Journal of Autism and Developmental Disorders, Vol. 35, No.

1, February 2005 ( 2005)

Psychometric Properties of the Chinese Version of the Psycho-educational Prole-Revised (CPEP-R)

Daniel T. L. Shek,1,4 Sandra K. M. Tsang,2 Lorinda L. Lam,3 Florence L. Y. Tang,3 and Penita M. P. Cheung3

The purpose of this study was to examine the reliability and validity of the Chinese version of the Psycho-educational Prole-Revised (PEP-R). The Chinese PEP-R (CPEP-R) was administered to 63 preschool children with symptoms of autistic disorder recruited from special child-care centers in Hong Kong. Results showed that the scales of the CPEP-R were internally consistent, reliable across raters and temporally stable. Regarding the concurrent validity of the CPEP-R, the developmental score and developmental age assessed by the CPEP-R were signicantly correlated with the MerrillPalmer Scale of Mental Tests and the Hong Kong Based Adaptive Behavior Scale. The Behavioral Scale of the CPEP-R was also signicantly related to the Childhood Autism Rating Scale. Besides replicating the ndings in the Western context, the present study suggests that the psychometric properties of the PEPR are stable across cultures and the related ndings support the cross-cultural reliability of the tool.
KEY WORDS: Psycho-educational ProleRevised (PEP-R); autistic disorder; Chinese; assessment; psychometric properties.

INTRODUCTION In Hong Kong, around 6,500 families have members with autism, which constitutes about a prevalence of 810 in 10,000 of the population (Health and Welfare Bureau, 1994, p.12). As no known cure has been found for this disorder, these individuals may become a source of disappointment to their families and burdens to the society, especially if proper and timely rehabilitation is not oered. However, psychological services for children with
1

3 4

Social Welfare Practice and Research Centre, Department of Social Work, The Chinese University of Hong Kong, Shatin, Hong Kong. Department of Social Work and Social Administration, The University of Hong Kong, Pokfulam Road, Hong Kong. Heep Hong Society, Hong Kong. Correspondence should be addressed to: Daniel T. L. Shek, Social Welfare Practice and Research Centre, Department of Social Work, The Chinese University of Hong Kong, Shatin, Hong Kong, e-mail: danielshek@cuhk.edu.hk

autistic disorders in Hong Kong were undeveloped before the 1990s. In 1995, a Hong Kong study team comprising government ofcials, senior executives of rehabilitation agencies and leading professionals in the eld visited overseas programs for autistic people, including the internationally renowned North Carolina TEACCH program (The Treatment and Education of Autistic and related Communication Handicapped Children; Schopler, 1997). While the Team fully recognized the promising rehabilitation potentials of the TEACCH program, it was explicitly stated that cultural hurdles had to be overcome before the TEACCH Program could be successfully adapted and transplanted to Hong Kong (Health and Welfare Bureau, 1995, p.10). As assessment is the rst step to devising appropriate training for autistic children, the importance of developing a valid and reliable tool is unquestionable. Unfortunately, there is a severe lack 37
0162-3257/05/0200-0037/0 2005 Springer ScienceBusiness Media, Inc.

