Brought to you by:
University of Florida Libraries
|Full Text + Links|
|·Full Size Images|
|PDF (701 K)|
|Save as Citation Alert|
Young children’s ability to use a computer mouse
Because there is little empirical data available on how well young children are able to use a computer mouse, the present study examined their proficiency in clicking on small objects at various positions on the screen and their skill in moving objects over the screen, using drag-and-drop and click-move-click. The participants were 104 children from Kindergarten 2 and Grade 1. The results show that children in Kindergarten 2 clicked and moved slower than children in Grade 1. Nearly all of the children were able to click within 3 mm horizontally and 6 mm vertically from the centre of a 3 mm target. The findings also demonstrate that in educational software drag-and-drop is the most appropriate movement procedure as it was found to be faster than click-move-click and resulted in fewer interaction errors. Interesting differences between horizontal and vertical movements were found. It is concluded that young children are generally well capable of using a mouse to operate educational software, making this a suitable input device for such applications.
Keywords: Human–computer interaction; Interface; Media in education; Navigation; Elementary education
Educational software is frequently used in schools as it offers children individualised instructions and practice. Apart from instructional content, presentation of content, and interaction procedures, the learning gain of educational software and the enjoyment that children experience when they are working with it seem to be related to the input device (Smith & Keep, 1986). Little research has been done to investigate the usability of the mouse, the most common input device in educational software. Therefore, the main purpose of this study was to examine how accurately young children are able to use the mouse (Hourcade, 2002). The mouse proficiency of children from Kindergarten 2 and Grade 1 was assessed by asking them to click repeatedly on a very small object at various positions on the screen and to move objects over the screen using two different procedures. It was expected that Kindergarten children would have difficulties in mouse handling and that Grade 1 children would use the mouse more smoothly, because of their more developed motor skills.
Different input devices may be used to operate educational software, but most developers and users choose keyboards, joysticks, and the mouse (Whitefield, 1986 and Wilton and McLean, 1984). An obvious advantage of the mouse, joystick, and keyboard is that they do not hinder the user’s view of the screen, unlike light-pens, for instance. To keep the input-device out of sight, however, the movements of the cursor need to be projected on the movements of the user, drawing heavily on eye-hand co-ordination. Nonetheless, children as well as adults appear to work faster and produce fewer errors when working with a mouse, instead of other input-devices (Joiner et al., 1998 and MacKenzie et al., 1991). Additionally, the computer mouse is the preferred input-device of children (Smith & Keep, 1986).
The motor skills of children are continuously developing. Children of 4 and 5 years old have less direct reaching trajectories than adults, children younger than 8 are not able to grab and lift objects as well as adults, and the speed of fast hand movements increases until children are 12 (Kuhtz-Buschbeck, Boczek-Funcke, Illert, Joehnk, & Stolze, 1999). The differences in motor development can cause a variety of difficulties for young children in executing subtle and delicate movements that require eye-hand co-ordination. Children may for instance not yet be able to tie their shoelaces or write letters, and similarly, they may not be able to adequately use computer input devices. Children practice their fine motor-skills intensively when they are learning to write, in the Netherlands at the age of seven. It is expected that this not only improves their performance on writing tasks, but also on other motor-tasks, such as handling a computer mouse (Lindemann & Wright, 1998). Of course, the mouse skills of children also improve with age because older children use the computer more regularly than younger children (Crook, 1992, Joiner et al., 1998 and Wilton and McLean, 1984). The present study investigated whether the differences in motor development between children who could, and could not write, were reflected in their use of a mouse, and whether particular inabilities of young children require specially designed educational interfaces.
The mouse as an input device is used to point at an object, to click it, or to move it over the computer screen. Pointing is an action consisting of three stages: rushing towards the target, reducing speed, and aiming precisely (Dennerlein & Yang, 2001). Clicking adds a fourth stage: pressing and/or releasing the mouse button while keeping the mouse stationary. To move an object, these four stages have to be completed twice: once to select the movable object and a second time to select the release-spot. Both clicking and moving can be evaluated by examining the speed at which children can execute the operation, the number of mistakes they make, and how comfortable children feel while using the mouse (Inkpen et al., 1996b and Whisenand and Emurian, 1999).
The most basic mouse-operation is a task where subjects have to aim precisely at an object. They generally try to do this quickly, but without making too many errors. A balance between speed and accuracy is attained by a strategic, recurrent process: subjects increase their aiming speed until too many mistakes occur, at which time they decrease their speed to a safe level and then start increasing their speed again. This continues until the subject has reached a speed at which he considers his rate of error acceptable to himself and the experimenter (Elliott, Hansen, Mendoza, & Tremblay, 2004). A number of aspects of the task may increase the speed that accompanies the rate of error acceptable to subjects. A square target, for instance, is reported to be faster to aim at than a round target, probably due to the fact that the surface of a square target is larger than the surface of a round target when their widths are equal (Crook, 1992, Inkpen et al., 1996b, Phillips and Triggs, 2001, Tränkle and Deutschmann, 1991 and Whisenand and Emurian, 1999). Obviously, large targets result in faster aiming speeds and more correct clicks. The present study investigated how large an object needs to be to ensure that young children are able to accurately select it. This was examined by recording how accurately children clicked on a small target that appeared at various locations on the screen.
