A comparison of two procedures for teaching language to autistic children

Discrete-trial procedures have been used successfully in language training with autistic children. Training is usually conducted by professionals in highly structured settings. A procedure called incidental teaching has been developed to train language in the natural setting using teachers as therapists. The purpose of the present study was to compare these two procedures and to deter mine which procedure was more effective in increasing autistic children's generalized spontaneous language in extra training settings.

Subjects were two autistic children, a five-year-old female and a ten-year-old male. The setting was a classroom divided into an academic and a preferred area. Two teachers were assigned to the classroom.

A multiple baseline across responses with reversal design was used. The dependent variables were: a) Total spontaneous language; and b) spontaneous use of target responses.Verb use was the targeted response for subject 1;labeling was the targeted response for subject 2. Independent variables were: Discrete-trial language training and Incidental teaching.

Observers measured spontaneous language in the preferred area and during snack, and the teachers maintained records of discrete-trial and incidental language training responses. Weekly reliability checks were conducted by a second observer.

Language training on each child's targeted response was con ducted during- four intervention periods: two using discrete-trial training and two using incidental teaching.

The major results showed that: a) Most changes in total spontaneous words were due to increases in the target responses; and b) incidental teaching was far more effective than discrete-trial training in increasing targeted responses. Future investigations should be undertaken to identify certain characteristics of an autistic child's language that would indicate a need for discrete-trial training in addition to incidental teaching.