38 of Chinese assessment tools and teaching materials that are designed specically for autistic children in Hong Kong. Furthermore, due to their non-compliance and peculiar developmental and behavioral characteristics, researchers have argued that the usefulness of standardized tests (e.g., intelligence tests) for assessing autistic children is rather limited and that the development of a local comprehensive, valid and reliable assessment tool tailor made for autistic children is highly imperative (Alpern, 1967; Lam & Rao, 1993; Parks, 1988; Rutter, 1978; Volkmar, Hoder, & Cohen, 1985). The Psychoeducational Prole-Revised (PEP-R) was developed by the TEACCH Division in North Carolina for the assessment and formulation of individualized training plan for children with pervasive developmental disorders and other related communication disorders (Schopler, Reichler, Bashford, Lansing, & Marcus, 1990). Compared with the existing standardized intellectual assessment tools that are commonly used in Hong Kong to assess autistic children, the PEP-R has several signicant advantages. These include: (a) appropriate selection of testing items and testing materials; (b) efcient sequencing and arrangement of testing items; (c) exibility in administration and tapping the ceiling of the childs performance and training potential; (d) ability to identify the idiosyncratic learning styles of the autistic child. Results of the PEP-R can be used to design individually tailored educational and behavioral intervention programs. The intervention program derived form the PEP-R is based on the premise that autism is a developmental disability that is best treated through a psycho-educational approach (Schopler & Reichler, 1979). By employing a reliable and valid assessment instrument, the resulting intervention program can provide a framework for ongoing evaluation and formulation of suitable education plan for the children. The PEP-R appeared to be a good candidate as an assessment tool to be used in the Chinese culture, especially if the cross-cultural application of the instrument could be demonstrated. Against this background, a validation study was conducted to examine the psychometric properties of the Chinese version of the PEP-R. The purpose of this paper is to report research ndings arising from this validation study. Previous Western studies have generally established the reliability and validity of the Psychoeducational Prole (Muris, Steerneman, & Ratering, 1997; Schopler & Reichler, 1979; Schopler, Reichler,

Shek, Tsang, Lam, Tang, and Cheung Bashford, Lansing, & Marcus, 1990; van Berckelaer-Onnes & van Duijn, 1993). Regarding the validity of the PEP, Schopler and Reichler (1979) reported that the Spearman correlation between the PEP Developmental Age and Mental Age of the MerrillPalmer Scale of Mental Tests was .85. With reference to inter-rater reliability, the two studies by Schopler, Reichler, Bashford, Lansing and Marcus (1990) and Muris, Steerneman and Ratering (1997) showed that the inter-rater reliability of the developmental items and the behavioral items was acceptable. There are also research ndings supporting the PEP-R outside the United States. Based on the Dutch version of the PEP-R, Steerneman, Muris, Merkelbach and Willems (1997) reported that the values of Cronbachs alpha ranged between .82 and .98 for the PEP-R scales. Similarly, van Berckelaer-Onnes and Van Duijn (1993) reported that the values of Cronbachs alpha ranged between .77 and .95 for the PEPR scales. However, one observation that can be highlighted from the literature is that inter-rater reliability of the Behavioral Scale was generally lower than that of the Developmental Scale. In addition, although the previous ndings on the inter-rater reliability of the PEP-R were encouraging, a cursory review of the previous studies showed that no information on the testretest reliability of the PEPR measures had been reported. If the PEP-R were to serve as a useful pre- vs. post-treatment outcome measure, it would be necessary to obtain information on its temporal stability. In particular, Lam and Rao (1993) argued that there was a need to establish the testretest reliability of the instrument if preand post-treatment assessment are administered by dierent persons. In short, the present study would attempt to examine the internal consistency, interrater reliability, and testretest reliability of the PEP-R scales in a Chinese context. With reference to the Chinese version of the PEP, the original Psychoeducational Prole (Schopler & Reichler, 1979) was translated by Lam and Rao (1993) who found that the developmental scores of the Chinese PEP were signicantly correlated with the MerrillPalmer Scale of Mental Tests (r = .71). With the release of the PEP-R in 1990, Heep Hong Society in Hong Kong formed a Task Force (with the rst author as the Chairman and the second, third and fourth authors as members of the Task Force) to develop the Chinese version of the PEP-R and the related materials on

