Printer Friendly

Interacting with notebook input devices: an analysis of motor performance and users' expertise.

INTRODUCTION

In recent decades the use of portable computers has continuously increased, following the fundamental change in the nature of the computer-related work environment. The demand for mobility and networking has resulted in a replacement of desktop personal computers with notebooks in both business and private applications. This still-ongoing development places more emphasis on functional and effective notebook technology as well as on a more natural and easy communication with interactive systems. The numerous input devices available on the market (e.g., mouse, trackball, joystick, graphic tablet, light pen, touch screen) differ distinctly in their shape, size, mode, and cursor velocity function.

Among the types of currently popular peripheral (non-keyboard) input devices, the mouse is still the most-used device with both desktop (97%) and laptop (64%) computers (Hastings, Woods, Haslam, & Buckle, 2000). According to Hastings et al., the integrated input technology in notebooks is fairly infrequently used, in contrast to the broad usage of notebooks among computer workers in general (touchpad: 31%; touch screen: 6%; and mini-joystick: ns). It is noteworthy that users of notebooks do not use the integrated input devices very often. That they prefer the mouse is in direct contrast to the demands for universal mobility imposed on notebook manufacturers.

Difficulties in controlling the notebook devices and a lack of precision were user-reported problems associated with the handling of the touchpad and the mini-joystick (e.g., the IBM TrackPoint[R]), which were integrated in the notebooks (Douglas & Mithal, 1997). The ongoing spread of mobile computers and electronic devices in work and private areas increases the pressure of usability needs and requirements on the design of input devices and the ease with which they are used. This is of crucial importance, as many mobile devices (e.g., personal digital assistants, mobile phones, and computer notebooks) are equipped with small input devices integrated into the hardware. Here it is of central importance that input devices can be handled quickly and easily but also be operated accurately, allowing the user to directly hit and not overshoot the targeted object.

The increasing variety of input devices has attracted closet; attention to better design and more effective human-computer interaction (for an overview see Douglas & Mithal, 1997). Since 1964, when the mouse was introduced in the computer workplace, several studies have been concerned with the usability of input devices with respect to the quality of cursor control (e.g., Card, English, & Burr, 1978: mouse, isometric joystick, keyboard; Fernandez, Cihangirli, Hommertzheim, & Sabuncuoglu, 1988: mouse, joystick, touch screen, trackball; MacKenzie, Sellen, & Buxton, 1991: mouse, tablet with pen, trackball; Sears & Shneiderman, 1991: mouse, touch screen; Sperling & Tullis, 1988: mouse, trackball; Ziefle, 2003: mouse, trackball). The experimental focus of these studies was on motor performance during the use of various types of input devices (e.g., mouse, trackball, touch screen, light pen. joystick) and different cursor control actions (pointing, selection, and manipulation tasks). which makes it hard to integrate the findings across the studies. As a theoretical base for usability evaluation, Fitts's law (Fitts, 1954) is commonly referred to as a standard research paradigm in input device ergonomics (e.g.. Armbruster, Sutter, & Ziefle, 2004; Card et al., 1978; MacKenzie, 1992; Sheikh & Hoffmann, 1994; Sutter & Ziefle, 2004b; Trankle & Deutschmann, 1991). It determines the difficulty of a movement by predicting the movement time as a log-linear function of target distance and target size.

Although most studies have been directed mainly at evaluating external input devices (e.g., mouse), there has been an increase in studies specifically concerned with the usability of input devices integrated in the keyboard or the chassis of computer notebooks, such as the touchpad and mini-joystick (e.g., Armbruster, Sutter, et al., 2004; Batra, Dykstra, Hsu, Radle, & Wiedenbeck, 1998: Douglas, Kirkpatrick, & MacKenzie, 1999; Douglas & Mithal, 1997: Phillips & Triggs, 2000; Sutter & Ziefle, 2003a, 2003b, 2004a, 2004b). From an ergonomic point of view, their exclusive inspection seems to be highly justified given that the technical design, location, and size of integrated devices are restricted by the chassis. Moreover, bearing in mind that the basic benefit of computer notebooks is the possibility of working anywhere and of being completely mobile, the experimental locus should stress the quality of the different internal notebook input devices.

Batra et al. (1998) reported one of the first studies of performance with different notebook input devices in point-click, point-drag, and drawing tasks. In comparing the touchpad, mini-joystick, and trackball, all integrated into the chassis of notebooks, they found cursor control to be 10% faster and 5% more precise with the touchpad and trackball as compared with the mini-joystick. The inferiority of the mini-joystick was found to be even more distinct when more complex tasks (point-drag, drawing tasks) were examined. The performance outcomes were mirrored by usability ratings: Most of the participants reported that they highly favored the touchpad and trackball over the mini-joystick, which ranked last. Unfortunately, the button use was not equal for the input devices examined: When using the mini-joystick and trackball, users confirmed object selections or drag actions by pressing the mouse button, whereas when using the touchpad they double clicked directly on it, without using the mouse buttons, and this may have systematically given an advantage to the touchpad. Moreover, only novice users were examined, so it is unclear whether the inferior mini-joystick performance would have been found if the participant group had represented a greater range of expertise with the devices. It is possible that cursor control with the mini-joystick is more difficult to learn, reaching the benefit of training only after a relatively longer period. Thus it is unclear if the performance differences would have remained if only expert performance had been examined.

Directly contrary to the findings of Batra et al. (1998), however, Douglas et al. (1999) reported a superiority of the mini-joystick over the touchpad. The participants, novices at using notebook input devices, were distinctly faster (by 18%) when using the mini-joystick in point-click tasks. The benefit of training was higher with the touchpad (50%) than with the mini-joystick (22%), which may support the assumption that a slower learning process is characteristic for the mini-joystick.

The outcomes of the mentioned studies are difficult to interpret in terms of a clear advantage or disadvantage of one input device over the other, given that the effects of users' expertise, level of motor skill performance, and different characteristics of input devices seem to be mixed up. For a further evaluation of the mini-joystick and touchpad, as the two currently most common notebook input devices, it seems useful to have a closer look at the level of user performance, the learnability of motor performance, and the different characteristics of notebook input devices.

It is known from many studies that the learnability and expertise level reached by training differs greatly among input devices (e.g., Card et al., 1978; Douglas & Mithal, 1997; Trankle & Deutschmann, 1991 ; Ziefle, 2003). As early as the late 1970s, Card et al. (1978) showed that after about 1400 to 1600 point-click actions, no further improvement of motor performance was observable. The relative improvement by training, however, varied distinctly among different types of input devices (mouse 41%, text keys 49%, joystick 29%, and step keys 24%). Furthermore, the expert levels reliably distinguished the different devices: Expert performance with the mouse was 22% faster than with the joystick, 51% faster than with text keys, and 79% faster than with step keys. In another study (Douglas & Mithal, 1997) the participants, who were experts with the mouse, showed a 70% advantage with the mouse over the mini-joystick, with which they were inexperienced.

The learnability of input device usage can be integrated into the framework of perceptual-motor skill learning (Fitts, 1964). Highly trained motor processes rely on well-organized receptor-effector processes in spatial as well as temporal terms. According to Fitts (1964), skill learning as a continuous process can be considered with respect to three phases, of which the following are the first two: (a) The early phase, which includes the first skill-learning processes, covers the time from understanding the task instruction to the completion of a few preliminary trials, up to the development of a mental representation of the task. (b) The intermediate phase comprises associative aspects of skill learning, with mediation of the formation of either task-specific associations (learning to respond to specific cues) or cognitive set learning.

