Carnage in the beginning…
In a recent post I bemoaned the fact that the fatal accident rate for private flying had gone up to 1.40 per 100,000 hours after remaining level in the 1.20 range for almost 20 years. Guess what it was when Air Facts started in 1938? Would you believe 16.6, or, a fatal accident about every 6,000 hours. In another place I saw it as high as 30.0. At the time, the airline rate was 2.0. The trash media would have a field day if it were that today.
It is interesting that the airline rate in 1938 was a bit worse than private flying is today. At that time the airlines were flying mostly no-tech DC-3s which had about the same performance as high-performance singles and light twins. They flew in most all weather, probably more than they fly in today, and did so over a lot of really rough terrain at night as well as day. The only advantage they had was a crew of two but in those days the co-pilot, as he was called, was usually told to shut up and not touch anything.
A reader commented that in the previous post I hadn’t given credit for all the improvement in private aviation safety prior to the last 20 years. Well, I just did. Another reader emailed that the current lethargy in general aviation has little to do with cost and the other familiar whipping boys but it has a lot to do with the public perception of danger in private flying.
Air Facts was all about safety from the start and it has always tended that subject. How did we evolve from that bloody 16.6 or worse in 1938 to the 1.2-1.4 range where it has been for the last 20 years?
It was slow going and has more to do with pilots than technology.
The accident rate has always varied pretty widely among airplane types. I have always thought that was true because different airplanes attract different pilots.
After World War Two the majority of the pilot population was either ex-military or trained under the GI-Bill. Where in 1938 there had been only about 9,000 “sportsman” pilots, as they were called then, the number moved quickly into the hundreds of thousands after the war.
Pilots of that day weren’t much interested in safety. It was a fatalistic group that had been to war and back and part of the romance of flying came from risking your life. Safety features in airplanes, or in cars, were far in the future and the public was not clamoring for such. “Cheated death again” was a pretty standard post-flight remark.
The Civil Aeronautics Board (forerunner of the NTSB in this regard) and Beechcraft both did a study of accidents in 1952. It gave a good picture of what was going on with the pilot population while flying airplanes built after the war. The overall private aviation fatal accident rate per 100,000 hours was at about 4.8 in 1952.
Think of these numbers in relation to the current 1.2-1.4 rate. The CAB and Beech numbers were pretty close and these are the CAB numbers.
The original Bonanza 35 had a fatal accident rate of 4.9 where the A, B and C35 had a rate of 2.5. Why were the newer airplanes better? I think it was because pilots were becoming more accustomed to what a bad idea it was to lose control of such an aerodynamically clean airplane.
I have flown all models of the V-tail Bonanza and can see no other reason why the newer ones had a better record. It was simply a matter of better pilots or at least pilots who were more aware of the characteristics of the airplane. Even though the Bonanza was frequently referred to as a V-tail doctor killer, the 35 was average and the newer ones above average.
The Model 18 twin Beech had the best record in 1952 at .90. It was usually flown by professional pilots who were simply about five times better than average at the art of flying.
The Cessna 170 and 170A came in at 4.8, or right at average. Today’s version of that airplane, the Skyhawk, always has one of the best records, substantially better than average, so why wasn’t that true in 1952? The only answer I can think of relates to the fact that the airplanes were largely owner flown then and were widely used for travel and, when the weather was marginal, for scud-running. Today, the Skyhawk is flown a lot for instructional purposes where the safety record is excellent, and by more sedate pilots for private flying.
Two-place airplanes were a large part of the fleet and the airplane that was designed to be safer (because it was stall-resistant and spin proof and had no rudder pedals to fool with), the Ercoupe, had the worst record at 10.4 fatal accidents per 100,000 hours. Shades of 1938. The Cessna 120/140 had the best two-place record at 3.0 and all the rest fell somewhere in between. The Taylorcraft was definitely on the high side at 8.7. I have instructed in both 140s and T-crafts and it is a mystery to me why the T-craft would have a record almost three times as bad as the 140.
Anecdotally, a jeweler who based a Taylorcraft at the airport where I worked in 1952 would fly his airplane in and out of some farm fields that were more suitable for helicopters. We were all convinced he would come to grief. He didn’t but an Ercoupe owner at the same airport did.
Why was the Ercoupe so bad? It could only be charged to human nature. Pilots thought the airplanes were “safe” where in truth Ercoupes were every bit as dangerous as any other airplane. Maybe they resisted stalls and wouldn’t spin but if flown too slowly they would develop a high sink rate, especially if there was even a little wind shear, and they could and did hit the gr
ound hard and nose down. There was a special certificate for two-control airplanes that required less training. That was a sad and tragic mistake.
The next time I sorted out accident rates by type was in the late 1970s. That was at the peak of the best-ever private aviation boom, when all the World War Two folks were in their peak earning years.
