People who keep saying ‘nice work’ usually miss the one line that actually builds skill

You just said “good job” again and watched it slide right off. The learner nodded, smiled, and changed nothing. That gap has a cause. Praise feels kind, but it rarely tells anyone what to do differently on the next try. The feedback that actually builds skill is the specific move you need to fix next — the elbow lifted two inches, the second step redone, the wrong word swapped for the right one. Research backs this up, and the size of the difference is bigger than most people guess.

Before the sections open, the figures this page stands on — each one carrying its own source.

Here is the honest question this page answers: does praise really build skill, or does something more specific do the heavy lifting? The short answer is that information beats applause. When feedback names the exact task and the exact fix, people learn more — across thinking skills and physical skills alike.

Does praise actually help someone get better at a skill?

A large 2020 review pulled this apart. Wisniewski, Zierer and Hattie revisited “The Power of Feedback” and combined 435 studies, 994 effect sizes, and about 61,000 people. The overall effect of feedback landed at d = 0.48 — a medium push once outliers were removed. That number sounds tidy, but the studies varied a lot. The reason they varied is the whole point.

What drove the impact was information content. High-information feedback — the elaborated kind, tied to the task — worked far better than plain praise or a bare grade. Telling someone their answer earned a B does little. Telling them which step went wrong, and what to try instead, does a lot. Praise and grades sit at the low-information end. They mark the outcome and stop there.

The effect was also larger for some skills than others. Feedback helped cognitive outcomes most, at d = 0.51. It helped physical and motor skills strongly too. It did least for motivational and behavioral outcomes, at d = 0.33. So if your goal is a sharper essay or a cleaner golf swing, specific feedback is your best lever. If your goal is to make someone feel encouraged, praise may do that — but do not expect it to build the skill itself.

The key stats below show how the numbers stack up, so you can see where feedback earns its keep.

The same page reads differently depending on where you stand. Tell it where you stand.

What does ‘high-information feedback’ look like in a real moment?

Picture a teacher handing back a paragraph. One version reads, “Nice work!” The other reads, “Your topic sentence promises three reasons, but you only give two — add the third or cut the promise.” Both take a few seconds to say. Only the second one tells the writer what to change next.

That is the line between praise and task-tied feedback. High-information feedback answers three quiet questions the learner is holding: Where am I going? How am I doing right now? What is my next step? Praise mostly answers the middle one, and even then only in mood, not in detail. The correction fills in all three.

It does not have to be harsh. A specific fix can be warm — “You are close; the elbow just needs to come up two inches.” Warmth and information can travel in the same sentence. What matters is that the learner walks away knowing the exact move to try, not just that you were pleased.

A page should show what it grows from. Here is the tree — every leaf quoted, every receipt attached.

How guessing first and sleep quietly change what feedback sticks to

Two other lines of research help explain why specific feedback lands where it does — think of them as background that supports the main idea. Both come with a catch worth knowing.

The first is guessing before you learn. A preregistered meta-analysis looked at prequestions — having people take a guess at an answer before they study it. The benefit was moderate for the exact material they guessed about, at g = 0.54 across 97 studies. But for the other, non-guessed material in the same lesson, the benefit was almost nothing, at g = 0.04 across 91 studies. Guessing narrows attention. It sharpens memory for the precise thing you reached for, and leaves the rest roughly where it was.

Notice how that mirrors feedback. A guess focused on one point primes you for correction on that same point. General guessing spreads thin. Targeted attention pays off; vague attention mostly does not. The pattern rewards precision on both ends.

The second line is memory replay during sleep. A meta-analysis of targeted memory reactivation — quietly re-presenting a learning cue while someone sleeps — pooled 91 experiments and 212 effect sizes across 2,004 people. The overall memory benefit was small but real, at Hedges’ g = 0.29. It was strongest during deep, non-REM sleep. During REM sleep or while awake, the effect was absent.

The takeaway for a learner is modest but useful. Sleep helps lock in what you practiced, and specific cues can nudge that along a little. It will not replace the correction itself. First you need the exact move to fix — the deep-sleep replay only firms up what is already there.

