Home / RBT Study Guide / Behavior Acquisition / Conditioned reinforcers and token economies
Domain C · Behavior Acquisition · Tasks C.2, C.11
Conditioned reinforcers and token economies
A conditioned reinforcer is something that became reinforcing by being paired with other reinforcers. A token economy puts that idea to work: the learner earns tokens for target behaviors and trades them for backup reinforcers.
Last reviewed November 10, 2026 against the BACB RBT Test Content Outline (3rd ed.)
Three kinds of reinforcers
Tasks C.2 and C.11 on the RBT outline cover establishing and using conditioned reinforcers and implementing token economies. Both rest on one distinction: where a reinforcer gets its value.
| Type | Where its value comes from | Examples |
|---|---|---|
| Unconditioned (primary) | No learning needed. It works because of biology. | Food, water, warmth |
| Conditioned (secondary) | Learned, through being paired with other reinforcers. | Praise, a sticker, a thumbs up, a high five |
| Generalized conditioned | Learned, through being paired with many different reinforcers. | Money, tokens, points |
Generalized conditioned reinforcers are especially useful because they do not depend on one kind of motivation. A snack loses its pull right after lunch, but a token that can be traded for a snack, a game or a walk still has value. (For why a single reinforcer rises and falls in value, see motivating operations.)
Establishing a conditioned reinforcer: pairing
Something neutral becomes a conditioned reinforcer through pairing. You present the neutral item together with, or just before, something that already works as a reinforcer, and you do it many times. Over time the once-neutral item starts to work as a reinforcer too.
In session
Tokens mean nothing to Mia yet. Her plan says: after each correct response, hand her one token and immediately trade it for a few seconds with her favorite spinner. After many pairings, the plan asks for two tokens before the trade, then three. The token has become worth working for.
Praise is often paired the same way. If your plan says to give praise every time you deliver another reinforcer, keep doing it even when it feels repetitive. That pairing is the procedure, and it is how praise becomes enough later on.
Common mix-up
A conditioned reinforcer can lose its value if the pairing stops for good. Tokens that can never be traded for anything stop working. Keeping the exchange going is part of keeping the tokens effective.
How a token economy runs
A token economy is a system in which a learner earns tokens for specific behaviors and exchanges them later. Your supervisor designs it. Every token economy has the same working parts:
- Target behaviors. A clear list of what earns a token, described so that anyone could tell whether it happened. See operational definitions.
- Tokens. Stars, stickers, checkmarks, points or chips. They are safe, easy to hand over quickly and hard to get any other way.
- Backup reinforcers. The things tokens buy, often chosen with preference assessments and shown on a menu.
- Exchange schedule. When and how tokens are traded: how many are needed and when the exchange happens (for example, once the board is full or at the end of an activity).
In session
Sam's board has five spaces. His plan lists two target behaviors: completing a worksheet row and asking for help with words. Each time one happens, you place a star on the board right away and say what earned it. At five stars Sam picks from his menu: three minutes of drawing, a short walk or a puzzle. Then the board is cleared and you start again.
Two schedules are at work here: how many responses earn a token, and how many tokens earn the exchange. Either one can be thinned over time, and that is the supervisor's decision. See schedules of reinforcement.
Common implementation errors
The design belongs to the supervisor. Running it correctly belongs to you. These are the errors to watch for:
- Delayed tokens. A token delivered a minute late may reinforce whatever happened in that minute.
- Tokens for the wrong behavior. Giving a token to stop whining, or "to be nice", breaks the contingency.
- Skipped or late exchanges. If earned tokens are not exchanged as scheduled, they lose value.
- Free access to backup reinforcers. If the tablet is available all day anyway, tokens for tablet time are worth little.
- Removing tokens that the plan does not allow. Taking tokens away is response cost, a punishment procedure. Do it only if it is written in the plan. See reinforcement vs punishment.
- Bargaining. Changing the price or the rules mid-session because the learner negotiates.
- Different rules with different staff. Everyone on the team runs the same system the same way.
If the system seems to have stopped working (the learner ignores tokens or refuses the menu), run it as written, record what you see and tell your supervisor. Changing the menu, the price or the targets is not an RBT decision.
How these tasks tend to be tested
A common question format describes an item and asks which kind of reinforcer it is. Ask whether it needed learning (conditioned) or not (unconditioned), and whether it can be traded for many things (generalized). Another format describes a token system with an error in it and asks what went wrong.
On the exam
The token is the conditioned reinforcer. The snack or activity it buys is the backup reinforcer. Questions often hinge on keeping those two labels straight.
Practice questions
Original questions in the style of the exam. Pick an answer before you open the explanation.
1. Praise does not yet work as a reinforcer for Leo. His plan has the RBT say 'nice work' and immediately hand over a favorite snack after each correct response, so that praise gains value. Which procedure is this?
- Pairing a neutral stimulus with a reinforcer
- Shaping closer approximations of a response
- Fading a prompt across successive trials
- Thinning the schedule of reinforcement
Show answer
A. Pairing a neutral stimulus with a reinforcer
Presenting praise together with an established reinforcer, again and again, is pairing, the way a conditioned reinforcer is established. It is not thinning, because nothing about how often reinforcement is delivered has changed.
2. Diego earns tokens that he can trade for snacks, tablet time, a walk or stickers. Because the tokens have been paired with many different reinforcers, they are best described as what?
- Unconditioned primary reinforcers
- Backup reinforcers for the system
- Generalized conditioned reinforcers
- Discriminative stimuli for requests
Show answer
C. Generalized conditioned reinforcers
A conditioned reinforcer that has been paired with many other reinforcers is a generalized conditioned reinforcer. The snacks, tablet time, walk and stickers are the backup reinforcers. The tokens are what Diego trades for them.
3. Ava's plan says she earns a token for each completed task. Halfway through a task she whines and asks for a token, and the RBT gives her one to keep the session calm. What is the problem?
- The exchange schedule was thinned too fast
- The token was not contingent on the target
- The backup reinforcers were not varied
- The token was delivered after a long delay
Show answer
B. The token was not contingent on the target
Tokens are delivered only when the target behavior listed in the plan occurs. Giving one for whining breaks the contingency and may reinforce whining. Delay is not the issue here, because the token came right away, just for the wrong behavior.
Common questions
What is the difference between a conditioned and an unconditioned reinforcer?
An unconditioned reinforcer works without any learning history, such as food, water or warmth. A conditioned reinforcer gained its value by being paired with other reinforcers, such as praise, stickers, points or tokens.
What are backup reinforcers in a token economy?
Backup reinforcers are the items, activities or privileges a learner gets by trading in tokens, such as a snack, time with a game or a walk outside. They are what give the tokens their value.
Can an RBT take tokens away for problem behavior?
Only if the behavior plan includes it. Removing tokens is a response cost procedure, which is a form of negative punishment, and the supervisor decides whether it is part of the plan. If the plan does not mention removing tokens, the RBT does not remove them.