Association for Behavior Analysis International

Despite the many intensive early intervention programs implemented every year, mainly with children diagnosed with autism, some aspects of regularly used techniques are under-researched, and thus best practice has not been identified. In this symposium we report studies on prompting strategies and error correction that seek to remedy limitations in the literature on these topics. The first paper describe experimental studies evaluating delayed prompting strategies in a conditional discrimination task. It was found that , across the studies, there were persistent variations in effectiveness in prompting strategies across participants, but with some evidence in later experiments of greater average effectiveness of progressively increasing prompt delays in reducing prompt dependence. The second study was a field study of error correction techniques used in several schools. It found substantial variation in the preferred techniques being used in classrooms , that appear unrelated to theory or an evidence base. The discussant will evaluate these contributions both in the light of the literature that evalautes featiues of practice in ABA classrooms and her knowledge of practice in ABA schools in Ireland.

Teaching conditional arbitrary discriminations with match-to-sample instruction to children with varying disabilities can involve either trial-and-error or errorless learning procedures. Progressive and constant prompt-delay (PPD, CPD) are two widely used errorless learning procedures but they are rarely systematically compared. Using an adapted alternating treatments design, the first experiment directly compared three errorless teaching conditions (PPD, CPD 2-s and CPD 5-s), with one trial-and-error control condition, across five participants with intellectual disabilities and autism. This was done using four sets of a three-stimulus array in a conditional discrimination table top task. Acceptable levels of inter-observer agreement and procedural integrity, collected in approximately 33% of sessions, were obtained. Prompt-delay conditions proved effective for two participants. The prompt-delay procedure was modified for the other three participants to include the use of differential reinforcement across two experimental conditions (CPD 5-s & Control). This strategy was effective in two of three participants. Overall, there was no strong contender for best strategy in this group of participants which was itself fairly typical of children in intensive intervention programs. In the second experiment, an adapted alternating treatments design using a 3-stimulus array conditional discrimination task was implemented with five participants. We sought to replicate and extend previous findings by directly comparing three variations of the prompt-delay procedure (PPD, CPD 5-s, CPD 2-s), in addition to a control, incorporating the use of differential reinforcement where necessary based on previous findings. For three participants, differential reinforcement was incorporated throughout. Acceptable levels of inter-observer agreement and procedural integrity were obtained across approximately 33% of sessions. As predicted, prompt delay alone was effective, in terms of acquisition, for two participants, while prompt-delay in addition to differential reinforcement was effective for two of the remaining three participants. For the remaining participant, successful acquisition was obtained using stimulus fading. A between subject analysis show progressive prompt delay (PPD) to be associated with the least average amount of training trials and errors to criterion as compared with other variations of this procedure.

Professionals working with children diagnosed with autism use a number of behavioural techniques to encourage students to succeed in their work. Although there has been a vast amount of research into the effects of different types and schedules of reinforcement on behaviour, there has been little or no research into the effects of another type of consequence; error-correction techniques. In the past, research has provided practitioners with guidelines to using reinforcement effectively, however there is a lack of published accounts of error-correction procedures that are well-specified , have been evaluated, and could be used for staff training. Often, considerable variation is seen among these error correction procedures across and within early intervention programmes and, as a result, their underlying behavioural mechanism is not clear. Further research is required in order to isolate the effective components and processes operating within different error-correction procedures. Based on an assessment of current staff, practices, this study categorized and assessed the variation in error correction procedures used within and across early intervention programs for young children with autism in a number of ABA schools in Ireland. It was found that considerable variation exists across these schools.This study is the initial stage of a program of research aimed at developing a teaching training package and a computer-based assessment of efficiency of various error-correction techniques.