Spatial dependence exists when the variation between observations is dependent on spatial location. In the present study, geostatistical methods were used to examine spatial dependence in adolescents' perceptions of their neighborhoods: whether adolescents living in close proximity perceived their neighborhoods more similarly than adolescents living further apart. Participants included 343 adolescents (53 % male; 91 % African American) enrolled in sixth grade, residing in Baltimore City, with geocoded home addresses. These addresses were used in combination with a measure of perceived neighborhood disorder and disadvantage to compute variogram plots, a geostatistical technique used to evaluate spatial dependence. Results indicated that perceptions of youth residing in close geographic proximity were more similar than perceptions of youth residing further apart. Adjusting for census-level characteristics, but not individual and family characteristics, resulted in decreased variation among participants but did not eliminate the spatial dependence. Findings highlight the utility of geostatistical methods for social science research. (C) 2004 Wiley Periodicals, Inc.