An experimental researcher selects students who had low scores on a math pre-test and high scores on a math post-test to participate in a study examining the influence of a new math curriculum in an elementary school grade. What potential threat to validity exists in the experiment?