How should we praise children? In a 1998 study, fifth graders solved puzzles. Some heard, "You must be smart." Others heard, "You must have worked hard."

Then everyone tried very hard problems and failed. The "smart" group lost interest and did worse later. The "hard work" group kept trying. But later studies did not always find the same result, so this is not a magic rule.

Telling a child "You're so smart!" seems kind, but a famous study suggests it may have side effects. In 1998, psychologists Claudia Mueller and Carol Dweck gave fifth graders puzzles and then praised them for either their intelligence or their effort.

Next, all the children faced difficult problems and failed. Children praised for intelligence tended to choose easy tasks, enjoyed the problems less and performed worse afterward. Children praised for effort kept trying. Dweck explains this with mindsets.

Praise for intelligence may suggest that ability is fixed, so failure feels like proof of being unintelligent. However, some later studies, including a series with about 600 children in China, did not find a clear difference between the two types of praise.

In a famous 1998 study, Claudia Mueller and Carol Dweck asked whether praising a child as "smart" could backfire. Across six studies involving about 400 fifth graders, children solved a set of problems and then received praise for their intelligence, praise for their effort, or no extra comment.

Next, all of them failed on a harder set. Those praised for intelligence were more likely to choose easy tasks, enjoyed the work less, gave up sooner and performed worse. Many also inflated their scores when reporting them to other children. Dweck explains the results in terms of mindsets. Praise for intelligence may encourage a fixed mindset, in which failure signals a lack of ability, while praise for effort supports a growth mindset.

The findings spread quickly through schools, but later evidence is less clear-cut. A 2019 series of studies, including a close replication, with roughly 600 Chinese children and a 2020 study of Dutch vocational students found no clear difference between the two kinds of praise, and a 2018 meta-analysis concluded that growth mindset effects on academic achievement are generally small. Other data point the other way. In an observational study, the amount of process praise parents gave their toddlers predicted the children's attitudes years later.