The cognitive aspects of skill learning refer specifically to the coherence of stimulus patterns on the one hand and response sequences on the other. As an illustration, a point-click task presented on a computer display (stimulus) is highly coherent once target distance, width, and movement direction are specified and the response patterns--the specific handling of the input device--are associated with these (perceptual) stimulus characteristics. Here, the two notebook input devices might have differential effects. The touchpad senses finger movements, which are translated into an analogous cursor movement on the display (linear transfer function). Thus, finger movements on the pad are mirrored as a movement on the display. In contrast, the mini-joystick senses force from the fingertip, which results in a cursor movement specified by a nonlinear transfer function. This nonlinear operation should be less coherent for the user and may result in a decelerated skill learning process, as found in the study of Douglas and Mithal (1997).

As a last step within Fitts's (1904) skill learning theory, the late phase (c) refers to skill learning over long periods of practice for which a slow continuous improvement of performance is assumed. In this last phase all responses are at a high level of skill performance based on highly organized and automated motor processes.

Thus, with regard to a meaningful and objective ergonomic evaluation of two different input devices, the comparison of only expert performance is methodologically reliable, as the specific motor programs necessary to properly use a specific input device are well established only when the late phase of skill learning is achieved, as in the case of the experts.

Research Issues and Experimental Logic

The two currently dominating notebook input devices were examined with respect to their usability and psychomotor efficiency. In order to meet demands of ecological validity, in two experiments different cursor control tasks were used that represent daily work: point-click tasks (Experiment 1) and point-drag-drop tasks (Experiment 2). Both task types are rather common when navigating through the World Wide Web and when using word-processing programs.

In order to integrate the handling of the input device into the theoretical framework of perceptual-motor skill learning (Fitts, 1964) and to evaluate the ergonomic quality of the different types of notebook input devices, we compared touchpad experts with mini-joystick experts (expert-expert comparison). As both groups were highly experienced, their performance was assumed to follow the late phase of skill learning, with highly organized motor programs and receptor-effector routines, and they should therefore show equal (high) performance. Conversely, the novice-novice comparison will show the difficulty of handling the mini-joystick and touchpad for users not yet possessing the specific motor routines necessary to handle a device with which they are unfamiliar.

In a contrast between novice and expert performance, both groups of experts also completed the tasks with the notebook input device with which they were not experienced (novices). Thus the mini-joystick experts also worked with the touchpad and, vice versa, touchpad experts a/so worked with the mini-joystick (expert-novice comparison). This systematic survey of inexperienced and experienced performance is of theoretical as well as practical interest. It is thus possible to study the late phase performance (expert state) as well as the performance of the early-intermediate phase at the same time (novice state). Theoretically, finding out whether the benefit of the late phase skills can be transferred to the performance of earlier states in the skill-learning process (carryover cost-benefit analysis) could provide great insight. Practically, this methodology may give some fruitful hints for experienced users when changing to a different input device.

The touchpad and mini-joystick differ distinctly in the way they are manipulated, which results in a higher or lower task coherence. High task coherence is present in the touchpad, in contrast to the mini-joystick. The difference in the extent of task coherence is assumed to interact with expertise: Touchpad experts, who are used to high task coherence, might have additional costs in terms of decreased performance when using a mini-joystick. In contrast, it is assumed that mini-joystick experts, who are accustomed to a complex cognitive set with respect to the association between stimulus and response patterns, will show a lower performance decrement when using the touchpad, with which they are inexperienced. This asymmetry should be present in rather simple motor tasks such as point-click tasks (Experiment 1) but should be even more distinct when the task is more complex, as in the case of point-drag-drop tasks (Experiment 2).

Based on time and accuracy measures, the effects of expertise and different input devices were looked at in two experiments. Moreover, usability ratings were surveyed as validation of performance measures. When comparing the efficiency of two different notebook input devices with respect to psychomotor performance, it is of central importance to match the actual physical speed at which the devices are run. This was pursued in an extensive preliminary testing procedure and is introduced in the next section. Then, the two experiments are described. In order to explore the transition of motor performance from novice to expert state, we conducted an additional experiment with pure novices who initially were experienced with neither the touchpad nor the mini-joystick. A general discussion follows the descriptions of the experiments, and the outcomes are taken as a basis for deriving ergonomic guidelines for optimized usage of the mini-joystick or touchpad.

PRETEST

Physical and technical differences between input devices are obvious and make it difficult to experimentally compare the devices in a precise and methodologically objective manner. The most important difference between the two types of input devices--mini-joystick and touchpad--is the transfer function with which the cursor is driven. For the mini-joystick it is the translation of finger force into cursor displacement, defined by a ratio of force to cursor motion. For the touchpad it is the extent of finger displacement on the pad and the displacement of the cursor, defined by a ratio of finger motion to cursor motion. If performance with different input devices is under study, one must ensure that the relative displacement of the cursor (control gain) is equivalent in both input devices.

Similar to the procedure of Pitrella and Holzhausen (1982), in the present study the control gain by cursor movement was adjusted by controlling the physical speed of the mini-joystick and touchpad by a special testing procedure. The different speed levels that can be chosen from the software drivers of input devices (provided as presettings in the notebook) were carefully tested with respect to the relative control gain on the display in terms of pixels per second. By using software specifically developed for this purpose, we measured the cursor displacement at different levels by either maximum speed of finger movement on the touchpad or maximum force of finger pressing on the mini-joystick. For 2 min each, 3 test participants moved the cursor as fast as possible horizontally on the display. This procedure was repeated for three system velocity levels (slow, medium, and fast) in different notebooks. The recorded data were analyzed, and cursor speed was calculated by the ratio of cursor movement to time (averaged over trials). Results are given in Table 1 for the maximum cursor speed (in pixels per second) in three cursor velocity levels (slow, medium, and fast).

As can be seen in Table 1, the software drivers of input devices provide a graduation of system adjustments up to about 10 different cursor speed levels. The enormous differences in the actual (physical) cursor speed within a system velocity level are rather striking: slow (minimum speed setting) = 66 to 1487 pixels/s, medium (medium range of the given scale) = 806 to 2333 pixels/s, and fast (maximum speed setting) = 1585 to 5447 pixels/s. Apparently, cursor speed specifications as preset by the different computer systems do not provide for a comparable physical speed, neither between drivers nor between input devices of different systems. For the present experiments, as appropriate and comparable levels, a medium cursor speed was chosen for both types of input devices: 1574 pixels/s for the mini-joystick (Toshiba Satellite 1700-300, Level 7) and 1570 pixels/s for the touchpad (Gateway 2000 Solo 9100, Level 6).

EXPERIMENT 1: PERFORMANCE IN POINT-CLICK TASKS

The first experiment focused on performance with the mini-joystick and touchpad in point-click tasks. The 30 participants (15 touchpad experts and 15 mini-joystick experts) completed 32 point-click actions with the device they were experienced with and another 32 point-click actions with the device with which they were inexperienced. Input device performance was measured by the time and accuracy of cursor control.

