Abstract
This study found that kindergarteners who performed poorly in spatial relationships, visual–motor Integration, manual dexterity, and attention were likely to have less legible Chinese handwriting and slow writing speed in the first grade.
Proficient handwriting allows students to take notes, complete assignments, and express their ideas effectively in written tests. However, mastering handwriting skills, particularly during the early elementary school years, can be challenging. Studies have shown that despite receiving handwriting instruction and practice, approximately 5% to 33% of school-age students still have poor handwriting legibility and slow writing speed (Karlsdottir & Stefansson, 2002; Overvelde & Hulstijn, 2011; Tseng, 1993). These handwriting problems may contribute to academic and psychological problems, such as low academic performance, feelings of frustration, less motivation, and passive attitudes toward handwriting tasks (Feder & Majnemer, 2007). Early detection and intervention for children at risk of handwriting problems are critical to prevent these issues, which can start as early as kindergarten (Kadar et al., 2020; Zylstra & Pfeiffer, 2016). However, which kindergarten characteristics are related to later handwriting performance is still uncertain.
In some countries, children begin formal handwriting instruction in kindergarten or even earlier (Tse et al., 2019). For children who have already learned how to write in kindergarten, evaluating their handwriting performance directly may be a feasible way to identify those at risk of handwriting difficulties early on (Marr & Cermak, 2003). However, in other countries, such as Taiwan, formal handwriting instruction and practice are not included in the kindergarten curriculum (Taiwan Ministry of Education, 2017), so assessing kindergarteners’ handwriting performance to identify at-risk children may be unreliable. Nonetheless, some developmental skills are crucial for proficient handwriting. Therefore, investigating whether performance of these skills in kindergarten can predict later handwriting performance could help develop a reliable way to screen and identify at-risk children early on.
Handwriting is a complex task that requires the integration of sensory, perceptual, cognitive, and motor systems (Schneck & Case-Smith, 2015). The most explored foundational skills related to handwriting performance are fine motor (FM), visual–motor integration (VMI), and visual perception (VP). It is well established that FM and VMI skills are strongly correlated with the legibility (Cornhill & Case-Smith, 1996; Volman et al., 2006) and speed (Tseng & Chow, 2000) of school-age children’s handwriting. Although the relationship between VP and handwriting among children seems inconsistent in alphabetic handwriting studies (Feder et al., 2005; Lee, 2021; Volman et al., 2006), a few studies on Chinese handwriting have consistently confirmed the correlation between performance of both (Tse et al., 2014; Tseng & Chow, 2000).
Although limited research has been conducted on the association between these fundamental skills and handwriting performance among kindergarten children, the results align with findings found with school-age children. Kindergarteners’ FM and Beery–Buktenica Developmental Test of Visual–Motor Integration (Beery™ VMI; Beery et al., 2010) scores also significantly correlate with their handwriting legibility (Daly et al., 2003; Frolek Clark & Luze, 2014 ; Seo, 2018; Tse et al., 2019). However, evidence of the importance of VP in kindergarteners’ handwriting performance is limited. A study on Hong Kong’s kindergarten children found that, of seven VP aspects tested by the Test of Visual Perceptual Skills–Third Edition (TVPS–3; Martin, 2006), spatial relationships were the best predictor of children’s handwriting legibility in both English and Chinese copying tests (Tse et al., 2019). This suggests that spatial relationships may play a crucial role in kindergarteners’ handwriting development.
The extent to which developmental skills performed in kindergarten can predict children’s future handwriting performance in school remains uncertain because of the limited research. Two studies on English handwriting have explored this issue. On the one hand, Marr and Cermak (2002) found that kindergarten Beery VMI scores significantly predicted first-grade letter legibility only among girls. On the other, van Hartingsveldt et al. (2015) examined how kindergarten Beery VMI and FM scores predicted first-grade handwriting skills and found a significant predictive value of both abilities for later handwriting speed but not for legibility. A recent study on Chinese handwriting by Hwang et al. (2020) found a significant relationship between kindergarten FM and VP and the risk of poor handwriting in first grade, mainly based on the ratings of first-grade teachers.
