This study employed design-based logistic regression to examine the extent to which attending a science, technology, engineering, and mathematics (STEM) magnet high school, school demographics, and student-level factors explain racial/ethnic differences in a student's intent to declare a STEM major in the first semester of college. Data were drawn from the High School Longitudinal Study of 2009 (HSLS:09). The results show only student-level factors (gender, ethnicity/race, highest math course level, and math achievement) are statistically significantly associated with a student's likelihood of intending to declare a STEM major. These findings suggest simply establishing a STEM magnet high school may not increase the likelihood of its graduates pursuing STEM careers. [This paper will be published in "The Educational Forum" v86 n4 2021.]