit city residence, respectively, above and beyond the contributions of the sociodemographic control variables, everyday unfair treatment and acute life events. Results shown in Model 1 of Table 5 indicate that race is signi®cantly associated with health outcomes, controlling for age, education and income ( p < 0.001). Model 3 shows that race remains signi®cantly associated with general health status once the e€ects of everyday unfair treatment and acute life events are included in the model ( p < 0.01), although the size of

6

B

The VHW sample was not included in this analysis because of the similarity in the size and signi®cance of the predictor variables between the VHW and the DAS sample of African American women in Detroit found in the preceding analyses. 7 Race and area of residence were not included in the same model in these analyses due to the high correlations between the two variables (0.830), and resultant high standard errors and unacceptably low tolerances.

the coecient is reduced. In Model 2, Detroit residence was signi®cantly associated with health outcomes, again controlling for age, education and income ( p < 0.05). Once the e€ects of subjective experiences with unfair treatment and acute life events are included in Model 4, residence within or outside of the Detroit city limits is no longer signi®cantly associated with general health status7. These results suggest that the relationship of race to general health status is associated with, but not entirely accounted for by, di€erentials in unfair treatment and acute life events. While residence inside or outside the city limits is signi®cantly associated with health status above and beyond the e€ects of age, education and income, this relationship is no longer signi®cant when the e€ects of everyday unfair treatment and acute life events are included in the model. Discussion and limitations The results reported in this paper generally support the hypothesis that women's subjective experiences of unfair treatment are associated with racial categories, and that their exposure to the acute life events assessed in this study are associated with both race and with area of residence. African American women report lower mean household incomes at any given level of education than white women, and this e€ect is exacerbated for African American women who live on the

ÿ0.194 (0.082) ÿ0.089 3.347 0.140 27.450 3.529 0.155 20.851

ÿ0.015 0.158 0.075 ÿ0.345 Age Education Income Race (0=white, 1=African American) Everyday unfair treatment Acute life events Detroit city residence (1=Detroit, 0=non-Detroit) Constant R2 F signi®cant

a African American and white women living in the city of Detroit and outside the city of Detroit in the metropolitan area were included in this table. VHW survey respondents were not included in this analysis due to their similarity to women in the city of Detroit as a whole. b p < 0.05. p < 0.01. p < 0.001.

4.250 0.213 30.042

ÿ0.170 (0.52) ÿ0.123 ÿ0.270 (0.044) ÿ0.225 ÿ0.110 (0.081) ÿ0.050 4.160 0.204 28.489

ÿ0.302 0.149 0.088 ÿ0.019 (0.002) 0.170 (0.44) 0.054 (0.24) ÿ0.015 (0.002) 0.157 (0.45) 0.088 (0.25)

b (SE) Independent variables

(0.002) (0.45) (0.25) (0.83)

ÿ0.247b 0.139 0.123 ÿ0.156

ÿ0.239 0.138 0.144

ÿ0.019 0.170 0.044 ÿ0.251 ÿ0.147 ÿ.266

(0.002) (0.44) (0.24) (0.82) (0.52) (0.043)

ÿ0.303 0.148 0.072 ÿ0.113 ÿ0.107 ÿ0.221

b(SE) b(SE) b(SE) B

B

Model 3 Model 2 Model 1

Table 5 Self-reported general health, regressed on age, race, household income, everyday unfair treatment, and acute life events (n = 681)a

