Applying regression analysis to model the risk of space flight and terrestrial activities

Abstract

This article presents a novel method for coarsely modeling space flight risk in the absence of vehicle-specific data. Risk and usage rates for several different modes of transportation (including space flight) and adventure sports activities (mountaineering, skydiving, and SCUBA diving) were correlated, and a line of best fit equation was derived. The strong, inverse correlation between number of fatal accidents per trip (i.e. risk) and number of trips per year (i.e. usage) (r = −0.90, p < 0.01), and the strong correlation between number of fatalities per participant and number of participants per year (r = −0.93), suggest that risk and usage may be inherently correlative, even across distinct modalities. As such, this quantitative relationship can be used to supplement traditional analysis techniques and serve as a sanity check for expert opinion—particularly during the early stages of vehicle development, when quantitative data is limited and cannot readily support alternative risk prediction techniques. In addition, this general relationship can provide an additional benchmark for tracking performance throughout the operational lifetime of a program, and offers a unique perspective for comparing the relative risk of spaceflight to more commonly experienced terrestrial activities.