Contribution to the handbook
Supporting · Assessed 2026-10-28
The evidence specifies a measurement pipeline: text representations were placed on a question-specific axis between affirming and negating statements, with resulting scores checked against participants’ ratings. That is useful for evaluating methods, but the correlation was imperfect and other text models or axis definitions may yield different measurements.
Does this challenge the goal?
In-scope gap
The question seeks detail on how written opinions were converted into position estimates, which can inform measurement and testing of deliberative mechanisms. It does not challenge the exploration’s goal.
The inquiry concerns one study’s measurement method and does not establish that this method is suitable or required for governance.
Why this category?
- test and measurement method
- implementation alternative
- limitations of evidence
The method and its validation are reported through AI interpretations of one shared article source. The supplied evidence does not establish that the position scores capture all meaning in written opinions or generalize to other settings.
This is an AI assessment of contribution to the goal, separate from whether the question has an answer.
Tests, stop conditions and unresolved work →
Read the exploration goalDescribe how written opinions were converted into question-specific position estimates.
What we can say so far
The researchers used Sentence T5 to turn written opinions into numerical text representations (described in the article as 768-dimensional vectors). For each question, they defined an agree–disagree axis using representations of affirming and negating position statements, then measured where each opinion fell along that axis. They compared the resulting scores with participants’ earlier position ratings; the scores were correlated, but not an exact match.
What remains open
No material gap is recorded within this part’s current scope.
Conditions and limits
- The descriptions are interpretations of the same article source, not independent corroboration.
- The article copy is described as fully readable, but supplementary materials are not included.
- The reported correlation with earlier ratings does not show that the scores capture every nuance of the written opinions.
- The authors note that other text models or axis definitions could yield different measurements.
Objections to consider
- The reported correlation does not establish that the numerical scores preserve every meaning or nuance in the written opinions.
- Alternative text models or position-axis definitions may produce different estimates; the supplied evidence does not compare those alternatives.
Evidence and related ideas
Supporting evidence, qualifications and context play different roles in this answer.
context
Participants ranked and critiqued statements Participants privately wrote a short paragraph about their opinion; they typically gave both their position and a reason. The AI helper generated candidate group statements for participants to rate and rank. After the top-ranked statement was critiqued, the AI generated revised candidates for another ranking. Participants compared the initial and revised winners and answered a brief survey about whether the process affected their views. They also rated their positions on the questions immediately before and after the full set of deliberation rounds.
Describes participants’ written opinions and the study procedure, but not the conversion of opinions into position scores.
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) ↗