The experiment is half a century old and still the cleanest way into the argument. Deci paid one group of subjects to solve an interesting puzzle and left another group unpaid; afterwards, in free time, the paid group played with the puzzle less (Deci, 1971). Payment had converted play into work, and when the payment stopped, so did the interest. Three decades later, a meta-analysis of 128 experiments confirmed the pattern: tangible rewards contingent on doing, finishing, or excelling at a task reliably undermined intrinsic motivation, and the effect was strongest in children (Deci, Koestner and Ryan, 1999). Verbal feedback, information rather than payment, tended to strengthen interest instead.
Grades are tangible rewards contingent on performance, administered to children, attached to tasks schools desperately want children to find interesting. The literature does not say grades communicate nothing; it says they communicate at a price, and the price is the student's relationship to the work itself. The A-student learns to ask "is this on the test?", which is the puzzle-player asking "am I still being paid?". Both questions are rational. Both are the sound of intrinsic motivation being crowded out.
What replaces the letter
The standard defense of grades is that there is no alternative, that without letters, nobody knows anything. This is empirically obsolete. The alternatives are not thought experiments; they are operating systems in daily use.
Mastery tracking replaces the averaged letter with a map: which specific competences has this child demonstrated, at what level, with what evidence. The Lumiar schools founded by Ricardo Semler run on this architecture, with advancement tied to demonstrated competence rather than age, and a digital record following each child's individual trajectory (Revista Educação, 2020). As Semler has told it, two decades of these schools, roughly 3,000 children across six countries, ran without grades for memorization at all: the assessment is whether the project shipped and what the child can now do.
Portfolios replace the transcript's compression with the work itself: the essays, the models, the recorded performances, curated over time. And observation-based reporting replaces the inspection ritual with continuous documentation by adults who know the child. What is striking is how mainstream this has quietly become at the small end of the market: in the National Microschooling Center's 2025 survey of 800 American microschools, only 29 percent used traditional letter grades; the rest ran on observation, portfolios, and mastery tracking (Soifer and Soifer, 2025). Hundreds of independent operators, mostly unaware of one another, converged on the same replacement.
The design move underneath
All the working alternatives share one structural feature: they collapse the distance between the assessment and the thing assessed. A grade is a proxy of a proxy, a letter summarizing a score summarizing a test approximating a competence. Each layer of compression discards information and invites gaming. A portfolio or a mastery record is closer to the object itself: here is what was made, here is what was demonstrated.
Semler ran the identical substitution on adults. As he has told it, after his collapse at 25 he stopped counting his workers' hours, the proxy, and started measuring whether the pump was delivered, the object; absenteeism fell. Schools face the same choice with the same asymmetry: proxies are cheaper to administer, objects are truer. The letter grade survives for the same reason the time clock did, administrative convenience at scale, and it distorts for the same reason: people optimize the measure, not the meaning.
What the skeptic is owed
Honesty requires the caveats. The undermining effect is a lab-anchored finding, and its size in messy classrooms is debated; a diploma system needs some interface with universities and employers, which is why serious gradeless schools keep rigorous records precisely so they can generate a transcript on demand; and gradeless assessment done lazily is worse than grades, because it removes the signal without adding the documentation. The alternative to the letter is not absence. It is more assessment, closer to the work, legible to the child.
But the burden of proof sits where it rarely gets placed. One side of this argument has a replicated experimental literature saying its instrument corrodes the motivation it is supposed to serve. The other side has habit and administrative convenience. In any other field, that would not be a close call.
Sources
- Deci, E. L. (1971). Effects of Externally Mediated Rewards on Intrinsic Motivation. Journal of Personality and Social Psychology 18, 105-115. Paid subjects showed reduced voluntary engagement with an interesting task afterwards. https://doi.org/10.1037/h0030644
- Deci, E. L., Koestner, R. & Ryan, R. M. (1999). A meta-analytic review of experiments examining the effects of extrinsic rewards on intrinsic motivation. Psychological Bulletin 125(6), 627-668. 128 experiments; tangible performance-contingent rewards undermine intrinsic motivation, strongest in children; verbal feedback tends to enhance it. https://doi.org/10.1037/0033-2909.125.6.627
- National Microschooling Center / Soifer, D. & Soifer, A. (2025). American Microschools: A Sector Analysis. Only 29% of 800 surveyed microschools use letter grades; the rest use observation, portfolios, mastery tracking. https://microschoolingcenter.org/news-blog/2025-american-microschools-sector-analysis
- Revista Educação (2020). Criador da escola Lumiar explica seu projeto (interview with Ricardo Semler). Competence-based advancement and individual digital learning records in Lumiar schools. https://revistaeducacao.com.br/2020/01/09/lumiar-entrevista/