Effects of Blended Integrity Failures on Reinforcement-Based Interventuions

Reinforcement-based interventions reduce problematic behavior when implemented as designed. However, the effectiveness of these interventions may decrease when deviations from treatment protocols (i.e., treatment-integrity errors) occur. Treatment-integrity errors differentially impact reinforcement-based interventions based on multiple factors, including how frequently errors occur and the intervention type. Even nominally acceptable integrity values (e.g., 80%) may be detrimental depending on the intervention. To evaluate this possibility and directly compare the effectiveness of multiple reinforcement-based interventions, we conducted two within-subject evaluations using laboratory arrangements. For both experiments, we recruited four undergraduate students to participate in a computer task that involved clicking on moving circles to earn points. During Experiment 1, we compared the effectiveness of Fixed-Time Schedules (FT) and Extinction (i.e., Noncontingent Reinforcement) and Differential Reinforcement of Alternative Behavior (DRA) when implemented at 80% integrity. During Experiment 2, we compared the effectiveness of ratio-based DRA (DRA with a ratio schedule maintaining alternative behavior) and interval-based DRA (DRA with an interval schedule maintaining alternative behavior). Results were idiosyncratic across participants. However, DRA with a ratio schedule was the only consistently effective intervention when implemented at 80% integrity; neither FT nor interval-based DRA were consistently effective when implemented at 80% integrity. Implications for research on effects of treatment-integrity errors and applied practice are discussed