The short answer
Good climate information shows its source, says how sure it is, separates measurements from opinion, and fixes mistakes in public. Weak information hides the source, speaks in absolutes, or leans on a single day's weather. Six questions, in the checklist below, will sort most claims in a few minutes.
The six-question check
| Question | What good looks like | Warning sign |
|---|---|---|
| 1. Who is the source? | A named institution or author, with a link to the original data or report | "Experts say" or a screenshot with no link |
| 2. How long a period? | A trend over years or decades, compared like with like | One hot or cold day used as proof |
| 3. How sure is it? | Uncertainty stated plainly, such as a range or a confidence level | "Proven beyond doubt" or "total hoax" |
| 4. Is the number checkable? | Unit, place, year and method are stated | A figure with no year, place or unit |
| 5. Who is behind it? | Funding, background and contact details are public | No author, no funder, no contact |
| 6. Are errors corrected? | A visible corrections policy and dated updates | Edits with no notice, or no way to report a mistake |
How the experts show their confidence
The Intergovernmental Panel on Climate Change (IPCC) uses fixed terms so readers can tell how strong a statement is. Its Sixth Assessment Report synthesis expresses confidence with five qualifiers: very low, low, medium, high and very high. It maps likelihood to probabilities: likely means 66 to 100%, and very likely means 90 to 100%.
That has a practical meaning. "Very likely" is not "maybe". It means at least a 90% chance, so at most a 100 minus 90 = 10% chance of the opposite. When the same report says that human activities have "unequivocally" caused global warming, with global surface temperature reaching 1.1°C above 1850 to 1900 in 2011 to 2020, it is making one of its strongest statements, backed by measurements.
When you see a claim that is more certain than the IPCC's own wording, or much less certain with no reason given, ask why.
What "scientific consensus" means
NASA's page on the scientific consensus lists statements from many scientific bodies. It attributes the figure "about 97% of climate scientists have concluded that human-caused climate change is happening" to the American Association for the Advancement of Science (2014). It also lists a 2009 statement from 18 scientific associations and a 2005 joint statement from 11 national academies. The studies it cites include Cook et al. (2013) and a 2021 paper by Lynas et al. titled "Greater than 99% consensus on human caused climate change in the peer-reviewed scientific literature".
The numbers differ because the studies use different methods and samples. A careful writer will name the study and year instead of saying "all scientists" or "most people". That is exactly the habit to look for.
What good fact-checkers promise
The International Fact-Checking Network (IFCN) Code of Principles lists five commitments. They are a handy test for any site, including ours:
- Non-partisanship and fairness: the same standard for every fact check.
- Transparency of sources: sources are given in enough detail that readers can replicate the work, and primary sources are preferred.
- Transparency of funding and organization: funding is disclosed, and readers can contact the organization.
- Transparency of methodology: how claims are chosen, researched, written and corrected is explained.
- An open and honest corrections policy: mistakes are corrected openly.
Three worked examples
These are made-up claims to practise on, not quotes from real posts.
- "It snowed in March, so warming is fake." Question 2 catches it. Snow on one date is weather. Compare with the difference between weather and climate and ask for a 30-year trend for that place.
- "Scientists say temperatures will rise 5 degrees by next year." Question 1 and 4 catch it. Which scientists, which dataset, which place, compared with what year? If none is given, you cannot check it.
- "It is very likely that the region will see more heatwaves." This is a calibrated statement. Using the IPCC's scale, it means at least 90% probability. You would still ask which assessment it comes from and for what region and period.
A checklist to copy
Before I share this:
1. I can name the source and open it. 2. It describes a long period, not a single day. 3. It says how sure it is. 4. It gives a unit, a place and a year. 5. I know who is behind it. 6. It has a visible way to report errors.
If I tick fewer than four, I will not share it until I find a better source.
If you find a mistake on our site, please email us at hello@theclimatetruthproject.org. Our editorial principles explain how we separate evidence from lived experience.
What this checklist cannot do
A checklist helps you spot weak claims, but it cannot replace expertise. A well-sourced article can still be wrong, and a poorly written one can still be right. For technical questions, go to the original report from an institution such as the IPCC, a national meteorological service, or a university. The NASA and IFCN pages describe US and international practice, and consensus estimates vary by study.
Key takeaways and quick answers
- Check source, time period, certainty, the number itself, who is behind it, and how errors are fixed.
- The IPCC says "likely" for 66 to 100% and "very likely" for 90 to 100%.
- Consensus estimates are reported as 97% by AAAS (2014) and above 99% in a 2021 study, because methods differ.
- A transparent corrections policy is a sign of trustworthy information.
What does the IPCC mean by "likely" and "very likely"?
In the IPCC's Sixth Assessment Report synthesis, likely means a probability of 66 to 100% and very likely means 90 to 100%. Confidence is expressed separately as very low, low, medium, high or very high.
Do 97% of climate scientists really agree?
The 97% figure appears on NASA's scientific consensus page, attributed to the American Association for the Advancement of Science in 2014. Other studies cited there report different figures, including one titled "Greater than 99% consensus". Estimates vary with method, but all point the same way.
Is a fact-checking site automatically trustworthy?
No. Look for published sources, funding, method and a corrections policy. The International Fact-Checking Network's Code of Principles sets these as commitments for its signatories.