Sound findings and appropriate statistics: Response to Snowling and Hulme

Abstract

In summary, Snowling and Hulme were entitled to claim that, despite the solid effect sizes, the case for the effectiveness of the DDAT treatment was 'not proven' owing to the imbalance in initial reading scores. The analyses of covariance fully resolve this issue, confirming that the significant differences remain even when the effects of any initial differences are removed. The treatment did therefore lead to significant effects, and our conclusions were indeed sound. Copyright© 2003 John Wiley & Sons, Ltd.