assets would eliminate more of the behavioral differences is not known. Even adding such controls in one study, however, leaves untouched a large part of the black disadvantage, relative to whites, in cardiovascular disease indicators (Rooks et al., 2002).

Another source of behavioral differences may be cultural, or more specifically, degree of acculturation among immigrants and their descendants (Espino et al., 1991; Jasso et al., 2000). The more acculturated a person—whether Hispanic, Asian, or black—the more prone the person is to smoking and obesity, whether acculturation is determined from United States versus foreign birth, duration of residence in the United States, or language spoken at home (Singh and Siahpush, 2002; Winkleby and Cubbin, 2004). Various cultural beliefs, such as the presumed attractiveness of moderately overweight women, have also been proposed as important (Stevens et al., 1994).

An additional factor that may produce behavioral differences is the residential environment. Neighborhoods provide stimuli, such as outlets for alcohol or illegal drugs, or limit options for healthy nutrition or exercise. Black neighborhoods appear more likely to suffer from such institutional risk factors as the proliferation of liquor stores and insufficient supplies of prescription drugs (LaVeist and Wallace, 2000; Morrison et al., 2000). However, since blacks have lower rates of drinking and smoking than whites—an advantage that increases when socioeconomic status is controlled—the implications of such neighborhood disadvantages are unclear. Differences do indeed appear across neighborhoods in smoking, dietary practices, physical activity, and substance abuse (Morenoff and Lynch, 2004; Winkleby and Cubbin, 2004). Whether these are due mainly to socioeconomic deprivation in poorer neighborhoods or actually reflect some effect of the neighborhood environment is difficult to verify.

Does variation in these behavior risk factors account for some portion of racial and ethnic differences in health? This issue has been insufficiently studied, especially for comparisons other than black and white. Behavioral factors clearly do not explain all differences, and how much they do explain is unclear.

Looking at black-white differences in mortality, Otten et al. (1990) combined behavior risk factors (smoking, BMI, and alcohol intake) and some health outcomes closely linked to behavior (systolic blood pressure, cholesterol level, and diabetes). Among persons 35-54 years old, those factors combined explained 31 percent of the excess mortality for blacks relative to whites. Slightly more of the mortality differential was explained by income, while a larger proportion remained unexplained.