2024 Are GA Accidents Rising or Falling?

[@twentyrothmans7308]

Happy New Year, Scott! I hope that the accident rate becomes so low that it's statistically insignificant.

[@tomdchi12]

I'm glad you both pointed out that AOPA should be more clear about their methodology and add information like error bars, but at the same time that their work/report is certainly a useful look at what's going on. My 30,000' view of their report and yours is that they largely agree - the current situation is that there is little improvement in reducing fatal accidents over the last 10 or so years, coming after massive improvements compared with what was going on in the 1970s and earlier. Student pilots are carefully monitored and limited but once they have their "license to learn" I'd expect that we may see accidents increase, unfortunately. I am of this recent pandemic crop, so I am all ears as to what we should be doing better! Thanks! One thing about year-to-year trends - the smaller the number of incidents (fatal accidents) the more noisy that year-to-year trend line will be.

[@adamdrain]

Great to see this video breaking down numbers. It may "seem" like more accidents purely on the concept that in 2025 we can filter out the news and stories we are interested in. If you're interested in aviation, airplanes, accidents etc, then your information streams can be catered to those. Informative as always Scott!

[@zvexevz]

Good idea for a video, and it's great that you went and found the better, more precise data yourself. Sadly those charts leave a whole lot to be desired in terms of clarity. It would help a lot if you labelled your axes, especially when using multiple y-axes in one graph. The graph at 8:35 appears to show over 100,000 accidents due to unlabelled axes. When your x-axis starts at 1, but it's referring to years, specify what is year 0. I'm guessing it was 1990 from contextual clues, but it would be better if they could be understood on their own. Without your script, those charts are nearly inscrutable. But also your dialogue sometimes clashes with the charts, in terms of the colours of various indicators and the variables discussed.

[@FlyWirescottperdue]

I didn't label the y axis because I normalized the data... how do you compare 800,000 of something to a percentage? The curve of the line is the important relationship over time.

[@jamesseddon1604]

Thanks for this Scott. Would you consider uploading a spreadsheet that contains the full data somewhere after your video series is over?

[@FlyWirescottperdue]

All the data I pulled from were previous Nall Reports and Bureau of Transportation Statistics. Available online.

[@gclaughlin1]

I had interaction with McSpadden, bad mojo, me retired F 16 flight test, CFII, MEI. Thank you Scott.

[@n6mz]

9:50 best laugh I've had all week, THANK YOU!!

[@DaddyRecon1]

Excellent review, completely agree, we should peruse the why! Happy New Year Scott

[@dermick]

Looking forward to more videos on this topic - anything we can do to reduce crashes and fatalities will be very welcome!

[@Byzmax]

Happy New Year. Great video, cutting through the cloud of statistics to get to data gathered under VFR so to speak.

[@TheFamilyman7]

Off the cuff, might insurance premiums be a metric of how GA accident rates are trending?

[@joakimlindblom8256]

Excellent review and analysis of the accident data! It's a bit surprising (and disheartening) that the accident rates appear not to have improved much, despite all the new technology enabled flying aids and more broadly available information on accidents and best practices (e.g. KZbin, etc.). Very much looking forward to your next video in this series -- thanks for your great work!

[@abramsalinas1004]

Car accidents are occuring a lot more than aircraft are. Concern for them should be talked about more. There's a lot of planes taking off and landing every minute. The ratio of safe to un safe where safe out paces the unsafe by a lot is higher for SAFE so no big deal. Thanks for the video

[@scottw5315]

Nobody talks about the 40k traffic deaths each ear or the 40k fentanyl deaths either. I've seen attempts at comparing auto/GA death rates. There are lots of variables there. For me, I feel much safer flying than driving. I see reckless driving nearly every minute I'm on an interstate. By comparison, I've flown coast to coast many times and never see another plane until transitioning to the landing pattern. If guys would stay out of weather neither they nor their plane can handle, avoid running out of gas and be careful with their currency, i.e., flying ten hours/year I think GA flying would prove far safer than driving.

[@abramsalinas1004]

@scottw5315 For the most part flying is safe. I'll tell you this, I felt a bit safer when I flew the Cessna 150 or 152 for my solo training. Unfortunately people get all flipped out during these instances.