In educational software, objects are rarely alone on the screen. Normally, software offers several choices, meaning that children have to click on an object that is accompanied by other objects. When a single object is on the screen, mistakes by users obviously do not have many consequences; children can just try again. However, when an object is surrounded by other objects, they may accidentally click on an unwanted object and consequently give a wrong answer. To prevent these mistakes, children have to decrease their rate of error. Adults can increase their aiming precision accordingly, sometimes at the expense of speed. Children on the other hand may not be able to aim very precisely or to hold the mouse steady during clicking and may as a result still miss the target (Crook, 1992 and Walker et al., 1993). In the present study, distracting objects were placed close to, or far from, the target to evaluate their impact on accuracy. The distance between the centre of the target and the location of the click was recorded as well as the reaction time of children. It was expected that children would consider fewer errors acceptable when the target was presented close to distracting objects and that they would need more time to attain this high level of accuracy.
In educational software children often have to reposition objects, for instance moving a target to a matching release spot. Roughly two procedures can be used to move objects on a computer screen: drag-and-drop or click-move-click. Drag-and-drop, also referred to as dragging, drag-drop, or point-drag, consists of holding the mouse over an object, pressing the mouse button to select it, repositioning the mouse while keeping the mouse button pressed, and releasing the mouse button after the cursor is positioned at the release-spot (Crook, 1992, Gillan et al., 1990, Inkpen et al., 1996a, Inkpen et al., 1996b, Joiner et al., 1998 and Whisenand and Emurian, 1999). In the alternative procedure the mouse is clicked to select the object, the mouse is repositioned, and the mouse is clicked again to release the object. In the present article this procedure is referred to as click-move-click, in the literature as point-select, point-click, point-and-click, or pointing (Inkpen et al., 1996a, Inkpen et al., 1996b, Joiner et al., 1998 and Whisenand and Emurian, 1999). There is still a debate whether adults are faster and more accurate using drag-and-drop or click-move-click (Joiner et al., 1998, MacKenzie et al., 1991 and Whisenand and Emurian, 1999). Drag-and-drop could lead to more errors, because it requires users to maintain pressure on the mouse button, which is motorically demanding (MacKenzie et al., 1991 and Strommen, 1993). On the other hand, this demand can actually prevent errors as it reminds users continuously of the action they are performing (Joiner et al., 1998). Because drag-and-drop is often used in current software, it may have become a routine for users, especially those who use computers regularly. When people have to deviate from their routines, they are likely to get sidetracked and errors will occur (Heckhausen and Beckmann, 1990 and Sternberg, 1996).
For children, it has been suggested that click-move-click is more suitable than drag-and-drop (Segers & Verhoeven, 2002). Especially young children moved objects faster using click-move-click than drag-and-drop if they had to move objects over distances larger than 200 pixels on a 640 × 480 screen (Joiner et al., 1998). However, in the study of Joiner et al. the object in the drag-and-drop condition (an animal) may have afforded slower movement than the (inanimate) cursor in the click-move-click condition (Phillips, Triggs, & Meehan, 2001). In another study children between 9 and 13 years of age were also faster using click-move-click than drag-and-drop (Inkpen et al., 1996b). The difference was very small (65 ms), but increased considerably when error-trials were included, which was probably due to differences in error handling between the two interaction procedures. When the children made an error in drag-and-drop, they had to perform the entire trial again, but when they made an error in click-move-click, they only had to perform the incorrect click again. This could also have affected the correct trials, as children may have adopted a more accurate, slower, strategy during drag-and-drop than during click-move-click to avoid time-consuming mistakes. Apparently, the dispute between drag-and-drop and click-move-click can not be resolved on the basis of past research. In the present study we attempted to resolve this issue by using carefully designed software in which drag-and-drop and click-move-click used the same cursors and error handling.
An often applied method to compare mouse-movements is Fitts’ Law, which calculates information processing rate (workload) on the basis of target variables and movement speed (Hourcade, 2002). However, Fitts’ law only applies to error-free behaviour and unlike adults, children are often not capable of moving the mouse without making mistakes. Therefore, the present study examined the reaction time and movement speed of children as well as three types of errors. First, selection errors may occur when children fail to accurately select the object. Secondly, the object may accidentally be dropped during movement (Inkpen et al., 1996b). Drop-errors mainly occur because children find it difficult to maintain pressure on the mouse button, but also because children click to release the object before they have reached the release-spot (Inkpen, 1997, MacKenzie et al., 1991 and Strommen, 1993). Is it too difficult for children to keep the mouse button pressed during dragging? If this is the case, more drop errors would be made when objects were moved over long distances than over short distances (Inkpen, 1997, MacKenzie et al., 1991 and Strommen, 1993). However, when drop errors are due to difficulties in selecting the release-spot, distance of movement would have no effect. The third type of error, interaction errors, occur when children try to use drag-and-drop when they are required to use click-move-click, or vice versa. When children have become more familiar with one of the movement procedures through their daily use of computers, they would probably make fewer interaction errors when using this procedure.
There are indications that adults are more comfortable moving objects from left to right – the direction in which we read and write – rather than from right to left, but whether there is a difference between horizontal and vertical movements is as yet unknown (Crook, 1992, Tränkle and Deutschmann, 1991 and Whisenand and Emurian, 1999). We expected that children who have learned to write would move objects faster in a left to right direction, opposed to a downward direction.
