What are the odds of dying as an astronaut on a spaceflight mission?
Evidence quality 4.88/5
Eight-dimension review score against the quality rubric . Each dimension scored 1–5.
- D1 Source grounding
- 5/5
- D2 Source authority
- 5/5
- D3 Arithmetic
- 5/5
- D4 Uncertainty
- 5/5
- D5 Scope
- 5/5
- D6 Prose
- 4/5
- D7 Perception honesty
- 5/5
- D8 Caveat completeness
- 5/5
Lifetime probability · lifetime, activity-specific
1 in 42
2.4% lifetime chance
range 1 in 50 to 1 in 24
● your factors — click this risk ▾ to reveal
≈ As likely as
Perceived
Most people sense that spaceflight is genuinely dangerous — one of the few popular fears where public intuition tracks the statistics reasonably well. Media coverage is anchored by the two Space Shuttle losses, Apollo 1, and Soyuz 1/11, and the general reaction to "would you take a 1 in 50 risk for a seat?" is a cautious no. We have not found a standalone survey isolating "fear of dying as an astronaut", so perceived risk is marked as editorial intuition. The interesting property of this fear is that it is among the few Likelier entries where the perceived risk is roughly calibrated against the actual number, and where informed consent — not risk blindness — is what explains the behavior.
Rough estimate: most people correctly guess spaceflight is in the low single-digit percent range per mission
Source: editorial intuition, not polled
Actual
~19 deaths across ~791 people flown into space (1961-2026)
all humans who have flown to space, 1961-2026
Show derivation
Reference subgroup: a crew member on a crewed orbital or suborbital spaceflight mission of any program (Vostok, Mercury, Voskhod, Gemini, Soyuz, Apollo, Skylab, Space Shuttle, Shenzhou, Crew Dragon, etc.) between the first human spaceflight in 1961 and April 2026. The Wikipedia list of spaceflight-related accidents and incidents, which aggregates the standard program-level fatality records, reports that 791 people had flown into space as of April 2026 and that 19 of them died in spaceflight-related incidents, for a headline rate of 19/791 ≈ 2.4 percent. The scope is declared as activity_specific_lifetime because this is per-person-per-mission risk for a specific activity, not a general-population lifetime risk, and it is not directly comparable to the population-lifetime figures on other Likelier pages. The 19-death figure includes the Apollo 1 ground-test fire (3 deaths), Soyuz 1 reentry (1), Soyuz 11 decompression (3), X-15 Flight 3-65-97 (1), Challenger STS-51-L (7), and Columbia STS-107 (7). Excluding Apollo 1, which happened during a ground plugs-out test rather than in flight, yields 16 deaths / 791 ≈ 2.0 percent, which matches Rick Hauck's Shuttle-era calculation of "closer to two percent" across the Shuttle program. Because the two canonical figures (2.0 percent in-flight-only and 2.4 percent including Apollo 1) bracket the same order of magnitude, we use 0.024 as the headline point estimate and a wider uncertainty band to reflect that this is a small-sample statistic dominated by a handful of catastrophic events, each of which killed its entire crew.
Caveats: Spaceflight fatality statistics are a small-sample problem. Nineteen deaths acro…
Spaceflight fatality statistics are a small-sample problem. Nineteen deaths across 65 years of human spaceflight is not enough data to support a stable per-mission rate in the frequentist sense; the headline 2.4 percent figure is dominated by five catastrophic events, each of which killed its entire crew, and would move substantially with a single future event in either direction. The figure also collapses risk profiles that differ by orders of magnitude: a Crew Dragon ferry flight to the ISS in 2025, a first-generation Vostok orbital flight in 1961, a Space Shuttle mission in 1986, and a future Mars transit mission are not drawing from the same risk distribution. The "1 in 42 per mission" headline should be read as a long-run historical average across the whole 65-year crewed-spaceflight program, not as a forecast for any specific future mission or vehicle. NASA's own internal probabilistic risk assessments for the late-Shuttle era settled on approximately 1 in 90 per flight, a figure that was itself a roughly tenfold revision upward from pre-Challenger management estimates of 1 in 100,000. The gap between engineering PRA and program management risk perception is one of the recurring themes of the Rogers Commission, CAIB, and Aerospace Safety Advisory Panel reports.
