Underlying functions associated with keyboarding performance of elementary-school students

Abstract

Background

Keyboarding (Typing) is a major writing mode in educational settings in addition to, or as an alternative to, handwriting. Therefore, it is important that occupational therapists become experts on this activity, to support students’ performance. Yet, the knowledge of keyboarding performance of elementary-school students, and the underlying functions it entails, is limited.

Aim

To compare keyboarding performance (speed and accuracy) of 4^th-grade students in copying and dictation keyboarding tasks, and to examine the role of underlying functions in predicting keyboarding performance.

Material and Methods

The sample consisted of 57 4^th-grade students, recruited from 2 elementary schools. Students were tested for reading speed, attention shifting, fine-motor skills, kinaesthetic awareness, and keyboarding performance.

Results

Keyboarding performance differed in the copying and dictation tasks. Reading speed was the major underlying function predicting keyboarding performance in both tasks. Additionally, kinaesthetic awareness had a low, negative correlation with dictation accuracy.

Conclusions

When occupational therapists assess students’ keyboarding performance, they should use various tasks. Additionally, therapists should consider students’ reading speed and kinaesthetic awareness, as they may explain keyboarding performance. This knowledge may also support decision-making when considering keyboarding as an alternative writing mode for students with handwriting difficulties.

Keywords:

• Typing
• reading
• cognitive
• motor
• primary-school

Introduction

Handwriting is a major activity that is crucial for students’ development and performance [1], as well as their participation in different school activities. It enables them to express their knowledge, thoughts, and experiences while engaging in schoolwork such as note- or test-taking, writing essays, and stories. Therefore, students with handwriting difficulties are at risk for a limited engagement in participating in school activities that involve written assignments. This, in turn, may affect their grades and achievement [2,3]. For several decades, occupational therapists have been involved in treating students with handwriting difficulties, with the attempt to improve their handwriting performance or provide supports and accommodations [3,4]. One of the main solutions they offered students with handwriting difficulties was the use of computers via keyboarding as an alternative writing mode [5,6].

In the past few decades, there has been a great increase in the use of computers in educational settings [7], for school purposes, as well as other daily activities such search for data or engaging in social media [1,8]. This phenomenon, which has been exacerbated by the COVID-19 pandemic, highlighted new problems, such as students having difficulty becoming proficient keyboarders. Paradoxically, however, as opposed to handwriting which has been quite extensively researched (for reviews see [1,9,10]), the knowledge relating to the keyboarding process and factors that predict proficient keyboarding is scarce [11], limiting the ability of educators and occupational therapists to support students with poor keyboarding skills. To address this knowledge-gap, this study focussed on the keyboarding performance of fourth-grade students and the factors associated with keyboarding performance.

Keyboarding performance and associated factors

Only few studies addressed the keyboarding performance of elementary-school students. These studies indicated that, as opposed to handwriting, keyboarding is not systematically taught [6,12,13]. Additionally, using the Child version of the Assessment of Computer Task Performance (ACTP), which was developed to assess strengths and limitations of children (ages 5-11) in using assistive devices, Dumont and Mazer [14] have demonstrated that computer performance (both keyboarding and mouse use) improved with age. However, most studies examined students’ keyboarding performance in one task, usually copying or writing the alphabet from memory (i.e. the 'Alphabet task’; [15,16]). Yet, different writing tasks (e.g. copying or dictation) may place varied demands on the typists [14,17,18], and therefore, examining students’ keyboarding performance in more than one task may yield different results, and expand the knowledge of students’ performance and the functions underlying keyboarding.

Additionally, the theoretical model describing the keyboarding process suggests that it requires the orchestration of linguistic, cognitive, and sensory-motor functions [19,20]. Preliminary studies have supported this postulation. A recent study, for example [21] compared higher-education students with specific learning disorders, including handwriting disorders, who did and did not have keyboarding difficulties. They reported that the students who also had keyboarding difficulties had lower scores in phonological decoding and morpho-orthographic knowledge, compared to their peers without keyboarding difficulties. Among elementary-school students, only a handful of studies examined linguistic processes related to keyboarding, and most based their findings on the 'Alphabet task’. Overall, their findings implied that keyboarding abilities of students with handwriting difficulties were similar to those of normally achieving students, but those with spelling difficulties typed the alphabet slower [11,16]. In a later study, Ben-Moyal-Zada [22] found that reading speed was associated with keyboarding speed among fourth-grade students, as measured in a paragraph- copying task.

