Objective To investigate the effect of stage duration on peak physiological response to arm ergometry incremental exercise testing in individuals with spinal cord injury or disease (SCI/D) in support of developing a more individualized approach to testing. Design Parallel reliability study. Setting University research laboratory. Participants Individuals with SCI/D (N=38) performed 2 peak oxygen consumption assessments comprised of 1-minute or 3-minute stages on 1 day. Interventions Not applicable. Main Outcome Measures Heart rate and oxygen consumption were recorded continuously from baseline through recovery. Every 1 minute or 3 minutes, resistance increased at an individualized level, and participants continued until volitional fatigue. The highest 30-second average was selected as peak oxygen consumption (VO2peak) and peak heart rate. The highest workload achieved for ≥30 seconds was recorded as peak power output (PO). Results Intraclass correlation coefficients (ICCs) were >.95 for the following variables: absolute VO2peak (mean ± SD, 1min: 1.2±0.7, 3min: 1.2±.62L/min; 95% CI,.91–.98), relative VO2peak (mean ± SD, 1min: 16.6±9.4, 3min: 15.8±8.0mL/kg/min; 95% CI,.90–.98), absolute peak PO (mean ± SD, 1min: 61±41, 3min: 66±41W; 95% CI,.92–.98), and relative peak PO (mean ± SD, 1min:.80±56, 3min:.90±.55W/kg; 95% CI,.91–.98). There was no significant difference in peak heart rate between 3 minutes (132±27 beats per minute [bpm]) and 1 minute (130±28 bpm); however, the ICC lower limit was <.75 (0.67–0.90). Total test time was significantly longer in the 3-minute (8:17±2:52min:s) than in the 1-minute (6:29±2:21min:s) protocol. Conclusions For measuring peak physiological values in an incremental arm ergometry test, stage duration is of little consequence because similar values were obtained for the main outcome measures in both protocols.