Occupational therapists have long been committed to the assessment and treatment

Transcription

1 Reliability and Validity of the Test of Visual-Perceptual Skills (Non-Motor) Revised for Chinese Preschoolers Phoebe L. C. Chan, Susanna M. K. Chow OBJECTIVE. The reliability and validity of the Test of Visual-Perceptual Skills (Non-Motor) Revised (TVPS R) were examined for its usefulness on Hong Kong Chinese preschoolers. METHOD. Content validity was evaluated by six experts. Test retest reliability, construct validity, and criterion validity were examined on 66 typically developing preschoolers while concurrent validity was examined on these preschoolers plus 52 preschoolers with visual-perception difficulties. RESULTS. Test retest reliability was high for the total score (intraclass correlation coefficient [ICC] =.88), but not for all subtests (ICC ranged.38 to.77) or for individual items (mean kappa = 0.32). The standard error of measurement (SEM; 1.53) and internal consistency (Cronbach s alpha =.90) were satisfactory. Evidence supporting the test s validity included a significant developmental trend (F = 4.99, p <.001), a lack of gender bias (F =.04, p =.84), and positive known-group differentiation (Wilks s lambda = 52.42, p <.001). The correlation between the Motor-Free Visual-Perceptual Test Revised composite score and the TVPS R composite was moderate at r =.60. CONCLUSIONS. When based on the total scores, the TVPS R was reliable and valid. The use of subtest scores and item scores for decision making or treatment planning is not recommended. Chan, P. L. C., & Chow, S. M. K. (2005). Reliability and validity of the Test of Visual-Perceptual Skills (Non-Motor) Revised for Chinese preschoolers. American Journal of Occupational Therapy, 59, Phoebe L. C. Chan, MSc, BSc, is Occupational Therapist, Occupational Therapy Department, Kowloon Hospital, Mong Kok, Hong Kong, Special Administrative Region, China. Susanna M. K. Chow, PhD, MA, MSc, BSc, was Assistant Professor, Department of Rehabilitation Sciences, The Hong Kong Polytechnic University, Hung Hom, Hong Kong Special Administrative Region, China; susannamkchow@yahoo.ca Occupational therapists have long been committed to the assessment and treatment of visual-perception dysfunction because of the importance of visual perception for competent occupational performance. Schneck (2001) contended that visual-perception deficits negatively impact on children s self-care, work, and leisure performance. In self-care tasks, children may have difficulties manipulating eating utensils or shoelaces. Work and productive activities such as house chores and school activities also present problems for those with visual-perceptual problems. For example, these children may be slow to master the alphabets and numbers. In play, children with visual-perceptual problems may demonstrate difficulties in negotiating obstacles in the playground and doing puzzles. Tseng and Chow (2000) found visual sequential memory to be one of the best predictors for handwriting speed in children with handwriting dysfunctions. In order to identify visual-perceptual factors that limit occupational performance and participation, occupational therapists often ask and observe how visual-perceptual difficulties affect daily occupation. A norm-referenced test is almost always used in addition to interviews and clinical observations. Of the normreferenced tests available, the Test of Visual-Perceptual Skills (Non-Motor) Revised (TVPS R; Gardner, 1996) is widely used by occupational therapists, including those in Western countries (Schneck, 2001) and in Hong Kong (Child and Adolescent Psychiatry Project Group, 1998). Although there is extensive The American Journal of Occupational Therapy 369

2 documentation of how cultural experiences influence visual perception (see for example, Levine, 1987; Zimbardo, 1992), little has been done to examine if visual-perceptual tests such as the TVPS R, which was developed for American children, are valid for children with other cultural backgrounds. Hence, the purpose of the present study was to investigate the usefulness of the test for Chinese children in Hong Kong. The TVPS R assesses visual-perceptual strengths and weaknesses of children 4 through 12 years and 11 months of age. The test consists of 112 items grouped in 7 subtests (i.e., visual discrimination, visual memory, visual-spatial relationships, visual-form constancy, visual-sequential memory, visual figure ground, and visual-closure). Each subtest consists of simple verbal instructions of either one or two sentences (e.g., Look at this form and remember it so that you find it on another page. Find it