- What reaches them
- A one-page script with the objection, the minute it lands, and the three things the answer is scored on.
- What it means from there
- What the panel disagreed about was never the candidate — it was which objection each interviewer raised and how hard. Fixing the prompt leaves only the answer to disagree about.
- What happens next
- Writes the script once per role, and holds the panel to scoring the response rather than the room.
Critique injected on script in the collaborative exercise
The collaborative round carries a written script: the same objection is raised at the same point for every candidate, and the panel records what was said in response rather than how the room felt.
What it looks like when this intervention is in place
Composite excerpts, written to be typical rather than copied. No particular message, company or person is quoted here.
- 00:18 Scripted at minute 18, verbatim for every candidate: "I think this design is wrong — you have put the queue in the wrong place."
- 00:19 Recorded: asked which failure the objection was about; restated the trade-off; proposed moving it and named what that would cost.
- 00:31 Not recorded: tone, warmth, whether the room enjoyed it. The rubric has no column for it, so no column carries it.
What to notice: Friction in a collaborative exercise is a real signal, and this does not deny it — it makes it comparable. Unscripted, the objection arrives when an interviewer happens to feel like raising it, in words chosen for that candidate, and what comes back is scored against a memory of how other rounds went. Scripted, the same prompt produces answers that can be set side by side.
What this targets
Expected effects
- mech.communication_or_working_style_friction is measured against a common prompt rather than inferred from an interviewer's chosen one
- Interviewer variance stops entering the verdict as if it were candidate behaviour
What to measure
scripted_prompt_coveragepanel_inter_rater_agreement
Reached without being linked
The relations above are one step. This is what the entry reaches only by walking through them — the shape of it, since the list runs long in a registry this cross-referenced.
- 9
- Observation2Barrier7
From each side
The same entry from inside each position that meets it: what reaches them, what it means from there, and what happens next given what they control.
- What reaches them
- An objection that arrives as part of the exercise, and a rubric that names what the answer is being read for.
- What it means from there
- The pushback is the test, not a reaction to something already gone wrong — so it can be answered on the merits instead of managed.
- What happens next
- Answers the objection rather than reading the room for what caused it.
- What reaches them
- Panel write-ups that name a scripted moment and quote the response, instead of arriving as impressions.
- What it means from there
- A write-up quoting a common prompt is the only kind that can be compared between candidates or defended after the fact.
- What happens next
- Returns the quoted exchange as the feedback, and refuses write-ups that score the room.
ATS vendorNot visible from here
From this position the entry is not visible. That is not a gap in the atlas — it is the finding: the decision is made where nobody in this position can see it.
ClientNot visible from here
From this position the entry is not visible. That is not a gap in the atlas — it is the finding: the decision is made where nobody in this position can see it.
Employer policyNot visible from here
From this position the entry is not visible. That is not a gap in the atlas — it is the finding: the decision is made where nobody in this position can see it.
Public policy and industry standardsNot visible from here
From this position the entry is not visible. That is not a gap in the atlas — it is the finding: the decision is made where nobody in this position can see it.