Open ellisp opened 7 years ago
See http://stoppress.co.nz/features/newshub-polling re changing methodologies. Also even without that perhaps variance in the estimates of house effects should be added to the variance of the point estimates at the final stage.
This is somewhat resolved if we use the state space model, which is more rigorous and coherent in treating the whole thing at once.
See http://stoppress.co.nz/features/newshub-polling re changing methodologies. Also even without that perhaps variance in the estimates of house effects should be added to the variance of the point estimates at the final stage.