Psychometric Properties of the Chinese Version teaching strategies and activities guides, which formed the Training Package for Autistic Children (TPAC). Tang (2000) made some initial eort to establish the reliability of the Behavioral Scale of the Chinese version of PEP-R. Results showed that there was support for the testretest reliability (intra-class correlation coecients ranged from .75 to .99, n=15), inter-rater reliability (intra-class correlation coecients ranged from .73 to .94, n =12) and internal consistency of the related measures. Although Tang (2000) provided pilot data, the ndings were unpublished ndings based on small samples and a more complete study of the psychometric properties of the CPEP-R was needed. In particular, there was a need to establish the validity of the CPEP-R. Against this background, the present study was conducted to examine the reliability (including internal consistency, inter-rater reliability and temporal stability) and concurrent validity of the Chinese version of the PEP-R. METHOD Instruments The Chinese version of the Psychoeducational Prole-Revised (CPEP-R) was used. The CPEP-R consists of a Developmental Scale and a Behavioral Scale derived from the PEP-R of the TEACCH Program. The Developmental Scale, with a total of 153 items, yields a developmental score that covers seven important domains of the childs development: Imitation (16 items), Perception (13 items), Fine Motor (16 items), Gross Motor (18 items), Eye-Hand Coordination (15 items), Cognitive Performance (26 items) and Cognitive Verbal (27 items). The Developmental Scale also yields a measure of developmental age. The Behavioral Scale (43 items) is used to identify the degree of behavioral abnormality and atypical behavioral characteristics of children with autism. The scale is also intended to yield information on the childs disorganized and disturbed behavior. There are four areas covered in the Behavioral Scale: Relating and Affect (12 items), Play and Interest in Materials (8 items), Sensory Responses (12 items) and Language (11 items). A working group with a psychologist, an occupational therapist, a special educator, and a special child-care worker rst translated the PEP-R items into Chinese. The draft was then sent to the Task Force for review and modication. Compared with

39 the PEP-R, major changes in the CPEP-R were in the areas of language and use of stimuli. Some adaptation and modications were made in consideration of cultural and language factors. For examples, Chinese words were used to replace the English ones in the items for letter matching, naming and sorting, and some of the pictures more culturally suitable were adapted and replaced. The translated test is administered in Cantonese. The items in the Developmental Scale were coded as Passed=2, Emerging=1, and Failed=0. The items in the Behavioral Scale were coded as Appropriate=1, Mild=2, Severe=3, and N.A.=0. To assess the concurrent validity of the CPEP-R, the following measures were employed: (1) MerrillPalmer Scale of Mental Tests (Stutsman, 1948). The scale is characterized by a wide range of visual-motor tasks and a few verbal items and it has been widely used by practitioners in assessing the cognitive functioning of pre-school children with language and cognitive delays. A smallscale local study was conducted to develop a local norm based on children from 24 months to 48 months in Hong Kong (Department of Health, 1995). This norm was used in assessing the cognitive ability of the subjects involved in this study. (2) Hong Kong Based Adaptive Behavior Scale (Kwok, Shek, Tse, & Chan, 1989). The Hong Kong Based Adaptive Behavior Scale (HKBABS) was modeled after the Vineland Adaptive Behavior Scale (Sparrow, Balla, & Cicchetti, 1984) and its norms were based on over 1000 children in Hong Kong. The scale covers four major domains of adaptive functioning: communication, daily living skills, socialization, and motor skills domains. The scale could yield supplementary information on childrens social-cognitive functioning and employment of this scale can address the criticism that developmental scale or IQ scores alone are not adequate to determine functioning level and outcome treatment eect. There were studies showing that the scale provided additional information on childrens social-cognitive functioning that could facilitate dierential diagnosis and assessment of severity in autism (Gillham, Carter, Volkmar, & Sparrow, 2000; Sparrow, Balla, & Cicchetti, 1984). (3) Childhood Autism Rating Scale (CARS). The CARS is a behavioral rating scale developed by the TEACCH Division (Schopler, Reichler, DeVellis, & Daly, 1980; Schopler, Reichler, & Renner, 1988). The CARS consists of ratings in 15 dierent areas of functioning signicant for autism: Relating to People; Imitation; Emotional Response; Body Use; Object

