Originally Posted by
megan
According to the report at the link:
Mid Air Collision Rate per 100,000 in US GA in 1981-1990: 0.067. Australia: 0.049
Mid Air Collision Rate per 100,000 in US GA in 1991-2003: 0.047 Australia: 0.048
The report also says:
The overall midair collision accident rate [in GA in Australia] was higher in the period 1961-1980 (0.074 per 100,000 flight hours) compared with the subsequent period 1981-2003 (0.048 per 100,000 flight hours), but the difference was not statistically significant.
If a difference between 0.074 and 0.048 per 100,000 flight hours is “not statistically significant”, good luck in spinning the difference between 0.047 (Australian GA 1991-2003) and 0.048 (US GA 1991-2003).
Overall, for the period 1981-2003, the US had a lower number of hours flown per collision (930,000 hours) for general aviation aircraft relative to Australia (1,260,000 hours), indicating that general aviation in the US may have had a slightly higher risk for midair collisions. However, this difference was not statistically significant.
Quite so. One wonders, then, why the ATSB nonetheless gratuitously speculated on what “may” be “indicated” by differences that are not statistically significant.
Unfortunately, there were no figures available for the number of hours flown or number of movements in or near the circuit area versus away from the circuit area for either Australia or the US. Based on the assumption that the proportion of flying in or near the circuit area was the same in Australia and the US, it was found that:
• Australia had a general aviation midair collision accident rate 1.1 times higher than that of the US in or near the circuit area during the period 1981-2003. This comparison excluded collisions involving aircraft deliberately flying close to each other. This difference was not statistically significant.
• The US had a general aviation midair collision accident rate 5.8 times higher than that of Australia away from the circuit area during the period 1981-2003. This comparison excluded collisions involving aircraft deliberately flying close to each other. The test of the difference between the two countries was not significant. However, this test had a low level of power to detect differences in the rates as the Australian rate was based on only one event (Moreton Bay, 1988). The fact that only one event also occurred during the period 1961-1980 (Cecil Plains, 1980) suggests that the Australian rate may be reasonably reliable. There was a significant difference in the hours flown per collision involving general aviation between the two countries, indicating that there was a higher collision risk in the US away from the circuit area relative to Australia.
“5.8 times higher”? Yeah right. Best to revisit the validity of that assumption.