Wearable hip protectors represent a promising strategy for preventing hip fractures. However, there is lack of agreement on biomechanical testing standards and subsequent uncertainty about the ability of hip protectors to attenuate impact force during a fall. To address this issue, we designed a fall impact simulator that incorporated a "biofidelic" surrogate pelvis, which matched the surface geometry and soft tissue stiffness measured in elderly women (n=15). We then used this system to measure the attenuation in peak femoral neck force provided by two commercially available soft shell protectors (Safehip Soft and Hipsaver) and one rigid shell protector (Safehip Classic). Finally, we examined how the force attenuation provided by each protector was influenced by systematic changes in fall severity (impact velocity), body size (pelvis size), and soft tissue stiffness. With the biofidelic pelvis, the force attenuation averaged over all impact velocities was 27% for Safehip Soft, 17% for Safehip Classic, and 19% for Hipsaver. However, the rank order of hip protectors (and especially the performance of Safehip Classic) varied with the test conditions. Safehip Classic attenuated force by 33% during a low velocity (1 ms) fall, but only by 8% for a high velocity (4 ms) fall. In the latter condition, improved attenuation was provided by the soft shell hip protectors (19% by Safehip Soft and 21% by Hipsaver). As soft tissue stiffness increased from softest to most rigid, the attenuation provided by Safehip Classic increased 2.9-fold (from 26% to 76%), while Safehip Soft increased 1.7-fold (from 36% to 60%) and Hipsaver increased 1.1-fold (from 36% to 38%). As pelvis size decreased from largest to smallest, the attenuation provided by Safehip Classic increased 8-fold, but for a high velocity fall and moderate tissue stiffness, never exceeded that provided by Safehip Soft and Hipsaver. Our results indicate that, under biofidelic testing conditions, the soft shell hip protectors we examined generally provided greater force attenuation (averaging up to 27%) than the hard shell protector. Measured values of force attenuation were highly sensitive to variations in impact velocity, pelvic size, and pelvic soft tissue stiffness. This indicates the need to develop international testing standards to guide market approval, the selection of protectors for clinical trials, and the design of improved hip protectors.