In Reply Our study showed that, among individuals with severe mental illness, the rates of tobacco, alcohol, and other substance use are dramatically higher than in the general population.1 The criticisms made by Knapp and Hayat highlight long-standing statistical controversies: effect size vs P value and proper adjustment for confounding variables.
Hartz SM, Bierut LJ, Pato MT. Very Small P Values—Reply. JAMA Psychiatry. 2014;71(8):969. doi:10.1001/jamapsychiatry.2014.707