Experiments begin with random assignment to conditions ... In the same way that random selection makes your sample representative and eliminates bias, randomly assigning subjects to conditions ...