A detailed look at the Guardian’s framing—and why Michael Mann’s Stalinism comparison is flat‑out wrong.
The Guardian has a flair for drama. Their latest entry, headlined “Scientists decry Trump energy chief’s plan to ‘update’ climate reports: ‘Exactly what Stalin did’”, is a case study in framing, innuendo, and the casual misuse of history to shut down legitimate scrutiny. The article targets U.S. Energy Secretary Chris Wright for saying the administration will review and “update” the National Climate Assessment (NCA) reports, calling forth a predictable chorus of “the science is under attack.” The capstone is Michael Mann’s emailed comment likening the move to Joseph Stalin.
Let’s slow down the rhetoric, and then examine what’s really going on.
What the Guardian actually reports:
According to the article, Secretary Wright said the administration is “reviewing national climate assessment reports published by past governments” and that, after review, “we will come out with updated reports on those and with comments on those reports.” The Guardian characterizes the NCA as “the gold standard report of global heating and its impacts” and then asserts that the Department of Energy recently produced a report arguing concern over the climate crisis is overblown—“slammed by scientists for being a ‘farce’ full of misinformation”—without, of course, addressing any specific claims or counterclaims on the merits.
Here is the passage the Guardian:
“In recent weeks the Trump administration deleted the website that hosted the periodic, legally mandated, national climate assessments. Asked about Wright’s comments on the national climate assessment reports, respected climate scientist Michael Mann said in an emailed comment to the Guardian: ‘This is exactly what Joseph Stalin did.’”
The piece then amplifies a statement from an activist-policy director at the Union of Concerned Scientists warning that the administration plans to “bury the scientific evidence” and “replace it with outright lies,” and urges Congress to “safeguard the integrity” of the NCA process. It adds that the NCA is published by NOAA, notes a DOE spokesperson told CNN Wright was not suggesting he would personally alter past reports, and recounts parallel regulatory battles, including EPA’s proposed moves on the 2009 endangerment finding.
The non-problem with “updating” government reports:
Government reports are not scripture. They’re periodic documents produced by committees—committees populated by people with their own priors, selection criteria, and preferred models. Reports get updated all the time; definitions, datasets, and assumptions change; and methodological critiques emerge. If you think that’s scandalous, you haven’t worked in or around government very long.
The Guardian treats the idea of reviewing and updating the NCA as inherently sinister, as if the earlier versions descended from Sinai on stone tablets. But this is exactly what administrations do: they review the work of prior administrations and amend, correct, or supplement it. The only real question is whether the revisions are transparent, documented, and evidence-based. That’s the standard we should demand—every time, from any administration.
The article doesn’t engage with a single methodological critique of prior NCAs. That omission matters because there are many long-standing issues worth debating:
- Scenario bias: heavy reliance on worst-case scenarios (historically, RCP8.5 and its rebranded equivalents) as a central narrative rather than an outlier risk case.
- Model–observation divergence: several model ensembles have run hot versus observed lower-troposphere trends and surface station networks, especially when quality control of station siting and urban heat island influences is applied.
- Attribution framing: complex multivariate extreme-event attribution framed with overconfident headlines rather than transparent treatment of uncertainties, confounders, and data inhomogeneities.
- Economic damage functions: speculative, model-driven projections pushed as policy imperatives while bracketing enormous uncertainty bands as afterthoughts.
You’d think a newspaper so invested in “holding power to account” might welcome a scrub of past claims. Instead, the Guardian’s reflex is to protect the incumbent narrative by analogizing scrutiny to dictatorship. That’s not journalism; that’s narrative enforcement.
About that Stalin comparison:
Let’s tackle the heart of the piece. Dr. Michael Mann is quoted as saying, “This is exactly what Joseph Stalin did.” Strong words. Also wrong—spectacularly wrong.
Of course, this isn’t the first time he’s gone off the rails, back in 2019 he said something similar:
What Stalin actually did was purge dissidents, centralize total control over social institutions, imprison and execute political opponents, censor and rewrite encyclopedias, and erase entire human beings from photographs and history. That was enforced via the NKVD, show trials, gulags, and famine. The stakes were life and death.

By contrast, reviewing a set of climate reports in a democratic system—reports that are already public, archived, widely distributed, and subject to FOIA—does not erase them, does not criminalize dissenting scientists, and does not prevent contrary publications. It adds a new edition, a commentary, or an alternative framing—precisely the kind of back-and-forth that defines open societies. To equate these two things is not “rhetorical flourish”; it’s historical malpractice.
Using Stalin as a cudgel here is meant to end debate, not inform it. The tactic is simple: label your opponent a tyrant, claim emergency, and you no longer have to engage with the underlying scientific disputes—scenario selection, model skill, detection and attribution uncertainties, observational biases, or cost-benefit tradeoffs of proposed policies.
To put it bluntly: Mann’s comparison is demagogic nonsense. It cheapens real historical horrors and poisons the public square. If everything you don’t like is “Stalinism,” then nothing is.
“But the website was deleted!”
The article claims: “In recent weeks the Trump administration deleted the website that hosted the periodic, legally mandated, national climate assessments.” As someone who has watched federal web infrastructure get replatformed, archived, consolidated, and re-skinned for decades, I can tell you “the website was deleted” is a slogan, not an argument. Government web content moves. URLs change. Archives exist. Documents get mirrored on multiple servers and non-government sites. The Guardian itself says the “most recent report is hosted on the Guardian website.”
