On Ecological Fallacy and Assessment Errors Stemming From Misguided Variable Selection: Investigating the Effect of Data Aggregation on the Outcome of Epidemiological Study

In behavioral studies, ecological fallacy is a wrong assumption about an individual based on aggregate data for a group. In the present study, the validity of this assumption was tested using both individual estimates of exposure to air pollution and aggregate air pollution data estimated for 1,492 schoolchildren living in the in vicinity of a major coal-fired power station in the Hadera sub-district of Israel. In 1996 and 1999, the children underwent subsequent pulmonary function (PF) tests, and their parents completed a detailed questionnaire on their health status, and housing conditions. The association between children's PF development and their long-term exposure to air pollution was then investigated in two phases. During the first phase, the average rates of PF change observed in small statistical areas in which the children reside were compared with average levels of air pollution detected in these areas. During the second phase of the analysis, an individual pollution estimate was calculated for each child covered by the survey, using a "spatial join" tool in ArcGIS. While the analysis of aggregate data showed no significant differences in the PF development among the schoolchildren surveyed, the comparison of individual pollution estimates with the results of PF tests detected a significant negative association between changes in PF results and the estimated level of air pollution. As argued, these differences are attributed to the fact that average exposure levels are likely to cause a misclassification bias of individual exposure, as further demonstrated in the study using pattern detection techniques of spatial analysis (local Moran's I and Gettis-Ord statistic). The implications of the results of the analysis for geographical and epidemiological studies are discussed, and recommendations for public health policy are formulated.