Updated on
September 2, 2026
What intrinsic motivation to learn means, how researchers measure it, and what studies show about learning outcomes, rewards and the limits of inference.
What is intrinsic motivation to learn?
Intrinsic motivation to learn means taking part in learning because the activity itself is interesting or satisfying, rather than only for a separable outcome such as a grade, reward or approval. It is distinct from identified regulation, where a learner values an outcome without necessarily enjoying the activity itself.
Intrinsic motivation to learn means learning because the activity itself is interesting or satisfying. The main reason is not a separate outcome, such as a grade, reward or approval. This differs from identified regulation. There, a learner values the outcome but may not enjoy the activity.
The research does not split motivation into simply “good” and “bad” forms. Studies use different measures and populations. They also study different outcomes.
This article keeps theory, findings and possible school use apart. It sits within the wider theories of motivation and fundamental theories of learning.
Intrinsic motivation means doing an activity for its own interest or satisfaction (Ryan & Deci, 2000). The key question is why the learner takes part. A learner may read about volcanoes because the topic is absorbing. That reason is intrinsic.
The same learner may read because the topic matters for a future qualification. This reason can be personal and strongly held. Yet it is separate from enjoyment of the activity. It is therefore identified regulation.
These cases can look the same from the outside. In both, the learner reads. The reason for the act is what sets them apart.
Intrinsic motivation is not a fixed learner type. Interest can change by topic, setting and time. A learner may also have several reasons for one action. Enjoyment, value, grades and approval can all play a part.
Behaviour alone cannot reveal a motive. Optional work may show interest. It may also reflect personal value, anxiety or an expected test.
Questions and persistence can add context. They do not prove why a learner acts.
Self-determination theory separates several forms of motivation. External regulation involves rewards, punishments or demands. Introjected regulation involves inner pressure, such as guilt. Identified regulation means the learner accepts the value of an activity or outcome.
Identified regulation can be quite autonomous. It remains extrinsic because the reason lies in what the activity leads to (Ryan & Deci, 2000b).
The wider self-determination theory proposes three basic psychological needs. These are autonomy, competence and relatedness (Ryan & Deci, 2000b, 2020). The theory looks at how social settings support or frustrate those needs.
It is not a diagnostic checklist. Need satisfaction, motivation, attainment and wellbeing remain distinct ideas.
Intrinsic Motivation: The activity itself is interesting or satisfying.
Identified Regulation: The outcome matters personally, even if the activity is not enjoyable.
External Regulation: An external demand, reward or consequence is the main reason for acting.
Different reasons can coexist. Behaviour alone does not reveal motive.
Researchers cannot see intrinsic motivation directly. They use several measures. Each one answers a different question.
Scroll horizontally to compare the measures.
| Measure | What it records | Main limit |
|---|---|---|
| Self-reported interest or enjoyment | What a person says about the task | It cannot predict later action or performance by itself. |
| Free-choice persistence | Time spent after a task or reward ends | It cannot reveal the sole reason for staying. |
| Observed task behaviour | Choices, return to a task or persistence | It cannot rule out other motives or limits. |
This helps explain why studies may seem to disagree. A reward may change later free-choice behaviour but not reported interest. A study of one puzzle also has a narrow reach. It cannot show how learners will respond to a full term of required work.
No one measure is a full test. A score can support a claim, but it cannot settle it on its own. Using more than one measure can add context, while none establishes motive on its own.
Large reviews usually find positive links with learning or performance. However, intrinsic motivation is not the only route to learning. Group averages also cannot predict one learner's result. The size of each link depends on what a review measures.
Scroll horizontally to compare the evidence.
| Source | Sample | Finding | Limit |
|---|---|---|---|
| Cerasoli, Nicklin and Ford (2014) | 183 studies and 212,468 people across school, work and physical tasks | Corrected links with performance ranged from ρ=.21 to .45. Intrinsic motivation had a stronger link with quality. Incentives had a stronger link with quantity. | The review mixed domains. It was mainly predictive, not a test of one school intervention. |
| Howard et al. (2021) | 344 samples, 223,209 learners and 26 outcomes | Intrinsic motivation linked with success and wellbeing. Identified regulation linked especially with persistence. For objective performance, average links were ρ=.13, .11 and -.03 for intrinsic, identified and external regulation. | These links do not prove cause. They also do not reveal one learner's motive. |
| Bureau et al. (2022) | 144 studies and more than 79,000 learners, from primary school to university | All three needs linked with self-determined motivation. Their links were not equal. Competence and autonomy had more weight than relatedness in the intrinsic model. | No need must be fully met first. Behaviour cannot diagnose an unmet need. |
| Wang et al. (2024) | A review of 36 interventions with 11,792 people. The meta-analysis covered 31 studies, 137 effects and 9,433 people. | The mean effect on intrinsic motivation was g=.58. Effects were 1.14 for autonomy and .48 for competence. Relatedness was .44 and not statistically significant. | The programmes varied. The result does not prove one method works in every class. |
The evidence supports a careful conclusion. Intrinsic motivation is linked with learning and performance. The link may be stronger for the quality of work.
