Give the dog a big bone: Magnitude but not delivery method of food impacts preference and reinforcer efficacy in dogs.

Abstract

Despite the prevalence of reinforcement-based training practices in animal training, little research has investigated how to enhance the efficacy of delivered consequences; effective reinforcers are critical for maintaining long chains of behavior, such as in working animals, and for competing with environmental reinforcers. Two potentially easy methods to increase efficacy of reinforcers used in animal training are increasing magnitude (e.g., deliver 4 pieces of food instead of 1) or altering delivery (e.g., offer reinforcers 1-by-1 with simultaneous praise or all at once without praise). In Experiment 1, we evaluated dogs' preference for (a) several treats delivered 1-by-1 with simultaneous praise or all treats delivered at once without praise and (b) large or small food magnitudes. In Experiment 2, we assessed reinforcer efficacy of these delivery methods and magnitudes by measuring dogs' break points using Basis 2 progressive ratio 1 schedule of reinforcement. We also evaluated whether delivery of praise when delivering treats impacted reinforcer efficacy of the treats. Results showed that magnitude reliably increased reinforcer efficacy, with large magnitude reinforcers producing break points nearly twice those of small magnitude reinforcers. In contrast, delivery method had small, inconsistent effects on reinforcer efficacy. The provision of simultaneous praise had no effect on responding. Generally, preference tests predicted relative reinforcer efficacy. Our results suggest small treats can be used in training for maintaining short behavior chains but larger magnitude treats are likely warranted for longer behavior chains.

Social interaction can function as a reinforcer for dogs: Effects of stimulus duration and session parameters

“…That we found no differences in responding between 4-s and 30-s social interactions was not expected given that the magnitude of reinforcement often affects reinforcer value (e.g., Hoch et al, 2002;Hodos & Kalman, 1963;Nader & Woolverton, 1991), including in dogs (Feuerbacher et al, 2022). However, there might be differences that we did not detect; in an effort to reduce satiation as a possible factor, we capped sessions at 10 trials.…”

“…Prior research has suggested that petting, but not vocal praise, can function independently as a reinforcer for dog behavior (McIntire & Colley, 1967). More recent research has found vocal praise is not a preferred reinforcer for dogs (Feuerbacher & Wynne, 2015) and does not influence the reinforcer value of food when delivered simultaneously with food (Feuerbacher et al, 2022). Additionally, investigating different forms of petting and its possible interaction with response topography would be useful; we used more active massaging and scratching petting.…”

“…The recruited dogs differed in their prior history with the experimental response (nose touch). Two dogs (Iorek and Ninja) had extensive training with emitting a nose touch to their owner's hand to receive food and have been subjects in prior research using nose touch as a response that produced food (Feuerbacher et al, 2022). Nintendo and Leica had some but little training by their owner to emit a nose touch to the owner's hand for food, and the owner had not been engaged in any specific training of this response for at least 2 months prior to the experiment.…”

“…The research discussed thus far has used relatively short social interaction. It is possible that longer social interaction could function as a more effective reinforcer; dogs have been shown to be sensitive to the magnitude of other reinforcers (food; Feuerbacher et al, 2022), and humans have been found to be sensitive to different magnitudes of social interaction (e.g., Trosclair-Lasserre et al, 2008). Fonberg et al (1981) evaluated longer social interaction, which was 30 s of petting plus a vocal praise statement ("good dog") as a reinforcer and reported it was as effective as food in that it produced the same percentage of correct for the dogs' behavior of placing and holding their foreleg on a food tray.…”

Reinforcer efficacy of grain for horses

“…As with training, the target response was a muzzle touch to the target, though during testing the horse had to walk back and forth across the testing stall to access the target. Horses worked on a Basis 2 Progressive Ratio 1 schedule (Feuerbacher et al, 2022 ; Vicars et al, 2014 ), such that the horse had to complete each response requirement twice before the response requirement was increased by one touch. Response requirements began at one touch at the beginning of each testing session.…”