I got an email reminder yesterday about this weekend’s Equiosity webinar. This reminded me that I still hadn’t posted my thoughts from the March webinar!
For the March webinar, Alexandra Kurland and Dominique Day invited Dr. Jesús Rosales-Ruiz to join them for a discussion about schedules of reinforcement. This can often be a rather dry topic, but it was actually a delightful conversation!
Toward the end of the conversation, I even got to share a little bit about a project I’ve been working on related to DRO schedules of reinforcement.
Here are a few of my notes and thoughts from the webinar. If you missed the webinar, it’s available for purchase from the Equiosity website.
1) What gets reinforced?
We often talk about reinforcers “strengthening” behavior. However, one thing that Dr. Jesús Rosales-Ruiz discussed in the webinar is that reinforcement may strengthen a particular dimension of the response, rather than the whole response.
By “dimension,” Jesús meant factors such as speed, duration, frequency, magnitude, height, or location.
Alex gave an example of teaching a horse to shift his weight and take a step back. Your click could be reinforcing the direction of travel, the speed of the leg lift, the height of the leg lift, or something else entirely.
So, how do you know what dimension or dimensions are being reinforced?
This is why it’s so important to look at patterns of behavior over time. If you just look at one or two instances of the behavior, you can’t be sure what was really reinforced.
However, when you look at multiple instances of the behavior and how it is changing, you’ll get a clearer picture of what you are reinforcing.
2) Schedules don’t create motivation
Sometimes, trainers suggest using certain schedules of reinforcement because a particular schedule will create more motivation.
However, in the lab, it is very easy to design experiments in which two animals are equally hungry, but one responds very quickly and the other very slowly.
The schedule will create a certain rate and pattern of responding. However, it does not reflect the animal’s level of motivation or create more motivation.
3) Think in terms of units
Many types of schedules of reinforcement involve intermittent reinforcement. That is, not every response is reinforced. For example, the rat will receive a food pellet every fifth time that it presses the lever.
However, Jesús suggested that it is more helpful to talk about reinforcing units of behavior, rather than a certain individual response.
So, in the rat example, it is more helpful to talk about reinforcing a unit of five responses, instead of reinforcing the fifth response.
This was an important point that came up several times during the conversation.
4) Units help you build duration
Thinking in terms of units will help you build longer behaviors. You can think of shaping the unit, just like you would shape any other response.
So, instead of thinking “I’m going to reinforce the second response each time,” you can instead think, “I’m going to reinforce a unit of two responses.” Later in the training process you might reinforce a unit of three responses, then five responses, and so on.
Thinking in terms of units will help you build cleaner behaviors.
During the webinar, Jesús showed a graph of a pigeon from an experiment in the 1950s. The pigeon pecked 20,000 times for one reinforcement. It took the pigeon more than three hours to peck this many times. That’s a lot of pecks and required some careful shaping!
Where to learn more
If you missed the webinar, the recording is available for purchase from the Equiosity website. You can also sign up for the webinar that Dr. Jesús Rosales-Ruiz will be doing this weekend with Alexandra Kurland and Dominique Day. The topic is “Cues in Context,” which should generate some great conversation.