Regional breakdown
The headline figure averages across very different populations. Here’s how the probability varies by geography or context:
| Region / context | Lifetime probability | Notes |
|---|---|---|
| All crewed missions (1961-2026, per-person-per-mission) | 1 in 42 |
Headline figure. 19 deaths / 791 unique people flown. Dominated by five catastrophic events. |
| Space Shuttle era (per seat, 135 missions) | 1 in 63 |
14 deaths across ~850 seats filled over 135 Shuttle missions, or roughly 1.6 percent per seat. Two of 135 missions (Challenger STS-51-L, Columbia STS-107) were loss-of-crew events. |
| Soyuz program (per seat, 1967-2024) | 1 in 333 |
4 deaths (Soyuz 1, Soyuz 11) across more than a thousand cosmonaut/astronaut seat-flights on Soyuz crewed missions. The Russian crewed program has had no in-flight fatalities since 1971. |
| SpaceX Crew Dragon era (2020-2026) | 1 in 10,000 |
No fatalities across ~50 crew seats flown on Dragon as of April 2026. Small sample; the upper bound of the 95 percent confidence interval on a zero-event Poisson denominator of 50 is still above 5 percent per seat. Reported here as 'effectively zero so far'. |
| NASA internal PRA estimate, late Shuttle program | 1 in 91 |
NASA's post-Challenger probabilistic risk assessment settled on roughly 1 in 90 per flight for mature Shuttle operations; an earlier internal estimate put the first nine flights at roughly 1 in 9. Per-seat risk is similar because loss-of-crew events killed the whole crew. |
Risks at similar odds
Other risks with roughly the same likelihood — useful for calibration.
Cruise ship norovirus
What are the odds of being on a cruise ship voyage that has a norovirus outbreak?
Driving after cannabis
What are the odds of causing a fatal crash by driving within a few hours of using cannabis?
Driving at 0.10% BAC
What are the odds of causing a fatal crash when driving with a 0.10% blood alcohol level?
Driving on sedating meds
What are the odds of causing a fatal crash by driving while on a sedating prescription medication?
Recently viewed on this device
Stored locally — clear anytime.
Pick challenger
The headline number here is striking: roughly 1 in 42 per astronaut per mission across the entire history of human spaceflight. As of April 2026, 791 people have flown to space and 19 have died in spaceflight-related incidents, a rate of about 2.4 percent. For a professional activity selected and trained at the frontier of engineering capability, with entire state bureaucracies devoted to catching every failure mode before launch, a per-trip fatality rate in the low single-digit percent range is extraordinary. Flight engineer Rick Hauck, computing the figure just after the Columbia loss in 2003, told an AIAA audience that at a known four percent death rate he would not have flown — and that his personal comfort threshold was closer to one percent, “the typical American has over his life of dying in an ordinary traffic accident.”
What’s unusual about this fear is that perception and reality are already in rough agreement. Most people correctly sense that going to space is genuinely risky; the numbers bear that sense out. Astronauts accept the risk knowingly, under informed consent, because the mission purpose justifies it to them. That makes this one of the few Likelier entries where the rational answer to “is this fear calibrated” is yes. The only comparable peacetime modern activity with a similar routine per-event mortality is high-altitude mountaineering — per-summit death rates on Everest and K2 are in the same order of magnitude, and Hauck explicitly drew that comparison. A closer historical analogue is wartime: US heavy-bomber crews in the 1943-1944 Eighth Air Force faced per-mission fatality rates in the same low single-digit percent range across a 25-to-30-mission tour of duty. Spaceflight and strategic bombing are the two modern activities where a 2-percent per-event death rate is treated as a working cost of doing business.
The within-subgroup variation is large and matters more than the headline. The Space Shuttle era produced 14 of the 19 deaths across 135 missions, for a per-mission loss-of-crew rate around 1.5 percent and a post-Challenger NASA probabilistic risk assessment of roughly 1 in 90 per flight. The Russian Soyuz program has had no in-flight fatalities since 1971, across more than fifty years and several hundred seats. SpaceX Crew Dragon has flown dozens of seats since 2020 with no fatalities, though the sample is still small enough that a zero-event history is consistent with anything from ~0.2 percent down to essentially zero per flight. Pre-Challenger NASA management famously estimated Shuttle risk at 1 in 100,000 per flight while its own working engineers, per Feynman’s Rogers Commission appendix, estimated roughly 1 percent — a gap of three orders of magnitude on the same hardware. The historical 2.4 percent is the number for the whole program across 65 years. It is not the number for any specific vehicle, and it is explicitly not a forecast for any future mission.
Claim ledger
Every number below is what each source reported, with the verbatim quote we relied on and how we arrived at our figure. Click any link to verify directly.
