The effects of three independent variables on eight measures of driving simulator discomfort were investigated using a high-fidelity, moving-base driving simulator. The between-subjects simulator variables were: (1) simulation of lateral acceleration (LAT)—by lateral translation (standard method) versus by angular rotation; (2) presence or absence of delay in the visual and motion systems (DEL)—nondelayed (normal) versus delayed; and (3) simulator platform (CAB)—open (normal) versus enclosed. Sixty-four subjects were divided into eight groups, each group having equally distributed scores on a test of field independence-dependence. Each group was then assigned to one of the eight simulator conditions. After subjects drove the simulator, a multivariate analysis of variance was performed on the data and resulted in significance for each main effect and the LAT x DEL interaction. Subsequent analyses demonstrated that dependent measures of pallor, skin resistance, respiration rate, yaw deviation, and steering reversals were each reliably sensitive to at least one of the simulator independent variables. It is concluded that future simulator designs should avoid: rotation of the platform to simulate translation, delay in the system dynamics, and complete enclosure of subjects.