Chronometric counting is a prevalent issue in the study of human time perception as it reduces the construct validity of tasks and can conceal existing timing deficits. Several methods have been proposed to prevent counting strategies, but the factors promoting those strategies in specific tasks are largely uninvestigated. Here, we modified a classical two-interval duration discrimination task in two aspects that could affect the tendency to apply counting strategies. We removed the pause between the two intervals and changed the task instructions: Participants decided whether a short event occurred in the first or in the second half of a reference duration. In Experiment 1, both classical and modified task versions were performed under timing conditions, in which participants were asked not to count, and counting conditions, in which counting was explicitly instructed. The task modifications led to (i) a general decrease in judgment precision, (ii) a shift of the point of subjective equality, and (iii) a counting-related increase in reaction times, suggesting enhanced cognitive effort of counting during the modified task version. Precision in the two task versions was not differently affected by instructed counting. Experiment 2 demonstrates that—in the absence of any counting-related instructions—participants are less likely to engage in spontaneous counting in the modified task version. These results enhance our understanding of the two-interval duration discrimination task and demonstrate that the modifications tested here—although they do not significantly reduce the effectiveness of instructed counting—can diminish the spontaneous tendency to adopt counting strategies.

计时计数是人类时间感知研究中的一个普遍问题，因为它降低了任务的构造有效性并且可以隐藏现有的计时缺陷。已经提出了几种方法来防止计数策略，但在特定任务中促进这些策略的因素在很大程度上未被研究。在这里，我们在两个方面修改了经典的两间隔持续时间辨别任务，这可能会影响应用计数策略的趋势。我们删除了两个间隔之间的停顿并更改了任务说明：参与者决定在参考持续时间的前半段还是后半段发生短事件。在实验 1 中，经典任务版本和修改任务版本都是在时间条件下执行的，其中参与者被要求不计数，计数条件，其中明确指示计数。任务修改导致（i）判断精度普遍下降，（ii）主观平等点的转移，以及（iii）与计数相关的反应时间增加，这表明在修改后的任务中增加了计数的认知努力版本。两个任务版本中的精度没有受到指示计数的不同影响。实验 2 表明，在没有任何与计数相关的指令的情况下，参与者不太可能在修改后的任务版本中进行自发计数。