Method

Variables. The study was based on a two-factorial design with repeated measurement. As independent variables, expertise (expert/novice state) and input device (mini-joystick/touchpad) were examined. First, based on Fitts's (1964) theory of perceptual-motor skill learning and in order to represent a wide range of users, we investigated psychomotor performance of both inexperienced and experienced users. Second, the two input devices we studied varied in task coherence. High task coherence is obtained with the touchpad, which is manipulated by finger displacement (finger movements on the pad are similar to cursor movements), whereas low task coherence occurs with the mini-joystick, which is manipulated by finger force (it requires that finger force be translated into cursor movement).

The dependent variables were time and accuracy of cursor control as well as the judged usability of the device. The time measure of performance was the total time to complete a task, which consisted of two parts: homing time, defined as the time from the initial press of the space bar to the onset of cursor movement, and moving time, which was the interval from the onset of cursor movement to the final button click, the defined end of the task. The ergonomic arrangement of the keys and the ease of reaching a pointing device are described by the homing time, whereas movement time reflects cursor positioning and fine adjustment.

For accuracy, we measured incorrect usage of the left mouse button, or click error. This error occurred when participants clicked while the cursor was outside the target.

The usability of input device was measured first by participants' estimates of task difficulty associated with each device on a 4-point scale (4 = difficult, 3 = moderately difficult, 2 = moderately easy, 1 = easy). Second, effort was rated on a 2-point scale (1 = low, 2 = high), indicating the users' judged cost and benefit from device usage.

Apparatus and materials. Two notebooks (the Toshiba Satellite 1700-300 with a mini-joystick and the Gateway 2000 Solo 9100 with a touchpad as input technology) were used in this study. The mini-joystick, a small isometric joystick that is sensitive to strain gauges, was placed between the "G," "H," and "B" keys on the keyboard, and its two mouse buttons were located in the wrist rest. The touchpad is a fiat 6.0- x 4.4-cm touch-sensitive panel placed in front of the keyboard with two mouse buttons underneath (Figure 1). The actual cursor speed in both devices was fixed at about 1570 pixels/s (taken from the pretesting procedure, as described previously). Both notebooks were connected to an external 15-inch (38-cm) thin film transistor flat screen (liyama TXA 5841J) monitor with a 1024 x 768 resolution to control for possible confounding effects from different display qualities.

[FIGURE 1 OMITTED]

Participants. The user group investigated consisted of 16 university students and 14 employees in the services and consulting sector (11 women, 19 men). The participants were highly experienced with one of the two integrated notebook input devices. In preexperimental questionnaires, the degree of previous experience with different types of input devices was carefully assessed. For the two experimental groups, we selected only those participants who fulfilled two criteria: having expertise with one of the two input devices (all had been full-time notebook users for more than 2 years, with an average use of 3.6 hr/day, SD = 2.5) and having no experience with the other. The touchpad experts and the mini-joystick experts were both comparable (ns) with respect to their extent of self-reported expertise. Thus the expert and the novice state of participants in both groups was controlled. The participants were 21 to 39 years of age (M = 27), and all but one were light-handed. If necessary, corrective lenses were worn throughout the experiment. Participants responded to a call posted in a newsletter and volunteered for the study. There was no compensation, but the graduate students fulfilled a course requirement with their participation.

Task and software design. The point-click task represents a typical demand of computer work--that is, to select objects at different locations on the screen by pointing and clicking. To meet demands of ecological validity, we varied cursor control actions from easy (big size and short distance) to medium (big size and long distance, or small size and short distance) to difficult (small size and long distance). Participants had to select a target object (two sizes: 6.25 or 25 [mm.sup.2]) that was at one of two distances (2.5 or 5 cm) from the starting position of the cursor by moving the cursor to the square target and clicking inside its boundaries (Figure 2, left panel, shows a 6.25 [mm.sup.2] object 2.5 cm from the starting cursor position). If the participant hit it correctly, the black target changed to green, providing visual feedback. The selection confirmation with the left mouse button completed one trial, and a new display was presented immediately. The task started with a self-paced press of the space bar, and the target always appeared centrally. To exclude confounding effects of movement directions, we placed the starting position of the cursor at one of eight different locations (45[degrees], 90[degrees], 135[degrees], 180[degrees], 225[degrees], 270[degrees], 315[degrees], or 360[degrees]) relative to the target. There were 36 trials presented, of which 4 were for training purposes.

[FIGURE 2 OMITTED]

Time, spatial information for cursor movements (x, y coordinates), and button actions were recorded on line with a logging and analyzing software tool that was developed specifically for this experiment by our work group.

Procedure. In total, the experiment lasted 45 min. Participants completed the point-click actions with both input devices, mini-joystick and touchpad. The order of conditions was counterbalanced over participants.

A screening questionnaire completed in the beginning detailed the users' expertise with input devices and provided demographic data. Participants were instructed to work as fast and accurately as possible. In the four training trials, they were familiarized with the procedure and the respective task. In order to control for different working styles, we instructed the participants to handle the input device and mouse buttons exclusively one-handed.

When participants finished the point-click task with one input device, they rated its usability (ease of use). After a short break they performed the second condition (change of input device), beginning with four training trials; after completing the 32 point-click actions, they rated that device's usability.

Results

Results were analyzed by a two-way repeated-measures analysis of variance (ANOVA) with input device and expertise as the two main factors. To assess expert-expert and novice-novice differences, t tests were performed. The level of significance was set at p < .05 (values reaching p < .1 are reported as marginally significant). If the variances were shown to be not homogenous, the adjusted degrees of freedom were taken.

Time parameters. The total time comprised the two parts of the movement: homing time and moving time (Figure 3).

[FIGURE 3 OMITTED]

Although homing time was similar for both input devices (543 ms for the mini-joystick and 517 ms for the touchpad), participants needed distinctly longer to move the cursor to the target with the mini-joystick (M = 2565.5 ms) than with the touchpad (M = 1799.5 ms). This difference yielded a significant effect, F(1, 28) = 14.1, p < .01. The superiority of the touchpad over the mini-joystick was also observed in the total time, F(1, 28) = 12.7, p < .01, showing that the touchpad (M = 2304 ms) was handled 34% faster than the mini-joystick (M = 3107.5).

In looking at the differences between experts and novices, we found a significant advantage of expertise for homing time, F(1, 28) = 10.6, p < .01. The expertise advantage was also found for moving time, F(1, 28) = 9.1, p < .01, as well as for total time, F(1,28) = 10.2, p < .01. Thus experts were faster in completing the point-click tasks than were novices, which was expected. However, it was of high ergonomic interest to analyze the time costs that occurred when experts with one input device worked with the other input device. Significant interactions of input device and expertise were found for all the time measures--homing time: F(1, 28) = 5.4, p < .05; moving time: F(1, 28) = 6.2, p < .05; and total time: F(1,28) = 7.2, p < .05--clearly showing that there are indeed additional costs if experts with one input device work with an unknown input device. Carryover costs were asymmetrical, however: The inexperienced mini-joystick users (the touchpad experts using a mini-joystick) worked less efficiently than did the mini-joystick experts using the touchpad. The results for time measures are shown in Figure 4.

[FIGURE 4 OMITTED]

In Figure 4, performance is shown as a function of expertise (experts vs. novices) and input device (mini-joystick vs. touchpad). With respect to homing time, mini-joystick experts were significantly faster (M = 449 ms) than mini-joystick novices (M = 636 ms). In contrast, touchpad experts (M = 501 ms) were, statistically, as efficient as touchpad novices (M = 532 ms). The homing time of touchpad versus mini-joystick experts (expert-expert comparison) and of touchpad versus mini-joystick novices (novice-novice comparison) showed no significant differences.

