When studies on school chess are claimed to have a positive effect, it is no big deal. Very few interventions in school show no effect or a negative effect. Size matters. On the average, the effect size of an educational intervention in school is 0,4. An effect size above 0,4 is therefore seen as a desirable outcome. This has been established in a synthesis of more than 800 metaanalyses of studies on educational interventions directed by John Hattie from the University of Auckland in his seminal “Visible Learning” (2009).
Giovanni Sala and Fernand Gobet (pictured above at our 2013 conference) from the University of Liverpool have picked up Hattie´s challenge and conducted the first real metaanalysis of studies on chess in school. According to seven criteria they included 24 studies based on more than 5000 students in the experimental and control groups. They found an average effect size of 0,338. It was less in reading abilities and slightly bigger in maths abilities, but smaller than Hattie´s treshold of 0,4. However, after excluding studies with less than 25 hours of chess Sala and Gobet found an average effect size of 0,428, which is quite satisfiable.
As Sala pointed out in his presentation at the conference, none of the studies reached the highest methodological standard of comparing chess not only with no intervention but also with another intervention. Their metaanalysis has been accepted and will soon be published by a journal. We are delighted that they gave us permission to publish a preprint version. If you want to quote it, please contact Giovanni.Sala@liv.ac.uk for an updated reference.