The testing and measurement of Autism Spectrum Disorders (ASDs) is a topic of interest in research. As a result, critical issues in testing and measurement of the ASDs have been the center of debate and research. Therefore, this paper will outline the five critical issues in testing and measurement of ASDs.
We will write a custom Term Paper on Autism Spectrum Disorders: Testing and Measurement specifically for you
301 certified writers online
It will begin with the rationale that explain why the issues are critical, followed by the reasons for unavailability of the solutions to the issues and the consequences of the issues. For example, diagnostic issues are vital because Autism Spectrum Disorders are associated with intellectual and language skills that affect their presentation (Garcin & Burack, 2008).
Secondly, source of information is a critical issue because accurate information lead to an accurate diagnosis (Goldstein et al., 2008). Thirdly, the impact of research on diagnostic instrument is crucial because classifications of ASDs are research based rather than empirical studies (Jessa & Carr, 2009). Fourthly, methodological issues are critical because validity and reliability affect the tests for ASDs (Simpson & Myles, 2008). Finally, practical issues are imperative because they affect the choice of the measurement scale (Luyster et al., 2008).
The paper continues by analyzing and reviewing the critical issues in line with the current literature. To begin with, diagnostic issues include issues concerning measurement of intellectual capacity, language delay, the association between autism and language impairment and the early and late diagnosis of ASDs. Secondly, issues with sources of information entail history taking, observation and the use of diagnostic instruments. Thirdly, the issues of the impact of research on diagnostic instrument analyze the relationship between research and clinical diagnosis together with the development of new diagnostic instruments. Fourthly, methodological issues analyze the reliability and validity of the test measurement. Lastly, practical issues outline the limitation of the measurement scales.
Finally, the paper end with a conclusion that analyze and review the discussed issues and recommend the way forward. For instance, the clinician uses of multiple sources of information increase the accuracy of diagnosis. On the other hand, the paper recommends the development of a policy that will enhance accessibility of diagnostic and intervention for children with ASDs.
Rationale of the critical issues
Diagnostic issues are critical because children with Autism Spectrum Disorders have an association with a range of intellectual and language skills that affect their presentation (Garcin & Burack, 2008). Moreover, the disorder begins early in life and progresses to adulthood hence, it is hard to identify a particular behavior that can lead to its diagnosis (Goldstein et al., 2008). For instance, a delay in milestone development, which is a presentation of autism, does not distinguish it from other developmental disorders (Simpson & Myles, 2008). Moreover, the age at which health care professions diagnose children with ASDs portray some controversy (Simpson & Myles, 2008).
The selection of sources of information is important in making a diagnosis of ASDs because accurate information lead to a proper diagnosis while the vice versa is true (American Medical association, 2011). In this case, the challenge is that different sources may provide different information hence discrepancies in the diagnosis (American Medical association, 2011). Therefore, different sources require critical analysis before a conclusion on which source to chose (Akshoomoff, 2009). At some point, the clinician may require to use a variety of sources because of the belief that accuracy in diagnosis increases with the use of multiple sources (Jessa & Carr, 2009). For instance, the clinician can combine history taking, observation and the use of diagnostic instrument (Simpson & Myles, 2008).
The classification of ASDs uses the researches done rather than experimental scientific studies (Jessa & Carr, 2009). Although neurobiological factors can lead to a diagnosis of autism, heterogeneity in biology may occur in other disorders like Aspergers disorder hence the clinician will depend on the description of social and other behaviors (Akshoomoff, 2009).
Therefore, this becomes a critical issue because a clear demarcation between autistic spectrum disorders and other developmental disorders does not exist especially in terms of changes in the developmental milestone (Volkmar et al., 2005). Finally, in the near future clinicians expect that information about development and symptoms that emerge from studies will influence the classification of ASDs because diagnostic instruments should change as more information is acquired (Jessa & Carr, 2009).
Methodological issues are very important because validity and reliability of the test affect the diagnosis of children with ASDs (Akshoomoff, 2009). Besides, many of the diagnostic instruments have not effectively addressed the issue of validity and reliability (Akshoomoff, 2009). On one hand, the lack of information is understandable because of the difficulties in getting a large sample while on the other hand; it is a reflection of the limitation of the development of the instrument (Volkmar et al., 2005).