For moving time, mini-joystick novices (M = 3210 ms) showed the worst performance; users in all other conditions were faster (mini-joystick experts: M = 1922 ms; touchpad experts: M = 1738 ms; touchpad novices: M = 1861 ms). Although moving time did not differ" significantly between mini-joystick experts and touchpad experts, the novice-novice comparison yielded significant differences, t(16.48) = 3.4, p < .01 : Mini-joystick novices had distinctly longer movement times (M = 3210 ms) than did touchpad novices (M = 1861 ms).

The same result pattern was found for the total time. Mini-joystick experts needed on average 2371 ms to complete the point-click tasks. When they worked with the touchpad, the unknown device, their performance was comparably good (in fact, numerically, they were even faster: M = 2368 ms). However, tremendous additional costs occurred when touchpad experts had to use the mini-joystick: On average, they completed the task in 3846 ms, whereas they needed only 2239 ms with the touchpad, the device with which they were experienced. The surplus in total time was 42%. Again, with respect to the total time, mini-joystick experts were not significantly different from touchpad experts, but the novice-novice comparison yielded significant differences in the total time, t(16.34) = 3.41, p < .01, with the mini-joystick novices (M= 3846 ms) being 38% slower than the touchpad novices (M = 2368 ms).

Errors. No significant results were found for effects of input device or expertise on the number of click errors, which was overall very small (on average 5.2 click errors occurred with the mini-joystick and 4.4 with the touchpad), hinting at the participants' rather accurate working style. Accordingly, the expert-expert and the novice-novice comparisons were nonsignificant.

Usability rating. When participants finished the point-click tasks, they rated two usability aspects of both input devices: the difficulty in processing the task (ranging from 4 = difficult to 1 = easy) and the handling effort (on a 2-point scale, with 1 = low and 2 = high). In Figure 5 the mean judgments are shown for both measures.

[FIGURE 5 OMITTED]

Processing the tasks with the mini-joystick was judged as significantly more difficult, F(1, 28) = 20.5, p < .01, than with the touchpad (mini-joystick: M = 2.1; touchpad: M = 1.3). Furthermore, expertise significantly affected difficulty in processing the tasks, F(I, 28) = 25.5, p < .01. Experts reported task difficulty as rather easy (M = 1.4), whereas novices judged it to be only moderately easy (M = 2.0). Again, user judgments were different among experts using the unknown device (significant interaction of Expertise x Input Device), F(1,28) = 7.9, p < .01.

Touchpad experts rated task difficulty as easy (M = 1.2) when using the touchpad but as much more difficult (M = 2.6) when using the mini-joystick. Asymmetrically, mini-joystick experts rated task difficulty using the input device they are trained with as only moderately easy/as = 1.7) and they rated the touchpad, the device with which they were inexperienced, as slightly more easy (M = 1.5). In looking at the judgments of both expert groups (expert-expert comparison) and both groups of novices (novice-novice comparison), we found significant differences. Mini-joystick experts (M = 1.7) rated the task as significantly more difficult, t(28) = 2.8, p < .01, than did touchpad experts (M= 1.2). Also, mini-joystick novices (M = 2.6) found it significantly more difficult, t(28) = 4.5, p < .01, to process the task in general than did touchpad novices (M = 1.5).

With respect to handling effort, the touchpad was rated as more comfortable than the mini-joystick (touchpad: M = 1: mini-joystick: M= 1.4), F(1,28) = 38.5, p < .05. Again. looking at carryover costs, we found a significant interaction of Expertise x Input Device, F(1, 28) = 38.5, p < .05. Touchpad experts using the touchpad estimated handling effort as low (M = 1.0), but when using the mini-joystick, they judged handling effort as much higher (M= 1.8). Conversely, mini-joystick experts rated handling effort as easy (M = 1.0), independent of the type of input device. Both groups of experts (expert-expert comparison) rated handling effort as equally low (M = 1.0); however, this was not the case in the novice-novice comparison. Mini-joystick novices (M = 1.7) rated the effort as significantly lower, t(14) = 6.2, p < .01, than did touchpad novices (M = 1.0).

EXPERIMENT 2: PERFORMANCE IN POINT-DRAG-DROP TASKS

In Experiment 2, the same 30 participants as in the first experiment volunteered to take part. With both input devices, they executed point-drag-drop tasks. This task type also represents typical daily work actions, as it mirrors text manipulations in word-processing programs (highlighting text passages, dragging the highlighted object, and dropping it elsewhere in the text). As in Experiment 1, according to a two-factorial design with repeated measurement, independent variables were user expertise (expert or novice) and type of input device (mini-joystick or touchpad).

Method

Variables. As dependent variables, we analyzed the three time measures used in Experiment 1. Beyond the homing time, movement time here was defined as the interval required for the point-drag-drop task, including the point-highlight and drag-drop processes. Total time consisted of homing time plus moving time. Accuracy was calculated by click and release errors. In addition to click errors, highlight errors were defined as errors within the highlight process (i.e., starting the highlight process at the wrong position or releasing the cursor at the wrong position). Drag errors (button pressed but drag object missed) and drop errors (button released outside the drop box) were calculated as well.

Task and software design. The text manipulation task is a more complex serial task, as it consists of several single actions to be executed one after another: An object had to be selected, highlighted, dragged, and dropped. Each trial began by a self-paced press of the space bar. A string of letters appeared on the screen (Figure 2, right). The cursor had to be moved from the start to the underlined letters in the center of the string. By dragging the cursor over the target string with the left mouse button pressed, the participant highlighted the target. Successful highlighting of the string was confirmed by visual feedback (the string changed to green). The mouse button was then released. After a correct button release, a target box for a drag-drop action appeared. Next, the highlighted object had to be dragged by clicking in the highlighted area and moving it into the target box with the left mouse key pressed (drag and drop). The task was finished if the target box changed to green (signaling a successful procedure) and the pressed mouse key was released. Then the target disappeared, and the next string appeared. Similar to the point-click task in the first experiment, the target box (6.25 or 25 [mm.sup.2]) appeared in eight locations (45 [degrees], 90 [degrees], 135 [degrees], 180 [degrees], 225 [degrees], 270 [degrees], 315 [degrees], 360 [degrees]) around the starting point. The distance between the starting point and the target was either 2.5 or 5 cm.

Results

Again, results were analyzed by a two-way repeated measures ANOVA with significance set at p < .05. As in the first experiment, time measures for both input devices will be described, followed by the effects of expertise and the carryover costs. Finally, the various error types in the point-drag and drag-drop tasks will be presented.

Time parameters. The experimental task required the participants to highlight the underlined object and to drag and drop it as quickly and accurately as possible into a target box. Although homing time did not differ between the two input devices (M = 532 ms for the mini-joystick and M = 524 ms for the touchpad), both input devices differed significantly in moving time, F(1, 28) = 16.6, p < .01 (mini-joystick M = 9926.9 ms, touchpad M = 7243 ins), as well total time, F(1, 28) = 16.2, p < .01. In total, the tasks were completed 36% faster with the touchpad (M = 7767 ms) than with the mini-joystick (M = 10,559 ms; Figure 6).

