For switch trials, there were no significant effects involving SwitchProb, whereas the analysis of repetitions showed an interaction of SwitchProb and RCI which approached significance, F (1, 47) = 4.27, p = .044 (Bonferroni-corrected p = .088), η p 2 = .083.
Statistical power considerations Given the uncertainty regarding the diagnostic interaction between switch probability, switch cost and RCI in the error analyses in Experiment 1 (it was marginally significant in the ANOVA, but only “anectodal” in the Bayesian test), we sought to augment the statistical power by increasing the number of participants.