Abstract
This paper presents a quantitative comparison techniquefor evaluating similarities between sets of data resul-ting from a common random process. The normalized inte-gral square error (MISE) criterion is derived using cor-relation techniques and is applied to deceleration-timehistories obtained from six vehicle crash tests in an at-tempt to relate the vehiele response to the crash event.Two sets of data resulting from computer simulations ofvehicle crashes are also analyzed.