In this study, we aimed to further identify the independent predictive power of kindergarten FM, VP, and VMI skills for Chinese handwriting legibility and speed, as assessed through direct handwriting assessments in first grade. Additionally, we explored the predictive power of behavioral traits such as inattention, hyperactivity, and impulsiveness for first-grade handwriting performance, given their importance in handwriting learning, as shown among children born prematurely (Feder et al., 2005; Shih et al., 2018).
Method
Study Design
This study used a 1-yr longitudinal, prospective, observational design with two data collection points. In the first year, data on kindergarten children’s visual–perceptual and motor abilities and behavioral traits were collected. Approximately 1 yr later, in the first grade, their handwriting legibility and speed were assessed.
Participants
Participants were recruited from kindergartens located in Tainan City, Taiwan. The inclusion criteria for participation included (1) being age 5 or 6 yr, (2) starting the first grade in September (the month of elementary school enrollment in Taiwan) in the same year as data collection, and (3) having parents who were native Chinese speakers. Children who had (1) chromosomal or genetic deficits; (2) neuromuscular or developmental disabilities (e.g., cerebral palsy, intellectual disabilities, autism spectrum disorder, attention deficit hyperactivity disorder); (3) significant neuromuscular or orthopedic injury (e.g., finger fractures); (4) visual or hearing impairments; or (5) any remediation intervention (e.g., occupational therapy, speech therapy, taking medicines for learning problems, interventions for handwriting) during the time of the study were excluded. This study was reviewed and approved by the Ethics Committee of National Cheng Kung University Hospital (B-ER-104–119). Written informed consent was obtained from all parents and teachers before data collection.
Measures in Kindergarten Year
Bruninks–Oseretsky Test of Motor Proficiency–Second Edition
We used the Fine Motor Precision (FMP) and Manual Dexterity (MD) subtests of the Bruininks–Oseretsky Test of Motor Proficiency–Second Edition (BOT–2; Bruininks & Bruininks, 2005) to measure children’s FM skills. The FMP subtest consists of five time-free tasks, such as connecting dots and folding a paper, whereas the MD subtest has five time-limited tasks, such as making dots in circles and transferring pennies. Both subtests demonstrate excellent reliability and validity, as reported in the test manual.
Test of Visual Perceptual Skills–Third Edition
The TVPS–3 is a widely used test with sound psychometric properties for measuring different aspects of VP. It consists of seven subtests: Visual Discrimination (e.g., identifying a design that is identical to the target), Visual Memory (e.g., recognizing a design that was just seen), Spatial Relationships (e.g., identifying a design with a different orientation from the others), Form Constancy (e.g., identifying a design that may be smaller, larger, darker, or reversed compared with the target), Visual Sequential Memory (e.g., recognizing shapes in the same order as they were seen), Visual Figure–Ground (e.g., identifying a shape hidden in pictures the same as the target shape), and Visual Closure (e.g., recognizing a design that is identical to the target when its lines are connected). Each subtest contains 16 questions. For each question, children are presented with four to five design choices and asked to select the correct one.
Beery–Buktenica Developmental Test of Visual– Motor Integration
The Beery VMI (Beery et al., 2010) is a well-known international test with strong reliability and validity. It is a pencil-and-paper test in which the child is asked to copy 24 geometric forms without a time limitation. A Chinese version with norms for Taiwanese children is available (Liu & Lu, 1999). For the current study, two trained raters scored the participants’ test booklets, and their interrater reliability was excellent (intraclass correlation coefficient [ICC] = .97).
Attention-Deficit/Hyperactivity Disorder Test
The Attention-Deficit/Hyperactivity Disorder Test (ADHDT; Gilliam, 1995) consists of 36 items that are divided into three subscales: Inattention, Hyperactivity, and Impulsiveness. Teachers use a 3-point Likert scale to rate the severity of a child’s problematic behaviors. Higher scores indicate more severe attention deficit hyperactivity disorder (ADHD) symptoms and traits. Children whose subscale standard score is higher than 12 or whose ADHD quotient is higher than 110 are highly suspected of having ADHD (Cheng, 2007). According to the test manual, the ADHDT–Chinese version has sound psychometric properties.
