A Positive Affirmation Marker to Improve Response Rate and Reduce Non-target Behaviours During Reinforcement Thinning in Dogs

Most dog training protocols use food as an unconditioned reinforcer. To maintain a learned behaviour with reduced reinforcement it is necessary to transition from continuous reinforcement to a leaner schedule, during which it is desirable to minimise problematic behaviour by the dog who is receiving fewer rewards. Experimental evidence for the most ethical and effective way to manage this transition is lacking. This paper describes a proof-of- concept study of using a conditioned verbal “positive affirmation marker” (PAM) to support the transition from a continuous to a variable ratio schedule of reinforcement. Four dogs were trained to perform a new task before introducing PAM by higher order conditioning. The reward schedule was then thinned using PAM, and behaviours were compared on a variable ratio schedule of reinforcement under two conditions; with PAM or without, in a within- subject study design. We observed an increased response rate and a reduction of non-target behaviours, particularly those indicative of frustration, in the PAM condition compared to without PAM for three of the dogs but not the fourth. The reasons behind the variability observed, as well as limitations and possible implications and uses of this conditioned reinforcer, are discussed. We propose PAM as a useful addition to the dog training toolkit.