Although studies suggest that air pollution is linked to perinatal outcomes, the geographic characterization of exposure to pollution differs between the studies. We compared neighborhood-and county-level measures of air pollution exposure, while examining the association between particulate matter less than 2.5 mm in aerodynamic diameter (PM 2.5 ) and birth weight among full-term births in California in 2000. To reduce the effects of demographic variability, our analysis was limited to two populations of 8579 non-Hispanic white and 8114 Hispanic mothers who were married, between 20 and 30 years of age, completed at least a high school education, and gave birth for the first time. Measurements from the nearest monitor, and average and distance-weighted average of monitors within a 5-mile radius from each mother's residence (constituting neighborhood metrics) and the mean of monitors within each mother's county of residence were considered. PM 2.5 measurements, provided by the California Air Resources Board, were calculated to correspond to each mother's 9-month gestation period. Although metrics within the 5-mile radii and the county were highly correlated (r 2 ¼ 0.78), the county-level metric provided a stronger association between PM 2.5 and birth weight (b ¼ À4.04, 95% confidence interval ¼ À6.71, À1.37) than the metric for the average of all monitors within 5-miles (b ¼ À1.38, 95% confidence interval ¼ À3.36, 0.60) among non-Hispanic white mothers; similar results were observed among the Hispanic sample of mothers. Consequently, inferences from studies using different definitions of air pollution exposure may not be comparable.