Mean human response curves and associated biomechanical response corridors are commonly developed from human subject test data to guide the design of anthropomorphic test devices (ATDs) by providing “target” biomechanical responses to impact. Since differences in anthropometry and physical characteristics within a group of human test subjects can result in widely varying response data, the first step in developing target biomechanical responses is typically to normalize the responses to a certain “standard” anthropometry representing the dummy to be designed or evaluated. The normalization procedure should remove variation associated with anthropometric differences, thus collapsing the group of curves so that a single mean response can be more accurately established that characterizes the human response of the “standard” anthropometry. Several methods for normalizing human subject test data can be found in the literature, but there is no consensus as to which is the most effective. In this study, the two most common existing normalization techniques, as well as some newly developed methodologies, were evaluated by applying them to both a side impact PMHS sled test data set, and a component-level PMHS thoracic pendulum impact data set. The efficacy of the normalization techniques was assessed for each group of common signals by calculating the cumulative percent coefficient of variation (%CV) for time- history curves, and the cumulative ellipse error for two-dimensional force-deflection curves. Both of these measures provide a quantifiable assessment of the similitude of the group of curves (i.e., the normalization technique resulting in the lowest cumulative %CV value or cumulative ellipse error most effectively collapses the curves). The normalization technique found to consistently perform the best is a newly developed extension of impulse momentum-based normalization in which the stiffness ratio was determined from effective stiffness values calculated from the test data, rather than using characteristic lengths. Utilization of this normalization methodology in the development of mean human response curves should prove useful in more accurately characterizing the target human response to aid in the design of more biofidelic dummies.