Table 2.
Main effects linear mixed model on the probability of the conditional P(if p then q).
| Effect | Parameter | F | df | F-scaling | p |
|---|---|---|---|---|---|
| (Intercept) | 0.14 | 8.70 | 1, 60.35 | 1 | 0.005 |
| Inference | 0.43 | 3, 10.16 | 0.84 | 0.74 | |
| P(q|p) | 0.78 | 86.81 | 1, 88.14 | 1 | <0.001 |
| P(p ∧ q) | 0.00 | 0.00 | 1, 90.91 | 1 | >0.99 |
| P(¬p ∨ q) | −0.01 | 0.01 | 1, 88.26 | 1 | 0.91 |
| P(q|¬p) | −0.00 | 0.00 | 1, 81.59 | 1 | 0.98 |
The model was fitted with restricted maximum likelihood. Model df = 20, AIC = −98.42, BIC = −42.67, deviance = −138.42, Ω20 = 0.84 (explained variance against the intercept only model; Xu, 2003). The values in the table note are based on a model with variables centered on 0 to be comparable to the Supplemental material.