The study, Van terroristische aanslag tot genocide (From Terrorist Attack to Genocide), was conducted by media researcher Nel Ruigrok, journalist Coen Ramaer, and journalism students from Erasmus University Rotterdam. Its findings were published Wednesday in Trouw and discussed in other Dutch media.
The researchers found that Israeli sources were quoted about one and a half times as frequently as Palestinian sources, that Israeli statements were presented with greater certainty, and that language describing Palestinian deaths was more likely to be euphemistic. They also criticized the media's reluctance to describe Israel's actions as genocide.
But a closer examination of the study's methodology reveals something rather different. In two important areas, the researchers appear to have built their own assumptions into the definitions of bias. What they regard as evidence of pro-Israel bias is often simply responsible journalism. And journalists following the researchers' preferred standards could end up producing reporting that is biased against Israel.
Consider how the study defines euphemistic language.
In its coding instructions, the researchers identify words equivalent to "died," "killed," "liquidated," and "fallen in combat" as potential euphemisms when used instead of "murdered."
That is exactly backwards. "Killed" is a neutral term, "murdered" attaches a moral component that often cannot be established by a journalist.
In a war, people are killed. Soldiers are killed in combat. Civilians are killed in airstrikes. Some are deliberately targeted, others die as a consequence of attacks on military objectives. Some killings are lawful under international humanitarian law; others are war crimes.
"Murdered" is not a more accurate or less euphemistic synonym for "killed." It is an accusation that a killing was unlawful, generally involving intentional wrongdoing. A reporter who says that someone was killed is describing what happened. A reporter who says that someone was murdered is making an additional claim that requires additional evidence.
This is Journalism 101. You don't presume a crime merely because someone died violently.
Yet the researchers' methodology risks treating precisely that restraint as evidence of bias. If journalists describe Palestinians as having been killed rather than murdered, the study can classify their language as euphemistic.
So when someone says "70,000 were killed in Gaza" the report would code that as pro-Israel bias, it implies that the proper term should be "murdered." Which is a flat out lie. To suggest that journalists should use "murdered" instead of "killed" for every death without establishing the circumstances is to prejudge the very questions that journalists are supposed to investigate.
The same problem is even with targeted strikes of specific targets, like Hamas leaders. It is accurate to say they were "assassinated" but not that they were "murdered" because murder is a crime, assassination is a disputed method.
The study's treatment of genocide raises a similar problem.
The researchers created two categories for articles discussing genocide: DEBAT, referring to genocide as a disputed allegation, and DAAD, referring to genocide as an established occurrence.
That distinction sounds reasonable until you examine how they applied it.
Their coding instructions allow an article to be classified as treating genocide as an established occurrence when an authoritative institution, relying on experts and research, concludes that genocide occurred.
This effectively allows organizations such as Amnesty International to confer factual status on a disputed legal conclusion.
The researchers even acknowledge an extraordinary consequence of their methodology: some articles in De Telegraaf reported genocide accusations by Amnesty International or Human Rights Watch and immediately rejected those accusations. Nevertheless, the study classified them as articles treating genocide as an established occurrence because of its coding rules.
An article can reject a genocide accusation and still be coded as treating genocide as a fact!
The researchers also classified an article as DAAD when it contained both an established-genocide passage and a passage treating genocide as a debate.
The entire controversy concerns whether Israel committed genocide, a crime requiring proof of specific intent to destroy a protected group, in whole or in part, as such. Amnesty International's conclusions are not self-validating simply because Amnesty is an established human rights organization. Its interpretation of the legal requirements is itself subject to criticism and scrutiny. In fact, Amnesty's Gaza report controversially argued that the conventional interpretation of genocidal intent was too restrictive. That is a substantive legal dispute, not something journalists should be expected to resolve by accepting Amnesty's authority.
A journalist who reports that Amnesty accused Israel of genocide, explains the accusation, and includes contrary evidence or legal arguments is not necessarily displaying pro-Israel bias. That journalist may be doing exactly what journalism demands.
To be precise, the Dutch study does not explicitly say that every journalist who rejects Amnesty's conclusion is biased toward Israel. But its coding system blurs the distinction between reporting an accusation and accepting it as established fact. That makes its findings about the media's treatment of genocide much less meaningful.
In both cases, the researchers appear to confuse journalistic neutrality with insufficient willingness to adopt accusations against Israel.
There are other problems with the study, too.
It found that Palestinian suffering appeared in headlines about two and a half times as often as Israeli suffering, which hardly sounds like an obvious pro-Israel bias. The researchers then divided the number of suffering-related headlines by the number of deaths and concluded that Israeli victims received vastly more attention per death.
But why should news coverage be proportional to fatalities? Israeli hostages remained newsworthy throughout their captivity, even on days when nobody was killed. A massacre occurring in a single day and a prolonged war naturally generate different patterns of coverage. Dividing headlines by death counts cannot establish what fair coverage would look like.
Similarly, counting Israeli and Palestinian sources without adequately accounting for their roles, the subjects they discuss, or the accuracy of their claims cannot establish which side receives more favorable treatment. The report divides sources into Israeli, Palestinian and other. But that doesn't indicate bias; what they say does. An Israeli quoted criticizing Netanyahu is still an Israeli source. A UN official accusing Israel of genocide is not a Palestinian source. Source nationality is not the same thing as editorial viewpoint. Bias must be demonstrated against an appropriate factual baseline, not against the researchers' preferred conclusions.
The study shows how easily a research project can manufacture an appearance of bias by defining ordinary journalistic standards as suspect.
This is a broader problem in contemporary discussions of Israel. Accusations are increasingly treated as established facts, and demands for evidence are interpreted as attempts to excuse wrongdoing. The distinction between describing an allegation and endorsing it becomes blurred.
The result is an epistemic trap: if journalists adopt the accusation, they are considered objective; if they insist on evidence or use legally neutral terminology, they are accused of bias.
The Dutch researchers set out to determine whether the media was biased toward Israel. In significant parts of their methodology, they appear to have assumed the answer to questions that journalists were supposed to investigate.
