A large-scale study challenges the idea that children’s reading difficulties are primarily explained by intelligence or visual processing. Instead, it identifies recognizing speech sounds and building vocabulary and background knowledge as especially strong predictors of reading performance.
Led by Daniel Hajovsky, an associate professor of educational psychology at Texas A&M University, the research combined 50,000 correlations from 137 cognitive and achievement test batteries. The team included researchers from Texas A&M, the University of Kansas, St. John’s University, Montclair State University and Temple University.
The analysis found no meaningful relationship between visual processing and reading ability. General intelligence was associated with reading, but its influence was smaller than some earlier research had suggested. Auditory processing and comprehension-knowledge had stronger, more specific links to reading, including decoding and comprehension.
Auditory processing is the ability to recognize and distinguish the sounds in spoken words. It helps children connect those sounds to letters as they work out how written words are pronounced. Comprehension-knowledge includes vocabulary, concepts and background information that support understanding words, sentences and longer passages.
Memory, reasoning and processing speed also contributed to reading development, but played more supportive roles, according to the study. The findings suggest that reading draws on a set of connected abilities rather than serving as a simple measure of intelligence or visual skill.
The researchers say the results could help educators identify the particular skills behind a student’s reading difficulties and tailor instruction accordingly. They also indicate that struggling readers are not necessarily less intelligent; some may need targeted support with specific foundational abilities.
The study, led by Hajovsky, was published in the *Journal of Intelligence* in 2025 under the title “Toward a Consensus Model of Cognitive–Reading Achievement Relations Using Meta-Structural Equation Modeling.”
Sources: ScienceDaily.