The goal of this study was to assess the sensitivity of (corrective) driving inputs on the risk of rollover of a sedan and the resulting touchdown conditions in case of soil trip rollover crashes. The driving inputs include the initial steer that leads to the departure from the roadway, the corrective steer in an attempt to gain back the control and the initial travel speed of the vehicle. Latin hypercube sampling was used to uniformly sample various combinations of driving inputs and corresponding simulations were run using a multibody model of a sedan validated for aggressive driving maneuvers and quasi‐static suspension tests. Logistic regression model was fit to predict the probability of the binary outcome of rollover with the driving input as predictor variables. The model involving interaction between the predictor variables had a better predictive capability than the main effects model. The initial travel speed and the first steer angle were the most influential in affecting the risk of rollover of the vehicle. The touchdown parameters varied depending on the peak lateral acceleration at trip and the trip location from the road. The peak lateral acceleration and trip location in turn varied depending on the driving inputs. The study established a methodology to estimate the sensitivity of the risk of rollover of a sedan and distribution of corresponding touchdown parameters to driving inputs in case of corrective maneuvers.