Running rounds
This page is the procedure for one round after the study has launched. It is not the results explanation. How to read a status and choose the next step is Interpreting results. What the measures mean is Consensus analytics.
You are done with a round when it is Closed, which is a read-only record, and either another round is being prepared or the study has no further round to open.
Before you start
The study is past draft. Launch created the rounds. It did not open round 1, and it did not send an invitation. In the study navigation, round 1 is Building and each later round is Planned. On the Round design screen, before launch, those same places are labelled Preparing and Template. People answer only after a round is open, and only if they are enrolled and eligible. See Panel management.
The round plan can store a questionnaire on a later round. The study navigation treats a questionnaire as a questionnaire only when it is round 1. A later round stored as a questionnaire is labelled Rating. Round 1 questions are Round 1 questionnaire.
How a round moves
One round is in progress at a time. The control you see is the next move, and only for a seat that holds it.
| From | Control | Who | To |
|---|---|---|---|
| Planned | Start preparing this round | Principal investigator | Building |
| Building | Open this round | Principal investigator or Coordinator, after a valid approval | Open |
| Open | Close collection | Principal investigator or Coordinator | Reviewing |
| Reviewing | Move to the decision | Principal investigator or Steering committee (full access) | Decision |
| Decision | Close this round | Principal investigator or Steering committee (full access) | Closed |
A later round cannot be prepared while an earlier one is still in progress. Start preparing this round is refused until the earlier round has closed.
1. Approve, then open
Approve this round is the principal investigator only. The dialog title is Approve this round? Approval records the decision for this content, these rules, the consent form, and this schedule. It does not open the round. A relevant change makes the approval out of date. The button is then Approve again. The line under a missing approval says "Only the principal investigator approves a round." Nothing on a schedule approves a round. A schedule opens an approved round on its own only when Round design is set to Open approved rounds on their own at the scheduled time and the round is set to open On the scheduled date, and the approval is still valid. On the scheduled date together with I open each approved round myself does not open the round.
Open this round fixes the round's panel from the people who are enrolled and eligible at that moment. The dialog states how many panelists it goes to. If nobody is eligible, the round cannot open. The message tells you to enroll people or to fix the Panelists page.
Panelists are emailed that the round is open only if the Round open message has already been saved, or if you tick Save this default wording as version 1 before continuing in the open dialog. That tick applies to later rounds too.
Panelists answer through their round link. Answers save as they go and can be changed until the cutoff, which is the deadline plus any grace window. A late answer is refused. A new study is set to close At the deadline. The schedule then closes collection once that cutoff has passed. Until that job runs, the study home can say Past deadline, close collection. If the round is set to close When you close it, a person uses Close collection.
2. Close collection and read results
Close collection ends collection and moves the round to review. It does not compute results. The dialog reports how many panelists have responded and how many have not, and it repeats the Setup rule for people who do not respond. A response rate does not block reporting, and it is not a stopping rule. Stopping advice does not close the round or change a disposition.
On the round's Results page, Compute results builds the figures from submitted responses. The principal investigator, Steering committee (full access), and Biostatistics can run it. If stored results cannot be trusted, the page says so and does not show figures. Move to the decision is a separate control after review.
A later rating round cannot be approved or opened until the earlier round has closed, its decisions are confirmed, and it has computed results. That check does not freeze the figures. After opening, panelists read the earlier round's current verified results, which can differ from what they would have read when the round opened. There is no feedback lock.
3. Confirm the researcher decision
The software does not choose the scientific outcome. On the Decision stage a person confirms the plan. The dialog title is Confirm Round N's decisions? The confirm control is Confirm decisions.
While another round follows, the plan tiles use the study's labels, including Rated again for a statement that goes on unchanged. That disposition is Advance. Revise is a reword, not the automatic path for an uncertain or polarized statement. Hold means the statement is not rated again. The map from a class to those tiles is From a result to the next step.
Who confirms a rule-determined outcome is the choice on Study team: The principal investigator or Biostatistics, or Biostatistics only. Outcomes the rules send to steering review are decided by the steering committee. A decision that departs from what the rules imply is an override and needs a reason of at least 20 characters.
On the last planned round, the Decision tiles use Final: no consensus for a statement that would have been rated again and did not come from a clarity review, and Final: deferred, still unclear for a clarity keep with no round left. The shorter label No consensus is still a decision label: it is what Rated again is called when no next round exists, and the override line prints it. A count line can say "no consensus" for the Final: no consensus group. The protocol class No consensus is a different label. It covers uncertain and polarized. See After the last round.
Close this round makes the round read-only. You can then prepare the next planned round.
What this procedure will not do
- It will not open a round that the principal investigator has not approved.
- It will not treat a response-rate row as a bar on reporting or as a reason to add a round. Stopping advice does not add a round. The principal investigator can add one further rating round as a recorded amendment. That rule is on the governed study record.
- It will not show panelists a classification. What they see is Panelist experience.
If the round will not open
Use Round will not open.
Next step
Read the closed round on Interpreting results, then prepare the next round or complete the study from The Delphi lifecycle.