Evaluate without discouraging

The scene is repeated every year, the first week of December, in almost every school in the country. A meeting room, twelve teachers bearing the exhaustion of the first term, a screen with the list of 2nd B. Two hours for thirty students: that comes out to four minutes per head. "Lucía: good in everything, don't let her get complacent." "Hugo: three failed subjects, let's see if he wakes up." The meeting progresses, the minutes are filled in, and on Friday each family opens the report card. At Lucía's house no one mentions it, because Lucía gets B's and B's are not discussed. At Hugo's house, the report card falls onto the kitchen table with the weight of a medical report. And amid all the hustle and bustle, almost no one asks the question that really matters that week: what is Hugo going to do differently on Monday?

That question is the heart of what educational literature calls formative assessment, an expression with bad luck. It sounds like workshop jargon, has ended up naming things it is not, and has been dragged into the debate over standards, where each side uses it as ammunition. It deserves to be rescued, because beneath the label lies an ancient and very sensible distinction: assessment has two jobs. One is to measure what has been learned and certify it—grades, report cards, transcripts. The other is to help learning: telling each student where they are and what their next step is. Both are legitimate and both are needed. What happens is that the first has a uniform, a calendar, and a box on the platform, while the second lives on borrowed time. In this notebook we wrote it regarding PISA and it is just as valid inside the classroom: a good thermometer is a blessing; a judge who only sentences is a plague. The December report card can be both, and the difference is not in the document but in what the student can do with it.

To see what is at stake there, it is worth traveling to Jerusalem, 1988. Ruth Butler, from the Hebrew University, proposed a series of tasks to 132 fifth and sixth-grade students spread across twelve classrooms and returned the corrected exercises in three ways. One-third of the kids received only a comment: a specific sentence about their work, with an indication for improvement. Another third received only a grade. The last third received both a grade and a comment together, which is what any of us would call the perfect correction. The results have been making readers uncomfortable for almost forty years. The group with comments only improved in the following sessions and maintained interest, both the bright students and the struggling ones. With the grade alone, performance remained lower and the interest of the struggling students plummeted. And the surprise lay in the third group: the grade with a comment did not work like the comment, but like the grade. On the main task, the high-achieving students who received both ended up scoring worse (11.8) than those with the grade alone (16.5) and a world away from those with comments only (24.3). Butler's explanation carries no cruelty: the kids remembered the grade and forgot the sentence. The teacher who spends Sunday writing meticulous corrections alongside a 4.75 does not fail because of bad writing; they write on a paper that the number has rendered illegible.

Ten years later, two researchers from King's College London put large numbers to that small intuition. Paul Black and Dylan Wiliam reviewed some 250 papers published in more than 160 journals and summarized them in a title that became highly successful: Inside the Black Box. Educational policies, they argued, treat the classroom as an opaque box from which results are demanded without looking at what happens inside, and what happens inside—how questions are asked, how corrections are made, what the student does with the correction—drives learning more than almost any reform. Their numbers were dazzling: typical effect sizes between 0.4 and 0.7, among the highest documented for an educational intervention; with a 0.7, they calculated, a country in the middle of the international rankings would have climbed to the leading group. And a finding that should be on the door of every staff room: the improvement of formative assessment helps low-achieving students the most, thereby raising the average by narrowing the gap. Their rule of thumb fits in two lines: the comment must speak to the qualities of the specific work and how to improve it, and must avoid comparison with other students. What is needed, they wrote, is a culture of success sustained by the belief that everyone can achieve it.

And now, the part that enthusiasm usually skips. Those numbers aged. Subsequent meta-analyses with stricter criteria brought the average effect size much lower, to around 0.20—respectable, but no longer miraculous. The British EEF, which maintains the most sober summary of evidence in the profession, currently ranks feedback among the most cost-effective levers that exist—equivalent to about six months of additional progress for a very low cost—and in the same paragraph warns that the range is wide and that there are studies where feedback makes things worse. Wiliam himself has spent a large part of his subsequent career regretting what his creation became: a label so elastic that it ended up naming anything, from a well-crafted Socratic question to a twenty-row rubric that no one reads. Let's put it bluntly: an assessment reform that translates into more boxes to check is dead on arrival. Spanish teachers already report administrative tasks as their main source of stress—according to TALIS 2024, and as we have reported here—and the feedback that works is not the one that is most documented, but the one that the student actually uses.

There remains the fundamental objection, which we take especially seriously in this notebook because it is voiced by a respected figure close to home. Gregorio Luri has spent years warning against the great sin of modern pedagogy: "pursuing emotional well-being by lowering academic standards." Is this assessing without discouraging—lowering the bar so that no one suffers? It is exactly the opposite, and Luri himself gives the key with his distinction between demand and harshness: demanding is taking the student's capacity seriously. Look again at Butler's study. What was discouraging was not the difficulty of the task, but the format of the verdict. And what sustained interest was not some emotional coddling: it was a sentence that called for work. A precise comment is more demanding than a grade, for the one who writes it and the one who receives it. The grade allows one to settle in—I am a C student, I am a failing student—and remain the same. The comment challenges: this is achieved, this is not, and this is what you do on Monday. Along the same lines is the frequent, low-stakes quiz that Luri champions as training and retrieval practice supports: more assessment, not less—but assessment that teaches. Discouraging is not an excess of standards. It is their cheap substitute.

And what can a teaching staff do from now until December, without new platforms and without spending a single euro? Three actions are enough to start. Comment on the task and the next step, not on the person: "this paragraph loses the thread here; rewrite it linking the second idea with the first" works; "you must try harder" is merely decorative. Separate the comment from the grade when regulations permit: if the number erases the sentence, let them at least not arrive together—first the sentence and the work on it, then the grade. And in the evaluation meeting, spend at least one sentence on each student regarding what they need to learn, not just the count of what they have failed: four minutes per head is not much, but it is enough for that. None of the three actions adds bureaucracy; all three ask for something more difficult, which is to reach an agreement. That is why assessment is, at its heart, a matter of core values. A school that proclaims in its educational project that every student is a promise and then only speaks to them in numbers contradicts itself in the very document that families read most. A teaching staff's way of correcting is their view of the human being put in writing, thirty times per term.

Hugo's report card will say in December what it has to say, and it is right that it does: an honest thermometer is also a form of respect. The difference between one school and another will not be in that number, but in whether anyone, before the holidays, has told Hugo what to do on Monday—and in whether Hugo, upon opening the report card, remembers the sentence or only the grade.

Cover image generated with artificial intelligence.