The keyboarding process also requires cognitive skills, specifically attention. Skilled typists direct their visual attention to the screen to monitor their typed text, but also shift their attention to the keyboard when necessary [13,20,23]. In contrast, non-skilled typists, mainly gaze at the keyboard to find the required keys and control their finger movements, and only occasionally shift their attention to the screen to monitor their keyboarding. Yet, a study examining this hypothesis showed that among fourth-grade typists there were no significant correlations between attention shifting and keyboarding speed, as compared to sixth-grade students, where a significant correlation between these variables was reported [22]. Nevertheless, data on this issue is limited and requires further investigation.

Analysis of the keyboarding process further suggests that it requires sensory-motor functions, yet the empirical findings are inconclusive. Researchers reported significant correlations between fine-motor skills and keyboarding speed among fourth-grade students [22] and keyboarding accuracy among fifth-grade students [24]. On the other hand, among young adults, Rosenberg-Adler and Weintraub [21] and Weintraub et al. [25] did not find significant correlations between fine-motor scores and keyboarding speed. Similarly, although most researchers agree that kinaesthetic awareness plays a role in the keyboarding process, especially among touch-typists [20,26], surprisingly, the evidence relating to the association between these factors is insufficient. Neither Berninger et al. [27], among elementary-school students, nor Weintraub et al. [25], among adults, found a significant correlation between performance on the Finger Succession test (which measures kinaesthetic awareness) and keyboarding speed. By contrast, Preminger et al. [24] reported a significant correlation between kinaesthetic ability and keyboarding accuracy among fifth-grade students. The limited and equivocal results regarding the different underlying functions of keyboarding performance calls for further investigation, to enhance the understanding of keyboarding performance.

An additional factor that may be associated with keyboarding performance is students’ gender. Handwriting studies have indicated clear gender differences in speed and legibility [1,28]. The shared underlying functions of handwriting and keyboarding [21,24] suggests that gender may also be related to keyboarding performance. Nevertheless, only a few studies examined the gender effect on keyboarding. In these studies, no significant associations were noted between gender and keyboarding speed both among higher education [21,29] and school-aged students [22]. By contrast, other researchers did find a gender effect [30,31]. Thus, this issue also requires further attention.

This study had two objectives: (a) To compare elementary-school students’ keyboarding performance (speed and accuracy) in two tasks: copying and keyboarding to dictation; and (b) to examine the relationship between underlying functions (reading speed, attention shifting, fine-motor skills, and kinaesthetic awareness) and keyboarding performance in the two tasks. We selected these two keyboarding tasks because they represent different common school activities, each placing different demands on the typists [17,30], and thus, may require different underlying functions. Moreover, whereas a copying task is commonly used in keyboarding studies, the use of dictation for examining keyboarding performance is rare. Based on the limited existing evidence, we hypothesised that each of the underlying functions will be associated with keyboarding speed, accuracy or both, and will have a unique contribution in predicting these performance measures. We could not, however, predict the effect of gender on keyboarding performance.

Materials and methods

Design and participants

This study followed a cross-sectional and correlational design, employing a convenience sampling method. The sample consisted of 57 fourth-grade students, who were recruited from two elementary schools in central Israel (moderate socioeconomic status). Students were included if their parents gave written consent and the students gave written assent to participate. Students were excluded from the study if they: (a) had learning or attention disorder, (b) were receiving special education services, and (c) scored at the 10^th percentile or lower on the Raven’s Coloured Progressive Matrices test (CPM) [32]. Eighty (52.6%) parents of the 152 students in the cohort consented to their child’s participation in the study, and their children assented. From these, 19 did not meet the study’s criteria and 4 were absent during the data collection days. The 57 participants were between the ages of 9-10-years-old (Mean age = 9.21, SD = .41), and 33 were female (57.9%).

Instruments

Background Questionnaire

This is a self-report questionnaire, developed for research purposes that served to gather information on students’ age, gender, mother tongue. Additionally, students were asked if they had a computer at home and if they knew how to touch-type.

