chang & yu (2009)

Perceptual and Motor Skills, 2009, 109, 3, 770-782. © Perceptual and Motor Skills 2009

DOI 10.2466/PMS.109.3.770-782 ISSN 0031-5125

DISCRIMINANT VALIDITY OF THE VISUAL MOTOR INTEGRATION TEST IN SCREENING CHILDREN

WITH HANDWRITING DYSFUNCTION1, 2

SHAO-HSIA CHANG

Department of Occupational Therapy I-Shou University

NAN-YING YU

Department of Physical Therapy I-Shou University

Summary.—To examine the discriminant validity of the Visual Motor Integra-tion test in screening children with handwriting dysfunction, 599 children in Grade 2, including 41 children with handwriting dysfunction identified by their teachers and 558 typically developing children, were assessed. The Visual Motor Integration test, with an area under the receiver operating characteristic curve (0.894), showed high accuracy regarding screening purposes. Judging from the values for sensitiv-ity, specificity, positive likelihood ratio, negative likelihood ratio, hit rate, Youden’s index, and odds ratio, a standard score of 85 was the best cutoff point for screening children for handwriting dysfunction.

From literature reviews of occupational therapy in the school sys-tem, a common problem for referred children is handwriting dysfunc-tion (Tseng & Cermak, 1993; Jongmans, Linthorst-Bakker, Westenberg, & Smits-Engelsman, 2003; Denton, Cope, & Moser, 2006). Handwriting is a fundamental skill for schoolchildren; however, it is a complex activity with many essential components. Fluent and legible handwriting requires maturation and integration of cognition, visual perception, kinesthesia, motor planning, visual-motor integration, and fine motor skills (Maeland, 1992; Tseng & Cermak, 1993; Tseng & Murray, 1994; Weil & Amundson, 1994). Berninger and Ruberg (1992) proposed that fine motor movements as measured by the finger-succession task are the best indicators of hand-writing performance, explaining 52.5% of the variance in the evaluation of handwriting performance through differences in fine motor control. In the development of fluent handwriting, visuomotor integration and in-hand manipulation were found to be significant predictors of handwriting test scores, accounting for almost 73% of the variance (Cornhill & Case-Smith, 1996). Since handwriting has underlying performance components, evalu-ating a handwriting problem and identifying the related components for an effective intervention program are challenging.

Teachers usually assess the legibility of children’s letter and handwrit-1Address correspondence to Nan-Ying Yu, RPT and Ph.D., Assistant Professor, Department of Physical Therapy, I-Shou University, No. 8, Yida Rd., YanChao Township, Kaohsiung County, Taiwan, 82445, ROC or e-mail ([email protected]).2The authors thank the National Science Council of the Republic of China for supporting this work financially under Contracts No. NSC-94-2614-E-214-005 and NSC-95-2221-E-214-008.

VISUAL MOTOR INTEGRATION, HANDWRITING DYSFUNCTION 771

ing forms (regular or cursive connected script) using a standard evalu-ation. This evaluation involves having a child copy an article or a para-graph within a certain time frame and, notably, involves a subjective examination of characters, numbers, and sentences. However, the factors that contribute to legible handwriting are multifactorial and complex and the test-retest reliability of these subjective methods remains question-able (Diekema, Deitz, & Amundson, 1998; Stefansson & Karlsdottir, 2003; Feder, Majnemer, Bourbonnais, Blayney, & Morin, 2007). Actual perfor-mance in the classroom is the key factor in observing handwriting prob-lems (Chang & Yu, 2005), and so teachers’ ratings of handwriting legibility have been utilized to verify the validity of the Evaluation Tool of Chil-dren’s Handwriting–Manuscript (ETCH–M). However, a weak correlation was found between scores on the ETCH–M and teachers’ ratings of legi-bility. Importantly, identification by a teacher, using a general observation rather than a one-time assessment, has been shown to be a valid method for identifying children with handwriting difficulties (Sudsawad, Trom-bly, Henderson, & Tickle-Degnen, 2001).

In evaluating handwriting, the Beery-Buktenica Developmental Test of Visual–Motor Integration (VMI) is most commonly used (Doyle & Goy-en, 1997; Feder, Majnemer, & Synnes, 2000). Findings indicate that of all perceptual-motor skills, visual-motor integration correlates best with handwriting performance (Maeland, 1992; Tseng & Murray, 1994; Weil & Amundson, 1994; Tseng & Chow, 2000; Daly, Kelley, & Krauss, 2003), so visual-motor integration is an important component of handwriting skill; however, it explained only 30.5% of the variance (Tseng & Murray, 1994) in handwriting performance. Further, previous studies’ focus was only on correlations, so no data are available on how visual-motor integration con-tributes to handwriting performance. Whether visual-motor integration scores could predict handwriting performance adequately is not clear.