[FIGURE 6 OMITTED]

Is the benefit of expertise also verifiable in the more complex point-drag-drop task? The data show that this clearly is the case (Figure 7). Regarding homing time, experts (M = 481 ms) were significantly faster, F(1,28) = 5.4, p < .05, than novices (M = 576 ms; 20%). Regarding moving time, the effect of expertise was corroborated, F(1, 28) = 9.1, p < .01. Experts' moving time averaged 7524 ms, whereas novices needed nearly 30% more time (M = 9646 ms). Total time was also significantly affected by expertise, F(1, 28) = 9.5, p < .01 (expert M = 8005 ms, novice M = 10,222 ms). Although the interaction of expertise and input device was not significant for homing time, it was marginally significant for movement time, F(1, 28) = 3.5. p < . 1, and total time, F(1,28) = 2.7, p =. 1. Homing time reached 446 ms when mini-joystick experts used the mini-joystick and increased by 16%, to 533 ms, when they worked with the unknown device. Touchpad experts had a rather similar pattern: With the input device with which they were experienced, touchpad users had a homing time of 515 ins, and this increased by 10%, to 617 ins, when using the unknown mini-joystick. Neither expert-expert comparisons nor novice-novice comparisons yielded significant differences.

[FIGURE 7 OMITTED]

For mini-joystick experts, moving time with the mini-joystick was, on average, 8307 ms, and it decreased to 7746 ins when they completed the point-drag-drop task with the touchpad. Thus there was a benefit of 7% when they used the unknown device. Touchpad experts using the touchpad, in contrast, yielded an average movement time of 6741 ms, but they showed carryover costs of 40% (M = 11,546 ms) when using the mini-joystick. These carryover costs can be also seen in the total time (Figure 7).

For moving time as well as homing time, the performance of experts differed significantly when compared with novices. Mini-joystick experts needed significantly longer (moving time M = 8307 ms), t(28) = 1.9, p < .01 (total time M = 8754 ms), t(28) = 1.7, p < .01, than did touchpad experts (moving time M = 6741 ms: total time M = 7256 ms). In contrasting both groups of novices, again the mini-joystick was inferior: Mini-joystick novices were significantly slower in both movement time (M = 11,546 ins), t(28) = 3.5, p < .001, and total time (M = 12,164 ms), t(28) = 3.5, p < .001, than were touchpad novices (moving time M = 7746 ms; total time M = 8279 ms).

Errors. In the point-drag-drop task, various errors can be differentiated: click error, highlight error, drag error, and drop error. In Table 2, errors are summarized for each input device.

In both devices, click errors occurred comparably often (touchpad 2.7: mini-joystick 2.2). For highlight errors, significant differences, F(1, 28) = 8.1, p < .01, between the two input devices were observed (touchpad 8.6: mini-joystick 18.0). Furthermore, dragging, F(1, 28) = 7.5, p < .05, and dropping, F(1, 28) = 5.4, p < .05, actions led to significantly more errors when using the mini-joystick: On average, 11 drag and 11 drop errors were observed with the mini-joystick, whereas only 5 drag and 6 drop errors occurred in the touchpad.

In looking at carryover costs, we found that the interaction of Input Device x Expertise was not statistically significant for any of the error types, even though, numerically, carryover costs seemed to be present. Again, the asymmetrical cost-benefit pattern was found when expert users with one input device used the unknown device: Mini-joystick experts had an overall (all error types combined) score of about 41 errors with their familiar input device (mini-joystick). When they used the unknown touchpad, however, about 32% fewer errors were observed (M = 28). In other words, the mini-joystick experts made more errors with their familiar input device than with the unfamiliar device! In contrast--and this fits with the asymmetrical carry-over costs--the touchpad experts had an error score of only 16 errors when using the touchpad, but their accuracy was 63% lower (M = 43 errors) when using the mini-joystick. Specifically, accuracy in completing the drag-drop procedure was distinctly lower for the mini-joystick experts than for the touchpad experts. On average, mini-joystick experts made 10 drag errors, 10.6 drop errors, and 19 highlight errors, whereas touchpad experts had only 3.4 drag errors, 4.4 drop errors, and 6.9 highlight errors: drag error, t(28) = 2.7, p < .05; drop error, t(18.05) = 2.6, p < .05; highlight error, t(28) = 2.3, p < .05.

The novices did not differ significantly with respect to their accuracy in the point-drag-drop task, even though, numerically, the superiority of the touchpad over the mini-joystick always remained.

Although the total error scores differed, the proportions of error types were found to be similar for both input devices, as illustrated in Figure 8. Comparatively, click errors represented the smallest proportion in both devices (touchpad 12%, mini-joystick 5%). The dragging and dropping errors amounted to a proportion of about 26% (touchpad: drag 22%, drop 26%; mini-joystick: drag 26%, drop 26%). Definitively, the most substantial role was found for errors occurring during the highlight procedure (releasing an object too early or too late), with 43% errors with the mini-joystick and 40% with the touchpad. Statistical analysis revealed significant differences between the frequency of errors, [chi square](5) = 215.4, p < .01. Click errors (M = 8.5%) occurred significantly the least, and highlight errors occurred the most (M = 41.5%), whereas the drag (M = 24%) and drop errors (M = 26%) did not differ from each other but did significantly differ from the click and highlight errors.

Usability rating. As in Experiment 1, participants rated task difficulty on a 4-point scale (4 = difficult, 1 = easy) and handling effort on a 2-point scale (1 = low, 2 = high) after finishing the point-drag-drop tasks. In Figure 9 the mean judgments are illustrated for both measures.

[FIGURE 9 OMITTED]

Task difficulty with the mini-joystick was rated as significantly more difficult (M = 3.3) than with the touchpad (M = 2.3), F(1, 28) = 7.3, p < .05. Experts rated task difficulty as significantly less difficult (M = 2.5) than novices did (M = 2.9), F(1, 28) = 5.9, p < .05. The interaction of Expertise x Input Device was not found to be significant. With respect to the judged handling effort, no significant differences were obtained for the different input devices (touchpad M = 1.6, mini-joystick M = 1.7), but expertise was shown to significantly affect handling effort, F(1,28) = 11.5, p < .0 l (expert M = 1.5, novice M = 1.8). Mini-joystick experts judged task difficulty and handling effort as comparably high as touchpad experts did, yielding no significant expert-expert differences. However, mini-joystick novices reported the task difficulty (M = 3.4) and handling effort (M = 1.9) of the mini-joystick to be very high, whereas touchpad novices worked much better with the touchpad: task difficulty, M = 2.4, t(28) = 3.5, p < .001: handling effort, M = 1.7, t(28) = 1.9, p < .01.

META-VIEW: TRANSITION FROM EXPERT TO NOVICE

So far, we have compared two experimental groups with regard to their motor performance: Experts were defined by a highly trained usage of either the mini-joystick or touchpad. In contrasting expert performance, we also examined these same participants as novices in a transfer condition (transfer novices). The results so far revealed highly asymmetrical carryover costs in only one group: When a touchpad expert had to use the mini-joystick, performance dropped dramatically. Conversely, mini-joystick experts profited even when using the touchpad. These results can be fully interpreted within Fitts's (1964) motor learning theory only by validating the motor performance of "pure" novices (i.e., novices in both the mini-joystick and touchpad conditions) and contrasting the performance. This was undertaken by examining a third group of 22 participants (11 women, 11 men; age M = 25.5 years) without experience with the mini-joystick or touchpad. Of these 22 participants, 11 (6 women, 5 men; age M = 23.4 years) were investigated with the touchpad, and the other 11 (5 women, 6 men, age M = 25.6 years) operated the mini-joystick. Both groups performed 32 point-click and 32 point-drag-drop tasks. To contrast these different expertise groups, we performed a meta-view comparing movement time across the three groups (experts, transfer novices, and pure novices) for both input devices. The results will be presented first for the point-click task and then for the point-drag-drop task.

