Changes in Error Rates during Mouse Pointing Tasks Paced by Estiamted Movement Time Given by Fitts'-Law

Abstract

Fitts' law can provide HCI engineers with a means to predict the time length necessary for the completion of such tasks because the predicted movement time by the law is generally robust and accurate. These studies, however, did not discuss the issue with the relationship between Fitts' law parameters and error rates when the tasks are performed. The objective of this study is to show empirically and discuss changes in error rates during pointing tasks paced by varying movement time including T_EST, estimated movement time given by the IDs. Two sets of experiments were conducted. The objective of the first experiment was to obtain constants of the Fitts' law equation by the linear regression model through various IDs. The second experiment was to show the relationship between time constraint and error rates. Results of the first experiment showed that the individual ID-Movement time relationship was well-fitted to the model with an average R² of 0.959. The second experiment showed that the average error rate was 21% at the estimated movement time, which was much higher than that reported by Card, et al. (1983). The data were then analyzed by using the logistic regression. The results implied that a feedback movement in the pointing task is significantly disturbed by the attention to the pace during the task.