Does variability in human performance outweigh imprecision in response devices such as computer keyboards?

Abstract

In chronometric research, it is well-known that many response devices used with personal computers, especially computer keyboards, suffer from measurement inaccuracies due to infrequent polling. In this article, it is investigated whether it is worth being concerned by this added error, given that human performance inherently exhibits a considerable degree of variability. By comparing inaccuracy due to response device imprecision with the level of variability already present in human performance, it is shown that the former is very unlikely to have a negative impact on statistical outcomes. In other words, millisecond resolution of response devices, although desirable, is not a general requirement. © 2010 The Psychonomic Society, Inc.