This piece first appeared in the May 28, 2024 issue of our newsletter.
In our previous post, we discussed how a perceived error in a learner’s response does not represent a failure on the learner’s part, and can highlight a skill deficit that can often be addressed through antecedent adjustments.
Along the same lines, we always recommend giving your learner multiple options to contact reinforcement during training sessions, so that you can get a good read on how engaged your learner really is in the session. For example, adding a snuffle mat to a dog’s training session gives the dog a true choice to opt in or out of the session. He can leave the person and access the same reinforcement independently. When a dog leaves the session to access his own reinforcement it gives clear communication to the trainer that something in the behavior loop isn’t working. The rate of reinforcement could be too low or the shaping steps are too large and difficult. It could be that the session is too arousing and the dog needs time sniffing in the mat to settle her brain. As ethical trainers, we frequently assess the level of stress we’re observing and adjust our training to keep the animal successful. Frequent failure is frustrating to the teacher, but even more frustrating to the learner.
We can achieve great behavior results with minimal frustration and low latency through “errorless learning.” The term was first introduced in the 1950s by Charles Ferster, then tested in a notable 1963 study that compared an errorless approach against a “trial-and-error” approach to teaching pigeons to discriminate between a red and green light (Terrace, 1963). The study found that pigeons taught using an errorless approach were not only more successful in the task and quicker to learn, but also significantly less frustrated by the process. Within the last year, new research has come out to suggest that this may also be an effective approach to training dogs with minimal frustration (Handley et al., 2023).
For example, if a dog immediately jumps up every time he approaches despite you yelling “sit”, the cue obviously isn’t working. He isn’t naughty, rude, or dominant. Instead, shift the focus of your ABCs to train through an errorless approach. Drop food to reinforce the approach, then you have a behavior of “nose to ground” (while dog is eating), which is incompatible with jumping up, and you no longer need to try to cue a sit or catch the dog before she launches at you. You can also build enrichment into your differential reinforcement procedure – check out how we use frisbees to capture calm greeting behaviors!
Through our work with reducing arousal in shelter dogs, we’ve consistently found that foraging for food also tends to lower arousal, which can reduce some of the jumping up behavior. In this way, changing the antecedents immediately alters which behavior is offered, and nearly eliminates the “problem” behavior of jumping up. Will this exact approach work for all dogs who jump? No. We also have to look at the function of the behavior, what consequence is maintaining it, and what antecedent is triggering it. It could be that the approach is not the antecedent that needs changing, but the attempted interaction after the approach needs adjustment. Errorless learning reduces stress, but it also reduces the number of reps needed for skill acquisition. If you learn it “right” the first time, you don’t have to unlearn it later.
And if this works for pigeons and pups, why not people, too? Blaming a person for not being able to understand what is being taught to them is useless and counterproductive to the end goal. Just like with our canine learners, we need to meet individuals where they are and provide support through careful antecedent arrangement. Shelter staff and volunteers should have clear protocols and standards of care from the onset of onboarding and training. Clear communication (TAGTeach anyone?), and adherences to objective, measurable criteria, can reduce confusion and increase opportunities for reinforcement. One way we can promote growth and shift culture is reinforcing our human staff for meeting animal emotional needs over a “get it done” mentality. Adjust antecedents and consequences to reinforce Fear Free practices over how many animals were vaccinated in a day. Quality care takes more time and effort. Adjusting distant antecedents (e.g., ratio of staff to animals, resources to reduce stress, more time per patient, continuing education on body language and low stress handling) can set the stage for meaningful growth and development across staff, which ripples out to affect the entire population in care.
References:
Handley, K., Hazel, S., Fountain, J., & Fernandez, E. J. (2023). Comparing trial-and-error to errorless learning procedures in training pet dogs a visual discrimination. Learning and Motivation, 84, 101944. https://doi.org/10.1016/j.lmot.2023.101944
Terrace, H. S. Errorless Transfer of a Discrimination across Two Continua. Journal of the Experimental Analysis of Behavior, vol. 6, no. 2, Apr. 1963, pp. 223–232, https://doi.org/10.1901/jeab.1963.6-223.
コメント