Practical considerations in ASDs are critical issues because scale developers have outlined various advantages and disadvantages of the test measurement (Jessa & Carr, 2009). Furthermore, some distinctions are excellent but they do not have accessible statistics for support (Rojahn & Matson, 2010). For example, a test involving a yes or no answer can produce specific information yet one cannot conclude that it can accurately diagnose ASDs (American Medical association, 2011).
Additionally, researchers have not done studies that address this hypothesis (Rojahn & Matson, 2010). Another issue that emerges in practical consideration is the distinction made between instruments that are used for research purposes and for clinical purposes (Akshoomoff, 2009). There is no verification although some scale developers propose that research instrument can diagnose children with ASDs in spite of the fact that their development is for research purposes (Jessa & Carr, 2009).
Measurement of intellectual capacity in children is imperative in diagnosing Autism Spectrum Disorders (Volkmar et al., 2005). Although varieties of instruments can test general intelligence such instruments are vanity in testing ASDs because of the occurrence and the unpredictability of the disorders (Rojahn & Matson, 2010). Moreover, cognitive tests use the demography of population to determine average but in the case of autism determination of average is difficult because children with autism are usually fewer. If many of them are included in a study before follow up the result will be erroneous (Rojahn & Matson, 2010).
On the other hand, studies that compare distribution of different samples can provide vital information concerning the consistency of the diagnosis (Akshoomoff, 2009). Moreover, in a recent research, Andrien & Barthelemy (2010) explains that there is a new Social Cognitive evaluation battery that assesses both mental and the social growth in children with autism spectrum disorders. In addition, this instrument is very constructive and appropriate in assessing the mental development in children with ASDs and it has been tested and proved to be valid and reliable (Andrien & Barthelemy, 2010).
Get your first paper with 15% OFF
Language delay needs consideration in diagnosing autism because children with ASDs have severe language delay than other children of the same developmental level (Akshoomoff, 2009). Therefore, diagnostic criteria that depend on expressive and receptive language must consider this language delay (Rojahn & Matson, 2010). Nevertheless, it is complicated to develop such a diagnostic instrument because controlling language may control the autism leading to an invalid comparison (Akshoomoff, 2009).
For instance, it is invalid to compare a 2-year-old child who has autism with an 8-month old child who is normal but with the same receptive language skills (American Medical association, 2011). On the other hand, measurement of language in children with autism spectrum disorders is through two approaches that are standardized testing and parent reporting. These two approaches offer valuable information about children’s linguistic skills (Luyster et al., 2008). Standardized testing involves ADOS-G, the Mullen Scale of Early Learning, the Imitation Battery, and Early Socio Economical Scales. Parent reporting involves the use of ADI-R, the Vine Land Adaptive Behavior Scale, and a Completion of Parent Questionnaire Booklet (Luyster, et al., 2008).
The association between autism and impairment of language is complex because a person with little spontaneous speech may not portray many language abnormalities like a person with autism who has fluent speech (Andrien & Barthelemy, 2010). As a result, language abnormality score is by addition of a number of ways in which ones language is unusual.
For example, clinicians can assess echolalia and neologism concurrently and a person with a complex score is more abnormal than a person who cannot speak (Andrien & Barthelemy, 2010). Furthermore, researchers and scientists have developed various tools that measure language impairment. These tools include Peaboy test that assess the vocabulary level in young children, language assessment through clinical trials and Preschool Language Tests (Ozonoff et al., 2008).
Although these tools measure both receptive and communicative speech, a medical appointment with a pathologist who has specialized in speech is important because the pathologist is in a position of knowing the child’s problem and can provide the appropriate interventions (Rojahn & Matson, 2010). On the other hand, assessment of pragmatic information includes analysis of non-verbal behaviors like facial expressions, body languages, and gestures (Dyches, 2011). Therefore, health care professionals can use the following tests, which assess and evaluate practical language in children with autistim spectrum disorders. The test of language proficiency and the scrutiny of communicative ability test (Ozonoff, et al., 2008).
Additionally, some diagnostic instruments for children with ASDs are more accurate for school-going children and the accuracy decreases as one approaches adulthood (Dyches, 2011). Unfortunately, these groups of children together with others like children with Pervasive Developmental Disorders and Asperger’s Disorders require diagnostic instruments (Cosden et al., 2010). Therefore, it is significant that people who use diagnostic instruments familiarize themselves with the biases that instruments can possess (Andrien & Barthelemy, 2010).
