Theme
“Counting the dead in Gaza”
Every argument about the war leans on a number: 'more than 60,000 dead', '186,000', 'most of them women and children', 'half were fighters'. This page sets out where each figure comes from, what it can show and what it can't, with the critiques from both sides.
What's true
What the independent estimates say: the total is more likely too low than too high.
Two studies using different methods both point to an undercount of violent deaths. A capture–recapture analysis in The Lancet (January 2025), matching three lists (the Ministry's hospital list, its online family reports and social-media obituaries), estimated about 64,000 traumatic-injury deaths to 30 June 2024, roughly 41% above the Ministry's count for that period. The first household survey inside Gaza (The Lancet Global Health, 2026; 2,000 households, fieldwork 30 December 2024 – 5 January 2025) estimated 75,200 violent deaths to 5 January 2025, about 25,000 more than the Ministry's figure, with women, children under 18 and people over 64 making up 56% of them. Both studies have critics. Two of the capture–recapture study's three lists come from the Ministry, and its plausible models range from about 47,000 to 88,000 deaths; a statistical response by Abraham Wyner argues that a few influential bad records drive the result and puts the undercount much lower ⚠. The survey's sampling and interviewer effects have been questioned in the same journal ⚠. Their supporters point out that two different methods reaching a similar order of undercount is itself a form of confirmation.
dataJamaluddine et al., 'Traumatic injury mortality in the Gaza Strip from Oct 7, 2023, to June 30, 2024: a capture–recapture analysis', The Lancet (2025)
Peer-reviewed; 64,260 (95% CI 55,298–78,525); MoH undercount ~41%.
data'Violent and non-violent death tolls for the Gaza conflict: new primary evidence from a population-representative field survey', The Lancet Global Health (2026) (2026) ⚠
75,200 violent deaths to 5 Jan 2025; 56.2% women, children and over-64s; ~12,200 missing (figures via AOAV and CBC summaries; the article returned 403).
dataAbraham Wyner, 'A Statistical Response to Traumatic Injury Mortality in the Gaza Strip… A Capture-Recapture Analysis' (SSRN, 2025) (2025) ⚠
What's left out
What the critiques of the Ministry's lists found.
Analysts of the named lists, notably the Henry Jackson Society (2025), found errors: impossible ages, men recorded as women, deaths from natural causes and from before the war, and about 3,400 previously listed names that disappeared without explanation in March 2025 ⚠. These errors matter most for the composition of the dead, the share of women and children, which is the figure most often used to argue about intent. They matter much less for the overall scale, which the independent studies suggest is if anything underestimated.
advocacy / viewHenry Jackson Society, 'Questionable Counting: Analysing the Death Toll from the Hamas-Run Ministry of Health in Gaza' (2025) (2025) ⚠
Israel's figures: combatants killed.
The Israeli army has periodically stated how many militants it has killed; the figures are not published with names or methods and can't be checked from outside. An investigation by +972 Magazine, Local Call and the Guardian (August 2025) reported that the army's own intelligence database listed about 8,900 fighters as killed or probably killed by May 2025, which would imply a much higher civilian share than the army's public statements; the army disputed the reporting ⚠. So the combatant–civilian split, the number that matters most for the laws of war, is the least verifiable number of all.
press+972 Magazine / Local Call / The Guardian, investigation into the IDF intelligence database of militants killed (2025)
21 August 2025; IDF told the Guardian the figures were incorrect.
The picture
Where the Health Ministry's figures come from.
The Gaza Ministry of Health, part of the Hamas-run government, publishes two things: a running total and, periodically, named lists with each person's name, ID number, age and sex. The lists are built from hospital and morgue records and, since 2024, from families' reports through an online form. The total also includes bodies not yet identified. Neither separates combatants from civilians. UN OCHA passes the figures on 'according to the Ministry of Health', and from there they reach the media, where the attribution often disappears (see the casualty-chain diagram on 'The citation map of the genocide claim'). In earlier Gaza wars, later counts, including Israeli ones, came out broadly close to the Ministry's totals ⚠.
The number that isn't a count: '186,000'.
In July 2024 a letter in The Lancet took the then-current direct toll of 37,396, applied a ratio of four indirect deaths per direct death taken from a 2008 report on other conflicts, and called up to 186,000 deaths 'not implausible'. It was correspondence, not a study, and one of its authors later called the figure illustrative and said it had been misquoted. It still circulates as 'the Lancet study says 186,000' (see 'Proxy presented as outcome').
Traps here:◆ Proxy presented as outcome◆ Citation loop
advocacy / viewKhatib, McKee & Yusuf, 'Counting the dead in Gaza: difficult but essential', The Lancet (correspondence) (2024)
Letter, not a study; 4:1 indirect-death multiplier; later described by a co-author as illustrative.
Where it leads if pressed
What each number can show, and what it can't.
The totals and the independent estimates show the scale, and on the evidence the scale is enormous. The named lists show who died, with errors that matter for the demographic arguments. Neither shows how many were combatants, and neither shows why anyone died: whether a strike was lawful, disproportionate or deliberate, which is what the war-crimes and genocide questions turn on (see 'Israel is committing genocide in Gaza.'). Honest use of the numbers: give every figure with its origin (Ministry-reported, independently estimated, army-claimed), give ranges rather than the convenient point, and don't ask a count to answer a question about intent.