[@abramsalinas1004]

I hope videos like this don't stop. The more the better.

[@medea27]

It's a bit of a misnomer to compare two vastly different modes of transport on an arbitrary measure of "safety." I know that it's a good statistic to quote to people afraid of flying, but the reality is that the barriers to entry in obtaining a pilot licence & owning/operating an aircraft are two very strong factors in the relative safety of flying. In contrast, anyone can own & drive a car - even when they are wholly unsuitable for driving and even when they don't have a licence... and addressing that issue is a whole different kettle of fish to aviation. But they are both important issues... safety isn't something you can say "achieved ✅️" and move onto the next topic, it requires constant monitoring. So if you feel vehicle accidents or fentanyl deaths aren't talked about enough then go start those conversations... but I wouldn't treat _any_ statistic showing deaths as "no big deal" because it sure-as-hell is a big deal to the couple of hundred people in it.

[@LTVoyager]

It is possible to get pretty accurate flying hours if the FAA wanted to do accurate statistics. My car registration requires me to supply the odometer reading each year and, similarly, my annual car inspection records the odometer so the state DOT has a couple of ways to track miles driven each year. Every airplane requires an annual inspection or 100 hour inspection. All the FAA has to do is require the A&P to submit the hours on the airplane at each inspection. This could be done in 5 minutes or less with a simple online system. In this day and age there is no reason to not have fairly accurate GA activity data.

[@jmizzonini]

Right but implementing this in the FAA will cost millions of dollars annually, is it worth the cost?

[@LTVoyager]

@ Millions annually? That’s hilarious. This requires an app that would cost at most $20K to write. The cost to summarize this data annually is not even one full-time employee, so maybe $150K tops. Is that data worth that much? That is a judgement call I can’t answer, but the FAA spends a lot more than this on safety programs with a lot less utility.

[@Saml01]

@@LTVoyageryou don't need an app when you can complete an 8710 in iacra for the flight review and all you need is two more fields, N number and hours. That along with the applicants times is all you would need.

[@LTVoyager]

@ If the 8710 was made a requirement for a flight review that would work. The only downside is your data is up to two years behind and that is one of the problems today with both AOPA and NTSB is the long delay in getting data and recommendations out to pilots. Using the airplane annual cuts that in half and even better for commercial aircraft that require 100 hour inspections. It would be a start though.

[@xenadu02]

Get ready for the conspiracy nuts to scream about BIG GUBBMIT and refuse. They're already starting up in the replies to the video.

[@robertlough4090]

Since the 19thC the poissonic distribution function has been used to estimate the probability of accidents from independent events. Estimating the mean to be used in the Poisson function, should have some rigor, but in the context of calculating probability trends using the Poisson function, it is still a statistical estimate. It would be interesting to observe, through back testing, whether the actual annual mean is consistent with the estimate used for the ratio in these accident reports and discussions. Mid air collisions, by definition not independent events, would have to be addressed in the analysis.

[@FlyWirescottperdue]

Thanks! I did not know of this function!

[@wassermutt7805]

A relative constant % of fatal would suggest that the types and ratios of accidents aren't changing, just the quantity. Your information on types of accidents will be interesting to see.

[@FlyWirescottperdue]

That is what I expect.

[@landonsock6408]

Scott, thanks for the in-depth analysis. Not to pull the pin and throw the grenade into the comment section, but what are your thoughts on AQP training for GA guys? I'm really looking forward to the next video.

[@FlyWirescottperdue]

Thanks. I think a modified version of AQP is a great idea.

[@skid2151]

Gunny! Appreciate the New Year safety briefing! Perhaps all G.A. pilots should be grounded until they have watched your briefing!! Happy 2025! Cheers! Skid

[@FlyWirescottperdue]

Thanks Skid!

[@mikedx2706]

Scott is letting us know that the emperor has no clothes. The government agencies' GA "accident database" has to be seen for what it is, just a broad guesstimate of what is really going on in the general aviation world.

[@FlyWirescottperdue]

Exactly my friend!

