Table 2.
Estimated sampling error rates for 18 combinations of data used, assumed reading effects β, and assumed dependence between specimen effects.
| Study | β | Assumed dependence, σ1 with σ2 | Sampling error estimate, % (95% confidence interval) |
−2l* | |||
|---|---|---|---|---|---|---|---|
| ŝ10 | ŝ21 | ŝ32 | ŝ43 | ||||
| S1 | −1 | None | 0 (0, 100) | 3 (0, 31) | 45 (21, 72) | 14 (2, 54) | 28.01 |
| Estimated | 0 (0, 100) | 3 (0, 28) | 46 (22, 72) | 14 (2, 53) | 27.82 | ||
| Full | 0 (0, 100) | 2 (0, 41) | 43 (19, 69) | 13 (2, 54) | 29.99 | ||
| 0 | None | 0 (0, 100) | 3 (0, 30) | 42 (17, 72) | 12 (1, 58) | 28.40 | |
| Estimated | 0 (0, 100) | 3 (0, 30) | 43 (18, 72) | 12 (1, 58) | 28.12 | ||
| Full | 0 (0, 100) | 2 (0, 38) | 37 (14, 67) | 10 (1, 59) | 30.48 | ||
| +1 | None | 0 (0, 100) | 3 (0, 26) | 17 (1, 87) | 6 (0, 95) | 28.50 | |
| Estimated | 0 (0, 100) | 3 (0, 24) | 26 (3, 80) | 7 (0, 88) | 28.11 | ||
| Full | 0 (0, 100) | 1 (0, 43) | 6 (0, 99) | 5 (0, 97) | 30.00 | ||
| S2 | −1 | None | 18 (7, 41) | 0 (0, 100) | 46 (14, 81) | 25 (15, 40) | 52.49 |
| Estimated | 18 (7, 41) | 0 (0, 100) | 46 (14, 81) | 25 (15, 40) | 52.49 | ||
| Full | 17 (7, 39) | 0 (0, 27) | 42 (16, 73) | 25 (14, 39) | 53.83 | ||
| 0 | None | 16 (5, 43) | 0 (0, 100) | 45 (13, 82) | 26 (15, 40) | 51.13 | |
| Estimated | 16 (5, 43) | 0 (0, 100) | 45 (13, 82) | 26 (15, 40) | 51.13 | ||
| Full | 15 (5, 39) | 0 (0, 100) | 39 (13, 72) | 25 (15, 39) | 52.93 | ||
| +1 | None | 15 (3, 47) | 0 (0, 100) | 31 (5, 80) | 28 (16, 44) | 51.44 | |
| Estimated | 16 (4, 49) | 0 (0, 100) | 34 (4, 87) | 28 (16, 43) | 51.41 | ||
| Full | 0 (0, 29) | 9 (3, 24) | 15 (1, 80) | 26 (15, 41) | 51.96 | ||
*
−2 times the maximized log likelihood.