Research Platform

Open science for emotional technology research

METHODOLOGY

Research Methodology

How TEG-Blue research is conducted. Open science principles, pre-registration, open data, transparent methodology.

Open science principles

TEG-Blue research aims to follow open science principles:

  • Pre-registration of studies before data collection (where applicable)
  • Open data sharing (anonymized) via Zenodo
  • Open access publication of all findings
  • Transparent reporting of methodology and results
  • Reproducible analysis pipelines documented in public repositories

Status note: These are our working standards. Not all studies to date have been fully pre-registered. We are transparent about where current work meets these standards and where it does not yet.

Status ladder

Every claim, tool, and result in TEG-Blue carries a status label:

Established

Existing theories, measures, and findings from independent research across multiple fields. TEG-Blue builds on these but did not create them.

Proposed synthesis

The way TEG-Blue connects established theories into one interoperable map. This is the original contribution — and the part that most needs testing.

Preliminary evidence

Initial studies, pilot data, and computational analyses completed to date. See Publications for details and limitations

Open to validation

Constructs, tools, and claims that need independent replication, psychometric validation, cross-cultural testing, or external benchmarking.

Validation approach

Our validation approach aims to use multiple methods:

Inter-rater reliability

Independent raters assess the same samples to test consistent identification of regulatory states.

Convergent validity

New measures compared against established instruments (e.g., DERS, AAQ-II) to verify they capture related constructs.

Discriminant validity

Testing that measures differentiate between distinct regulatory states, not just general distress.

Ecological validity

Studies use naturalistic language samples and real-world contexts, not just laboratory settings.

Status note: The initial validation study used computational analysis of natural language. Formal psychometric validation studies using these methods are planned but not yet completed. We need collaborators to design and run these studies.

How TEG-Blue was developed

The architecture

Developed by Anna Paretas-Artacho over nearly two years of independent research, drawing on a lifetime of observing patterns in human behavior, systems thinking, personal experience, and cross-disciplinary reading.

The literature mapping

Once the architecture was established, AI research tools (including the deep thinking models of Claude, Perplexity, and Microsoft Copilot) were used to systematically identify which established theories and researchers align with each framework's propositions. The architecture determined the connections. The AI tools helped locate and organize the corresponding academic literature.

What this means

The theoretical mapping is a working hypothesis — a starting point for deeper scholarly validation, not a finished academic work. Human researchers are needed to verify accuracy, correct errors, and deepen the analysis.

Limitations

Some literature connections may be inaccurate or oversimplified. Researchers may disagree with how their work is represented. Corrections are welcomed and the mapping is updated based on scholarly feedback.

Ethical standards

All research involving human participants follows ethical guidelines:

  • Informed consent obtained before participation
  • Right to withdraw at any time without consequence
  • Data anonymization before analysis and sharing
  • No deception in study design
  • Debriefing provided after participation
  • Mental health resources offered to all participants

We take particular care with vulnerable populations and ensure appropriate support structures are in place.

Trauma-informed data architecture

The system assumes many difficult behaviors started as Protection Mode survival responses. Data systems built on this framework should not be designed to shame, profile, or exploit. This is an architectural constraint, not just an aspiration.

AI-readable research

All publications are designed for both human and AI consumption:

  • Structured JSON-LD metadata on every page
  • Semantic HTML with proper heading hierarchy
  • Dublin Core and Schema.org annotations
  • Clear, consistent terminology throughout

Where current methodology stands honestly

AreaStatus
Open science principlesWorking standard; not all work meets full pre-registration yet
Validation study (n=10,000+)Completed; computational analysis of natural language
Psychometric validation of toolsNot yet started; collaborators needed
Cross-cultural replicationNot yet started; collaborators needed
Independent replication of four-mode classificationNot yet started; collaborators needed
AI schema evaluationEarly implementation; needs formal evaluation

If you can help with any of these, see Collaborate →

TEG-Blue Research Consortium · Open Science · CC BY-NC-SA 4.0