It does not fitts my data! analysing large amounts of mobile touch data

Abstract

Touchscreens are the dominant input device for smartphones and learning about smartphone users' touch behaviour became even more important. We developed a game for Android phones to collect a truly large amount of touch data from diverse devices and players. A part of the game is designed as what we expected to be a Fitts' law task. By publishing the game in the Android Market we collected 5,359,650 micro tasks from 63,154 installations of the game. Using Fitts' law to find a model for these tasks we found a very weak correlation and an implausible high index of performance across different devices. Further analysis shows a similar correlation between time and distance as with Fitts' law but only a very weak correlation with the targets' width. © 2011 IFIP International Federation for Information Processing.