Air Safety

Keywords: independent events, binomial probability

Description

Publicity given recently to several serious safety incidents on Qantas flights has lead to a renewed interest in air safety statistics. The article Fasten Seatbelts by Neale Maynard on pages 21-22 of the Courier Mail, April 29, 2000, read

According to aircraft manufacturer Boeing, serious accidents are exceedingly rare in the jet age.

"The risk of being involved in a commercial jet aircraft accident where there are multiple fatalities is approximately one in 3 million. To put this in perspective, you'd have to fly once every day for more than 8200 years to accumulate 3 million flights," Boeing says.

University of Queensland statistician and senior lecturer in mathematics, Dr Gordon Smyth, says that based on Boeing's figures, a passenger would have a 37 percent chance of going for 3 million flights without being involved in a multiple fatality accident.

"That means the probability is slightly more than 50:50 that you would have an accident in 3 million flights," Smyth says.

Increase your travel tally to 6 million flights and Smyth says the odds of avoiding a multiple fatality crash worsen to 14 percent, and fall to just 5 percent for 9 million flights.

"You've got a smaller and smaller probability of going more and more flights without an accident," Smyth says, conceding that increased safety is a variable which will improve a passenger's odds.

He says that if Boeing's figures are translated to a more achievable number of flights, such as the number of flights taken by a flight attendant flying four times a day for 230 days a year over 30 years, the probability of being involved in a fatal crash would be 1 percent.

But he points out that just because the statistics suggest the likelihood of an accident within a certain number of flights, that does not mean that an accident will occur.

On the Qantas issue, Smyth says that establishing whether the airline is now statistically more likely to face a serious accident than it was previously would be a major statistical exercise and, in any event, chance could also play a part.

The Rome and Bangkok incidents, could, potentially, be a chance cluster, he explains.

"It's always possible to get more accidents or fewer accidents, regardless of what the safety standards are," Smyth says.