I am conducting community/school research to exploring the health impact of environmental exposures (e.g., air pollution). The exposure is predicted based on the address of the community/school centroid using spatio-temporal model. Therefore, there is no difference in exposure among participants in the same school/community based on the prediction results.

However, considering the clustered data structure (the health data is aggregated at school/community level), we incorporated the ID of community/school as random effects in the regression model.

health outcome ~ exposure + confounder + (1 | ID)

Since the exposures vary only between community/school, is there a problem in fitting the model in this way? Is there a tendency to producing a false-negative estimate for exposures?

Besides, in a PANEL STUDY, when the exposure is based on the air monitoring station, the individual exposures do not have differences on the same day in the same area, and there are only differences at different times. Considering the independence of data for the same person, we model subjects(ID) as random effects. Is it suitable to do so? We are concerned there might be a false-negative estimate for exposure)


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.