Examples; Oxford: Smooth Fit to Log-Odds Ratios
An example from OpenBUGS openbugs:2014:ex
and Breslow and Clayton breslow:1993:AIG
concerning the association between death from childhood cancer and maternal exposure to X-rays, for subjects partitioned into 120 age and birth-year strata.
Deaths are modelled as
where ri0 is the number of deaths among unexposed subjects in stratum i, ri1 is the number among exposed subjects, and yeari is the stratum-specific birth year (relative to 1954).
oxford.jl
Iterations = 2502:12500
Thinning interval = 2
Chains = 1,2
Samples per chain = 5000
Empirical Posterior Estimates:
Mean SD Naive SE MCSE ESS
beta2 0.005477119 0.0035675748 0.00003567575 0.00033192987 115.519023
beta1 -0.043336269 0.0161754258 0.00016175426 0.00133361554 147.112695
alpha 0.565784774 0.0630050896 0.00063005090 0.00468384860 180.944576
s2 0.026238992 0.0307989154 0.00030798915 0.00302056007 103.967091
Quantiles:
2.5% 25.0% 50.0% 75.0% 97.5%
beta2 -0.0010499046 0.0028489198 0.0056500394 0.0077473623 0.013630865
beta1 -0.0745152363 -0.0543180318 -0.0434425931 -0.0321216097 -0.009920787
alpha 0.4438257884 0.5238801187 0.5675039159 0.6051427125 0.695968063
s2 0.0007134423 0.0033352655 0.0146737037 0.0397132522 0.118202258