How The Intercept laundered a book teaser into 'proof of pro-Israel bias' in 96 hours

On Tuesday, May 12, at 4:58 in the morning Eastern time, The Intercept published Adam Johnson’s column We Analyzed Thousands of News Articles: Here’s the Proof of Pro-Israel Bias in Mainstream Media. The piece previewed seven statistical findings — bar charts comparing how often US outlets used “right to defend itself,” “slaughter,” “massacre,” “human shields” for Israelis versus Palestinians — drawn from Johnson’s book How to Sell a Genocide, published by Pluto Press on April 21. The teaser claimed coverage of “over 12,000 articles” from seven named outlets plus “5,000 TV segments” from CNN and MSNBC.

By Tuesday night the same numbers were running at Common Dreams as a book-launch story. The same night, IBTimes UK added the two samples together and headlined “Analysis of 17,000 Reports” — combining 12,000 articles and 5,000 TV segments into one round number without acknowledging they were coded differently. Middle East Eye filed a live-blog item the same day. By Friday, May 15, Middle East Monitor called the same seven charts “sweeping new analysis.”

In ninety-six hours a book-promotion preview had become “proof.” No outlet in the chain re-ran the methodology. None could have: the methodology was not published.

What the Intercept piece disclosed, and what it did not

The teaser names one author — Adam Johnson, media analyst and co-host of the Citations Needed podcast. It names seven outlets for the print sample and two for the TV sample. It states the window as the first year of the war with emphasis on the first hundred days. It gives a sample size in round numbers. It presents seven comparative ratios. That is what is disclosed.

What is not disclosed: the coding rubric — the operational definition that tells a coder when “right to defend itself” in a paragraph counts as pro-Israel framing rather than a quotation, paraphrase, denial, or setup for the opposite point. Who the coders were. Whether the coding was done by Johnson alone, by Johnson and his co-researcher Othman Ali (named on the 2024 piece but not this one), by paid coders, by graduate students, or by an LLM. Whether two coders read the same article and agreed — the inter-rater reliability number that separates a study from a press release. Whether the dataset or codebook is downloadable. Whether a second team could replicate.

The Intercept itself published a May 15 correction — three days into the cascade — acknowledging that the “Emotive Words on TV” caption had misstated which Sunday shows were analyzed and that two other graphics had wrong scales. The piece kept the headline. The downstream amplifiers kept their pieces.

How the seven charts traveled

The cascade unfolded on the timetable Robert Cialdini named in Influence in 1984 as social proof: when a claim has already been published by a peer outlet, the cost of accepting it falls toward zero, and the cost of independently testing it rises with deadline pressure. Aligned outlets do not need to coordinate; they need only to read each other. The alignment in this case is the prior conviction that US coverage is pro-Israel; the cascade is the mechanism by which that prior conviction picks up the appearance of empirical confirmation it has not actually earned.

At The Intercept the claim was empirical. At Common Dreams the same day it was evidence corroborating a book launch. At IBTimes UK the same day it was “analysis of 17,000 reports” — the round number that signaled scale to a reader who would not click through. At Middle East Eye the same day it was an “analysis.” At Middle East Monitor three days later it was a “sweeping new analysis” that found US media “complicit in selling Gaza genocide.” Nowhere along the chain did an outlet either publish a methodology critique or attempt to replicate any of the seven ratios on a re-coded sample.

Johnson and Ali’s earlier January 9, 2024 Intercept analysis of 1,100 articles had already been absorbed into the secondary literature. In March 2024, George Washington University’s William Youmans cited it as the empirical backbone of a DAWN essay without himself replicating the coding. Two years on, the May 12 follow-up enters a peer ecosystem already primed to receive Johnson’s findings as established, by Johnson, with no independent test of either round.

What a disclosed-methodology study looks like

The contrast is not hypothetical, and it is not partisan. Researchers across the political spectrum have produced disclosed-methodology studies of Israel/Gaza coverage in the same window.

