The 10-year follow-up examination in 1995-1996 to the population-based Coronary Artery Disease Risk Development in Young Adults Study was used to compare the strength with which socioeconomic indicators at the individual and area levels are related to smoking prevalence and to investigate contextual effects of area characteristics. When categories based on similar percentile cutoffs were compared, differences across area categories in the odds of smoking were smaller than differences across categories based on individual-level indicators. In Whites, there was evidence of a significant contextual effect of area characteristics on smoking: Living in the most disadvantaged area quartiles was associated with 50-110% higher odds of smoking, even after controlling for individual-level socioeconomic indicators. Clear contextual effects of area characteristics were not present in Blacks, but there was evidence that contextual effects may emerge at higher levels of individual-level socioeconomic position. Similar results were obtained for census tracts and block groups. Even in the presence of contextual effects, area measures may underestimate associations of individual-level variables with health outcomes. On the other hand, as illustrated by the presence of contextual effects, area- and individual-level measures are likely to tap into different constructs.
Available at: http://works.bepress.com/catarina_kiefe/105/