In this study, we use large-scale, longitudinal data to model the growth in student reading across 3rd, 4th, and 5th grade in three states using multi-level models for change (Singer & Willett, 2003) to fit individual growth models for students nested in schools and districts. Students who were identified as gifted had 3rd grade reading scores that were nearly a full standard deviation higher than their non-gifted peers. However, students identified as gifted showed reading growth that was either similar to, or slightly lower than, their non-gifted peers.