Fig. 3
From: Visual search under physical effort is faster but more vulnerable to distractor interference
Posterior distributions of the hierarchical Bayesian ex-Gaussian parameters. The curves indicate the fits of a nonparametric kernel density estimates to the posterior samples for each group-level parameter. The triangles on the x-axis indicate posterior means. The posterior distributions of each parameter across all four conditions (a) can be decomposed into the main effect of singleton distractor presence versus absence (b), physical load (d), and the interaction effect (d). For the interaction effect (d), shaded areas represent 95% highest density interval (95% HDI) of posteriors. Only the exponential parameter Ï„ yielded a credible interaction effect with the lower bound not crossing zero