The Analogical Reasoning Learning Test: Theoretical and Empirical Foundation of a Diagnostic Tool for Individuals with Moderate Mental Retardation

Abstract

Research as well as practice has demonstrated that classical testing procedures are inappropriate for individuals with moderate to severe mental retardation (MSMR). Several characteristics of the population, such as lack of understanding of the task instructions and demands, short attention span, weak communication skills, and slow information processing, lead to a general floor effect on traditional intelligence tests. Thus, the results are neither reliable nor valid. In addition to the problem of not being able to evaluate their cognitive competencies in a reliable way, there is a deep-rooted belief that these individuals cannot go beyond a concrete level of reasoning and that it is extremely difficult to improve their intellectual functioning. Consequently, individuals with IQs lower than 50-55 are often treated as one large homogeneous group and tend to be taken care of in special institutions, without differentiation between those who are able to develop their cognitive competencies and those for whom educational goals may be limited to the development of social and lifeskill competencies.The Analogical Reasoning Learning Test (ARLT; Hessels-Schlatter, in press; Schlatter, 1999; Schlatter & Büchel, 2000) has been especially constructed for the assessment of individuals whose IQ, as measured with a traditional test, would be below 50-55. The aim of the ARLT is to provide a reliable and valid estimate of the learning capacities of individuals with MSMR, i.e., to distinguish persons who can profit from cognitive training programs and more demanding schooling from those for whom such an approach would have little value.The ARLT is a dynamic procedure. It consists of 2x2 analogical matrices and is divided into three phases. The first one is a pre-training phase intended to familiarize students with the task demands and to teach them some cognitive prerequisites. The second one is a learning phase, whose aim is to teach students to solve analogical matrices, with the help of specific, standardized, and hierarchically ordered hints. The third phase takes place one week after learning. This phase is designed to evaluate maintenance and transfer capacity of the learned rules and is applied in a more static way. Learning capacity is categorized at three levels: gainer, non-gainer, or undetermined. The reliability and validity of the test were analyzed.The study involved a total of 58 participants, 38 male and 20 female, all educated in special schools for MSMR students. The mean chronological age was 13-11 (min=6-6, max=19-10). The experimental plan followed a pretest - retest - training - posttest design. The complete ARLT was administered as a pretest to all participants. After a delay of four weeks, the third testing phase was repeated in order to assess test-retest reliability. To evaluate the predictive validity of the ARLT, the participants were matched on the basis of their ARLT performance and were randomly assigned to either an experimental group (EG) with training in inductive reasoning or a control group (CG) without training. The training for the EG consisted of 8 to 12 lessons of about one-half hour each over a period of four weeks. The lessons included different kinds of tasks: Standard analogies (involving similar kinds of relations as in the ARLT), analogies requiring the induction and application of other kinds of relations, such as “lives in,” “one as opposed to many,” “is part of,” and classifications. A post-test including items corresponding to the training tasks was designed to evaluate the training effects. To estimate the discriminant validity, we also administered the Raven Kurzzeit-Lerntest (RKL) of Frohriep and Guthke (1992; Frohriep, 1978). The RKL is a learning test based on the Coloured Progressive Matrices, originally developed to detect developmentally delayed kindergarten children.The research showed that the ARLT provides a highly reliable and valid estimate of learning capacity for this population. The participants classified as non-gainers on the ARLT showed no improvement after a one-month training, in contrast to participants defined as gainers. A familiarization effect could, however, be found on the classification tasks. It should be stressed that the training in this study was relatively short. Further research is needed to find out whether the reasoning abilities of the non-gainers can be enhanced with longer training and perhaps other kinds of intervention. Comparison between the ARLT and the RKL showed that it is indeed analogical reasoning that is assessed and not simpler perceptive abilities.We have shown that the ARLT, by differentiating gainers from nongainers in a population classified as moderately to severely mentally retarded, leads to useful and beneficial information for educational purposes. It allows one to establish differentiated educational programs and to stimulate the students who prove to be able to benefit from a cognitive training (gainers) to attain higher levels of cognitive functioning. The test can also help to counter the general low expectations of teachers, educators, and psychologists with regard to individuals whose IQ’s range is below 55.