556 Cultural Aspects of Health Psychology
research that explicitly focuses on the potential sources of
SES differences, the overwhelming majority of studies
designed to delineate the determinants of health tend to sta-
tistically control for the effects associated with SES. From a
clinical perspective, the observation that SES groups differ
with respect to a number of health indices, although informa-
tive, does not lead logically to the more proximal variables
that are related to biobehavioral processes, which may be
more amenable to prevention and treatment strategies. Sec-
ond, the assessment of SES has historically been rather crude.
The most frequently used proxies for SES include income,
education, and occupation, with income showing the
strongest relationship to health (Stronks, van de Mheen, Van
Den Bos, & Mackenbach, 1997). It is important to note that
within SES groupings (whether assessed by income, educa-
tion, or occupation), the major U.S. ethnic groups are differ-
entially distributed, with African Americans and Hispanics
being disproportionately represented in the lowest SES
groups, and Asian or Paci“c Islanders being disproportion-
ately represented in the highest SES groups (NCHS, 1998;
D. Williams, 1996). Third, in most empirical investigations,
SES is measured cross-sectionally. This methodological
limitation is particularly noteworthy, given that an emerg-
ing body of literature suggests that changes in socioeco-
nomic status (Hart, Smith, & Blane, 1998; Lynch, Kaplan, &
Shema, 1997; McDonough, Duncan, Williams, & House,
1997) and early life experiences (D. Barker, 1995; Peck,
1994; Rahkonen, Lahelma, & Huuhka, 1997) are predictive
of health outcomes.
SES and Health Status
The medical expenditures associated with negative health
outcomes are exceedingly high in the United States. For ex-
ample, the estimated medical costs associated with treating
onlythree of the major chronic diseases (heart disease, lung
cancer, and diabetes mellitus) were $131 billion in 1995
(NCHS, 1998). Research delineating factors related to nega-
tive health outcomes has the potential of better informing
prevention and intervention efforts, and as a result, reduces
health care costs. Socioeconomic status is one such factor
that has been explored extensively by research scientists.
The observation that individuals with fewer social and
economic resources generally have more negative health out-
comes than their more •resourcefulŽ counterparts is reported
to be at least 2,000 years old (Lloyd, 1983; Sigerist, 1956).
With the exception of some cancers (Gold, 1995; Kelsey &
Bernstein, 1996) and heart disease mortality during the
“rst half of the twentieth century (Marmot, Shipley, &
Rose, 1984), more contemporary studies continue to
document inverse relationships between SES and morbidity
and mortality. This SES-health gradient has been observed
across ethnic, gender, and age groups for all-cause and disease-
speci“c mortality and an array of chronic diseases, communi-
cable diseases, and injuries (Breen & Figueroa, 1996;
Cantwell, McKenna, McCray, & Onorato, 1998; Gissler,
Rahkonen, Jarvelin, & Hemminki, 1998; JNC, 1993; Litonjua,
Carey, Weiss, & Gold, 1999; Liu, Wang, Waterbor, Weiss, &
Soong, 1998; NCHS, 1998; Ogle, Swanson, Woods, &
Azzouz, 2000; Robert & House, 1996). These data indicate
that persons of lower SES are disproportionately burdened by
negative health outcomes.Interactions of Ethnicity, SES, and HealthBecause African Americans and Hispanics have lower me-
dian household incomes, educational attainments, and occu-
pational positions, as well as poorer outcomes for a number
of medical ailments (NCHS, 1998; U.S. Department of
Health and Human Services, 1985), it was once believed that
if SES were controlled (via strati“cation or statistically), the
between-ethnic group health disparities would be eliminated.
That is, if poorer health is secondary to a relative lack of re-
sources for nutritional needs, access to, and use of, quality
health care and adequate housing (controlling for SES)
should •even the playing “eld,Ž thereby eliminating
between-group disparities. Although intuitively appealing, an
emerging body of literature suggests that adjustments for
SES may substantially reduce or eliminate these disparities
for some (Cantwell et al., 1998; Litonjua et al., 1999) but not
all health outcomes (Kington & Smith, 1997; Lillie-Blanton
& Laveist, 1996; NCHS, 1998; Schoenbaum & Waidmann,
1997; Schoendorf, Hogue, Kleinman, & Rowley, 1992;
D. Williams, 1996).
A number of hypotheses have been presented to explain the
persistence of these between-group disparities (N. Anderson
& Armstead, 1995; Kington & Nickens, 1999; D. Williams,
1996). For example, R. Clark, Anderson, Clark, and Williams
(1999) proposed two reasons to help explain “ndings that the
prevalence of hypertension and all-cause mortality are higher
for African Americans than European Americans at compara-
ble educational levels (Pappas et al., 1993). First, within-
SES group •protectionŽ may not be comparable across ethnic
groups (N. Anderson & Armstead, 1995; D. Williams &
Collins, 1995). As such, attempts to compare African
Americans and European Americans at any given educational
level, for instance, would not take into account the observa-
tion that African Americans earn signi“cantly less than their
European American counterparts at every level of education
attainment (NCHS, 1998). Second, if African Americans