Researchers alter the amount of light reaching one set of plants while leaving a comparable set under usual lighting. They then compare growth. What makes this design relevant to causal direction?