Bayesian parameter estimation and aggregation of evidence across experiments.
Violin plots depicting the posterior distributions for the main effects of memory strength and retention interval as well as the two-way interaction. Light gray violins depict posterior distributions for the parameters given non-informative priors and data from the separate experiments. Dark gray violins depict posterior distributions for the parameters when aggregating evidence across experiments using the posterior of the previous experiments as the prior for the later experiments. White dots represent the center of the posterior distribution, and black rectangles depict 95% high posterior density intervals. The dashed red line marks the zero effect line. (A) Posterior distributions of the main effect of memory strength. (B) Posterior distributions of the main effect of retention interval (i.e. forgetting). (C) Posterior distributions for the retention interval and memory strength interaction. The distributions of the interaction effect overlap the zero effect line, supporting the notion that there was no difference in forgetting rate between the conditions.