Round design
Set the statements, the rating rules, and how the rounds run, and write Round 1's questionnaire only after launch creates the round.
Who can open this: Principal investigator, Steering committee (full access), Biostatistics, Coordinator
Who this is for
Round design is the third Setup group. Its sidebar label is "Round design", and its line is "Configure statements, rating scales, and feedback between rounds." On the screen the bar is "What is rated" ("Item bank", "Rating & consensus") and "How rounds run" ("Round structure", "Schedule", "Participant feedback").
Only the Principal investigator can edit Setup. Steering committee (full access), Coordinator, and Biostatistics can read it. The item bank and the rating page say "Read-only: you can't change setup for this study." A save from anyone else is "You do not have permission to change this." Coder, Adjudicator, and Steering confirmer do not get Setup.
Steps
Review & launch counts Item bank, Rating & consensus, and Rounds & feedback as three of the ten steps. An error is "Blocks launch". A warning is "Needs attention" and still counts as ready. Round structure, Schedule, and Participant feedback are three pages of the one Rounds & feedback step.
Table: which pages belong to each Round design step.
| Step | Pages |
|---|---|
| Item bank | Item bank |
| Rating & consensus | Rating & consensus |
| Rounds & feedback | Round structure, Schedule, Participant feedback |
Item bank. The heading is "Item bank": "Create and manage the statements for your study. Import, edit, organize and track versions. A statement can be used in several rounds and revised as needed; a revision keeps the original." The buttons are "Add a statement" and "Import".
When Round 1 is a rating round, launch waits on "Add at least one statement." and "Assign statements to Round 1." When Round 1 is a questionnaire, an empty bank is "Required later": the statements come from the panel's answers, and the empty bank does not block launch. A statement left outside this study's areas blocks launch. The line ends "Move it before you lock." and names the study's own word for an area. A draft is a warning: "1 statement is still a draft. Only final statements can go into a round." "Two statements have the same wording." is a warning.
Removing a statement is refused: "A statement cannot be removed from the Item bank. Retire it instead." After a register is locked, "A locked register cannot be changed or removed."
Round 1's questions are not written on this page. Round structure says "Round 1 is created when you launch the study." The study's rounds are empty until then. After launch, a questionnaire round opens "Round N questionnaire": "The preliminary questionnaire. It is separate from the Item bank and does not use the consensus rating rules." Saving questions before that round exists is refused with "That round was not found." A round that is not being built is refused with "Round 2 is not being built yet, so its content can no longer be changed here." A round that has opened or finished is "Round 2 is open or finished, so its content can no longer be changed here." Saving questionnaire content onto a rating round is "Round 2 is a rating round. Its statements come from the Item bank." See Round 1 questionnaire.
Rating & consensus. The heading is "Rating & consensus": "The scale panelists rate on, and the rules that decide whether a statement reaches consensus. Zones and consensus thresholds lock when the study launches." The sections are "Dimensions", "Rating scale", "Consensus rules", "Preview", and "Advanced methodology".
Launch waits on "Confirm the rating scale and consensus rules before launch." The button is "Confirm the rating rules". A later edit asks "The rules changed after they were confirmed. Confirm them again." The scale also has to be one the organization allows, with a minimum below its maximum, zones that sit inside the scale and do not overlap, a named "Unable to rate" choice when that choice is on, and a named primary dimension ("Choose a primary dimension for consensus.").
Panelists read neutral anchors. A decision word in a draft anchor is a warning, and the warning names the words. Stability and the subgroup comparison are labeled "Not computed yet". The numeric thresholds are listed on Review & launch; see Thresholds.
After launch the page says "The scale and the consensus rules are locked: the study has launched. Stability advice can still change. The stopping rule changes only by a recorded amendment with a reason." Consensus numbers can change from the next unopened round by a recorded amendment. There is no amendment for the scale, the zones, or RAND/UCLA.
Round structure. The heading is "Round structure": "How many rounds you run and what each one is. Rounds become real one at a time, as each one opens." "Round sequence" says "Rounds run in this order and any round can be a questionnaire or a rating round. You can add up to 10. The sequence can be adjusted before launch." Each round is "Questionnaire" or "Rating". Round 1's line is "Round 1 is created when you launch the study." A later round's line is "Round 2 is created after Round 1 closes, then opens only after you approve it." (and the same shape for each following round).
"Pilot test" is "A small test of the questions and instructions before Round 1. It is not a round, does not count toward the permitted rounds and its answers are not analyzed." A planned pilot that is not marked done blocks launch: "The pilot with N testers is planned but not marked done. Finish it before Round 1." A sign-off the plan has outgrown blocks launch: "The rounds changed after the pilot was signed off. Check the changes and sign the pilot off again." The step also blocks launch with "Add at least one rating round."
Schedule. The heading is "Schedule": "How long each round runs, the time between rounds, and when reminders go out." "Round defaults" says "Apply to every round unless you override them below." Short rounds, long rounds, a short gap, and missing or crowded reminders are warnings. They need attention and do not block launch.
Participant feedback. The heading is "Participant feedback": "What panelists see about the group between rounds." "Controlled feedback" says "What panelists see about the group before they rate again. Applies to every round after the first rating round. Rounds already open keep their own copy." Leaving out a group number, hiding a panelist's own earlier rating, or skipping investigator approval is a warning. So is "Approved rounds open on their own at the scheduled time. Each round still needs your approval first. With manual opening you choose the moment as well."
What launching locks, listed on Review & launch, includes "Pilot and Round 1 comment policy. What participants were told about comments in the first round." "Consent wording locks when recruitment opens, Round 1's time zone when it opens, and the late-response policy when its collection ends."
What the app will not do
Setup will not save Round 1's questionnaire. The round does not exist until launch, and saving questions before it exists is refused with "That round was not found."
The Item bank will not delete a statement. Retire it. A locked register cannot be changed or removed.
An empty Item bank does not block launch when Round 1 is a questionnaire. It does when Round 1 is a rating round.
Schedule and feedback warnings do not block launch. A planned pilot that is not marked done does.
Stability and the subgroup comparison stay labeled "Not computed yet".
The scale, the zones, and RAND/UCLA have no amendment after launch. The stopping rule and the consensus numbers change only by a recorded amendment.
Participant feedback has no switch for a statement's status. Panelists are never shown one. A save that tries to change that stored switch is refused: "Panelists are never shown a statement's status, so that setting cannot change."
Related
Checked against the app on 2026-10-31.