Contribution to the handbook
Priority · Assessed 2026-10-28
This is a useful test of whether a deliberative system’s group statements stay within the range of participants’ views—a potentially consequential safeguard against overstating consensus. The evidence reports that 96% of statement scores were within members’ range, but this score-based result does not show that members agreed with every part of the wording. The text-position measure and its limitations are relevant to designing and testing citizen-input processes, not proof of suitability for public decisions.
Does this challenge the goal?
None identified
The question examines how to measure whether group statements reflect members’ positions; it does not contest the handbook’s goal or boundary.
The study’s measurement concerns group statements in its own setting, not whether citizens would regard a statement as fair or accept its wording.
Why this category?
- Consequential representation and consensus-testing measure
- Evidence includes a quantified result and a limitation on its interpretation
The records derive from the same source root, so they are not independent corroboration. The supplied evidence does not establish how these measures perform in other populations or real governance decisions.
This is an AI assessment of contribution to the goal, separate from whether the question has an answer.
Authority, accountability and resilience →
Tests, stop conditions and unresolved work →
Read the exploration goalState whether, and how consistently, group-statement positions fell within the range of members’ positions.
What we can say so far
Yes. The article reports that 96% of group-statement scores fell within the range of the corresponding group members’ scores. This indicates that nearly all statements were within that measured range. The result concerns text-derived position scores; being within the range does not establish that every member agreed with every part of a statement’s wording.
What remains open
- The supplied evidence does not give a question-by-question breakdown of the 96% result.
Conditions and limits
- The 96% result is from the supplied account of the article; the evidence items share source ancestry and are not independent corroboration.
- Positions are measured using text-derived scores, and a score within the range does not prove agreement with all aspects of the statement.
Objections to consider
- The text-score measure’s match with earlier ratings was correlated but not exact, leaving uncertainty about how fully the scores capture the written opinions.
- The reported result depends on a particular text model and axis definition; the supplied evidence notes that alternative choices are possible but does not show whether they would change the 96% finding.
Evidence and related ideas
Supporting evidence, qualifications and context play different roles in this answer.
qualifies
Text scores used a question-specific scale The researchers encoded each written opinion as a 768-number text vector, then rotated those vectors onto a question-specific line running from agreement to disagreement. They checked the resulting position scores against participants’ earlier ratings; the scores were correlated, but the match was not exact.
Describes the text-score measure and notes its correlation with earlier ratings was not exact, qualifying interpretation of the measured positions.
AI can help humans find common ground in democratic deliberation
Read full source: AI can help humans find common ground in democratic deliberation (opens in a new tab) ↗Read source discussion: AI can help humans find common ground in democratic deliberation →