40 Use; Adaptation to Change; Visual Response; Listening Response; Taste, Smell, and Touch Response and Use; Fear or Nervousness; Verbal Communication; Nonverbal Communication; Activity Level; Level and Consistency of Intellectual Response and General Impressions. Childrens behaviors were observed and rated on a 7-point scale (4 anchor points and 3 half steps) with clear behavioral anchors. Schopler and Reichler (1979) reported that the Behavioral Scale of the PEP was related to the Childhood Autism Rating Scale (r =.80). Participants Sixty-three children (59 boys and 4 girls) aged between 3 years and 5 years 11 months old with the diagnosis of Autistic Disorder or Pervasive Developmental Disorder Not Otherwise Specied (e.g., Atypical Autism) based on the diagnostic criteria of the DSM-IV (American Psychiatric Association, 1994) were recruited from eleven special child care centres at Heep Hong Society with parents formal written consent. Amongst the participants, children were randomly assigned to assess the testretest reliability (N=30) and the concurrent validity (N=33) of the CPEP-R. Out of the 63 participants involved in the evaluation of the testretest reliability and concurrent validity of the test, 30 children were randomly selected for the study of inter-rater reliability. There were more boys than girls participating in the study. This situation was in part due to the fact that there was a higher incidence rate of boys than girls suffering from autistic spectrum disorders in the general population. In addition, because there were more boys than girls admitted to those centers participating in this study and the participants were randomly selected to participate in the study, the number of boys to girls was extremely high. For the concurrent validity study (N=33), the age distribution of the participants was as follows: 3 years to 3 years and 11 months (18 boys); 4 years to 4 years and 11 months (11 boys and 1 girl); 5 years to 5 years and 11 months (3 boys). The mean functioning level of the participants indexed by the CPEP-R Developmental Scale at the rst rating was 20.49 (SD=5.03, range = 1230). In terms of severity of autism, the mean score of the participants on the CARS was 32.77 (SD=6.88, range = 2147), which suggests that the average level of autism of the participants was in the mild to moderate range. Procedures

Shek, Tsang, Lam, Tang, and Cheung For the testretest reliability study (N=30), the age distribution of the participants was as follows: 3 years to 3 years and 11 months (7 boys); 4 years to 4 years and 11 months (10 boys and 2 girls); 5 years to 5 years and 11 months (10 boys and 1 girl). The mean functioning level of the participants indexed by the CPEP-R Developmental Scale at the rst rating was 25.27 (SD=12.69, range = 350). In terms of severity of autism, the mean score of the participants on the CARS was 67.07 (SD=9.81, range = 4792), which suggests that the average level of autism of the participants was in the severe range.

A total of six researchers, including two educational psychologists, one clinical psychologist, one occupational therapist, one speech therapist and one project assistant with special child care experience acted as administrators and raters in the study. Before the formal launching of the validation study, a panel involving these six researchers was formed to review and clarify the administration procedures and scoring criteria of the CPEP-R. Researchers went through the items one by one and counter-checked with the original English version when discrepancies among members arose. The administration procedures and scoring criteria were then rewritten and elaborated on the discrepant items. After the clarication process, the administrators administered the claried version of the CPEPR to one normal child and one autistic child. Five co-raters also rated the case. Meetings were then held to compare the results and clarify the discrepancies among the administrators and co-raters. In order to equip the researchers with the experience of administering the test to normal subjects and to test out the nal version for the study, the CPEP-R was administered to 12 normal pre-school kindergarten children. Subjects were recruited from two normal kindergartens, consisting of two girls and two boys from each age group of 34 years, 45 years, and 56 years old normal children. Each researcher had hands-on experience of testing the CPEP-R on two normal subjects as administrators and another two subjects as co-raters. The nal version for the study was conrmed after these steps. For the testretest reliability study (n = 30), test and retest of the CPEP-R were administered at an interval of 10 to 32 days. The variation of the interval was due to intermission by school holidays or

Psychometric Properties of the Chinese Version diculties in arrangement involved (e.g., childs sickleave and holding of functions). For the concurrent validity study (n =33), all six researchers administered the CPEP-R and the three psychologists administered the standardized intellectual assessment, including the MerrillPalmer Scale of Mental Tests and the Childhood Autism Rating Scale. For the Hong Kong Based Adaptive Behavior Scale, it was administered with either the parent or teacher as informant. The participants were randomly assigned to the researchers. To assess the inter-rater reliability of the CPEPR, 30 participants were randomly drawn from the 63 participants of the test-retest and concurrent validity studies. Each subject was administered the CPEP-R independently by one administrator and two co-raters as observers. The administratorco-raters grouping was randomly assigned by a research assistant who was blind to the characteristics of the subjects. In order to keep impartiality through the process, testing conditions were similar for all the children and the researchers involved in the case would rate the childs performance independently and handed the protocols to the research assistant directly for data entry.
Table I. Internal Consistency, Inter-rater Reliability, and Testretest Reliability of the CPEP-R Inter-Rater Reliabilityb

