What 30 years of praise research actually shows
You have probably been told to praise effort, not intelligence, because it builds a growth mindset. The original studies were real. The effect, when tested at scale, turned out to be a fraction of the story — and knowing that makes you better at this, not worse.
What happened to the growth mindset
In the late 1990s a set of experiments found that children praised for intelligence became more fragile in the face of difficulty than children praised for effort. It was a striking result, it was replicated in small samples, and it turned into a global movement — posters, assemblies, teacher training, curriculum.
Then the field did what fields should do and pooled everything. Two meta-analyses in 2018 found that holding a growth mindset correlated with achievement at about r = 0.10 — around 1% of the variance — and that interventions designed to install one produced an average effect of about 0.08 standard deviations.
1% of the variance, drawn to scale
One square in a hundred. The rest is teaching, sleep, home, health, prior knowledge, luck and everything else. This is what a real but small effect looks like when you stop reading it in a headline.
Small, cheap and harmless is a perfectly good deal. It is just not the lever it was sold as.
So what do you actually say?
Here is the useful residue of all this. Praise that labels a child — clever, gifted, a natural — gives them something to protect, and protecting it means avoiding hard things. That mechanism is plausible and worth avoiding regardless of effect sizes.
But “you worked so hard” has its own failure mode: children hear it as consolation when they have not succeeded. The version that survives both problems is description. Say what happened. It is specific, it is true, and it hands them something repeatable.
Gives them an identity to defend. The rational response to a fragile identity is to avoid tests of it.
Well-meant, and sometimes heard as “you didn’t manage it, but never mind”. Not harmful, just oversold.
Describes the move. Repeatable, honest, and it works whether they succeeded or not.
“You went back and checked” or “you tried the smallest one first”. That is a sentence a child can act on tomorrow.
“That approach ran out at step three” is information. “Never mind, you tried” is a full stop.
Praise is manners and information, not an intervention. The heavy lifting is done by what you give them to think about.
Questions parents actually search for
Does growth mindset actually work?
Barely, on average. Mindset beliefs explain roughly 1% of variation in achievement, and interventions average about d = 0.08. A 2023 review of 122 studies argued even that is likely attributable to design and reporting problems — though other researchers dispute this and point to narrow benefits for specific groups.
Should I praise effort or intelligence?
Neither is the best option. Describe the specific thing the child did — the strategy, the check, the retry. Labelling ability gives them something to protect; praising effort alone can read as consolation.
So was Dweck wrong?
The mechanism she described — that beliefs about ability change how children respond to difficulty — is plausible and still studied. What was wrong was the scale of the promise, and the idea that a short intervention could move achievement much.
What matters more than praise?
What you give a child to think about, and whether being wrong is safe. Both of those move learning more than the wording of your encouragement.
One finding a week, in plain words, with its caveat attached
That is the whole newsletter. Two minutes on a Friday, nothing sold, unsubscribe in one click.