In the point-click task, the ANOVA revealed significant effects for the main factors of expertise, F(2, 76) = 7.5, p < .01, and input device, F(1, 76) = 21.9, p < .01. Moreover, the interaction of expertise by input device yielded significance, F(2, 76) = 4.1, p < .05. As illustrated in Figure 10 Cleft panel), experts' movement time (M = 1829 ms) was found to be distinctly lower; by about 700 ms. than that of the transfer novices and pure novices (M = 2535 and 2568 ms, respectively). No performance difference between the novice groups was observed. Overall, performance with the mini-joystick was inferior to that with the touchpad (M = 2752 vs. 1889 ms, respectively). However. the interaction of expertise and type of input device revealed that this was true only for the transfer novices, t(16.48) = 3.4, p < .01, and pure novices, t(17.4) = 3.5, p < .01, not for the experts, who did comparably well with both input devices (ns).

[FIGURE 10 OMITTED]

For the point-drag-drop tasks, again, the analysis of movement time showed significant effects for both factors, expertise, F(2, 76) = 9.1, p < .01, and input device, F(1, 76) = 23.9, p < .01, but their interaction was not significant. As illustrated in Figure 10 (right panel), experts' performance (M = 7524 ms) was considerably superior to that of transfer and pure novices (M= 9646 and 10,703 ms, respectively). Again, no significant differences between the two novice groups' performance was found. Thus it can be assumed that both transfer novices and pure novices perform under equal conditions of motor learning behavior. The comparison of the two types of input devices showed again that performance with the mini-joystick was on average 38% inferior to that with the touchpad (M = 10,801 vs. 7781 ms), even with experts.

The benefit of this meta-view analysis turned out to be twofold. One is a methodological benefit, the other a benefit with respect to evaluation. From a methodological point of view, the performance of the transfer novices, as described in Experiments 1 and 2, can now be unequivocally classified and interpreted: In fact, the transfer novices performed like pure novices, as both showed comparable motor performance in point-click and point-drag-drop tasks. From an evaluation point of view, the participants who were touchpad experts and transfer novices with the mini-joystick were apparently unable to capitalize on their general motor expertise with the usage of an input device--rather; they fell back to an early stage of motor learning (Fitts, 1964). The same was true for the mini-joystick experts, who were transfer novices on the touchpad. At first glance they seemed to profit from their general motor experience, because they exhibited no transfer costs and even showed slight benefits. A detailed look, however; showed that the same benefit of the touchpad applied to pure novices as well. Thus it can be concluded that it is the easy and intuitive handling of the touchpad itself that benefits performance. Conversely, the mini-joystick--especially for novices--entails high performance costs, as can be seen from the overall lower performance with the mini-joystick when compared with the touchpad (nearly 40% lower; when considering both tasks and all users).

DISCUSSION

The aim of the present study was to ergonomically evaluate the usability of a mini-joystick and a touchpad, the two most common notebook input devices, by focusing on cursor control performance in point-click and point-drag-drop cursor actions. Touchpad experts were compared with mini-joystick experts to determine motor performance in the late phase of skill learning, thus allowing a direct comparison of the quality of the two input devices. Both groups of experts completed the tasks not only with their expert device but also with the device they were inexperienced with (carryover costs). In order to gain a deeper insight into how their expertise with the familiar device may or may not have transferred to performance with the unfamiliar device (as transfer novices), we examined a third group consisting of pure novices with both input devices. Speed and accuracy of performance were analyzed, and usability ratings regarding the difficulty of task processing and ease of use when handling the input devices were surveyed. The outcomes are now summarized and discussed in the context of knowledge provided by the literature.

Method. Even though they were not the central experimental focus, it seems noteworthy that the presettings of cursor velocities available in notebooks do not provide a valid physical speed graduation, either within a type of input device or between different software drivers. If the actual cursor displacement is under experimental study, it is recommended that preliminary tests be carried out in order to determine the actual physical speed in pixels pet" second. The successful matching procedure for the physical speed allowed the overall comparison of two input devices that are rather different in their mode of manipulation.

Input device and task coherence. It was assumed that the stronger task coherence, defined as the concordance of response and stimulus patterns, is manifested in the touchpad condition, in which finger movements are translated into a linear movement on the display. This assumption was confirmed by the performance data. Overall, there was a clear superiority of touchpad performance over that of the mini-joystick. For movement time, the superiority reached 34% in point-click tasks and 36% in point-drag-drop tasks. The errors that occurred in the more complex point-drag-drop tasks give further insights into the lower task coherence present in the mini-joystick condition. In highlighting the target, participants' error rate with the mini-joystick was more than double that with the touchpad. The same was true for the drag and drop process: With the mini-joystick, participants made about twice as many drag and drop errors as with the touchpad.

The participants' usability ratings revealed that task processing was much easier and handling comfort was better with the touchpad than with the mini-joystick, an input device that demands sophisticated motor skill coordination of force and fingers. According to participants' (mini-joystick novices') remarks after the experiment, the main problem with the mini-joystick was in balancing finger force, making it highly problematic to control the cursor Interestingly, participants felt that the mini-joystick ran "much faster" than the touchpad, even though the physical cursor speeds of both input devices were definitively equal. Presumably, the participants thought that their greater difficulty with the mini-joystick was caused by its "faster" cursor speed.

Expertise. In order to make a reliable comparison of the different input devices with respect to usability, expert performance with both the mini-joystick and the touchpad was contrasted. According to theory (Fitts, 1964), experts are assumed to have established a proper cognitive and motor skill coordination and, thus, should exhibit perfectly harmonized response and stimulus task patterns when executing cursor actions with the input device with which they are experienced. From this, two expectations were derived. One was that expert performance should outperform novice performance, which is rather trivial. The second expectation is less trivial, however: that the performance of both groups of experts (touchpad and mini-joystick) should be equally good. As the data show, the overall superiority of expert over novice performance was definitively present, and this is in close accordance with Fitts's (1964) motor learning theory. Experts were faster, by 26%, in both the simple clicking actions and the more complex point-drag-drop task. Given that the experts' superiority was equal in both task types (each 26%), task complexity did not interact with the benefit of expertise (otherwise, the benefit of expertise should have been distinctly bigger in the more complex point-drag-drop task).

The second assumption--that experts should show comparably good performance, independent of the type of input device--was found not to be true. In the point-click task, the expert-expert differences did not reach statistical significance, thus confirming the assumption at first glance. In the more difficult point-drag-drop task, however; the differences between the mini-joystick and touchpad experts became broadly evident: mini-joystick experts were 1.4 s slower than touchpad experts (20%) and made 2.5 times the errors that the touchpad experts made, a striking performance difference. Hence, the speed and accuracy with the mini-joystick's cursor control was distinctly inferior, even though the users were highly experienced. This result cannot be explained by Fitts'(1964) motor skill learning theory, as it contrasts expert versus novice performance but does not differentiate different expertise levels in the late phase of skill learning. As both expert groups possessed the same amount of experience with their input device (as was statistically proven), it can be derived that an expert state on one device is not necessarily on a par with an expert state on another device but, rather, that the usability of input devices moderates expert performance.

