To the Editor.
—The randomized controlled trial by Dr Grossman and colleagues1 of a violence prevention curriculum is mistitled; the study is neither a randomized trial nor an investigation of violence. Second, procedural flaws, data errors, and conjectural interpretations cast doubt on the curriculum's reported success.Contrary to common usage, words like "aggression" did not signify physical attack but milder conduct like stomping one's feet or sticking out one's tongue.2 In theory, such conduct may induce violence. However, the present study tracked nonviolent behavior in 8-year-olds, so phrases like "violence prevention" are speculation. The accompanying Editorial went further, praising the curriculum for "addressing intentional injuries" and potentially yielding "safer homes, safer schools, and safer communities."3More troubling was the unfounded depiction of this research as a randomized controlled trial. In this study neither children nor schools were randomized. The pairing of schools followed by random assignment within
Green J, Wintfeld N. A Violence Prevention Curriculum. JAMA. 1997;278(12):979-980. doi:10.1001/jama.1997.03550120039024