Overjustification Effect

An expected reward for something someone already enjoys can quietly replace the reason they wanted to do it.

3 min read

What Is It?

The overjustification effect describes what happens when an external reward is attached to a behavior someone was already doing for internal reasons: enjoyment, curiosity, a sense of purpose. Rather than simply adding to the person’s motivation, the reward can replace the internal reason for the behavior, so that once the reward is removed, the behavior drops below its original level rather than returning to it. Edward Deci’s 1971 experiments were an early demonstration: people paid to solve puzzles spent less time on the same puzzles during a later, unpaid free-choice period than people who were never paid at all. Notably, Deci’s own results already hinted at a boundary condition: controlling rewards like cash undermined later interest, while purely informational positive feedback tended not to. Mark Lepper, David Greene, and Richard Nisbett’s 1973 study made the effect vivid. Preschoolers who already enjoyed drawing were split into three groups: one told in advance they’d receive a “Good Player Award” for drawing, one given the same award as a surprise afterward, and one given nothing. Weeks later, during free play, the children who’d expected the reward spent significantly less time drawing than either of the other two groups. The anticipated reward hadn’t just failed to reinforce the behavior. It had crowded out the enjoyment that was sustaining it. The theoretical account has evolved since. Early explanations leaned on Daryl Bem’s self-perception theory: people infer their own motivations from watching their own behavior and its apparent cause, so when a reward is a visible, sufficient explanation for why someone did something, they conclude “I did it for the reward” rather than “I did it because I wanted to.” Later work, particularly self-determination theory, points to a more specific mechanism: rewards that are experienced as controlling undermine a person’s sense of autonomy over the activity, while rewards experienced as purely informational tend not to. The two accounts overlap more than they compete, but the autonomy framing explains why some rewards crowd out motivation and others don’t.

Why Does It Matter?

Organizations routinely try to reinforce behavior that’s already happening for intrinsic reasons: a naturally helpful colleague, a team that already collaborates well, an engineer who’s genuinely curious about a hard problem. Adding formal recognition, a bonus, or a contest to that behavior can be exactly the wrong move. The risk is strongest when the reward is expected in advance, tangible, and experienced as controlling, especially when it becomes the obvious reason for doing an activity that was already intrinsically interesting. It can convert something self-sustaining into something that only continues under continued reward, often landing lower than where it started once the reward stops or people habituate to it. This is a different mechanism from Goodhart’s Law, even though the two get invoked together. Goodhart’s Law describes a measure getting gamed once it becomes a target, people optimize for the number rather than the underlying goal. The overjustification effect requires no gaming at all. Nobody is cheating the system. The person has genuinely stopped wanting to do the thing, because the reward quietly became the reason they were doing it in the first place. The shift isn’t only internal to the person being rewarded. Formalizing a reward can change not only someone’s own explanation for their behavior, but everyone else’s around them. What once signaled generosity, curiosity, or professional pride can start to look transactional, weakening the social meaning that helped sustain it in the first place.

What Changes Once You See It?

You stop assuming that recognition or incentive layered onto an already-motivated behavior is automatically a free win, and start asking whether the behavior was already intrinsically sustained before deciding to formalize a reward around it. You start distinguishing between behaviors that need external motivation to happen at all, where incentives genuinely help, and behaviors that are already happening and mainly need to not be undermined, where a formal reward structure is the more dangerous tool of the two. You also start watching for a specific pattern: a formal recognition program followed by a drop in the very behavior it was meant to celebrate, once the program ends or becomes routine. That’s the overjustification signature, distinct from a metric quietly being gamed.

Common Misunderstandings

It is not a claim that extrinsic rewards are generally harmful, or that organizations shouldn’t pay people. Most work isn’t purely intrinsically motivated to begin with. The effect specifically concerns undermining pre-existing intrinsic interest in a particular task, not compensation broadly.
It doesn’t mean recognition and praise are always risky. The effect is strongest for rewards that are expected in advance, tangible, and experienced as controlling. It’s weaker or absent for unexpected recognition, purely informational feedback, and tasks that had little intrinsic interest to begin with.
It is not the same mechanism as Goodhart’s Law, even though the two are often mentioned in the same breath. Goodhart’s Law is about a measure being gamed once it’s a target. The overjustification effect is about the reward genuinely replacing the internal reason for the behavior, with no gaming required.
It isn’t an iron law that guarantees crowding-out whenever a reward appears. The research itself is contested: some studies find substantial undermining for expected tangible rewards, while other meta-analyses find the effect is more limited and depends heavily on reward type, contingency, and how controlling the reward feels. Treat it as a real, well-documented risk to watch for, not a certainty in every case.

Diagnostic Question

Was this behavior already happening for its own sake, and if so, does formalizing a reward around it risk becoming the only reason it keeps happening?

Explore Further

Field Notes

None yet.

Related Field Guide

Origin

Edward Deci’s 1971 experiments on rewarded puzzle-solving first documented the effect. Mark Lepper, David Greene, and Richard Nisbett’s 1973 study using a drawing task with preschoolers gave the effect its clearest and most widely cited demonstration.