Context: Concussion is common among military service members. In response to advancing clinical practice guidelines regarding concussion management, service members, like athletes, complete a baseline assessment prior to participating in high-risk activities. While several studies have established test stability in athletes, no investigation to date has examined the stability of baseline assessment scores in military cadets. Objective: To evaluate the test-retest reliability of the Immediate Post-Concussion Assessment and Cognitive Testing (ImPACT) among military cadets from year-to-year. Design: Reliability study. Setting: United States Service Academies. Patients or Other Participants: 10,603 cadets (19.69 ± 1.41 yr, 69.72 ± 3.66 in, 167.67 ± 28.26 lbs) enrolled in the Concussion Assessment, Research and Education (CARE) Consortium at three United States Service Academies. Cadets diagnosed with a concussion (n = 800) and those that changed their concussion history (n = 315) between baseline assessments were removed from the data set yielding a sample of 9,488. Interventions: All cadets participating in the CARE investigation completed the standard baseline battery of assessments that included the ImPACT. At the time of this analysis, 3,965 had completed a baseline ImPACT in year one (Y1) and year two (Y2) and 195 cadets had completed a baseline in Y1 and year three (Y3). Main Outcome Measures: Baseline assessments were used to assess test stabilityfromY1toY2andY1toY3ofthe CARE Consortium. Dependent variables included verbal memory, visual memory, motor speed, and reaction time. Pearson product moment correlations (r) and intraclass correlation coefficients (ICC3,1) were used to measure the stability from Y1 to Y2 and Y1 to Y3. ICC values range from 0.0 to 1.0 with higher values representing more stable performance. Test-retest reliability strength was interpreted as poor (<0.50), moderate (0.50-0.75), good (>0.75-0.90), and excellent (>0.90). Reliable change indices (RCI) using 95% confidence intervals were calculated to determine the percentage of cadets that showed significant change between baselines. Results: Motor speed (r = 0.663, ICC = 0.662, RCI = 9.55) was the most consistent ImPACT measure from Y1 to Y2 followed by visual memory (r = 0.517, ICC = 0.517, RCI = 22.9), reaction time (r = 0.433, ICC = 0.422, RCI = 0.19), and verbal memory (r = 0.418, ICC = 0.418, RCI = 19.49). Across composite scores, the percentage of cadets who exceeded the RCIs was small (2.35% Declined and 2.48% Improved). From Y1 to Y3, motor speed (r = 0.590, ICC = 0.590, RCI = 10.52) remained the most consistent followed by reaction time (r = 0.510, ICC = 0.490, RCI = 0.18), visual memory (r = 0.478, ICC = 0.478, RCI = 23.80), and verbal memory (r = 0.448, ICC = 0.448, RCI = 18.98). The percentage of cadets (3.21% Declined and 3.85% Improved) that exceeded the RCI values from Y1 to Y3 increased slightly. Conclusions: ImPACT scores demonstrated less than optimal stability from yearto-year in uninjured participants with the strength of the values ranging from weak to moderate. These results suggest that baseline cognitive performance as measured by the ImPACT may fluctuate over a 3-year period but have a nominal effect.