Measures in First Grade
Chinese Handwriting Legibility Test
The Chinese Handwriting Legibility Test (CHLT) was developed by our research group (Hsiao & Hwang, 2023). Forty-five target words of the CHLT were selected from the words most often used in first-grade Chinese language textbooks employed by Taiwanese elementary schools. Each target word is included in a three- or four-word sentence, and each word in the sentence is appended to its Mandarin phonetic symbol. The tester read each word to the child and then asked the child to write the target word in the blank grid (2 cm × 2 cm) from memory to complete the sentence without a time limitation.
Chinese handwriting legibility can be classified into two dimensions: character accuracy and character construction (Tseng, 1993). Therefore, the CHLT produces scores on character accuracy and character construction. The scoring criteria for the accuracy and construction of each target word were modified according to the criteria of Tseng’s Handwriting Problem Checklist (Tseng, 1993) and the Chinese Handwriting Evaluation Form (CHEF; Chang & Yu, 2012).
The scoring criteria for accuracy included character formation (i.e., malformation of strokes or radicals) and the number of strokes within the character (i.e., missing or adding strokes), rated on a 2-point scale (0 = incorrect, 1 = correct). The total accuracy score ranges from 0 to 45 points, with a higher score indicating more words written correctly.
For character construction, we selected from the 45 target words the 10 that most participants wrote correctly and asked three elementary school teachers to rate the construction of each word on a 5-point scale (ranging from 1 = very poor to 5 = excellent). The scoring criteria were based on (1) inappropriate size, spacing, and position of strokes or radicals in a character and (2) inability to write a character within the grid (i.e., out of grid). Incorrectly written characters were not scored on character construction. Two of the three teachers randomly scored the child’s character construction, and their average score for per character represented the child’s construction score, ranging from 1 to 5 points. A higher score indicates better character construction.
The CHLT has good test–retest (accuracy, r = .71; construction, r = .55), intrarater (accuracy, r = .98; construction, r = .81–.91), and interrater (accuracy, r = .94; construction, ICC = .77) reliabilities. Additionally, the CHLT significantly correlates with the handwriting T score on the Battery of Chinese Basic Literacy (BCBL; Hung et al., 2003; accuracy, r = .72, p < .001; construction, r = .2, p = .036), as well as with the Accuracy and Construction dimensions of the CHEF (Chang & Yu, 2012; accuracy, r = −.49, p < .001; construction, r = −.52, p < .001; Hsiao & Hwang, 2023).
Chinese Word Copying Test
This test consists of 25 Chinese characters that were also selected from the words most often used in first-grade Chinese language textbooks. The sample sheet has five columns with five stimulus words in each column. The child was asked to copy the stimulus words into the 2 cm × 2 cm grids on the test sheet, sequentially from top to bottom and right to left, as quickly and accurately as possible. The total time taken to complete the near-copying test was recorded. The number of strokes written per minute represents the child’s writing speed. The test–retest reliability (r = .81) of this test is good. The children’s copying speed was found to be significantly correlated with their handwriting score on the BCBL (r = .52, p < .001) and the Speed dimension of the CHEF (r = −.24, p = .018; Hsiao & Hwang, 2023).
Procedure
Kindergarten children came to our department to complete the BOT–2 FMP and MD subtests, TVPS–3, and Beery VMI tests. All tests were administered by one of two trained occupational therapists (one of whom was Ying-Lu Hsiao). Simultaneously, mothers were asked to fill out the demographic questionnaire onsite. Children’s kindergarten teachers were asked to fill out the ADHDT–Chinese version and a questionnaire about the content of their handwriting instruction in school. After a year, when children were in the first grade, they returned to our department for the handwriting assessments. A break was given between handwriting assessments to prevent hand muscle fatigue. Onsite, mothers were requested to answer a few questions regarding their child’s medical condition within the past year.
Data Analysis
All data were analyzed with R software (Version 4.0.2). Descriptive statistical analyses were used to describe the characteristics of the present sample. Spearman correlations or χ2 tests were used to determine the correlation between all explanatory variables, including child characteristics and test scores obtained in kindergarten. Univariate linear regression analysis was conducted to identify significant factors related to first-grade handwriting performance. After model-fitting techniques (i.e., simultaneously considering the contribution of each significant factor to the handwriting outcome and their correlation and collinearity) were used (Daniel & Cross, 2018; MacFarland, 2013), several potential multivariate linear regression models for each handwriting outcome were identified. In this study, we used only the model with the largest adjusted R 2 to predict each handwriting outcome. Two-sided tests were performed to evaluate model parameters, and p < .05 was set as the level of statistical significance for all analyses.