The present study first investigated how accurately and fast young children were able to use the mouse by asking them to click repeatedly on a very small target. The influence of distracting objects on the accuracy and speed of the children was investigated by presenting distracting objects at various positions and distances next to the small target. The computer recorded the reaction time and accuracy of the children. Secondly, children were asked to move objects (characters) with click-move-click and drag-and-drop over long and short distances, horizontally and vertically. The speed and the number of selection, interaction, and drop errors were recorded. The children also indicated which of the two procedures they favoured. The children would probably prefer drag-and-drop if this was the procedure they used by default. If they found it difficult to keep the mouse button pressed during drag-and-drop, they might favour click-move-click.
In the study by Inkpen et al. (1996b) girls made fewer errors during click-move-click than during drag-and-drop, whereas for boys these differences were much smaller. Boys possibly had more experience in using a mouse and therefore in using drag-and-drop techniques. Today, girls are likely to have as much experience as boys and therefore, the differences between drag-and-drop and click-move-click may have disappeared. However, because children in Kindergarten 2 may have less computer experience than children in Grade 1, they may use click-move-click more comfortably than drag-and-drop.
Dutch children from Kindergarten 2 (mean age 6 years and 0 months) and Grade 1 (mean age 7 years and 0 months) of three schools in the wide area of Amsterdam participated in this study. The children in Grade 1 had received at least six months of formal training in reading and writing. In total 104 children participated, 29 boys and 24 girls from Kindergarten 2, and 26 boys and 25 girls from Grade 1. Most of the children (N = 89) were right-handed. All but one of the left-handed children used the mouse with their right hand, without noticeable unwillingness or discomfort. This is also common for left-handed adult mouse users and does not influence their abilities negatively (Peters and Ivanoff, 1999 and Woods et al., 2003). One left-handed child used the mouse with her left hand, with the mouse still on the right side of the keyboard. She indicated that this was her usual way of using the mouse. Because her performance was well within the range of obtained data, we did not exclude her from the analyses.
The children used an infrared mouse with a control-display ratio of 1:4 (Jacob, 1996). The tasks were presented on a laptop computer with a 15-in. screen and a resolution of 640 × 480 pixels, resulting in pixels of 0.54 mm wide and tall. This paper will describe distances in pixels, the original unit of measurement, but for generalisations it may be useful to convert these to millimetres.
In the aim and click task three dots with a diameter of 6 pixels were presented on a yellow background in a horizontal or vertical array. Two of these dots were white, the other was red. The red dot could be presented at any of the three positions. The distance between the dots was 10 or 20 pixels. The children were asked to click on the red dot with the cursor, presented as a slanted arrow of 11 pixels wide and 19 pixels tall, similar to those used in typical graphical interfaces. When the children clicked, the dots disappeared and after a short interval (1 s) the next array of dots was presented at a randomly determined position at least 130 pixels horizontally and vertically from the previous position. The software did not reposition the cursor.
In the movement task the children performed an exercise taken from Leescircus (Reitsma, 1999), a software program that provides various exercises in beginning reading. The children were instructed to compare a word (for instance maak, make) with another word that was presented with a picture (for instance maan, moon), see Fig. 1. Because the words could be compared in a simple manner, the exercise was not considered to be too difficult for children in Kindergarten 2. The one letter in which the first word differed from the second, which was always the last letter of the word, had to be moved either horizontally or vertically to a waste bin, located at 149 pixels or 299 pixels from the centre of the letter. It was not possible to move any letter other than the altered letter. If the altered letter was moved to a place other than the waste bin, the computer returned it to its original position, a drop error was recorded, and the child had to start again. The trial continued until the child had moved the letter to the waste bin, at which time the next trial was presented. A new picture and new words were displayed and the waste bin was repositioned. The cursor in the program was the same arrow as in the aim and click task, but it changed to a pointing hand of 21 pixels tall and 19 pixels wide as soon as it was positioned on the letter that had to be moved. During movement, the cursor changed to a closed hand of 17 pixels tall and 19 pixels wide, attached to the letter. The software did not reposition the cursor at any moment. These cursor procedures are identical to those generally used in educational software for young children. The children were asked to perform this task using both the drag-and-drop and the click-move-click procedure.
Fig. 1. A screenshot of the movement task. Children have to move the “k” to the waste bin.
The children performed the tasks in a quiet room in their school. The order in which the three tasks (aim and click, drag-and-drop, and click-move-click) were presented was counterbalanced to avoid training-effects and effects of strain and fatigue. The children always performed the two movement tasks successively, but half of the children first used drag-and-drop and the other half click-move-click. The aim and click task was presented either as the first task or as the last task. One or two trial items and a short explanation preceded each task. The experimenter did not refer explicitly to either accuracy or speed requirements.
In the aim and click task children were asked to click on a small red dot that was presented together with two white dots. The position of the red dot, the orientation of the accompanying white dots, and the proximity of the dots varied to produce 12 different conditions. These conditions were counterbalanced per factor and presented again in the reverse order to avoid training effects. For each trial, the difference between the x-coordinate of the child’s click and the centre of the red dot determined horizontal accuracy and the difference between y-coordinates determined vertical accuracy. Reaction time was defined as the duration of time (in seconds) from the presentation of the red dot to the moment of the click. The data were subjected to a 3 (position) × 2 (orientation) × 2 (proximity) multivariate repeated-measures analysis of variance (MANOVA) with Gender and Grade as between-subject variables. Only the effects that included the proximity of the dots were reported. A one-way MANOVA was used to examine the differences between right-handed and left-handed children. Correlations between reaction time and horizontal and vertical accuracy were calculated to identify a possible trade-off between speed and precision.
