According to the National Highway Traffic Safety Administration, over 6 million traffic accidents claim more than 41,000 lives each year in the United States. Causes of accidents and related injury severity are of special interest to traffic-safety researchers. Such research is aimed not only at reducing the number of accidents but also the severity of injury. One way to accomplish the latter is to identify the most profound factors that affect injury severity. Understanding the circumstances under which drivers and passengers are more likely to be severely injured (or killed) in an automobile accident can help improve the overall driving safety situation. Factors that potentially elevate the risk of injury severity of vehicle occupants in the event of an automotive accident include demographic and/or behavioral characteristics of the person (e.g., age, gender, seatbelt usage, use of drugs or alcohol while driving), environmental factors and/or roadway conditions at the time of the accident (e.g., surface conditions, weather or light conditions, the direction of impact, vehicle orientation in the crash, occurrence of a rollover), as well as technical characteristics of the vehicle itself (e.g., vehicle’s age, body type).

Questions for Discussion
1. How does sensitivity analysis shed light on the black box (i.e., neural networks)?
2. Why would someone choose to use a blackbox tool like neural networks over theoretically sound, mostly transparent statistical tools like logistic regression?
3. In this case, how did neural networks and sensitivity analysis help identify injury-severity factors in traffic accidents?