Dyscalculia’s Rival Definitions, and What Each Tells a School

What the review found

Under the ICD-10, dyscalculia is a serious impairment in the learning of basic numerical and arithmetical skills in a child whose intelligence and schooling are otherwise adequate, established by standardised testing that shows poor calculation alongside normal intelligence. The requirement to demonstrate that gap was already being withdrawn when Liane Kaufmann, of UMIT in Hall in Tyrol, and Michael von Aster, of the University Children’s Hospital of Zurich, surveyed the field for Deutsches Ärzteblatt International in 2012. The fifth edition of the DSM, then a year from publication, would group the learning disorders as dimensions within a single category and drop the discrepancy rule, on the grounds that children given this diagnosis differ too much from one another for one rule to sort them.

A third term circulates in education research. A child is said to have a mathematical learning disability when a score on a mathematics test falls below a percentile of a study population chosen arbitrarily by the researchers, with no reference to intelligence. Kaufmann and von Aster note that these competing criteria produce study populations and testing procedures that cannot readily be compared with one another, which is part of why the treatment literature remains as unsettled as it is. Their article surveys the field selectively, with no stated search protocol, rather than pooling studies. Both authors declare licence fees and royalties on standardised arithmetic tests, which is a stake in the category of instrument the article goes on to recommend.

The condition itself is common and steady in its rate. About 5 percent of primary school pupils are affected, a figure that has held across the countries where it has been measured, and it matches the rate for dyslexia. Between 20 and 60 percent of children with dyscalculia have a second learning difficulty, most often dyslexia or ADHD. Among children with learning disorders generally, 30 to 50 percent have a diagnosable mental health condition, against 8 to 18 percent of children without.

The instruments in use divide into two kinds, and they were built to answer different questions. Curricular tests establish whether a child has met the learning objectives for the current grade, credit only the correct answer, and can be given to a whole class at once. Neuropsychologically oriented tests are given to one child at a time and record processing speed and the strategies used to arrive at an answer, with the aim of producing a description of the child instead of a rank. Most tests of either kind place the diagnostic threshold at the tenth percentile. Because children with dyscalculia often perform far below their own grade level, Kaufmann and von Aster warn that a curricular test alone can miss where the difficulty actually lies, and an intervention pitched at the wrong level follows.

What it means for you and your child

A diagnosis of dyscalculia establishes that a child’s arithmetic falls below a threshold and that low general ability does not account for it. The description of how that child calculates comes from a different instrument, and the review is explicit that intervention planning should begin from the performance profile.

The evidence on what to do next is genuinely mixed, and the review does not disguise it. A meta-analysis by Kroesbergen and van Luit pooled 58 intervention studies with primary school pupils; the samples ran from 3 children to 376, and the effect sizes from minus 0.44 to above 3. Within that spread, interventions aimed at basic numerical skills outperformed those aimed at precursor skills or problem solving, and instruction delivered in person outperformed computerised programmes. A separate meta-analysis of individualised intervention in English-speaking countries reported an effect of 0.91 and found that teaching a child a strategy worked considerably better than presenting material to be absorbed.

On duration the two analyses contradict each other. Kroesbergen and van Luit found shorter programmes more effective, while Ise and colleagues, pooling eight studies from German-speaking countries, found longer and more intensive delivery better, for an overall effect of 0.50 the review calls intermediate. Most German-language teaching programmes for arithmetic have never been tested for efficacy at all.

Rubinsten and colleagues gave methylphenidate to children with ADHD, some of whom also had a mathematics difficulty, and found improvement in the parts of calculation that lean on working memory, such as carrying in multi-digit sums, with no effect on basic numerical skills. Those children, the authors concluded, need specific learning therapy as well as a prescription.

What this makes worth asking at a school assessment meeting is narrow and answerable. Which test was administered, whether it was curricular or neuropsychologically oriented, and what percentile the score represents are all a matter of record. The question that determines teaching is which numerical skills the child has acquired and which are missing, and what strategies the child used to reach the answers that were right. Untreated, this difficulty persists into adulthood, so the answers matter beyond the current year. If the school’s answer is that only a curricular test was given, then the profile does not yet exist, and that is the finding you take away from the meeting.