AIC/BIC difference | Expected (Raftery’s) | Sample size |
---|
400 | 800 | 1600 | 3200 | 6400 |
---|
0-2
| 50-75% | 59.3% | 60.4% | 61.1% | 62.9% | 64.9% |
2-6
| 75-95% | 82.3% | 85.4% | 86.4% | 86.5% | 87.1% |
6-10
| 95-99% | 96.2% | 97.3% | 97.8% | 97.5% | 97.9% |
>10
| >99% | 99.4% | 99.8% | 99.2% | 100.0% | 100.0% |
- Proportion of correct classifications using the AIC/BIC in terms of the observed BIC difference between models. The proportions shown are averaged over all the scenarios considered in Tables 3 and 4. The expected proportions of Raftery are also listed for these differences as a reference.