Objective - The objective of this study was to estimate the association between certain characteristics of the residential environment and obesity, after adjusting for age, sex, and socioeconomic status among residents of the Paris metropolitan area. Methods - The analyses were conducted using data collected in 2005 by the SIRS (Health, Inequalities, and Social Disruptions) cohort from a random sample of 3,023 individuals representative of the French-speaking adult population of Paris and its inner suburbs. Obesity is defined as a BMI of 30 or higher, based on self-reported data. Characteristics of the residential environment were measured at the IRIS level or within a 500-meter radius, depending on the availability of external sources. The analyses utilized a geographic information system and multilevel regression models. Results - After adjusting for age, sex, education level, and income level, there are strong and significant associations between the risk of obesity and the average household income level in the residential Iris, the proportion of residents with a higher education level, the average distance to the nearest food retail stores, the proportion of fast-food establishments among restaurants, and the number of neighborhood shops and services. Discussion-Conclusion - Such an approach to "contextual epidemiology" appears highly informative for diseases such as obesity—where multilevel models allow for comparing the "effect" of individual characteristics with that of residential context characteristics—at a time when regional health policy discussions and actions are developing. (R.A.)