This study presents findings from a randomized controlled trial of an afterschool program intended to develop mathematics identity for students from grades 4 and 5 in groups underrepresented in STEM. Mathematics identity refers to the ways that students think about themselves in relation to mathematics and the extent to which they have developed a commitment to, and have come to see value in, mathematics. While the impact analyses showed no effects of the intervention on mathematics identity or achievement, the exploration of the longitudinal data collected over 2 years provided several insights. On average, student mathematics identity remained constant over the study period; however, the overall averages mask large variations in individual students and sites. Some students saw improvement in mathematics identity, while others saw decreases. Counter to findings in previous literature, we found no overall differences by gender suggesting that boys and girls report similar mathematics identity. Importantly, we found a positive relationship between mathematics identity and achievement. This finding holds in both directions and suggests that boosting mathematics identity could lead to improving mathematics achievement and vice versa. This study contributes to our understanding of mathematics identity, how it is measured, how it evolves over time, what relationship it has to mathematics achievement, and what its potential for development in afterschool environments could be. (author abstract)