The Combination Analysis Of Ecological Data And Samples With Covariate Data

Most ecological research is based on groups as the basic unit of data collection or analysis in order to study the relationship between exposure and disease.As we only know the number of the exposed and non-exposed and the number of sick and non-sick.But we don't know how many of the disease occurred in the exposed and non-exposed.This means that ecology can not be linked to exposure and disease in a particular individual.It would create ecolgical bias. Sebastien J.-P.A.Haneuse and Jonathan C.Wakefield proposed a hybrid design that reduced the bias and increased the efficiency in 2008.However,they did not consider ecological data containing some covariate.It can not eliminate the influence of the potential variable on the disease that result from the loss of covariate information in ecology study.This artical will analyze the relationship between the ecological data and samples with covariate data.We establish a hybrid model and develop the likelihood for the design.Then we make statistical inference for the model parameters.