In the movement tasks the children were asked to move letters to a waste bin. The direction of movement and the distance between the centre of the letter and the centre of the waste bin were combined to produce four conditions. These conditions were counterbalanced for the direction of movement and presented again in a different order to counterbalance for distance. The children completed the resulting eight trials first with one interaction procedure and after a short explanation with the other procedure. For each trial, the duration of time between the first time the children pressed the mouse button and the last time they released the mouse button defined the total time, the reaction time including the time spent on errors. Movement speed, the reaction time corrected for distance, was calculated per item by dividing the movement time (total time without time spent on errors) by the distance between the letter and the waste bin. Selection errors occurred when children did not accurately select the correct letter. When the mouse moved fewer than 25 pixels between mouse-down and mouse-up, this was identified as a mouse slip. When it moved more than 25 pixels it was classified as a drag-attempt. Drag-attempts were recorded as drop errors during drag-and-drop, but as interaction errors during click-move-click. Mouse slips were recorded as interaction errors during drag-and-drop and were regarded as part of a correct movement sequence during click-move-click. Drop errors during click-move-click were recorded when the children clicked to release the object outside the waste bin. After the completion of both movement tasks, the children indicated their procedure preference. The data were subjected to a 2 (procedures) × 2 (distance) × 2 (direction) multivariate repeated-measures analysis of variance (MANOVA) with Grade and Gender as between-subject variables. A one-way MANOVA was used to determine the differences between right-handed and left-handed children. Correlations between movement time and the number of errors were calculated to identify a possible trade-off between speed and accuracy.
The results indicate that an object needs to be only 12 pixels wide (6.5 mm) and 22 pixels tall (11.9 mm) in order to include 95% of the mouse-clicks of children in either Kindergarten 2 or Grade 1. Children clicked on average 2.1 pixels right or left from the centre of the red dot and 3.5 pixels above or below it. The difference between horizontal and vertical accuracy, t(103) = 9.45, p < 0.001, indicates that aiming was more difficult in a vertical than in a horizontal direction. The precision of children is remarkable, but it took quite some effort to attain this level of precision. They needed on average 2.95 s to click on the red dot. The mean speed and accuracy are presented in Table 1 as a function of the independent variables. Reaction time did not increase at the expense of speed or vice versa.
Descriptive statistics for the results of the aim and click task
Reaction time (s) Horizontal accuracy (pixels) Vertical accuracy (pixels) Mean 2.95 2.12 3.54 Kindergarten 2 3.33 2.21 3.60 Grade 1 2.58 2.01 3.28 Far distracting objects 2.84 2.21 3.62 Close distracting objects 3.07 2.01 3.26
Children in Grade 1 did not click more accurately than children in Kindergarten 2, suggesting that they considered the same rate of error acceptable. However, children in Grade 1 clicked faster, F(1,97) = 22.40, p < 0.001, which indicates that young children needed more time to realise this error rate. Differences between boys and girls were not significant, but there was a three-way interaction between Grade, Gender and Distance for reaction time, F(1,97) = 7.04, p < 0.01. In Kindergarten 2, boys clicked faster on the more easy items with the dots far apart than on items with the dots close together, the more difficult items; girls clicked equally slow on both. In Grade 1, boys clicked as fast on difficult items as on easy items; girls clicked slower on difficult items than on easy items, see Fig. 2. This suggests that the ability to click on difficult items develops for boys sooner than for girls. There were no effects of hand preference.
Fig. 2. The effect of group, gender and distance on reaction time.
As expected, placing dots close together resulted in longer reaction times, F(1,97) = 24.38, p < 0.001, higher horizontal accuracy, F(1,97) = 5.95, p < 0.05, and higher vertical accuracy, F(1,97) = 11.82, p < 0.01. An interaction between the position of the red dot and the distance between the dots, F(2,194) = 5.38, p < 0.01, indicated that the vertical accuracy for red dots at the left or top position was not affected by the distance between the dots, see Fig. 3. This suggests that the children searched the array of dots from left to right and from top to bottom, confirmed by the fact that both vertical accuracy, F(2,194) = 13.91, p < 0.001, and horizontal accuracy, F(2,194) = 7.65, p < 0.01, were higher when children clicked on dots at the middle, right, or bottom position than at the left or top position. The interaction between proximity and the position of the red dot for reaction time, F(2,194) = 3.20, p < 0.05, showed that, although the reaction time of children was longer for items with the distracting objects close to the target than for items with the distracting objects farther from the target, the children clicked equally fast on dots at the bottom or right position when the dots were close together as when they were far apart, see Fig. 3.
Fig. 3. The effect of the proximity of objects and the position of the red dot on reaction time and on vertical accuracy.
