Table 2 Coverage probabilities for nominal (90%/95%) intervals based on 10,000 simulations, based on three population settings of (p, q, w). Setting 1: p = 0.96, q = 0.06, w = 0.04. Setting 2: p = 0.9, q = 0.15, w = 0.04. Setting 3: p = 0.8, q = 0.2, w = 0.5.
From: Interval estimation of genetic susceptibility for retrospective case-control studies
Sample size | 10 | 50 | 100 | 500 |
|---|---|---|---|---|
(p, q, w) settings | Â | Â | Â | Â |
| Â | Asymptotic Frequentist | |||
1 | 0.888/0.951 | 0.919/0.962 | 0.916/0.960 | 0.896/0.948 |
2 | 0.942/0.883 | 0.906/0.962 | 0.909/0.954 | 0.896/0.953 |
3 | 0.909/0.954 | 0.908/0.954 | 0.900/0.947 | 0.899/0.951 |
| Â | Asymptotic Bayesian | |||
1 | 0.888/0.883 | 0.918/0.920 | 0.907/0.951 | 0.901/0.948 |
2 | 0.866/0.920 | 0.900/0.950 | 0.904/0.946 | 0.894/0.951 |
3 | 0.879/0.907 | 0.901/0.950 | 0.895/0.944 | 0.899/0.948 |
| Â | Exact Bayesian | |||
1 | 0.911/0.979 | 0.890/0.922 | 0.894/0.954 | 0.897/0.946 |
2 | 0.945/0.967 | 0.900/0.950 | 0.908/0.950 | 0.896/0.951 |
3 | 0.865/0.954 | 0.900/0.948 | 0.896/0.946 | 0.895/0.950 |