among these forms. ). In all subtests, children were required to indicate the correct answer out of either four or five choices by any means (e.g., by pointing or verbal indications). The child is allowed 15 seconds for each item. Each subtest ends when the child makes three errors on four consecutive items on a subtest with four choices, or four failures on five consecutive items on a subtest with five choices. The raw scores of each subtest and all of the subtests were summed to yield subtests and total scores, respectively. The norms established on 1,032 nondysfunctional American children recruited from regular classrooms were used to yield scaled scores, perceptual quotients, and percentile ranks. The racial and ethnic sample of participants was representative of the 1990 U.S. Census (Gardner, 1996) and consisted of 62.1% white, 14.8% black, 13.1% Hispanic, 8.2% Asian, and 1.8% other. A number of positive psychometric findings were reported in the manual and in two independent studies. This includes a wide range of item passing rate (between.35 and.96) (Gardner), high internal consistency (Cronbach s alpha between.83 to.91) (Gardner) and small standard error of measurement (SEM ) (between 4.48 to 6.20) (Gardner). A lack of gender bias (Gardner), low to moderate intersubtests correlation (between.18 to.51) (Klein, Sollereder & Gierl, 2002), moderate to high subtest-total correlation (between.59 to.76) (Klein et al.), and low to moderate correlation with intellectual tests (between.37 to.52) and visual-motor tests (.42) (Klein et al.) were also reported. Test retest reliability was not reported in the manual but was examined by McFall, Deitz, and Crowe (1993) on 30 children 6 to 8 years of age with learning disabilities. They reported satisfactory results for the total test scores (intraclass correlation coefficient [ICC] =.81) but not for all subtests. Item-level reliability was not studied. In addition to the lack of item-level information available, a few other areas deserve attention. First, the studies of both Klein et al. (2002) and McFall et al. (1993) were conducted on children 6 years of age or older, but not on younger children who were likely to be more difficult to test reliably as they were less able to concentrate or to follow instructions (Sattler, 1992). Second, although the test results were shown to be lower in children with learning disabilities (Gardner, 1996), the known-group comparison had not been investigated on children with identified visual-perceptual difficulties. Mercer (1997) noted that the Individuals with Disabilities Education Act Amendments of 1991 did not consider visual-perceptual dysfunction a prominent or diagnostic feature of learning disabilities. Third, the test has not been examined against other measures of nonmotor visual perception to provide important information on convergent validity. Fourth, as mentioned briefly earlier, visual-perception performance is influenced by our cultural and environmental experiences (Levine, 1987; Zimbardo, 1992). There are at least two major cultural and environmental experiences that are different between American and Hong Kong children. First, children in Hong Kong are exposed to a more varied written language system as they are biliterate in Chinese and English (Chow, Choy, & Mui, 2003). Whereas Chinese is predominately orthographic in nature, English is made up of strings of alphabetical scripts. Also, children in Hong Kong are exposed to writing script much earlier as most children begin learning to write at 3 years of age (Opper, 1996). Because the TVPS R was developed for and on children in the United States, different culture experiences may affect the usefulness of the test for Hong Kong children, for example, in terms of the relevance or representativeness of the items. Taking these into account, the present study aimed to provide a review of the psychometric characteristics of the TVPS R on children 4 to 5 years of age. Specifically, the study focused on four psychometric areas: (1) content validity (i.e., the degree to which the items were relevant, representative, and comprehensive) (2) reliability (i.e., test retest reliability and internal consistency), (3) construct validity (i.e., item passing rate, developmental changes, gender bias, and known-group comparison), (4) criterion validity (i.e., convergent validity with the Motor-Free Visual-Perception Test Revised [MVPT R]). The test retest reliability and internal consistency were examined on item scores in addition to total and subtest scores. 370 July/August 2005, Volume 59, Number 4

