Millions of children with exceptional potential in science, technology, engineering and maths may be going unnoticed in classrooms, according to a new review of global research into how gifted STEM pupils are identified. The study suggests that many schools still rely too heavily on traditional intelligence tests, even though these measures often fail to capture creativity, spatial reasoning and real world problem solving skills. The findings were published in the Journal of Intelligence.
Researchers from Tsinghua University in Beijing examined 26 peer reviewed studies published over recent decades, drawn from nine major academic databases. The systematic review, published in the Journal of Intelligence, followed the PRISMA framework to assess how STEM talent is identified in pupils aged five to eighteen across countries including the US, South Korea, Turkey, Taiwan, and Saudi Arabia.
The findings show a gradual shift away from relying solely on IQ scores towards a broader mix of approaches. These now include creativity assessments, spatial ability testing, academic achievement records, portfolio evaluations, subject competitions such as science Olympiads, and recommendations from teachers and parents.
Despite this progress, the review highlights a significant gap in early identification. None of the studies analysed focused on kindergarten or primary school pupils, with research concentrated instead on secondary and high school students. Given that talent development experts widely agree that ability begins forming in early childhood, the authors argue this represents a missed opportunity to support young learners before disengagement sets in.
The review also points to persistent inequalities in how STEM talent is recognised. Socioeconomic background, cultural context, and gender expectations all appear to shape who gets identified as gifted, often independently of a child’s actual ability. Pupils from lower income households and certain minority backgrounds remain underrepresented in gifted programmes, while girls continue to face lower rates of recognition in STEM fields despite comparable potential.
Spatial reasoning, described in the review as a crucial yet frequently overlooked cognitive skill, was found to be a strong predictor of creativity and innovation, particularly in engineering and computer science. Yet it remains rarely tested in mainstream gifted education systems, which still lean towards verbal and mathematical reasoning assessments.
The authors also note that most existing research originates from the US, raising concerns about a Western centric bias in how giftedness is defined and measured globally. They call for future studies to include non English language literature and to explore how emerging technologies, including artificial intelligence and data analytics, might be used to identify STEM talent more fairly and accurately.
For parents and educators, the message is clear. According to Stride Psychology, standardised testing alone cannot capture the full range of abilities that make a child suited to future careers in science and technology. A more holistic approach, one that values creativity, problem solving and early encouragement, may be essential to nurturing the next generation of engineers, scientists and innovators.
