Interval Estimation for Messy Observational Data
Gustafson, Paul ; Greenland, Sander
Statist. Sci., Tome 24 (2009) no. 1, p. 328-342 / Harvested from Project Euclid
We review some aspects of Bayesian and frequentist interval estimation, focusing first on their relative strengths and weaknesses when used in “clean” or “textbook” contexts. We then turn attention to observational-data situations which are “messy,” where modeling that acknowledges the limitations of study design and data collection leads to nonidentifiability. We argue, via a series of examples, that Bayesian interval estimation is an attractive way to proceed in this context even for frequentists, because it can be supplied with a diagnostic in the form of a calibration-sensitivity simulation analysis. We illustrate the basis for this approach in a series of theoretical considerations, simulations and an application to a study of silica exposure and lung cancer.
Publié le : 2009-08-15
Classification:  Bayesian analysis,  bias,  confounding,  epidemiology,  hierarchical prior,  identifiability,  interval coverage,  observational studies
@article{1270041259,
     author = {Gustafson, Paul and Greenland, Sander},
     title = {Interval Estimation for Messy Observational Data},
     journal = {Statist. Sci.},
     volume = {24},
     number = {1},
     year = {2009},
     pages = { 328-342},
     language = {en},
     url = {http://dml.mathdoc.fr/item/1270041259}
}
Gustafson, Paul; Greenland, Sander. Interval Estimation for Messy Observational Data. Statist. Sci., Tome 24 (2009) no. 1, pp.  328-342. http://gdmltest.u-ga.fr/item/1270041259/