For instance, Baley Scale is excellent in assessing cognitive skills in young children and it provides an estimate of intellectual quotient in children according to their specific age (Dyches, 2011). However, Baley Scale cannot provide a separate standardized score for both verbal and non-verbal skills (Rojahn & Matson, 2010). As a result, clinicians prefer to use Mullen Scale of Early Learning because infants and children up to 68 months can benefit from Baley Scale (Akshoomoff, 2009).
Recent research has shown that it is difficult to diagnose ASDs before the age of approximately 6 years and that no method can diagnose autism at the age of 6 months to 1 year (Volkmar et al., 2005). On the contrary, other investigators have developed methods that can diagnose autism earlier in life and they believe that the earlier the diagnosis the greater the long-term benefits (Rojahn & Matson, 2010). For instance, standardized measure of ASDs can diagnose autism in children less than 18 months but CHAT and STAT scaling instrument can diagnose autism in children who are 1.5 to 2 years old (Prelock & Brooke, 2007).
Moreover, though early interventions are promising standardized measures of ASDs have methodological shortcomings like lack of control that matches the measures and the outcomes of the measures of the specific symptoms of ASDs (American Speech-Language-Hearing association, 2011). Therefore, data can be promising at present but they are far from conclusion (Rojahn & Matson, 2010). According to DSM IV, diagnosis of autism is possible at the age of 30 months (Akshoomoff, 2009). Nonetheless, the controversy is whether an accurate diagnosis is possible at this age especially if the presenting symptoms of autism are hidden (Andrien & Barthelemy, 2010).
Issues with sources of information
Although the parent and children with Autism Spectrum Disorders report usually concur, it is only possible for older children who can describe their own symptoms and feelings (American Speech-Language-Hearing association, 2011). Moreover, the parent report can be more reliable and valid in issues of developmental milestone (Prelock & Brooke, 2007).
For instance, a parent can accurately report about the child’s abnormal development in motor or speech at an early stage of life (American Speech-Language-Hearing association, 2011). On the other hand, an accurate indicator can be direct observation like observing the behavior of very young children with autism or self-reporting, which takes place when a child with autism reports about his or her feelings (Dyches, 2011). Therefore, it is important to use multiple sources of information (Dyches, 2011).
The use of multiple sources of information helps in increasing the accuracy of the diagnosis because it places the diagnosis in both social and development context (Andrien & Barthelemy, 2010). For example, information about limited social interaction in early childhood can be responsible for social isolation in adulthood (American Speech-Language-Hearing association, 2011). Moreover, observation of a child’s response when being called by his or her parents can be helpful in determining the child’s attempt to get attention (Akshoomoff, 2009).
Additionally, the use of a diagnostic instrument makes diagnosis of autism easier because it maximizes the use of observation and description while obtaining first hand information directly from the person with ASDs (Gotham et al., 2007). On the other hand, researchers and clinicians need to know that diagnostic instruments cannot replace history or observation (Prelock & Brooke, 2007). Moreover, clinicians should structure observations of social communication behavior so that understanding and treatment of ASDs becomes easier (Gotham et al. , 2007). On the other hand, combining information from multiple sources is intricate (Prelock & Brooke, 2007). For example, determination of severity of autism might be to consider the repetition of different information in different sources (Rojahn & Matson, 2010).
Instruments for diagnosing ASDs usually vary in the degree in which they measure the presence of abnormalities or the absence of normality (Rojahn & Matson, 2010). In a research done, an argument on convergent validity occurred because there was a correlation between SCEB and Brunet-Lezine developmental level scores because a correlation between the two existed although the two instruments measure developmental domains (Andrien & Barthelemy, 2010).
The expectation was that a moderate correlation would exist because the two instruments do not measure the same component of the domain and their adaptation to children with ASDs is not equal (Andrien & Barthelemy, 2010). In young children, absence of behaviors like social smiles may be a prediction of an outcome than abnormalities and besides, it could have a correlation with chronology and cognitive age (Dyches, 2011). Finally, diagnostic instruments may measure the same features but use different perspectives (Melissa & Matson, 2007).
