Testing to Fail: Exacerbating disability through inappropriate assessment Short abstract Many school-age children with little or no speech are assessed, or labelled, as having IQs less than 50 or as being untestable. Typically in such assessments a test (such as the Weschler Intelligence Scale for Children) requiring both fluent reliable speech and normal fine motor skills is administered to a child who has neither. Typically the child without speech provides few if any answers and the assessing psychologist thus estimates their cognitive function as moderately or severely impaired. This guess is then used to determine the appropriate educational placement for that child. Such assessments commonly affect the provision of both academic and communication programs, not to mention parental expectations. Children placed in programs for students with IQ scores less than 50 are usually restricted to AAC strategies that could never allow them to challenge the psychologist s estimate. We have re-assessed some of these children using standardised tests such as the Peabody Picture Vocabulary Test (targeting receptive vocabulary and general knowledge) or Raven s Progressive Matrices (targeting problem-solving skills). Both can be administered to anyone with functional sight and hearing who can point accurately. Most of those re-assessed scored in the average range for their age on a Peabody or Raven s test and all scored significantly above the previous assessments based on tests requiring speech. This presentation will provide more information about the tests we used and include video of students undertaking these tests, together with data from researchers trialling other cognitive assessments with children with minimal speech. Long abstract Most children with little or no speech are assumed to have severe learning difficulties, and are treated accordingly. The likelihood of misassessment has been recognized from the early days of AAC, but practices have proved very hard to change. The rigid procedures of our Department of Education and Training (DET) demonstrate this. DET directs psychologists to use intelligence tests requiring fluent reliable speech to assess the cognitive potential of children who have minimal speech, specifically the Weschler Intelligence Scale for Children (WISC), for children aged 6 to16, and the Weschler Adult Intelligence Scale (WAIS) for students over 16. This is as sensible as using a tartan-matching test to assess a blind child. Children with minimal speech, who typically have diagnoses of severe autism, cerebral palsy or Down syndrome, often score zero because they cannot say the answers, and do not have the fine motor skills to do the performance tasks. They may also find the test situation stressful and be uncooperative,
DET advises the psychologists who administer the IQ tests that if students are untestable, for any reason, their intelligence should simply be estimated. Most children with minimal speech are consequently labeled as having IQs below 50 based on a psychologist s guess. Our experience with hundreds of school students with minimal speech indicates that most are significantly underestimated. Children who are untestable due to speech issues should be automatically referred for AAC intervention, and given more appropriate tests when they can communicate, however, this does not happen. Instead the psychologists initial guesses determine, often for their entire school careers, these children s school placements and the education they are offered. Negative expectations, generated by the low scores they are given, affects AAC provision as well. Most of these children have either very basic communication aids or none at all. In 2014 we saw four students with autism and no functional speech from a large school for children with IQs less than 50. Two of these students are 12 and two are 18. School admission was based on estimated WISC scores as all four were deemed untestable. According to their parents, the school made no effort to provide these students with any communication aids, despite government-funded aids costing up to $7,000 being readily available. One speech pathologist provides services for 100 children with minimal speech. When a parent requested AAC for her child, the school psychologist asked What s an AAC? The Peabody Picture Vocabulary Test (PPVT-M) is a standardized general knowledge and language test that does not require speech or fine motor skills. All four students tested in the average range on a PPVT-M administered in the standard manner after working on the concentration and pointing skills required for multiple-choice tests. (The pre-test training familiarized them with both location and tester). All four also demonstrated useful literacy skills, picked up from ipads, You Tube, and being read to at home. All now use communication aids to generate complex sentences. The damage to these students and their families from their previous mis-assessments is incalculable. Consequent frustration and stress have contributed to clinical depression and severe behaviour problems in the students. We are continuing to assess more students using the PPVT or Raven s Progressive Matrices (RPM). The results of students tested are compared with their DET assessments. A detailed summary will be available at the ISAAC conference. The issue of inappropriate assessment has been raised regularly for children with autism (Wilkinson & Rosenquist, 2006, Dawson et al, 2007, Kasari et al, 2013). A recent study -- Autistic children at risk of being underestimated: school-based pilot study of a strength-
informed assessment -- by Courchesne et al (2015) compared various assessments. Thirty autistic children (6 to 12 years) with little or no spoken language, attending specialized schools for autistic children with the highest levels of impairment, were assessed using Wechsler Intelligence Scale for Children (WISC-IV), Raven s Colored Progressive Matrices board form (RCPM), Children s Embedded Figures Test (CEFT), and a visual search task. The results were significant: None of the autistic children could complete WISC-IV; only six completed any subtest. In contrast, 26 autistic children could complete RCPM, with 17 scoring between the 5th and 90th percentile. [That is, in the normal range or better] In view of this evidence, continued reliance on the WISC for student assessments is indefensible. Rather than hunting for more appropriate one-off assessments, a different approach is possible. The US Individuals with Disabilities Education Improvement Act (2004) promotes Response To Intervention, or RTI, for children presenting with learning difficulties, rather than the use of standardized IQ tests. In the RTI process, service delivery is typically divided into three levels of support. Approximately 80% to 85% of the general student body meet grade level norms without additional assistance. Students who do not perform at the expected level are provided with supplementary small group instruction interventions at Tier 2. Approximately 3% to 6% of students will continue to have difficulties after Tier 2 interventions and these students will then receive Tier 3 oneon-one intervention services. As Grether and Sickman (2008) say Children with complex communication needs who use augmentative and alternative communication (AAC) are at risk for failure in the classroom and can benefit from the educational supports provided through RTI. While the RTI approach is not unproblematic, and has its critics, it at least moves away from one size fits all standardized IQ tests and the lifetime labels that often result, and points in a more promising direction. Mis-assessment of children with minimal speech has serious implications for the children and their families. It also has major implications for much of the published research on AAC. Many articles categorize their subjects by IQ or Mental Age (MA), a construct based on IQ. If the testing on which the categorization was based was unreliable, what does that say about the research? This is an issue that the AAC community needs to address urgently, both for the sake of potential and current AAC users, and to maintain the credibility of the field itself. 990 words Courchesne, V., Meilleur, A., Poulin-Lord, M., Dawson, M., Soulières, I., (2015) Autistic children at risk of being underestimated: school-based pilot study of a strength-informed
assessment Molecular Autism (2015) 6:12:1-10 Dawson M, Soulières I, Gernsbacher AM, Mottron L. (2007) The level and nature of autistic intelligence. Psychol Sci. 18:657 62. Grether SM1, Sickman LS. (2008) AAC and RTI: building classroom-based strategies for every child in the classroom. Semin Speech Lang. 29(2):155-63 Kasari C, Brady N, Lord C, Tager-Flusberg H. (2013) Assessing the minimally verbal school-aged child with autism spectrum disorder. Autism Res. 6:479 93. Wilkinson, K. & Rosenquist, C. (2006) Demonstration of a method for assessing semantic organization and category membership in individuals with autism spectrum disorders and receptive vocabulary limitations, Augmentative and Alternative Communication, 22:4, 242-257 The authors disclose they have no financial or other interest in objects or entities mentioned in this paper.