Thresholds

Read the Thresholds table for this study's numbers and for whether each one is enforced, reported, advice, not applied yet, or off, and change a number on the Setup page the row links to.

Who can open this: Principal investigator, Steering committee (full access), Biostatistics, Coordinator

Who this is for

Anyone who can read Setup sees the table. The principal investigator can edit the Setup pages the rows link to. Steering committee (full access) and Biostatistics can read Setup. Coordinator can read Setup and does not edit Method, Rating & consensus, or the Analysis plan there. Coder, Adjudicator, and Steering confirmer do not open Setup, and the round fold is hidden for them because it is shown only to a viewer who may read Setup.

The table changes nothing by itself. A seat that cannot edit the linked Setup page can still read the effect in the table.

Steps

1. Open Review & launch, at the end of Setup, and read "Thresholds". The introduction says: "Every number that classifies, gates, reports or protects this study's results, where it is set, and what it does. Enforced changes an outcome by itself; Reported only is shown and a person decides; Not applied yet is set in Setup but nothing acts on it."
2. On a round, open the first Review tab. A rating round uses Results. A questionnaire round uses Review. The fold is "Thresholds in force for Round {n}", with "Show" and "Hide".
3. The columns are "Threshold", "Value", "Set in", "Effect", and "Applies to". "Set in" links to Method, Rating & consensus, or Analysis plan. "Applies to" is "Classification", "Round gate", "Reporting", "Privacy", "Quality indicator", or "Stopping".
4. Read the Effect badge. The five words are "Enforced", "Reported only", "Advice only", "Not applied yet", and "Off". "Enforced" changes an outcome by itself. "Reported only" is shown and a person decides. "Advice only" produces a suggestion and nothing moves; the stopping row's own note, when that switch is on, is "Shown as a suggestion once results exist; nothing moves a round." "Not applied yet" means the number is set and described in Setup, and no code acts on it yet. "Off" means the switch is off, or the method this study chose does not use that number. Many off rows say "Switched off."
5. Use the Value cell as this study's number. For Retain, Exclude, the interquartile range, and the minimum number of numeric responses, the built-in seeds, before a workspace default replaces them, are 70, 70, at most 2, and 7. Those seeds are not a universal standard. The table is the number in force after anyone edits Setup.
6. Treat every "Not applied yet" row as described and not yet acting. When the structural decision rules are on, these two use the note "Set and described on Method; no code applies it to responses yet." When those rules are off, both are "Off" with "The structural decision rules are off for this study."
   • "Missing structure: independent panelists naming the same gap". The recommended start is 3.
   • "Areas needing restructure before structural rating stops". The recommended start is 3. While the structural rules are on, the effect is "Not applied yet" and the note is "Set and described on Method; no code applies it to responses yet." Nothing counts areas and stops the study by itself. A recorded stop blocks a later round.
7. "Subgroup comparison: concordant / borderline / minimum responses" is "Not applied yet" while the Analysis plan reports the subgroup comparison. The note is "Set and described in Rating & consensus; no subgroup comparison is computed yet." On Rating & consensus the badge is "Not computed yet" and the card says "Not computed yet: results do not show the subgroup comparison. These settings are kept for when they do." A new study seeds concordant from 70%, borderline from 50%, and a minimum of 5 responses, and a new Analysis plan leaves the subgroup comparison switched on, so a new study shows "Not applied yet". When the Analysis plan does not report it, the effect is "Off" and the note is "The Analysis plan does not report the subgroup comparison."
8. "Stability: panelists changing zone between rounds" is "Not applied yet" while stability is switched on. The note is "Set and described in Rating & consensus; stability is not computed yet." On Rating & consensus the badge is "Not computed yet" and the card says "Not computed yet: no round measures stability, so stopping advice and results do not use it. These settings are kept for when it is." A new study seeds stability on and "at most 15%" changing zone. When the switch is off, the effect is "Off" and the note is "Switched off."
9. Under the table, "Groups below the reporting floor" warns when a Size & mix target is below "Smallest reportable group". The badge is "Below floor". The suggested fixes are "Raise the target" on Size & mix (Setup, Panel design) or "lower the floor" on the Analysis plan (Setup, Governance). A new plan's floor starts at 5 people. The row itself is "Enforced": "Smaller cells show no figures in results, feedback, reports and data files; small groups are pooled as “Other”."
10. "Operational limits (not research rules)" sits under the same card. They are warnings, not Effect badges. "Spam complaints above 0.1% of sends" and "Permanent bounces above 2% of sends" each warn the study team once at least 50 messages have been sent, and "sending is never paused automatically." "Planned attrition above 50%" warns that the study would enrol more than twice its target. "Panel target below 10" warns that the panel is small for a Delphi study.

Table: the rows a new study shows as "Not applied yet", and the note on the row.

Threshold	Starting value	Note
Missing structure: independent panelists naming the same gap	3	Set and described on Method; no code applies it to responses yet.
Areas needing restructure before structural rating stops	3	Set and described on Method; no code applies it to responses yet.
Subgroup comparison: concordant / borderline / minimum responses	concordant from 70%, borderline from 50%, from 5 responses	Set and described in Rating & consensus; no subgroup comparison is computed yet.
Stability: panelists changing zone between rounds	at most 15%	Set and described in Rating & consensus; stability is not computed yet.

What the app will not do

The Thresholds card does not edit a number. "Not applied yet" means the value is stored and described, and the row's note says nothing acts on responses. That is every row in steps 6 to 8, for as long as that row's effect stays "Not applied yet". Choosing stop on the structural gate still requires at least the areas-to-stop number of areas; that check does not classify responses or choose the branch. "Advice only" does not move a round. "Reported only" does not choose the committee's outcome, including the minimum-response figures and the advisory structural-gate comparison. The response-rate row whose name is "Response rate: “block reporting” below" says "Despite its name, nothing is blocked: results and the consensus report show a note." Operational limits never pause sending by themselves.

Related

• Study design
• Round design
• Review and launch
• Coding and adjudication
• Roles and access

Checked against the app on 2026-10-31.