1 Introduction

Touchscreen technology has gained increasing popularity over the last decade due to its natural and convenient human-technology interaction. It is widely used in a variety of personal, public and occupational settings [1, 2], including varied complex information systems implemented in healthcare facilities, aviation industry, and nuclear power stations [3, 4].

The use of touchscreen technology can bring a number of advantages. First, touchscreen can be easily accessed and operated by a wide range of users, including inexperienced and disable user [5,6,7,8]. Second, touchscreen can reduce physical dimensions of a device, as physical keys that are usually applied in traditional input devices such as keyboards or mice can be replaced by on-screen virtual keys [7, 9, 10]. Third, it appears that the intuitive human-technology interaction provided by touchscreen can make the technology more attractive [6]. Finally, the design of touchscreen interface can be easily adjusted to provide only the keys that are relevant in a specific task at a given time (e.g., providing digit keys only in digit input tasks) [11].

In spite of its convenience and potential benefits, touchscreen interfaces, if poorly designed, are likely to result in frustration and irritation for users, in inefficiency and disruption in work process, and in a higher likelihood of committing errors [12, 13]. These negative outcomes might rise safety issues to both systems and users [7]. Therefore, it is important to investigate the effects of interface design factors that may affect the use of the technology so as to provide usable and safe touchscreen interfaces.

Previous studies have addressed several important design factors for touchscreen interfaces, such as key size and gap between keys [11, 14,15,16,17,18,19]. For example, Pfauth and Priest suggested that key size is one of the most important factors in touchscreen use when the touchscreen interface involved a hierarchical menu display [20]. Chen et al. investigated the effects of key size and gap size on touchscreen input performance by individuals with varied motor abilities [14]. Their results indicated that as key size increased, the performance for disable participants improved, while the performance for non-disable group plateaued at a key size of 20 mm. Chourasia et al. evaluated the effect of posture (i.e., sitting and standing) on touchscreen performance and touch characteristics during a digit entry touchscreen task among individuals with and without motor-control disabilities [15]. They found that standing affected touchscreen performance only at smaller key sizes (i.e., key sizes smaller than 20 mm) and led to greater exerted force and impulse compared with sitting. Colle and Hiszem investigated effects of key size (i.e., 10 mm, 15 mm, 20 mm and 25 mm) and gap size (i.e., 1 mm and 3 mm) on touchscreen numeric keypad performance [11]. They found that task input time was longer and error rate was higher for smaller key sizes, while the performance plateaued as key size increased up to 20 mm. Jin et al. examined a number of key sizes (i.e., 11.43 mm, 13.97 mm, 16.51 mm, 19.05 mm, 21.59 mm, and 24.13 mm) and gap size (i.e., 0 mm, 3.17 mm, 6.35 mm, 12.7 mm and 19.05 mm) for touchscreen interfaces with older adults [19]. Their results indicated that 19.05 mm size yielded the highest accuracy rate. In addition, these studies consistently found that gap size did not affect user performance [11, 14, 15, 19]. However, they did not consider scenarios where there is no gap between keys.

Although key size and gap between keys have been widely examined [14, 15], consensus on design guidelines for the two factors seems lacking [14]. For instance, International Organization for Standardization (ISO) recommends that the size of the touch-sensitive area should be at least equal to the breadth of the index finger distal joint for the ninety-fifth percentile male, while the Electronic Industries Association (EIA) recommends 19.05 mm as the minimal touch-sensitive size. Moreover, the America National Standards Institute (ANSI) suggests a minimal key size of 9.5 mm with a 3.2 mm gap, and states that key sizes larger than 22 mm lead to no performance improvement.

Besides key size and gap, there exist other important factors (e.g., the location of key characters on keys) that may affect the usability of touchscreen but have not been previously investigated. Theoretically, key characters could be presented in any location on the key area. In practice, the location of key characters vary in different types of keys. For example, it is widely encountered in traditional physical keyboards that key characters located in upper left area for letter keys, in upper left, central or lower right area for some function keys, and in upper or lower left for numeric keys. However, there seems to have few design guidelines to guide the design practice for the location of key characters. In addition, the location of key characters is likely to provide users for behavioral cues, especially in touchscreen input tasks. It is possible that users may hit the character on a key directly rather than other areas of the key when they are required to click the key. Hence, the location of key characters may affect a user’s decision on the area that their fingertips would touch on the keys, and thereby affect the input accuracy. Moreover, users’ hands and fingers could cover key characters when they touch the keys. The degree to which key characters are covered may be different as the characters are located in varied locations within the key area. For example, key characters in lower right could be fully covered when users’ fingers get close to the touchscreen interface, while key characters in upper left could be well seem until users’ fingers press the interface. Therefore, the location of key characters may have important impacts on the usability of touchscreen devices in input tasks; but this speculation requires further confirmation.