3 Methodology Participants Content Validity. According to Geisinger (1994), an expert panel review is an effective technique to evaluate content validity. A panel of six experts including four occupational therapists and two clinical psychologists, recruited by purposive sampling, served as participants to examine the content validity of the test. The panel members had 6 to 13 years of experience with children with visual-perceptual problems, and were familiar with the TVPS R. Reliability. Using the Statistical Design Analysis System and setting.7 as the reliability coefficient and 95% as the level of confidence, the minimum number of participants required was 65. A stratified random sampling method was used to select the sample. The sampling process was conducted in three stages. First, 4 of the 18 districts in Hong Kong were randomly selected. Then, one kindergarten was selected within each selected district. Finally, 18 children were randomly drawn from each selected kindergarten based on age and gender as well as on the inclusion criteria. Nine children (five girls and four boys or four girls and five boys) from the group of 4-year-olds (4 years 1 month to 5 years 0 months of age) and nine children (four girls and five boys or five girls and four boys) from the group of 5-yearolds (5 years 1 month to 6 years 0 months of age) were randomly selected from each kindergarten. None of the selected children had been diagnosed with physical, intellectual, or sensory impairments, all were Cantonese-speaking and were attending mainstream schools, and none had any identified special educational needs. The parents or guardians of the children gave their written consent prior to the assessment. Two participants withdrew from the study and 4 other participants data were excluded as they showed poor comprehension of the instructions during the test. Hence, the total number of participants was 66. There was an equal number of 4- and 5-year-olds (mean age = 5.16, SD =.55) and an equal number of boys and girls. Construct and Criterion Validity Children with and without visual-perceptual deficits were recruited. This included the 66 children without visual-perceptual deficits who participated in the reliability study. In addition, 52 children with visual-perceptual deficits were also recruited through occupational therapy clinics by purposive sampling method. These children met the following selection criteria. They: (1) were 4 years 0 months to 6 years 0 months of age, of either gender, (2) had visual-perception problems that had been identified by an occupational therapist and confirmed with the MVPT R, (3) were able to follow simple instructions and indicate correct answers, and (4) had not been diagnosed with mental retardation, cerebral palsy, epilepsy, or autism. In addition, if primary auditory or visual deficits were reported, these were corrected. The participants came from three occupational therapy clinics and were diagnosed with developmental delay. Thirty-seven participants were male and 15 participants were female, with a mean age of 5.11 years (SD =.54). Instruments A content validity questionnaire was designed to obtain expert judgment on the quality of the test with a 4-point Likert scale with 1 being strongly disagree and 4 being strongly agree. The experts were asked to rate the relevance and comprehensiveness of each of the seven subtests. The overall representativeness of the whole test for visual perception was also rated. In addition to the quantitative rating, comments and recommended modifications were invited. The TVPS R consists of a test book on which children indicate their answers. These were recorded by the tester onto a record form. The Chinese instruction established by Chan (2002) through forward and backward translation was used in this study. For concurrent validity, the MVPT R (Colarusso & Hammill, 1996) was used. MVPT R is a standardized test containing 40 items that can be administered individually or in a group. The test included items from five visualperception areas (i.e., spatial relationships, visual discrimination, figure ground, visual-closure, and visual memory) and yields a general visual-perceptual score. The MVPT R was standardized on 912 American children 4 to 11 years of age. Positive psychometric properties of the test included satisfactory item difficulty level ( ), total-item correlation ( ), test retest reliabilities (.77.83), and internal consistency (minimum.71). The Chinese instructions for the MVPT R developed by Wong (1999) were used in this study. Procedure The panel members were invited to participate in the study after an explanation of the purpose of the study. They were then sent the content validity questionnaire, an invitation letter, and a consent form. The panel members were asked to complete the questionnaire individually to ensure the independence of the opinions, and were requested to return the questionnaire in 2 weeks. After the content validity study, a pilot test was conducted with 10 children to refine the testing procedures prior to the other studies. The American Journal of Occupational Therapy 371

