Counting does not improve the accuracy of long time productions

It is common to use counting strategies to produce time intervals accurately. Does counting improve the accuracy (deviation of produced duration from veridical duration) and precision (variability of produced duration) of time productions in the range of seconds to minutes? In a series of experiments, we compared chronometric counting to intuitive timing (no counting) and to attentional control (simultaneously performing mental arithmetic). In a field experiment, participants had to produce time intervals of 60-s duration in a classroom setting. Relative to intuitive timing, counting did not improve the accuracy (absolute error) of time productions but led to overproduction of duration (larger constant error). In four laboratory experiments, we tested the effects of counting on time production of 10-, 30-, 45-, 60-, and 90-s intervals. Consistently, counting did not improve the accuracy of time production. The relative overproduction due to counting was replicated for long intervals (60 and 90 s) but disappeared at shorter intervals. However, across all intervals tested, counting had positive effects on the precision of time production. As expected, mental arithmetic impaired accuracy and precision and led to overproduction of duration, indicating that participants followed instructions. Based on the experimental data, the overproduction of longer intervals due to counting can neither be explained in terms of attentional processes nor by means of a word-length effect when counting multisyllabic numbers or when participants switch their pace of counting from mono- to multisyllables.