6.4: Schedules of Reinforcement

Fixed Ratio Schedule (FR).

With this schedule, we reinforce some set number of responses. For instance, every twenty problems (fixed) a student gets correct (ratio), the teacher gives him an extra credit point. A specific behavior is being reinforced — getting problems correct. Note that if we reinforce each occurrence of the behavior, the definition of continuous reinforcement, we could also describe this as an FR1 schedule. The number indicates how many responses have to be made and, in this case, it is one.

FR schedules are characterized by a high rate of performance with short pauses called post-reinforcement pauses. A rat, for instance, will press the lever quickly and until a food pellet drops, eat it, walk around the chamber for a bit, and then return to the business of lever pressing. The pause to walk around the chamber represents taking a break, and when rats (or people) are asked to make more behaviors the break is longer. In other words, the pause after making 50 lever presses (FR 50) will be longer than after making 20 lever presses (FR 20). See Figure 6.2 for the typical response pattern on an FR schedule.

We might decide to reinforce some varying number of responses such as if the teacher gives the student an extra credit point after finishing an average of 5 problems correctly (VR 5). We might reinforce after 8 correct problems, then 5, then 3, and then 4 problems. The total number of correct problems across trials is 20. To obtain the average, divide by the number of trials which is 4 and this yields an average of 5 problems answered correctly. This is useful after the student is obviously learning the material and does not need regular reinforcement. Also, since the schedule changes, the student will keep responding in the absence of reinforcement.

VR schedules yield a high and steady rate of responding and typically produce fewer and shorter post-reinforcement pauses. Like an FR schedule, the payoff is the same. An animal on an FR 25 schedule will receive the same amount of food pellets as a different animal on a VR 25 schedule. See Figure 6.2 for the typical response pattern on a VR schedule.

Fixed Interval Schedule (FI).

With an FI schedule, you will reinforce the first behavior made after some set amount of time. Let’s say a company wanted to hire someone to sell their products. To attract someone, they could offer to pay them $10 an hour 40 hours a week and give this money every two weeks. Crazy idea but it could work. Saying the person will be paid every indicates fixed, and two weeks is time or interval. So, FI. Rats in a Skinner box will be reinforced when they make the first lever press after 20 seconds of time has passed. Hence, the rat would be on a FI 20(sec) schedule and would not receive any food until another 20 seconds has passed (and then it pushed the lever).

FI schedules produce a response pattern best described as scalloped shape. At the beginning of the interval, responses are almost non-existent but as time passes the rate of responding increases, especially as the end of the interval nears. The FI schedule has a post-reinforcement pause, like the FR schedule. Consider your study habits for a minute. When a new unit starts, you are not likely to study. As the unit progresses, you may start to study your notes, and as the unit nears the end, your rate of studying will be very high as you know the exam will be soon. This process is repeated when the next unit starts. See Figure 6.2 for the typical response pattern on a FI schedule.

Variable Interval Schedule (VI).

Finally, you could reinforce the first response after some changing amount of time. Maybe employees receive payment on Friday one week, then three weeks later on Monday, then two days later on Wednesday, then eight days later on Thursday. Etc. This could work, right? Maybe not for a normal 9-5 job, but it could if you are working for a temp agency. VI schedules reinforce the first response after a varying amount of time. Let’s say our rat in the FI schedule example was moved to a VI 20(sec) schedule. He would be reinforced for the next response after an average of 20 seconds. The actual intervals could be 30 seconds, 15 seconds, 25 seconds, and 10 seconds (30+15+25+10/4 = 20).

VI schedules produce almost no post-reinforcement pauses and a steady rate of responding that is predictable. For humans, waiting in line at the DMV is a great example of a VI schedule. Each person ahead of us that completes his or her business and leaves reinforces our waiting, but how long each person ahead of us takes when they make it to the counter will vary. We may be reinforced for our waiting after just 30 seconds with one customer but after several minutes with another customer. The amount of time we wait for each customer to have their problem solved produces an average amount of time after which we are reinforced for our response of waiting (and gradually sliding forward). See Figure 6.2 for the typical response pattern on a VI schedule.

