Those responsible for designing cars to meet the European Side Impact Test Procedure are tending to find it more difficult than they originally expected. The conventional wisdom has been that strengthening the car and providing padding should improve protection. However, many are finding that attempts to meet the test requirements, by reducing door intrusion or intrusion velocity, are proving ineffective. Extensive research carried out at TRL has helped to explain why strengthening the car side is ineffective. It has also given an insight into the side impact injury process and provided design guidelines which can be used to ensure that cars meet the test requirements.The difficulties experienced by car designers may explain why alternative test procedures have been suggested and proposals have been made which would reduce the severity and effectiveness of the full scale test procedure. This paper reports on the extensive research carried out by TRL, explains why the conventional wisdom was flawed, details the design guidelines and provides reasons why the beneficial use of quasi-static and component testing of doors is likely to be limited.