In a study on sleep deprivation and reaction time, if the lab is noisier on some testing days than others, noise becomes an extraneous variable that could distort results independently of the sleep manipulation.