Blog

Shaped by the Data We See

Jan Gunash · Co-Founder & CGO · June 1, 2026

The best questions are not always the comfortable ones. Like: if the physical evidence for climate change has never been stronger, and the energy transition has never moved faster, why are more people than ever convinced that none of it will be enough? And if that pessimism is partly a data problem, what would it actually take to fix it?

New data published this week from Pew Research Center, the European Commission's Eurobarometer, and a range of global attitude surveys paint a picture that is more interesting than a simple story about public doubt. Different countries are arriving at very different conclusions about the same physical reality. That divergence is not random. It reflects the data environments people live in, not just the facts on the ground.

What the Global Data Shows

Pew Research Center surveyed 3,524 U.S. adults in March 2026. About six in ten Americans say countries around the world will not do enough to avoid the worst effects of climate change. About half say technology companies can do a lot to address climate change, but few expect them to follow through (source: Pew Research Center).

The European picture is strikingly different. The 2025 Eurobarometer found that 85% of EU citizens believe climate change is a serious problem, 81% support the EU's goal of becoming climate neutral by 2050, and almost nine in ten support more EU action to boost renewables and energy efficiency. A growing share, now 38%, report feeling personally exposed to environmental and climate-related risks (source: European Commission Eurobarometer).

China tells a different story again. Chinese respondents are notably more optimistic than either Europeans or Americans, with 77% believing climate action will have a positive impact on their quality of life, compared to 65% of Americans and 61% of Europeans (source: EIB Climate Survey).

And yet beneath Europe's strong headline numbers sits a more complex reality. A latent class analysis across 23 European countries identified four distinct public groups: Engaged at 18%, Pessimistic at 18%, Indifferent at 42%, and Doubtful at 21% (source: ScienceDirect). The largest single group in Europe is not concerned, not pessimistic, and not doubtful. It is indifferent. High concern at the survey level does not translate automatically into active engagement.

The Paradox in the Data

These surveys land in a specific context that makes them more interesting than a simple story about public attitudes. The same week that Americans are expressing record doubt about whether the transition will happen, the physical evidence shows it is already happening faster than most models predicted.

China and India cut coal simultaneously for the first time in 52 years in 2025. The EU crossed the renewable energy threshold for the first time, with wind and solar generating more electricity than fossil fuels. Global battery storage installations hit record levels. Solar is on track to become the world's largest electricity source within six years.

The physical transition and the public perception of it are moving in opposite directions in some countries and in different directions entirely across countries. That divergence is not a communication problem in the conventional sense. It is a data infrastructure failure with a specific mechanism.

What the Data Actually Shows

Most conversations about climate perception treat it as a communication problem. The right message, the right messenger, the right platform. But new data published this week by Pew Research Center, combined with an original analysis we ran on Panthaion using the Yale Climate Opinion Maps 2025, suggests the problem is more specific and more structural than that.

We took state-level climate opinion data for all 50 U.S. states and built a simple measure: the gap between the percentage of people who believe climate change is happening and the percentage who say they are worried about it. Nationally that gap is 6 percentage points. 69% of Americans believe it is happening. Only 63% are worried about it (source: Yale Climate Opinion Maps 2025, analyzed on Panthaion).

That 6 points is the story.

The scatter plot below places every U.S. state as a labelled dot. The x-axis shows belief. The y-axis shows support for renewable energy funding. Dot size reflects how worried people are. Colour shows the belief-worry gap: cyan means the two are well aligned, orange means far more people know than feel urgency.

Three findings stand out from the analysis. First, the gap between knowing and feeling is widest in some of the states most physically exposed to climate risk. Florida, Maine, and Montana all show gaps above 7.5 percentage points. Second, despite that worry gap, support for funding renewable energy is strong in every single state, ranging from 71% to 84%. The policy window is open even where urgency is not. Third, and most importantly for the data infrastructure argument: the states with the widest gaps are not the ones with the weakest belief. They are the ones where the data about climate impacts has not yet translated into felt personal relevance.

That is not a communication failure in the conventional sense. It is a data infrastructure failure. The physical evidence is accumulating faster than the systems that make it personally legible. A farmer in Montana knows climate change is real. The data that would show him what it means for his specific watershed, his specific growing season, and his specific energy costs does not reach him in a form he can act on.

The full methodology, state-by-state findings, and interactive version of this analysis are published on Panthaion at panthaion.org/p/when-the-data-does-not-reach-the-people-504975f0.

The Data Problem Underneath the Perception Gap

The gap between physical climate reality and public perception of it reflects a fundamental asymmetry in how different kinds of climate data travel.

The physical evidence for the energy transition reaching inflection points is being produced by a small number of analysts synthesizing data from multiple national energy agencies, satellite systems, and scientific databases. Carbon Brief's analysis of the China and India coal finding was produced by a single analyst working across data sources that do not speak to each other. The Ember analysis of the EU renewable threshold aggregated across 27 national reporting systems with different methodologies and different lags. The IEA's landmark data on renewable deployment requires paid subscriptions or institutional access for the detailed datasets.

