Modelling injury severity of helicopter accident: do contributing factors have heterogeneous effects?

It is crucial to understand the mechanism of injury and fatality of helicopter accidents. However, the explanations of risk factors may be affected by unobserved heterogeneity due to failures in collecting critical features in accident investigations. This study models the severity of 1,241 helicopter accidents from the National Transportation Safety Board (NTSB) database. Four types of random parameter logit models are used to capture unobserved heterogeneity. The results indicate that factors have different types but slight heterogeneous effects in affecting fatal accidents, while no heterogeneous effects are found in modelling injury accidents. The random parameter logit with heterogeneity in means (RPLHM) model achieves the best data-fitting. Accident features associated with fatal accidents are analyzed, including the instrument meteorological conditions, loss of control in flight, low-altitude operation event/collision, etc. Analysis also indicates that the active emergency locator transmitter, reciprocating engine(s), helicopters with rated power of 320-450 horsepower, and aircraft fire/explosion have heterogeneous effects.