Applications for the autumn mentorship cohort are open until 30 September. Apply now
Peer review examines how reasoning was built - the evidence, the assumptions, the risks it skipped. It never examines whether the reviewers agreed with the conclusion.
A member posts their draft reasoning on a governance action, or a narrower question they want tested. Nothing is public at this stage.
Reviewers respond with structured comments: weak evidence, unexamined assumptions, unaddressed risk, missing context, and the strongest counterargument they can construct.
Each comment gets a response. Rejecting a comment is a legitimate outcome - it just has to come with reasoning.
Review never produces a recommendation, a consensus position, or a tally of what reviewers would do.
The member may publish the review thread alongside their rationale, so delegators can see how the reasoning was tested.
How expert review works
Dr. Naomi Feld · Review Panel · animated poster, production target 90 sec
What happens between submitting an analysis and defending a vote: a reviewer categorises challenges as evidence, assumption, risk, missing context, or counterargument - and the DRep is free to update or reject every one of them.
Visible only to the author.
Reviewers are adding comments.
The author has replied to each comment.
Review complete; kept private.
Thread published with the rationale.
Demo data - a fictional review thread showing the two-pane workspace.
GA-2026-041 · Ada Okonkwo (demo member) · updated 2026-07-19
Original analysis
Reviewer comments
The 92 ADA per-participant figure comes from the pilot's self-report. Is there a third-party attendance record, or is that number unverified?
Author response
Correct - it is self-reported. I have added that caveat and asked the proposer for attendance exports.
Updated analysis
Assuming comparable partner rates across all six regions ignores large differences in local delivery cost. That assumption may be doing more work than the analysis admits.
Author response
Agreed. I replaced the flat-rate assumption with a per-region range and noted the uncertainty explicitly.
Updated analysis
If funding lapses at month nine, what happens to the regional partners mid-cohort? The analysis does not consider partial-delivery failure.
Author response
Added a short section on wind-down risk and what evidence would reduce it.
Updated analysis
Two prior education programmes were funded in 2025. Their reported outcomes are relevant baseline context and are not referenced here.
Author response
Reviewing those reports now; will cite them before publishing the rationale.
Open
An argument against abstention: the disclosure is public and the relationship is minor, so a reasoned vote with the conflict on the record may serve delegators better than silence.
Author response
I considered this and still prefer abstention here. The relationship touches the delivery budget directly, and I would rather be over-cautious and say why.
Rejected with reasoning
Reviewers challenge the reasoning. They do not vote, recommend, or record a preferred outcome.
Members choose when to submit and whether to publish. Reviewers choose what they have the standing to challenge.