ObjectiveTo determine whether normalizing spatial-temporal gait data for walking speed obtained from multiple walking trials leads to differences in gait variability parameters associated with a history of falling in people with transtibial amputations. DesignCross-sectional study. SettingRehabilitation center. ParticipantsPeople with unilateral transtibial amputations (N=45; mean age ± SD, 60.5±13.7y; 35 men [78%]) were recruited. InterventionsNot applicable. Main Outcome MeasuresParticipants completed 10 consecutive walking trials using an instrumented walkway system. Primary gait parameters were walking speed and step-length, step-width, step-time, and swing-time variability. A retrospective 12-month fall history was obtained from participants. ResultsSixteen amputees (36%) were classified as fallers. Variation in gait speed across the 10 walking trials was 2.9% (range, 1.1%–12.1%). Variability parameters of normalized gait data were significantly different from variability parameters of nonnormalized data (all P<.01). For nonnormalized data, fallers had greater amputated limb step-time (P=.02), step-length (P=.02), swing-time (P=.05), and step-width (P=.03) variability and nonamputated limb step-length (P=.04) and step-width (P=.01) variability. For normalized data, only 3 variability parameters were significantly greater for fallers. These were amputated limb step-time (P=.05), step-length (P=.02), and step-width (P=.01) variability. ConclusionsNormalizing spatial-temporal gait data for walking speed before calculating gait variability parameters may aid in discerning variability parameters related to falls histories in people with transtibial amputations. This may help focus on the initial rehabilitation efforts of amputees with a fall history.