B

Model 4

B

A. Schulz et al. / Social Science & Medicine 51 (2000) 1639±1653

1649

east side of Detroit. These ®ndings are consistent with theoretical and empirical work that suggests that racially-based residential segregation combined with economic disinvestment in a€ected areas contributes to decreased access to economic resources among residents of those areas (Massey & Denton, 1993; Sugrue, 1996; Wilson, 1987). African American women in the Detroit metropolitan area, whether inside or outside the city limits, reported signi®cantly more frequent experiences with unfair treatment than white women. African American women reported similar frequencies of experiences with unfair treatment regardless of whether they lived on the east side, city-wide or outside the city limits in the Detroit metropolitan area; that is, unfair treatment was associated with race regardless of area of residence. The third hypothesis, that acute life events are structured by both race and residence, received quali®ed support on the basis of the results reported here. African American women who live outside the city reported more frequent acute life events than white women who live outside the city, and fewer acute life events than African American women living in the city, although neither of these di€erences were statistically signi®cant. However, African American women in the city of Detroit reported signi®cantly more acute life events than white women who lived outside the city. This result may have been in¯uenced by the speci®c acute life events measured for this study, some of which were related to race regardless of area of residence (e.g. likelihood of being a victim of a physical attack), while others were associated with both race and area of residence (e.g. whether someone close had died in the past year). Further research to disentangle those acute stressors that are most closely associated with neighborhood environment versus those that are organized around the social constructs of race or gender, regardless of neighborhood context, would help to further clarify the social pathways that contribute to racial disparities in women's health. There are several limitations to this study. One clear limitation is that the data are cross-sectional. The results cannot con®rm, for example, that discrimination or acute life events lead to poorer health outcomes. However, these results are consistent with a large body of research, both cross-sectional and longitudinal, indicating the cumulative e€ects of life stress on health status (Heaney & Israel, 1997; House et al., 1988; Israel & Schurman, 1990; Thoits, 1995). Though these cross-sectional analyses are important in their own right, longitudinal studies will be necessary to establish causation. Relatively small sample sizes for some of the comparison groups, for example, the small number of African American women who lived outside the city limits

1650

A. Schulz et al. / Social Science & Medicine 51 (2000) 1639±1653

of Detroit (n = 64), may have contributed to a lack of statistical signi®cance for indicators within this group. Even more problematic was the very small number of white women sampled within the Detroit city limits (n = 15), which we have already noted precluded comparisons of this group which would have further explicated the in¯uence of residential and racial di€erences as they contribute to di€erentials in health status. Furthermore, the study used two data sets with slightly di€erent sampling strategies and eligibility criteria. Most importantly, the focus in the East Side Village Health Worker survey on women with responsibility for the care of children resulted in a relatively younger study sample (mean age=38.9) than that of the Detroit Area Study (mean age=46.2). Age was controlled in all of the regression analyses included in the paper, and the ®nal series of regression analyses included only respondents from the DAS sample. These strategies should minimize potential bias introduced by these di€erences in the age of the study samples. Our measure of acute life events was limited to ®ve items that appeared in both data sets. As already noted, di€erences in the distribution of these acute life events suggest that racial di€erentials in women's health status may occur through multiple pathways. Some of these disproportionately a€ect African American women regardless of where they live, and some are more likely to a€ect African American women because of their disproportionate representation in urban communities characterized by economic marginalization. Measures of acute life events that contained additional items could extend our understanding of the variety of ways that stressors are patterned by race and area of residence, and the relative contributions of these di€erential distributions to racial di€erences in women's health status. It is dicult to disentangle the e€ects of place of residence on the individuals who reside there from individual or household factors that may contribute to residence in di€erent areas. The categories of residence used in these analyses were relatively broad, given di€erences in residential segregation and socioeconomic status that exist within both the city of Detroit and in the outlying metropolitan area. Further analyses using data with large enough samples to disentangle the e€ects of race and residence within smaller geographic areas (e.g. between African American and white respondents living in the same neighborhoods) could help to assess the e€ects of race and place more thoroughly. In addition, the data used here were gathered in the Detroit metropolitan area and care should be taken in generalizing to other metropolitan areas where economic patterns as well as patterns of residential segregation may di€er. Finally, these results, which focus on stressors and

their relationship to general health status, present only a part of a larger picture. Not all communities with high concentrations of poverty, and not all individuals within such communities, experience negative health e€ects to the same extent (Geronimus et al., 1996; Israel, James, Schulz, Parker & Farquhar, 2000; Parker et al., 2000). The stress process model suggests that protective factors not included in these analyses, such as social support and social integration, may play a key role in modifying the negative relationships shown here between exposure to chronic and acute stressors and health outcomes (Israel & Schurman, 1990; Israel et al., 2000; Williams, Spencer & Jackson, 1998). Examining factors such as household composition, the availability of social support, and community level factors, would contribute to an understanding of the ways that these indicators may in¯uence the health of women within various contexts, above and beyond the e€ects of life stressors. Concluding comments Despite these limitations, these results contribute to a growing body of evidence that there are multiple pathways through which discrimination in¯uences health. African American women in this study had fewer economic resources at their disposal (both overall, and at any given educational level), reported more frequent experiences with unfair treatment and experienced a greater number of acute life events in the past year than white women. Experiences of unfair treatment and acute life events both made independent and signi®cant contributions to, although they did not completely account for, racial di€erences in health status. Household income at a given level of education, and the likelihood of experiencing some acute life events were associated with residence in the city, and had a greater impact on African American women because they were more likely to live in the city. In contrast, experiences of unfair treatment did not vary signi®cantly for African American women by area of residence: that is, regardless of where they lived, African American women were more likely than white women living outside the city to experience unfair treatment. Whether through subjectively experienced discrimination or through structural conditions that contribute to increased risks of certain acute life events, the cumulative e€ects of these di€erential exposures to stressors contributes to longstanding racial di€erences in women's health status. Detailing the speci®c stressors experienced by African American and white women, and their relationship to race and to aspects of the environment or context, helps to enhance our understanding of the social determinants of longstanding racial disparities in health.

