Assessing the effect of different input modalities on error recovery

It is often claimed that touch-capable devices are more usable and faster than their non-touch-capable counterparts. Empirical studies however, have shown the increase in speed to be only marginally significant, while the cost of this increase is reflected in increased error rates. To limit the severity of problems introduced by using touch-capable devices, an account of the users' cognitive limitations is required in order to guide future design of said devices and software packages. Reaction times and mousetracking data was obtained using both an online pilot study and the main experiment, wherein participants played a simple probe matching game. Through the online pilot experiment, touch pad data could be obtained along with mouse data. The main experiment obtained mouse and touch screen data. Results confirm empirical evidence on the effects of input modality (touch screen versus computer mouse), and introduces previously uninvestigated evidence for a larger than expected increase in error recovery speed using a touch screen. Additionally, the mousetracking data conforms to data found in response conflict experiments using either eye-tracking or hand/arm-tracking. Based on the results, and their relation to theories on human cognition, recommendations are made for future research and development in usability engineering and interaction design.