[@xenadu02]

Uhh how is the topic of this video (the AOPA's report) related to government agencies? AOPA is a private organization. The FAA's actual reported estimates of GA trips, landings, and hours are available and so is the methodology.

[@richardseton7014]

Good solid analysis. However what we see is the so called 'safety authority's ' increasing the overreach of onorous but ineffective regulations, then unjustly congratulating themselves, thus incentivising more inappropriate non beneficial regulations further driving up costs whithout any sfety benefits. Looking forward to your further focussed analysis!

[@xenadu02]

Where are we seeing that exactly? What specific rules has the FAA implemented to constrain GA pilots in the past 10 years? Which of those rules were based on bad data?

[@brentdykgraaf184]

You sir..Dan..and Juan are making a difference

[@jamesb7807]

Awesome as usual!!

[@C56-d8h]

Insurance companies have this data, at least in the UK. I need to estimate the hours for next year and report the hours for the previous. They know how many hours I have on type, my ratings, how long I have held an licence and they will have records of all accidents. Of course, in the UK data protection will prevent this being shared - and I suspect this is the case even if anonymised.

[@FlyWirescottperdue]

The issue is not every airplane is insured.

[@C56-d8h]

@@FlyWirescottperdue That’s interesting. I am in the UK and insurance is mandatory. I guess there may be a few rogue owners but probably not many.

[@davecat1458]

The majority of student pilots at 2 local airports in my area are foreign. They end up flying in another country. Statistically, you can massage any data pool to corroborate any thesis.

[@TheBullethead]

Good points.

[@AlyssaM_InfoSec]

Love the idea of tossing out dimensions that aren't reliable and using data that we can count on. I always wonder how skewed the % fatal is since as you noted many non-fatals don't get reported but pretty much every fatal does.

[@Big_Tex]

I’ve been hearing of a lot more accidents … ever since I started following aviation channels on KZbin 🤣

[@FlyWirescottperdue]

Not actually more, just higher profile.

[@AnarchyEnsues]

With adsb-out requirements, there should be a fairly simple way to gather all this digital information, not only hrs but take off and landings..

[@FlyWirescottperdue]

A couple of things make that idea problematic. First not all airplanes are required to have ADSB... it is required to fly in certain airspace, if you don't fly there you don't need it. Second ADSB in the US is primarily ground based which means it is highly dependant on how far away the ground station sis from you to get line of sight. At my airport 1000 ft is generally the lowest you'll get an ADSB hit.

[@AnarchyEnsues]

@FlyWirescottperdue only extremely old and rarely flown aircraft in a limited amount of locations qualify for exemption.(That can be your error bars) And to be honest, you having a data set that covers 99% of the fleet, it's good enough. You are viewing adsb from services that gain adsb from ground sources that are free, flight radar 24 has thousands of volunteers and airports that add one of their adsb antennas to their home or airport. So free information Adsb is also collected via satellite(from that same adsb source that transmits normally to the ground), but obviously satellite vendors do not offer that information for free(unfortunately) Just as the juju 737 crashed the other day and we don't have the adsb below 1000ft, the satellite vendors do, aireon is main provider. I'm sure the investigation team and airline would have that information as they sign up for such coverage.

[@jimmydulin928]

Thanks for your work, Scott. When we see fewer accidents but almost the same total number of fatalities, we have a training problem don't we? In my experience with crop dusting, I saw the much larger number of accidents in piston Cub, Sterman, Pawnee, CallAir, and Ag Wagon than later in turboprop airplanes costing a million dollars. Fatalities also dropped with the better equipment. In the piston end of general aviation, I see fewer accidents (hull cost is part of it here too) but the same number of fatalities every year. In my opinion, the low altitude orientation of Ag training saves lives. The speed of the modern turboprop airplane in the crop field scares the pee out of me, but what Wolfgang says of the experienced pilot makes a world of difference on takeoff and coming out of the crop field: "The experienced pilot will therefore point his nose up only as high as absolutely necessary to clear obstructions. In fact, if the take-off is from a very tight field, he will often point the ship's nose actually at the obstructions--maybe a half height of the trees. This does not help the airplane gain height, but it does help it gain speed. And the experienced pilot does not worry, because he knows (without thinking about it) that speed and height are two forms of the same thing, and for various reasons he prefers to have that thing in the form of speed. And he knows he can always at the last moment convert speed into height by pulling the nose up; the airplane will then 'zoom' and clear the obstructions." Stick and Rudder page 85.

