The Effects of Delayed Positive Reinforcement on Learning in Dogs
Browne, C. M. (2015). The Effects of Delayed Positive Reinforcement on Learning in Dogs (Thesis, Doctor of Philosophy (PhD)). University of Waikato, Hamilton, New Zealand. Retrieved from http://hdl.handle.net/10289/9808
Permanent Research Commons link: http://hdl.handle.net/10289/9808
Domestic dogs (Canis familiaris) have lived closely with humans for thousands of years. Successful dog training is important for dogs to fulfil the many roles that they play within human societies, and to aid good human-dog relationships and thus the welfare of both parties. The overall aim of this thesis was to investigate the interactions that take place between humans and dogs during training, with a particular focus on the timing of positive reinforcement delivery. The first two studies examined sources of dog training information. Dog owners and trainers rated their own personal experiences as being the most valuable source of dog training information, followed closely by free or low-cost resources such as books, discussions with other owners or trainers, and the Internet. The second study examined the content of five popular dog training books, and found that some of the information on aspects of learning theory and human-given cues presented in these books was supported by academic literature; however, inconsistencies were found across the books. Not all of the books contained enough information to enable readers to replicate their training advice. The third study in this thesis used video observations of owners and their dogs at dog training clubs to investigate the timing of positive reinforcement during training. Dogs failed to respond to 44.20% of their owners’ commands. Typically, conditioned reinforcement (e.g., verbal praise) was delivered first (average delay was 0.62 s), followed by unconditioned reinforcement (average delay was 0.98 s). This suggests that delayed reinforcement is commonplace in real-life dog training. Study four examined the effects of delays on dogs’ learning in a laboratory setting. When unconditioned positive reinforcement was delayed by 1 s, only 25.00% of dogs learned the task, whereas 60.00% of dogs who received immediate unconditioned reinforcement learned. Forty percent of dogs who received immediate conditioned reinforcement and unconditioned reinforcement delayed by 1 s, learned the task. These findings show that dogs’ learning is affected negatively by delayed positive reinforcement. The final study aimed to determine how dogs are able to learn in everyday dog training situations if delayed positive reinforcement is both a common occurrence and detrimental to dogs’ learning. Observations of owners training their dogs revealed that owners gave signals in the form of body movements such as hand gestures prior to providing any intentional feedback (e.g., verbal praise) to their dogs in 75.26% of trials (average delay was 0.31 s), which may bridge the temporal gap between dogs’ responses and delayed reinforcement, thus aiding dogs’ learning. Overall findings from this thesis suggest that delays to positive reinforcement are detrimental to dogs’ learning. Advice to dog owners and trainers should emphasise speed and consistency when delivering feedback to dogs.
University of Waikato
All items in Research Commons are provided for private study and research purposes and are protected by copyright with all rights reserved unless otherwise indicated.
- Higher Degree Theses