To compare the findings between Experiments 1 and 2, we categorized reproduced duration as "Short" or "Long" relative to 1.2 s. Figure 4A presents psychometric curves that reveal an assimilation bias toward previous durations only in the prior Time condition. In the prior Time condition, the PSE for prior long and short intervals were 987 ± 59 ms and 1111 ± 52 ms, respectively. In the prior Direction condition, these values were 1124 ± 52 ms and 1114 ± 61 ms, respectively (Figure 4B). A two-way repeated measures ANOVA revealed a main effect of the previous Duration, F(1,23) = 5.407, p = .029,\(\eta_{p}^{2}\) = 0.011, and a main effect of the prior Task, F(1,23) = 6.150, p = .021,\(\eta_{p}^{2}\) = 0.017, and a significant interaction effect between these factors (F(1,23) = 5.479, p = .028, \(\eta_{p}^{2}\) = 0.015). Further analysis revealed a significant assimilation effect in the task-relevant (Time) condition (t(23) = 3.465, p = .004, BF10 = 18.385) but not in the task-irrelevant (Direction) condition (t(23) = 0.239, p = 1, BF10 = 0.22).
Additionally, to assess the decisional carry-over effect, we grouped trials based on the preceding reports ("Short Response" or "Long Response"). Figure 4C displays psychometric curves for each group, revealing a distinct difference based on prior responses. As indicated in Figure 4D, the PSE values for the prior "Long Response" was 934 ± 45 ms, significantly shorter than the prior "Short Response" (1166 ± 55 ms), revealing a significant decisional carry-over effect (t(23) = 3.457, p =.002, d = 0.939).