By this time the fatal accident rate was down to 2.02. Did the fact that it was twice as favorable as in 1952 have anything to do with the booming sales? Nobody will ever know, but by this time the average age of pilots had reached a much more mature number and pilots might have started giving more thought to safety. By the late 1970s, more attention was being paid to crashworthiness and shoulder belts had come to new general aviation airplanes.
Against that 2.2 rate, how did some of the airplanes of the day stack up? An NTSB study pegged the Bellanca as the worst, at 5.68. The best was the Piper Navajo at 1.13.
Something that had been pointed out in Air Facts years before was verified in this NTSB study: when high performance singles other than the Bellanca were compared with similar light twins, Bonanzas v. Barons for example, the rates were similar and most fell just above the 2.2 that was average for the time.
A lot of thought was given to that Bellanca record because there was no valid reason for it to be so high. It could have been an aberration caused by a relatively small fleet size and inaccurate estimate of hours flown. Also, when the safety record was calculated in other ways the Bellanca came out much closer to the other retractables.
The fixed-gear singles all came in under the average 2.2 number with the Cessna 150 best at 1.34, the 172 at 1.47 and the Cherokee at 1.97.
The stall/spin is a leading killer on these simple airplanes. The Cherokee has the tamest stall characteristics of the bunch, especially when the older 150/172 airplanes are considered, yet it had a worse record. What was the story there?
One semi-explanation was/is the fuel system. In a Cherokee you select between wing tanks. Both Cessnas can or do draw from both tanks at once and no fuel system management is required. And, you guessed right, the Cherokees had a bunch a fuel system mismanagement accidents where the Cessnas had none. Other than that, the conclusion might be that more docile stall characteristics don’t mean that an airplane will have a lower accident rate.
The most recent time that I delved into this was in the early 2000s, when the fatal accident rate had settled down to the 1.2 per 100,000 hours rate. Much of the following is from research I did for the book “The Next Hour.”
Against the average, higher performance airplanes again did not fare as well as lower performance airplanes. The worst record was in the Cessna P210 at 2.33 followed by the Piper Malibu/Mirage at 2.04. Make it a simpler retractable, without pressurization, and the Saratoga, 210 and Mooneys were all close to the average. The Bonanza A36 was a little high at 1.81. The Cirrus is not a retractable but has similar performance and the SR20/22 was slightly above average at 1.52.
The Cessna 172 was best at .56, the 182 next at .74 and the Piper Archer below average at 1.06.
More aggressive pilots fly more aggressive airplanes and this is clearly illustrated in the Cirrus record of that time. The SR22 is clearly the more aggressive of the two Cirrus airplanes and it appeared to have a four-times greater involvement than the SR20 in fatal accidents.
The economy took a humongous hit right after those numbers were developed and the price of 100LL went through the roof. Flying activity went way down. Most of the airplanes in the fleet have gotten a lot of years older, too, and older airplanes tend to fly less. Thus any current calculation of hours flown by type would have to be taken with so many grains of salt that it would not be worth much.
It is my opinion that nothing has changed much since that last calculation. I think the Cirrus is likely doing better, maybe even substantially better, because of educational programs that include more aggressive use of the airframe parachute.
Whenever the overall private aviation accident rate is mentioned there was always a chorus of “yes, but” comments. A primary one is about the fact that accidents in experimental airplanes count here and the record there is worse. We are still all members of the same community and experimental airplanes and pilots are an integral and important part of that community. They are simply part of what we do so they count.
The NTSB has in the past shown rates for various fleets of airplanes. In that early-2000s period I was just talking about they showed fatal accident rates per 100,000 hours as follows: experimental 4.65, single-engine piston 1.50, piston twin 1.95, turboprop .69 and jet .24. Most experimental airplanes are single-engine piston types so those numbers could be used to suggest that singles are a lot safer than twins.
With the accident rate all but stagnant for the least twenty years, and with more high-tech innovation during that period than any other in our history, I can only conclude that there is no safety advantage to all those wonderful gadgets and gizmos. I always had the latest and greatest in my airplane and I loved having them. I never deluded myself into thinking they reduced my personal risk except in one area, collision avoidance. Having traffic information on the panel was, to me, pretty wonderful. Maybe my thinking on that is skewed by the fact that the only accident I had in 57 years of pilot-in-command flying was a midair collision.
If anything stands out as a growing problem today, it is piston engine failures, and my thoughts on this are anecdotal. I look at the press reports of airplane accidents every day and the number of airplanes damaged or destroyed in forced landings is almost mind-boggling. Where the old airframes are doing okay, the old engines might be giving up the ghost too frequently. This is something that needs to be watched closely, to see if there is a disturbing trend here.
If I could pick one factor that I think resulted in many years of improvement and now almost 20 years of no further improvement, it would be the nature of pilots. We became steadily more risk-averse over the years and then that got as good as it was going to get and no other factor had a measurable influence on safety
In the end, there’s only one question that really counts: “How safe is my flying?” Those of us who work in private aviation safety might not be able to do anything to further improve the overall accident rate but we might be able to help individuals come up with the best possible answer to that important question. So we will keep trying.