Raven’s Coloured Progressive Matrices (CPM)

The CPM is a standardised test designed to measure nonverbal general intelligence in children aged 5-11-years-old [32]. It includes 36 items presented in three sets, in increasing difficulty. Each item is scored dichotomously; as correct/incorrect. Raw scores are transformed into percentile scores. The CPM has high internal consistency (α = .80-.94) and test-retest reliability (r = .82, p < .01) [33].

Aleph-Taph

This is a norm-referenced test for elementary-school students, that assesses reading, verbal memory, and linguistic skills in Hebrew [34]. In this study, we administered the reading text test (a paragraph consisting of 100 words), which was scored for speed (number of words read per minute). The Aleph-Taph has high internal consistency (α = .88) and reflects developmental trends of reading ability (i.e. significant increase in reading speed from 1^st to 6^th grades).

Children’s color trails test (CCTT)

The CCTT is a neuropsychological norm-referenced test designed to measure attention shifting and sustained attention among children and adolescents (8-16 years old). It includes two tasks (CCTT1 and CCTT2) [35]. In this study, we administered the CCTT2, which is more complex, and requires the student to connect numbers and colours, alternatingly. It is scored for total completion time which was found to have moderate test-retest reliability (r = .46-.68) in a sample of children with attention deficits (ADHD) [36]. Additionally, good discriminant validity between children with and without ADHD was reported [37].

Purdue pegboard test (PPT)

The PPT measures fine-motor skills (i.e. finger dexterity, and uni-manual and bi-manual hand coordination) [38]. It involves placing pegs into holes through four subtests (Dominant hand, Non-dominant hand, Both hands together, and Assembling). In this study, we used the 'Both hands’ subtest (both hands are used together to place pegs in hole in the same row). The score is calculated as the number of peg-pairs placed in the holes within a 30-sec period. Raw scores are converted to age-related Z scores. This test has medium test-retest reliability (.58 < r < .91) [38,39].

Finger Succession

This test examines fine-motor skills (i.e. sequential finger movement) and kinaesthetic awareness [40]. Students are asked to perform a sequence of finger movements, touching each finger with their thumb, beginning with the little finger and moving to the index finger, as quickly as possible until told to stop (after five cycles), while holding both hands out of peripheral vision. The sequence is performed once with the dominant hand and once with the non-dominant hands. Scoring is based on the duration (seconds) of completing five correct cycles. Lower scores reflect better finger sequencing and kinaesthetic awareness. This task has high inter-rater reliability .89 > r > .97.

Hebrew keyboarding Assessment (H-KBAT)

This H-KBAT examines keyboarding performance of fourth to sixth-grade students [41]. It consists of three tasks: (a) a 10-min paragraph-copying task, where students copy the text from a paper next to them at their own pace; (b) a 3-min dictation task, where the text is sounded through the computer audio, and (c) a 5-min composition task. In this study, we administered the first two tasks. Performance is measured by speed (mean number of characters typed per minute) and accuracy (percent of correct keys typed). The H-KBAT is performed on a 14" laptop, using the Microsoft® Word word-processing software. It has been found to have convergent validity for speed; a strong correlation, r = .9, p < .001, between the copying and dictation keyboarding speed. It was also found to show a developmental trend; 6^th-grade students typed significantly faster and more accurately than 4^th-grade students. Additionally, a high inter-rater reliability was found in the accuracy measure in both the copying (ICC = .99; 95% CI = .985-.997) and the dictation tasks (ICC = 1.00; 95% CI = 1.00-1.00) [17].

Procedure

This study is part of a larger study relating to various aspects of elementary-school students’ keyboarding performance. After obtaining approval from the Hebrew University’s ethics review board (#23082018) and consent from the Ministry of Education in Israel (AL10373), administrators of two elementary schools agreed to participate in the study. Next, a letter describing the study purpose, data collection procedure, and a consent form was sent to the parents. Students whose parents consented to their participation, and met the study criteria, were asked to sign an assent form. Next, the test battery was administered by paediatric occupational therapists who were trained to administer the test battery. Testing took place during two sessions (45 min each). The first session was carried out in the classroom, where students filled out the Background Questionnaire and completed the H-KBAT test, while being observed for their keyboarding method. In the second session, students were individually administered the other tests.

