Part II of Tests

Autism Behavior Checklist


The Autism Behavior Checklist (ABC) is a general measure of autism. It is not as reliable as the CARS or ADI-R. Correlations between the ABC and CARS ranged from 0.16 to 0.73 in a study by Eaves and Milner (1993). The CARS correctly identified 98 percent of the autistic subjects; it identified 69 percent of the possibly autistic as autistic. The ABC correctly identified 88 percent of the autistic subjects, while it identified 48 percent of the possibly autistic as autistic. In two separate studies, teachers' ratings on the ABC failed to reveal a common set of characteristics of students with high functioning Autistic Disorder (Myles, Simpson, & Johnson, 1995) and Asperger's Disorder (Ghaziuddin, N., Metler, Ghaziuddin, M., Tsai, & Luke, 1993).


Checklist for Autism in Toddlers


The Checklist for Autism in Toddlers (CHAT) is a screening instrument designed to detect core autistic features to enable treatment as early as eighteen months. The most effective treatment currently available for autism is early educational intervention, beginning as soon as possible after a child's diagnosis. Unfortunately, intervention rarely begins before the age of three years because few autistic children are diagnosed before they reach preschool age. CHAT offers physicians a means of diagnosing autism in infancy so that educational programs can be started months or even years before most symptoms become obvious. According to the authors, "We stress that the CHAT should not be used as a diagnostic instrument, but it can alert the primary health professional to the need for an expert... referral."


This first study (Baron-Cohen, Allen, & Gillberg, 1992) using the CHAT re- vealed that key psychological predictors of autism at thirty months are showing two or more of the following at eighteen months: (a) lack of pretend play, (b) lack of protodeclarative pointing, (c) lack of social interest, (d) lack of social play, and (e) lack of joint-attention. The CHAT detected all four cases of autism in a total sample of 91 eighteen-month-old children. The authors recommend that if a child lacks any combination of these key types of behavior on examination at eighteen months, it makes good clinical sense to refer him or her for a diagnostic assessment by a specialist with experhse m auQsm.


A second study (Baron-Cohen, Cox, Baird, Swettenham, Nightingale, Morgan, Drew, & Charman, 1996) concluded that "consistent failure of three key items from the CHAT at eighteen months of age carries an 83.3 percent risk of autism, and this pattern of risk indicator is specific to autism when compared to other forms of developmental delay." In the second study, research data on 16,000 children suggested that children who failed three items on the CHAT are at high risk of being autistic. The items include protodeclarative pointing (pointing at an object to direct another person's attention to it – not to obtain the item, but simply to share an interest in it); gaze monitoring (turning to look in the same direction as an adult is looking); and pretend play. The false positive rate for detection of autism using the CHAT is estimated at 16.6 percent.


Real Life Rating Scale


The Real Life Rating Scale (RLRS) (Freeman, Ritvo, Yokota, & Ritvo, 1986) is a scale used to assess the effects of treatment on 47 behaviors in the motor, social, affective, language, and sensory domains among autistic persons. The RLRS is applicable in natural settings by nonprofessional raters, is rapidly scored by hand, and can be repeated frequently without affecting inter-observer agreement. Data are presented on inter-rater agreement among novice and experienced observers. Instructions for the scale, target behaviors, and definitions are appended to the journal article.