The impact of research on diagnostic instrument
A difference in priority of the results of the diagnosis of Autism Spectrum Disorders exists in both clinical and research purposes (Dyches, 2011). On the one hand, clinical diagnosis helps children with autism and their significant others access information about ASDs (Akshoomoff, 2009). Furthermore, it allows the service provider to attend to them because clinicians use clinical diagnosis to allocate limited resources and to ensure that children with ASDs receive the required services (American Speech-Language-Hearing association, 2011).
On the contrary, researchers prefer narrower diagnosis because it decreases the probability of false positive and it provides reliability because it reduces the overlap of the control groups (American Speech-Language-Hearing association, 2011). Additionally, researchers usually include a small population in their study and their objective is to make good use of the children who are involved in the study (American Medical association, 2011). As a result, all the above factors affect the objective of the diagnostic instrument and the way in which the instrument is used (Dyches, 2011).
A need for diagnostic instrument to test other developmental disorders apart from ASDs exists because there is a minimal difference between ASDs and other disorders such as Aspegers Disorder and PDD-NOS (Rojahn & Matson, 2010). Furthermore, a clear diagnostic criterion for these disorders does not exist because there is no reliable and valid instrument and the lack of empirical data affects the ability to distinguish these disorders from autism (Dyches, 2011). Although numerous diagnosis criteria for Aspergers Disorders exist, these criteria do not address the overlap with autism (Rojahn & Matson, 2010).
On the other hand, DSM-IV and ICD-10 diagnostic criteria express the relationship that exists between the two disorders while providing little conceptualization (Melissa & Matson, 2007). Additionally, there is conceptualization in disorders like non-verbal learning disabilities but an explanation of their relationship with autism is absent (American Speech-Language-Hearing association, 2011). The development of an assessment tool that can test Aspergers Disorder is complex because the definition of the disorder is not clear and there is still a debate as to whether the disorder is distinct from ASDs or it is its subtype (American Medical association, 2011).
Despite the fact that diagnosing ASDs presents difficulties, the process has gained a lot from research because different researchers are constantly studying the ASDs and their management (Melissa & Matson, 2007). For instance, Williams & Atkins (2009) carried a study on autistic spectrum disorders in the community and they focused on the discrepancies in classification. Moreover, both experimental and descriptive researches have tried to find a solution to these difficulties by developing a theory of mind that discriminates ASDs from other disorders (Akshoomoff, 2009).
Finally, the perception of ASDs has shifted from theory to practice thus affecting the content of the diagnostic instruments (Canadian Department of Developmental Services, 2008). For example, scientists have developed several diagnostic instruments for testing autism and related disorders. These instruments include Pervasive Developmental Disorders Screening Test (PDDST), the checklist for testing autistic spectrum disorders in children below one year and a tool for screening autism spectrum disorders in children above two years (Rojahn & Matson, 2010).
Reliability is the degree of which a score is free from error and a number of factors influence it such as cross raters, time, and the nature of the instrument (American Speech-Language-Hearing association, 2011). Additionally, correlation measures whether the ranking of an individual is the same across rater’s and it estimates the reliability (Rojahn & Matson, 2010). For example, Melissa & Matson (2007) in their study reviewed extant literature and summarized Autism Spectrum Disorders symptoms over a period to ensure that the results were reliable.
The challenge of correlation is that the results can differ resulting in to different diagnosis even though the correlation is high (Cosden et al., 2010). Additionally, if a certain threshold determines the diagnosis the score can result in different diagnosis for the same person (American Speech-Language-Hearing association, 2011). Therefore, correlations are not accurate in making a diagnosis specifically when threshold is involved (Rojahn & Matson, 2010). Consequently, clinicians use percentages of agreement between larger groups of raters instead of correlation (Rojahn & Matson, 2010). For instance, Jessa & Carr (2009) used percentages in their study about functional assessment of behavioral problem in children with ASDs.
During the development of a diagnostic instrument, the scientist should consider item-level inter-rater agreement because it facilitates an experiment that determines its validity (American Speech-Language-Hearing association, 2011). According to Williams & Atkins (2009 ), most of the diagnostic tools for testing ASDs do not have this level because they rely on domain scores whose internal consistency are well documented. For example, Wechsler Test and the Vinland Adaptive Behavior Scale do not possess this level (Gotham et al., 2007). On the other hand, some testing tools that are reliable exist and the example is the Diagnostic Assessment Scale for the Severely Handicapped (Melissa & Matson, 2007).