Data analysis

We analysed the data using IBM SPSS Statistics (Version 25; IBM Corp., Armonk, NY). Descriptive statistics were performed for demographic characteristics and to describe the central tendency and variability of measures. Statistical significance was set at p < .05. One-sample Kolmogorov-Smirnov test was used to assess normality. Results showed that all keyboarding performance variables (dependent variables) except for copying accuracy were normally distributed. Repeated measure ANOVA was conducted to examine the difference in students’ keyboarding performance while copying and dictation. Pearson and Spearman correlation analyses were used to examine the relationship between underlying functions, and keyboarding performance. Finally, linear regression analysis was administered to determine keyboarding performance predictors.

Results

Sample characteristics and possible confounding variables

Students’ reports revealed that all students owned a home computer, and none learned to touch type; which was confirmed during the observations. Students’ keyboarding performance (speed and accuracy) in each of the tasks is presented in Table 1. As hypothesised, we found that students’ keyboarding was significantly faster (more than double the speed) in dictation compared with copying (F(1,56) = 206.38, p < .001), but they were significantly more accurate in copying compared with dictation (F(1,56) = 85.01, p < .001). Further examination of the texts indicated that the low accuracy in the dictation task was mostly due to the students’ omission of text because they were not able to keep up with the dictated pace.

Table 1. Distribution of students’ typing performance measures.

Typing Performance	M	SD	Median	IQR
Copying
Speed	28.74	12.12	26.20	17.00
Accuracy	93.91	6.67	95.65	3.256
Dictation
Speed	59.54	23.58	57.00	38.84
Accuracy	62.34	24.60	60.15	42.48

IQR: Interquartile Range, Speed: Mean number of typed characters per minute, Accuracy: Percent of correct keys typed; Did not show normal distribution.

Next, to examine if gender was a possible confounding variable, we compared keyboarding performance of males and females. We found no significant differences (p > .05) except for copying speed; girls typed faster (M = 31.50, SD = 12.23) compared to boys (M = 24.95, SD = 11.10; t(52.25) = 2.10; p = .04). Therefore, in subsequent analysis, in predicting copying speed, we controlled the gender effect.

Relationship between underlying functions and keyboarding performance

Students’ scores in the underlying functions (reading speed, attention shifting, fine motor, and kinaesthetic awareness) are presented in Table 2. In examining the correlation between the underlying functions and keyboarding performance (speed and accuracy) in the copying and dictation tasks, results showed (Table 3) that reading speed was significantly moderately correlated with both copying (r = .54, p < .001) and dictation speed (r = .49, p < .001), as well as with dictation accuracy (r = .52, p < .001), indicating that the faster the students read, the faster they typed and were more accurate in dictation. We also found a negative low significant correlation between kinaesthetic awareness (Finger Succession) in the dominant hand and dictation accuracy (r = −.25, p < .01), suggesting that the higher the kinaesthetic awareness the more accurate was the students’ keyboarding in the dictation task. By contrast, we did not find any relationship between fine motor and attention shifting and keyboarding performance.

Table 2. Descriptive statistics of students’ underlying functions (N = 57).

Underlying functions	M	SD	Median	IQR
Aleph-Taph	76.67	17.89	74.43	24.77
CCTT2	89.79	15.07	94.00	21.00
PPT	4.45	1.97	4.04	3.44
Finger Succession
Dominant hand	10.74	5.08	10.00	3.59
Non-dominant hand	11.08	4.67	10.80	3.00

IQR: Interquartile Range, Aleph-Taph: Reading speed; number of words per minute, CCTT2: Children’s Colour Trails Test; Attention shifting, PPT: Purdue Pegboard Test; Fine motor skills - Z scores, Finger Succession: Kinaesthetic awareness; time in seconds – lower scores reflect better awareness.

Table 3. Correlations between underlying functions and typing measures (N = 57).

	Copying		Dictation
Underlying functions	Speed	Accuracy	Speed	Accuracy
Aleph-Taph	.54***	−.03	.49***	.52***
CCTT2	.03	.10	−.02	−.01
PPT	.08	.13	.10	.08
Finger succession
Dominant hand	−.14	−.08	−.15	−.25**
Non-dominant hand	−.03	−.02	.02	−.10

Aleph-Taph: Reading speed; number of words per minute; CCTT2: Children’s Colour Trails Test; Attention shifting; PPT: Purdue Pegboard Test; Fine motor skills - Z scores, Finger Succession: Kinaesthetic awareness; time in seconds – lower scores reflect better awareness.

