Ejection during a vehicle crash increases fatalities and injuries. A recent analysis of crash data focused on the factors influencing ejection in a fatal crash. Factors studied included safety belt use, average speed, driver age, percentage of light trucks involved in crashes, and rollover.

John Winnicki reported on the results of this study in "Analysis of Ejection in Fatal Crashes," a National Highway Traffic Safety Administration (NHTSA) Technical Report, published in November 1997.

Data Summary

The study examined data collected by the Fatality Analysis Reporting System (FARS) database from 1982 to 1996. The research included fatally injured drivers and passengers over age 12. The variables studied showed that:

The use of safety belts was the most significant indicator of ejection in a crash. Among belted crash victims in the database, 2.49 percent were ejected. Among unbelted victims, however, 29.43 percent were ejected. Table 1 compares the ejection rates in fatal crashes of belted and unbelted occupants.

Rollover dominated the variables for both unbelted and belted ejections. The most common related variable in rollovers was traveling speed, followed by vehicle type (light truck), and collision with another vehicle. In non-rollover ejections, the most common related variable was the driver's age, followed by vehicle type.

The differences between the rollover and non-rollover ejections also included:

• Weather, which was a more significant factor in rollovers. These ejections occurred .607 times less during wet weather, possibly due to lowered speeds.
• Time of day was more significant in rollovers but showed no significance in non-rollovers. Ejections occurred in rollovers less frequently at night, possibly due to lower speeds.
• Speeding was a more significant factor in the non-rollover ejections.
• Driver's sex was a more significant factor in rollovers, with 20 percent more ejections for female drivers.

The data showed that rollover crashes ejected 50.72 percent of the occupants, while non-rollover crashes ejected only 11.12 percent. Speed was the most significant factor among rollover ejections. Younger drivers were more likely to be involved in rollover ejection crashes. Table 2 shows the ejection rates in fatal crashes for occupants of vehicles that rolled and vehicles that did not roll.

The data also demonstrated a slight increase over time in the prevalence of ejections in fatal crashes. As a cause, the researcher suggested changes in vehicle crashworthiness or in human behaviors--such as more aggressive driving or risk-taking.

Analysis of Trends in Ejection in Fatal Crashes

The ejection rates for fatally injured people from 1982 to 1996 changed little. During the same time, use of safety belts by fatally injured occupants increased sharply--a 2.5 percent increase per year. This consistent rate of ejection despite increased use of safety belts prompted further study of why ejections did not decrease with the increased use of safety belts. The following graph compares ejection rates among belted and unbelted persons.

A comparison of the unbelted and belted data showed that ejection was rare among people who used their safety belts. Among the belted occupants, ejections occurred more often when using automatic belts. The researcher suggested this may be due to improper use of automatic belts, such as not using the lap belt. Also, automatic belts became more common in the late 1980s, possibly explaining the increase in ejection rates among belted persons between 1988 and 1993.

In both belted and unbelted ejections, rollover was the primary factor in fatal crashes. The data showed little trend in the overall fraction of rollovers. After 1984, however, rollover fatalities for belted occupants slowly and consistently increased. Among unbelted occupants, the data showed an increase of rollover fatalities of approximately .64 percent per year. The researcher found this comparison "consistent with the conjecture that the population that remained unbelted were those at a higher risk of being involved in rollover (and consequently being ejected)."

An increasing ejection rate (1.5 percent per year) among safety belt users suggested other factors might be influencing ejection in rollover crashes. Travel speed was one of the most significant variables. Speed data showed that belted occupants traveled at an average speed of about 46 mph. For unbelted occupants, the traveling speed increased from 51 mph in 1982 to 55 mph in 1996. Those who continued to travel without using their safety belts, in spite of safety belt use laws and public information campaigns, seemed more likely to be involved in more severe crashes at higher speeds.

Because rollover was closely associated with ejection, the average speed in both types of crashes was about the same--with the average speed in ejection fatal crashes tending to increase over time. Speed data indicated that after 1985 the average speed in rollover fatalities for unbelted individuals was 1 to 3 mph faster than among belted persons. The researcher suggested that the large difference in the early years may have been due to low sample sizes and a lower rate of safety belt use during those years.

Lights trucks showed an increase in fatal injuries averaging .7 percent per year. Further analysis of light truck involvement in fatal crashes indicated an increase in safety belt use of more than 23 percent, with a slight decline in ejection rates. Ejection rates continued to be higher for light trucks (at around 35 percent) than for passenger vehicles in general. The rollover rate increased for unbelted light truck occupants but decreased for belted occupants. Average speed in fatal crashes increased among unbelted occupants of light trucks but stayed constant for belted occupants.

Occupant's age was the last variable analyzed. The increase in the average age of fatally injured individuals followed the increase trend in average age of the overall U.S. population. Both unbelted and belted occupants showed an increase in average age, with the unbelted population about 10 years younger. The unbelted data revealed about half the rate of age increase as the general population data. Further data suggested that the average age of fatally injured individuals in ejections and rollovers followed similar patterns, increasing during each year of the study. Fatally injured light truck occupants were younger than the overall population of fatally injured passenger vehicle occupants.

Conclusions