Effects of Varying the Quality of Therapist-Provided Social Interaction During Discrete Trial Instruction

Background: A behavioral approach has shown to be an effective method for the assessment and treatment of skill deficits in children diagnosed with autism and other developmental disabilities (Lovaas, 1987).  The literature regarding quality of social interaction provided to children with autism and other developmental disabilities during discrete trail training is limited.  However, behavior intervention manuals focusing on evidence based treatment for autism and related disabilities recommend high rates of enthusiastic interaction when conducting discrete trial instruction (Sundberg & Partington, 1998; Lovaas, 2003). 

Objectives: The objective was to examine the quality of social interaction and reinforcement provided during discrete trial instruction.

Methods: Two studies with two children with developmental disabilities were conducted.  Participants were a 5-year old male with Shaken Baby Syndrome and a 6 year old male with PDD-NOS with autistic-like features and language delays.  The first study was conducted in an ABAB reversal design across high and low quality social interaction during instruction with easy, mastered tasks.   The dependent variables included combined inappropriate behaviors and responding to instructions.  Combined inappropriate behaviors were calculated as behaviors per minute and were defined specifically for each participant.  Responding to instructions was calculated as a percentage of the number of correct plus incorrect responses to therapist provided instructions and excluded non-responses.  The independent variables included high quality social interaction and low quality social interaction.  The high quality interaction phase included smiling, physical attention such as hugs, tickles, or squeezes, and “Motherease.”  During the low quality social interaction phase the therapist was directed to avoid smiling, provide no physical attention, and to not speak in a “Motherease” manner (i.e., monotone).  The therapist presented the instructions in either high or low quality style and provided a consequence in the same high or low quality manner. Study 2 was also conducted in an ABAB reversal design across high and low quality social interaction, with the exception that the therapist provided instruction to complete more difficult, non-acquired tasks. The dependent variables for Study 2 remained the same.  Therefore, Study 2 was a systematic replication of study one with the addition of an evaluation of task difficulty in the form of non-acquired instructions.

Results: Data from both studies indicate idiosyncratic results suggesting that some individuals respond better with high quality, high energy instruction and interaction, while others may respond better to low quality, low energy interaction.  The participant who responded with more problem behavior during the high quality condition in experiment 1 showed the opposite effect when tasks were more difficult suggesting an effect of task difficulty. 

Conclusions: The results of these two studies provide some support for high quality social interaction when presenting difficult instructions to children with developmental disabilities.   This study suggests the importance of considering the quality of social interaction in teaching environments and indicates that individual differences may exist.  Therefore, blanket recommendations for high quality social interaction may not be warranted for all individuals.