Results
The data from 106 children (53 girls and 53 boys) who met the criteria and had completed the necessary assessments were analyzed. More than two-thirds of the mothers (69.8%) had a bachelor’s degree or higher. The majority of children had a typically developing score on kindergarten visual–perceptual and motor tests. Only a small number of children had scores below 85 (1 SD below the mean) on the Beery VMI (n = 5) and TVPS–3 (n = 1) and scores below 10 (1 SD below the mean) on the FMP (n = 3) and MD (n = 1) subtests. None of the children had a suspected ADHD quotient score (>110), but 6 children had subscale scores (>12) indicating suspected ADHD. In the first grade, 92.5% of the children were right-handed. On the CHLT, the average character accuracy score was 30.8, ranging from 15 to 44, and the average character construction score was 2.9, ranging from 1.8 to 4.1. The average writing speed was 74.4 strokes per minute.
Table 1 displays the results of the univariate regression analysis. Factors that showed significant correlations with the accuracy score included Beery VMI, TVPS–3 Visual Memory and Spatial Relationships subtest scores, Inattention and Impulsiveness scores (subscales of the ADHDT), and duration of time spent in first grade. Additionally, the child’s sex, mother’s education level and Beery VMI, BOT–2 FMP subtest, TVPS–3 Spatial Relationships and Sequential Memory subtests, and three ADHDT subscale scores (Inattention, Hyperactivity, and Impulsiveness) were significantly correlated with the character construction score. Furthermore, the BOT–2 MD score, the child’s age at entry into elementary school, and the duration of time in first grade showed significant correlations with the copying speed.
Predictors of First-Grade Handwriting Performance Using Univariate Regression Analysis
Note. ADHDT = Attention-Deficit/Hyperactivity Disorder Test; BOT–2 = Bruninks–Oseretsky Test of Motor Proficiency–Second Edition; FMP = Fine Motor Precision; MD = Manual Dexterity; ref. 5 reference group; TVPS–3 = Test of Visual Perception–Third Edition; VMI = Beery–Buktenica Development Test of Visual–Motor Integration.
*p < .05. **p < .005.
Table 2 presents the multivariate linear regression model with the highest adjusted R2 for each handwriting outcome, as determined by model-fitting tests. The duration of time in first grade, VMI, spatial relationships, and inattention accounted for 19.9% of the variability in character accuracy, F(4, 101) = 7.517, p < .001. All factors except for VMI (p = .069) were independent predictors for this handwriting outcome variable. The multivariate regression model for character construction, consisting of sex, mother’s education, VMI, spatial relationships, and inattention, accounted for 24.2% of the variability, F(5, 100) = 7.696, p < .001. Spatial relationships (p = .125) was the only variable that did not demonstrate a significant independent prediction for this handwriting outcome. Additionally, the multivariate regression model including age at school entry, months in first grade, and MD score accounted for 21.2% of the variability in copying speed, F(3, 102) = 10.392, p < .001. All factors showed a significant independent contribution to copying speed (Table 2).
Predictors of First-Grade Handwriting Performance Using Multivariate Regression Analysis
Note. CI = confidence interval; MD = Manual Dexterity; ref. = reference group; VMI = visual–motor integration.
*p < .05. **p < .005.
Discussion
This study explored the long-term prediction of kindergarten visual–perceptual and motor abilities and behavioral traits on children’s handwriting legibility (character accuracy and construction) and speed in the first grade. After controlling for children’s biological and social factors (e.g., sex, age at elementary school entry, maternal education levels), multivariate regression results indicated that kindergarten spatial relationships and inattention were the most significant predictors of first-grade character accuracy, whereas VMI and inattention were the most significant predictors of character construction. Additionally, MD was the best predictor of handwriting speed.