**p < .01; ***p < .001.

Due to the few correlations found between the different underlying functions and keyboarding performance measures, we examined two regression models. In the first model, predicting copy-keyboarding speed (Table 4), we used hierarchical regression analysis. In the first step, we entered gender, with the purpose of controlling for its effect, and in the next step, we entered reading speed. Results showed that the final regression model was statistically significant (F(2,56) = 13.30, p < .001). Although in the first step, gender was found to be a significant predictor, accounting for 5.6% of the variance in copying speed, it no longer had a unique significant contribution after entering reading speed. Together they accounted for 30.5% of the variance of keyboarding copying speed. In the next model we examined the functions predicting dictation accuracy using linear regression (Table 5). Results showed that the regression model was statistically significant (F(2,56) = 11.17, p < .001). Together, reading speed and kinaesthetic awareness (Finger Succession) accounted for 26.6% of the variance in keyboarding accuracy in the dictation task, but only reading speed had a significant unique contribution.

Table 4. Regression model for predicting copying speed.

Predictors	B	SE	Βeta	Adjusted R^2	p Value
Step 1				.056	.04
Gender	−6.56	3.16	−.27
Step 2				.305	<.001
Gender	4.85	2.74	.20		.08
Aleph-Taph	.35	.51	.51		<.001

Aleph-Taph: Reading speed; number of words per minute.

Table 5. Regression model for predicting dictation accuracy.

Predictors	B	SE	Βeta	Adjusted R^2	p Value
				.266	<.001
Aleph-Taph	.68	.16	.49		<.001
Finger-Succession	−.64	.57	−.13		.27

Aleph-Taph: Reading speed; number of words per minute, Finger Succession: Kinaesthetic awareness; time in seconds.

Discussion

Although the use of computers in educational settings has been gradually increasing, becoming the main form of written expression, the knowledge concerning the keyboarding process and underlying functions associated with this important skill is very limited [11]. This gap of knowledge may limit occupational therapists ability to treat students who encounter difficulties in becoming proficient typists. This study aimed to address this gap by comparing elementary-school students’ keyboarding performance (speed and accuracy) in two tasks: copying and keyboarding to dictation, and examining the relationship between various underlying functions and keyboarding performance. As opposed to most keyboarding research, in this study we measured students’ performance in two tasks, because they entail different functions, which may affect their keyboarding performance and the underlying functions associated with keyboarding.

As expected, we found a task effect on students keyboarding performance. Students typed faster in the dictation compared to the copying task. These findings are aligned with those reported by Khoury-Shaheen and Weintraub [18], who described the development of two keyboarding assessments in Hebrew, for 4^th- and 6^th-grades and in Arabic, for 4^th- and 5^th -grades. These findings may not be surprising because in the dictation task, students had to type at the dictated speed, whereas in the copying task, they typed at their own speed. Similar to findings in handwriting [42], these findings suggest that students are capable of adapting their keyboarding speed to the task requirements (e.g. keyboarding faster while taking exams or notes compared to writing compositions), which is encouraging. However, the increase in speed may tax other aspects of keyboarding performance such as accuracy.

In fact, our findings revealed a speed-accuracy trade-off [43]. In the copying task, students’ accuracy was high (over 93%), but the speed was relatively slow. By contrast, in the dictation task, the accuracy dropped to 62%, but the speed doubled. This phenomenon is common among individuals learning a motor skill [44], including novice typists, who are in the process of optimising their speed-accuracy balance, and adjusting it to their skill-level, often consciously slowing down their performance to achieve better precision [45].

Next, we examined the contribution of underlying functions in predicting keyboarding performance. Our findings showed that only reading speed was associated with keyboarding speed. However, because we also found that gender was associated with keyboarding speed, in predicting this measure, we first entered gender and found a significant contribution. Yet, after entering reading speed, gender no longer had a significant contribution. Hence, it appears that similar to previous studies [21,22,29], keyboarding is not affected by gender. A possible explanation for this finding is that copying performance depends on reading ability, and elementary-school girls have been found to read faster than boys [46]. Thus, perhaps the gender effect on copy-keyboarding speed reflected the girls’ higher reading speed and not necessarily their keyboarding speed; explaining why we did not find a gender effect in the other keyboarding measures.

