Table 3. Minimum and maximum bootstrap selection model selection rates of the 3 fitted 2-group models in relation to AIC differences observed in samples drawn from the population.
| Δ AIC | Model 1 | Model 2 | Model 3 | ||||||
|---|---|---|---|---|---|---|---|---|---|
| N=200 | |||||||||
| min | max | #samp | min | max | #samp | min | max | #samp | |
| 0-2 | 0.136 | 0.518 | 53 | 0.564 | 0.862 | 59 | 0.112 | 0.457 | 56 |
| 2-4 | 0.181 | 0.538 | 20 | 0.501 | 0.908 | 89 | 0.196 | 0.515 | 37 |
| 4-6 | 0.265 | 0.606 | 6 | 0.619 | 0.906 | 81 | 0.255 | 0.623 | 17 |
| 6-8 | 0.297 | 0.509 | 4 | 0.534 | 0.935 | 105 | 0.436 | 0.544 | 3 |
| 8-10 | 0.37 | 0.783 | 3 | 0.694 | 0.968 | 88 | 0.388 | 0.651 | 3 |
| 10-12 | 0.643 | 0.643 | 1 | 0.514 | 0.974 | 73 | NA | NA | 0 |
| 12-15 | 0.643 | 0.643 | 1 | 0.55 | 0.982 | 102 | NA | NA | 0 |
| 15-20 | NA | NA | 0 | 0.683 | 0.992 | 100 | NA | NA | 0 |
| >20 | NA | NA | 0 | 0.687 | 0.996 | 94 | NA | NA | 0 |
| N=350 | |||||||||
| 0-2 | 0.179 | 0.353 | 25 | 0.69 | 0.862 | 48 | 0.084 | 0.319 | 27 |
| 2-4 | 0.242 | 0.449 | 20 | 0.688 | 0.9 | 67 | 0.165 | 0.443 | 8 |
| 4-6 | 0.348 | 0.487 | 7 | 0.783 | 0.943 | 71 | 0.262 | 0.371 | 4 |
| 6-8 | 0.47 | 0.542 | 2 | 0.824 | 0.954 | 74 | 0.459 | 0.459 | 1 |
| 8-10 | 0.511 | 0.511 | 1 | 0.859 | 0.967 | 64 | NA | NA | 0 |
| 10-12 | NA | NA | 0 | 0.898 | 0.981 | 81 | NA | NA | 0 |
| 12-15 | NA | NA | 0 | 0.787 | 0.988 | 97 | NA | NA | 0 |
| 15-20 | NA | NA | 0 | 0.914 | 0.996 | 151 | NA | NA | 0 |
| >20 | NA | NA | 0 | 0.815 | 1 | 252 | NA | NA | 0 |
| N=700 | |||||||||
| 0-2 | 0.274 | 0.348 | 2 | 0.734 | 0.734 | 1 | NA | NA | 0 |
| 2-4 | 0.317 | 0.364 | 3 | 0.79 | 0.868 | 3 | NA | NA | 0 |
| 4-6 | NA | NA | 0 | 0.842 | 0.855 | 2 | NA | NA | 0 |
| 6-8 | NA | NA | 0 | 0.907 | 0.954 | 4 | NA | NA | 0 |
| 8-10 | NA | NA | 0 | 0.875 | 0.968 | 6 | NA | NA | 0 |
| 10-12 | NA | NA | 0 | 0.91 | 0.964 | 8 | NA | NA | 0 |
| 12-15 | NA | NA | 0 | 0.922 | 0.988 | 22 | NA | NA | 0 |
| 15-20 | NA | NA | 0 | 0.932 | 0.996 | 54 | NA | NA | 0 |
| >20 | NA | NA | 0 | 0.962 | 1 | 895 | NA | NA | 0 |
Note: Δ AIC is the AIC difference between the best and second-best fitting model in samples drawn from the population. Min and max are the minimum and maximum bootstrap selection rates for samples within a given range of Δ AIC presented for each winning model, and #samp is the number of samples within that range favoring that model. For instance, for N=200, there were 53 samples (out of P=1000) with an AIC difference between 0 and 2 favoring model 1, and the minimum and maximum bootstrap selection rates for model 1 were 0.136 and 0.518, respectively.