FCC as a Research Instrument¶

How to use the FCC Agent Team Framework as a research methodology tool for structured inquiry, multi-perspective analysis, and collaborative knowledge construction.

Conceptual Foundation¶

The Find-Create-Critique cycle maps naturally to the scientific method:

Scientific Method	FCC Phase	Activities
Observation	FIND	Literature review, data collection
Hypothesis formation	CREATE	Theory building, model design
Experimentation	CREATE	Implementation, data generation
Analysis	CRITIQUE	Evaluation, statistical testing
Peer review	CRITIQUE	Expert review, feedback incorporation
Iteration	REFINE	Revise based on critique
Publication	DELIVER	Final output, dissemination

Multi-Perspective Analysis¶

Using Personas for Research Perspectives¶

FCC personas can represent different research perspectives on the same problem. This enables structured multi-perspective analysis:

from fcc.personas.registry import PersonaRegistry
from fcc.simulation.engine import SimulationEngine

registry = PersonaRegistry.from_package_data()

# Select personas representing different research perspectives
perspectives = {
    "empirical": "research_catalyst",       # Data-driven perspective
    "theoretical": "domain_expert",         # Theory-driven perspective
    "methodological": "build_champion",     # Method-focused perspective
    "ethical": "responsible_ai_guardian",    # Ethics-focused perspective
}

for name, persona_id in perspectives.items():
    persona = registry.get(persona_id)
    print(f"\n{name.upper()} perspective ({persona.name}):")
    print(f"  Role: {persona.riscear.role}")
    print(f"  Style: {persona.riscear.style}")
    print(f"  Constraints: {persona.riscear.constraints}")

Structured Disagreement¶

Use the cross-reference matrix to identify natural tension points between perspectives:

from fcc.personas.cross_reference import CrossReferenceMatrix

matrix = CrossReferenceMatrix.from_package_data()

# Find feedback relationships (bidirectional critique)
entry = matrix.get("research_catalyst")
feedback_partners = entry.by_type.get("feedback", [])
print(f"Research Catalyst receives feedback from: {feedback_partners}")

Research Design Patterns¶

Pattern 1: Systematic Review Pipeline¶

FIND:     Research Catalyst -> Competitive Intelligence Analyst
CREATE:   Build Champion (synthesize findings)
CRITIQUE: Domain Expert -> Governance Auditor (quality check)
DELIVER:  Documentation Generator (write-up)

Use this pattern for literature reviews and meta-analyses.

Pattern 2: Mixed Methods Research¶

FIND:     Research Catalyst (qualitative) + Data Engineer (quantitative)
CREATE:   Build Champion (integrate findings)
CRITIQUE: Domain Expert (methodological review)
          + Privacy Tech Evaluator (ethical review)
DELIVER:  Documentation Generator (final report)

Use this pattern when combining qualitative and quantitative methods.

Pattern 3: Action Research¶

Cycle 1:
  FIND:    Observe current practices
  CREATE:  Design intervention
  CRITIQUE: Evaluate with stakeholders
  REFINE:  Adjust intervention

Cycle 2:
  FIND:    Observe effects of intervention
  CREATE:  Refine intervention
  CRITIQUE: Evaluate outcomes
  DELIVER: Document findings

Use this pattern for participatory and action research.

Pattern 4: Grounded Theory¶

Open Coding:      FIND phase with Research Catalyst
Axial Coding:     CREATE phase with Build Champion
Selective Coding:  CRITIQUE phase with Domain Expert
Theory Building:  Knowledge graph construction

Use this pattern for qualitative research with emergent theory.

Knowledge Graph for Research¶

Modeling Research Concepts¶

from fcc.knowledge.graph import KnowledgeGraph

kg = KnowledgeGraph()

# Research question
kg.add_node("rq1", node_type="CONCEPT", metadata={
    "type": "research_question",
    "text": "How does X affect Y?",
})

# Hypotheses
kg.add_node("h1", node_type="CONCEPT", metadata={
    "type": "hypothesis",
    "text": "X positively correlates with Y",
})
kg.add_edge("rq1", "h1", edge_type="DECOMPOSES_TO")

# Methods
kg.add_node("m1", node_type="METHOD", metadata={
    "name": "Survey analysis",
    "sample_size": 500,
})
kg.add_edge("h1", "m1", edge_type="TESTED_BY")

# Findings
kg.add_node("f1", node_type="RESULT", metadata={
    "text": "Significant positive correlation (p < 0.01)",
})
kg.add_edge("m1", "f1", edge_type="PRODUCES")
kg.add_edge("f1", "h1", edge_type="SUPPORTS")

Exporting for Publication¶

Export your research knowledge graph for supplementary materials:

from fcc.knowledge.serializers import JSONLDSerializer

serializer = JSONLDSerializer()
jsonld = serializer.serialize(kg)

# Include in supplementary materials
with open("supplementary/knowledge_graph.jsonld", "w") as f:
    f.write(jsonld)

Collaboration for Peer Review¶

Setting Up a Review Session¶

from fcc.collaboration.engine import CollaborationEngine
from fcc.collaboration.models import SessionConfig

config = SessionConfig(
    title="Manuscript Review - Study on X and Y",
    personas=["domain_expert", "research_catalyst", "responsible_ai_guardian"],
    approval_threshold=0.8,
)

engine = CollaborationEngine()
session = engine.create_session(config)

# Add reviewer comments as turns
engine.add_turn(session.id,
    persona_id="domain_expert",
    content="Methodology is sound but sample size justification is weak."
)

engine.add_turn(session.id,
    persona_id="responsible_ai_guardian",
    content="Ethical implications section needs expansion."
)

Metrics for Research Quality¶

Use the observability layer to track research process metrics:

from fcc.observability.metrics import FccMetrics

metrics = FccMetrics()

# Track research process metrics
metrics.record("papers.reviewed", 45)
metrics.record("hypotheses.tested", 3)
metrics.record("methods.applied", 2)
metrics.record("iterations.completed", 4)
metrics.record("review.score", 0.85)

Citing FCC in Publications¶

When using FCC as a research instrument, cite it as:

FCC Agent Team Framework (v1.0.1). INFORMATION COLLECTIVE, LLC.
https://github.com/rollingthunderfourtytwo-afk/l2_fcc_agent_team_ext

FAIR Workflow -- FAIR compliance
Literature Review Agents -- Automated reviews
Reproducibility Guide -- Reproducible workflows
Notebook 16_knowledge_graphs.ipynb -- Knowledge graph construction
Notebook 17_rag_pipeline.ipynb -- RAG pipeline for research
Demo open_science -- Open science workflow demonstration