Table 2. Model fits (-lnL and BIC) for the best fitting model for each observer and duration.
| Long Duration | |||||||||||
| Parameters | AR | DC | DO | ||||||||
| Model | c | k | w | λ | M | -lnL | BIC | -lnL | BIC | -lnL | BIC |
| Fully Flexible | TxB | TxB | TxB | TxB | 160 | 7307 | 15588 | 5819 | 12612 | 5763 | 12499 |
| Fixed λ | TxB | TxB | TxB | F | 121 | 7307 | 15350 | 5819 | 12375 | 5763 | 12262 |
| Flexible Specificity | TxB | T | B | F | 63 | 7426 | 15236 | 5944 | 12272 | 6007 | 12398 |
| Flexible Criterion | B | TxB | B | F | 81 | 7368 | 15228 | 5909 | 12311 | 5911 | 12314 |
| Fully Constrained | F | F | F | F | 4 | 9260 | 18544 | 10531 | 21086 | 10095 | 20214 |
Notes: c - specificity parameter, k -criterion parmeter, w - attention weight and λ, response exponent. TxB - Task x Block, T - Task, B - Block, F - Fixed across task and block. M – number of parameters, -lnL - Negative log-likelihood, BIC - Bayesian Information Criterion.
The best fitting model (i.e., lowest BIC) is shown in bold.