experimental method designed to reduce bias, typically accomplished by randomly allocating subjects to two or more groups, with one being a control group