Abstract In a technologically driven society, math and science students in the United States are falling further and further behind their international counterparts, resulting in an influx of STEM focused, reformed K-12 schools, including schools focused on project-based learning (PBL). This article reports a study of the effectiveness of PBL on high school students' performance on state mandated standardized mathematics and science achievement measures. Manor New Tech High School is a nationally recognized model STEM school, with a diverse student population, where all instruction is delivered through PBL. Although there is ample research suggesting that PBL is advantageous for increasing STEM learning compared to conventional teaching approaches, there is a lack of studies randomly assigning students to receive PBL. Further, some of the effects observed for students attending project-based schools could be due to a self-selection bias for students or parents that choose such an alternative learning environment. This study addresses both of these concerns and found that students taught through PBL, as a group, matched performance of conventionally taught students on all science 11th grade and mathematics 9th, 10th, and 11th grade TAKS achievement measures and exceeded performance by a scale score increase of 133 for the 10th grade science TAKS measure by (B = 133.082, t = 3.102, p < 0.05). One possible explanation of the differences observed in this study could be the TAKS instrument used to capture student math and science achievement that interprets "real-life applications" of content differently between math and science questions. These results align with literature on the effects of PBL and deepen our understanding of these effects by providing a controlled study with random assignments to the PBL experience. Future research looking at the effect of PBL on achievement on the PISA could be beneficial in identifying benefits of PBL implementation in schools.