Racial and ethnic achievement gaps contribute to the lack of underrepresented minorities in STEM-related careers. This research is grounded in the expectancy-value model of achievement motivation which posits that motivation is influenced by social-cognitive variables such as self-efficacy and beliefs about the usefulness or utility of the task. These social-cognitive variables (self-beliefs and task beliefs), in turn, are influenced by numerous ecological factors such as parental involvement. Parent involvement can promote children's self-efficacy and task beliefs which is important for STEM interest and persistence. This study utilized data from the High School Longitudinal Study: 2009 (HLS: 2009) to examine links among parent involvement and underrepresented students' STEM self-efficacy, utility, interest and achievement. Concurrent and longitudinal models conducted in MPlus tested whether parent involvement in 9th grade predicted STEM interest and achievement directly and indirectly via self-efficacy and utility. Concurrent models showed that parent involvement in STEM was significantly related to adolescents' STEM self-efficacy which in turn was significantly related to STEM interest and achievement. Longitudinal models showed that parents' STEM involvement in 9th grade predicted adolescents' STEM efficacy in 11th grade which in turn predicted adolescents' cumulative GPA in STEM courses. Parent involvement in STEM was more strongly and consistently linked to self-efficacy than to utility. These results suggest that parent involvement in STEM helps adolescents to feel more confident in their STEM abilities but it does not necessarily contribute to adolescents' STEM utility values.