Subitizing-Counting Analogue Observed in a Fast Multi-tapping Task

摘要

A widely-used method for entering texts with mobile phones is the multi-tapping one. To understand the internal processes of humans while doing multi-tapping, we designed an experiment with time pressure, that is, to multi-tap a key for the prescribed number of times as fast as possible. We observed three types in time series data for inter-tap intervals: Type I (Plan and Do), Type II (Do and Adjust), and Type III (Mixture of Type I and II). And, for the distribution of errors in the tapping, we found a subitizing-counting analogue. That is, if the instructed number of tapping was smaller (< 4), the error rate was smaller, and if the number was larger (> 4), the error rate rose abruptly. These findings could lead to the model of human cognition and manipulation of the number, hence to the design of the usable human interface.