Why Praise Builds Dependency in Children and How to Use Feedback for Real Confidence

Carol Dweck's landmark studies at Stanford, published across several papers between 1998 and 2007, found that children praised for being "smart" after a task performed worse on harder tasks than children praised for their effort. The smart-praised group avoided difficult problems to protect their reputation. The effort-praised group sought them out. One word, smart, was enough to shift a child from curiosity to self-protection.
This is the dependency Dweck's research keeps returning to: praise tied to identity ("you're so talented") teaches a child that their worth is a fixed quantity, one that a single failure can reduce. The child stops taking risks. They need the next compliment to confirm the last one still holds.
Chanakya, writing in the Arthashastra, warned that a king surrounded by flatterers was more isolated than one with no counsel at all, because flattery produces the illusion of information while delivering none. The principle scales down perfectly. A child raised on identity-praise is similarly isolated from accurate feedback about their own capabilities.Dependency from praise doesn't announce itself as insecurity. It arrives looking like high performance. The child who always finishes first, always raises their hand, always needs to be the best in the room, and then shuts down completely when they aren't, is showing the signature. The praise worked as a motivator. It failed as a foundation.
In Indian classrooms and homes, this pattern has a specific texture. The child who scores 95 and is told "you're a genius" will often refuse to attempt a subject where 95 is not guaranteed. The motivation is not learning. The motivation is the next round of validation. Remove the audience and the effort drops with it.Dweck's research draws a clean line between two types: person praise ("you're brilliant") and process praise ("you worked through that carefully"). Process praise rewards the behaviour that actually produces results, sustained attention, willingness to revise, trying a different approach when the first one fails. Person praise rewards the outcome, which the child often cannot control and cannot repeat on command.

The practical shift is smaller than it sounds. "You got it right" becomes "you kept going even when it got hard." "You're so creative" becomes "that solution you found in the second attempt was better than the first, what changed?" The child now has information they can use. They know what to repeat.
Chanakya's Arthashastra is direct on the value of honest counsel over comfortable agreement: a minister who tells the king only what pleases him is an enemy in disguise. The same logic applies to feedback given to children. Comfortable praise that doesn't connect to anything real is not kindness, it is noise dressed as support.None of this means praise is harmful. The research finding is not that praise damages children, it is that undifferentiated praise does. Specific, earned, process-linked feedback builds genuine confidence because it gives the child a model of their own competence they can test and verify.

Validation is appropriate when something genuinely difficult was attempted, when a child corrected their own mistake without being told to, or when persistence produced a result that earlier effort hadn't. In those moments, naming what happened precisely, "you rewrote that paragraph three times and the third one is much clearer", is not flattery. It is accurate feedback that happens to feel good.
The distinction matters for parenting and for teachers working with children across different learning environments. A child who receives accurate positive feedback builds an internal model of what competence feels like. A child who receives indiscriminate praise builds a model of what approval feels like. These are not the same thing, and they produce different adults.Real confidence, the kind Dweck's research consistently associates with a growth mindset, is characterised by one specific behaviour: the willingness to attempt something where failure is possible. Children with this quality will try a harder math problem, submit a story they're not sure about, pick up an instrument and sound bad at first. They do this not because they expect to succeed but because they've been taught that the attempt has value independent of the result.

This is the goal of process praise: not to make a child feel good, but to make the attempt feel worthwhile. The child who has internalised this does not need an audience. They work when nobody is watching. The dependency breaks not because the child stopped wanting approval, but because the approval is no longer the point.
The two patterns, identity-praise producing dependency, process-praise producing confidence, are not opposites on a spectrum. They are different feedback loops running in parallel, and the one you feed more often is the one that grows. Dweck's data and Chanakya's counsel arrive at the same place from different centuries: what you reward shapes what the person believes they are. Reward the effort, and the child learns they are someone who can try. Reward the result, and they learn they are only as good as the last one.