What the evidence shows

This study addresses a practical gap in climate communication research: while communicating the scientific consensus ("97% of climate scientists agree") is known to shift public perception, no prior research had tested whether that effect holds up in the presence of real-world misinformation — or how to protect it if it does not. The authors frame the problem using inoculation theory, borrowed from medicine: just as a weakened pathogen can prime the immune system, forewarning people about misleading tactics and pre-emptively refuting specific misinformation ("prebunking") may confer psychological resistance.

Study 1 identified the most persuasive and widely known piece of climate misinformation in circulation — the Oregon Global Warming Petition Project, which claims over 31,000 scientists signed a statement denying a climate consensus. Study 2 then tested whether two versions of an inoculation message (a brief general warning about politically motivated misinformation tactics, and a longer detailed refutation of the Oregon Petition specifically) could protect the positive effect of consensus messaging when participants were subsequently exposed to that misinformation.

The study is notable for its political scope: results are disaggregated by party affiliation (Democrat, Independent, Republican), allowing the authors to test whether inoculation works across the political spectrum or backfires among sceptical audiences — a question with direct implications for advocates trying to communicate across partisan lines.

Key Findings

• Misinformation fully cancels the consensus effect when presented sequentially. When participants read the consensus message followed by the misinformation countermessage, the net change in perceived consensus was near zero (Mdiff = 0.51) — effectively wiping out the +19.72 point gain from the consensus message alone.

• Both inoculation messages successfully protected the consensus effect. The general inoculation preserved roughly one-third of the original consensus gain (Mdiff = 6.47); the detailed inoculation preserved approximately two-thirds (Mdiff = 12.71). Both were statistically significant versus the un-inoculated sequential condition (p < 0.001).

• Inoculation worked equally across the political spectrum. Republicans, Independents, and Democrats all showed similar protective effects from both inoculation messages — directly contradicting the "cultural cognition" hypothesis that consensus messaging backfires among ideologically sceptical audiences.

Standout data points

• +19.72 percentage points: the increase in perceived scientific consensus from the consensus-only message (CT) — a large effect (Cohen's D = 1.23 vs. control), establishing the baseline value of consensus communication that inoculation then seeks to protect.

• −8.99 percentage points: the drop in perceived consensus caused by misinformation alone (CM-only condition, Cohen's D = 0.48) — quantifying the real damage a single disinformation message can cause without any counter-framing.

• Two-thirds effect preserved: the detailed inoculation (In2 | CM) recovered approximately 64% of the original consensus-treatment effect (Mdiff = 12.71 out of 19.72), holding across Democrats (11.08), Independents (12.79), and Republicans (10.75) — the strongest evidence for cross-partisan inoculation efficacy.

Examples of how Understories uses this research

• Clarity The study shows that clear, simple consensus framing (+19.72 points) is powerful — but only when paired with an inoculation frame that explains why someone might try to muddy it. Messages that state facts without anticipating counter-narratives are vulnerable. Understories should flag unprotected consensus claims in high-misinformation contexts.

• Credibility This is foundational research for the Credibility criterion. It provides the strongest available evidence that consensus messaging is effective and that it must be defended against misinformation to remain so. 

• Solution focus Inoculation messaging addresses the credibility environment rather than solutions themselves — but restoring accurate consensus perception removes a key barrier to solution uptake.

• Autonomy The entire premise of inoculation is to give audiences the cognitive tools to evaluate information themselves, rather than simply telling them what to believe. This is the most autonomy-respecting communication approach in the library to date. Recommendations derived from this paper are consistent with Understories' principle that AI suggests, humans decide.

• Explainability Effect sizes, standard errors, confidence intervals, and ANCOVA statistics are all reported. Understories can cite specific numbers for each experimental condition. The six-condition design also allows granular comparison (general vs. detailed inoculation) that directly informs practical recommendation levels.

— non-exhaustive.

Key quote

"Both inoculation messages proved effective in protecting the positive effect of the consensus message and shifted the opinions of Republicans, Independents, and Democrats alike in a direction consistent with the conclusions of climate science."

© The Author(s) 2024. Licensed under a Creative Commons Attribution 4.0 International License (CC BY 4.0)

Last update: 15 September 2026

——

Want to stay ahead of the curve? Join the Reading List.