The children needed on average 3.2 s to move the altered letter to the waste bin, including time lost on errors. Their movement speed was 109 pixels per second. On average, the children made a selection error in 16% of the items, an interaction error in 19%, and a drop error in 5% of the items. However, because children could make more than one mistake per item, this does not mean that they made errors in 40% of the items. There was no trade-off between movement speed and the number of errors. The mean reaction time, movement speed and number of errors are presented in Table 2 as a function of the independent variables.
Descriptive statistics for the results of the movement tasks
Reaction time (s) Speed (pixels/s) Proportion of errors in Selection Dropping Interaction Mean 3.07 110.14 0.14 0.04 0.17 Kindergarten 2 3.75 98.05 0.20 0.07 0.25 Grade 1 2.63 120.59 0.11 0.03 0.13 Drag-and-drop 2.67 120.62 0.15 0.03 0.12 Click-move-click 3.71 98.02 0.17 0.07 0.27 Short distance 3.05 82.51 0.17 0.05 0.19 Long distance 3.32 136.13 0.15 0.05 0.20 Horizontal move 3.09 112.57 0.16 0.05 0.18 Vertical move 3.29 106.06 0.16 0.05 0.20
Children in Grade 1 made fewer selection errors, F(1,100) = 6.12, p < 0.05, interaction errors, F(1,100) = 6.29, p < 0.05, and drop errors, F(1,100) = 4.18, p < 0.05, than children in Kindergarten 2. Children in Grade 1 also moved the mouse faster than children in Kindergarten 2, F(1,100) = 20.92, p < 0.001, and needed less time to complete trials, F(1,100) = 15.05, p < 0.001. There were no significant differences between boys and girls or between left-handed and right-handed children.
As predicted, moving targets over long distances resulted in faster movement speeds than moving them over short distances, F(1,100) = 1294.49, p < 0.001. This effect was a little more pronounced for children in Grade 1 than for children in Kindergarten 2, F(1,100) = 10.89, p < 0.01. Long distances also took more time to complete, F(1,100) = 4.60, p < 0.05. An interaction between distance and direction showed that the difference in reaction time between long and short distances was only significant for items where the target had to be moved horizontally, F(1,100) = 14.36, p < 0.001. Moving objects horizontally generally resulted in faster movement speeds than moving them vertically, F(1,100) = 20.56, p < 0.001, but this effect appeared only for short distances, as indicated by the interaction between distance and direction on speed, F(1,100) = 12.61, p < 0.01, see Fig. 4. These effects suggest that horizontal movements are started more quickly than vertical movements and that vertical movements can be continued faster once the movement has been initiated, which may be related to the difficulty children had to aim vertically in the aim and click task. The interaction of distance and direction for the number of drop errors children made, F(1,100) = 4.17, p < 0.05, confirms this hypothesis. Children make more drop errors in a short vertical move than in a short horizontal move. For long moves, this effect is reversed, see Fig. 4.
Fig. 4. The effect of direction and distance on movement speed and on the number of drop errors.
Children did not make more drop errors when using drag-and-drop than click-move-click. Additionally, children needed less time to complete an item when using drag-and-drop than when they used click-move-click, F(1,100) = 26.83, p < 0.001. Drag-and-drop also resulted in higher movement speeds than click-move-click, F(1,100) = 57.72, p < 0.001, indicating that the difference was not due to the time the children spent on errors. That the effect of movement procedure on speed was more pronounced for long moves than for short moves, F(1,100) = 4.43, p < 0.05, suggests that the difference not only originates in the beginning and end of the move, but also in the speed during the movement.
More interaction errors occurred when children used click-move-click than when they used drag-and-drop, F(1,100) = 9.48, p < 0.01, an indication that drag-and-drop has become a routine for them. The distance of 25 pixels that was set for mouse slips was arbitrary, of course. Had we allowed for larger mouse slips, fewer slips would have been identified as interaction errors during click-move-click or as drop-errors during drag-and-drop. The results of the aim and click task, however, show that children rarely make mouse slips of more than 11 pixels, indicating that the criterion for mouse slips in this study was rather too lenient than too strict.
There was no difference between the two interaction procedures in the number of selection errors the children made, but there was an interaction between the movement procedure and the distance over which objects had to be moved, F(1,100) = 7.01, p < 0.01. When the object had to be moved over short distances, more selection errors were made using click-move-click than drag-and-drop. For long distance items, this effect was reversed. Fig. 5 shows a screenshot of the Movement Task where the positions of the selection errors are marked by black dots. Some of the children tried to click on the waste bin or the picture, probably out of curiosity, and one child clicked near the hand (bottom left) to indicate that he did not want to practice anymore, but most of the children who made selection errors found it difficult to select the letter. Some of the letters may have been too small. For instance, the “l” and the “i” were 32 pixels tall, but only four pixels wide. A one-way ANOVA did not reveal that some letters resulted in more selection errors than others, but the letters did affect reaction time, F(4,1691) = 3.38, p < 0.01, and movement speed, F(4,1691) = 2.86, p < 0.05. The “s” resulted in higher reaction times than the “f” and in higher movement speeds than the “k”, p’s < 0.05. This may have affected the number of selection errors, because the items that resulted in many selection errors, contained relatively few “s” targets and many “k” and “f” targets.
Fig. 5. A screenshot of the movement task. The selection errors made by the children are presented as black dots.
