Risk Tree Analysis
The paper focuses on criticizing the risk analysis method used in Mok and Savage (2005) article “Why has safety improved at rail- highway grade crossing”. The risk cause analysis process in the article uses a negative binomial regression model for analyzing causes of increased safety at public railway crossing. Negative binomial regression analysis is used for the analysis of the number of collisions between trains and motor vehicle incidents, and for the number of deaths occurring in these incidents. The aim of the analysis was to identify the correct decisions implemented for reduction of the number of accidents (Mok & Savage, 2005).
The sample was taken to 49 states, and for the years between 1975 and 2001. This provides sufficient data for the analysis. Therefore, it can be regarded as a representative of the incidents and death trend in the 49 states. The Negative binomial regression method is justified by the high dispersion of the data used. In addition, the data for the period between the years 1975-2001 was not updated in a regular manner, and could not form an historical data. Also, the measures were not implemented simultaneously; thus, the data had greater variation. Therefore, the choice of the method is not appropriate. Alternatively, fault tree analysis method should have been used in the study. This is because it would allow the use of qualitative data in the analysis of possible measures taken, alongside quantified probabilities of success of the measures as calculated in the study (Mok & Savage, 2005).
Lastly, the article does not provide recommendations on the best measures that were implemented to reduce the number of incidents and fatalities, but rather, estimates the benefits and cost of the implemented measures. In this regard, a solid recommendation to the decision that needs to be implemented is required. A tree risk analysis provides a step by step process from the head event to the last event. This provides an overview of events from which decision can be made.
In conclusion, the fault tree risk analysis method is more suitable for analyzing the causes of accidents compared to other methods, such as a negative binomial regression method.
Mok, & Savage (2005). Why has safety improved at rail highway crossing grading? Risk
Analysis, 25 (4), 867-880.