Fixed and variable duration schedules.

In a fixed duration (FD) schedule, the organism has to make the behavior continuously for a period of time after which a reinforcer is delivered. An example would be a child practicing shooting hoops for 60 minutes before he receives a reinforcer such as playing a video game (FD 60-min). In a variable duration (VD) schedule, the behavior must be made continuously for some varying amount of time. A coach may have his players running drills on the football field and provide the reinforcement of a refreshing drink at different times. This could occur after 5, 12, 7, and 16 minutes of practice but on average, the athletes will practice 10 minutes before they receive reinforcement (VD 10-min). Interestingly, as I write this module, I realized that I have done all of my textbook writing on an FD schedule. I generally sit down and write for a designated period of time each day. Once this time is complete, I provide myself reinforcement such as getting to go to the gym, reading my current science fiction novel, or taking a break to watch a movie. The reinforcement comes no matter how productive I am, and some days I am not as productive as others. I can write 20 pages in 2 hours when properly motivated but other days may just write 5 pages in the same period. Should I write according to an FR schedule though? Instead of writing for a fixed period of time (FD 120-min), I could decide to complete a certain number of sections in a module. For my current project, I may decide to write 2 sections in a module per day or a FR 2-sections schedule, before receiving one of the aforementioned reinforcers. How do you tackle studying or writing a large term paper? Do you use an FD or FR schedule and does it work for you?

Noncontingent schedules (FT/VT).

Noncontingent Reinforcement Schedules involve delivering a reinforcer after an interval of time, regardless of whether the behavior is occurring or not. In this respect they are response-independent schedules which should make sense given the fact that they are noncontingent, meaning there is no If-then condition. In a fixed time (FT) schedule an organism receives reinforcement after a set amount of time such as a pigeon having access to food after 30 seconds (FT 30-sec) whether it is pecking a key or not. In an FI 30-sec schedule, the pigeon would have to peck the key after the 30-second interval was over to receive access to the food. FT schedules are useful in token economies as you will come to see in Module 7.

In a variable time (VT) schedule, reinforcement occurs after a varying amount of time has passed, regardless of whether the desired behavior is being made. In a VT 20-sec schedule, the rat receives a food pellet after an average of 20 seconds whether it is pushing the lever or not. This reinforcement could occur after 18, 34, 7, 53, etc. seconds with the average being 20 seconds.

Progressive schedules.

In a progressive schedule, the rules determining what the contingencies are change systematically (Stewart, 1975). One such schedule is called a progressive ratio (PR) schedule (Hodos, 1961) and involves the requirement for reinforcement increasing in either an arithmetic or geometric way, and after each reinforcement has occurred. For instance, a rat may receive reinforcement after every 4, 6, 8, 10, and 12 lever pushes (arithmetic) or after 6, 12, 24, 48, and 96 times. The reinforcer may even change such as the quality of the food declining with each reinforcement. Eventually, the rate of behavior will decrease sharply or completely stop, called the break point. Roane, Lerman, and Vorndran (2001) utilized a PR schedule to show that reinforcing stimuli may be differentially effective as response requirements increase. Stimuli associated with more responding as the schedule requirement increased were more effective in the treatment of destructive behaviors than those associated with less responding. The authors note, “practitioners should therefore consider arranging reinforcer presentation according to task difficulty based on the relation between reinforcer effectiveness and response requirements” (pg. 164).

DRA or Differential Reinforcement of Alternative Behavior

This is when we reinforce the desired behavior and do not reinforce undesirable behavior. Hence, the desired behavior increases and the undesirable behavior decreases to the point of extinction. The main goal of DRA is to increase a desired behavior and extinguish an undesirable behavior such as a student who frequently talks out of turn. The teacher praises the child in front of the class when he raises his hand and waits to be called on and does not do anything if he talks out of turn. Though this may be a bit disruptive at first, if the functional assessment reveals that the reinforcer for talking out of turn is the attention the teacher gives, not responding to the child will take away his reinforcer. This strategy allows us to use the reinforcer for the problem behavior with the desirable behavior. Eventually, the child will stop talking out of turn making the problem behavior extinct.

