Association for Behavior Analysis International

Social reinforcement is a topic of enormous conceptual and applied significance. The basic mechanisms of social reinforcement are not well understood, however. The goal of this session is to bring together several lines of research designed to assess the efficacy of social reinforcement across a range of species and different reinforcer types. Feuerbacher & Wynne will discuss research aimed at measuring the reinforcing efficacy of human social interaction for dogs, including analyses of reinforcer duration. Pinkston and colleagues will describe research on social/sexual reinforcement with rats, as a baseline against which to measure the response weakening effects of antidepressant drugs. Call and colleagues will present data comparing the relative efficacy of social and non-social reinforcers in autistic and non-autistic children, putting a sharper quantitative point on general methods for assessing social deficits in autism. Hackenberg and colleagues will discuss research exploring the effects of social familiarity on preference for social reinforcement in rats. Together, the work illustrates some promising methods for assessing and quantifying the efficacy of social reinforcement across species, settings, and reinforcers a first step in a comparative analysis of social reinforcement.

Whether human social interaction can function as a reinforcer for domestic dog behavior remains unclear, but is an important question for owners hoping to maintain desirable behavior in their dogs through social interaction. Previously, we demonstrated that brief human interaction did not function as an effective reinforcer for dog behavior. However, others suggested longer interaction might, although confounding contingencies prevented clear conclusions. Thus, we examined whether 30 s of social interaction would function as a reinforcer for dog behavior. We saw little effect and no difference compared to dogs that received 4 s of interaction. To investigate a transient response spike in some subjects, we provided some dogs 4 min of presession noncontingent interaction. This did not affect responding. Finally, we implemented a multielement design in which dogs alternated within and across days between abolishing operation (presession attention) and establishing operation (presession ignore) conditions. Half of the dogs showed a reinforcement effect of 30 s social interaction, although with little distinction between abolishing operation and establishing operation conditions. The other half showed no reinforcement effect. Our results suggest that for some dogs, longer social interaction can function as reinforcer for their behavior, but session spacing might be a critical variable.

There is growing support for the theory that disruptions in the degree to which social interactions are reinforcing may constitute a root cause of autism spectrum disorder (ASD). This paper will present studies that have attempted to develop methods for quantifying the degree to which social interactions function as reinforcers for children with ASD. This will include the results of a study that used PR schedules to compare the relative reinforcing efficacy of social attention and leisure items in children with ASD (n=8), Williams Syndrome (WS; n=4), and typically developing peers (n=9). Participants in the ASD group exhibited higher breakpoints and Omax for leisure items than for attention, whereas children in the typically developing and WS groups exhibited the opposite pattern. Results will be discussed in terms of how these methods compare to other approaches to quantifying the reinforcing efficacy of social interactions in children with ASD.

In a concurrent schedule procedure, female rats made repeated choices between two options, each of which opened into side compartments in a three-compartment apparatus. Responses on levers in the middle chamber opened guillotine doors separating the middle from the side compartments, permitting 45-s access to the side compartment adjacent to target lever. In baseline conditions, one side compartment contained a familiar female rat (the focal rats cagemate, with which it was housed outside the experiment), whereas the other side compartment was empty. In a second condition, the empty chamber was replaced with a new female rat, unfamiliar to the focal rat, thereby permitting a choice between a familiar and unfamiliar rat. This was followed by a return to baseline conditions, with choice between a familiar rat and an empty chamber. The focal rats showed a clear and consistent preference for the familiar rat over the empty chamber in baseline conditions, but reversed their preference in favor of the unfamiliar rat when pitted against a familiar rat. The methods show promise as an experimental paradigm for evaluating and quantifying preference between qualitatively different social reinforcers.