4 Permission to recruit participants with typical development and visual-perceptual deficits was obtained from the persons in charge of the kindergartens and clinics, respectively. An invitation letter to the center in charge, an invitation letter to the children s parents or guardians, and a consent form were sent to the school or the occupational therapists of the center. The first author assessed all of the selected children individually. All directions for administrating the TVPS R followed the guidelines stipulated in the TVPS R manual, including those on testing time, test environment, and the participant preparation procedures. The test was administered again in 2 weeks at the same time of the day as the first test, and the same procedures were adopted by the same tester. Data Analysis The content validity, mean, standard deviation, and percentage of agreement were calculated from the experts ratings on the 4-point scale. The comments made were also analyzed qualitatively. The test retest reliabilities of the total and subtest scores of the nondysfunctional group were estimated by a two-way mixed model of the ICC. According to Portney and Watkins (2000), coefficients below.75 suggested poor to moderate reliability while those above.75 were considered good. To examine the test retest reliability of each item, the kappa statistics were used. values of.8,.6 and.4 are cutoffs for good, moderate, and poor levels of agreement (Portney & Watkins, 2000). However, since TVPS R items were scored as either pass or fail, a deflated value could result even when the agreement was high (Cicchetti & Feinstein, 1990). Therefore, the percentage of agreement was also presented, even though the latter can be inflated by chance agreement. Construct validity was examined on children without visual-perceptual deficits, whereas concurrent validity was examined on children both with and without visual-perceptual deficits. For internal consistency, Cronbach s alpha was used to evaluate the homogeneity of the subtests. Acceptable alpha coefficients are between.70 and.90 (Portney & Watkins, 2000). To examine developmental (age) changes, gender bias, and the possible interaction of these two factors in the nondysfunctional group, multiple analysis of variance (MANOVA) was adopted. Concurrent validity with MVPT R was estimated with Pearson correlations. According to Portney and Watkins, correlations ranging from 0.00 to.25 indicate little or no relationship; those from.25 to.50 suggest a fair degree of relationship; values of.50 to.75 are moderate to good; and values of above.75 are considered good to excellent. The knowngroup comparison was based on multiple analysis of covariance (MANCOVA) with age being the covariate. Results Content Validity As shown in Table 1, the mean ratings assigned by the panel members on the relevance and comprehensiveness of each of the seven subtests ranged from 3 to At least 83.3% of all ratings were 3 or above. In terms of the representativeness of the items, the mean rating was 3, with 83.3 the panel members again rating 3 or above. Qualitatively, the raters comments varied and were not shared by others. For example, one rater commented on the difficulty of items in terms of form constancy, sequential memory, and visual-closure, whereas another rater viewed the items to be too easy in visual discrimination and spatial relationship. Test Retest Reliability The ICC of the test retest reliability for the total scores was.88 (p <.001), with a 95% confidence interval ranging from.81 to.92, indicating good test retest reliability. The ICC for each subtest ranged from.38 to.77 (see Table 2). Only one subtest (visual-sequential memory) showed good stability over time. The reliabilities of the other six subtests were moderate to poor. Again from Table 2, only 6 items showed good reliability over time (all k = 1), 3 items showed a moderate level of reliability (.64.66), whereas the remaining 103 items showed poor reliability (k <.60). The overall mean kappa was also low at.32. In terms of the percentage of agreement, the mean for all the items was 76.6%. Of the 112 items, 39 had an agreement of more than 80%, 67 items had an agreement of between 60% and 80%, and 6 items of below 60%. Internal Consistency The internal consistency of the TVPS R was excellent with a Cronbach s alpha of.90. The subtest total score correlation was moderate to good (ranging from.53 to.86). At the item level, the Cronbach s alphas of the seven subtests were good (.72 to.91). However, the corrected item-subtest correlations were mostly poor. As can be seen from Table 3, almost 50% (54) of the items had a poor correlation with their respective subtests (<.5), and only three showed an acceptable correlation (i.e., (.7). The Standard Error of Measurement The SEM of the total score was 1.53 that was 2.68 the mean. The SEM for seven subtests scores ranged from July/August 2005, Volume 59, Number 4