What if we instead need to eliminate a problem behavior – i.e. reducing it down to no occurrences? DRO is the strategy when we deliver a reinforcer contingent on the absence of an undesirable behavior for some period. We will need to identify the reinforcer for the problem behavior and then pick one to use when this behavior does not occur. Determine how long the person must go without making the undesirable behavior and obtain a stopwatch to track the time. Do not reinforce the problem behavior and only reinforce the absence of it using whatever reinforcer was selected, and if it is gone for the full-time interval. If the problem behavior occurs during this time, the countdown resets. Eventually the person will stop making the undesirable behavior and when this occurs, increase the interval length so that the procedure can be removed.

For instance, if a child squirms in his seat, the teacher might tell him if he sits still for 5 minutes he will receive praise and a star to put on the star chart to be cashed in at a later time. If he moves before the 5 minutes is up, he has to start over, but if he is doing well, then the interval will change to 10 minutes, then 20 minutes, then 30, then 45, and eventually 60 or more. At that point, the child is sitting still on his own and the behavior is not contingent on receiving the reinforcer.

There are times when we don’t necessarily want to completely stop a behavior, or take it to extinction, but reduce the occurrence of a behavior. Maybe we are the type of person who really enjoys fast-food and eats it daily. This is of course not healthy, but we also do not want to go cold turkey on it. We could use DRL and decide on how many times each week we will allow ourselves to visit a fast-food chain. Instead of 7 times, we decide that 3 is okay. If we use full session DRL we might say we cannot exceed three times going to a fast-food restaurant in a week (defined as Mon to Sun). If we eat at McDonalds, Burger King, and Wendy’s three times on Monday but do not go again the rest of the week we are fine. Full session simply means you do not exceed the allowable number of behaviors during the specified time period. Eating fast-food three times in a day is definitely not healthy, and to be candid, gross, so a better approach could be to use spaced DRL. Now we say that we can go to a fast-food restaurant every other day. We could go on Monday, Wednesday, and Friday. This works because we have not exceeded 3 behaviors in the specified time of one week. If we went on Sunday too, this would constitute four times going to a fast-food restaurant and we would not receive reinforcement. Spaced DRL produces paced responding.

There are times when we need to substitute Behavior A with Behavior B such that by making B, we cannot make A. The point of DRI is to substitute a behavior. If a child is made to sit appropriately in his seat they cannot walk around the room. Sitting is incompatible with walking around. DRI delivers a reinforcer when another behavior is used instead of the problem behavior. To say it another way, we reinforce behaviors that make the undesirable or problem behavior impossible to make. DRI is effective with habit behaviors such as thumb sucking. We reinforce the child keeping his hands in his pocket. Or what if a man tends to make disparaging remarks at drivers who cut him off or are driving too slowly (by his standard). This might be a bad model for his kids, and so the man’s wife tells him to instead say something nice about the weather or hum a pleasant tune when he becomes frustrated with his fellow commuters. These alternative behaviors are incompatible with cursing and she rewards him with a kiss when he uses them.

DRH or Differential Reinforcement of High Rates of Responding.

In this type of differential reinforcement, we reinforce a behavior occurring at a high rate or very often or seek to increase a behavior. Many jobs use this approach and reward workers who are especially productive by giving them rewards or special perks. For instance, I used to work for Sprint Long Distance when selling long distance to consumers was a thing. Especially good salespeople, such as myself, would win quarterly trips paid for 100% by Sprint and then the annual trip if we were consistent all year. We won the trips because we sold large numbers of long-distance plans, and other products such as toll-free numbers, in a designated period of time (every 3 months or every 12 months). Hence, a high rate of responding (i.e. engaging in the behavior of selling lots of plans) resulted in the reinforcement of quarterly or annual trips, and other benefits such as time off work and gift cards for local food establishments.