AbstractThe aim of this paper is to investigate an iterative statistical procedure, based on a small and censored sample of impact test experiments, useful for interval estimation of head impact safety parameter as critical fall height of protective devices. An adaptive testing routine was developed that was mainly constituted by a series of at least four impact test experiments, followed by the comparison of at least two parameter estimates based on incremental exponential regression fittings and a final confirmation experiment. A total number of 23 protective devices, mainly made of polyethylene foam, were investigated in order to validate the adaptive routine. The routine, applied to critical fall height of protective devices, was 19 times convergent within a maximum of 6 impact test experiments. 4 times the sample was censored because the iterative procedure has exceeded the available number of specimens. Confidence intervals at the 90% level were always less than 0.18 m. The applicability of the adaptive routine was satisfactory demonstrated with reference to devices made of PE-foam and safety threshold of peak acceleration a-max equal to 200 g. The target of a confidence interval below the state-of-art was achieved.