In reliability, sample-size determination is imperative (Dyches, 2011). For instance, ASDs affect children who possess a variety of differences in language and cognition and if the researcher takes the samples carelessly, the overlap in scores will decrease (Williams & Atkins, 2009). Moreover, if the development of the instrument was from a small population it may not produce reliable results when used on a large population (American Speech-Language-Hearing association, 2011).
Therefore, it is vital for test users to consider the context of instrument development before making a sound decision on the use of the instrument (American Speech-Language-Hearing association, 2011). In addition, the entire testing instrument should have information about their stability and the changes expected across multiple administrations (Canadian Department of Developmental Services, 2008). Nevertheless, the test user should remember that although other test users have used the instrument in many studies its reliability on an individual basis needs considerations (Canadian Department of Developmental Services, 2008).
The parents’ awareness about their children with ASDs requires consideration because these parents are familiar with how their children fit into the diagnostic scheme (Ozonoff et al., 2008). For example, assigning value to families of children with ASDs leads to identification of cultural and linguistic diversity affecting these children, which leads to proper intervention (Canadian Department of Developmental Services, 2008). Nonetheless, if the parent’s information is for preliminary diagnosis then teachers or caregivers who lack familiarity with the disorder should determine its reliability (Williams & Atkins, 2009).
On the other hand, a diagnostic instrument should be valid and this means that other evidence should support its results (Goldstein et al., 2008). For instance, if the results of an autistic child’s behavior measured using the Behavioral Assessment System for Children (BASC) concurs with the results when using strength and difficulties questionnaire then the BASC is valid (Cosden et al., 2010).
Additionally, content validity refers to the degree in which the different test instrument represents the criteria used in diagnosis of ASDs (Akshoomoff, 2009). For instance, many instruments analyzed predated the liberation of DSM-IV and ICD-10 criteria for ASDs and as a result, they do not correspond with the three-domain approach illustrated in the diagnostic system (Melissa & Matson, 2007). Moreover, interpretation of the findings from ADI and ADOS/PL-ADOS influenced the strategies tested in the trial field together with the ICD-10 amendments (Dyches, 2011). On the contrary, the convergent validity of ADOS, CARS and ADI has been excellent (American Speech-Language-Hearing association, 2011).
Although some measures emphasize the fact that they are very detailed, the detail does not matter what matters is the ability of the test to obtain relevant information that is useful in the diagnosis of Autism Spectrum Disorders (Jessa & Carr, 2009). For instance, DSM-IV is a very detailed test yet general and as a result, it cannot provide a clear distinction between ASDs and other developmental disorders. Thus, clinicians cannot fully rely on it and they have to use other tests to confirm the diagnosis (Luyster et al., 2008).
Practical issues like the complexity of a measuring scale can limit the use of a particular scale (Akshoomoff et al., 2009). For example, ADOS and ADI-R are very good scales but they require one to attend a workshop in order to know their use or to procure a video tape that illustrates their use (Akshoomoff et al., 2009). Therefore, it becomes expensive and time consuming to use these scales resulting in clinicians opting for cheaper and easier measurements scales (Dyches, 2011). For instance, it is easier to use scales like DSM-IV and ICD-10 (Akshoomoff et al., 2009).
In conclusion, the critical issues discussed for testing and measuring Autism Spectrum Disorders require great attention because they affect the diagnosis of the disorders and possible interventions. For example, diagnostics issues affect the accuracy of the diagnosis, which determine the accuracy of the intervention.
Moreover, the sources of information determine how reliable the information is in the diagnosis of the disorder. Furthermore, the scientists can use the impact of research information in the diagnostic instruments to develop valid and reliable test measurement criteria. In addition, methodological issues are imperative in determining and choosing a valid and reliable measurement scale. Finally, these practical issues require attention because they facilitate the choice of a measurement scale that influences the accuracy of the result of the diagnosis of ASDs.
Due to the criticality of the issues in the testing and measurement of Autism Spectrum Disorders, the following recommendations are proposed. First, scientists should develop a valid and reliable diagnostic instrument while the researchers develop a valid and reliable measurement scale. Secondly, during the assessment of a child with autism, the clinician should collect a detailed history from both the parent and the child, use observed information from the caregiver and finally use diagnostic instruments. This is because multiple sources increase the accuracy of the diagnosis.
