Contributions
- Directly tests the "backfire effect" hypothesis (Nyhan & Reifler 2010) across 52 distinct political issues in five separate experiments.
- Enrolls 10,100+ subjects and finds zero backlash corrections across all conditions.
- Shows that factual correction works even on high-salience, polarizing issues (e.g., WMD in Iraq, tax policy, abortion).
- Demonstrates that both liberals (85% correction rate) and conservatives (83%) accept factually accurate information despite partisan origin of the claims.
- Tests hypotheses about when backfire might occur (high polarization, counter-argument effects, issue complexity) and finds none supported.
Method
The authors designed five experiments using real statements from political leaders (Clinton, Obama, Trump, etc.) on ideologically charged topics. In each experiment:
- Subjects read an original political claim designed to diverge from factual reality.
- Subjects were randomly assigned to read a correction citing neutral government sources, or not.
- Subjects rated their agreement on a five-point Likert scale.
Issues tested included immigration, abortion, healthcare, defense spending, job statistics, and Iraq WMD — all high-salience for partisan audiences. The authors deliberately selected topics and speakers (e.g., Trump on tax policy, Obama on immigration) where backfire would be most likely if it existed.
Each study used a randomized latin squares design. Subjects saw statements presented as if from news articles; corrections were framed as retractions from neutral government sources.
Results
- Experiment 1 (8 issues): Among liberals, 85% of issues saw factual correction; among conservatives, 83%.
- Experiments 2–5: No single issue, across all 52 tested, produced evidence of backfire.
- Corrections reduced agreement with false claims regardless of ideological alignment.
- Correction size and correction proximity (whether the correction was proximal or distal to the claim) did not predict backfire.
- No subset of respondents showed systematic backfire on any issue.
The key finding: When citizens encounter factual information that contradicts a political leader's claim, they heed the correction even when it cuts against their ideology and even when it challenges a sympathetic figure. The "backfire effect" does not appear.
Notes
This paper is a landmark replication and extension of Nyhan & Reifler (2010). It is one of the most cited empirical challenges to the backfire effect, and its null findings have significantly shifted how researchers and practitioners think about fact-checking efficacy. The paper is careful to note that backfire may occur in specific conditions (e.g., at moments of high polarization or in certain demographics), but it does not appear as a general phenomenon across the conditions tested. The authors argue that ideology serves as a heuristic filter, not an irrevocable barrier to factual learning.