This study aimed to test the validity of a non-motorised treadmill (NMT) for the measurement of self-paced overground endurance running performance. Ten male runners performed randomised 5-km running time trials on a NMT and an outdoor athletics track. A range of physiological and perceptual responses was measured, and foot strike was classified subjectively. Performance time was strongly correlated (r = 0.82, ICC = 0.86) between running modes, despite running time being significantly longer on the NMT (1264 ± 124 s vs. 1536 ± 130 s for overground and NMT, respectively; P < 0.001). End blood lactate concentration and rating of perceived exertion were significantly higher on the NMT compared to overground. Integrated electromyography was significantly lower on the NMT for three muscles (P < 0.05), and mean stride rate was also significantly lower on the NMT (P = 0.04). Cardiorespiratory responses of heart rate, oxygen uptake and expired air volume demonstrated strong correlations (r = 0.68-0.96, ICC = 0.75-0.97) and no statistical differences (P > 0.05). Runners were consistently slower on the NMT, and as such it should not be used to measure performance over a specific distance. However, the strong correlations suggest that superior overground performance was reflected in relative terms on the NMT, and therefore, it is a valid tool for the assessment of endurance running performance in the laboratory.