In fact, reading speed was the major underlying factor of keyboarding performance in most measures; it was a significant predictor of copying and dictation speed as well as keyboarding accuracy in the dictation task. This finding is similar to those by Ben-Moyal-Zada [22] among fourth-grade students, suggesting that the speed of reading a text may play an important role in the process of keyboarding [23] since it entails similar linguistic functions such as phonological, morphological, orthographic, and semantic knowledge [19,47].

As opposed to reading, we did not find a significant association between fine-motor skills (as measured by the 'Both Hands’ task in the PPT) and keyboarding speed and accuracy in either of the tasks. This result is aligned with that by Preminger et al. [24], but contrasts the finding by Ben-Moyal-Zada [22], who also used the PPT. However, in the latter study, the researcher administered different PPT sub-tests (i.e. the individual hand tests and sequencing). Thus, perhaps that 'Both Hands’ task of the PPT does not tap the fine-motor skills required for keyboarding performance. However, the evidence concerning the fine-motor skills associated with keyboarding performance is still in its initial stages, and requires further examination.

Contrary to our expectations, our results also showed no significant correlations between attention shifting and keyboarding performance. A possible explanation for this finding is that, as opposed to experienced typists, who often shift their attention from the screen to the keyboard, novice typists spend most of their time attending to the keyboard. Therefore, the ability to shift attention is less relevant for this population. Support for this hypothesis may be found in the study by Ben-Moyal-Zada [22], who reported that a significant correlation between attention shifting and keyboarding speed was only found among sixth, but not fourth-grade students. These findings suggest that as students become more proficient typists (e.g. more frequently shifting their gaze from the keyboard to the screen), the ability to shift attention may play a more significant role while keyboarding [48].

Finally, our results showed no significant correlations between kinaesthetic awareness and keyboarding performance, excluding dictation accuracy. This result is consistent with most of the research findings, especially among novice typists [25,27]. This brings to question why we did find a significant correlation between kinaesthetic awareness and dictation accuracy. Perhaps, when students had better kinaesthetic awareness, they had a higher capacity to rely on kinaesthetic feedback while keyboarding. Consequently, they were able to shift their gaze from the keyboard to the screen, and monitor their typed text, leading to higher accuracy.

Implications for practice

The results of this study provide new data, which enhance the understanding of the keyboarding process and performance of elementary-school students. The findings suggest that when occupational therapists and educators evaluate students’ keyboarding performance, multiple tasks should be employed, because students’ performance varies in the different tasks. Moreover, students’ reading speed and kinaesthetic awareness should also be examined, as it appears to contribute to keyboarding performance. Finally, this study’s findings imply that at elementary school, students are still at their early stages of keyboarding acquisition, and are limited in their ability to balance speed and accuracy. Therefore, occupational therapists and teachers should adjust their expectations. For example, requesting students to perform keyboarding tasks that require high-speed (e.g. taking exams, or writing time-limited essays) may affect their accuracy as well as their ability to express their knowledge (as was found in the high omissions during dictation).

Limitations and future research

One limitation of our study is the sample size and uniform age. Expanding the sample size and age may enhance the external validity of the study’s results and provide a developmental perspective. Another limitation might be regarding the instruments used. The results of this study indicated that other than reading speed, there is a gap between presumed underlying functions of keyboarding (based on analysis of the keyboarding activity) and the results, which are based on behavioural tests that tap these underlying functions. For example, whereas there is a general agreement that keyboarding requires attention and specifically attention shifting, the behavioural test that measures the kind of attention related to keyboarding performance is still not clear. This gap calls for further research on this topic.

Conclusions

Our study aimed to compare elementary-school students’ keyboarding performance in a copying and keyboarding to dictation tasks, and to explore the underlying functions associated with keyboarding performance. This topic has rarely been studied, particularly among elementary-school students. Findings showed that students’ keyboarding performance varied in the different tasks, which suggests that students’ keyboarding performance evaluation should include multiple tasks. Furthermore, reading speed was associated with keyboarding performance in the two tasks, implying that when evaluating keyboarding readiness or performance, students’ reading speed should be considered. This knowledge may assist occupational therapists in their clinical reasoning and decision making when working with students with keyboarding difficulties or when considering keyboarding as an alternative writing mode for students with handwriting difficulties.

Acknowledgements

The authors thank the schools’ principals, teachers, and students for their time and cooperation in collecting the data.

Disclosure statement

No potential conflict of interest was reported by the author(s).