Categories

AOAV: all our reportsExplosive violence in Gaza

Analysis of new death data from Gaza’s Health Ministry reveals several concerns

On April 1, 2024, the Gaza Ministry of Health (MoH) released a third detailed list of people it claims have been killed in the war. Here Professor Mike Spagat follows up on the articles he wrote about the first and second detailed data releases with an analysis of this new dataset.

We are fortunate that the MoH provides such detailed, real-time information about deaths in an ongoing war. Nothing of comparable quality is available (e.g, for the devastating wars in Ethiopia, Sudan or even Ukraine).

Yet the trend toward declining data quality has continued. By mid-February only three of eight Gaza hospitals meant to report deaths into what I will call the “Central Tracking System” (CTS) were doing so, although the MoH has managed to offset its deteriorating hospital coverage with deaths reported via a publicly available form.

The newly released list contains 21,703 deaths but 440 have duplicate IDs, 470 have no IDs and 792 have the wrong number of digits in their IDs. A further 1,486 have invalid IDs even though they do have the requisite 9 digits.[1] The remaining 18,515 deaths all have listed sexes but 219 are missing ages. In short, roughly 1/7th of the entries in the new data release have quality problems.

The following table defines children as under the age of 18, organises the data by ID status and breaks down the 20,076 deaths with listed ages into men, women, boys and girls.

9-digit IDs predominate overwhelmingly. Thus, the percentage breakdowns by age and sex, provided in the second table below, are virtually identical for total deaths and for the subset of total deaths with 9-digit IDs.

This second table shows that women and children constitute 56.4% of the listed deaths. If we include only deaths with valid IDs, then the percentage of women and children drops to 53.3%.

The third table (below) shows that deaths reported through the new form include a far higher percentage of adult males than do deaths reported through the hospital system. Zaher al Wahaidi of the MoH’s Health Information Centre argues, plausibly, that the form draws in, disproportionately, many widows who are motivated to report the deaths of their husbands as a necessary first step toward collecting benefits. Yet one could, arguably, also expect widows to attenuate this bias toward reporting adult males by also reporting anyone else they know to have been killed.

The MoH has stated repeatedly that since mid-November it has supplemented its substantially disabled CTS with further deaths culled from “reliable media sources.” As hospital reporting has declined this supplementary data-capture channel gained importance and now accounts for more than 1/3 of the 32,845 total deaths that were claimed by the MoH through April 1.

Sky New reporter Ben van der Merwe pressed Mr. al Wahaidi to provide a database and methodology for this supplementary data collection but received neither. Instead, Mr. al Wahaidi asserted that the supplements integrate not just media sources but also reports from first responders.

Here are some tentative conclusions we may gain from observing this new data set.

First, the percentage of women and children killed does seem to be very high, roughly 60%, but the oft-cited claim that 70% of the Gazans killed in the conflict are women and children seems increasingly untenable. Indeed, in apparent disavowal of the 70% claim, Zaher al Wahaidi labelled this figure a “media estimate” that he could not explain.

Second, the announced total number of Gazans killed in the war, now exceeding 33,000, may seem plausible but it is not a documented fact. This figure includes roughly 13,000 deaths that have, apparently, been entered into an unavailable database using an unknown methodology. The short description of sources contributing to this figure has just shifted from “reliable media sources” to that plus first responders. First responders can, potentially, provide useful estimates of numbers of people, (e.g., trapped under rubble). However, victims covered by such estimates might eventually be captured by the hospital system and/or be reported through the publicly available form. Thus, we should dismiss the common claim that, because many of the dead are trapped under rubble or are missing for other reasons, the announced totals are undercounts. To the contrary, there seem to be at least two channels, aside from hospitals, through which such deaths can be captured.

A corollary of the above two points is that the MoH – whilst acknowledging the tremendous strains it must be operating under – risks losing some of its credibility.

Operating under the most difficult of circumstances, the MoH has supplied the world with mountains of precious information on the carnage in Gaza. I have argued previously that Israeli attacks on hospitals have compromised the MoH’s CTS, in violation of one of the provisional measures issued in January by the International Court of Justice. Because of this damage, the MoH information collection system lags, agonisingly, behind reality.

Nevertheless, the MoH should, if the chaos of the conflict at all permits, return to the practice it established earlier in the war: that of reporting its information in the most straightforward and transparent manner possible. The alternative is for its data to lose credibility, over time, and this would not serve the survivors nor the dead of Gaza well.


[1] 1,413 out of these 1,486 invalid IDs begin with the digits 8046 – what the significance of this is, is unclear.

Michael Spagat is a Professor of Economics at Royal Holloway College, University of London. Spagat’s published papers on armed conflict have been published in Nature, Scientific Reports, the New England Journal of Medicine, the Journal of Peace Research, the Journal of Conflict Resolution and PLoS Medicine among other places.  His current research addresses universal patterns in modern war, fabrication in survey research, the Dirty War Index civilian casualties in the Iraq conflict and problems in the measurement of war deaths.