Skip to main content

Architecture of Delphi studies in Delphi Studio

Delphi Studio is designed around the core principles of the Delphi method: structured iteration, anonymity, controlled feedback, and transparent, pre-specified rules. Rather than a traditional software architecture document, this section describes the methodological architecture — how Delphi Studio organizes studies, participants, items, and rounds to support rigorous and defensible consensus research.

Overview​

A Delphi study in Delphi Studio is built around a clear, logical structure that mirrors best practices in consensus methodology:

  • A Study defines the overall research question, methodology, and rules.
  • A Panel of experts is recruited and managed.
  • An Item bank contains the statements, competencies, or scenarios to be evaluated.
  • Rounds provide the iterative structure through which panelists rate items and receive feedback.
  • Controlled feedback and analysis drive convergence toward consensus.
  • All decisions and data are captured in an auditable, reproducible record.

Studies also draw on organization-level building blocks: an expert directory for recruiting panelists across studies, reusable study templates that carry taxonomy, policy, and configuration defaults (never item statements), and optional layers for multi-tier rating designs.

This structure ensures that the study remains transparent, fair, and aligned with established standards such as CREDES.

Core entities and their relationships​

The diagram below shows the main conceptual components of a Delphi study and how they connect:

This model keeps the researcher in control of the overall process while the platform automates repetitive tasks (statistics, feedback generation, round timing) and enforces methodological safeguards.

See Key concepts for round states, ratings, zones, and collaborator roles.

The round lifecycle​

Each round represents one complete cycle of rating and feedback. Delphi Studio manages rounds through a clear sequence of states that supports both flexibility and rigor:

  1. Draft — The round is being prepared. The researcher defines which items will be included and sets the timing.
  2. Scheduled / Open — Panelists receive access and independently rate items and provide comments.
  3. Closed — No further ratings are accepted. The round is finalized.
  4. Analyzed — The system calculates statistics and classifies items. The researcher reviews results and prepares feedback.

This lifecycle ensures that every round follows a consistent, documented process. Researchers can open and close rounds manually or let the platform handle timing automatically. See Running rounds for operational detail.

Controlled feedback and iteration​

Between rounds, panelists receive controlled feedback — a structured summary showing:

  • Group-level statistics (median, spread, distribution)
  • Their own previous rating for comparison
  • Carefully reviewed, de-identified comments from other panelists

Importantly, feedback is locked before it is released. Every panelist sees the same information, which protects fairness and allows the study to be audited later. This controlled, anonymized feedback is the primary mechanism that enables the panel to move toward consensus over multiple rounds without the biases of live discussion.

Researcher oversight and governance​

While Delphi Studio automates much of the technical work, the researcher retains meaningful oversight at key points:

  • Defining the methodology and consensus rules before the study begins
  • Reviewing analysis results after each round
  • Approving the content of controlled feedback
  • Deciding whether to revise items based on panel input
  • Determining when to stop the study

This human-in-the-loop design ensures that important methodological and interpretive decisions remain with the research team, while the platform handles calculation, consistency, and record-keeping.

Auditability and reproducibility​

Delphi Studio is built to support transparent and reproducible research. Key features include:

  • All consensus rules and thresholds are frozen at the start of the study and cannot be changed without audit logging.
  • Every round's data, feedback, and classifications are recorded in detail.
  • Item revisions are tracked with full version history.
  • The complete study record can be exported in formats suitable for manuscripts and regulatory review.

These capabilities make it straightforward to demonstrate how conclusions were reached, which is essential for publication and for studies that may inform policy or clinical practice. See Audit trail and Reporting & exports.

Panelist experience​

From the panelist's perspective, the process is intentionally simple and low-burden:

  • They receive a secure magic link to access their ratings (no account creation required).
  • They rate items independently and may add comments where helpful.
  • Between rounds, they receive clear, structured feedback.
  • Their identity remains protected from other panelists.

This design respects experts' time while still capturing the rich quantitative and qualitative input needed for high-quality consensus work.