Table 3. Best parameters, deviance, and AIC for each model.
It also presents the critical value of χ2 for the Likelihood Ratio Test for nested models.
Model | Parameters | |||||||||
biasall | biasnothing | biasleft | biasin | α | β | γ | δ | ϵ | ζ | |
MBiases | 0.13 | 0.076 | 0.058 | 0.005 | 0 | 0 | 0 | 0 | 0 | 0 |
wsls | 0.1 | 0.05 | 0.018 | 0.002 | 38 | 30 | 4.6 | 0 | 0 | 0 |
fra | 0.06 | 0.05 | 0.003 | 0.000 | 40 | 30 | 15 | 0.5 | 30 | 0.954 |
Conventions: * signifies that χ2 has 3 d.o.f. and that p < 0.001. Moreover, ** signifies that χ2 has 6 d.o.f. and that p < 0.001. Finally, † signifies strong support that the winning model is closer to the true generating process over and above the complexity introduced by the extra parameters.