Figure 1
Qualitatively different regions of effects. A. The three regions are positive, zero, and negative effects, and they correspond to three distinct classes of theories. B. The distributions of individual effects can either be in one region or span several regions. The distribution labeled “No”, which has no qualitative individual differences, is constrained to the positive region. The distribution labeled “Yes” spans all three regions.
Figure 2
Observed and model-estimated effects. The observed effects are shown as crosses, and the variability of these estimates reflects both trial noise and true variability across people. The model-estimated effects are shown as circles, and they account for trial noise reflecting only true variability across people. A. Stroop-effect data from Von Bastian et al. (2015). B. Stroop-effect data from Rey-Mermet et al. (2018).
Figure 3
The three models are shown as bivariate distributions across two individuals’ true values (denoted θ1 and θ2). Darker areas show greater concentration of density.
Figure 4
Analyses of orientation strength and bias effects. The data are from Rouder et al. (2010). A. Mean response time for correct responses as a function of tilt angle. Both strength and bias effects may be observed. B. Empirical and model estimates of individual strength effects. The evidence favors quantitative individual differences. C. Empirical and model estimates of individual bias effects. Although there is a significant overall rightward bias, the evidence here favors qualitative individual differences. D. Posterior probability that each participant has a rightward bias. Horizontal lines are for 1-to-1, 3-to-1, and 10-to-1 odds.