What Does Not Supported Verdict Mean in Fact Checking
“huh? confidence low? verdict: not supported? what does that mean?”
Summary
A low confidence rating signals that the fact‑checkers have limited certainty about the claim, usually because the evidence is sparse, ambiguous, or contradictory. When the verdict reads “not supported,” it means the claim lacks sufficient reliable evidence to be confirmed and is therefore treated as unsubstantiated. This interpretation aligns with standard fact‑checking methodology and statistical confidence reporting.
Sources 57 searched
- Cognitive Biases in Fact-Checking and Their Countermeasures: A Review - ScienceDirect
Overconfidence Effect (Dunning, Griffin, Milojkovic, & Ross, 1990). To be too confident in one’s own answers.
- Crowdsourced Fact-checking: Does It Actually Work? - ScienceDirect
Nevertheless, each of these groups of workers achieves a lower accuracy compared to the whole crowd from Our Study. This result confirms the importance of diversity in the background and political beliefs of workers when a statement is evaluated. We also observe statistically significant correlations ( ... 005) between the self-reported workers’ confidence and the truthfulness assessments of Soprano et al.
- Scaling up fact-checking using the wisdom of crowds | Science Advances
Error bars indicate 95% confidence intervals. To assess how well the layperson judgments correlate with the fact-checkers’ in absolute terms, we use the correlation between the fact-checker ratings as a benchmark (34). Perhaps surprisingly, even with a small number of layperson ratings per article, the correlation between the laypeople and the fact-checkers is not significantly lower than the correlation among the fact-checkers.
- Cross-checking journalistic fact-checkers: The role of sampling and scaling in interpreting false and misleading statements - PMC
The critical issues evaluated in this work are how much two high profile fact-checkers observing U.S. politics—The Washington Post’s Fact Checker and PolitiFact—agree on what to fact-check and, when they do check the same facts, the degree of veracity agreement. Should two (or more) fact-checkers overwhelmingly agree on what is worth fact-checking and agree in their assessments of the factual accuracy, news consumers should have trust and confidence in a consistent set of facts from news providers to stay informed.
- Community notes increase trust in fact-checking on social media - PMC
The 95% confidence intervals (error bars) were derived using the bootstrap method for 1,000 resamples. Full results are in Tables S10 to S16. Figure 5a shows that under the control condition (in which no fact-check was shown), participants performed better in correctly identifying misleading ...
- The effectiveness of short‐format refutational fact‐checks - PMC
Our first main hypothesis was that detailed refutations would be more effective than simple retractions. We specified that: (H1a) False‐claim belief will be lower after refutations than retractions. (H1b) Inferential‐reasoning scores will be lower, reflecting lower endorsement of the contested claim, after refutations than retractions.
- The Principles of the Truth-O-Meter: How we fact-check
Decisions on coverage, what facts to check, and Truth-O-Meter ratings are determined solely by PolitiFact’s independent journalists. PolitiFact seeks to present the true facts, unaffected by agenda or biases. Our journalists set their own opinions aside as they work to uphold principles of independence and fairness.
- Harvard Kennedy School Misinformation Review1 October 2023, Volume 4, Issue 5
Snopes and PolitiFact in their fact-checking verdicts during the studied period.
- HKS Misinformation Review“Fact-checking” fact checkers: A data-driven approach | HKS Misinformation Review
Rating discrepancies are attributed to various systematic factors: 1) differences in the granularity of verdict ratings, 2) differences in focus between two fact checkers, 3) similar claims but subtle differences in the key information to ...
This fact check is free and donation-based. $1 powers ~30 fact-checks.