The Visual Motor Integration test is commonly used in the evaluation of handwriting, as found by Doyle and Goyen (1997) in a survey of pedi-atric occupational therapists in the New South Wales region of Australia. Feder, et al. (2000) conducted a survey of 50 senior Canadian pediatric oc-cupational therapists, most of whom evaluated gross/fine motor, percep-tual skills, motor planning, movement quality, and sensory functioning for children having handwriting problems, while less formal assessments were conducted for psychosocial or environmental aspects. Use of the Vi-sual Motor Integration test was more frequent as compared to a standard-ized evaluation of handwriting performance.

The Visual Motor Integration test is not designed particularly to screen for handwriting dysfunction. Despite the convenience of admin-istering the test and the results being correlated with handwriting skill,

S-H. CHANG & N-Y. YU772

whether the test can provide additionally useful information in handwrit-ing assessment is still in question. The discriminant validity of the test for handwriting dysfunction is an important topic.

Comprehensive studies are required to assess whether the Visual Mo-tor Integration test is effective in screening children with handwriting dysfunction. The present study applied the receiver operating characteris-tic (ROC) analysis to explore discriminant validity on screening handwrit-ing dysfunction. This analysis was chosen as it was developed to assess test performance as a trade-off between sensitivity and specificity, with the goal of providing an appropriate cutoff score. In addition, the present Visual Motor Integration test results can provide clinicians as an objective reference for screening children’s handwriting.

MethodParticipants

Participants were 601 schoolchildren from 19 Grade 2 classes of three elementary schools in southern Taiwan. From the data provided by the school health center, children with any physical, psychological, or emo-tional problems, such as visual or hearing impairment, epilepsy, atten-tion/hyperactivity deficits, or mental retardation, were excluded from this study. One of the 601 children was diagnosed as having mild mental retar-dation, and another one was repeating the grade (age = 102 mo.), so data for these two children were excluded from the present study. Children with handwriting dysfunction were selected objectively through a pro-cess in which the class teachers were given an evaluation questionnaire (Chinese Handwriting Evaluation Questionnaire) as a referenced criteri-on. The Chinese Handwriting Evaluation Questionnaire is composed of six major dimensions, including legibility, accuracy, speed, pencil grip, gross movement, and attitude (Chang & Yu, 2005). According to the test manual, the presence of problems in at least two of the six dimensions is the criterion for the identification of a handwriting deficit. Forty-one chil-dren were referred by teachers as having some level of handwriting dys-function; Table 1 presents the ages and handwriting status by sex for the 599 participants.Tests

The Beery-Buktenica Developmental Test of Visual–Motor Integra-tion (VMI) was originally developed by Beery in 1967. The Chinese ver-sion employed in this study was prepared by Liu and Lu in 1997 (Beery, 1997). This test is a developmental one for assessing visual-motor integra-tion, visual perception, and motor coordination, with an applicable age range of three years to adulthood. The test requires the respondent to copy


geometric figures with paper and pen, three figures by imitation and 24 geometric forms directly copied from those in a developmental sequence, totaling 27 items. The test can be administered to children collectively or individually in about 10 to 15 minutes. The administration manual sug-gests that younger children should be tested individually, while children above Grade 1 could be tested collectively in a class. This study adopted this assessment tool and used only the Visual Motor Integration subscale.

Scoring was done according to the criteria set for each of the figures listed in the manual to apply a pass or fail, while 1 point was scored per item passed for a total of 27 points. In addition to raw scores, from a table in the manual, scores could be transformed to the standard, scaled, or age-equivalent ones. This study adopted the standard score for analysis.

