Schedules of Reinforcement

One of my dreams has always been to run an extremely successful comic book store. There’s always something that sparks a conversation, be it powerscaling, morality discussions, or character motives and themes. Honestly, it’s a joy to watch people get that invested. So, for the sake of understanding Schedules of Reinforcement, let’s assume I’ve actually managed to open one.

Let’s fast forward to opening day. I’ve opened the doors early, stocked the shelves, got everything ready, and now I’m standing behind the counter waiting for my first customer. Unfortunately, it’s not exactly going according to plan. The store is quiet, and I’m beginning to wonder whether anyone is actually going to show up. Thankfully, four of my scientist buddies helped spread the word about the store’s inauguration and customers slowly begin pouring in. Given that it’s opening day, I decided to try out four different discount strategies for my customers. Which strategy a customer receives depends entirely on the first letter of their name. I divide my customers into four groups: Group 1 consists of customers whose names begin with A-F, Group 2 is G-L, Group 3 is M-R, and Group 4 is S-Z.

The important thing to understand is that I am using these discounts as reinforcers. In operant conditioning, reinforcement is the increased likelihood that a behaviour will occur again because of a consequence that follows it. In my store, the behaviour I’m interested in is purchasing, while the discount acts as the reinforcer. A Schedule of Reinforcement simply tells me when and how that reinforcer will be delivered.

Group 1 is first up! All members belonging to that group will receive a 20% discount on every fifth purchase. It doesn’t matter how long their visit lasts or how much time passes between visits. Every time someone from Group 1 enters my store and purchases something, that purchase counts toward their discount journey. Once they have made five purchases in total, they receive a 20% discount, putting Group 1 on a Fixed Ratio, or FR 5, Schedule of Reinforcement. "Ratio" tells me that the reinforcer's delivery is dependent on the number of purchases. In this case it would be 5 purchases. ”Fixed" simply means that the requirement of 5 behaviours remains the same throughout the schedule. The consistent pattern is therefore very straightforward and consistent: a customer completes five, and on the fifth purchase, they receive their discount. Because my customers know exactly how many purchases are required, I notice that Group 1 develops a relatively high rate of purchasing. They know what they have to do to earn the discount, so they keep working towards it. However, after a while, I notice that after receiving their discount, some customers seem to disappear. Sometimes I won't see them again for two or three weeks. This is known as a postreinforcement pause. It is a temporary reduction in responding immediately after a reinforcer has been delivered.

And it kinda makes sense when I think, you know? They've just received their discount so there isn’t much reason to immediately begin working toward the next one. They know that another five purchases are required, so they may take a break before starting the process again. But I don’t want that! So I decide to manipulate the size of the reinforcer and see what happens. When I increase the discount from 20% to 30%, the postreinforcement pause shortens. The larger reward gives customers more incentive to begin working toward the next one. Simple as!

Group 2 is my favourite. Instead of giving them a discount after a fixed number of purchases, I vary the requirement. Sometimes a customer receives a discount after their third purchase. Sometimes after their second. Sometimes they have to make six purchases. Occasionally, they might even receive a discount on the very next purchase after their previous one. The important thing is that they cannot predict exactly when reinforcement will occur. This puts Group 2 on a Variable Ratio, or VR 5, Schedule of Reinforcement. Just like the Fixed Ratio schedule, reinforcement is based on the number of responses. The key difference here is that the number of responses required changes following every visit. For example, an FR 5 schedule would look like this:

5, 5, 5, 5, 5

A VR 5 schedule, instead, could instead look like this:

5, 2, 7, 3, 8

The requirements are different each time, but the average is still five. This unpredictability is what makes the Variable Ratio schedule so effective at producing a high and steady rate of responding. The Group 2 customers don’t know when exactly reinforcement will occur so there they continue shopping for a discount. As a result, Group 2 becomes my most frequent and profitable customers! They keep coming back, keep making purchases, and rarely take a significant break after receiving reinforcement. How wonderful! More people like this, please!

The one thing I would need to be careful about is suddenly making the purchase requirement far too high. This could result in something called ratio strain, which can cause customers to respond negatively. Don’t need negative reviews so early into my career as a comic book store owner!

There is something I noticed about customers belonging to Group 3 after a while. They aren't particularly interested in making frequent purchases. Instead, they seem to enjoy spending time in my store browsing through the comics and sharing conversations with those that come along with them or those already present in the store. Somehow, all four of my scientist friends fall into this group too. They spend hours looking at comics, ask me endless questions, and then somehow still have the audacity to haggle over the price after I've already given them the Friends and Family discount. Maybe I need new friends -- getting off topic. Anyway, because Group 3 isn't my target audience, I don't want to waste my discounts trying to encourage them to increase their frequency of shopping. Their discounts will be delivered based on time amount of time passed since previous discount. Suppose I decide that a discount becomes available to them every five days. It that case, it doesn't matter how many purchases the customer makes during those five days; they could make twenty purchases or fewer than five. What matters is that once those five days have elapsed, the first purchase they make is a discounted one.

This puts Group 3 on a Fixed Interval 5-day, or FI 5-day, Schedule of Reinforcement. “Interval" here means that reinforcement depends on the passage of time rather than the number of times purchasing occurs. This produces a very different pattern of behaviour from the ratio schedules. Because my customers know that a discount is about to become available after a predictable period of time, they don't necessarily have much reason to purchase immediately after receiving a discount. They can simply spend their time browsing through the comics and once the discount period arrives, their frequency of purchase increases. For my purposes, that's perfectly fine. Group 3 is mainly there to make my store look busy.

Finally, I have Group 4. I decide to use a strategy similar to Group 3, but I remove the predictability. Instead of making a discount available after exactly five days, I vary the amount of time that has to pass. It could be one day, three days, a week, or even just five hours. Group 4 has no way of knowing exactly when the next discount will become available. The only way to find out is to come back to my store and make a purchase. This puts Group 4 on a Variable Interval 5-day, or VI 5-day, Schedule of Reinforcement.

Like the Fixed Interval schedule, reinforcement is dependent on the passage of time. However the amount of time is unpredictable. The "5 days" represents the average interval, rather than a fixed requirement. This unpredictability changes the customer's behaviour. If the discount were always available exactly five days after the previous one, customers could simply wait until the fifth day before coming back. But if the interval could be five hours, one day, three days, or a week, they can't predict when reinforcement will become available.

As a result, Variable Interval schedules tend to produce a steady and consistent rate of responding. They are also relatively resistant to postreinforcement pauses because the customer cannot simply predict when the next opportunity for reinforcement will occur. That means Group 4 keeps checking to see if a discount is available.

Now the final issue I need to fix are these friends of mine. How would I change their behaviour of haggling? Given that they’re all working at Caltech, they’re quite intelligent so I can’t make it quite as obvious as Punishment… Maybe Differential Reinforcement? But which kind…?

References:

Cooper, J. O., Heron, T. E., & Heward, W. L. (2020). Chapter 13: Schedules of Reinforcement. In Applied Behavior Analysis, 3rd edition. Pearson Education, Inc.