The rating measures a journalist’s reliability based on the accuracy of their claims. It considers true, false, unverifiable, and pending claims, but with additional weight applied to unverifiable claims.
The rating is calculated using the following formula:
((T - F - (Wu × U)) / (T + F + U)) * (1 - e^(-α * (T + F + U))) * 100
Where:
Positive numbers mean higher reliability (mostly true claims). Numbers closer to zero means mixed reliability (true and false claims). Negative numbers mean lower reliability (more false claims than true).
The term (Wu × U) adjusts the impact of unverifiable claims on the rating. Unverifiable claims are included because they introduce uncertainty about a source's reliability. If a source frequently makes claims that cannot be proven true or false, it raises questions about the accuracy of their reporting. By applying a weight factor (Wu), these claims decrease the score slightly to account for this uncertainty, ensuring that journalists who consistently provide verifiable information are rated more favorably than those who do not.
This emphasizes the importance of unverifiable claims in assessing overall reliability, while still acknowledging their more neutral nature compared to outright false claims.
Pending claims are neutral and don’t affect the score until resolved.
The confidence factor prevents inflated ratings for journalists with few claims. It increases with the number of claims, making the rating more reliable as more claims are made.