Autonomous forms of extrinsic motivation also matter. Results vary by measure, outcome and setting. Much of the evidence shows association, not cause.
The evidence does not support a simple divide. Intrinsic motivation does not always lead to deep, creative or happy learning. Extrinsic motivation does not always lead to shallow or brief effort.
Identified regulation has useful links of its own. Learners often act for several reasons at once.
These reviews ask broad questions. They pool many groups and tasks. This can show a trend, but it can hide wide gaps between studies.
People often claim that rewards destroy intrinsic motivation. The evidence is narrower and disputed. Starting interest, reward type, timing, age and the outcome measure all matter. No one result covers all these conditions.
The overjustification effect article covers this dispute in more depth.
Deci (1971) ran two lab studies with 24 undergraduates in each. He also ran a field study with eight people. Each study had three periods. Rewards began in the second period and later stopped.
Money tended to reduce later free-choice puzzle activity. Positive feedback tended to increase it. These small, early studies do not establish a general rule.
Lepper, Greene and Nisbett (1973) studied 51 preschool children. They were 40 to 64 months old. Each already showed high interest in drawing.
The children received an expected award, an unexpected award or no award. One to two weeks later, the expected-award group chose less drawing time. It also fell from its own starting level.
The study does not show that certificates reduce all children's interest. Its sample was narrow, mostly white and middle class. The researchers also excluded three Black children from the analysed group. This limits how far the result can reach.
Deci, Koestner and Ryan (1999) reviewed 128 experiments. They kept free-choice behaviour separate from reported interest. For interesting tasks, expected tangible rewards reduced later free-choice engagement. The effects were about d=-0.40 for engagement, -0.36 for completion and -0.28 for performance.
Effects on reported interest were smaller. They were about d=-0.15 for engagement and -0.17 for completion. Positive feedback raised free-choice engagement by about d=0.33. It raised reported interest by about d=0.31.
Cameron, Banko and Pierce (2001) reviewed more than 100 experiments. Their results were more favourable in some settings. Rewards could raise engagement in low-interest tasks.
Verbal rewards were positive for high-interest tasks. Rewards tied closely to a performance standard could increase engagement or make no difference.
Deci, Koestner and Ryan (2001) challenged that review. They disputed its study choices, labels, control groups and pooled results. The dispute concerns methods as well as theory. Neither review supports one rule for all rewards.
A reward is not one fixed thing. Cash, praise and a prize can send quite different signals. A reward for taking part also differs from one for a set standard.
The best conclusion is conditional. Expected tangible rewards can reduce later free-choice engagement for an already interesting task. The effect is not uniform across ages, rewards or measures.
Positive feedback can raise interest or engagement. Incentives may also support the amount of work done. No checked study shows that a change in merit-point wording will reliably change a learner's motive.
This evidence can help schools judge claims about motivation, but it cannot sort learners into fixed types. Optional work, curiosity and persistence are not diagnostic signs. Such acts can have more than one cause.
They may show interest. They may instead show value, anxiety, confidence, prior knowledge or outside pressure.
School staff cannot diagnose a psychological need from one observation. Bureau et al. (2022) report group-level links, so their findings describe broad patterns rather than a test for one child.
Wang et al. (2024) combine varied programmes. These findings can guide careful thought about learning settings. They do not guarantee a result for one learner, class or SEND plan.
For staff, the safest use is to test a claim, not label a child. Ask what changed in the task. Ask what the learner says. Then check what else may explain the act.
Three boundaries follow from the evidence:
Important gaps remain. Trials have not isolated every common school method. Short tasks may not represent motivation across a term.
Culture and age need direct study, not stereotypes. Claims about poverty, ADHD, neurodivergence or generative AI need separate evidence. They do not follow from the studies reviewed here.
For related context, see motivation in education, mindset, motivation and resilience, and strategies that promote learner wellbeing. These pages cover their own concepts and evidence. Their claims are not imported into this review.
Flow is also a different idea. It describes deep absorption in an activity. It is not a general reason for starting that activity. See the separate flow-state account for that distinction.
