mryan75 said:
Here's an even more fascinating study:
...
So 92% of the AMEs surveyed had 1 or 0 fatalities over a 10-year period, and one guy had 25. He either had some seriously bad luck, was someone who dealt with extremely tricky cases (and should promptly be removed from such duties), or he was the AME that all the pilots with otherwise disqualifying conditions could go to and get a medical with a wink and a nod.
Interesting study yes. But fairly peripheral to the proposition being discussed. Have a look at figure 2 in that report. It shows there is a distribution of fatality crash rates amongst AMEs with outliers on both high and low crash rates. There is always going to be some AME who is the most outlying. But there is a distribution there so this is not just a single datapoint on the more fatalities side.
What the data in this paper showed is that there is such a distribution in the likelihood of fatalities by AME. I would not find that result surprising. People differ in the care and detail with which they apply a set of rules. Pulling out the single datapoint at the tail on one side and contrasting it with the middle range only, omitting the relatively continuous nature of this distribution is a form of fairly selective citation to the data.
Quite a speculative leap to assert that this means both that pilots with disqualifying medical conditions are going to a specific subset of AMEs with looser criteria AND that those disqualifying conditions then resulted in the fatalities.
I truly do applaud taking the time to do the searches and find some of the data which may bear on this questions. Many kudos. Thanks for pulling them out.
Interpreting such data with some level of scientific objectivity often takes a fair amount of time and an attempt to consider all the data and all of the paper in their appropriate balance. I think these deserve more scrutiny and thought. Based on a brief review, I certainly would not conclude they demonstrate that FAA medical certification improves the safety of flight.
The authors themselves appear to share some of this reservation. While the discussion in this paper is fairly long and sort of mushes around on both sides of the issue, it does state "However, further study is needed to determine what role, if any, the medical certification process played. The procedure used in this study was able to identify a small group of AMEs with atypical high fatality rate per exam; however, the significance is yet to be determined." This strongly suggests the authors themselves do not think that ""the answer is patently obvious".
If you want to continue the discussion of objective interpretation of these papers, I am happy to do so and explore how they bear on the proposition being debated here. However, I will ask that the personal attacks and assumptions of some sort of malicious motive come to an end. May I suggest another way of looking at it? That when people with different perspectives look at data it can lead to honest differences of opinion about the meaning of that data. The best and most useful way for such people to have a discussion is to focus on the data and the arguments themselves, leaving aside any differences in personal characteristics. I appreciate the effort to bring up these other datasets and interpretations. I hope they can continue in a civil manner.