The study was barely scientific. The researcher was his own test subject, and the result has never been replicated. Also, most crashes probably occur under conditions where the driver can't be aware of whether the cyclist is wearing a helmet or not. "The cyclist suddenly came out of nowhere" is a common defense.