The Checklist and the Score
An audit's question list is built in the survey designer and kept as a checklist template. When an audit chooses a template, it takes its own copy of the questions — so editing the template later never changes an audit that has already been answered.
Three properties make a question scorable
The designer's property grid carries three extra properties on top of the usual survey ones. They are what turn a questionnaire into an audit checklist.
| Property | Where it sits | What it does |
|---|---|---|
| Score | On each choice | What that answer is worth. |
| Conformity | On each choice | How the answer reads against the criterion: Conformant, Non-conformant, Observation, or Not applicable. |
| Weight | On the question | How much this question counts relative to the others. Defaults to 1. |
A question with no choices — free text, a file upload, a heading — carries no score and is simply skipped.
How the total is worked out
For each scored question:
- Earned is the score of the chosen answer. For a multi-select, the sum of the chosen answers.
- Maximum is the best the question could have done: for a single choice, the highest-scoring option; for a multi-select, the sum of every option worth having.
- Both are multiplied by the question's weight before being added to the totals.
The audit then carries earned, maximum and the percentage between them.
Not applicable drops out of the denominator
A choice marked Not applicable does not score zero — it removes the question from the calculation entirely, on both sides. A checklist written for the whole organization can therefore be used on a unit that genuinely has no such process, without that unit being punished for it.
Unanswered questions
A question left unanswered counts towards neither side. The audit's score reflects what was actually assessed.
An answer that is no longer among the choices — because someone edited the template after the fact — is treated as unanswered rather than scored, so a choice removed later cannot quietly change a past audit's number.
Multi-select ceilings
If the chosen options on a multi-select add up past what the question could be worth, the earned figure is capped at that ceiling rather than refused. The schema is the organization's to write, and the score should not exceed 100% because of how it was written.
The score is the server's answer
The survey library totals a score in the browser as the auditor works, and that running total is useful. It is not the number that gets stored.
The stored score is calculated on the server, from the audit's own copy of the schema and the answers as saved. An audit score is evidence, and evidence is not taken from the client.
A template can be marked not scored. Its questions are then a checklist in the ordinary sense — answered, recorded, used to raise findings — with no number at the end.
Raising a finding from a question
A non-conformant answer is not automatically a finding. Someone decides.
The checklist offers to raise a finding from the question being answered, and a finding raised this way keeps the question's text, copied at the time. Questions live inside the audit's own copy of the schema rather than as records of their own, so carrying the text is the only way the finding can still say what it was about once the list has moved on.