-
[1] Aerospace (MDPI) / Schmitz J, Komorowski M, Russomano T, Ullrich O, Hinkelbein J — Sixty Years of Manned Spaceflight — Incidents and Accidents Involving Astronauts between Launch and Landing
Sixty Years of Manned Spaceflight — Incidents and Accidents Involving Astronauts between Launch and Landing- Statistic
327 manned spaceflights from 1961-2020, 1,294 astronaut-missions, 19 astronaut deaths; fatality rate 5.8% per spaceflight- Excerpt
“"The number of astronauts who have died during spaceflight is represented by n = 19." "The current statistical fatality rate is 5.8% (deaths per spaceflight) with the highest fatality rate in the 1960s (0.013 deaths/day spent in space), and the lowest rates in the 1990s and the period from 2010 until the present (no deaths)." ”
- Source data from
- 2022-11-01
- Accessed
- 2026-04-12 · archived copy
- Calculation
- Schmitz et al. count 1,294 person-launches across 327 missions, confirming 19 deaths. The per-mission rate (19/327 = 5.8%) is higher than the per-unique-person rate (19/791 ≈ 2.4%) because many astronauts flew multiple missions. Likelier uses the per-unique-person rate as the headline since the entry frames risk per astronaut. The 19-death roster matches the Wikipedia aggregate exactly.
- Independence
- Peer-reviewed open-access paper drawing on NASA, Roscosmos, and ESA mission records. Independently compiled and methodologically distinct from the Space Review secondary reporting of Hauck's AIAA lecture.
-
[2] American Journal of Medicine / Freiberg AS, Zhou S — Celestial Versus Terrestrial Travel — An Analysis of Spaceflight Fatalities and Comparison to Other Modes of Transportation
Celestial Versus Terrestrial Travel — An Analysis of Spaceflight Fatalities and Comparison to Other Modes of Transportation- Statistic
18 fatalities in 4 fatal missions across 326 launches (1961-2019); per-trip fatality rate 1.2%, per-person rate 1.4%- Excerpt
“"There has yet to be a fatality in orbit, and there have been none on any space flight since 2003." ”
- Source data from
- 2020-11-01
- Accessed
- 2026-04-12 · archived copy
- Calculation
- Freiberg and Zhou count 18 deaths (likely excluding X-15 Flight 191 depending on definition) across 326 launches through ~2019, giving a per-trip rate of 1.2% and per-person rate of 1.4%. The lower count vs Schmitz et al. (18 vs 19) reflects the definitional boundary around X-15 suborbital flights. Both papers independently confirm zero in-orbit fatalities and no deaths since Columbia (2003).
- Independence
- Fully independent of Schmitz et al. — different research group, different journal, different publication year, same underlying agency records but independently compiled.
-
[3] The Space Review (reporting on Rick Hauck's AIAA lecture) — Weighing the risks of human spaceflight (page 2)
Weighing the risks of human spaceflight (page 2)- Statistic
18 of 430 humans who have flown in space had died, for a fatality rate of just over four percent; Shuttle-specific rate closer to two percent across ~600 seats on 113 flights- Excerpt
“"18 of the 430 humans who have flown in space have died" ... "a fatality rate of just over four percent." ... "Would I have flown if I had known there was a four percent chance of death? No, I don't think I would have flown." ”
- Source data from
- 2003-11-10
- Accessed
- 2026-04-11 · archived copy
- Calculation
- Hauck's 2003 AIAA analysis used a person-basis denominator (unique individuals flown) rather than a seat-basis denominator, producing a ~4 percent per-person figure immediately after the Columbia loss. Over the next two decades the denominator grew much faster than the numerator (no in-flight fatalities since 2003), pulling the figure down to ~2.4 percent by 2026. Hauck's separate seat-basis Shuttle calculation across "over 600 seats filled on 113 flights" produced "closer to two percent", which matches the modern 19/791 figure within rounding and is used as the cross-check that sets the low end of the uncertainty band. The Space Review's article is a secondary reputable-reference summary of a former Shuttle commander's technical talk, which is why it is classified reputable_reference rather than primary_study.
- Independence
- Hauck's person-basis denominator (430 people flown as of 2003) and Wikipedia's current 791-person denominator both derive ultimately from space-agency astronaut rosters, so the two sources are not independent on the denominator. They are independent in methodology — Hauck's lecture was computed ad hoc from program data in 2003, while the Wikipedia roster was aggregated from subsequent mishap-report publications. Treat them as method cross-check rather than two independent measurements.