This study evaluates the impact of the Retired Mentors for New Teachers program, a two-year mentoring program at the elementary school level developed by Aurora Public Schools in Colorado. Many of the district's schools serve a large percentage of economically disadvantaged children, experience high teacher turnover, and hire newer, less experienced teachers. The program addresses these challenges using master educators who recently retired from the district to provide tailored one-on-one mentoring to new teachers. The program requires mentees to meet weekly one-on-one with their mentor and monthly in school-level groups over the course of two years. This study was undertaken by Regional Educational Laboratory (REL) Central in collaboration with Aurora Public Schools. It used a randomized controlled trial to assess the impacts of the Retired Mentors for New Teachers program on student achievement, teacher retention, and teacher evaluation ratings during the 2013/14 and 2014/15 school years. As part of the study, the district's elementary school teachers were randomly assigned to either a group that received only the district's typical mentoring support (the business-as-usual group) or a group that received both typical mentoring support and added support from a retired mentor under the Retired Mentors for New Teachers program (the program group). The business-as-usual support involved first-year teachers being assigned to work with a more experienced "buddy" teacher for about 15 hours over one school year. The study team collected data on resources and costs associated with running the program and generated a return on investment estimate. The study team also used teacher and mentor surveys and focus groups to investigate whether the program was implemented with fidelity to its model. The following key findings emerged from the study. The first four are causal findings; the last four are exploratory findings from exploratory analyses: (1) At the end of the first year math achievement was significantly higher among students taught by teachers in the program group than among students taught by teachers in the business-as-usual group; (2) While the differences were not statistically significant, reading achievement was also higher among students taught by teachers in the program group than among students taught by teachers in the business-as-usual group; (3) The program had no effect on teacher evaluation outcomes; (4) Although more teachers in the program group than teachers in the business-as-usual group left the district after two years, the effect of the program on teacher retention was not significant; (5) Increased hours of mentoring were associated with higher teacher retention in the second year among teachers who participated in the program; (6) The program had an average annual local cost of approximately $171 per student; (7) Exploratory analysis suggested that the program could yield a return on investment that may pay back the annual cost of the program more than 15 times over through increased student earnings over time; (8) Overall, the program was implemented with fidelity to its intended model. The following are appended: (1) Data, outcome measures, baseline equivalence, and methodology; (2) Sensitivity analysis; (3) Elements considered in the analysis of program cost; (4) Implementation fidelity; and (5) Regression results.