If the reports were truly being disappeared, you wouldn’t be reading about them—or reading them—so easily. The real grievance isn’t that the documents vanished; it’s that the official homepage no longer serves the narrative at the velocity activists prefer. That’s a communications fight, not a historical erasure. Let’s keep our terms straight.
The NCA is not a holy writ—it’s a consensus product with known limitations.
The Guardian calls the NCA “the gold standard.” That’s a press phrase, not a scientific finding. The NCA is a consensus compilation—often useful as a reference, but not a substitute for the primary literature. Everyone in the field knows consensus documents tend to smooth over outlier results that complicate the storyline and to highlight material that is legible to policymakers and media. That’s not a smear; it’s how these reports are built.
Consider scenario usage. For a decade, worst-case scenarios were presented as though they were central expectations rather than outlier tails. While there’s an active effort in parts of the community to recalibrate away from that habit, much of the public-facing narrative still leans hard on “if it bleeds it leads.” If a new NCA review rebalances scenario weighting, good. That’s not “burying science”; it’s catching up to a literature that has been trying (not always successfully) to self-correct.
Consider model evaluation. Multi-model means are not thermometers. Skill varies by region, parameter, and timescale. Lower-troposphere temperature trends, tropical mid-troposphere amplification, cloud feedbacks, and aerosol forcing uncertainties all matter. Elevating ranges of observed data that undercut hot-running models is not “denial”; it’s basic due diligence.
Consider observational quality. The surface record is highly sensitive to siting, station moves, time-of-observation changes, and urbanization. Homogenization algorithms attempt to address this, but those algorithms are not magic. We should welcome deeper, transparent treatment of siting biases—especially in light of multiple surveys showing siting compliance problems. That’s called improving the data pipeline, not censoring it.
“They plan to bury evidence and replace it with lies!”
The Guardian quotes an activist policy director claiming the administration will “bury the scientific evidence” and “replace it with outright lies.” That’s a serious charge. But the article offers no specifics, no contested datasets, no disputed equations, no transparent side-by-side comparisons. Just adjectives.
If the fear is that the new process will remove or downplay certain claims, the antidote is simple: side-by-side version control. Post the old and the new, with diff-style change logs, clear sourcing, and hyperlinks to the underlying datasets and code. Invite external replication. If an administration attempts to bury legitimate evidence, this approach will expose it. And if critics are exaggerating, this approach will expose that, too. Either way, sunlight wins.
By the way, the article itself notes a DOE spokesperson said Wright “was not suggesting he personally would be altering past reports.” That undercuts the whole “Great Leader personally rewrites history” vibe the piece is trying to evoke. Bureaucracies—love them or loathe them—have processes. Use them. Audit them. But spare us the gulag metaphors.
The Guardian’s framing problem; two aspects of the article deserve special mention.
First, the piece gives the reader a single narrative arc: brave scientists and journalists versus a regime intent on erasing the truth. The rhetorical high point is Mann’s Stalin line. But this arc collapses under basic scrutiny. The reports in question are already public. They will remain public. Other journals and institutions will keep publishing. Universities will keep teaching. NGOs will keep organizing. None of that is being criminalized. This is a fight over what gets elevated as “official” and how uncertainties are presented. That is not Stalinism; that is politics in a pluralistic society.
Second, the article repeatedly appeals to authority (“gold standard,” “peer-reviewed,” “respected climate scientist”) while skirting the hard questions about how the sausage is made—scenario selection, parameterization, tuning, observational inhomogeneities, and the translation of model outputs into headlines. Authority is not a substitute for transparency. If your argument depends on scaring readers with historical tyrants rather than persuading them with data and method, you don’t have much of an argument.
I’ll be blunt: Michael Mann’s Stalin comparison is reckless, historically illiterate, and frankly beneath the level of discourse the public deserves.
It’s the rhetorical equivalent of hitting the fire alarm because someone asked to check the library and found a book missing. It trivializes actual totalitarian crimes while simultaneously attempting to delegitimize any audit of climate orthodoxy. That’s not how science advances; that’s how dogma defends itself.
Strong language is warranted here because the allegation is strong. If you accuse your political opponents of doing “exactly what Stalin did,” you had better be ready to point to the prisons, the purges, the shootings, the disappearances, the shuttered presses, the banned books, and the criminalization of curiosity. None of that is happening. What is happening is a bureaucratic review of a bureaucratic report. If that feels threatening, maybe the report is less solid than advertised.
We’ve been told for years that “the science is settled,” usually as a way to short-circuit questions the public is entitled to ask. But science is never settled in the way press releases claim. Methods evolve. Data improve. Models are tuned, tested, and sometimes discarded. Updating a report to reflect that constant churn is not a threat to science; it’s the sign that science—and policy informed by it—is still alive.
So yes, let’s review the National Climate Assessment. Let’s correct excesses. Let’s embrace uncertainties. Let’s publish code and data. And let’s stop using the darkest chapters of the 20th century as a rhetorical shield against the most basic tool of the scientific enterprise: skeptical inquiry.
If the Guardian wants to argue that the forthcoming review is misguided on the merits, they should roll up their sleeves and engage the details. Until then, the Stalin talk is sound and fury—signifying nothing but a fear of debate.