Abstract
Conditionally specified logistic regression (CSLR) models p binary response variables. It is shown that marginal probabilities can be derived for a CSLR model. We also extend the CSLR model by allowing third order interactions. We apply two versions of CSLR to simulated data and a set of real data, and compare the results to those from other modeling methods.
Data Availability Statement
Source data cannot be shared publicly due to tribal regulations on data sharing set out by the Navajo Nation Human Research Review Board. Investigators interested in using the data included in this analysis may contact the corresponding author (Curtis Miller, [email protected]) or the contact listed on the Navajo Nation Human Research Review Board website (www.nnhrrb.navajo-nsn.gov) to submit an application for access to the data if the request is consistent with the IRB-approved protocol.
Acknowledgments
Funding for the DiNEH Project survey and its analysis was provided by National Institute of Environmental Health Sciences (NIEHS) grants R01 ES014565, R25 ES013208, and P30 ES-012072, NIH/NIEHS P42-ES025589 for the UNM METALS Superfund Center, and NIH UG3 OD023344 for the ECHO program.