Table 4.
Percentage of simulated data sets where SRR was estimated as 0 (% ρ=0 ) and the bias of the corresponding SCR estimates taken as the percentage in relation to the true SCR
| Population | Sample size | SCR = 0.0036 | SCR = 0.0108 | ||
|---|---|---|---|---|---|
| % ρ=0 | Bias (%) | % ρ=0 | Bias (%) | ||
| Africa | 250 | 28.6 | −25.7 | 16.0 | −26.1 |
| 500 | 21.9 | −25.3 | 8.4 | −26.6 | |
| 1,000 | 13.1 | −25.6 | 2.0 | −25.6 | |
| 2,500 | 3.7 | −25.2 | 0.1 | −26.4 | |
| 5,000 | 0.6 | −24.4 | 0.0 | N/A | |
| 10,000 | <0.1 | −22.4 | 0.0 | N/A | |
| Elsewhere | 250 | 27.0 | −29.6 | 14.6 | −30.5 |
| 500 | 20.1 | −29.4 | 6.1 | −30.2 | |
| 1,000 | 11.2 | −29.2 | 1.5 | −30.7 | |
| 2,500 | 3.0 | −28.6 | <0.1 | −32.2 | |
| 5,000 | 0.4 | −28.8 | 0.0 | N/A | |
| 10,000 | 0.0 | N/A | 0.0 | N/A | |
Bias was defined as the difference between the mean of the corresponding estimates and the true value of SCR. The true SRR that generated the data sets was fixed at 0.017.