Extrinsic rewards can crowd out intrinsic motivation

What it is

The overjustification effect — the undermining or crowding-out effect — is the finding that offering expected tangible rewards for an activity people already enjoy reduces their willingness to do it once the rewards stop, often below the level of people never rewarded. The reward does not add to motivation; it rewrites its source, from "I like this" to "I do this for the payoff," and when the payoff ends, the behavior goes with it.

Why it happens

The mechanism is a rewrite of self-attribution. People explain their own behavior with the most salient reason available: once "this earns points/money/gifts" is on the table, "I simply like it" retreats, and the perceived locus of causality shifts from internal to external. The richer original reasons — curiosity, felt mastery — get covered by a single thin one; the behavior is over-justified. Payment also carries a market signal: if someone pays me to do this, the activity itself is probably not worth much. Which rewards trigger the rewrite depends on how controlling they are perceived to be — a finer distinction treated separately.

Studying it

• Paradigm: participants with measured interest are randomized to expected-reward / no-reward conditions, engage in the activity, then a free-choice period (reward settled, session "over") measures whether they continue.
• Landmark findings: the children's drawing experiments (expected-award groups subsequently drew less on their own) and adult puzzle experiments established the effect; meta-analyses find the undermining is most robust for expected tangible rewards, positive for verbal praise, and moderate in size.
• Methodological cautions: pre-measure and match task interest — "undermining" means nothing for tasks nobody likes; meta-analytic conclusions are sensitive to how rewards are categorized (expected vs. unexpected, engagement- vs. performance-contingent), so check the coding scheme before comparing; the alternatives available during free choice amplify or shrink the effect.

Where it stops holding

The effect concentrates in tangible, expected, behavior-contingent rewards. Unexpected gifts, rewards decoupled from the specific act, and verbal appreciation are not in the undermining class. It does not occur for tasks that were never interesting. The effect size is moderate and the literature contains genuine disputes about its breadth — it is not a law that any reward instantly destroys interest. In live products, rewards arrive mixed with social meaning and identity, so pure crowding-out is hard to isolate.

Applying it

• For behaviors users already show interest in (creating, practicing, contributing content), do not stack "do X, get Y" settlement ledgers; reinforce with competence information ("you cracked the three hardest problem categories this week") rather than redeemable currency.
• If rewards are commercially necessary, make them after the fact, unexpected, and not tied to individual acts — framed as thanks, not wages.
• Put "is this behavior already interesting to this user?" on the incentive-design review checklist, and baseline no-incentive retention before introducing settlement.
• To validate: turn a reward off for a holdout group and compare core-behavior retention; if the behavior collapses with the reward despite a prior organic baseline, crowding-out has occurred.

Related

• Same group: P2.01.1 Intrinsic motivation comes from the activity itself · P2.01.3 Behavior often drops after rewards are withdrawn · P2.01.4 Autonomy, competence, and relatedness are the three sources of intrinsic motivation · P2.01.5 Informational feedback does not crowd out; controlling feedback does · P2.01.6 Tasks with no intrinsic interest have no crowding-out problem · P2.01.7 Unexpected after-the-fact rewards harm less than promised ones
• Nearby: P2.02 Gamification elements · P2.10 Gamification misuse
• Search terms: overjustification effect · undermining effect · tangible rewards