The norms on the Visual Motor Integration for Taiwanese Children were constructed in 1972. The test adopted in this study is the fourth ver-sion. According to age, the norms of this version were constructed on a sample of 2,977 participants, for whom the age interval was set to be every 2 mo. from ages 3 to 17 years. From a sample of 200 children in Grade 2, the coefficients of split-half reliability and the Cronbach α were estimated at .77 and .69, respectively. From a sample of 492 with 82 children in each of six grades, and another sample of 612 with 102 children in each grade, the test-retest and interrater reliability coefficients were estimated at .91 and .96, respectively. In an analysis of concurrent validity, the correlation coefficient with the Bender Gestalt Test was –.65. The correlation of scores with chronological age was .86, the correlation with WISC–R IQ standard score was .47, that with performance in Chinese language class was .57, and the correlation with the performance in mathematics class was .66 (all ps < .01; Beery, 1997). Procedure

After obtaining the permission of schools, class teachers, parents, and the children, the Visual Motor Integration test was administered in the morning to the whole class. Following the criteria listed in the manual, all of the answer sheets were scored by the first author of this article who was blind to group classification and to the children’s age. The answer sheets from 30 children were sampled randomly from the total group and

TABLE 1The Recruited Participants

Handwriting Dysfunction

Number of Participants Age, mo.Boys Girls Total M SD

No 306 252 558 93.08 3.37Yes 20 21 41 93.25 3.55Total 326 273 599 93.09 3.39


rescored by the second author to test the interrater reliability of the test. Analysis

This study used SPSS-PC (Version 10.0) for the receiver operating characteristic (ROC) and other statistical analyses. In this study, sensitiv-ity is defined as, for handwriting dysfunction, the proportion of children with a positive Visual Motor Integration test result, while specificity is defined as, in cases without handwriting dysfunction, the proportion of children with a negative result on the Visual Motor Integration test. In ROC analysis, the amount of a true positive fraction (sensitivity) and a false positive fraction (1-specificity) at different cutoff points of test scores would be depicted to identify the most appropriate cutoff as the point at which both sensitivity and specificity approached the value of 1 (Pagano & Gauvreau, 2000).

In the first step of the analysis, the area under the ROC curve (AUC) was computed to examine the discriminant validity of the Visual Motor Integration test in screening children for handwriting dysfunction (the larger the ROC, the more accurate). Second, during analyses for cutoff points, in addition to adopting –1, –1.5, and –2 standard deviations as of-ten recommended for developmental tests, values were extended 0.5 stan-dard deviations upwards because the Visual Motor Integration test was not designed to assess handwriting dysfunction.

The other indicators most often used in diagnostic testing, i.e., posi-tive likelihood ratio, negative likelihood ratio, hit rate, and Youden’s index, were applied to specify the optimal cutoff point of Visual Motor Integra-tion test score for screening children with handwriting dysfunction. The positive likelihood ratio reflects how much the odds of the dysfunction in-crease when a test is positive. The negative likelihood ratio refers to how much the odds of the dysfunction decrease when a test is negative. The hit rate represents the proportion of respondents who are correctly diagnosed with or without a dysfunction in the total group; this can be used as an estimate of agreement between the screening test and the confirming di-agnoses by other means. Youden’s index is the probability of the accurate test score after removing both false positives and negatives, to measure the overall effectiveness of the sensitivity and specificity. The odds ratio com-pares whether the probability of a certain event is the same for two groups. It is simply a ratio of odds of an event occurring in the exposed versus the unexposed groups. In this study, handwriting dysfunction was the event whether it occurred in the group with positive test scores.

ResultsInterrater Reliability and Discriminant Validity

On the interrater reliability test, a Pearson correlation of the two sets


of scores rated by two different examiners for 30 randomly selected test sheets was analyzed. The correlation coefficient value was .88 (p < .001).

Regarding the test of discriminant validity, standard scores on the Vi-sual Motor Integration test obtained from two groups of children with or without handwriting dysfunction were examined. The group with hand-writing dysfunction had a mean of 77.1 (SD = 10.6); among children with-out handwriting dysfunction, the mean score was 95.1 (SD = 9.6). Fig. 1 shows the histograms of standard scores on the Visual Motor Integration

0

20

40

60

80

100

60 65 70 75 80 85 90 95 100 105 110 115 120 125 130

Freq

uenc

y

VMI Standard Scores

120

Fig. 1. Histogram of VMI standard scores of children with () or without () hand-writing dysfunction

test, calculated separately for children with and without handwriting dys-function. However, there was still overlap between the distributions of the two groups and further analyses were necessary. The values for the ROC analysis in Fig. 2 indicate that the area under the ROC curve (AUC) was .89. The area represents the validity of the test as well as the discriminant power. As recommended by Chong and Karlberg (2004), when AUC is larger than .80, the test possesses usefulness for diagnosis or screening. In the present analysis, the discriminant power of Visual Motor Integra-tion scores in screening handwriting dysfunction met this fundamental requirement.

