It was expected that a significant difference in performance would be seen that could be related to the size of the mouse pad. The relationship would be inversely proportional, with performance improving (completion time decreasing) as the size of the mouse pad is increased due to the user's less frequent need to reposition the mouse to avoid falling off the edge of the mouse pad. This relationship was not expected to be linear however, with the improvement between Medium and Large mouse pads being smaller than the improvement between Small and Medium pads. This relationship was expected to be seen in both the clicking and dragging portions of the experiment. An illustration of the expected results is shown below.
The size of the mouse pads are: (SMALL, MEDIUM, LARGE) Each subject group was presented the pads in a different order (SML,SLM,MSL,MLS,LSM,LMS).
2.2.2 Dependent VariablesThe performance times of the user to complete clicking and dragging tasks for each size mouse pad were tabulated. In addition to the performance times, each subject was asked to fill out a subjective questionnaire for each size pad and then a final questionnaire afer using all three pads.
Three individuals were tested, one for each size mouse pad. As a result of this study, several key changes were made. First off, the foam being used for the mouse pads was too sticky and did not create the appropriate mouse pad surface. So, a soft material was glued on top of the foam to create a smooth surface. In the materials packet, spaces to write down performance times were included so that users could work at that own pace. In the program, a bug was found that prevented the user from running through a series of tasks more than once. The program was debugged so that a user could run through a series of tasks as many times as desired. (for the experiment, this was three)
Thirty students from the College Of Computer Science at the University of Maryland were utilized for our experiment. All of our subjects had a minimum of two to three years of computer experience and working knowledge of a mouse.
A total of 18 rectangular mouse pads were used in the experiment, 6 each of 3 sizes: Small (15 cm X 18 cm), Medium (20 cm X 24 cm), and Large (25 cm X 30 cm). These dimensions were suggested so that the Medium pad was approximately the size of most commonly available mouse pads, with the Small pad 25% smaller in linear dimensions and the Large pad 25% larger. Mouse pads were constructed of a fabric surface affixed to a 1/4" foam rubber base (the most common thickness of commercial mouse pads).
Six Gateway Pentium 166 PCs with 17" SVGA monitors and Gateway mice much like any common mouse were used to conduct the experiment. All systems were equipped with the Windows 95 operating system, running in SVGA (1024 X 768) mode. Monitor position was fixed by the desks used in the laboratory. Mouse pads were placed in identical positions for all users, with the center of the mouse pad positioned consistently for all 3 sizes of mouse pads. Mice were centered on the mouse pads at three times during each trial (see discussion of procedure) 2.5.2 SoftwareA simple testing program was written in Visual Basic. Sample screen captures can be seen below (shown at reduced size - view for full size), and the program itself is available here (Win95/NT executable). The program included two basic screens, one with a clicking task, another with a drag and drop task. In the clicking task screens, twenty-five buttons appeared on screen. One button would appear red and clicking on the red button would cause another button to turn red. The order that the buttons turned red was random. After all twenty-five buttons had been clicked in the proper order a number, the subject's completion time in seconds, appeared on screen along with instructions. The drag and drop screens were similar to the clicking screens, but with the center button on the clicking screens replaced by a larger black box. Instead of clicking on the red button, subjects to drag the button to the black box, where it was dropped. As before, upon completion of the task the subject's completion time appeared on screen.
Three versions of each of the basic task screens were used in the program. Each version had a different order of activated buttons, achieved through the mirroring or flipping of the activation order. This ensured all three versions would require the same amount of movement in the same order. For example, the first two activated buttons of the first version could be the upper left corner and lower right corner, the upper right corner and lower left corner for the second version, and the lower left corner and upper right corner for the third version - though movements went in different directions, the magnitude and axis of motion were the same. The result of this approach is three effectively identical version of the same task, but with little risk of subjects being able to predict which button would next become activated, a problem if the order of activation was not altered. Subjects completed all 3 versions of each task, alternating tasks as they progressed (clicking, drag and drop, clicking, and so on)
The program also included several screens of brief instructions to the tasks. A Clicking on the OK button of each instruction screen launched the next task and activated the timer. Additional instruction screens appeared after each pair of tasks (one clicking screen, one drag and drop screen) instructing subjects to wait for the proctors to switch the subject's mouse pad before continuing with the program. (To get a clear view of the screen shots, right click on the shot and select view image.)
Subjects also completed questionnaires during the experiment. Background questions on computer and mouse usage were asked, followed by many subjective questions about the mouse pads and places to record the completion times which appeared on screen. An abbreviated text version can be found here .
Upon entering the laboratory subjects were instructed to sit at workstations running the test program, sign a consent form, and read the first page of the questionnaire packet. Once all the subjects had arrived, they were given verbal instructions on what to expect during the experiment. These instructions elaborated on the ones in the questionnaire packet and on the workstation screen.
Once given instructions, subjects began the experiment. Following instructions on the screen, the user performed the clicking task first. When the user finished the clicking task, a time appeared on the screen - this time was recorded in the space provided in the packet. Following more on-screen instructions the user performed the dragging task. Upon completion, the time was recorded and subjects completed all questions appearing on that page of the questionnaire. During this time the administrators went to each workstation and replaced the mouse pad just used by the subject with a different size pad. This procedure was repeated for each size of mouse pad. After completing the final task and the questions on that page, another page of subjective questions were completed. A short debriefing was given after all subjects had finished.
With three sizes of mouse pads there were six possible permutations of order which had to be used in order to eliminate any learning effects. All subjects present in any one session would be given mouse pads in the same order, for example small, medium, large, the order of pads changed with each session. In the end five subjects had completed the experiment for each permutation of size order.