Fluid reasoning and verbal knowledge stand out as math predictors in a meta-analysis of 122 test batteries

The useful design choice here is the pooling. A single battery, such as the Wechsler test commonly used to evaluate children's intellectual abilities ^[14], measures abilities in its own way, so a link found in one instrument can belong to that instrument. Putting 122 distinct batteries into one analysis ^[4] tests which abilities keep turning up beside math skill whoever built the test. Fluid reasoning and comprehension knowledge came out as two of the most consistent ^[3].

Consistency is about agreement across instruments. Size is a separate question. The published summary does not report effect sizes, so a reader cannot yet tell how much of the spread in math skill these two abilities account for.

Comprehension knowledge covers language, vocabulary and verbal reasoning. Fluid reasoning is deductive or inductive reasoning ^[5]. Daniel Hajovsky, an associate professor of educational psychology at Texas A&M and a co-author of the paper ^[1][2], said of comprehension knowledge: "It's important for reading, it's important for writing and we see now it's important for mathematics." ^[6] For fluid reasoning he offers a test question: "How well can a student learn a game when there are no prior rules or background knowledge required for it?" ^[7]

Hajovsky also described the finding that complicates any single-ability account. "Performance on one mental task is positively correlated with performance on a different mental task," he said ^[10]. Such tasks are "both tapping into this mental engine that drives performance on these unique cognitive tasks," in his words ^[16]. When every ability correlates with every other, I'd expect part of any link between reasoning and math to come from that shared engine and not from reasoning alone.

Number sense sits beside the two cognitive abilities. It was "just as important and a strong foundation for later math skills," Hajovsky said ^[8]. He added: "Cognitive abilities are important for learning and academic outcomes, but it's not what fully makes up someone's math ability." ^[9]

The study relates scores across existing tests ^[4]. That design shows which abilities travel with math skill. It cannot show that training fluid reasoning would raise a child's calculation or problem-solving scores. Answering that takes a trial with a control group.

Hajovsky's own proposal stays close to what the data support. He wants screening that looks at "the pattern of errors" a student makes ^[11], to work out "whether math challenges stem from conceptual difficulties like language, reasoning or efficiency issues" ^[15]. He argues for acting early because math classes build on earlier topics ^[12]. He also said these skills keep developing inside and outside the classroom and can be built through time and effort ^[13].