41

Domain Developmental scale Imitation Perception Fine motor Gross motor Eye-Hand coordination Cognitive performance Cognitive verbal Developmental score Behavioral scale Relating and aect Play and interest in materials Sensory responses Language
a

Alphaa

Testretest Reliabilityc

.90 .81 .88 .87 .92 .94 .96 .98 .85 .83 .74 .90

.98* .98* .98* .98* .99* .98* .99* .99* .87* .88* .85* .84*

.97* .89* .95* .93* .94* .98* .99* .99* .76* .84* .76* .92*

RESULTS Reliability Concerning the internal consistency of the different measures of the CPEP-R, the alpha values of the different domains of the CPEP-R are shown in Table I. Based on the ratings of the researchers conducting the concurrent validity study, analyses based on coecient alpha (Cronbach, 1951) showed that the internal consistency of the dierent measures was acceptable. Analyses were also carried out to examine the inter-rater reliability of different measures in the Developmental Domain and the Behavioral Domain. A review of the literature shows that different methods can be used to assess inter-rater reliability (Hartmann, 1977). However, to make the present ndings comparable to those reported previously, the procedures adopted by Schopler et al. (1990) were followed where intraclass correlation analyses (Guilford & Fruchter, 1978; Shrout & Fleiss, 1979) were performed. The ndings presented in Table I generally show that the CPEP-R was reliable across raters. For the Developmental

N=33. Assessment of the internal consistency of the CPEP-R measures was based on the ratings of the chief conductor of the concurrent validity study. b N=30. For the measures under the Developmental Scale, the analyses were based on the total number of Passed items in each domain. For the measures under the Behavioral Scale, the analyses were based on the total number of Appropriate items in each area. c N=30. For the measures under the Developmental Scale, the analyses were based on the total number of Passed items in each domain. For the measures under the Behavioral Scale, the analyses were based on the total number of Appropriate items in each area. * p < .001.

Domain, the related values of the intra-class correlation coecients ranged from .98 to .99. For the Behavioral Domain, the related values of the intra-class correlation coecients ranged from .84 to .87. Regarding the temporal stability of the CPEP-R, Pearson correlation analyses were performed (DeVellis, 1991). Results showed that the dierent domains in the CPEP-R were temporally stable. The values of Pearson correlation coecients ranged from .89 to .99 for dierent measures in the Developmental Scale. For the dierent areas assessed by the Behavioral Scale, the values of Pearson correlation coefcients ranged from .76 to .92. The ndings on the testrest reliability of the various measures of the CPEP-R are presented in Table I. The present ndings showed that the various measures of the CPEP-R were generally internally consistent, stable across raters and temporally stable. The present ndings thus establish the reliability of the Chinese version of the PEP-R.

42 Validity For the concurrent validity of the CPEP-R, the Developmental Score and the Developmental Age of the CPEP-R were found to have signicant correlation with the MerrillPalmer Scale of Mental Tests (r=.71 and .67, respectively, p < .001). Furthermore, the Developmental Score and the Developmental Age of the CPEP-R were signicantly correlated with the various domain scores of the Hong Kong Based Adaptive Behavior Scale (see Table II). Descriptive information on the participants scores on the CPEPR Developmental Scale and Behavioral Scale, MerrillPalmer Scale of Mental Tests and the Hong Kong Based Adaptive Behavior Scale is presented in Table III.

Shek, Tsang, Lam, Tang, and Cheung Behavioral Scale of the CPEP-R with CARS Results showed that the average total scores of the Behavioral Scale of the CPEP-R were signicantly related to the total scores of the CARS (r = .70, p < .001).

