“You’re so smart” sounds like confidence in sentence form. A child succeeds, an adult names the quality that seems to explain it, and everyone leaves the moment feeling good. The problem may not appear until the same child meets work that does not come easily.
A line of research beginning in the 1990s suggests that ability praise can make success feel like proof of intelligence and struggle feel like evidence of its absence. When being smart becomes the prized identity, a hard problem carries a second risk. The child is not only trying to solve it; they may feel they are being tested on who they are.
Process praise moves attention toward something more workable: the strategy used, the revision made, the care shown, the decision to persist or the willingness to ask for help. That does not mean one “you’re so smart” damages a child, nor that chanting “great effort” produces resilience. The evidence concerns the meaning repeated feedback can convey, not a forbidden phrase or a parenting trick.
The Artful Age
A weekly letter on aging well, family across generations, and the creative life after the kids leave home.
The original experiments changed one sentence after success
Claudia Mueller and Carol Dweck reported six studies in a 1998 Journal of Personality and Social Psychology paper. Fifth graders first completed a set of problems and received a single sentence of praise after doing well. Some were praised for intelligence, some for effort, and some received praise focused on the outcome.
The differences emerged in what children wanted next and how they responded when the work became difficult. Children praised for intelligence cared more about performance goals, meaning the chance to demonstrate ability, than learning goals, meaning the chance to develop it. They were more likely to choose tasks that protected the appearance of being smart rather than harder tasks from which they might learn.
After a deliberately difficult set of problems, the intelligence-praised children showed less task persistence and enjoyment, made more low-ability explanations and performed worse than children praised for effort. They were also more likely to describe intelligence as fixed. In one part of the research, more children in the intelligence-praise group misreported their scores when writing anonymously to unfamiliar peers.
These were controlled, short-term experiments, not years-long observations of families. They show that a small change in feedback can alter children’s immediate goals and post-setback behavior under the study conditions. They do not establish that every child who hears “smart” develops fragile confidence.
Why a flattering identity can become a burden
The proposed mechanism is not that compliments are bad. It is that person labels turn a result into evidence about a stable self. “You solved that because you’re smart” creates an attractive explanation while the answer is right. The same logic becomes threatening when the answer is wrong.
If easy success proves intelligence, difficulty can look like disproof. Avoiding the difficult task then makes psychological sense: it protects the label. Giving up can also protect it, because trying fully and failing feels more diagnostic than not really trying.
Process feedback leaves more routes forward. “You compared the two examples and noticed the pattern” names a strategy. “You caught the mistake when you checked your work” names monitoring and revision. “You stayed with that even after the first idea failed” recognizes persistence without claiming persistence alone guarantees success.
This is the distinction in the title between a sturdy and fragile kind of confidence. Sturdy confidence is not certainty that every task will be easy. It is confidence that difficulty supplies information and that the next response can change.
Younger children showed a related person-versus-process pattern
Melissa Kamins and Carol Dweck studied five- and six-year-olds using role-play tasks. Their 1999 Developmental Psychology paper included one experiment on criticism and another on praise. In the praise study, 64 children acted out successful tasks, received person, outcome or process feedback, and then encountered a setback.
Children who had received person praise, feedback aimed at the whole self, showed more helpless responses afterward than those who received process praise. Their reactions included more self-blame, poorer self-assessments and less persistence. Person criticism produced a related pattern in the other experiment.
The design matters. Children were role-playing scenarios arranged by researchers; parents were not observed speaking naturally at home. The study supports the idea that whole-person judgments can make worth feel contingent on performance, but it does not measure the long-term effect of a real parent’s vocabulary.
Home observations add realism, not causal certainty
Elizabeth Gunderson and colleagues moved outside the laboratory. For a 2013 paper in Child Development, researchers recorded parent-child interactions when children were 14 to 38 months old, coded the praise parents spontaneously used and assessed the children’s motivational frameworks five years later.
Among 53 families, a greater proportion of process praise in early childhood predicted more incremental, challenge-oriented beliefs later. The association was especially clear for praise aimed at effort, strategies and actions rather than fixed qualities.
This is a small observational study. Parents were not randomly assigned to a praise style, so process praise could travel with other differences in teaching, language, expectations or family life. The result makes the laboratory pattern more plausible in ordinary homes, but it cannot show that substituting one sentence caused the later mindset.
It also should not turn a conversation into anxious self-monitoring. Children hear much more than praise: how adults respond to their own mistakes, whether questions are welcomed, whether help is available, and whether struggle brings curiosity or shame. Those conditions give any single phrase its meaning.
Good process praise is more than praising effort
The easiest way to misuse this research is to replace one automatic line with another. “You worked so hard” can be just as empty when the adult did not notice the work, when the child did not work hard or when effort was poured into an ineffective strategy.
A 2002 Psychological Bulletin review by Jennifer Henderlong and Mark Lepper argued that praise may enhance, undermine or have no effect on intrinsic motivation depending on how it functions. It is more likely to help when it feels sincere, provides specific information, points to controllable causes, supports autonomy and conveys standards the child can realistically meet.
That makes strategy and revision as important as effort. A parent might notice that a child organized materials before beginning, tried a smaller version, asked a precise question, returned after a break or recognized that the first method was not working. Sometimes the most useful response is neutral information: “The tower falls when that narrow block is at the bottom. What do you notice?”
Process praise is not a demand to admire endless struggle. A child may need instruction, a different tool, a rest or permission to stop. Persistence with no path forward can become frustration. The adult’s job is not to make effort sacred, but to keep ability from becoming a verdict.
The Artful Age has also examined research suggesting that inflated praise can land poorly for children with low self-esteem. Louder approval is not automatically more believable or more helpful.
The wider growth-mindset claims are more modest than the slogan
Praise experiments helped inspire a much larger growth-mindset movement in education. That broader evidence is mixed. A 2018 pair of meta-analyses in Psychological Science examined 273 effects involving 365,915 participants for the relationship between mindset and achievement, plus 43 intervention effects involving 57,155 participants. The average relationships and intervention effects were weak, although some academically at-risk or lower-income groups appeared more likely to benefit.
A later systematic review and meta-analysis of growth-mindset interventions also challenged sweeping achievement claims. Believing ability can develop does not replace good teaching, subject knowledge, time, sleep, learning support or a strategy that fits the task.
That limitation does not erase the praise findings. It puts them in scale. Feedback can influence how a child interprets success and failure, but no sentence can carry an entire educational environment.
The careful takeaway is not to ban “smart” or become frightened of ordinary affection. It is to avoid making intelligence the identity a child must keep proving. Specific, honest attention to process gives them somewhere to go when the work becomes difficult: back to the method, the next useful change and the possibility that learning is still happening.