Among the children who were able to tell us which one of the two procedures they used most often (N = 42), the majority (N = 29) mentioned drag-and-drop, χ2(1) = 6.10, p < 0.05. A non-significant majority of the children who were able to indicate which procedure they preferred (N = 103), preferred click-move-click (N = 60), χ2(1) = 2.81, n.s. Children did not explicitly like or dislike the procedure that they used most often.
Research has indicated that input devices are important for the usability of education software, but empirical data in this area is rather limited (Hourcade, 2002 and Smith and Keep, 1986). Therefore, the aim of the present study was to investigate the suitability of the mouse as an input device for young children. The results clearly showed that young children are capable of using the mouse to aim precisely and to move objects, even though they need quite some time to complete some of these tasks.
Surprisingly, children in Kindergarten 2 clicked as accurately on small objects as children in Grade 1 did: a large majority of the children was able to accurately select an object of 7 mm wide and 12 mm tall. As a comparison, when Microsoft Word is presented on a 17-in. screen with a resolution of 1024 × 768 pixels, the save-button is 6 mm wide and long. Even though the children in K2 and G1 were equally accurate, the children in Kindergarten 2 needed more time to aim and click than the older children in Grade 1. This suggests that children in K2 and G1 considered the same rate of error acceptable, but that the children in K2 had to try hard to reach this rate of error (Elliott et al., 2004). In the movement task, children in Kindergarten 2 not only needed more time to complete the moves, but they also made more errors than children in Grade 1. This suggests that children in K2 considered a higher rate of error acceptable than children in G1 or that the children in K2 were unable to achieve the rate of error that was generally considered acceptable.
A trade-off between speed and accuracy was present in neither of the tasks. The trade-off may have been obscured by large differences in motor skills between children. Alternatively, the design of the task may have caused all children to value accuracy more than speed: whereas the small size of the dots and the presence of distracting objects encouraged them to be accurate, they were not stimulated to perform the tasks quickly. In contrast to the tasks in this study, the games that children play at home often focus on speed, and so do some educational programs. In software that focuses on speed, the objects may need to be somewhat larger, because children will want to spend less time on aiming precisely. Nevertheless, the present study clearly indicated that if speed is not required, the mouse can be used to aim and click quite accurately, even on very small objects, by children both in Kindergarten 2 and Grade 1.
In the aim and click task, the vertical accuracy of children was lower than their horizontal accuracy, possibly because the arrow that was used as a cursor was slanted. However, children presumably had sufficient experience to know how to use this cursor (Phillips et al., 2001). Alternatively, vertical aiming may have drawn more heavily on motor skills than horizontal aiming. Yet another explanation might be that children tended to move the mouse vertically during their mouse press (Crook, 1992). The results from the movement task shed some light on this issue. Over short distances children move objects faster horizontally than vertically, suggesting difficulties in moving the mouse vertically. However, over longer distances they moved objects faster vertically than horizontally, suggesting that horizontal moves are more easily initiated, but slower continued than vertical moves. This also explains the difficulties children experienced in vertical aiming, because aiming consists of making small movements with the mouse (Dennerlein & Yang, 2001). The difference between initiating and continuing a move might be physical. To move the mouse over a short distance, the wrist is used and horizontal moves may be made more comfortably by the wrist than vertical moves, as reflected by the problems children had with vertical aiming. Moving the mouse over a long distance requires the use of the whole arm and this may be easier vertically than horizontally.
These differences were also reflected in the number of drop errors children made. Children made more drop errors in short vertical moves than in short horizontal moves, and in long horizontal moves than in long vertical moves. This implies that vertical drop errors are made close to the original position of the target or to the release-spot, whereas horizontal drop errors are made during the move. However, to investigate the cause of drop errors further, more data on drop errors should be recorded while children move objects over diverse distances and directions. This will also provide more information on the suitability of movement directions and distances for young children.
In previous research, it was suggested that click-move-click is a more appropriate procedure for young children to move objects than drag-and-drop (Inkpen et al., 1996b and Joiner et al., 1998). However, in the studies by Inkpen, Booth, and Klawe, and by Joiner et al., the advantage of drag-and-drop may have been due to cursor size or error handling. In the present study, the only difference between the two conditions was the movement procedure. The data show that drag-and-drop is faster than click-move-click and that this was not only due to different speeds of picking objects up and releasing them, but also to the speed of movement itself. Also, drag-and-drop caused fewer interaction errors than click-move-click.
Drag-and-drop is supposed to be difficult because it would be motorically demanding to keep the mouse button pressed during movement (MacKenzie et al., 1991 and Strommen, 1993). However, in the present study there were no differences between the two movement procedures in the number of drop errors children made. Keeping pressure on the mouse button during movement does not seem to be difficult for children. However, the criterion for mouse slips in the present study may have been too lenient. Had we allowed for smaller mouse slips, more errors during drag-and-drop that were presently identified as interaction errors would have been identified as drop errors. The cause of drop errors should be examined in a future study, but the present study does not support the hypothesis that they are caused by premature releases of the mouse button. Children do not appear to make more drop errors when they drag objects over long distances than when they drag objects over short distances. Generally, the results indicate that drag-and-drop is a more suitable movement procedure for young children than click-move-click.
