Does Gamification Work for Learning? Yes and No

Does gamification work for learning? The evidence

Start with the best available number. In 2020, Sailer and Homner published a meta-analysis of gamified learning in Educational Psychology Review. Pooling experimental studies, they found significant but modest effects: g = 0.49 for cognitive outcomes, g = 0.36 for motivational, g = 0.25 for behavioral. One detail matters more than the headline figures. When the analysis was re-run on only the methodologically rigorous studies, the cognitive effect held. The motivational and behavioral effects wobbled.

Read that honestly and two things are true at once. Gamification is not snake oil: points, badges and quests produce a measurable average gain. And gamification is not a method. The effects vary widely between studies, and the averages get dragged around by what the game elements were attached to. Sailer and Homner’s own moderator analyses point the same way — outcomes shifted with the specific elements and combinations used, and bare reward-and-status designs underperformed designs built around challenge, fiction and social play.

So “does gamification work for learning?” is two questions wearing one coat. Does it make you show up more? Mostly yes. That is the motivational slice, and every streak app on your phone has industrialized it. Does it tell you what you learned? Not by itself. Points are a pointer. Their value is decided by exactly one thing: the event that fires them.

Later syntheses repeat the pattern — positive average effects, high heterogeneity, everything sensitive to design. Nobody serious reads this literature and concludes gamification is a myth. Nobody serious concludes it is a method, either. It is an amplifier, and an amplifier has no opinion about what it amplifies.

The part that works: gamification as a habit engine

Showing up is a real bottleneck. Most abandoned skills die of absence, not of bad pedagogy — the practice simply stops happening. Game mechanics attack exactly that failure mode. A streak makes tomorrow’s absence feel like a loss. A quest turns a vague intention into a finishable unit. XP makes invisible effort visible on the day you put it in. As habit machinery, this is legitimate engineering, and the behavioral effect in the meta-analysis is the sound of it working.

And the habit layer matters even by our own strict standard. Retrieval has to be spaced to work, and spacing is just attendance repeated on a schedule. You cannot recall anything on a day you never showed up. So the mechanics that get you in the door are doing real pedagogical work — right up until the moment someone reads the attendance record as an ability record.

It works well enough to overshoot. In a 2023 Journal of Consumer Research study, Silverman and Barasch found that once people log a streak, maintaining it becomes a goal in its own right — separate from the activity it was supposed to track. The record stops describing the practice and starts steering it. That is a strange property for a metric. Your thermometer does not give you a fever.

None of this makes streaks evil. It makes them a habit instrument that keeps getting sold as a progress instrument. We wrote a full teardown of the streak-as-metric problem, and its one-line version has survived every rewrite since: “Streaks measure habit. We measure learning.” Keep the chain if it gets you in the door. Refuse to read it as a report card.

If you want the habit engine without the fake report card, that split is exactly what Plan2Skill is built around — XP that fires on retrieval, not on attendance. Create a free account, or keep reading and run the checklist below against whatever app you already use.

When gamification stops working for learning

Call the failure mode XP-as-theater: points attached to attendance events. You opened the app — points. You watched the video — points. You re-reviewed a card you have known for a year — points, plus a pleasant sound. Every one of those events can occur with zero new encoding, so the score climbs whether or not anything was learned. The number is not lying, exactly. It is answering a different question: how often do you show up and touch easy things?

For a streak-app refugee, this is a familiar biography. Two hundred days, a glowing chain, a level with a crown on it — then a real conversation, a real bug, a real exam, and nothing comes out without the app’s scaffolding. The theater did not fail to measure learning by accident. It was never wired to the retrieval events where learning would have shown up. If that describes your last app, the finished-the-tree playbook covers what to do with the habit you did build — because the habit, at least, is real.

The theater version is worse than neutral, and the reason is calibration. A rising score is evidence, and your brain files it as evidence about skill. When the score actually tracks attendance, every session quietly updates your confidence with someone else’s engagement metric. That is how a person holds a year-long record and a sincere belief they are advanced: the instrument was pointed at the wrong quantity, and it never once read low. Noise you know is noise is harmless. Noise dressed as signal compounds.

XP-as-map: what points look like when they measure

Full disclosure: Plan2Skill hands out XP, quests and streaks too. We are not against game mechanics — we are against unmeasured theater. The difference is a wiring decision: which event fires the points. Five checks separate a map from a stage set, in our app or anyone else’s:

• Points fire on retrieval. XP for producing the thing cold — from memory, under a little difficulty — is attached to a learning event. XP for logging in is attached to a calendar event. Same mechanic, different meaning.
• Misses are recorded, not hidden. A failed recall is the most informative event in the whole system: it locates the gap. Theater buries misses because they hurt engagement. A map marks them as the next destination — that is desirable difficulty doing its job.
• The score can go down. Memory decays, so an honest metric decays with it. A number that only rises is counting accumulated sessions, not the current state of the skill.
• The number translates into a sentence. “What can I produce without help, today?” If the score cannot be converted into an answer to that question, it is decoration.
• The removal test. Delete the points and look at what remains. A map still describes which connections held under testing, and when each one was last checked. A stage set leaves attendance logs.

Wired this way, the same mechanics stop being a bribe and become an instrument. A quest is a scoped, finishable claim about ability. XP is a coordinate: where on your skill graph retrieval succeeded, and how recently. The streak still exists — it just goes back to its honest job of keeping the habit alive while the edges do the measuring. That is the version of gamification the evidence supports, and the research behind our method is public.

That also gives a streak-app refugee a clean decision rule. You do not need to swear off game mechanics; the habit they built is an asset. You need a second instrument. Either your app records retrieval events, or you keep a cold log next to it and let the app handle attendance. One instrument per quantity — that is the whole reform.

FAQ

Is gamification good or bad for learning?

Neither, as a category. The Sailer and Homner (2020) meta-analysis found small positive average effects, and the cognitive effect survived the rigorous-study subset. The value is decided by wiring: mechanics attached to retrieval and challenge help, while mechanics attached to attendance produce engagement without measurement. Judge the event that triggers the reward, not the presence of rewards.

Do XP and points improve learning outcomes?

Indirectly — they increase time on task and consistency, which is the motivational and behavioral effect in the research. Points add no encoding by themselves. If the point-earning activity is retrieval under difficulty, more of it means more learning. If it is easy review, more of it just means a higher score.

Should I quit apps that use streaks and XP?

Not for using mechanics — quit if the mechanics are the only measure the app has. Keep whatever keeps you practicing daily. Then add one external check the app cannot fake: once a week, produce the skill cold, without hints and without the app. If the score rises while the cold test stalls, the score is theater.

Keep the XP. Change what it counts

Plan2Skill attaches every point to a retrieval event — quests, XP and a skill map that can tell habit apart from learning.

See your map →