5 Table 1. Judges Ratings of the Relevance, Comprehensiveness, and Representativeness of the Test of Visual-Perceptual Skills (Non-Motor) Revised % Strongly % Strongly Cumulative Disagree % Disagree % Agree Agree Percentage Mean SD < 3 3 Relevance 1. Visual Discrimination Visual Memory Visual-Spatial Relationship Visual-Form Constancy Visual-Sequential Memory Visual Figure Ground Visual-Closure Comprehensiveness 1. Visual Discrimination Visual Memory Visual-Spatial Relationship Visual-Form Constancy Visual-Sequential Memory Visual Figure Ground Visual-Closure Representativeness Seven Subtests Table 2. The, Percentage of Agreement of Test Items, and Intraclass Coefficient Correlation of the Subtest Test Retest Reliability Scores (N = 66) 1. Visual Discrimination 2. Visual Memory 3. Visual-Spatial Relationship 4. Visual-Form Constancy Item k p Agreement k p Agreement k p Agreement k p Agreement 1.66 <.001*** 98.5% 1 <.001*** 100% % % ** 87.9% ** 80% %.44 <.001*** 75.8% % % ** 93.9%.35 <.001*** 83.3% % % * 80.3%.40 <.001*** 71.2% ** 77.3%.43 <.001*** 78.8% ** 74.2% ** 72.7% ** 71.2% %.44 <.001*** 74.2% ** 69.7% % % %.41 <.001*** 71.2% 8.42 <.001*** 77.3% % % * 69.7% % %.51 <.001*** 78.8% * 71.2% ** 66.7% % % * 74.2% ** 66.7% ** 68.2% ** 66.7%.65 <.001*** 90.9% * 63.6% * 63.3% * 65.2% % * 63.6% * 68.2%.39 <.001*** 74.2% % * 69.7%.39 <.001*** 69.7% % * 84.8% ** 68.2%.47 <.001*** 86.4% ** 68.2%.44 <.001*** 87.9% ** 75.8% % % % Total ICC (p).56 (<.001***).59 (<.001***).61 (<.001***).52 (<.001***) [95% CI] [.38.71] [.40.73] [.44.74] [.33.68] 5. Visual-Sequential Memory 6. Visual Figure Ground 7. Visual-Closure Item k p Agreement k p Agreement k p Agreement 1.49 <.001*** 87.9% ** 93.9% 1 <.001*** 100% * 81.8%.49 <.001*** 97%.46 <.001*** 83.3% 3.39 <.001*** 71.2%.44 <.001*** 74.2% ** 74.2% % ** 69.7%.41 <.001*** 71.2% 5.57 <.001*** 83.3% * 63.6% * 66.7% * 68.2% % * 69.7% % % * 62.1% * 62.1%.51 <.001*** 75.8% ** 74.2% 9.64 <.001*** 83.3% % % % % % <.001*** 74.2%.53 <.001*** 86.4% % <.001*** 78.8% % 1 <.001*** 100% % % * 87.9% % * 78.8% % <.001*** 84.8% 1 <.001*** 100% 1 <.001*** 100% % % 1 <.001*** 100% Total ICC (p).77 (<.001***) 49 (<.001***).38 (<.001***) [95% CI] [.65.85] [.28.66] [.16.57] Note. ICC = intraclass coefficient correlation. *.05. **.01. ***.001. The American Journal of Occupational Therapy 373

6 Table 3. Correlation of Subtest Scores With the Total Test of Visual-Perceptual Skills (Non-Motor) Revised Score and Correlation With Motor-Free Visual-Perception Test Revised Correlated Item- Correlation With Subtest Total Correlation MVPT R Visual Discrimination.831***.589*** Visual Memory.780***.243 Visual-Spatial Relationship.792***.434*** Visual-Form Constancy.529***.402** Visual-Sequential Memory.860***.186 Visual Figure Ground.612***.398** Visual-Closure.624***.284* Note. MVPT R = Motor-Free Visual-Perception Test Revised *.05. **.01. ***.001. to.41. All were below 10 the mean. The results suggested that both the total scores and subtests scores of the TVPS R had a reasonable stability level. Passing Rate An examination of the passing rate in Table 4 revealed that the items were arranged in decreasing pass rate, with only a few exceptions in each subtest. This is to be expected as children were discontinued after making a specific number of consecutive errors in the subtest. Although the visualdiscrimination, visual-memory, and visual-spatial relationship subtests were easiest (with overall proportion of passing between.69 and.60), the remaining were more difficult (.36 to.49). Developmental Changes and Gender Bias Figure 1 showed that the age change was largely linear over the preschool period and that girls performed better than boys in general. The MANOVA results confirmed a significant age effect, (F = 4.99, p <.001) but not a gender bias (F =.04, p =.84). Age by gender interaction (F =.77, p =.97) was not significant either. The parameter estimates indicated that the children s total TVPS R score improved by each year during the preschool years. Known-Group