A. Schulz et al. / Social Science & Medicine 51 (2000) 1639±1653

These results suggest that what have been termed fundamental causes of health or disease (Link & Phelan, 1995) include not only inequalities in socioeconomic status at the household level, but also individual and institutional forms of discrimination that operate through multiple pathways, including di€erential exposure to multiple stressors. The results reported here suggest that African American women are more likely than white women to experience the e€ects of unfair treatment, lower household incomes, and acute life events and that the cumulative e€ects of these di€erentials result in poorer health at younger ages. These results suggest that e€orts to address underlying social determinants of health are an important aspect of addressing persistent racial disparities in health. Approaches that address these underlying stressors imply attention to the health-related implications of social policies that are rarely discussed in these terms. Examples of such policies include minimum wage, child care for working parents, equal education and equal wages for comparable work programs (Marmor, Morris & Evans, 1994; Thoits, 1995). Policies and practices that actively work to counter both institutional and interpersonal forms of discrimination, whether that discrimination is enacted on the basis of racial or gender categories, are linked to health through these pathways. In addition, policies that address the concentration of stressors in localized areas, for example, economic and social stressors, combined with physical environmental hazards found in some urban communities, are important health-related policies when considered within the framework of the stress process model (Corin, 1994; Evans, Barer & Marmor, 1994). Although this article has focused on relationships between stressors linked to the social categories of race, and has focused on their implications for racial disparities in women's health, many such policies are likely to in¯uence community health more broadly.

Acknowledgements The authors would like to acknowledge the Butzel Family Center, Detroit Health Department, Kettering Butzel Health Initiative, Mack Alive!, Warren Conner Development Coalition, Henry Ford Health System and Friends of Parkside for their assistance with the implementation of the East Side Detroit survey. We thank Lora Lempert, Debora Paterniti, Pat Smith and Deb Street for comments on earlier versions of this manuscript. In addition, we thank Sue Andersen for her contribution to the development of this manuscript. The research reported here was supported in part by a grant from the Centers for Disease Control

1651

and Prevention, and by grant MH-57425 from the National Institute of Mental Health. References Aneshensel, C. A. (1992). Social stress: theory and research. Annual Review of Sociology, 18, 15±38. Auerbach, J. D., & Figert, A. E. (1995). Women's health research: public policy and sociology. Journal of Health and Social Behavior, (extra issue), 115±131. Avison, W. R., & Gotlib, I. H. (1994). Introduction and overview. In W. R. Avison, & I. H. Gotlib, Stress and mental health (pp. 3±14). New York: Plenum Press. Barnes, A., & Ephross, P. H. (1994). The impact of hate violence on victims: emotional and behavioral responses to attacks. Social Work, 39, 247±251. Cohen, S., Kessler, R. C., & Gordon, L. U. (1997). Strategies for measuring stress in studies of psychiatric and physical disorders. In S. Cohen, R. C. Kessler, & L. U. Gordon, Measuring stress: a guide for health and social scientists (pp. 3±26). Oxford: Oxford University Press. Collins, C. (1997). The relationship between segregation and mortality for Blacks and Whites. Paper presented at the Annual Meeting of the American Sociological Association, August 9±13, 1997, Toronto, Ontario. Collins, J. W. Jr, & David, R. J. (1990). The di€erential e€ect of traditional risk factors in infant birthweight among black and white women in Chicago. American Journal of Public Health, 80, 679±681. Corin, E. (1994). The social and cultural matrix of health and disease. In R. G. Evans, M. L. Barer, & T. L. Marmor, Why are some people healthy and others not?: The determinants of health of populations (pp. 93±132). New York: Aldine de Gruyter. Dion, K. L. (1975). Women's reactions to discrimination from members of the same or opposite sex. Journal of Research in Personality, 9, 294±306. Dion, K. L., Dion, K. K., & Pak, A. W. (1992). Personalitybased hardiness as a bu€er for discrimination-related stress in members of Toronto's Chinese community. Canadian Journal of Behavioural Science, 24(4), 517±536. Duleep, H. O. (1995). Mortality and income inequality. Social Security Bulletin, 58, 34±50. Duncan, G. J., Brooks-Gunn, J., & Klebanov, P. K. (1994). Economic deprivation and early childhood development. Child Development, 65, 296±318. Edin, K., & Lein, L. (1996). Making ends meet: how single others survive welfare and low-wage work. New York: Russell Sage Foundation. Essed, P. (1990). Everyday racism: reports from women of two cultures. Claremont, CA: Hunter House. Evans, R. G., Barer, M. L., & Marmor, T. L. (1994). Why are some people healthy and others not?: the determinants of health of populations. New York: Aldine de Gruyter. Feagin, J. R. (1991). The continuing signi®cance of race: Anti-black discrimination in public places. American Sociological Review, 56, 101±116. Feagin, J. R., & Sikes, M. (1994). Living with racism: the black middle class experience. Boston, MA: Beacon Books. Freudenberg, N. (1998). Community-based health education

