However, they still come out on top because: a) the statistical issues are still simpler when you can control for just about any factor except the ones you are really studying, and, b) since the experimental setups are more standardized, the few people that do grasp statistics have provided tools and rules of thumb that actually work well for the cases where they are used.