Understanding results

Evidence and transcripts

How answers are collected, and how to trace any figure back to who gave it.

Collection

Experts answer in one of three ways, set by collection_method: a written form, an AI-moderated interview, or an interview with a Synquery researcher. With any, Synquery chooses per expert. For interviews, Synquery extracts the values for your answer fields from the transcript. Whatever the method, the expert reviews the final structured answer and confirms it before it is released to you. responses[].collection says which method was used.

Transcripts and recordings

For interview answers, the transcript and recording are available in the Synquery app, and quote_start / quote_end locate each quote in them. They are absent for written answers. Transcripts can also be requested through support.

Tracing a figure to its source

Every value in summary is calculated from the fields of the responses on that request. Each response carries a stable respondent_id, the expert's own quote, and the basis and basis_scope of their answer, so any figure can be traced back to who gave it and what it rests on.

Who the experts are

Identity, employment history, expertise and conflicts are checked before an expert answers. See Expert verification.

Next

API reference: Submit a question →

Quote with transcript positionjson
{
  "respondent_id": "exp_4k9a2",
  "collection": "ai_interview",
  "quote": "Every fellow we hire now expects a robot in the room. That's what's moving our numbers.",
  "quote_start": "00:04:12",
  "quote_end": "00:04:31",
  "basis": "first_hand",
  "basis_scope": "Own institution; national figure is an estimate"
}
Written answer: no transcriptjson
{
  "respondent_id": "exp_7m3q8",
  "collection": "written",
  "quote": "We bought the robot to keep two surgeons from leaving, not because the volume justified it.",
  "quote_start": null,
  "quote_end": null
}