Praise vs. specific feedback: which one to reach for

Both have a place, and it helps to see the trade honestly. Praise builds warmth and can steady a nervous learner. Task-tied feedback builds the skill. The comparison here lays out where each one earns its spot, using the effects the research reported.

The practical rule is simple. When someone needs to feel safe enough to keep going, kindness matters. When someone needs to get better, name the next move. Most good coaching does both in the same breath — a word of encouragement, then the precise fix.

You have read enough about minds in general. This one maps yours — drawn live from your answers, with a citation under every claim.

The exact words that turn ‘good job’ into a next move

Vague praise is a habit, and habits change when you swap the actual words. Here are three before-and-after pairs. Say the first out of reflex; say the second on purpose.

Reflex line: “Good job!” Stronger line: “Your intro grabbed me — now cut the last two sentences so the point lands sooner.” The first marks the outcome. The second hands over the exact next edit.

Reflex line: “That was better.” Stronger line: “Your backswing stayed level that time; keep the wrist firm through contact and it will carry.” One notes improvement in general. The other names the single thing to hold onto.

Reflex line: “You’re so smart.” Stronger line: “You caught the sign error in step three — do that same check on step five.” Praise for the person teaches nothing repeatable. Naming the move gives the learner something to run again tomorrow.

Notice the shape. Each stronger line points at one concrete action on the very next try. That is what high-information feedback means in plain speech — it is specific, it is doable, and it respects the learner enough to be honest.

Here is who says these lines, and one thing each can do today. The teacher grading a stack of papers: on the next essay, write one margin note that names the exact sentence to change, instead of a smiley at the top. The coach at practice: pick a single mechanic — the elbow, the wrist, the footwork — and give one cue for it before the next rep. The parent helping with homework: when your child gets a problem right, point to the step that worked and ask them to use it on the next one. Small, specific, and aimed at the next attempt.

The answers to the questions people ask most about this are gathered below, so you can settle the common doubts in one place.

What to remember when you feel the urge to just say ‘nice work’

The pull toward praise is strong because it is easy and it feels warm. The research is steady on what actually moves skill. Across 435 studies, feedback that carries real information — the exact move to fix next — outperforms praise and grades, and it helps thinking and physical skills the most. Guessing first sharpens memory for the precise point you reached for. Sleep firms up what you already practiced. In every case, precision is what pays.

So the next time you catch yourself reaching for “good job,” add the one sentence that names the next move. That single sentence is the difference between a nod and a new skill.

Remember the note you left? It has been waiting for you.

Praise vs. task-tied feedback

• Names the exact next move — No — Yes
• Best for building skill — Weaker — Stronger (d = 0.48 overall)
• Effect on cognitive skills — Low-information end — d = 0.51
• Effect on motivational outcomes — Limited — d = 0.33
• Can also feel warm — Yes — Yes, when worded kindly

And if a nerve got touched just now, the help below is real and free, open at any hour.

Feedback that builds skill: your questions

• Is praise useless, then? — No. Praise can steady a nervous learner and build warmth. It just does little to build the skill itself — for that, name the exact move to fix next. The 2020 review found feedback works best when it carries real information about the task.
• How big is the difference, really? — Across 435 studies and about 61,000 people, feedback landed at a medium overall effect of d = 0.48. It was strongest for cognitive skills at d = 0.51 and weakest for motivational outcomes at d = 0.33. Information content is what drove the gains.
• Does guessing before I learn help? — It helps the exact thing you guessed about. A meta-analysis found a moderate benefit of g = 0.54 for the specifically guessed material, but almost none — g = 0.04 — for the rest of the lesson. Guessing narrows your attention to that one point.
• Can sleep improve what I practiced? — A little. Re-presenting a learning cue during deep, non-REM sleep gave a small memory benefit of Hedges’ g = 0.29. During REM sleep or while awake, the effect was absent. Sleep firms up what you already practiced; it does not replace the correction.

One last move before you go: press this page into a single sentence of your own — the when and the how, decided now.

And to honour the receipts above: here is how this page itself was built, device by device.

If part of your situation reaches past this page, the guides below cover the next step directly.

This is general information about the mind, not therapy or a diagnosis. If things feel hard, please consult a professional. In a crisis, reach a free, confidential crisis hotline right away; findahelpline.com lists one for your country.