The arXiv preprint by Bedoor AlShebli, Bruno Gabriel Salvador Casara, and Anne Maass at NYU Abu Dhabi’s Psychology Science Division analyzed roughly 14,000 articles from NYT, BBC, CNN, and Al Jazeera English over October 7, 2023 to October 7, 2024. It names the authors, the framework (Human Empathy and Narrative Taxonomy), the LLM-coding pass, the statistical model (GEE logistic regression), and the validation step against prior human-coded responses. It also does not publish its own inter-rater reliability for the LLM pass or release dataset and code — limitations that show up inside the paper rather than as a May-15 correction. It is a real study with real limits, named.

Andrew Fox’s April 2025 Henry Jackson Society report on 1,378 articles across eight named outlets names the author, the chapter editor (Tania Glezer of the International Institute for Social and Legal Studies), the sample frame, the outlets, and the denominators behind every percentage. Its conclusions cut in the opposite political direction from Johnson’s: 84 percent of the sample failed to distinguish combatant from civilian deaths — a confound for any “civilians killed” claim, because Hamas places its fighters and military assets in hospitals, schools, and apartment blocks — and 98 percent cited casualty figures from the Hamas-controlled Gaza Ministry of Health as if those figures were independently verified. They are not. The document still discloses what it did, who did it, and how a critic would replicate.

The Center for Media and Public Affairs at George Mason University, founded in 1985 by S. Robert Lichter, has published forty years of quantitative content analyses of US news coverage of Israel — Second Intifada in 2002, Israel-Hezbollah in 2006 — with disclosed sample frames, coding protocols, and named research staff. CMPA findings have cut in multiple directions over four decades because the methodology, not the conclusion, was the discipline.

What separates these from Johnson’s preview is not the political valence of the findings. It is whether the document discloses what would be required to test it. Johnson’s does not. None of the May 12–15 amplifiers asked it to.

Where to check every claim

Read the Intercept piece itself. Note what is presented — seven bar charts, two sample sizes, one named author — and what is not: codebook, coders, inter-rater reliability, dataset link, replication path. Note the May 15 correction at the bottom and ask why none of the downstream outlets updated their pieces.

Walk the cascade by date. Common Dreams, May 12. IBTimes UK, May 12, “17,000 reports.” Middle East Eye, May 12. Middle East Monitor, May 15, “sweeping new analysis.” Did the outlet re-run the coding or paraphrase the original? In every case, paraphrase.

Compare to disclosed-methodology work. The NYU Abu Dhabi preprint. The Henry Jackson Society’s 1,378-article report. The Center for Media and Public Affairs on forty years of content analysis. Named authors, named outlets, stated denominators, disclosed coding frames — and two of those three reach sharper or different conclusions than Johnson’s.

Then ask what the cascade did not: how would a second team, starting from the same corpus, test whether Johnson’s seven ratios reproduce?

---

The world does not get better when honest readers accept “proof” as a synonym for “this outlet agrees with that outlet.” It gets worse. Cascades of social proof, accelerated by digital reach and aligned partisan deadlines, are the engine of every collapsing media environment in living memory. The right side of history has never been built by accepting numbers from people who refuse to show their work.

Adam Johnson may be right about one or more of his seven ratios. He may be wrong about all of them. We do not know — and as long as the codebook, the coders, the inter-rater reliability, and the dataset remain unpublished, neither does he in any defensible sense, neither does Common Dreams, neither does IBTimes UK, which inflated his sample size by forty percent in twelve hours without noticing, and neither does Middle East Monitor, which three days later called a magazine preview “sweeping new analysis.” The honest phrase here is undocumented anti-Israel claim, not proof of pro-Israel bias.

CMPA and Henry Jackson Society have shown for decades that disclosed-methodology bias studies are publishable on this beat. The Henry Jackson Society’s findings cut sharply against the Intercept narrative — 84 percent of mainstream articles failed combatant/civilian distinction, 98 percent uncritically cited Hamas-supplied casualty figures — and those findings travel in nothing like the same cascade because the cascade is not built to carry them. Johnson and The Intercept can publish a disclosed-methodology study any time they choose. Until they do, MissingBridge will keep counting the difference between a study and a cascade, and naming whose case the cascade is built to make.

Read with us. Share when we earn it. Tell us when we miss.