Validity of estimates of the mean effect size
| Publication bias | N | τ2 | I2 (%) | Bias | MSE |
|---|---|---|---|---|---|
| None | 10 | 0.02 | 17 | 0.00 | 0.01 |
| None | 10 | 0.12 | 56 | 0.00 | 0.03 |
| None | 10 | 0.90 | 90 | 0.00 | 0.11 |
| None | 20 | 0.02 | 17 | 0.00 | 0.01 |
| None | 20 | 0.12 | 56 | 0.00 | 0.01 |
| None | 20 | 0.90 | 90 | 0.00 | 0.05 |
| None | 50 | 0.02 | 17 | 0.00 | 0.00 |
| None | 50 | 0.12 | 56 | 0.00 | 0.01 |
| None | 50 | 0.90 | 90 | 0.00 | 0.02 |
| None | 100 | 0.02 | 17 | 0.00 | 0.00 |
| None | 100 | 0.12 | 56 | 0.00 | 0.00 |
| None | 100 | 0.90 | 90 | 0.00 | 0.01 |
| RR=4 | 10 | 0.02 | 17 | 0.07 | 0.02 |
| RR=4 | 10 | 0.12 | 56 | 0.16 | 0.06 |
| RR=4 | 10 | 0.90 | 90 | 0.50 | 0.36 |
| RR=4 | 20 | 0.02 | 17 | 0.06 | 0.01 |
| RR=4 | 20 | 0.12 | 56 | 0.16 | 0.04 |
| RR=4 | 20 | 0.90 | 90 | 0.49 | 0.30 |
| RR=4 | 50 | 0.02 | 17 | 0.06 | 0.01 |
| RR=4 | 50 | 0.12 | 56 | 0.16 | 0.03 |
| RR=4 | 50 | 0.90 | 90 | 0.49 | 0.27 |
| RR=4 | 100 | 0.02 | 17 | 0.06 | 0.01 |
| RR=4 | 100 | 0.12 | 56 | 0.16 | 0.03 |
| RR=4 | 100 | 0.90 | 90 | 0.49 | 0.25 |
| RR=10 | 10 | 0.02 | 17 | 0.16 | 0.05 |
| RR=10 | 10 | 0.12 | 56 | 0.34 | 0.15 |
| RR=10 | 10 | 0.90 | 90 | 0.83 | 0.78 |
| RR=10 | 20 | 0.02 | 17 | 0.16 | 0.04 |
| RR=10 | 20 | 0.12 | 56 | 0.34 | 0.13 |
| RR=10 | 20 | 0.90 | 90 | 0.83 | 0.73 |
| RR=10 | 50 | 0.02 | 17 | 0.16 | 0.03 |
| RR=10 | 50 | 0.12 | 56 | 0.34 | 0.12 |
| RR=10 | 50 | 0.90 | 90 | 0.83 | 0.71 |
| RR=10 | 100 | 0.02 | 17 | 0.16 | 0.03 |
| RR=10 | 100 | 0.12 | 56 | 0.34 | 0.12 |
| RR=10 | 100 | 0.90 | 90 | 0.83 | 0.70 |
All estimates were based on 10 000 samples.
Bias, the average estimate of the mean effect size; MSE, mean squared error.