Comparison Since age had a significant effect on the test scores, this was controlled for in the known-group comparison. The MANCOVA analysis shown in Table 5 indicated that the performances of the typically developing preschoolers and the visual-perceptual deficit group differed significantly when their seven subtest scores and total scores were compared simultaneously (Wilks s lambda = 52.42, p <.001). The two groups differed on the total score (F = , p <.001) and on all seven subtests (F ranged from to , all p <.001). Concurrent Validity The correlation of the TVPS R and the MVPT R on 52 children with visual-perceptual deficits indicated a moderate relationship between the two tests (r =.60). As shown in Table 3, the correlation of the TVPS R subtests to the MVPT R ranged from a low of.19 (sequential memory) to a high of.59 (visual discrimination). Discussion With at least 83 the judges having assigned a positive score, the seven subtests of the TVPS R were, on the whole, judged to be relevant, comprehensive, and representative. It is interesting to note that little concern was expressed by the judges about the relevance of the test items for children in this study who are all Chinese. Perhaps the exposure of Hong Kong Chinese preschoolers to the English writing system in addition to the Chinese system provide visual- Table 4. Proportion of Children Passing the Test Items (N = 66) 3. Visual- 4. Visual- 5. Visual- 6. Visual 1. Visual 2. Visual Spatial Form Sequential Figure 7. Visual- Item Discrimination Memory Relationship Constancy Memory Ground Closure Subtest July/August 2005, Volume 59, Number 4

7 Figure 1. The total Test of Visual-Perceptual Skills (Non-Motor) Revised scores of boys and girls from 4 to 6 years of age. perceptual experience more akin to those of American children. Indeed Chow et al. (2003) found a very high correlation (Pearson r =.87) between the ability to produce English and Chinese scripts in their sample of 333 children. Reliability encompasses three components: precision, stability, and consistency (Sim & Wright, 2000). It can be argued that if a test cannot be shown to remain stable every time it is administered, it cannot be consistent. Similar to the findings of McFall et al. (1993) on 30 slightly older American children, our results confirmed that the retest reliability was excellent for the total score but not so for the subtests. At the item level, our results revealed rather poor reliability. Although the test manual did report poor to moderate item consistency (Gardner, 1996), only the minimum and maximum values were reported. The lack of item-level information makes it difficult for test users to appreciate the quality of the items. Also in the test manual, a low-item consistency was attributed to the inclusion of items with a high passing rate. While it is true that kappa values can be paradoxically low for items with either a very high or very low passing rate (Cicchetti & Feinstein, 1990), our results revealed that this may not be the key reason for the low reliability of the items. We found a high Table 5. Multiple Analysis of Variance Results of the Known-Group Comparison F Visual Discrimination Visual Memory Visual-Spatial Relationship Visual-Form Constancy Visual-Sequential Memory Visual Figure Ground Visual-Closure Total Score All p <.001 passing rate (.75) in only nine items, but the consistency was unacceptably low in 106 items. Even when adopting the less-stringent measure of percentage of agreement, only 35 the items were found to have an agreement of more than 80%. Poor reliability may be related to the sheer length of the test (consisting of 112 items) or to the uniform and, thus, monotonous nature of the items and the response set. On a more positive note, provided that only the total score was used, there was evidence to support the test s validity. First, the presence of a linear age trend in the preschool years concurs with the test s claim of the developmental nature of the test scores. An inspection of the passing rate of the items also confirmed that the level of difficulty of each subtest increased. This is quite important, as the validity of the procedure of terminating the test depends on the correctness of the difficulty of the items. Second, our results showing that the TVPS R score was free from gender bias supports the use of the test for children of both genders without the need for separate items or norms. The lack of gender differences was consistent with the results reported by the test manual and by studies of visual perception conducted on children in the West (Canning, Orr, & Rourke, 1980) and in Asia (Opper, 1996), as well as on adults (Trahan & Quintana, 1990). In view of the less than satisfactory reliability of the subtest and item scores, we have to concur with others that test users should avoid using subtest and item scores when making clinical decisions or planning treatments. Indeed, previous studies also questioned the validity of the subtests commonly used in tests of visual perception in general. Based on a confirmatory factor analysis, inter-item correlation and internal consistency, Klein et al. (2002) concluded that there was a lack of validity in the grouping and summing of the items into the seven subtests in the TVPS R. Treatment programs or research findings based on the validity of subcategories of visual perception, such as the ones in TVPS R, must be treated with great caution. Acknowledgments The authors gratefully thank the children, the parents, the heads, and teachers for their participation; Dr. Morrison F. Gardner and Psychological and Educational Publications for their permission to translate the test instructions into Chinese; Professor Sing Kai Lo and Mr. Peggo Lam for their advice on statistics; and Mrs. Serena Cheng, Mr. Bobby Ng, Ms. Sally Choy, Ms. Alice Kwoo, Ms. Anita Fong, Ms. Miranda Ng, Ms. Viola Wong, Ms. Theresa Chen, Mr. Ng Nai Kong, Ms. Donna Chan, Ms. Amy Chiu, and Ms. Cindy Chan for their assistance in the study. The American Journal of Occupational Therapy 375

8 References Canning, P. M., Orr, R. R., & Rourke, B. P. (1980). Sex differences in the perceptual, visual-motor, linguistic and concept formation abilities of retarded readers? Journal of Learning Disabilities, 13(10), Chan, L. C. (2002). A study of the reliability and validity of the Test of Visual-Perceptual Skills Revised for preschool children. Unpublished master s thesis, The Hong Kong Polytechnic University, China. Child and Adolescent Psychiatry Project Group. (1998). Report on child and adolescent assessment tools. Hong Kong: Hong Kong Occupational Therapy Coordinating Committee. Chow, S. M. K., Choy, S., & Mui, S. (2003). Assessing handwriting speed of children biliterate in English and Chinese. Perceptual and Motor Skills, 96(2), Cicchetti, D. V., & Feinstein A. R. (1990). High agreement but low kappa: II. Resolving the paradoxes. Journal of Clinical Epidemiology, 43(6), Colarusso, R. P., & Hammill, D. D. (1996). Motor-Free Visual- Perceptual Test Revised. Novato, CA: Academic Therapy Publications. Gardner, M. F. (1996). Test of Visual-Perceptual Skills (Non- Motor) Revised manual. Hydesville, CA: Psychological and Educational Publications. Geisinger, K. F. (1994). Cross-cultural normative assessment, translation and adaptation issues influencing the normative interpretation of assessment instrument. Psychological Assessment, 6(4), Individuals with Disabilities Act Amendments of Pub. L , 20 U.S.C Klein, S., Sollereder, P., & Gierl, M. (2002). Examining the factor structure and psychometric properties of the Test of Visual-Perceptual Skills. Occupational Therapy Journal of Research, 22(1), Levine, R. E. (1987). Culture: A factor influencing the outcomes of occupational therapy. Occupational Therapy in Health Care, 4, McFall, S. A., Deitz, J. C., & Crowe, T. K. (1993). Test retest reliability of the Test of Visual-Perceptual Skills with children with learning disabilities. American Journal of Occupational Therapy, 47(9), Mercer, C. D. (1997). Students with learning disabilities (5th ed). Upper Saddle River, NJ: Merrill. Opper, S. (1996). Hong Kong s young children: Their early development and learning. Hong Kong: Hong Kong University Press. Portney, L. G., & Watkins, M. P. (2000). Foundations of clinical research: Applications to practice (2nd ed.). Upper Saddle River, NJ: Prentice Hall Health. Sattler, J. M. (1992). Assessment of children (3rd ed.). San Diego, CA: Author. Schneck, C. M. (2001). Visual perception. In J. Case-Smith (Ed.). Occupational therapy for children (pp ). St. Louis: Mosby. Sim, J., & Wright, C. (2000). Research in health care: Concepts, designs and methods. Cheltenham, England: Stanley Thornes. Trahan, D. E., & Quintana, J. W. (1990). Analysis of gender effects upon verbal and visual memory performance in adults. Archives of Clinical Neuropsychology, 5(4), Tseng, M. H., & Chow, S. M. K. (2000). Perceptual-motor function of school-aged children with slow handwriting speed. American Journal of Occupational Therapy, 54(1), Wong, O. C. (1999). Validity and reliability study of Motor-Free Visual-Perception Test Revised. Unpublished master s thesis, The Hong Kong Polytechnic University, China. Zimbardo, P. G. (1992). Psychology and life (13th ed.). New York: Harper Collins. 376 July/August 2005, Volume 59, Number 4