The study aimed to assess a 54-item instrument that measures perceptions of environmental contaminant exposure, focusing on five domains: home and hobby, school, community, occupation, and exposure history.
Interviews were conducted with child-bearing-age minority women in Nashville, and data were analyzed using Support Vector Machine (SVM) modeling and logistic regression to explore the relationship between environmental exposure and respondents' ZIP code.
The findings revealed that SVM modeling and traditional logistic regression yielded nearly identical rankings of important variables, marking the first evidence that SVM analysis can effectively analyze complex spatial relationships in environmental exposure questionnaires.