This work considers the effect of four years of chess lessons in Germany. The test school's results in cognitive and behaviour assessments were significantly better than a control school with similar demographics, but the schools were not randomly allocated to test and control groups. Read more here.