It cannot be fully derived from the present data what exactly makes the mini-joystick's handling so difficult. What can be taken from our detailed motor analyses is that the performance inferiority of the mini-joystick is attributable to its longer movement times, accompanied by its greater inaccuracy of cursor control (drag and drop errors). Useful hints can be derived from the study of Douglas and Mithal (1997), who analyzed the microstructure of cursor movements. They suggested that use of the mini-joystick is characterized by a movement tremor that causes involuntary changes in cursor velocity, resulting in prolonged movement times. Our current research is concerned with precisely analyzing cursor routes and motor routines, from the starting point to the target area, and observing any detour routes that participants undertake.

Transfer of skill learning. Another point of theoretical as well as practical import was the analysis of carryover costs and benefits--that is, when experts worked with the input device with which they were inexperienced. The question was whether the late phase skills can be transferred to intermediate or early phases of motor skill learning. In other words, does expertise with one input device benefit novice performance with another device? The outcomes clearly show that the benefits and costs are asymmetrical and depend on the type of input device. This was shown by examining a group of users who were novices with both input devices. Transfer novices (i.e., novices with one input device but experts with the other) exhibited performance similar to that of pure novices (i.e., users with no experience at all with either device).

The carryover costs, however, are not equal but are less pronounced in the mini-joystick experts (as the mini-joystick is much more difficult to use, anyway). Here, mini-joystick experts had a clear advantage, as they performed even better when using the touchpad, the input device with which they were inexperienced. However, tremendous additional costs occurred when touchpad experts used the mini-joystick: In point-click tasks, they needed 71% longer to complete the simple object selection with the inexperienced input device (mini-joystick) and made three times more click errors. In contrast, mini-joystick experts managed to keep their expert performance level with the unknown touchpad. For the more complex tasks, such as the point-drag-drop task, the asymmetry of carryover benefits and costs are again corroborated. Touchpad experts using the mini-joystick were 40% slower and made nearly three times more click and release errors than they did with the touchpad. Directly contrary to this were the performance outcomes of mini-joystick experts. Using the touchpad, they enhanced their expert mini-joystick level by almost 5% and made distinctly fewer errors than they did with the mini-joystick, the device with which they were experts. Thus it cannot be concluded that the expert state leads to the best performance in any case.

Apparently, the motor routines that mini-joystick experts have already developed must be regarded as very capacious, comprising rather complex motor skills and including even the skills necessary to use the touchpad as an expert. Argued from the other side, however, the touchpad seems to be an input device that is overall very easy to use and that is, in contrast to the mini-joystick, more usable from an ergonomic point of view as it requires less motor skill and therefore covers the needs and skills of a wider range of users.

In conclusion, it can be stated that the touch-pad is highly effective and can therefore be recommended as a usable notebook input device.

The touchpad provides for a stable, good performance, being rather robust at different user expertise levels, and is recommended for beginners as well as experts. Specifically', its high task consistency enables intuitive and easy handling. Thus the touchpad is regarded as suitable for a wide range of customer segments.

Limitations of the study. Some final remarks concern potential limitations of the study with respect to the generalizability of the results. One should keep in mind, however, that in our study we examined only young, healthy adults, who surely do not represent the total variety of characteristics among computer users, and thus the results we obtained probably underestimate the situation that would be found with a wider variety of users.

First, all our participants differed with respect to their degree of expertise with one input device, but all were highly trained mouse users. The relative impact of this high mouse experience cannot be estimated, and thus we cannot exclude the possibility that experience with an input device (the mouse) influenced the results, positively and/or negatively. However, from the meta-view analysis amending the knowledge about novice performance, it can be reliably derived that being experienced with one input device did not positively influence performance with the other device under study (for mini-joystick experts it was the touchpad, and vice versa). Thus, transferring this logic to the present question, it has to be assumed that the experience with the mouse all participants possessed did not affect their motor performance when using the mini-joystick or touchpad, in either the point-click or the point-drag-drop task. Ultimately, however, this can be proven only by performing a cross-check with mouse novices (if that is at all possible, as mouse novices have become extinct within this young user group). However, for the results presented here, this possible limitation does not play a role in the credibility of the findings, as it was held constant over all conditions.

A second note of caution concerns age as a main user characteristic that modulates motor performance; users' motor abilities decrease with age. In recent studies (Armbruster, Sutter, et al., 2004: Armbruster, Ziefle, & Sutter, 2004), a distinct motor performance deterioration of about 52% in handling internal input devices was found in older adults as compared with younger adults. To what extent the considerable carryover costs shown here would be exhibited by older adults cannot be estimated on the basis of the present results and will have to be analyzed separately in future experiments.

A final limitation with respect to the generalizability of the findings refers to the health state of input device users. The alarming, increasing numbers of muscular disorders reported in current health statistics (e.g., Keir, Bach, & Rempel, 1999) make it necessary to validate the results for users with different kinds of repetitive strain injuries (e.g., carpal tunnel syndrome), perhaps specifying which of the internal input devices may fit which kind and extent of muscular problem.

Future research. Current studies are focusing on the nature of the inferior performance found in the mini-joystick condition. One approach is to study whether there is an optimal (lower?) cursor speed that matches the specific characteristics of the mini-joystick. Another approach is to track the actual cursor path, from the starting point to the target, and to precisely analyze the cursor speed as well as the acceleration of the cursor in different positions on the way to the target. By looking at both experienced and inexperienced user performance and measuring the detour routes participants take with the different notebook input devices, some useful insights should be gained. Furthermore, training experiments need to focus in detail on the characteristics of novice performance and the transition from the novice to the expert level. Here it will be of central interest to determine whether the expert level with the mini-joystick can be achieved in the same time as with the touchpad and if it can ever reach a comparable extent. Moreover, the training processes of children and older adults (also novices with respect to the usage of input devices) when using the two notebook input devices need to be examined, identifying the developmental changes (ascending in the case of child users and/or descending for older users) with respect to motor performance when handling internal input devices.

ACKNOWLEDGMENTS

Our special thanks go to Philipp Brauner for programming the sophisticated experimental and analysis software. Furthermore, many thanks to Preethy Pappachan, Julia Pollmann, and Judith Strenk for their help in analyzing the data and preparing the figures. Thanks also to Anna Dyckhoff, Gregor Giesen, Azadeh Nikookhesal, Dietmar Sommer, and Michael Wagner for further support. Finally, for their critical comments on earlier versions of this manuscript, thanks to Sarah Hatfield, Silke Molnar, Kim Petsch, Eduardo Salas, and two anonymous reviewers.
TABLE 1: Mean Maximum Notebook Cursor Velocity (Pixels/s)

                                Toshiba       Toshiba       Toshiba
                  Gateway      Satellite     Satellite     Satellite
Cursor             2000        Pro 4200        1730        1700-300
Velocity         Solo 9100       Mini-         Mini-         Mini-
(Presetting)     Touchpad      Joystick      Joystick      Joystick

Slow               66 (l)      1487 (l)        78 (l)        76 (l)
Medium           1570 (6)      2333 (5)       806 (5)       880 (5)
                                                           1574 (7)
Fast             2960 (11)     5447 (10)     1585 (10)     2366 (10)

Note. The velocity scale level preset in the notebook is given in
Parentheses

TABLE 2: Mean Errors in the Point-Drag-Drop Task

                    Touchpad     Mini-Joystick

Click error            2.7            2.2
Highlight error        8.6           18.0
Drop error             5.8           11.1
Drag error             4.9           10.8
Total errors          22.0           42.1

                    error types in    error types in
                  the mini-joystick    the touchpad

highlight error          43%                40%
drop error               26%                26%
drag error               26%                22%
click error               5%                12%

Note: Table made from pie chart.