[@FlyWirescottperdue]

Great comment Jimmy!

[@jmizzonini]

What about the theory that the fatal rate is flat due to simply being a certain number of situations over terrain that is impossible for even the best trained to make it out alive? Some fatalities are unavoidable due to where you’re flying when the unfortunate occurs

[@FlyWirescottperdue]

I'm guessing you are describing Controlled Flight Into Terrain... in mountainous terrain. That would be a subset of overall accidents, and I'd bet a small percentage as well.

[@Zaephyrs]

Hi Scott, Do you know the reason there was data missing for 2011? Interesting that recompute happened after that. A good and interesting video, thanks!

[@dp980]

It would be interesting to compare these data to the data the insurance companies collate (if they were willing to give it up). They have a vested interest in knowing improving accident rates.

[@FlyWirescottperdue]

How do we figure out how many accidents are in uninsured airplanes?

[@dp980]

@ Non-insured aircraft would be missed and be a similar limitation to those listed for your (AOPA?) data. However, what is the most important question to answer: what is the accident rate trend over time or what is the exact number of accidents year over year? I would argue the former is most important for AOPA especially given the data limitations. Both would be useful for the insurers because they translate to dollars. As you pointed out, there are data limitations. That will always be true for the reasons you stated: incomplete reporting, poor database design, lack of data validation, possible AOPA introduced bias etc. I’m being long winded but insurance companies have a monetary incentive to get the best possible data hence a comparison of their data/assumptions with AOPA’s would be of interest to see the same trends they see. It will be impossible to know the absolute truth so we are left with lies, damn lies and statistics.

[@JSFGuy]

Let's find out.

[@xenadu02]

My take is if the fatal rate has only decreased 5% of the total but the total has gone down a lot it means we have improved overall but we need more data to determine the causes of fatalities specifically to focus on those. AFAIK no one has ever done a good study on survival in light aircraft accidents like we have for cars. Maybe one NASA study in the 70s? We have no clue what changes to airframes, fuel tanks, instruments/user interface, etc might improve survivability and survivability varies a LOT seemingly on the most minor of differences. Crashes that eat energy like coming down on the tail or clipping one wing to turn forward momentum into angular momentum seem to be more survivable. Having headrests and shoulder harnesses seems to help. Could light aircraft have crumple zones or roll cages? Would a full airbag system help? No one actually knows. It's just armchair quarterbacks regurgitating useless trivia and opinion. "Too much energy"? Really? How do you know the average momentum carried into a crash to calculate the effectiveness of roll cages and airbags? "Too heavy!" Really? You have a PE stamp and did the simulations to arrive at the number "too much?" For jebus sake we don't even know if the much larger artificial horizon on glass cockpits reduces the rate of graveyard spirals! We know almost nothing compared to transport category aircraft or cars.

[@FlyWirescottperdue]

True. But what you are suggesting is replacing the GA fleet. That is exactly what happened with cars and airliners. What a great idea... I for one would be priced out of flying if I was required to install things you speculate on. I suspect well over 90% of pilots would be as well. Training and proficiency are the most effective solutions in the short term.

[@xenadu02]

@@FlyWirescottperdue I'm not proposing taking any action right now. Merely a government, industry, and pilot partnership to study what causes fatal injuries and what solutions might be available. Whether they are reasonable (weight, space, etc), cost-effective for new airplanes, or cost-effective to retrofit is a different question. It is possible such a study would end up telling us things are as good as they can be given weight and design constraints. Or maybe it will tell us new airframes can make certain changes that are just not reasonable to retrofit. OK, then pilots can make a decision on that but at least it is a reasoned decision based on a sound cost-benefit analysis. Or maybe such a study would tell us retrofit airbags are a good solution and the FAA could fast-track that to keep cost low. I agree with you that in the near term figuring out the human factors at play then improving training toward that end would be a good move. In fact we should do that even if what I propose happens tomorrow. I'm just saying humans will always make mistakes, especially in the GA world so we should also improve survivability in crashes - even stall-spin crashes - if the data shows it can be done reasonably effectively. Sure nothing will help a CFIT. Smashing into a mountain at 120kts is not survivable no matter what you do. But a lot of crashes are not at high speed and the survival varies a *lot* based on exactly how and where the airplane comes down. That tells me there is room to improve decel and energy dissipation. The airlines didn't cut hull losses (and fatalities) by an order of magnitude (or more) by only addressing pilot training. Improved bird ingestion testing for engines, improved fire suppression & detection, etc. They study what killed people in each accident and where appropriate change how new airplanes get built to address those issues. The FAA crashed a retired airliner in the desert (by remote control) to learn! No one has ever done that kind of work for GA.