DISCUSSION One of the purposes of the present study was to examine the Chinese version of the PEP-R (CPEP-R). There are several points that should be highlighted as far as the reliability of the CPEP-R is concerned. First, the ndings are consistent with those reported in the Western (e.g., Schopler et al., 1990; Steerneman, Muris, Merckelbach, & Willems, 1997; van Berckelaer-Onnes & Van Duijn, 1993) and local Chinese (Tang, 2000) contexts. Second, because there are no published ndings on the testretest reliability of the scale, the present ndings are important additions to the literature. In particular, the present ndings clearly show that the CPEP-R is reliable with respect to multiple indicators of reliability. Finally, consistent with the available research ndings, the present study showed that the inter-rater and test retest reliabilities of the Behavioral Scale of CPEP-R were generally lower than those of the Developmental Scale. This observation makes sense because it is reasonable to expect that behavioral ratings are less reliable than developmental scores. Nevertheless, it must be emphasized that the reliabilities of the Behavioral Scale domains of the CPEP-R were still quite high and respectable. Another objective of this study was to examine the validity of the CPEP-R. The present ndings showed that the Developmental Score and Developmental Age derived from the CPEP-R were related to measures of intellectual and adaptive functioning; the Behavioral Scale of the CPEP-R was also related to autistic behavior assessed by the CARS. The present ndings are generally consistent with the Western ndings reported previously (Schopler et al., 1990). Cross-culturally speaking, there are two aspects of the ndings that deserve our attention. First, utilizing a translated and adapted version of the PEP-R, the study replicated the Western ndings that the PEP-R possesses good psychometric properties. The replicated ndings strongly support the cross-cultural applicability and the stability of the PEP-R. Although the present study cannot give any clue to the question of whether the translated etic assessment

Table II. Pearson Correlation Coecients on the Relationships Between the CPEP-R (Developmental Score and Developmental Age) and the Dierent Domain Scores of the Hong Kong Based Adaptive Behavior Scale Domains scores of the Hong Kong Based Adaptive Behavior Scale Communication domain Daily Living skills domain Socialization-interpersonal Domain Motor skills domain Total domain scores * p < .001. CPEP-R Developmental Score .72* .83* .66* .63* .86* CPEP-R Developmental Age .75* .83* .66* .59* .85*

Table III. Descriptive Information on the scores of the participants of the Concurrent Validity study on the CPEP-R Developmental Scale and Behavioral Scale, MerrillPalmer Scale of Mental Tests and the Hong Kong Based Adaptive Behavior Scale Mean CPEP-R: developmental score CPEP-R: behavioral score HKBABS: sum of domains scores HKBABS: daily living skills HKBABS: motor skills HKBABS: communication HKBABS: socialization Merrill-Palmer scores CARS total scores 48.48 68.849 234.424 48.849 51.212 23.909 27.939 29.879 32.773 SD 16.56 10.435 23.179 14.576 10.703 10.887 8.562 17.197 6.875 Range 1574 4689 186288 1774 3378 348 1049 160 2147

Note: CPEP-R: Chinese version of the Psycho-educational ProleRevised (CPEP-R). HKBABS: Hong Kong Based Adaptive Behavior Scale. MerrillPalmer: MerrillPalmer Scale of Mental Tests. CARS: Childhood Autism Rating Scale.

Psychometric Properties of the Chinese Version tool is superior to an emic assessment tool (i.e., indigenously developed measure), the present ndings suggest that based on the assessment framework of the PEP-R, the CPEP-R performed very well on the psychometric level. Of course, it would be important to ask the questions of whether the CPEP-R represents an adequate measure of autism in the Chinese context and whether there is any need to develop an indigenous Chinese measure of autism. However, discussion of these questions is beyond the scope of this paper. Second, because there are very few validated measures of autism in different Chinese contexts, the present study has practical contribution for Chinese helping professionals working with autistic children. With reference to the arguments that measures developed in the West may not be applicable to nonWestern contexts (Shek, 2001a; Shek & Lai, 2001) and that validation studies must be carried out when using Western measures in the Chinese context (Shek, 1998, 2001b), the present study yields robust validation data suggesting that the original English items of the PEP-R, with some adaptations, can be objectively used by Chinese researchers and helping professionals working with autistic children. As argued above, having a valid and reliable assessment is the rst and foremost step in devising any suitable and individualized treatment and educational plan for autistic children. Thus, the development of the CPEP-R represents a signicant advancement in the services provided to Chinese autistic children. With the selection of suitable and valid assessment items and materials that are tailored to the interests and characteristics of autistic children, the CPEP-R may help overcome the difculties of assessing autistic children due to their non-compliance under usual testing situations. The CPEP-R also provides a comprehensive framework for observing the childs processing modalities and decits that might hinder the childs learning (e.g., it taps information about the childs problems of motivation, organization and attention). There are several limitations of the present study. First, because the present ndings were based on Chinese children in Hong Kong, there is a need to replicate the ndings in other Chinese contexts. This is especially important because there is a growing demand for the CPEP-R and related materials in Mainland China and Taiwan. Second, the present ndings are based on a small sample and boys with a restricted range of intellectual level (mild to moderate grade mental retardation) were