The purpose of this study was to examine the effects of key size, gap and the location of key characters on the usability of touchscreen devices. While many previous studies focused on small, mobile devices [21,22,23] and examined digit input tasks [14, 15], we based our study on a large touchscreen interface and examined letter input tasks.

2 Methods

2.1 Experimental Design

A three-factor (4 × 2 × 5), within-subjects design was employed in this study, with key size (i.e., 10 mm, 15 mm, 20 mm and 25 mm), gap (presence (i.e., 2 mm) and absence), and the location of key characters (upper left, upper right, central, lower left and lower right) serving as independent variables, and sets of usability metrics (i.e., task completion time, accuracy rate and user preference) serving as dependent variables. Task completion time referred to the amount of time a participant needed to complete a task. Accuracy rate was calculated as the proportion of correctly input characters in a task. User preference was assessed using a paper-based questionnaire that asked participants to choose their preferred button design from varied levels of the three examined factors.

2.2 Participants

Fourteen undergraduate students (8 male and 6 female), aged from 21 to 23 years, participated in this study. All participants were right-handed and reported having normal or corrected to normal vision. The study protocol was approved by the institutional review board of Shenzhen University.

2.3 Materials and Tasks

A DELL All-In-One touchscreen computer (screen size: 23 in., resolution: 1600 × 900) used to present our experimental tasks. The touchscreen was titled back with a 70-degree angle from the horizontal level, as suggested by previous studies [14, 15]. Task scenarios were created by Visual Studio 2012 in a Microsoft Foundation Classes operating system. The experimental interface contained a target box, an input box and an experimental keyboard (See Fig. 1 for an example) with square keys. White was applied to the background of the touchscreen keyboard interface, with grey for key characters. The task required the participants to enter a random six-letter string shown in the target box using the experimental keyboard.

Fig. 1.
figure 1

Screenshot for an experimental touchscreen interface (Experimental condition: 20 mm key size, the absence of gap and key characters located in the upper left area. English words in parentheses are used for explanation only and would not show in the test).

Full size image

2.4 Procedures

Before the experiment, participants provided informed consent and were given detailed information of test procedures. They were instructed to sit and adjust the chair according to their preference. Following several practice tasks to familiarize themselves with the test, participants were asked to click a start button on the center of the screen to initiate the main experimental tasks. Participants were asked to respond with their index fingers as quickly and accurately as possible. The combinations of key size, gap and the location of key characters were randomized in a full factorial design. There were three letter tasks for each of the combinations. Upon the completion of all tasks, the preference questionnaire was administered to elicit participants’ preference on button design. The whole experiment could be completed within half hour.

2.5 Data Analysis

Three-way repeated measures analyses of variance (ANOVAs) were used to determine the main and interaction effects of key size, gap and the location of key characters on user performance. Past hoc analyses were performed with Bonferroni adjustment where necessary. Chi-square test was performed to examine the difference in user preference. Level of significance was set at ɑ = 0.05. Statistical analyses were performed using SPSS Version 22.

3 Results

3.1 Task Completion Time

Table 1 presents ANOVA analysis results for task completion time. There was a significant main effect of key size on task completion time while gap and the location of key characters did not yield any effect. On average, the task completion decreased by 8% as the key size increased from 10 mm to 20 mm. The location of key characters had a marginal interaction effect with key size (F (12, 156) = 1.729, p = 0.065) (Fig. 2), but not with gap (Fig. 3).

Table 1. Effects of key size, gap and the location of key characters in task completion time (s).
Full size table
Fig. 2.
figure 2

Task completion time (s) by key size and the location of key characters.

Full size image
Fig. 3.
figure 3

Task completion time (s) by gap and the location of key characters.

Full size image

3.2 Accuracy Rate

Table 2 presents ANOVA analysis results for accuracy rate. Key size was found to have a significant effect on accuracy rate while gap and the location of key characters alone did not yield any effect. On average, the accuracy rate increased from 95.2% to 97.6% as the key size increased from 10 mm to 25 mm. Accuracy rate plateaued at 15 mm key size with little improvement for larger key sizes. The location of key characters had a significant interaction effect with key size (F (4, 52) = 2.634, p = 0.044) (Fig. 4), but not with gap (Fig. 5).

Table 2. Effects of key size, gap and the location of key characters on accuracy rate (%).
Full size table
Fig. 4.
figure 4

Accuracy rate (%) by key size and the location of key characters.

Full size image
Fig. 5.
figure 5

Accuracy rate (%) by gap and the location of key characters.