S-H. CHANG & N-Y. YU776 VISUAL MOTOR INTEGRATION, HANDWRITING DYSFUNCTION 777

Visual Motor Integration Cutoff Score Based on the ROC analysis, the most appropriate cutoff point should

be that with both a higher sensitivity and a lower false positive fraction. As shown in the ROC curve in Fig. 2, the cutoff point with a standard score of 85 (–1 standard deviation) was optimal.

Table 2 depicts the results of the sensitivity, specificity, false posi-tive fraction, positive likelihood ratio, negative likelihood ratio, hit rate, Youden’s index, and odds ratio (OR) at four different cutoff points (–0.5, –1.0, –1.5, –2.0 standard deviation). The odds ratio shows that the 99% confidence intervals for all four cutoff points do not include the ratio of 1.00; thus, the four cutoff points can statistically significantly screen out those children more likely to have handwriting dysfunction. Among these four cutoff points, the one at 92.5 (–0.5 standard deviation) has the highest sensitivity (.93) and the lowest negative likelihood ratio (.14). However, it has the lowest specificity, positive likelihood ratio, and hit rate, with the highest false positive fraction. The sensitivity, specificity, and hit rate of the cutoff point 85 (–1 standard deviation) are all greater than .80, with the highest Youden’s index (.70). Although the cutoff point of 77.5 (–1.5 stan-

0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.00

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1.0

SS = 92.5

SS = 85

SS = 77.5

SS = 70

1-Specificity

Sen

sitiv

ity

Fig. 2. Values of ROC analysis in the four cutoff points of the Visual Motor Integration test

S-H. C

HA

NG

& N

-Y. YU

776V

ISUA

L MO

TOR IN

TEGR

ATIO

N, H

AN

DW

RITIN

G D

YSFUN

CTIO

N777

TABLE 2The Statistical Measures at Four Different Cutoff Points

Cutoff point, SD units SS Handwriting Dysfunction SEN SPE FPF LR+ LR- Hit YI OR*Yes No Total

–0.5 < 92.5 38 (a) 226 (c) 264 .927 .595 .405 2.29 .12 .62 .52 18.61 ≥ 92.5 3 (b) 332 (d) 335

–1.0 < 85 34 (a) 70 (c) 104 .829 .875 .125 6.61 .20 .87 .70 33.86 ≥ 85 7 (b) 488 (d) 495

–1.5 < 77.5 27 (a) 11 (c) 38 .659 .982 .018 36.75 .35 .96 .64 105.69 ≥ 77.5 14 (b) 547 (d) 561

–2.0 < 70 8 (a) 2 (c) 10 .195 .996 .004 54.44 .81 .94 .19 67.39≥ 70 33 (b) 556 (d) 589

Total 41 558 599Note.—a, b, c, and d are used to index the terms in the formula for calculating the statistical measures. SS = Standard Score; SEN = sensitivity = a/(a+b); SPE = specificity =d/(c + d); FPF = false positive fraction = 1-SPE; positive likelihood ratio (LR+) = SEN/(1-SPE); LR– = negative likelihood ratio = (1-SEN)/SPE; Hit = (a + d)/(a + b + c +d); Youden’s index (YI) = SEN + SPE-1; OR = odds ratio = (a × d)/(b × c). *No 99%CIs of ORs included OR = 1.00.


dard deviation) has the highest hit rate and odds ratio as well as higher specificity and positive likelihood ratio, the sensitivity is only .66. The cut-off point 70 (–2 standard deviation) has the highest specificity (.996) but the lowest sensitivity (.195), so setting a cutoff point of 70 (–2 standard de-viation) as optimal was excluded.

The other three cutoff points were compared in advance. Based on the recommendation of screening children at risk, sensitivity should be above .75 (Barnes, 1982). If the cutoff point is set at 92.5 (–0.5 standard deviation) or 85 points (–1 standard deviation), the sensitivities are .927 and .829, re-spectively, and both meet the recommended criterion. If setting the cutoff point at 77.5 (–1.5 standard deviation), sensitivity is .659, which does not meet the criterion. For specificity, the recommended criterion should not be lower than .80 (Carran & Scott, 1992). If the cutoff point is set at 77.5 (–1.5 standard deviation) or 85 (–1 standard deviation), the specificities would be .982 and .875, respectively, both of which meet the recommend-ed criterion, but the specificity for the cutoff point set at 92.5 (–0.5 stan-dard deviation), then, is only .54, lower than the criterion recommended by Carran and Scott (1992). For these three cutoff points, only the score of 85 (–1 standard deviation) meets both criteria.