43 over-represented. Further replication of the present ndings is in order. Third, although the MerillPalmer Scale of Mental Tests was used to establish the concurrent validity of the original PEP-R, it can be criticized that because its norms are out-of-date, the use of the scale as a criterion measure would not be appropriate. However, while this argument is valid in the Western contexts, one should note that a local study was conducted in Hong Kong in 1995 to address some of the loopholes of the original MerrillPalmer Scale of Mental Tests and to develop new norms based on local children (Department of Health, 1995). As such, the criticism that the use of the Chinese Merrill Palmer is inappropriate because its norms are outdated might not be a fatal one. Finally, the present investigation of the validity of the CPEP-R was conned to concurrent validity only. It would be illuminating if other aspects of the CPEP-R, particularly its factorial validity, would be investigated in the future using conrmatory factor analyses. Despite these limitations, the present study is a constructive response to the warning that cultural hurdles had to be overcome before the TEACCH Program could be successfully adapted and transplanted to Hong Kong (Health and Welfare Bureau, 1995, p.10).

ACKNOWLEDGMENT The authors have contributed equally to the development of this paper. They would like to thank the Providence Foundation Limited (major sponsor) and the Committee on Conference and Research Grants, The University of Hong Kong for their nancial support for the implementation of the project, and the Research Grants Council of the Hong Kong Special Administrative Region (Grant CUHK4293/03H) for its nancial support for the preparation of this paper. The authors would also like to thank Amy A.M. Chow and Natty S.Y. Chan for their assistance in collecting the data. REFERENCES
American Psychiatric Association. (1994). Diagnostic and statistical manual of mental disorder (4th ed.). Washington, DC: Author. Alpern, G. D. (1967). Measurement of untestable autistic children. Journal of Abnormal Psychology, 72, 478496. Cronbach, L. J. (1951). Coecient alpha and the internal structure of tests. Psychometrika, 16, 297334.

