The New Blueprint for Learning: What Science Says Schools Are Getting Wrong

A Curious Child Walks Into a Classroom

Picture a 10-year-old in Cardiff who loves asking "why." She's genetically predisposed toward strong academic ability — but whether that potential ever becomes achievement depends less on her IQ than on something harder to measure: her curiosity, her self-belief, her hunger to keep asking questions.

That's not philosophy. That's data.

A major study led by Queen Mary University of London, published in Nature Communications, followed more than 5,000 children in England and Wales between ages 7 and 16. Researchers combined genetic analyses with developmental and psychological data and found that noncognitive skills — motivation, academic curiosity, self-concept, attitudes toward learning — explain a substantial portion of why children differ in educational achievement. Crucially, the influence of those skills grew stronger as children got older. Genetics loads the gun, but curiosity pulls the trigger.

"The strongest effects were linked to education-focused characteristics such as academic curiosity, motivation, self-concept and attitudes toward learning," said lead author Quan Zhou, a postdoctoral researcher at Queen Mary. The implication is clear: education systems that pour all their energy into test prep and content delivery may be missing the deeper levers of student success.

The Intervention Window — and Who Gets Access to It

Half a world away, a different group of researchers was asking a related question: if mindset and engagement matter so much, when and where do we intervene?

A study co-authored by Sarah Cohodes, associate professor at the University of Michigan's Ford School of Public Policy and published in the Journal of Human Resources, looked at STEM-focused summer pipeline programs designed to increase diversity in science and engineering fields. The findings were striking. High school students who participated were significantly more likely to enroll in — and graduate from — elite colleges with a STEM degree. Their predicted earnings rose by anywhere from 3% to 15%.

This matters enormously in the post-affirmative-action landscape. Since the U.S. Supreme Court's 2023 decision ended race-conscious admissions, universities have scrambled for legal, effective ways to diversify their campuses. These pre-college programs, Cohodes argues, represent exactly that. "When we intervene matters," she said. The window before college applications — when students are still choosing which schools to dream about — may be the most powerful moment of all.

AI Is Already in the Classroom. The Question Is How.

While researchers debate how to get more students into higher education, a separate transformation is reshaping what happens once they arrive.

Researchers at James Madison University, publishing in Research & Practice in Assessment, used the Generative AI Literacy Assessment Test (GLAT) to measure incoming students' understanding of AI tools. What they found should give every university administrator pause. Students were already using AI — but their underlying understanding of how those tools work, and how to critically evaluate them, was "still very limited," according to Stuart Miller, assistant director of Academic Data Acquisitions and Reporting. The study's authors argue that institutions cannot assume AI literacy and must build broad-based education programs for all students from day one.

Meanwhile, researchers from the University of Phoenix College of Doctoral Studies published findings in the International Journal of AI in Pedagogy, Innovation, and Learning Futures showing that doctoral students' attitudes strongly shape their ChatGPT usage — and that students who view AI-generated responses as superior to their own work use the tools more frequently. Critically, no gender gap was found, but significant differences emerged across fields of study, suggesting that disciplinary culture shapes how students relate to AI.

The takeaway isn't that AI is dangerous in classrooms. It's that we're handing students a powerful tool without teaching them how to hold it.

Leaders, Lanes, and the Infrastructure of Change

This theme — that the environment shapes outcomes as much as individual ability — runs through the week's research like a thread.

Texas A&M researchers Seung Won Yoon and Kyung Nam Kim, publishing in Human Resource Development Review, reviewed more than 600 articles on leadership, job crafting, and career development. Their conclusion: as AI reshapes the workforce, strong leaders who actively help employees redesign their roles will be the difference between organizations that thrive and those that flounder. "Leaders are not simply people who influence how jobs can evolve in the workplace," Kim noted. In the AI era, that guidance becomes existential.

In New York City, the infrastructure argument is playing out on asphalt. An NYU Tandon School of Engineering study, published in npj Sustainable Mobility and Transport, analyzed approximately 72 million Citi Bike trips recorded between 2013 and 2024. The finding was unambiguous: protected bike lanes — those physically separated from traffic by curbs or barriers — increased ridership by an average of 18%. Painted lanes and sharrows? No statistically significant effect. Safety infrastructure that feels real changes behavior. Safety theater does not.

That lesson echoes in a Columbia University Mailman School of Public Health study, conducted with Yale School of Public Health and published in the American Journal of Epidemiology, examining year one of New York City's Congestion Relief Zone. Overall monthly crash rates declined substantially after tolling began in 2025. Travel times dropped. Air pollution fell. The physical redesign of a city's relationship with cars, it turns out, saves lives.

What the Horses Know

Perhaps the most quietly radical finding of the week came from Gothenburg, Sweden, where University of Gothenburg researchers fitted heart rate monitors to eight Gotland Russ horses at the city's Children's Zoo. The question: does being petted by hundreds of strangers stress them out?

The answer, published in Zoo Biology, was no. Human contact didn't raise their heart rates at all. The excavator parked nearby did.

Lead author Isidora Dundjerovic put it simply: "They actually don't seem to mind all that much." It's a small finding, but it carries a lesson worth keeping. Stress comes from environments we can't predict or control — not from connection itself. The horses aren't distressed by being seen and touched. They're distressed by chaos.

The Common Thread

From Cardiff classrooms to New York bike lanes to Gothenburg pony parks, this week's research is telling one story: the environment we build determines the outcomes we get. Give curious kids the tools to stay curious. Give underrepresented students a pathway before the gate. Give workers leaders who help them adapt. Build bike lanes that actually feel safe. Design cities where people — and horses — aren't overwhelmed by noise.

The science keeps arriving, piece by piece, pointing toward the same conclusion. We already know enough to build better. The question is whether we'll choose to.