US Fatalities per 100K per year: Fishing and related workers: 116 or 0.116%; Logging workers: 91.9; Aircraft pilots and flight engineers: 70.6; Farmers and ranchers: 41.4; Coal mining: 38.9; Roofers: 32.4; Refuse and recyclable material collectors: 22.8; Driver/sales workers and truck drivers: 21.8; Police and sheriff's patrol officers: 18.0; Electrical power-line installers and repairers: 15.6.
Those are "normal" jobs.
High-altitude mountaineering: "23 deaths among 535 mountaineers beween 1968 and 1987", which is 4% or 214/100K per year. Roughly 450 people have flown into space or, like Challenger, tried to. Some 34+ have died. That's 7.5% (over the 5% you say is "politically unfeasible for NASA"), and scales to 151/100K per year.
About 375 American soldiers have died per year for the last few years in Afghanistan, out of 43,000 in the country. That's about 870/100K per year.
23 out of 535 Mountaineers is indeed 4%, but its over the course of 1968-1987. So that's an actual rate of 0.2% per year (1.2 deaths per year).
You need to do the same math with space travel---dividing the # of deaths by the # of years space travel has been in existance.
But to apply a notion from finance, space travel is such a small 'market' that its liable to fall prey to black swan events---e.g. maybe space travel is very very safe or very very dangerous, but the data set we have is non-typical and we'll start seeing 'typical' data once we ramp up to 100,000+ people per year travelling in space.
I including the number of years in the calculation. I wrote "which is 4% or 214/100K per year". That latter number includes dividing by 20 years. Specifically, "23/535 / 20 * 100000 = 214.9" or 215/100K (oops! I rounded down instead of to nearest).
Similarly, with 50 years of human space travel, 35 deaths/450 people/ 50 years * 100000 people = 156 deaths / 100,000 per year. I don't know why I wrote "151/100K", I meant to write "150/100K" and not be so precise. In any case, it looks like the source I used wasn't that accurate. See http://en.wikipedia.org/wiki/List_of_spaceflight-related_acc... for better details. I quote from there:
"About two percent of the manned launch/reentry attempts have killed their crew, with Soyuz and the Shuttle having almost the same death percentage rates. Except for the X-15 (which is a suborbital rocket plane), other launchers have not launched sufficiently often for reasonable safety comparisons to be made."
That implies that there is enough statistics.
In any case, I recall Feynman's essay about the Challenger explosion. He recounts how one person estimated Shuttle disasters as being about 10x safer than the more common unmanned ones, which works out to something like 1 out of every 50-100 Shuttle launches. Thus, other observations help strengthen the possibility that this isn't a black swan event.
From the list of astronauts in http://en.wikipedia.org/wiki/List_of_astronauts_by_name there are ~500 (some of the 554 listed didn't fly). Most of them flied between 2 and 4 times (The maximal I could find is 7.) So 3 flights per astronaut seams to be a good average. Then we get 500*3=1500 space travel seats, and 34 casualties, that is 2.2% = 2200 per 100K.
This is suspicious similar to the 1/50 (or 1/100) crash probability that Feynman gets in his report about the Challenger disaster. If 1 in 50 of the ships explodes, you have a 1 in 50 chance to die.
The 60 yeas space flight period cancels out , if we suppose that none of them flied twice the same year. We have and average of 25 space travel seats per year and .55 dead per year.
You get a 7.7% = 7700 per 100K, but didn't take in account that most astronauts flight many times.
18 people died in spacecraft, defined as in vehicles that could reach over 100km altitude. 517 have been to space. Jarvis, McAuliffe, and Michael J. Smith died in Challenger having never made it to space. However, another 11 died during space-related training, like Apollo 1. I'll use your 554 as the number of those trained as astronauts.
So I don't know where how that "34" I quoted earlier is justified, since I'm coming up with 29.
29 fatalities per 554 is 5% of the population, over 50 years of space flight. My hand-wavy calculation to get to "deaths per 100,000 per year is 29 fatalities / 554 people / 50 years * 100000 = 105/100K. (Previously I did it as 34 / 450 / 50 * 100000 =
151)
The numbers I'm computing are the effective deaths per 100,000 per year. That's different than the number you are calculating, which is the number of fatal launches per year, I think.
To simplify the calculations, I will only consider last year. How many astronauts were active last year? The number is not near 500 from the list. Some of them are dead. Some of them are behind a desk or doing public relation presentation. Some of them are retired or dead from natural events.
So, there are approximately 30 / 50 dead per year, from a population of 70 active astronauts, that is ~1%, or using a population of 30 flying astronauts we get a 2%.
http://en.wikipedia.org/wiki/NASA_Astronaut_Corps says "As of July 2011 the corps has 62 active astronauts, including 33 military officers, four medical doctors, and 15 with doctorates. The highest number of active astronauts at one time, was in 2000 when there were 149."
"only 339 candidates have been selected to date", but a couple of those candidates finished the program but did not go on with a career as an astronaut (one went into nuclear medicine instead).
"Twelve members of the Astronaut Corps were killed during spaceflight, Space Shuttle missions STS-51-L and STS-107. An additional seven were killed in training accidents. Sonny Carter died in a plane crash while traveling on NASA business."
So another way to reckon this is as 19 dead total from a total population of 339. 19/339 = 5.6% for the US program.
Assuming an average of 10 years, (1-p)10=(1-0.056) gives the probability of a person per year, which is 0.57%, or 580/100K/year.
I forgot to mention, your comment about the average number of years as an astronaut reminded me that I needed to use the right probability argument to get the statistics, rather than the coarse equation I used earlier. (HN ate my math: pow((1-p), 10) = (1-0.056) ) Thanks for pointing that out!
Those are "normal" jobs.
High-altitude mountaineering: "23 deaths among 535 mountaineers beween 1968 and 1987", which is 4% or 214/100K per year. Roughly 450 people have flown into space or, like Challenger, tried to. Some 34+ have died. That's 7.5% (over the 5% you say is "politically unfeasible for NASA"), and scales to 151/100K per year.
About 375 American soldiers have died per year for the last few years in Afghanistan, out of 43,000 in the country. That's about 870/100K per year.