Prediction of rib fractures using computational human body models in side impact remains an elusive challenge. Although cadaveric tests carried out with extensive instrumentation (kinematics, strains) and documentation of the injuries are available, the physical characteristics of the PMHS that need to be included in a computational human body model to improve its biofidelity is still on‐going. Therefore, the goal of this study was to assess the robustness of the principal and longitudinal strains as an indicator of the strain level in the ribs by running a parametric analysis where the posture of the subject and the strain measurement locations were varied. To do so, previously published side impact tests were simulated with THUMS (Total Human Body Models for Safety, v1.4), where three Post Mortem Human Subjects were postured in a nominal driving position before being struck unimpeded on the right side by a massive rigid wall travelling at 4.2 m/s. Nine models of the simulation were created to study the contribution of the posture, of the position of the arm, and of the location of the strain measurement on the ribs. The longitudinal and principal strains were computed from the simulations. It was found that the longitudinal and principal strains could be used interchangeably with THUMS, but that the predicted strains were greatly dependent on the posture and the arm angle.