44
Department of Health. (1995). Report on the ndings of a local study on the Merrill-Palmer Scale of Mental Tests. Hong Kong: Child Assessment Service, Department of Health, Government of Hong Kong. DeVellis, R. (1991). Scale development. Newbury Park, CA: Sage. Guilford, J. P., & Fruchter, B. (1978). Fundamental statistics in psychology and education. New York: McGraw-Hill. Gillham, J. E., Carter, A. S., Volkmar, F. R., & Sparrow, S. S. (2000). Toward a developmental operational denition of autism. Journal of Autism and Developmental Disorders, 30, 269278. Hartmann, D. P. (1977). Considerations in the choice of interobserver reliability estimates. Journal of Applied Behavior Analysis, 10, 103116. Health and Welfare Bureau. (1994). Report by the Working Group on Services for Autistic Persons. Hong Kong: Health and Welfare Bureau, Government of Hong Kong. Health and Welfare Bureau. (1995). Report on Overseas Study Visit on Autism. Hong Kong: Health and Welfare Bureau, Government of Hong Kong. Kwok, J., Shek, D. T. L., Tse, J., & Chan, S. (1989). Hong Kong Based Adaptive Behavior Scale. Hong Kong: Department of Applied Social Studies, City Polytechnic of Hong Kong. Lam, M. K. T., & Rao, N. (1993). Developing a Chinese version of the Psychoeducational Prole (CPEP) to assess autistic children in Hong Kong. Journal of Autism and Developmental Disorders, 23, 273279. Muris, P., Steerneman, P., & Ratering, E. (1997). Brief report: Interrater reliability of the Psychoeducational Prole (PEP). Journal of Autism and Developmental Disorders, 27, 621626. Parks, S. L. (1988). Psychometric instruments available for the assessment of autistic children. In E. Schopler, & G. B. Mesibov (Eds.), Diagnosis and assessment in autism (pp. 123136). New York: Plenum. Rutter, M. (1978). Language disorder and infantile autism. In M. Rutter, & E. Schopler (Eds.), Autism: A reappraisal of concepts and treatment (pp. 85104). New York: Plenum. Schopler, E. (1997). Implementation of TEACCH philosophy. In D. J. Cohen, & F. R. Volkmar (Eds.), Handbook of autism and pervasive developmental disorders (pp. 767795). New York: Wiley. Schopler, E., & Reichler, R. J. (1979). Individualized assessment and treatment for autistic and developmentally disabled children (Vol. 1): Psychoeducational Prole. Austin, TX: Pro-Ed. Schopler, E., Reichler, R. J., Bashford, A., Lansing, M., & Marcus, L. (1990). Individualized assessment and treatment for autistic and developmentally disabled children (Vol. 1): Psychoeducational Prole-Revised (PEP-R). Austin, TX: Pro-Ed.

Shek, Tsang, Lam, Tang, and Cheung

Schopler, E., Reichler, R. J., DeVellis, R. F., & Daly, K. (1980). Towards objective classication of childhood autism: Childhood Autism Rating Scale (CARS). Journal of Autism and Developmental Disorders,, 11, 201217. Schopler, E., Reichler, R. J., & Renner, B. R. (1988). The Childhood Autism Rating Scale (CARS). Los Angeles, CA: Western Psychological Services. Shek, D. T. L. (1998). The Chinese version of the Self-Report Family Inventory: Does culture make a dierence?. Research on Social Work Practice, 8, 315329. Shek, D. T. L. (2001a). Psychometric properties of the Chinese version of the Self-Report Family Inventory: Findings based on a longitudinal study. Research on Social Work Practice, 11, 485502. Shek, D. T. L. (2001b). The General Functioning Scale of the Family Assessment Device: Does it work in the Chinese culture? Journal of Clinical Psychology, 57 15031516. Shek, D. T. L., & Lai, K. (2001). The Chinese version of the SelfReport Family Inventory: Reliability and validity. American Journal of Family Therapy, 29, 207220. Shrout, P. E., & Fleiss, J. L. (1979). Intra-class correlations: Uses in assessing rater reliability. Psychological Bulletin, 86, 420428. Sparrow, S. S., Balla, D., & Cicchetti, D. V. (1984). Vineland Adaptive Behavior Scales. Circle Pines, MN: American Guidance Services. Steerneman, P., Muris, P., Merckelbach, H., & Willems, H. (1997). Brief report: Assessment of developmental and abnormal behavior in children with pervasive developmental disorders: Evidence for the reliability and validity of the Revised Psychoeducational Prole. Journal of Autism and Development Disorders, 27, 177185. Stutsman, R. (1948). Merrill-Palmer Scale of Mental Tests. Los Angeles, CA: Western Psychological Services. Tang L. Y. (2000). Reliability and validity of Psychoeducational Prole-Revised (Chinese Version). Unpublished Master of Science (Health Care) dissertation. Hong Kong: Department of Rehabilitation Science, The Hong Kong Polytechnic University. Van Berckelaer-Onnes, I., & Van Duijn, G. (1993). A comparison between the handicaps behavior and skills schedule and the Psychoeducational Prole. Journal of Autism and Developmental Disorders, 23, 263272. Volkmar, F., Hoder, L., & Cohen, D. (1985). Compliance, negativism and the eects of treatment and structure in autism: A naturalistic behavior study. Journal of Child Psychology and Psychiatry, 26, 865877.