1652

A. Schulz et al. / Social Science & Medicine 51 (2000) 1639±1653

for urban populations: an overview. Health Education & Behavior, 25(1), 11±23. Fullilove, M. T. (1998). Comment: abandoning ``race'' as a variable in public health research Ð an idea whose time has come. American Journal of Public Health, 88(9), 1297± 1298. Fullilove, M. T., Green, L., & Fullilove, R. E. I. I. I. (1999). Building momentum: An ethnographic study of inner-city redevelopment. American Journal of Public Health, 89(6), 840±844. Geronimus, A. T. (1992). The weathering hypothesis and the health of African-American women and infants: evidence and speculations. Ethnicity and Disease, 2, 207±221. Geronimus, A. T., Bound, J., Waidmann, T. A., Hillemeier, M., & Burns, P. B. (1996). Excess mortality among blacks and whites in the United States. The New England Journal of Medicine, 335, 1552±1558. Haan, M., Kaplan, G. A., & Camacho, T. (1987). Poverty and health: prospective evidence from the Alameda County study. American Journal of Epidemiology, 125(6), 989±998. Heaney, C. A., & Israel, B. A. (1997). Social networks and social support. In K. Glanz, F. M. Lewis, & B. K. Rimer, Health behavior and health education (2nd ed.) (pp. 179± 203). San Francisco, CA: Jossey-Bass. House, J. S. (1981). Work stress and social support. Reading, MA: Addison±Wesley. House, J. S., Landis, K., & Umberson, D. (1988). Social relationships and health. Science, 241, 540±545. Hummer, R. A. (1993). Racial di€erences in infant mortality in the U.S.: an examination of social and health determinants. Social Forces, 72(2), 529±554. Idler, E. L., & Benyamini, Y. (1997). Self rated health and mortality: a review of twenty-seven community studies. Journal of Health and Social Behavior, 38, 21±37. Israel, B. A., House, J. S., Schurman, S. J., Heaney, C. A., & Mero, R. P. (1989). The relation of personal resources, participation, in¯uence, interpersonal relationships and coping strategies to occupational stress, job strains, and health: a multivariate analysis. Work and Stress, 3(2), 163± 194. Israel, B. A., James, S. A., Schulz, A. J., Parker, E. A., Farquhar, S. (2000). The relationship between social support, stress and health among women on Detroit's east side. In preparation. Israel, B. A., & Schurman, S. J. (1990). Social support, control and the stress process. In K. Glanz, F. M. Lewis, & B. K. Rimer, Health Behavior and Health Education (pp. 187±215). San Francisco, CA: Jossey±Bass. James, S. A. (1993a). Forward: racial di€erences in preterm delivery. American Journal of Preventive Medicine, 9(Suppl. 6), v±vi. James, S. A. (1993b). Racial and ethnic di€erences in infant mortality and low birth weight: A psychosocial critique. Annals of Epidemiology, 3(2), 130±136. James, S. A., Schulz, A. J., vanOlphen, J. (1999). Social capital, poverty and community health: an exploration of linkages. Paper presented at the Conference on Social Capital, sponsored by the Ford Foundation, New York, NY, 13 March, 1999. Jaynes, G. D., & Williams, R. M. (1989). A common destiny:

