Table 1.
An example of a problem in Experiment 1, and its abstract form.
Table 2.
The conjunctive events of the 16 contents for the problems in Experiment 1, and their respective mean percentage probability estimates.
Table 3.
The two different orders of estimates in Experiment 1, the percentage of participants' violations of the JPD, and the latencies (in s) of participants' estimates of the three different probabilities.
Figure 1.
3D scatterplots of estimates of P(A), P(B), and P(A&B) and 2D scatterplots of estimates of P(A) and P(B) in Experiment 1.
Panel A shows the estimates of 2000 simulated runs of the computational model and its best fitting linear regression plane, and Panel B shows participants' estimates. Participants' estimates were separated by whether the estimate reflected zero, one, or two violations of the JPD. A violation was defined as a negative probability in the JPD extrapolated from the estimates. In the 2D scatterplots, estimates of P(A&B) correspond to the size of points such that larger points indicate larger estimates.
Figure 2.
3D scatterplots of estimates of P(A), P(B), and P(A&B) and 2D scatterplots of verbal scale (Panel A) and numerical scale (Panel B) estimates of P(A) and P(B) in Experiment 2 (see Figure 1 for an explanation of zero, one, or two violations).
Participants' estimates were separated by whether the estimate reflected zero, one, or two violations of the JPD. A violation was defined as a negative probability in the JPD extrapolated from the estimates. In the 2D scatterplots, estimates of P(A&B) correspond to the size of points such that larger points indicate larger estimates.
Table 4.
Model comparisons (R2 and RMSE) between mReasoner and two alternative models of probability estimates (Wyer's [32] equation 3a: {P(A)+P(B)/2+[P(A)*P(B)]}, and Fantino et al. [33] [P(A)+P(B)/2] against the data from Experiments 1 and 2.