Full size image

3.3 User Preference

Table 3 shows the user preference data on key size, gap and the location of key characters. The majority of participants preferred 15 mm key size (64%, χ2 = 16.286, p = 0.010), the presence of gap (57%, χ2 = 0.286, p = 0.900), and centrally located key characters (71%, χ2 = 27.429, p = 0.001).

Table 3. Distribution of user preference by key size, gap and the location of key characters.
Full size table

4 Discussion

This study examined the effects of three touchscreen design factors (i.e., key size, gap and the location of key characters) on user performance and preference during letter input tasks. In general, key size yielded a significant effect on touchscreen input performance, while gap and the location of key characters alone had no measurable effect. User performance improved as the key sizes increased up to 15 mm. The location of key characters was found to marginally interact with button size in terms of both task completion and accuracy rate. Users generally preferred 15 mm key size, the presence of gap between keys, and key characters that were centrally located.

4.1 Effects of Key Size

The results indicate that user performance improved as key size increased up to 15 mm. Little benefit could be obtained in larger key sizes. The value is within the recommended range of 15–20 mm for minimal usable touchscreen button sizes that are reported in previous studies [11, 14, 15, 18, 19]. However, it should be noted that the recommended range is relative broad, suggesting that variations exist, and the context of use should be considered in the touchscreen interface design. For example, Chourasia et al. found that while 15 mm button size was sufficient to achieve optimal performance during sitting, 20 mm button size was required during standing [15]. Chen et al. found that healthy adults could only get minimal gains from button sizes larger than 20 mm, while disabled adults continued to obtain benefits as button size increased above 20 mm [14].

4.2 Effects of Gap

Previous studies reported no effect of gap size on user performance [11, 14, 15, 19]. However, these studies only examined the presence of gap size (e.g., 3 and 5 mm in study by Chourasia et al. [15], and 1 and 3 mm in study by Chen et al. [14]), and ignored the absence condition (i.e., no gap between keys). Our study extended previous research by examining both presence (i.e., 2 mm) and absence of gap. Results indicated that the absence of gap size did not affect user performance. One explanation for this may be that the difference between the presence and absence of gap manipulated in our study was small so that no effect was detected. This finding has important implication in that gap between keys could be removed to obtain a minimal touchscreen keyboard area, especially in cases where the touchscreen interface is limited. However, it should be noted that users generally preferred touchscreen keyboards with the presence of gap.

4.3 Effects of the Location of Key Characters

Another important contribution of the present work to the literature is that we provided empirical evidence on the effects of the location of key characters. In particular, our study examined five commonly encountered locations of key characters (i.e., upper left, upper right, central, lower left and lower right). Although it is assumed that the location of key characters might provide behavioral cues for users and affect input performance, we found no effect of the location of key characters on touchscreen input performance. One reason for this could be that most of our touchscreen keys were large enough so that the participants are unlikely to hit the area outside the key scope. We found that the location of key characters interacted with key size in terms of accuracy rate. The accuracy rate was lower if the keys were located in the upper and lower left in the key area for the smallest key size (i.e., 10 mm). Another explanation may be that the key characters in our study were also large enough so that key characters could not be fully covered by users’ hands and fingers wherever the key characters was located. Users could always see the characters in varied location conditions and obtain behavioral cues. However, it should be noted that participants generally preferred centrally located key characters, which could be applied in design practice to improve user satisfaction.

4.4 Implications and Future Directions

The results of this study lead to the following implications for the design of touchscreen interfaces. First, a key size of 15 mm could be recommended as the minimal usable option across gap and the location of key characters. However, this value came from healthy young adults in a static, sitting posture. Whether it could be an equally optimal size for other user groups and across varied contexts of use is unknown and requires further confirmation. Second, gap between keys, though yielded no measurable effect on user performance, could be an important design factor in touchscreen interface design. Removing the gap between keys could save space for limited touchscreen interfaces and would not lower user performance, while the presence of gap is likely to increase user satisfaction. Finally, our results did not indicate an optimal location of key characters in user performance. In this case, a centrally located key character could be a better option as it was preferred most by users. However, more research efforts are required to examine the effects of the location of key characters in relation to other factors, such as tasks and use scenarios. In general, specific design guidelines regarding the examined factors in our study are lacking and should be established from future empirical evidence.

5 Conclusions

This study examined the effects of key size, gap and the location of key characters on touchscreen use. In general, key size had a significant effect on touchscreen input performance. The performance improved as the key sizes increased and plateaued at 15 mm. The presence of gap and the location of key characters, though having no measurable effect alone, affected the use of touchscreen interface through their interaction effects with key size and with each other. The factors examined in this study require further examination to determine their effects in a variety of touchscreen usage scenarios. The results may help with the design of more usable and safe touchscreen technology.