almost significantp = .084
In contrast, presenting the bars with perfectly stable (nonflashing) noise patterns did not produce any overestimations: When both, the bars and the pattern, were stable (Experiment 3c, Figure 2(k) ) the two-tailed Wilcoxon test resulted in an almost significant difference, W = 10, p = .084, r =.86, pointing, however, to underestimations on the noise side.