blacks and American society. Washington, DC: National Academy Press. Kaplan, G. A., Pamuk, E. A., Lynch, J. W., Cohen, R. D., & Balfour, J. L. (1996). Inequality in income and mortality in the United States: Analysis of mortality and potential pathways. British Medical Journal, 312, 999±1003. Katz, D., & Kahn, R. (1978). The social psychology of organizations (2nd ed.). New York: Wiley. Kennely, I. (1999). ``That single mother element'': how white employers typify black women. Gender and Society, 13, 168±192. Kessler, R. C., Mickelson, K. D., & Williams, D. R. (1999). The prevalence, distribution and mental health correlates of perceived discrimination in the United States. Journal of Health and Social Behavior, 40, 208±230. Kessler, R. C., & Neighbors, H. W. (1986). A new perspective on the relationships among race, social class, and psychological distress. Journal of Health and Social Behavior, 27, 107±115. Krieger, N. N. (1990). Racial and gender discrimination: risk factors for high blood pressure? Social Science & Medicine, 30, 1273±1281. Krieger, N. N. (1999). Embodying inequality: a review of concepts, measures and methods for studying health consequences of discrimination. International Journal of Health Services, 29(2), 295±352. Krieger, N., & Fee, E. (1994). Social class: the missing link in U.S. health data. Journal of Health Services, 24, 25±44. Krieger, N., Rowley, D. L., Herman, A., Avery, B., & Phillips, M. (1993). Racism, sexism and social class: implications for studies of health, disease and well-being. American Journal of Preventive Medicine, 9(Suppl. 6), 82± 122. Krieger, N. N., & Sidney, S. (1996). Racial discrimination and blood pressure: the CARDIA study of young black and white adults. American Journal of Public Health, 86, 1370±1378. Krieger, N. N., Williams, D. L., & Moss, N. E. (1997). Measuring social class in US public health research. Annual Review of Public Health, 18, 341±378. LaVeist, T. A. (1996). Why we should continue to study race. . .but do a better job: an essay on race, racism and health. Ethnicity and Disease, 6, 21±29. Lillie-Blanton, M., Parsons, P. E., Gayle, H., & Dievler, A. (1996). Racial di€erences in health: not just black and white, but shades of gray. Annual Review of Public Health, 17, 411±448. Link, B. G., & Phelan, J. (1995). Social conditions as fundamental causes of disease. Journal of Health and Social Behavior, (extra issue), 80±94. Lynch, J. W., Kaplan, G. A., & Cohen, R. D. (1994). Childhood and adult socioeconomic status as predictors of mortality in Finland. Lancet, 343, 524±527 (North American edition). Marmor, T. R., Morris, L., & Evans, R. G. (1994). The determinants of a population's health: what can be done to improve a democratic nation's health status? In R. G. Evans, M. L. Barer, & T. L. Marmor, Why are some people healthy and others not?: the determinants of health of populations (pp. 217±230). New York: Aldine de Gruyter. Massey, D., & Denton, N. (1993). American apartheid: segre-

