This paper replies to criticism by Cummings et al. [Cummings, P., Rivara, F.P., Thompson, D.C., Thompson, R.S., 2006. Accid. Anal. Prev. 38, 636-643] of an article [Curnow, W.J., 2005. The Cochrane Collaboration and bicycle helmets. Accid. Anal. Prev. 37, 569-573] disputing a conclusion of a Cochrane Collaboration review, namely, that it establishes scientific evidence that all types of standard bicycle helmet protect against injury to the brain. In response to the conclusion of Cummings that the review's case-control studies provide such evidence, I explain that their design is inadequate to do this.