When I was little, one of my favourite books was, The Little Engine that Could. In case you’ve not come across this paragon of (possibly feminist?) children’s literature, I’ll summarize: An adorable, blue engine pulls a train full of toys up a seemingly insurmountable hill, chanting a self-confidence-boosting, near-onomatopoetic mantra 0f “I think I can! I think I can!” until it succeeds.

Not Joyce or Shakespeare, perhaps, but the refrain stuck with me and saw me through some tough times (with the notable exception of high school chemistry, in which I thought I could and … definitely couldn’t). So, you can imagine how interested I was to see a new longitudinal study looking at how kids’ self-concepts are linked to their actual academic improvement in math and reading — in other words, whether “I think I can” leads to “I actually did.”

The little engine’s lesson now has scientific backing:

“Our study shows that youths’ perceptions of their abilities in middle childhood are important in promoting their later achievement in math and reading,” explains Maria Ines Susperreguy, assistant professor of education at Pontificia Universidad Catolica de Chile, who led the study. “This relation is not limited to students who perform at the top levels, but extends to students with different levels of achievement in math and reading. Even the lowest-performing students who had a more positive view of their math and reading abilities had higher levels of achievement in math and reading.”

The researchers examined measures of self-concept and standardized academic assessments for more than 15,000 British and American students between ages 5 and 18. (Self-concept was defined as kids’ perceptions of their capability to succeed in academic tasks.) Researchers controlled for other influencing factors such as children’s earlier achievement as well as birth weight, race/ethnicity, gender, age, their mother’s education, and more.

The team found that children’s beliefs about their personal math and reading abilities explain some of the variance in their later math and reading achievement. Specifically, across all achievement levels — poor, middling and high performing students — higher self-regard of mathematical ability predicted later math improvement, and higher self-regard of reading ability predicted later reading improvement. But the self-confidence was not transferable between subjects; self-concept in reading skills, for instance, did not affect achievement in math skills.

“When trying to understand the issues of low academic performance, we often examine what additional skills children need to succeed in school,” says Pamela Davis-Kean, professor of psychology and research professor at the Institute for Social Research at the University of Michigan, who co-authored the study. “Our findings, replicated across three data sets, show that it is important to understand the relation between children’s perceptions of their abilities and later achievement.”

In other words, instead of signing him up for an extra tuition, maybe the best thing you can do to help a kid improve his school performance is to get him to think he can help himself.*

*Unless it’s chemistry.