Variable-time schedules protect against effects of fidelity errors during noncontingent reinforcement.

Abstract

Noncontingent reinforcement (NCR) consists of response-independent reinforcer delivery according to a time-based schedule. Common application of NCR also includes withholding reinforcers following target behavior (i.e., extinction). Prior research suggests that inconsistent implementation (i.e., implementation with fidelity errors) of NCR programmed with fixed-time (FT) schedules results in degraded therapeutic outcomes. We conducted a human-operant evaluation to assess whether there were differences in responding (e.g., computer clicks) during reduced-fidelity NCR between FT and variable-time (VT) schedules. We randomly assigned participants to experience analogues of NCR with FT or VT schedules. Each participant experienced baseline, full-fidelity, and reduced-fidelity NCR in an ABAC design; FT or VT schedules varied depending on group assignment. Full-fidelity NCR was similarly efficacious at suppressing target behavior across the FT and VT groups, but VT schedules suppressed target behavior significantly better (p = .01) during reduced-fidelity NCR than FT schedules. Implications for researchers and practitioners are discussed.