Restraining children in cars is essential if their safety is to be increased. The experience with child restraint systems (CRSs) is positive, but improvements in system characteristics, handling and acceptance are necessary. Dynamic trolley tests have been carried out with different kinds of CRSs on the basis of previously published injury patterns in children. These tests showed that in addition to the force loading on children caused by the type of CR5 the major problem that exists is connected with the CRS’s characteristics for fastening it to the car. The relative belt geometry of the cars and problems of slack may often lead to unfavourable movements of the CR8, and these were observed in the tests. Some of the conditions of the compliance testing according to ECE-R 44 no longer represent actual real-life conditions. Proposals for improvement are given. Problems of misuse and CR8 handling and acceptance are discussed. These findings are based on extensive interviews with 1,282 parents on problems they had with CRSs. We received information on a total of 1,903 products. The results are subdivided into 9 different types of CRSs. Proposals for future CRS development which go beyond purely technical safety measures are also derived from this investigation.