How to Fade Classroom Rewards Without Losing the Behavior

Every teacher who builds a reward system eventually hears the warning, sometimes from a colleague and sometimes from the voice in their own head: what happens when the points stop? It's a fair question. A classroom that only runs when tickets are flowing is a rental. But the research answer is genuinely hopeful. Behavior falls apart when rewards end abruptly, and abruptly is a choice. Fading is the alternative, and it's a skill you can plan like any other part of teaching.

Why quitting cold turkey backfires

When you reward a behavior every single time, students learn something extra you didn't intend: they learn exactly what the deal is. The moment the deal disappears, so does the reason to keep the behavior, and it fades fast. Behavior analysts have known this for decades. In a classic study, Kazdin and Polster (1973) reinforced social interaction for two adults in a work setting: one on a continuous schedule, one intermittently. When all rewards stopped, the continuously rewarded participant's behavior dropped almost immediately, while the intermittently rewarded one kept going. Predictability had taught the first participant exactly when effort stopped paying.

None of this argues against reward systems. Dense, reliable reinforcement is exactly what you want when a behavior is new, as we cover in positive reinforcement in the classroom. It argues for treating "reward everything, every time" as a launch setting you plan to outgrow.

What the classroom research shows

Two studies map the path out.

Freeland and Noell (1999) worked with third graders on math accuracy. First, correct work earned immediate, reliable rewards. Then the researchers shifted to delayed, intermittent reinforcement: rewards came later, and only sometimes, so students couldn't tell which work sessions would count. When rewards were removed entirely, the students' performance held up. The unpredictability had done its job. Because any session might be the one that counted, students treated every session like it counted.

Jones and Kazdin (1975) went further and faded an entire token economy in a special education classroom without losing the gains. Their moves are worth copying: they gradually made tokens scarcer, then swapped them for reinforcers that lived naturally in the classroom, like peer recognition and access to preferred activities. When the tokens were gone, the behavior stayed, because by then the behavior was being paid in a currency the classroom prints anyway.

And if you're worried that using rewards at all quietly erodes students' love of learning: Cameron, Banko, and Pierce (2001) meta-analyzed 145 experiments and found rewards are generally not harmful to intrinsic motivation, and verbal praise reliably increases it. The risk they flagged is narrow: tangible prizes promised in advance and handed out regardless of quality. Read it as a caution against sloppy rewards. (For the fuller picture, see intrinsic vs. extrinsic motivation.)

A fading plan you can actually run

1. Start dense, then thin on purpose

Keep reinforcement generous while the routine is new. That part you're probably already doing. The shift is deciding, maybe three or four weeks in, that recognition gets less frequent and less predictable on purpose. Don't announce a downgrade. Just stop rewarding every instance and start rewarding the best ones.

2. Make the schedule unpredictable

This is the step teachers most often skip. Going from "every day" to "every Friday" is thinner, but it's still predictable, and students learn to perform on Fridays. Aim for variable: sometimes twice in a morning, sometimes not for two days. If you track points in a tool like Chou Chou Teach, it's easy to see at a glance whether your recognition has settled into a schedule students could set a watch by.

3. Add delay

Freeland and Noell's other lever was time. Instead of rewarding in the moment, acknowledge at the end of the block, then the end of the day: "I was watching during independent reading, and table three earned their points." The behavior starts to stand on its own between the doing and the noticing.

4. Let praise carry the weight

From the very first week, every point should arrive attached to behavior-specific praise: "you started your warm-up without being asked, and that's a point." Praise is the reinforcer that survives the fade. When tangible rewards thin out, the noticing continues at full strength, and by then the noticing is most of what students were working for.

5. Trade tokens for the classroom's own currency

Follow Jones and Kazdin: shift from prizes toward reinforcers your room produces for free: preferred activities, choice of seat or task, a positive note home, a job with real responsibility. These never run out, and no one has to buy them at the dollar store.

6. If behavior dips, back up a step

A wobble during fading usually means you thinned faster than the behavior could bear. Return to the previous step for a week or two, then try again. Fading is a ramp, and it's fine to pause on the ramp.

The bottom line

A reward system's job is to work itself out of a job. Reward generously while the habit is new, then make recognition less predictable, add delay, and let praise and the classroom's own payoffs take over. Done that way, the question "what happens when the points stop?" has a boring answer, which is the best kind: mostly, nothing. The warm-ups still start on time, the praise keeps flowing, and the points get to quietly retire.

References

• Kazdin, A. E., & Polster, R. (1973). Intermittent token reinforcement and response maintenance in extinction. Behavior Therapy, 4(3), 386–391.
• Jones, R. T., & Kazdin, A. E. (1975). Programming response maintenance after withdrawing token reinforcement. Behavior Therapy, 6(2), 153–164.
• Freeland, J. T., & Noell, G. H. (1999). Maintaining accurate math responses in elementary school students: The effects of delayed intermittent reinforcement and programming common stimuli. Journal of Applied Behavior Analysis, 32(2), 211–215.
• Cameron, J., Banko, K. M., & Pierce, W. D. (2001). Pervasive negative effects of rewards on intrinsic motivation: The myth continues. The Behavior Analyst, 24(1), 1–44.