There is a need to better understand the threshold for paediatric skull fracture. Historically, drop tests of cadaver heads have been used to estimate thresholds. However, societal and ethical considerations prevent tests with child cadavers. Therefore, researchers have attempted to estimate the threshold by scaling adult tolerance data. Recent work suggests that mass and material scaling of adult tolerance data is insufficient to develop robust estimates of paediatric skull fracture tolerance. Researchers have also attempted to develop finite element models to estimate skull fracture tolerance. These models require that both geometry and material properties of the skull and brain be known, but, while detailed geometry is known, lack of experimental material test data prevents development of reliable finite element models. This paper describes development of a method to estimate skull fracture tolerance using fall data collected in emergency room. The method depends on the observation that width of force versus time pulse recorded in head drop tests onto a given surface does not vary with height of fall or the mass of head. This observation is supported by analyzing data from adult cadaver head drop tests onto 50mm thick 90 Shore D, and 40 Shore D rubber pads. Next, data from neonatal head drop tests are used to estimate pulse width when an infant head is dropped onto a steel plate. This pulse width is used together with child fall data collected at the Children’s Hospital, Milwaukee to estimate forces needed to cause a simple linear fracture in an infant head. This paper describes the procedures used to obtain anthropometric data and fall data such as height of fall and type of surface that the head contacted. This physics based method can be used to analyse child fall data relatively easily to obtain robust estimates of child skull fracture.