Evidence and measurements
Reviewing extracted values
How to read a candidate value, the four review decisions, and who can give final approval.
Read the evidence
Each candidate is shown beside its evidence: the table cell, sentence or figure panel, with its page. Judge the evidence. The confidence number is only a guide to what needs attention and does not replace looking at the source.
Four decisions
A reviewer records one of four decisions on each value. Rejected and unclear values never enter the verified dataset.
- Accept: the evidence supports the value as extracted.
- Correct: the evidence supports a different value. Your correction is stored beside the reported value; the reported value and its evidence are never changed.
- Reject: the evidence does not support the value.
- Needs clarification: the value cannot be settled yet, and a note says why.
Flags name the problem
Quality-check flags say what to look at: a unit that does not fit the property, a missing condition, a possible duplicate, a missing page reference.
- Clear a flag by fixing the value, or by recording why it is right as it stands.
- A flag you cannot resolve is a reason to reject, not to accept with a note.
- Conflicting values for the same quantity go to conflict review; they are not averaged.
Who can approve
Final approval can be given by reviewers, owners, and editors whom the owner has marked "also reviewer". Approval records your name and the time, and both travel with the value into the exported dataset.
Some values cannot be approved in bulk: fields the study configuration marks as always needing a person, and every value read from a figure.
Related
Quality checks
Fixed rules, three levels of flag, and why open warnings also hold a value back.
The verified dataset
What a verified value is, how the dataset is organised, and how to filter it.
Extracting values
How candidate values are read from approved documents, what each one carries, and why a candidate is not yet data.
Figures and tables
Tables are read from their cells, plots are measured against their own axes, and every value read from a figure is reviewed by a person.