| Abstract: | In experiment 1, an extended functional analysis of self‐injury was conducted with a 21‐year‐old male diagnosed with autism and profound mental retardation. The multielement phase yielded undifferentiated results. Subsequent blocking of conditions plus the addition of a component allowing access to multiple sensory stimuli suggested that self‐injury was unrelated to programmed positive or negative reinforcement contingencies. The behavior appeared to be automatically reinforced; its occurrence decreased when access to alternative sensory stimuli was provided. Experiment 2 evaluated a treatment condition in which response‐independent access to these sensory stimuli was provided within the participant's everyday environment. Baseline and treatment frequencies of self‐injury were compared in a combined, multiple‐baseline‐across‐settings and ABAB design. The level of self‐injury decreased substantially during treatment. These results support the use of extended analog analyses of aberrant behavior in instances in which undifferentiated responding occurs in the initial analogue analysis. Additionally, a procedure is described for generalizing the intervention derived from the experimental analysis into the participant's everyday environment. Copyright © 2001 John Wiley & Sons, Ltd. |