Previous studies have shown that VP is crucial to Chinese handwriting performance (Tse et al., 2014; Tseng & Chow, 2000). Our study further demonstrated that among the seven subtests of the TVPS–3, Spatial Relationships was the best kindergarten predictor of the first-grade accuracy score on the CHLT, which requires children to write Chinese words from memory. This result highlights the importance of spatial relationships in correctly remembering the formation of Chinese characters for beginning handwriting. Two previous studies have also reported similar findings. McBride-Chang et al. (2008) found that spatial relationships were significantly associated with Chinese word recognition (r = .37) and rapid number naming (r = −.41) in preschool children at risk for dyslexia. Another study of Hong Kong children indicated that kindergarten Spatial Relationships scores were the only TVPS–3 subtest scores that predicted handwriting legibility in copying English and unfamiliar Chinese words (Tse et al., 2019).
However, the present results did not show any of the TVPS–3 subtests to have an independent predictive value for character construction. It should be noted that in the current study, all Chinese words used for the rating of character construction were ones that children wrote correctly, indicating that children were familiar with those words. This suggests that VP may not play a critical role in the spatial arrangement of strokes and radicals of familiar Chinese words.
Despite not being diagnosed with ADHD, the participants’ attention significantly predicted both character accuracy and construction in first grade. These findings indicate that insufficient attention during kindergarten activities may continue to affect handwriting learning as children progress through elementary school. Therefore, kindergarten teachers and parents should take steps to identify the causes of children’s inattention and improve their awareness of learning activities. However, our findings on inattention were inconsistent with those of van Hartingsveldt et al. (2015), who did not find that children’s sustained attention independently predicted their first-grade handwriting legibility. It is possible that attention may play a more critical role in Chinese handwriting because of the more complicated construction of Chinese characters compared with English words. Furthermore, different attention measurements used in both studies may contribute to the inconsistent results.
We found that kindergarten VMI independently predicted first-grade legibility, particularly for character construction but not for writing speed. In contrast, van Hartingsveldt et al. (2015) found that kindergarten VMI significantly predicted only first-grade writing speed, which our study did not replicate. One possible explanation for this discrepancy could be related to the varying levels of font complexity between Chinese and English.
Evidence suggests a correlation between FM and handwriting legibility among kindergarten children (Frolek Clark & Luze, 2014; Seo, 2018). However, van Hartingsveld et al.’s (2015) study and our study did not find such a correlation between kindergarten FM and first-grade handwriting legibility. The difference in results may be attributed to the fact that van Hartingsveld et al.’s study (2015) and our study had a 1-yr time interval between FM and handwriting assessments, whereas other studies measured the two variables concurrently. However, our results showed that kindergarten FM was a significant predictor of first-grade copying speed in Chinese, which is consistent with van Hartingsveldt et al.’s (2015) findings for English handwriting. Furthermore, of the two FM tests used in our study, BOT–2 MD score appeared to be a better predictor of writing speed than the FMP score, possibly because of the time limit requirement of the MD subtest. This finding suggests that time-limited FM training activities may be more effective in enhancing kindergarten children’s later writing speed than free-form training.
This study had several limitations. First, we did not collect data on the amount and content of handwriting assignments for each child in kindergarten and first grade because they can be challenging to estimate accurately. This lack of data may have had an impact on the present findings. Furthermore, the spatial construction of Chinese characters is quite distinct from that of alphabetic language systems. Therefore, caution should be taken when applying the present results to other languages.
Implications for Occupational Therapy Practice
The results of this study have the following implications for occupational therapy practice: Kindergarten FM, VP, VMI, and attention have different predictive powers for first-grade Chinese handwriting legibility and speed, suggesting that a comprehensive evaluation that incorporates assessments of these multiple abilities may yield a more accurate prediction of children’s Chinese handwriting performance in the first grade than one that assesses only a singular aspect of their abilities. In addition to FM and VMI, occupational therapists can improve the handwriting readiness program in kindergarten by incorporating training for spatial relationships and attention. This approach may support children’s handwriting performance, given the significant predictive power of these factors for first-grade Chinese handwriting legibility.
Conclusion
This study found that kindergarten spatial relationships, inattention, and VMI were significant predictors of first-grade Chinese handwriting legibility, and FM skills were the strongest predictor of first-grade writing speed. Identifying these predictors can help detect kindergarteners who may be at risk of future handwriting problems.
Footnotes
Acknowledgments
We thank the children, parents and teachers who participated in this study.