Data-driven heterogeneity in mathematical learning disabilities based on the triple code model.
Kids with math learning disabilities sort into distinct cognitive profiles—check for representational vs. verbal deficits when designing individualized math interventions.
01Research in Context
What this study did
Peake et al. (2017) ran a cluster analysis on kids with math learning disabilities. They used the triple-code model as a map. The model says math skills live in three brain codes: pictures of quantity, spoken number words, and written symbols.
The team looked for groups that matched each code. Younger and older grades were checked separately. No lessons were given; this was pure assessment work.
What they found
The numbers fell into clear groups. Older kids showed two clusters: one weak in pictures of quantity, the other weak in verbal number words. Younger kids added two more: a spatial group and a mixed low group.
Each cluster fits one part of the triple-code model. The finding says MLD is not one blanket label; it is several smaller problems wearing the same coat.
How this fits with other research
Granieri et al. (2020) looked at the same age and also used cluster analysis. They found only one big low-math group, not separate subtypes. The two studies seem to clash. The gap likely lives in the tests chosen and how strict the cut-off was for calling a kid MLD.
Vanbinst et al. (2014) tracked kids long before 2017. They showed that symbolic magnitude weakness predicts long-term math fact trouble. Christian’s verbal cluster lines up with that earlier warning sign.
Sajith et al. (2008) used a similar math-free method on autism and still found subgroups. The tool kit—cluster analysis—keeps pointing to hidden pockets inside wide diagnoses.
Why it matters
If you test a learner who can’t judge dot arrays quickly, picture-based tasks may unlock progress. If the child stumbles on spoken counting, switch to choral facts and rhythmic drills. Start with a short triple-code probe next week; five minutes can steer your goal choice better than a generic math goal bank.
Want CEUs on This Topic?
The ABA Clubhouse has 60+ free CEUs — live every Wednesday. Ethics, supervision & clinical topics.
Join Free →Give a one-page dot comparison and rapid naming probe; pick the weaker score to guide your first intervention target.
02At a glance
03Original abstract
Many classifications of heterogeneity in mathematical learning disabilities (MLD) have been proposed over the past four decades, however no empirical research has been conducted until recently, and none of the classifications are derived from Triple Code Model (TCM) postulates. The TCM proposes MLD as a heterogeneous disorder, with two distinguishable profiles: a representational subtype and a verbal subtype. A sample of elementary school 3rd to 6th graders was divided into two age cohorts (3rd - 4th grades, and 5th - 6th grades). Using data-driven strategies, based on the cognitive classification variables predicted by the TCM, our sample of children with MLD clustered as expected: a group with representational deficits and a group with number-fact retrieval deficits. In the younger group, a spatial subtype also emerged, while in both cohorts a non-specific cluster was produced whose profile could not be explained by this theoretical approach.
Research in developmental disabilities, 2017 · doi:10.1016/j.ridd.2017.10.005