Increasing the success of Underrepresented Students of Color (USC) in science, technology, engineering, and mathematics (STEM) is a central concern to many researchers, policymakers, and educators. To help understand STEM college student success, many studies have utilized Holland's (1966, 1973, 1985, 1997) person-environment fit framework applying it uncritically to all students. Using Quantitative Criticalism, this study engages the racial realities of USC while investigating several assumptions of Holland's theory and their implications for USC pursuing STEM fields. Utilizing a national, longitudinal dataset of 5,564 STEM bachelor's degree recipients drawn from the Cooperative Institutional Research Program's 2004 Freshman Survey and 2011 Post-Baccalaureate Survey, this study specifically examines students' interest in making a positive impact on society through socio-political action, or social agency, which Holland's typology suggests is incongruent with STEM environments. Findings show that USC may be more likely to be described as "incongruent" with Holland's classification of STEM environments, that the congruence assumption may not be fully applicable for understanding the long-term success of USC in STEM, and that the social agency of USC did not significantly change over the seven years while white students' significantly decreased. Implications for broadening participation and promoting equity in STEM fields are discussed.