Results Report

Based on this experiment, results favored the TouchPad. However, the degree is not significant enough to draw the conclusion of branding the TouchPad to be the optimum and favored notebook computer input device. The TouchPad performed the best of the 3 input devices. It had the fastest task completion time and lowest error rate. The TrackBall on the other hand fell behind on both task time and error rate. The AccuPoint yielded more errors than TrackBall but showed a faster performance time.

User preference also seemed to reflect the results of tasks. Most subjects preferred the TouchPad over the AccuPoint and TrackBall. Some subjects even preferred TouchPad over a regular mouse as noted in the subjective questionnaires.

However, it is imperative to note that the difference between the speed, accuracy and subjective preference is not significant, so a strong conclusion cannot be drawn on which of the 3 input devices is optimum.

This proved our original hypothesis to be false. Despite of the similar mechanical structure of the TrackBall compared to a standard mouse, the TrackBall performed slower and less accurate than the other devices.

We realized that rating these 3 input devices with 1 particular task is insufficient. From past experiments, certain input devices rated higher for some tasks and lower in others. As mentioned in the introduction, in the experiment conducted by MacKenzie, Sellen, and Buxton, the TrackBall, among 3 devices, rated the worst in simple pointing and dragging tasks. Whereas, in Epp's experiment, he found the TrackBall to perform the fastest among 6 devices in graphics editing tasks. Our task was simply a text editing task that measured manipulation in pointing, clicking, dragging, and releasing operations. Although our results showed to be inconclusive, similar experiments with these 3 input devices under different tasks could show a particular device to be significantly more appropriate.

Abnormality

An area of our experiment we found to be inconsistent resided in the subjects' personality type. Some subjects preferred to work fast, lowering their speed to complete the task, but increasing the error rate. Whereas, others preferred to work more carefully, concentrating on avoiding errors, but taking longer time to complete the task.

Even though the order of devices used to perform the tasks was rotated for every subject, subjects still performed relatively poorly on the first device used in comparison to the later devices. This can be caused by the frustration or unfamiliarity of the task on the first attempt. This could affect the subjective questionnaire in a way such that subjects tended to "dislike" the first device tested.

Due to the Dual Scan display of the notebook computers, subjects were not able to find their cursor at times. Some subjects had to pause their task and look for the cursor on the screen. This added some extra time to the completion of the tasks for some of the subjects. It also had an impact on the final statistic summary.

Department of Computer Sciences Direct questions and comments to the student editorial team