For this reason, the clinician should use both research diagnosis and clinical diagnosis to come up with an intervention. Furthermore, I recommend that the clinicians read and research about the reliability and validity of the available instruments so that they know which one to use in a particular situation. Finally, policy makers should develop policies that address the diagnosis and intervention of ASDs so that children with autism and their families know their rights. For example, policy makers can develop a policy that facilitate the accessibility of diagnosis and intervention measures for all children with autism because this will ensure that children with ASDs do not have any limitations in diagnosis and treatment.
Akshoomoff, N. (2009). Use of the mullen scale of early learning for the assessment of young children with autism spectrum disorders. Child Neoropsychology , 3 (12), 269-277.
Akshoomoff, N., Corsello, C., & Schimidt, H. (2009). The role of the autism diagnostic observation schedule in the assessment of autism spectrum disorders in school and community settings. Carlifornia Association of School Psychologists , 11 (1), 7-19.
American Medical association. (2011). Autism spectrum disorders: Assessment. Medical and Behavioural Heath Policy Manual (1), 43.
American Speech-Language-Hearing association. (2011). Assessing diverse students with autism spectrum disorder: Valuing families as an important member of multidisciplinary team will facilitate the accurate identification of culturallly or linguistically diverse-and, all-students with ASDs. ASHA Leader , 3 (12), 23-45.
Andrien, L., & Barthelemy, C. (2010). The social cognitive evaluation battery for children with autism: A new tool for the assessment of cognitive and social development in children with autism spectrum disorders. Autism Research and Treatment , 23 (4), 32-43.
Canadian Department of Developmental Services. (2008). Autism spectrum disorders: Best practice guideline for screening, diagnosis and assessment. Journal of Children and Family Services , 45 (916), 654-1000.
Cosden, M., Kern, L., Koegel, L., Greenwell, A., & Klein, E. (2010). Strength based assesment for children with autism spectrum disorders. Journal of Autism Development Disorder , 31 (2), 134-143.
Dyches, T. (2011). Assessing diverse students with autism spectrum disorders. ASHA Leader , 15 (1), 4-10.
Garcin, J., & Burack, K. (2008). Screning, assessment and diagnosis of autism spectrum disorders on young children: Canadian best practice guidelines. Mirian Foundation , 25 (514), 345-1300.
Goldstein, S., Naglieri, J., & Ozonnoff, S. (2008). Assessment of autism spectrum disorders. Journal of the Canadian academy of Child and Adolescent Psychiatry , 20 (1), 68-69.
Gotham, K., Risi, S., & Pickless, A. (2007). The autism diagnosis observation schedule:Revised algorithms for improved diagnostic validity. Journal of Autism and Development Disorder (37), 613-627.
Jessa, R., & Carr, L. (2009). Functional assesment of problem behaviour in children with autism spectrum disorders: A summary of 32 outpatient cases. Journal of Autism Developmental Disorder , 2009 (39), 362-372.
Luyster, L., Kadlec, M., & Carter, A. (2008). Language assesment and development in toddlers with autism spectrum disorders. Journal of Autism Development Disorder , 27 (38), 1426-1438.
Melissa, W., & Matson, L. (2007). Reliability and factor structure of autism spectrum disorders-diagnostic scale for intelectually disabled adults. Journal of Development and Physical Disability , 5 (19), 565-577.
Ozonoff, S., Goodline-jones, B., & Marjorie, S. (2008). Evidense-based assessment of autism spectrum disorders in children and adolescents. Journal of Clinical Child and Adolescent Psychology , 34 (3), 523-540.
Prelock, J., & Brooke, B. (2007). Interdisciplinary assessment of young children with autism spectrum disorders. American Speech Language Hearing Association , 34 (1), 194-202.
Rojahn, J., & Matson, J. (2010). Assesment and diagnosis of autism and spectrom disorders in children. Journal of Development and Physical Disabilities , 100 (22), 313-315.
Simpson, B., & Myles, R. (2008). Educating children and youth with autism: Strategy for efective practices. Austin, TX: Pro-Ed.
Volkmar, F., Rhea, P., Cohen, D., & Klin, A. (2005). Handbook of autism and perversive developmental disorders: Assessment, intervention, and policy. Hoboken,NJ: John Wiley and Sons.
Williams, M., & Atkins, M. (2009). Assessment of autism in community settings:Discripancies in classification. Journal of Autism and Development Disorders , 200 (39), 660-669.