Many studies have reported multiple rib fractures sustained by an Out-of-Position (OOP) driver subjected to a frontal airbag deployment, but the injury mechanisms and thresholds remain unclear. Two successive phases occur during the bag deployment:​ punch-out loading of the thorax, followed by a membrane effect ( Horsch et al. 1990 ). The aim of this study was to investigate the thoracic injuries generated by each phase separately. Tests of nine post-mortem human surrogates (PMHS) were carried out on a static test bench using a driver side airbag module described by Petit et al. (2003) . The steering wheel was replaced by a plate in order to increase the loading generated by the airbag. Three loading configurations were performed:​ membrane only, punch-out only, and both types combined. The membrane-only tests were performed with the thorax initially positioned at 13, 78 and 128 mm from the plate in order to vary the load magnitude. The punch-out and the combined tests were performed with the thorax initially 8 mm from the module. Accelerometers and angular rate sensors were fixed on the sternum and on the first, fourth, and eighth thoracic vertebrae of the PMHS. Ribs 2 to 6 were instrumented with strain gauges. The reaction force of the bag on the plate was measured using four 2-axis load cells. Results show that both pure punch-out and pure membrane loading can result in thoracic injuries. However, the rib fracture locations seemed to differ from one type of loading to the other. Moreover, for the same initial distance between the airbag module and the thorax, the injuries were more severe in the combined effect tests than in the pure punch-out or pure membrane. The nine PMHS tests formed a matrix allowing validation of a finite element model and further analysis of thoracic injury mechanisms and thresholds.