Phenomenon where observed intervention effects being more different from each other than one would expect due to random error (chance) alone.