The present study has provided the valuable insight that children are able to click precisely on small objects, especially when they are presented within a context of other objects. However, the target in the click task was much smaller than objects that are generally presented in educational software for children and this probably influenced the rate of error children consider as acceptable. If children have to click on larger objects, they may intuitively aim faster and less accurately. The number of selection errors the children made when they had to move characters demonstrated this. The letters probably suggested a larger active area than was actually present. However, the movement task was not designed to study the effect of letter-size on the number of selection errors. Therefore, a future study could investigate how accurately children click on objects of different shapes and sizes.
When children are required to click on large objects, the influence of distracting objects is probably smaller. In the present study, the effect of distracting objects indicated that children view the objects in the direction in which they read and write. However, these effects were small, possibly because accuracy was already high due to the small size of the objects. Distracting objects may be especially useful when children are inclined to click less accurately than necessary. An issue of research is therefore whether the accuracy of children also increases as a result of distracting objects when the objects are of a more realistic size.
The present study investigated mouse skills because the mouse is a popular input device. In the near future new input devices for children may be introduced in education. For instance, the game controller is a relatively new device that combines the advantages of both direct and indirect devices (Rosas et al., 2003 and Strommen, 1993). A more recent development is “nouse” technology, that allows a webcam to translate the movement of the nose of users into cursor movements (Gorodnichy & Roth, 2004). These devices may be more comfortable to use for young children than the mouse, but they should be the object of a study similar to the present study before they are used in educational software.
conclusion, the present shows that young children are clearly capable
of using a computer mouse. It is evident that they can click very
accurately on targets of 7 mm wide and 12 mm tall, even
though they need a lot of time to aim on objects. This indicates that
objects in educational software do not have to be much larger than
objects in software for adults, unless children are required to respond
quickly. Also, drag-and-drop is concluded to be a more suitable
procedure for young children to move objects than click-move-click with
the reservation that the number of drop errors in the present study may
have been affected by the definition of mouse slips and drag errors.
thank Femke Aalst for help in collecting the data and Pim Bongers for
his programming assistance. Inez Berends, Lotte Hoekstra, Bart Follink,
and Stuart Woods are thanked for their comments on earlier versions of
Crook, 1992 C. Crook, Young children’s skill in using a mouse to control a graphical computer interface, Computers and Education 19 (1992) (3), pp. 199–207. Abstract | Abstract + References | PDF (909 K)
Dennerlein and Yang, 2001 J.T. Dennerlein and M.C. Yang, Haptic force-feedback devices for the office computer: performance and musculoskeletal loading issues, Human Factors 43 (2001) (2), pp. 278–286. Full Text via CrossRef | Abstract + References in Scopus | Cited By in Scopus
Elliott et al., 2004 D. Elliott, S. Hansen, J. Mendoza and L. Tremblay, Leaning to optimize speed, accuracy, and energy expenditure: a framework for understanding speed-accuracy relations in goal-directed aiming, Journal of Motor Behavior 36 (2004) (3), pp. 339–351. Abstract + References in Scopus | Cited By in Scopus
Gillan et al., 1990 D.J. Gillan, K. Holden, S. Adam, M. Rudisill and L. Magee, How does Fitts’ law fit pointing and dragging?. In: J. Carrasco Chew and J. Whiteside, Editors, Proceedings of CHI’90 human factors in computing systems: Empowering people, ACM, New York, NY (1990), pp. 227–234. Full Text via CrossRef
Gorodnichy and Roth, 2004 D.O. Gorodnichy and G. Roth, Nouse ‘use your nose as a mouse’ perceptual vision technology for hands-free games and interfaces, Image and Vision Computing 22 (2004), pp. 931–942. SummaryPlus | Full Text + Links | PDF (625 K) | Abstract + References in Scopus | Cited By in Scopus
Heckhausen and Beckmann, 1990 H. Heckhausen and J. Beckmann, Intentional action and action slips, Psychological Review 97 (1990) (1), pp. 36–48. Abstract | Abstract + References | PDF (1527 K) | Full Text via CrossRef | Abstract + References in Scopus | Cited By in Scopus
Hourcade, 2002 Hourcade, J.P. (2002). It’s too small! Implications of children’s developing motor skills on graphical user interfaces. Technical Report No. CS-TR-4425, Department of Computer Science, University of Maryland, College Park, MD. Available online www.cs.umd.edu/Library/TRs/CS-TR-4425/CS-TR-4425.pdf.
Inkpen, 1997 K. Inkpen, Three important research agendas for educational multimedia: learning, children and gender. In: T. Muldner and T.C. Reeves, Editors, Proceedings of Ed-Media’97, AACE, Charlottesville, VA (1997), pp. 521–526 Available onlinewww.cs.sfu.ca/~inkpen/Papers/Ed-media_97.
Inkpen et al., 1996a Inkpen, K., Booth, K. S., & Klawe, M. (1996a). Interaction styles for educational computer environments: a comparison of drag-and-drop vs. point-and-click. Technical Report No. 96-17, Department of Computer Science, University of British Columbia. Available online www.cs.sfu.ca/~inkpen/Papers/TR17_DD.
Inkpen et al., 1996b Inkpen, K., Booth, K. S., & Klawe, M. (1996b). Drag-and-drop vs. point-and-click mouse interaction for children. Technical Report No. 96-20, Department of Computer Science, University of British Columbia. Available online www.cs.sfu.ca/~inkpen/Papers/TR20_DD.
