Abstract
Underreporting is a persistent limitation in research on environmental risk factors for dog cancer, impeding potential comparative investigations with human cancer. To address this challenge, we propose a regional modeling approach accounting for different reporting practices across the study area. In doing this, we demonstrate the need for new modeling strategies to improve statistical performance through more systematic assessments of spatial non-stationarity of statistical associations that can be linked to underreporting.
Boo, Gianluca