Comparing Delphi types
Delphi Studio offers three starting points on Create study. All three use the same platform workflow; they differ in default scales, consensus rules, and methodological emphasis.
Should you use Delphi?
Choose a Delphi design when structured judgment is needed because evidence is incomplete, inconsistent, or difficult to apply—for example, to prioritize outcomes, refine definitions, or develop recommendations with relevant stakeholders.
Choose another method when the main objective is:
- a representative estimate of population opinion;
- an empirical question that adequate data can answer directly;
- real-time negotiation or joint problem-solving; or
- endorsement of a conclusion already selected by the research team.
The defensible design tutorial provides a fuller fit assessment and protocol checklist.
Preset comparison
| Modified e-Delphi | RAND/UCLA Appropriateness | Custom | |
|---|---|---|---|
| Question type | Do experts agree? | Is this scenario appropriate? | You define |
| Default scale | 5-point Likert | 9-point appropriateness | Inherits e-Delphi |
| Primary dimension | Agreement | Appropriateness | Configurable |
| DI (Disagreement Index) | No | Yes | Optional |
| New item suggestions | Yes | No | Yes |
| Typical rounds | 2–3 | 2–3 | 2–3 |
| Best for | Guidelines, competencies, indicators | Clinical indication panels | Novel protocols |
Decision flowchart
Scale options (all presets)
After choosing a preset, the Rating & consensus step supports:
| Scale type | Use when |
|---|---|
likert_5 | Standard agreement (default for e-Delphi) |
likert_7 | Finer granularity; common in education research |
rand_ucla_9 | RAM appropriateness |
binary | Include / exclude item in a final list |
custom | Researcher-defined range with defined zones |
Stopping rules
All presets seed a maximum-round and stability configuration. After analysis, the platform produces an advisory stopping recommendation; the protocol and research team determine whether to continue, conclude with consensus, or report stable disagreement.
| Signal | Meaning | Where to see it |
|---|---|---|
consensus_reached | All active items reached a terminal classification and distributions are stable and response rate is at or above the floor | Stopping recommendation |
tentative_consensus_low_response | Consensus and stability criteria met, but response rate fell below the floor | Stopping recommendation |
stable_contested | Distributions stable but items still indeterminate | Stopping recommendation |
still_moving | Median or IQR shifting round-over-round | Stability panel |
max_rounds_reached | Configured maximum round count has been hit | Stopping recommendation |
insufficient_history | First round — no prior round to compare against | Stopping recommendation |
Anonymity model
Delphi Studio implements anonymous to peers: panelists never see each other's identities or individual ratings. Researchers use pseudonyms in routine study and analytic views; real identities are encrypted separately and may be revealed only through an authorized, reasoned, audited operational workflow.
Choosing Custom
Select Custom (explicit confirmation) when:
- You need a 7-point Likert with non-default zones
- You want multiple consensus dimensions with different thresholds
- Your IRB or registry requires documenting which defaults you reviewed
- You are piloting a novel stopping rule
Custom inherits modified e-Delphi defaults but marks every wizard section unconfirmed until you explicitly accept it — producing a stronger audit trail of researcher decisions.
Hybrid studies
Some teams run agreement on indicator importance (e-Delphi) and appropriateness on clinical scenarios (RAM) in separate studies rather than one mixed study. Delphi Studio supports one primary consensus dimension per study; use separate studies or sequential phases if both frameworks are needed.