Lack of meticulousness in the classification of children into research groups – a main reason for difficulties in making progress in the learning disability field.

As I read research about math learning disabilities, I feel a growing sense of frustration at the ways children are classified into research groups such as "math learning disabled", "low achievers in math who are not learning disabled" and "children with typical math achievement".   In each study, children are classified into these groups by entirely different criteria.  Moreover, many times there's no published data to support group classification.  As a result, it's difficult to reach reasonable conclusions about the existence of differences between these three groups and the reasons for these differences.

The following research is an example of that:

van Garderen, D., Scheuermann, A., & Poch, A. (2014).Challenges students with learning disabilities experience when using diagrams as a visualization tool to solve mathematics word problems. ZDM. The International Journal on Mathematics Education, 46, 135-

This study deals with a very interesting subject: differences between children with learning disabilities and children with high math achievement in the ability to use diagrams for problem solving in math.  Lack  of meticulousness (as I see it) in the classification of children into research groups makes it hard to draw conclusions from this study.  That's why I'll focus here on group classification and not on the research findings.

This is a qualitative research with a small sample, but it demonstrates fundamental problems I see also in quantitative studies with large samples.

The study compared a group of 16 students who had been diagnosed as learning disabled (LD) to a group of 12 students who were high achievers in math.  All children were in forth to ninth grade.  The authors write:  "Students with LD met their local district eligibility criteria and had a full-scale IQ score of 80 or higher on the WISC – IV (2003). The students with LD did not have to have a specific mathematics learning disability to participate.  However; all students with LD were lower achieving in mathematics than their peers without LD."

This raises a few questions:

What were the criteria for learning disabilities that these students met? The criteria are not published in the paper.  Since criteria for learning disability often change, it's important to publish them so the reader will understand the group's characteristics.

The full scale IQ score – since children with an IQ of 80 (which is equivalent to the 9^th percentile) were included in this group, it's hard to say that it's a group that includes only children with (at least) average IQ.  (an IQ score of 85 and above is considered  an average (or above average) IQ score, since 85 is a score lower than average by one standard deviation).  Admittedly, it's better not  to use the full scale IQ score for learning disability identification, but when using the discrepancy model, it's best to have in the learning disability group children who have IQ scores of 85 and above.

Since no requirement for specific math learning disability was required (it's not clear why), the authors did not ascertain that the "LD" group had significantly low achievement in math (for example, math scores that are one standard deviation and more below the mean, or that fit a grade level that is 2 grades below the student's grade level).  The authors write:  "All students with LD were lower achieving in mathematics than their peers without LD", but it's unclear whether they had significantly low achievement or were achieving in the low end of the average range, for example. 

The LD group apparently had learning disabilities in reading, and that must have been the main or common reason for their LD diagnosis, but this raises the question whether their reading achievement  was significantly below average.  This is especially important in research dealing with word problems, in which reading plays an important part.  Although the problems were read to the students, obviously children with reading difficulties rely more on their working memory to keep in mind the problem's details, compared with children who are good readers and can reread the problems.  No data is given in the study about the LD children's reading level.

The high math achievement group's reading level is also not reported.  This is important as they are being compared  with children who (probably) have LD that manifests in reading difficulties.  The authors didn't make sure that the high math achievement group did not have LD that manifests in reading.

This is one example of many of the frustrating situation with research in the LD field.