Table 1:

p Values for the SH test on the bias-corrected MSE summed across participants

PeakLevelLeaky (τ = 100 ms)Leaky (τ = 500 ms)Leaky (τ = 1000 ms)Full (τ = ∞)
RT0.0010.0070.0520.3040.4880.906
PSS00.0160.1240.3380.5090.889
  • The null hypothesis is that the corresponding model is not different from the best model, while the alternative hypothesis is that the corresponding model is worse than the best model. Bold type indicates the models within a confidence set of models.