[@GeorgeGeorge-yb2sz]

It sounds like AOPA is just as inefficient as NTSB. Two years? That is water under the bridge . . twice! It is now 2025 just in case AOPA hasn't changed their calendar in two years! I guess we are supposed to be kind to AOPA, but what have they contributed to GA? Oh I forgot, the Impossible Turn is not dangerous! How did that work out for you? Is your new director going to prove it is safe? Have you changed your opinion? Sorry, but reality is reality!

[@youbecha64]

I wonder if only physical injury/death or insurance claims make the accident list?

[@Qrail]

I hate to bring it up, but safety in the airplane, like airbags might help. Has anyone bothered to try it? What about fire bottles like on the dragsters? Wearing a helmet or motorcycle gear might help too. I wish the industry would think outside the box.

[@FlyWirescottperdue]

It pays to know the risk areas. As far as safety in cars goes it’s taken 50 years of improvements and car crashes still kill something like 50k a year.

[@joeshmooo5327]

@@FlyWirescottperdue thats a cop out answer, too many people burning to death due to crash. There has to be a better solution than to say "what about..."

[@Qrail]

@ I hear you. Thanks for all you do for the GA community. And Happy New Year

[@scottw5315]

I'd wager that by the time a manufacturer comes up with an airbag solution, it spends year in review at the FAA and many layers of liability insurance having been applied it could cost more than most light civil airplanes. Thirty years ago or so F Lee Bailey, celebrity attorney submitted an article to Smithsonian magazine. He happened to own Enstrom helicopters. He said that liability insurance doubled the price of his helicopters. This is a big reason why a new Cessna might run 500k. Being over lawyered in this country kills an enormous amount of innovation.

[@joeshmooo5327]

@@scottw5315 Flight Design F2 airplane has panel-mounted AMSAFE airbags

[@errorerror1337]

Fix the title please-something that rises moves from a lower position to a higher position. I think you mean "Is GA Accident Frequency Increasing or Decreasing?"

[@FlyWirescottperdue]

I kinda like it.

[@timthomson352]

N o w days with social media every blown tire is published widely. If you sneeze on approach the wolrld finds o ut

[@sandhill9313]

Thanks, Scott. It is quite a challenge you are taking on and I wish you well with it, I'm inclined to trust your conclusions even when couched in larger error bars more than some of the certainty we see from some sources, for sure.

[@nightcrawleroriginal]

So what is the purpose to all of this Numbers and Data stuff FW ? We dont learn from bars and graphs FW, we learn from Cause and How to avoid . Personal views. Take Care. 😊

[@FlyWirescottperdue]

It's simple my friend, you have to know what the problem is before you can address the cure.

[@nightcrawleroriginal]

Purdue ?? Hmm, sound familiar, any association to PMA ? Curious.😊

[@gracelandone]

Gryder’s scientific data analysis on GA Fatals is based on the number of hats he gives out. His intentions are good but his analysis is suspect.

[@LTVoyager]

Tracking actual fatalities is pretty precise. Since he is looking only at fatalities per year, regardless of flying activity, his data should be pretty solid. The problem is that he can’t say if a reduction in fatalities per year is due to increased safety or decreased flying. That doesn’t make his analysis suspect since he isn’t making any claims on accident rate per hour flown, only per calendar time.

[@sveneandersson7888]

Too much talking. Get to the essence and learnings. I shut it down half way.

[@keepyourbilsteins]

Too bad, could have learned something.