REFERENCES

Armbruster, C., Sutter; C., & Ziefle, M. (2004). Target size and distance: Important factors for designing user interfaces for older notebook users. In H. M. Khalid, M.G. Helander, & A.W. Yeo (Eds.), Work with computing systems 2004 (pp. 454-459). Kuala Lumpur, Malaysia: Damai Sciences.

Armbruster. C., Ziefle, M., & Sutter, C. (2004). Notebook input devices put to an age test: The usability of TrackPoint and touchpad for older adults. Manuscript submitted for publication.

Batra, S., Dykstra, D., Hsu, P., Radle, K., & Wiedenbeck, S. (1998). Pointing device performance for laptop computers. In Proceedings of the Human Factors and Ergonomics Society 42nd Annual Meeting (pp. 535-540). Santa Monica, CA: Human Factors and Ergonomics Society.

Card, S. K., English, W. K., & Burr, B.J. (1978). Evaluation of mouse, rate-controlled isometric joystick, step keys and text keys for selection tasks on a CRT. In R. Baecker & W. Buxton (Eds.), Readings in human-computer interaction (pp. 386-392). San Mateo, CA: Morgan Kaufmann.

Douglas, S. A., Kirkpatrick, A. E., & MacKenzie, I. S. (1999). Testing pointing device performance and user assessment with the ISO 9241, Part 9 standard. In Proceedings of the ACM Conference on Human Factors in Computing Systems--CHI '99 (pp. 215-222). New York: Association for Computing Machinery.

Douglas, S.A., & Mithal, A.K. (1997). The ergonomics of computer pointing devices. New York: Springer.

Fernandez, J.E., Cihangirli, M., Hommertzheim, D. L., & Sabuncuoglu, I. (1988). The effects of input devices on task performance. In F. Aghazadeh (Ed.), Trends in ergonomics/ human factors V (pp. 83-89). Amsterdam: Elsevier.

Fitts, P.M. (1954). The information capacity of the human motor system in controlling the amplitude of movement. Journal of Experimental Psychology, 47, 381-391.

Fitts, P.M. (1964). Perceptual-motor skill learning. In A.W. Melton (Ed.), Categories of human learning (pp. 243-285). New York: Academic.

Hastings, S., Woods, V., Haslam, R.A., & Buckle, P. (2000). Health risks from mice and other non-keyboard input devices. In P. T. McCabe, M.A. Hanson, & S. A. Robertson (Eds.), Contemporary ergonomics 2000 (pp. 312-316). London: Taylor & Francis.

Keir, P.J., Bach, J. M., & Rempel, D. (1999). Effects of computer design and task on carpal tunnel pressure. Ergonomics, 42, 1350-1360.

MacKenzie. I.S. (1992). Fitts' law as a research and design tool in human-computer interaction. Human-Computer Interaction, 7, 91-139.

MacKenzie. I.S., Sellen, A., & Buxton, W. (1991). A comparison of input devices in elemental pointing and dragging tasks. In Proceedings of the CHI' 91 Conference on Human Factors in Computing Systems (pp. 161-166). New York: Association for Computing Machinery.

Phillips. J.G., & Triggs, T. J. (2000). Cursor control device characteristics. Australian Journal of Information Systems, 7, 115-119.

Pitrella. F.D., & Holzhausen, K.-P. (1982). Selection and experimental comparison of computer input devices (Report No. 57). Wachtberg, Germany: Forschungsgesellschaft fur Angewandte Naturwissenschaften.

Sears, A., & Shneiderman, B. (1991), High precision touchscreens: Design strategies and comparisons with a mouse. International Journal of Man-Machine Studies, 34, 593-613.

Sheikh, I.H., & Hoffmann, E.R. (1994). Effect of target shape on movement time in a Fitts task. Ergonomics, 37, 1535-1547.

Sperling, B.B., & Tullis, T.S. (1988). Are you a better "mouser" or "trackballer"? A comparison of cursor-positioning performance. SIGCHI Bulletin, 19(3). 77-81.

Sutter, C., & Ziefle, M. (2003a). Health hazard from input devices: The diagnostics of muscular load and motor performance revisited. In H. Luczak & K. J. Zink (Eds.), Human factors in organizational design and management (pp. 525-530). Santa Monica, CA: International Ergonomics Association.

Sutter, C., & Ziefle, M. (2003b). How to handle notebook input devices: An insight in button use strategy. In R McCabe (Ed.), Contemporary ergonomics 2003 (pp. 241-246). London: Taylor & Francis.

Sutter, C., & Ziefle, M. (2004a). Accuracy measures of cursor movement path validated on notebook input devices. In H. M. Khalid, M. G. Helander, & A. W. Yeo (Eds.). Work with computing systems 2004 (pp. 525-528). Kuala Lumpur, Malaysia: Damai Sciences.

Sutter, C., & Ziefle, M. (2004b). Psychomotor efficiency in users of notebook input devices: Confirmation and restrictions of Fitts law as an evaluative tool for user-friendly design. In Proceedings of the Human Factors and Ergonomics Society 48th Annual Meeting (pp. 773-777). Santa Monica, CA: Human Factors and Ergonomics Society.

Trankle, U., & Deutschmann, D. (1991). Factors influencing speed and precision of cursor positioning using a mouse. Ergonomics, 34, 161-174.

Ziefle, M. (2003). Is the trackball a serious alternative to the mouse? A comparison of trackball and mouse with regard to cursor movement performance in manipulation tasks. In D. Harris, V. Duffy, M. Smith, & C. Stephanidis (Eds.), Human-centred computing: Cognitive, social and ergonomic aspects (pp. 158-162). Mahwah, NJ: Erlbaum.

Christine Sutter received her M.S. in psychology in 2001 from RWTH Aachen University, Germany, where she is a research assistant in the Department of Psychology.

Martina Ziefle is an associate professor at the Institute of Psychology at Aachen University. She received her Ph.D. in psychology from the University of Fribourg, Switzerland, in 1991.

Date received: July 3, 2002 Date accepted: January 18, 2005

Requests for reprints should be sent to Christine Sutter, Department of Psychology, RWTH Aachen University, Jaegerstrasse 17-19, 52056 Aachen, Germany; Christine.sutter@post.rwth.aachen.de.
COPYRIGHT 2005 Human Factors and Ergonomics Society
No portion of this article can be reproduced without the express written permission from the copyright holder.
Copyright 2005 Gale, Cengage Learning. All rights reserved.

Article Details
Printer friendly Cite/link Email Feedback
Author:Sutter, Christine; Ziefle, Martina
Publication:Human Factors
Date:Mar 22, 2005
Words:10869
Previous Article:Children using cellular phones: the effects of shortcomings in user interface design.
Next Article:The impact of hearing protection on sound localization and orienting behavior.
Topics:

Terms of use | Privacy policy | Copyright © 2019 Farlex, Inc. | Feedback | For webmasters