A. Schulz et al. / Social Science & Medicine 51 (2000) 1639±1653 gation and the making of the underclass. Boston, MA: Harvard University Press. Massey, D., Fischer, M. (1998). How segregation concentrates poverty. Paper presented at the American Sociological Association Meetings, August 1998, San Francisco, CA. McDonough, P., Williams, D. R., House, J. S., & Duncan, G. J. (1999). Gender and the socioeconomic gradient in mortality. Journal of Health and Social Behavior, 40(1), 17±31. Neter, J., Wasserman, W., Kutner, M. H. (1985). Applied linear statistical models, 2nd ed. Edition. Homewood, IL: Irwin. Noh, S., Belser, M., Kaspar, V., Hou, F., & Rummens, J. (1999). Perceived racial discrimination, depression, and coping: a study of southeast Asian refugees in Canada. Journal of Health and Social Behavior, 40, 193±207. O'Campo, P., Xue, S., Wang, M.-C., & O'Brien Caughy, M. (1997). Neighborhood risk factors for low birth weight in Baltimore: a multi-level analysis. American Journal of Public Health, 87(7), 1113±1118. Pappas, G., Queen, S., Hadden, W., & Fisher, G. (1993). The increasing disparity in mortality between socioeconomic groups in the United States, 1960 and 1986. New England Journal of Medicine, 329, 103±109. Parker, E. A., Eng, E. (1997). Conceptualizing community problem-solving capacity: results of a grounded theory study. Unpublished paper. Parker, E. A., Lichtenstein, R., Schulz, A. J., Steinman, K. J., Israel B. A., James, S. A. (2000). Disentangling community capacity. Under review. Parker, E. A., Schulz, A. J., Israel, B. A., & Hollis, R. (1998). Detroit's east side village health worker partnership: community-based lay health advisor intervention in an urban area. Health Education & Behavior, 25(1), 24±45. Pearlin, L. I. (1989). The sociological study of stress. Journal of Health and Social Behavior, 30, 241±256. Polednak, A. P. (1993). Poverty, residential segregation, and black/white mortality ratios in urban areas. Journal of Health Care for the Poor and Underserved, 4(4), 363±373. Ren, X. S., Amick, B. C., & Williams, D. R. (1999). Racial/ ethnic disparities in health: the interplay between discrimination and socioeconomic status. Ethnicity and Disease, 9(2), 151±165. Robert, S. A. (1998). Community-level socioeconomic status e€ects on adult health. Journal of Health and Social Behavior, 39(1), 18±37. Sarason, I. G., Pierce, G. R., & Sarason, B. R. (1994). General and speci®c perceptions of social support. In W. R. Avison, & I. H. Gotlib, Stress and mental health (pp. 153±178). New York: Plenum Press. Schulz, A. J., Israel, B. A., Becker, A. B., Parker, E. A., James, S. A., Schork, M. A. (1999). Stressors, protective factors and health: Results from a survey of African American women living on Detroit's east side. Paper presented at the American Sociological Association meetings, August 1998.

1653

Schulz, A. J., Williams, D. R., Israel, B. A., Becker, A. B., Parker, E. A., James, S. A., Jackson, J. (2000). Unfair treatment, neighborhood e€ects, and health in the Detroit metropolitan area. Journal of Health and Social Behaviour, Forthcoming. Sheak, R. (1988). Poverty estimates: political implications and other issues. Sociological Spectrum, 8, 277±294. Sugrue, T. J. (1996). The origins of the urban crisis; race and inequality in postwar Detroit. Princeton, NJ: Princeton University Press. Thoits, P. (1995). Stress, coping and social support processes: where are we? what next? Journal of Health and Social Behavior, (extra issue), 53±79. Thomas, J. M. (1997). Redevelopment and race: planning a ®ner city in postwar Detroit. Baltimore, MD: Johns Hopkins University Press. Turner, R. J., & Roszell, P. (1994). Psychosocial resources and the stress process. In W. R. Avison, & I. H. Gotlib, Stress and mental health (pp. 179±210). New York: Plenum Press. Turner, R. J., Wheaton, B., & Lloyd, D. A. (1995). The epidemiology of social stress. American Sociological Review, 60(1), 104±125. Turner, S. (1997). Barriers to a better break: employer discrimination and spatial mismatch in metropolitan Detroit. Journal of Urban A€airs, 19(2), 123±141. Wheaton, B. (1994). Sampling the stress universe. In W. R. Avison, & I. H. Gotlib, Stress and mental health (pp. 78± 114). New York: Plenum Press. Williams, D. R. (1997). Race and health: basic questions, emerging directions. Annals of Epidemiology, 7, 322±333. Williams, D. R., & Collins, C. (1995). U.S. socioeconomic and racial di€erences in health: patterns and explanations. Annual Review of Sociology, 21, 349±386. Williams, D. R., Spencer, M. S., & Jackson, J. S. (1998). Race-related stress and physical health: is group identity a vulnerability factor or a resource? In R. J. Contrada, & R. D. Ashmore, Self, social identity, and physical health: interdisciplinary exploration (pp. 71±100). New York: Oxford University Press. Williams, D. R., Yu, Y., Jackson, J. (1997a). The costs of racism: discrimination, race and health. Presented at the Joint Meeting of the Public Health Conference on Records and Statistics and Data User's Conference, Washington DC, 28±31 July. Williams, D. R., Yu, Y., Jackson, J., & Anderson, N. B. (1997b). Racial di€erences in physical and mental health: socioeconomic status, stress, and discrimination. Journal of Health Psychology, 2, 335±351. Wilson, W. J. (1987). The truly disadvantaged: the inner city, the underclass and public policy. Chicago, IL: University Press. Zax, J. S., & Kain, J. F. (1996). Moving to the suburbs: do relocating companies leave the black employees behind? Journal of Labor Economics, 14(3), 472±504.