Gait impairments have been documented following sport-related concussion. Whether preexisting gait pattern differences exist among athletes who participate in different sport classifications, however, remains unclear. Dual-task gait examinations probe the simultaneous performance of everyday tasks (ie, walking and thinking), and can quantify gait performance using inertial sensors. The purpose of this study was to compare the single-task and dual-task gait performance of collision/contact and noncontact athletes. A group of collegiate athletes (n = 265) were tested before their season at 3 institutions (mean age= 19.1 ± 1.1 years). All participants stood still (single-task standing) and walked while simultaneously completing a cognitive test (dual-task gait), and completed walking trials without the cognitive test (single-task gait). Spatial-temporal gait parameters were compared between collision/contact and noncontact athletes using MANCOVAs; cognitive task performance was compared using ANCOVAs. No significant single-task or dual-task gait differences were found between collision/contact and noncontact athletes. Noncontact athletes demonstrated higher cognitive task accuracy during single-task standing (P = .001) and dual-task gait conditions (P = .02) than collision/contact athletes. These data demonstrate the utility of a dual-task gait assessment outside of a laboratory and suggest that preinjury cognitive task performance during dual-tasks may differ between athletes of different sport classifications.