Stars on a chart, chips in a cup, points on a clipboard, punches on a card: token systems show up in clinics, classrooms, homes, and workplaces. This lesson covers how a token economy works, what your job is when you run one, and how to keep the tokens meaningful.
Why it matters
Many reinforcers are hard to deliver the instant a response happens. You cannot hand over 10 minutes of tablet time after every correct answer without stopping the lesson. A token solves that problem: you deliver it right away, without interrupting teaching, and the learner trades tokens for the bigger reinforcer later.
Token systems also help learners tolerate delays, work for longer stretches, and move toward the kinds of reinforcement used in everyday life, like grades, paychecks, and loyalty points. On the exam, expect scenarios about naming the parts of a token economy, delivering tokens correctly, and explaining why a token system stopped working.
Key ideas
- Tokens have no value of their own at first. A plastic chip is just a chip. It becomes a reinforcer because it is reliably exchanged for things the learner already wants.
- Every token economy has the same core parts. The plan defines (1) the target behaviors that earn tokens, (2) the tokens themselves, (3) the backup reinforcers tokens buy, and (4) the exchange rules: how many tokens, when, and where.
- Teach the value before you stretch it. For a learner new to tokens, the plan often starts with a quick exchange, sometimes after a single token, and then gradually raises the number required.
- Deliver tokens like any reinforcer. Give the token immediately after the target response, only when the response happens, and pair it with specific praise ("You put every block away!").
- Variety keeps tokens strong. A menu of backup reinforcers protects against satiation. If a learner can buy only one thing and gets tired of it, the tokens lose value too. Preferences shift as motivating operations change, so watch for signs that a backup reinforcer is losing its pull.
- Follow the plan exactly. Your supervisor sets the target behaviors, prices, exchange times, and any rules about losing tokens. Changing them on your own breaks the contingency and muddies the data.
- The long-term goal is to fade. Over time, supervisors thin the system by requiring more tokens, delaying the exchange, and leaning more on praise and natural reinforcers.
Definitions
- Token economy: a reinforcement system in which learners earn tokens for specified behaviors and later exchange them for backup reinforcers.
- Token: an object or symbol (chip, sticker, star, point, checkmark) delivered right after a target behavior and later traded in.
- Backup reinforcer: the item, activity, or privilege a learner buys with tokens, such as a snack, game time, or being line leader.
- Generalized conditioned reinforcer: a learned reinforcer paired with many different backup reinforcers. Because it does not depend on any single one, it is less affected by satiation or a lack of motivation for any one item. Well-run tokens and money are classic examples.
- Exchange schedule: the rules for trading tokens in: how many tokens each backup reinforcer costs and when exchanges happen.
- Response cost: an optional component in which a set number of tokens is taken away after a specified problem behavior. It is a form of negative punishment and is used only when the plan includes it.
Examples from a session
Three learners, one system
- Clinic: Leo, 5, is new to tokens. His plan starts him on a one-token board: he matches a color, gets a token with praise, and immediately trades it for 30 seconds of bubbles. Over the next weeks, the board grows to 3 and then 5 tokens.
- School: Amara, 11, earns a point on a card each time she raises her hand to ask for help in math. At the end of class she spends points from a menu that includes drawing time, a homework pass, and helping in the library.
- Day program: Ray, 24, earns a punch on a card for each completed packaging task at his vocational site. Ten punches buy a coffee-shop trip on Friday afternoon, much like a paycheck at the end of the week.
Look at what the RBTs did in each case: tokens were delivered right after the target behavior, paired with praise, and exchanged on a predictable schedule for things the learner actually wanted.
Common mistakes / exam traps
- Starting with a long wait. Requiring 10 tokens before a learner has ever traded one in gives the tokens no chance to gain value. Follow the plan's starting exchange rule.
- Delivering tokens late or for free. Tokens handed out at the end of the period, "on credit," or to calm a learner down no longer depend on the target behavior.
- Letting the menu go stale. When a learner stops working for tokens, check whether the backup reinforcers still matter to them, and report the change to your supervisor.
- Taking tokens away when the plan does not say to. Response cost is a punishment procedure. Removing tokens because a learner was rude, when the plan has no response cost, is a plan change you are not authorized to make.
- Leaving tokens within reach. If a learner can grab tokens from the bin, they get tokens without doing the target behavior. Keep them under your control.
- Changing prices on the fly. Dropping the price because a learner is upset, or raising it because they are doing well, is a supervisor decision. Share your data and observations instead.
Exam trap
Tokens are conditioned reinforcers, not unconditioned ones. If a scenario says tokens "stopped working," look first for a problem with the exchange: backup reinforcers that lost value, exchanges that stopped happening, or tokens available for free.
Exam trap
Earning rule vs exchange schedule: "a sticker for every 2 puzzles" says how tokens are earned; "10 stickers buy Friday recess" is the exchange schedule. A reinforcer given directly, with no token to trade, is a plain reinforcement schedule, not a token economy.
Quick check
A learner keeps earning stickers on a chart, but for weeks the planned Friday trade-in keeps getting skipped. The stickers no longer seem to motivate him. What is the most likely reason?