If we collect information on students' study practices and exam scores and find strong evidence of a positive association between not studying more than two hours and high grades, when can we generalize this finding to the general population?