Arrowsmith Program: What has changed? Not an awful lot.

Almost three years ago to the day, we asked such an important question on ABC Adelaide that it sparked a Media Watch segment.[1] Former Dyslexia SA Executive Member and current Director, Code REaD Dyslexia Network, Sandra Tidswell, asked this question: “Where is the published peer reviewed evidence that the Arrowsmith program works?” to controversial business promoter, Barbara Arrowsmith-Young.

Barbara Arrowsmith-Young is again touring Australia.

What has changed?  Not an awful lot.

The Arrowsmith program consists of a number of so-called ‘cognitive exercises’, developed by Arrowsmith-Young, which supposedly correspond to 19 areas of learning dysfunction. How are these 19 areas identified? Well, by a series of tests invented by the program’s creator. Are you seeing any red flags yet?

And if those red flags aren’t waving for you now, please consider that students in the Arrowsmith program typically attend for 40 weeks over three-to-four years and complete at least 200 sessions per year, at a cost up to $24,000 annually.[3][4]

Importantly no empirical work has been conducted to link these exercises to the areas of learning dysfunction she identified. Modern cognitive or neuropsychological theory would also beg to differ.[2]

But fear not, over 30 years since its invention, the author will now claim to have published peer-reviewed evidence.[5]

We will take a look at the report, but firstly, while still recruiting for the study, lead researcher Dr Lara Boyd, a stroke researcher from The University of British Columbia, released a study update in April 2016.

The stated aim was to recruit 90 children, across:

• an intervention group receiving the Arrowsmith program
• a control group of children with learning disabilities receiving other intervention
• a control group of typically developing children

This April 2016 update stated to have already recruited:

•  32 children from the Arrowsmith Program
• 14 children with learning disabilities who are not in the Arrowsmith Program
• 10 typically developing who do not have a learning disability[6]

Neuropsychology Professor Dorothy Bishop’s, comments on the October 2019 report are enlightening.

“The report describes data from 28 children undergoing the Arrowsmith programme, with no mention of controls, except in the first sentence of Analyses, where the sample characteristics are reported ‘relative to the typically-developing sample’. Yet no data are presented for that sample, nor for the more informative control group.

…it seems disingenuous to present data just from the Arrowsmith group, dropping crucial control data that might actually provide some indication of whether this method – which is extremely expensive in both time and money – is beneficial.

The question for the authors, therefore, is where are the control data, and why did they think it appropriate to drop them from this report?”[7]

The study sample of 28 participants, while having a history of a learning challenge that enabled their participation in the Arrowsmith program, seemingly the only requirement for the intervention group; the report notes that one-third of the Arrowsmith program participants did not perform below normative expectations at baseline.

An intervention can’t ‘work’ when there is no deficit that requires intervention.

Prof. Bishop explains further:

“Here is a description of the sample.  Even if the average score was 87.89 some unknown number had scores above 90 – not LD (Learning Difficulties/Disability).

Some unknown portion may have had an LD, most did not.  Note that the treatment did not “work” on single word reading, decoding etc.  as they were in the average range to begin with.

As a sample, the group also demonstrated Average (Standard Score = 90–109) single word reading, reading fluency, spelling, written expression, decoding, and maths problem solving skills. Their maths fluency (M = 87.89, SD = 17.74), reading comprehension (M = 89.00, SD = 16.15), and maths computation (M = 86.54, SD = 19.73) were in the Low Average range of performance (McGrew et al.,2007). Thirteen of the participants (46.4%) performed poorly (below a standard score of 85, lower than one standard deviation below the mean) on at least one measure of reading, suggesting a possible specific learning disorder or challenge in this area. Seventeen (60.7%) of the participants performed poorly (below a standard score of 85) on one measure of maths, suggesting a possible specific learning disorder or challenge in this area. Eleven of the participants (39.3%) performed poorly on at least one measure of writing (below a standard score of 85). Nine of the participants (32.1%) had no identified learning challenges in reading, maths, or writing. Nine (32.1%) had learning challenges in all three academic domains, five (17.9%) had only maths challenges, and two (7.1%) had maths and reading challenges.” (Developmental Disorders of Language and Literacy Network, 14.1.20)

The report also acknowledges the work was funded by private donations to Dr. Boyd’s Brain-Behaviour Lab to study educational neuroplasticity, and no conflict of interest was reported by the authors.[8] Where these private donations came from were not disclosed, but include, The Eaton Arrowsmith School.[9]

So here we are, three years on, and very little has changed. There is still no evidence that the Arrowsmith program works. 

[1] https://www.abc.net.au/mediawatch/episodes/controversial-claims-get-the-soft-treatment/9972760

[2] https://www.tandfonline.com/doi/full/10.1080/23735082.2019.1674908

[3] https://www.tandfonline.com/doi/full/10.1080/23735082.2019.1674908

[4] https://www.spelfabet.com.au/2015/05/brain-training/

[5] https://www.tandfonline.com/doi/full/10.1080/23735082.2019.1674908

[6] http://www.ppaap.org/wp-content/uploads/2017/02/UBC-Brain-Study-Update-April-2016.pdf

[7] https://pubpeer.com/publications/217BF88F6490215E4D36881FA362B2#1

[8]https://www.tandfonline.com/doi/full/10.1080/23735082.2019.1674908

[9] http://web.archive.org/web/20180615075820/http:/brain.rehab.med.ubc.ca/research/funding/