The interpretation of the positive and negative likelihood ratio has been presented by Jaeschke, Guyatt, and Sackett (1994). Based on their suggested criteria, the positive likelihood ratio value obtained at a cut-off point of 77.5 (> 10) achieved the best prediction and the one at a cutoff point of 85 could generate moderate prediction (5 to 10). The negative like-lihood ratio values of those with cutoff points at 85 and 92.5 show moder-ate prediction (0.1 to 0.2) while 77.5 is less (0.2 to 0.5). The higher the posi-tive likelihood ratio, the more effective the Visual Motor Integration test is in screening handwriting dysfunction. The lower the negative likelihood ratio, the more effective the test is for screening out absence of dysfunc-tion. For a screening test, a lower negative likelihood ratio value is espe-cially important because it means the probability of children with hand-writing dysfunction not being identified is reduced.

Moreover, when examining the hit rate and Youden’s index, the val-ues are the lowest of all, as the cutoff point is set at 92.5 (–0.5 standard deviation) with a hit rate which does not meet the recommendation pro-posed by Straus, Richardson, Glasziou, and Haynes (2005), i.e., larger than 0.8. The hit rates of the other two points (–1 and –1.5 standard deviation) are both above 0.8 and the highest Youden’s index (.70) can be obtained when setting the standard deviation at –1.

DiscussionMost researchers have stated that children’s handwriting performance

cannot be fully characterized with a single test. Most researchers recom-


mend taking environmental aspects into account during clinical evalua-tion. Especially with standardized testing, clinical observations should be included (Rodger, Ziviani, Watter, Ozanne, Woodyatt, & Springfield, 2003). However, for the purpose of screening, if an easy, time-saving, and valid screening tool is available, efforts of those who teach or evaluate children at risk of handwriting dysfunction would be supported.

In previous studies on the evaluation of handwriting dysfunction, the Visual Motor Integration test scores have been more highly correlated, especially for slow handwriters at a younger age (Tseng & Chow, 2000). The dysgraphic children had a mean Visual Motor Integration test score significantly lower than the mean for those without handwriting dys-function (Tseng & Murray, 1994; Lin, 2000). It was made clear that slow handwriters seem to rely more on visually directed processes, including sequence memory and visual-motor integration. In this study, the area un-der the ROC curve was .894 for the Visual Motor Integration test in screen-ing handwriting dysfunction, and so may provide an option for reliable screening of young children with handwriting dysfunction. Although no assumption of causality for visual motor integration and handwriting per-formance is possible, the test seems useful in assessment of handwriting dysfunction.

In this study, a cutoff point of 85 (–1 standard deviation) was an effec-tive screen for children at risk of handwriting dysfunction. This finding is consistent with that of Williams, Zolten, Rickert, Spence, and Ashcraft (1993), i.e., that the cutoff point of the Visual Motor Integration test should be set at 1 standard deviation below the mean score for effective screening of children. In 146 schoolchildren with learning disabilities, emotional, or behavioral problems, they found for screening with the cutoff point set at 85 (–1 standard deviation), that the OR was 7.6 (95%CI = 2.43–23.75). Since the interval does not contain the ratio of 1.0, the result supported the use of the Visual Motor Integration test in screening slower handwriters. They also recommended that further in-depth evaluation would be required for children with scores lower than 85.

Other researchers have reported that, even though Visual Motor In-tegration scores were significantly different in children with and without handwriting problems, the scores of the children with handwriting prob-lems still fell within the normal range (Weintraub & Graham, 2000; Goy-en & Duff, 2005). However, the present study showed that children with handwriting dysfunction had a mean Visual Motor Integration standard score of 77, which is more than 1.5 standard deviations below the mean score. Goyen and Duff (2005) used a cutoff point at 1 standard deviation below the mean score and reported low discriminant validity for children in Grades 4 to 6 with handwriting problems (Sensitivity, 34%); hence, they


did not recommend using the test for the assessment of handwriting dys-function on older schoolchildren. The present finding is quite different from that of Goyen and Duff (2005), perhaps reflecting differences in the ages of participants, sample size, or experimental design. The study of Goyen and Duff (2005) included children in Grades 4 to 6, while this work recruited from Grade 2. Since the Visual Motor Integration test is a devel-opmental test, it is linked to chronological age. Visual motor integration skills may be crucial when children start to learn handwriting, but reli-ance on visual feedback would be reduced for older children; therefore, the screening effectiveness of the Visual Motor Integration test might de-crease with age (Weintraub & Graham, 2000).

