How AP analyzed Gaza Health Ministry’s death toll data

The Associated Press analyzed the death toll in the Israel-Hamas war using data provided by the Health Ministry in the Gaza Strip. The analysis found that the proportion of women and children killed has declined over time.

The ministry, which is part of the Hamas-run government in Gaza, does not distinguish between civilian and combatant deaths, so the proportion of women and children killed is seen as the best available proxy for the civilian death toll.

The ministry has periodically released the underlying data upon which its daily updates to the death toll are based. The data sets released on social media in late October, early January, late March, and the end of April, named individuals the ministry says died because of the war.

The AP analysis was based on those individuals the Health Ministry identified with full names, genders, birth dates and Israeli-issued identification numbers. In April, nearly 23,000 deaths fit these criteria.

The death toll that the ministry publicizes daily, and which is often repeated by foreign media, is significantly higher and is not limited to those people who have been fully identified.

The ministry’s daily death toll – which stood at 34,622 at the end of April and 36,379 at the end of May – is not accompanied by the underlying data. Its count includes bodies that have not been claimed by families, or were decomposed beyond recognition, or whose records were lost in Israeli raids on hospitals — plus individuals with incomplete records.

The proportion of women and children who died in the war has declined over time even as the overall death toll has risen, the Health Ministry’s data shows: from 64% in October (4,344 out of 6,745 fully identified people), to 62% by early January (8,711 out of 14,117), to 57% in late March (11,296 out of 19,859) to 54% as of April 30 (12,479 out of 22,961.)

Women and children made up 38% of the newly and fully identified deaths in the month of April (1,183 out of 3,102), the Health Ministry’s data shows.

Beginning with its March snapshot, the Health Ministry data included dead bodies that were not fully identified, mainly because they lacked Israeli-issued ID numbers: There were 402 in March and 1,699 in April in this category.

The data provided to AP was imperfect in other ways. Among those people fully identified, some were listed twice by the ministry. There were 531 individuals counted twice in March, and 11 individuals counted twice in April.

The AP excluded from its analysis any dead bodies that were not fully identified, as well as the duplicates; filtering these out did not have a material impact (less than 1%) on the proportion of women and children killed.

When the war ends, the Health Ministry will likely be able to fully identify more bodies and get a clearer record of the war’s overall toll and its impact on different groups of people.