The result is that the most important positive data points about the energy transition exist but reach the public slowly, filtered through specialist media, condensed into headlines that lose the statistical context, and often drowned out by the immediate visibility of extreme weather events, which are vivid, local, and emotionally resonant in a way that a chart showing annual renewable deployment is not (source: PlanA.Earth).

Meanwhile the data infrastructure that produces climate concern travels fast, accessible, and emotionally compelling. Extreme weather events are immediate and visual. Satellite imagery of floods and fires is freely available and widely shared. The result is a systematic asymmetry: data that documents genuine transition progress travels slowly and narrowly, while data about climate impacts travels fast and far (source: Cambridge Global Sustainability).

The divergence between Chinese optimism, European indifference, and American pessimism partly reflects this dynamic playing out differently across different media and governance environments. China's state-coordinated communication about its own clean energy buildout, which is genuinely the largest in human history, produces a more consistent narrative about transition progress. Europe's exposure to extreme weather events, which are vivid and local, drives concern even where engagement is low. The U.S. data environment is fragmented and heavily weighted toward conflict and failure narratives on climate (source: Pew Research Center / Cambridge Global Sustainability).

The New Denial

There is a second layer to the data problem that the surveys do not fully surface. Research published in Cambridge's Global Sustainability journal documents a significant shift in climate misinformation over recent years. The conversation has largely moved away from denying that climate change exists toward condemning solutions as ineffective, risky, or harmful. This new denial targets the transition directly, attacking renewable energy, electric vehicles, and climate policy with arguments that are harder to counter because they engage with real tradeoffs and cost concerns (source: Cambridge Global Sustainability).

The data infrastructure to counter this shift does not yet exist at scale. AI detection systems are poorly equipped to handle synthetic climate-related imagery showing false disaster scenarios or fabricated infrastructure failures, a gap documented by the 2026 World Disasters Report (source: TechPolicy.Press). The COP30 Belém Declaration on Information Integrity and Climate Change, agreed in November 2025, was the first time information integrity was formally embedded in international climate negotiations. Declarations are not data systems, but the formal recognition matters (source: WACC).

What Mitigation Actually Looks Like

The perception gap is a solvable problem. The specific mechanisms are identifiable and the interventions are concrete.

The first requirement is public, accessible, real-time data on the energy transition in formats that do not require specialist knowledge to interpret. The IEA, Carbon Brief, Ember, and a small number of other organisations are producing this data. Most of it sits behind institutional access or in technical formats. Making it genuinely public and genuinely accessible is a governance choice, not a technical one. The data exists. The decision to make it available at the speed and accessibility level that social media operates at has not yet been made.

The second requirement is local and personal data. National and global statistics are abstract. The data that changes individual perception is the data that connects climate change to lived experience at a personal and community scale. What is the flood risk for a specific postcode? How much has the average temperature in a specific city changed in twenty years? How much would a household save in energy costs under different transition scenarios? This data exists in government datasets, satellite archives, and climate models. The infrastructure to make it personally relevant and instantly accessible does not (source: PlanA.Earth).

The third requirement is counter-narrative data infrastructure that operates at the speed of the misinformation it is countering. The ClimateCheck 2026 shared task, which brought together AI researchers to build fact-checking systems for climate claims, is a step in this direction. The systems produced are not yet operating at the scale or speed of the misinformation environment they are designed to address (source: ClimateCheck 2026 / arXiv).

The fourth and most structural requirement is mandatory, standardised, public reporting from the industries and institutions that most shape public understanding of the transition. If corporate climate commitments were verified against actual grid mix rather than purchased credits, the gap between claimed and actual progress would be visible rather than hidden. Transparency is a necessary condition for accurate public perception, even if it is not a sufficient one.

What the Surveys Are Actually Measuring

The Pew finding, the Eurobarometer, and the Chinese optimism data are not primarily measuring what people believe about climate science. They are measuring what data people have been exposed to and how that data reached them.

American pessimism about collective action is partly a product of a data environment that makes climate failures vivid and transition progress invisible. European indifference, despite high headline concern, is partly a product of the gap between what people are told in surveys and what they encounter in daily life. Chinese optimism is partly a product of a data environment in which the state's own extraordinary clean energy buildout is a consistent and visible feature of public information.

None of these is simply right or wrong. All of them reflect the data infrastructure that produces them. The Panthaion analysis makes this visible at the state level: the places where the belief-worry gap is widest are not the places where people know the least. They are the places where the data has not yet made the reality feel close enough to act on.

The transition is happening. The systems designed to make that legible to the people whose engagement it depends on are not keeping pace. That is the data gap this newsletter has been documenting since March.

Sources:

Want to contribute your thoughts or findings in data and AI with climate, energy, or nature? Message Jan Gunash on LinkedIn or send an email to jgunash@projectcoeus.com.