Joiner et al., 1998 R. Joiner, D. Messer, P. Light and K. Littleton, It is best to point for young children: a comparison of children’s pointing and dragging, Computers in Human Behavior 14 (1998) (3), pp. 513–529. SummaryPlus | Full Text + Links | PDF (803 K) | Abstract + References in Scopus | Cited By in Scopus
Kuhtz-Buschbeck et al., 1999 J.P. Kuhtz-Buschbeck, A. Boczek-Funcke, M. Illert, K. Joehnk and H. Stolze, Prehension movements and motor development in children, Experimental Brain Research 128 (1999), pp. 65–68. Full Text via CrossRef | Abstract + References in Scopus | Cited By in Scopus
Lindemann and Wright, 1998 P.G. Lindemann and C.E. Wright, Skill acquisition and plans for actions: learning to write with your other hand. In: D. Scarborough and S. Sternberg, Editors, An invitation to cognitive science: Methods, models, and conceptual issues, MIT Press, Cambridge, MS (1998).
MacKenzie et al., 1991 I.S. MacKenzie, A. Sellen and W. Buxton, A comparison of input devices in elemental pointing and dragging tasks. In: S.P. Robertson, G.M. Olson and J.S. Olson, Editors, Proceedings of CHI’91 human factors in computing systems: Reaching through technology, ACM, New York, NY (1991), pp. 161–166. Full Text via CrossRef
Peters and Ivanoff, 1999 M. Peters and J. Ivanoff, Performance asymmetries in computer mouse control of right-handers, and left-handers with left- and right-handed mouse experience, Journal of Motor Behavior 31 (1999), pp. 86–94. Abstract + References in Scopus | Cited By in Scopus
Phillips and Triggs, 2001 J.G. Phillips and T.J. Triggs, Characteristics of cursor trajectories controlled by the computer mouse, Ergonomics 44 (2001) (5), pp. 527–536. Full Text via CrossRef | Abstract + References in Scopus | Cited By in Scopus
Phillips et al., 2001 J.G. Phillips, T.G. Triggs and J.W. Meehan, Cursor orientation and computer screen positioning movements, Human Factors 43 (2001) (3), pp. 435–441. Full Text via CrossRef | Abstract + References in Scopus | Cited By in Scopus
Reitsma, 1999 P. Reitsma, Leescircus [Circus of reading], ThiemeMeulenhoff, Utrecht, The Netherlands (1999).
Rosas et al., 2003 R. Rosas, M. Nussbaum, P. Cumsille, V. Marianov, M. Correa and P. Flores et al., Beyond Nintendo: design and assessment of educational video games for first and second grade students, Computers and Education 40 (2003), pp. 71–94. SummaryPlus | Full Text + Links | PDF (341 K) | Abstract + References in Scopus | Cited By in Scopus
Segers and Verhoeven, 2002 E. Segers and L. Verhoeven, Multimedia support of early literacy learning, Computers and Education 39 (2002), pp. 207–221. SummaryPlus | Full Text + Links | PDF (966 K) | Abstract + References in Scopus | Cited By in Scopus
Smith and Keep, 1986 D. Smith and R. Keep, Children’s opinions of educational software, Educational Research 28 (1986) (2), pp. 83–88.
Sternberg, 1996 R.J. Sternberg, Cognitive psychology, Harcourt Brace College Publishers, Orlando (1996).
Strommen, 1993 E.F. Strommen, Is it easier to hop or walk. Development issues in interface design, Human–Computer Interaction 8 (1993), pp. 337–352. Full Text via CrossRef | Abstract + References in Scopus | Cited By in Scopus
Tränkle and Deutschmann, 1991 U. Tränkle and D. Deutschmann, Factors influencing speed and precision of cursor positioning using a mouse, Ergonomics 34 (1991) (2), pp. 161–174. Abstract + References in Scopus | Cited By in Scopus
Walker et al., 1993 N. Walker, D.E. Meyer and J.B. Smelcer, Spatial and temporal characteristics of rapid cursor-positioning movements with electromechanical mice in human–computer interaction, Human Factors 35 (1993) (3), pp. 431–458. Abstract + References in Scopus | Cited By in Scopus
Whisenand and Emurian, 1999 T.G. Whisenand and H.H. Emurian, Analysis of cursor movements with a mouse, Computers in Human Behavior 15 (1999), pp. 85–103. SummaryPlus | Full Text + Links | PDF (373 K) | Abstract + References in Scopus | Cited By in Scopus
Whitefield, 1986 A. Whitefield, Human factors aspects of pointing as an input technique in interactive computer systems, Applied Ergonomics 17 (1986) (2), pp. 97–104. Abstract | Abstract + References | PDF (1127 K) | Abstract + References in Scopus | Cited By in Scopus
Wilton and McLean, 1984 J.A. Wilton and R.S. McLean, Evaluation of a mouse as an educational pointing device, Computers and Education 8 (1984) (4), pp. 455–461. Abstract | Abstract + References | PDF (591 K)
Woods et al., 2003 V. Woods, S. Hastings, P. Buckle and R. Haslam, Development of non-keyboard input device checklists through assessments, Applied Ergonomics 34 (2003), pp. 511–519. SummaryPlus | Full Text + Links | PDF (372 K)
| Abstract + References in Scopus | Cited By in Scopus
Volume 48, Issue 4 , May 2007, Pages 602-617