A comparison that made the difference concrete for me. Two pieces covering the same subject, read for structure rather than for truth.
The higher-scoring one: claims attributed to named sources, numbers with methodology, disagreement acknowledged in the body rather than in a final paragraph.
The lower-scoring one: same topic, same broad conclusions, but sources referred to as “experts”, figures without provenance, and opposing views represented only as something “critics say”.
Neither article contains a false statement that I could identify. The gap is entirely in whether the reader is given the means to check.
That is the part I did not expect when I started reading this way: the difference between good and bad reporting on the same facts is mostly infrastructure — attribution, provenance, and whether disagreement is inside the argument or appended to it.
For those who evaluate sources professionally: do you weight attribution density explicitly, or is it something you assess holistically?