This study is the first time ROC and related statistical analyses have been used to explore the discriminant validity of the Visual Motor Inte-gration test in screening handwriting dysfunction and setting an optimal cutoff point in advance. However, the value is not an absolute criterion, since the results of any ROC analysis would differ depending on the char-acteristics of the recruited sample. As this research included children in Grade 2, there are major concerns. If the Visual Motor Integration test is used as a screening tool for handwriting dysfunction, it should be imple-mented for students at a younger age to permit earlier intervention. Since Grades 1 and 2 are the first two years of formal handwriting instruction, the teachers of Grade 2 classes would have already spent more than one year with the children and should have better knowledge of the children’s handwriting.Clinical Implications

Results of this study support the use of the Visual Motor Integration test as a screening tool for children with handwriting dysfunction. The data of this study suggest that the children with a VMI standard score of less than 85 should be further assessed. In addition to the cutoff point of 85, based on the present result from this study, the clinical implications of the other cutoff points are such that, if the children’s Visual Motor Integra-tion scores were higher than 92.5, the number of those with handwriting dysfunction is quite small, lower than 1% (3/335); however, if the Visual Motor Integration test cutoff score is lower than 77.5, the rate of those with handwriting dysfunction would be more than 71% (27/38), for whom the development of handwriting skills should not be neglected. Limitations

A possible limitation of this study is that the inclusion of subjects was based on referral by classroom teachers through practical concerns rather than from formative handwriting performance evaluations. By recruiting children in Grade 2 instead of Grade 1, the difficulty in identifying hand-


writing dysfunction can be greatly reduced. However, the inference of ef-fectiveness of the Visual Motor Integration test as a valid tool for hand-writing dysfunction may be limited to this specific age. Further research on screening criteria with children of different ages is needed.

REFERENCES

Barnes, K. E. (1982) Preschool screening: the measurement and prediction of children at-risk. Springfield, IL: Thomas.

Beery, K. E. (1997) The Beery-Buktenica Developmental Test of Visual–Motor Integration: administration, scoring and teaching manual. (4th ed., rev., H. S. Liu & L. Lu, Transl.) Taipei, Taiwan: Psychological Publishing.

Berninger, V. W., & Ruberg, J. (1992) Relationship of finger function to beginning writing: application to diagnosis of writing disabilities. Developmental Medicine and Child Neurology, 34, 198-215.

Carran, D. T., & Scott, K. G. (1992) Risk assessment in preschool children: research implications for the early detection of educational handicaps. Topics in Early Child-hood Special Education, 12, 196-211.

Chang, S-H., & Yu, N-Y. (2005) Evaluation and classification of types of Chinese handwriting deficits in elementary schoolchildren. Perceptual and Motor Skills, 101, 631-647.

Chong, D. S. Y., & Karlberg, J. (2004) Refining the Apgar score cutoff point for new-borns at risk. Acta Paediatrica, 93, 53-59.

Cornhill, H., & Case-Smith, J. (1996) Factors that relate to good and poor handwrit-ing. American Journal of Occupational Therapy, 50, 732-739.

Daly, C. J., Kelley, G. T., & Krauss, A. (2003) Relationship between visual-motor inte-gration and handwriting skills of children in kindergarten: a modified replication study. American Journal of Occupational Therapy, 57, 459-462.

Denton, P. L., Cope, S., & Moser, C. (2006) The effects of sensorimotor-based inter-vention versus therapeutic practice on improving handwriting performance in 6- to 11-year-old children. American Journal of Occupational Therapy, 60, 16-27.

Diekema, S. M., Deitz, J., & Amundson, S. J. (1998) Test-retest reliability of the evalu-ation tool of children’s handwriting manuscript. American Journal of Occupational Therapy, 52, 248-255.

Doyle, S., & Goyen, T. A. (1997) Measuring the difference: evaluation change for chil-dren with handwriting difficulties. Australian Association of Occupational Therapists 19th National Conference Proceedings, Melbourne, Australia. Pp. 269-273.

