The chronic disease burden associated with physical inactivity may disproportionately affect socially disadvantaged populations. Racial and ethnic minority populations are consistently shown to be less physically active during leisure time compared with whites (1). In 2000, the prevalence of having no physical activity during leisure time was 54% among Hispanics, 52% among blacks, and 35% among whites (1). Numerous population surveys have similarly reported a positive association between leisure-time physical activity and measures of social class (8,17,22). Approximately 72% of adults with less than a ninth-grade education do not regularly participate in leisure-time physical activity, compared with 25% of college graduates (1). The methodological challenges associated with valid measurement of physical activity using self-report measures (14,16) may be magnified in studies involving lower-income populations. Few self-report measures are designed to assess activity accumulated through nonleisure sources (e.g., domestic, occupational, transportation), which may be more common and account for a larger proportion of total physical activity among lower-income populations. Individuals in lower-income populations are overrepresented in physically active occupations (i.e., construction, maintenance, housekeeping, factory work). Particularly in urban settings, individuals of lower income may have a higher likelihood of engaging in ambulatory activity for transportation purposes (e.g., walking to work, school, or a bus stop). This is a key concern, because it is largely unclear whether the recommended guidelines for daily physical activity (18) are met or exceeded by those in lower-income populations through these often unmeasured activity sources. For example, in two large, randomized intervention studies conducted among more than 4000 lower-income, multiethnic adults, our group recently found much higher, accelerometer-validated, rates of self-reported physical activity than would be expected based on the available literature (10,21). Results of these and other (5) studies highlight the need to assess total accumulated physical activity in these groups, using reliable and externally valid measurement strategies. A variety of motion sensor technologies (i.e., pedometers, accelerometers) have been developed to overcome the compromised validity associated with many self-report physical activity indexes (28). Pedometers capture the vertical accelerations associated with normal ambulation and demonstrate excellent concordance with accelerometer-derived physical activity (median correlation of 0.86), and can objectively detect gradations in walking behavior (4,11,28). Walking is the most frequently adopted type of regular physical activity (16,28), particularly among some racial and ethnic minority groups, (16) yet it is among the least reliably recalled activity types (16,28). Given their accuracy, low unit cost ($10–20), and ease of use, pending continued rigorous validation, pedometers hold great promise as measures of total physical activity among lower-income, multiethnic populations. Emerging data provide expected reference values for daily steps by a variety of sociodemographic characteristics (20,27,29). Higher daily steps are generally found among men (24,29), those of younger age (20,27,30), and the employed (7,30). Some evidence suggests that pedometer-determined daily steps are inversely associated with measures of socioeconomic position and are generally higher among whites (27). Few studies, however, have measured physical activity using pedometers in well-characterized, population-based samples that include individuals from lower-income and racial and ethnic minority groups. Given the limitations of the current literature, the primary aimof the present study was to discern patterns of pedometer-determined physical activity by a range of sociodemographic characteristics in a predominately racial and ethnic minority population of low-income housing residents.