A series of our previous experiments (Exp. 1 – 3) with an interference task (the Multi-Source Interference Task, Bush & Shin, 2006) have been conducted to investigate control schedule learning, that is adjustment of executive function based on temporal regularity. This planed experiment (Exp. 4) aims at replicating the results of Exp. 2. Temporal position of interference stimuli in successive five trials was manipulated. In a uniform group, interference stimuli were presented uniformly on 2 – 5 trials. In a consistent group(s), they were presented on a fixed position (e.g., position 2 for the consistent 2 group or position 3 for the consistent 3 group). In a biased group, they were presented on the position 4 with the probability of 80%. In other words, interference stimuli were likely to appear on a specific position but did not always appear on that position. Exp. 1 had three consistent groups (the consistent 2, 3, 4 groups) and a uniform group with 4 blocks. Exp. 2 had a biased group and a uniform group with 4 blocks. Exp. 3 had the three groups (consistent, biased, and uniform) with 12 blocks. The sample size for previous experiments (Exp. 1 – 3) was determined in a heuristic manner: For Exp. 1 and 2 it was set to secure a statistical power of 80% by presuming a small effect size (f = .10) and for Exp. 3 it was arranged by a “rule of thumb” of 1,600 observations per condition in analysis with mixed effects models (Brysbaert & Stevens, 2018). The results of Exp. 1 – 3 were inconsistent with each other. Exp. 1 showed that accuracy on interference stimuli was higher for the three consistent groups than for the uniform group (the contrast of the three consistent groups vs. the uniform group). We did not find such a pattern in response times. For accuracy of interference stimuli on position 4, Exp. 2 did not indicate an advantage of the biased group over the uniform group but showed the interaction between group and block. The pattern of this interaction indicated that the advantage for the biased group over the uniform group became more pronounced as the block progressed, suggesting a learning effect. For response times, we did not find either an advantage for the biased group or interaction between group and block. For performance for interference stimuli on position 4, Exp. 3 did not demonstrate either the effect of groups or interaction on accuracy performance although it showed the interaction between group and stimulus type (interference vs. control stimuli on position 4) on RT performance. Given the inconsistency of results in Exp. 1 – 3, post-hoc power analyses were conducted regarding accuracy (i.e., observed power calculation), using simr package in R. For Exp. 1 the estimated power was approximately 70.50% for the contrast of the three consistent groups vs. the uniform group. For Exp. 2 it was approximately 36.90% and 55.80% for the main effect (group) and the interaction (group x block). For Exp. 3 it was approximately 33.60% and 22.90% for the contrast (consistent & biased vs. uniform) and interaction (the contrast x block). The estimated powers indicate that the previous experiments were under powered (< 80%). Therefore, Exp. 4 was designed to provide conclusive evidence that would resolve the inconsistency between the previous experiments. It aims to directly replicate Exp. 2 by following the procedure of Exp. 2 but increases the sample size to secure the statistical power of 80%. The sample size was determined as 700 participants to secure statistical power of 80% for the main effect of group (i.e., biased vs. uniform) and that of 93% for the interaction between group x block based on effect sizes calculated by the results of Exp. 2.