Feder, K., Majnemer, A., & Synnes, A. (2000) Handwriting: current trends in occupa-tional therapy practice. Canadian Journal of Occupational Therapy, 67, 197-204.

Feder, K. P., Majnemer, A., Bourbonnais, D., Blayney, M., & Morin, I. (2007) Hand-writing performance on the ETCH–M of students in a grade one regular educa-tion program. Physical and Occupational Therapy in Pediatrics, 27, 43-62.

Goyen, T. A., & Duff, S. (2005) Discriminant validity of the Developmental Test of Visual Motor Integration in relation to children with handwriting dysfunction. American Journal of Occupational Therapy, 52, 109-115.

Jaeschke, R., Guyatt, G. H., & Sackett, D. L. (1994) Users’ guides to the medical litera-ture: III. How to use an article about a diagnostic test. B. What are the results and will they help me in caring for my patients? The evidence-based medicine work-ing group. Journal of the American Medical Association, 271, 703-707.


Jongmans, M. J., Linthorst-Bakker, E., Westenberg, Y., & Smits-Engelsman, B. C. M. (2003) Use of a task-oriented self-instruction method to support children in pri-Use of a task-oriented self-instruction method to support children in pri-mary school with poor handwriting quality and speed. Human Movement Science, 22, 549-566.

Lin, C. H. (2000) The investigation of factors related to the handwriting problems for first grade students. Taipei, Taiwan: National Science Council. (Rep. No. NSC 89-2413-H-018-012)

Maeland, A. E. (1992) Handwriting and perceptual-motor skills in clumsy, dysgraph-ic, and “normal” children. Perceptual and Motor Skills, 75, 1207-1217.

Pagano, M., & Gauvreau, K. (2000) Principles of biostatistics. (2nd ed.) Pacific Grove, CA: Duxbury.

Rodger, S., Ziviani, J., Watter, P., Ozanne, A., Woodyatt, G., & Springfield, E. (2003) Motor and functional skills of children with developmental coordination disorder: a pilot investigation of measurement issues. Human Movement Science, 22, 461-478.

Stefansson, T., & Karlsdottir, R. (2003) Formative evaluation of handwriting qual-ity. Perceptual and Motor Skills, 97, 1231-1264.

Straus, S. E., Richardson, W. S., Glasziou, P., & Haynes, R. B. (2005) Evidence-based medicine: how to practice and teach EBM. (3rd ed.) Edinburgh, Scotland: Churchill Livingstone.

Sudsawad, P., Trombly, C. A., Henderson, A., & Tickle-Degnen, L. (2001) The rela-tionship between the Evaluation Tool of Children’s Handwriting and teachers’ perceptions of handwriting legibility. American Journal of Occupational Therapy, 55, 518-523.

Tseng, M. H., & Cermak, S. A. (1993) The influence of ergonomic factors and percep-tual-motor abilities on handwriting performance. American Journal of Occupational Therapy, 47, 919-926.

Tseng, M. H., & Chow, S. M. K. (2000) Perceptual-motor function of school-age chil-dren with slow handwriting speed. American Journal of Occupational Therapy, 54, 83-88.

Tseng, M. H., & Murray, E. A. (1994) Differences in perceptual-motor measures be-tween good and poor writers. Occupational Therapy Journal of Research, 14, 19-36.

Wang, T. M., & Liao, H. F. (2007) Assessment accuracy and cutoff points of Com-prehensive Developmental Inventory for Infants and Toddlers (CDIIT). Bulletin of Special Education, 32, 1-15.

Watkins, M. W., Kush, J. C., & Schaefer, B. A. (2002) Diagnostic utility of the Learning Disability Index. Journal of Learning Disabilities, 35, 98-103.

Weil, M. J., & Amundson, S. J. C. (1994) Relationships between visuomotor and hand-writing skills of children in kindergarten. American Journal of Occupational Therapy, 48, 982-988.

Weintraub, N., & Graham, S. (2000) The contribution of gender, orthographic, finger function, and visual-motor processes to the prediction of handwriting status. Oc-cupational Therapy Journal of Research, 20, 121-140.

Williams, J., Zolten, A. J., Rickert, V., Spence, G. T., & Ashcraft, E. W. (1993) Use of nonverbal tests to screen for writing dysfluency in school-age children. Perceptual and Motor Skills, 76, 803-809.

Accepted October 23, 2009.

chang & yu (2009)

Documents