Sufficient consensus
Information quality frameworks assess accuracy, completeness, currency, credibility and relevance. Guidelines for evaluating pages, including experience, expertise, authoritativeness and trust, are public and detailed.
Understand that it is the result that counts.
What the result pages leave out
These belong to the subject. People ask about them. They are missing from the agreed coverage.
- Verifiability by the reader. Most quality signals are about the publisher. Very few tell a reader how to check the claim themselves.
- Completeness for whom. A page can be complete for a subject and useless for the decision the reader arrived with.
- Confident wrongness. Fluency raises perceived quality independently of accuracy, and machine-generated text is fluent by construction.
- The stated limit. Naming where an approach fails is the strongest quality signal available and it appears almost nowhere.
- Age. Undated material is read as current. A large share of what ranks was written for a situation that no longer exists.
What people actually want to know
- How would I check this myself?
- When was this written, and what has changed since?
- Where does this approach not work?
- Is this complete for the subject, or complete for my decision?
Date it, state the limits, and give the reader a way to verify one claim without your help. The rest of the quality discussion follows from that.
More in social epistemology
Search engine epistemologyTen results that agree are ten results that were selected by the same criteria.Algorithmic biasNeutral procedure applied to unequal material produces unequal output.Filter bubblesThe result confirms the assumption that produced the query.Epistemic trustTrust is the mechanism.Echo chambersAdding facts to it changes nothing, because the facts arrive through a discredited channel.Distributed epistemologyThe part that is written down is the part that had a reason to be written down.Epistemic injusticeA search result inherits that and turns it into an ordering.