Table 4.
Bland and Altman statistics for sanitary risk observations made by four observers, relative to the ‘gold standard’ Observer A (JOO) (***Indicates significance at alpha = 0.001)
| Versus Observer B | Versus Observer C | Versus Observer D | Versus Observer E | Versus Observer F | |
|---|---|---|---|---|---|
| No. of water source visits | 139 | 146 | 53 | 147 | 152 |
| Lin’s concordance correlation coefficient (95% confidence intervals) | 0.320 (0.22–0.42) | 0.515 (0.40–0.63) | 0.372 (0.20–0.55) | 0.435 (0.33–0.54) | 0.230 (0.15–0.31) |
| Pearson’s r | 0.496 | 0.553 | 0.504 | 0.570 | 0.441 |
| Bias correction factor | 0.645 | 0.931 | 0.738 | 0.764 | 0.521 |
| Average difference in % risk scores (95% limits of agreement) | 15.4% (− 14.0% to 44.8%) | 5.2% (− 24.4 to 34.8) | 12.9% (− 20.5% to 46.4%) | 11.9% (− 18.2% to 42.1%) | 21.7% (− 15.0% to 58.5%) |
| Correlation between difference in risk scores and mean | − 0.14 | − 0.22 | − 0.32 | − 0.29 | − 0.39 |
| Bradley-Blackwood F test | 75.5 (***) | 12.9 (***) | 19.4 (***) | 54.4 (***) | 132.8 (***) |