Human Checkpoints
You Decide. Diverga Assists.
"Human decisions remain with humans. AI handles what's beyond human scope."
Why Checkpoints Exist
Human Checkpoints ensure you maintain full control over critical research decisions while Diverga handles the complex analysis work.
AI never makes decisions for you - every critical choice requires your approval
All decisions are documented for transparency and reproducibility
Complete audit trail for your research methodology section
Checkpoint Levels
Understanding the three levels helps you know when to expect a checkpoint and how to respond.
System STOPS immediately
Cannot proceed without your explicit approval. These are foundational research decisions.
System PAUSES
Strongly suggests your review before continuing. Important but not critical.
System ASKS
Offers you a choice with sensible defaults if you prefer to skip.
When You Will See Checkpoints
Checkpoints appear at key moments in your research journey:
| Stage | Checkpoint | Level | Decision |
|---|---|---|---|
| Starting | CP_RESEARCH_DIRECTION | π΄REQUIRED | Approve research question |
| Theory | CP_THEORY_SELECTION | π΄REQUIRED | Choose theoretical framework |
| Paradigm | CP_PARADIGM_SELECTION | π΄REQUIRED | Select research paradigm |
| Design | CP_METHODOLOGY_APPROVAL | π΄REQUIRED | Approve research design |
| Meta-Analysis | CP_META_GATE | π΄REQUIRED | Confirm meta-analysis approach |
| Analysis | CP_ANALYSIS_PLAN | π RECOMMENDED | Review analysis strategy |
How to Respond to Checkpoints
When a checkpoint appears, you have several options:
approve / μΉμΈ
Accept the current proposal and proceed
reject / μμ
Request changes or modifications
explain / μ€λͺ ν΄μ€
Get more detailed explanation before deciding
alternatives / λμ
See other options (VS methodology)
Real-World Scenarios
See how checkpoints work in practice:
Theoretical Framework Selection
π΄ CP_THEORY_SELECTIONYou're researching AI adoption in education
Situation:
The system suggests TAM (T=0.92) as the obvious choice, but VS methodology presents alternatives.
Options:
- βTAM (T=0.92) - Modal/Predictable
- βSelf-Determination Theory + TAM (T=0.6) - Emerging
- βCognitive Load Theory + Adaptive Ecosystem (T=0.4) - Novel
Your Role:
You decide which framework best fits your research context and contribution goals.
Ethics Review
π΄ CP_METHODOLOGY_APPROVALPlanning research involving human participants
Situation:
Before proceeding with data collection design, ethical considerations must be reviewed.
Options:
- βIRB protocol requirements identified
- βInformed consent procedures specified
- βData privacy measures proposed
Your Role:
You approve the ethical framework and ensure it meets your institutional requirements.
Low T-Score Warning
π΄ CP_THEORY_SELECTIONConsidering a highly experimental theoretical approach
Situation:
A T-Score below 0.3 indicates an experimental approach requiring strong justification.
Options:
- βProceed with awareness of justification burden
- βChoose a more established alternative
- βRequest hybrid approach combining established and novel
Your Role:
You acknowledge the risk and decide whether the innovation potential is worth the extra scrutiny.
Key Checkpoints Reference
Quick reference for the most important checkpoints:
| Checkpoint ID | Level | When | Decision |
|---|---|---|---|
| CP_RESEARCH_DIRECTION | π΄ REQUIRED | Finalizing research question | Approve research question and scope |
| CP_PARADIGM_SELECTION | π΄ REQUIRED | Choosing methodology | Select research paradigm (positivist, interpretivist, etc.) |
| CP_THEORY_SELECTION | π΄ REQUIRED | Selecting framework | Choose theoretical framework from VS alternatives |
| CP_METHODOLOGY_APPROVAL | π΄ REQUIRED | Design complete | Approve full research design and ethics |
| CP_META_GATE | π΄ REQUIRED | Starting meta-analysis | Confirm meta-analysis strategy and ES hierarchy |
| CP_ANALYSIS_PLAN | π RECOMMENDED | Before analysis | Review statistical analysis plan |
| CP_INTEGRATION_STRATEGY | π RECOMMENDED | Mixed methods | Review integration approach |
| CP_RESPONSE_APPROVAL | π RECOMMENDED | Peer review | Approve reviewer response letter |
| CP_VISUALIZATION_PREFERENCE | π‘ OPTIONAL | Creating diagrams | Choose visualization style |
Decision Audit Trail
Every checkpoint interaction is automatically documented:
- Timestamp and checkpoint ID for each decision
- Your choice with rationale (if provided)
- VS alternatives that were considered
- T-Score information for theory/framework choices
This audit trail can be directly incorporated into your methodology section, supporting transparency and reproducibility requirements.
Connection to VS Methodology
Human Checkpoints are the final phase (Phase 5) of the VS Methodology process:
Modal Awareness
Identify predictable/obvious choices
Long-tail Sampling
Explore less common alternatives
Context Matching
Evaluate against your specific context
Differentiated Presentation
Present options with T-Scores
Human Checkpoint
YOU make the final decision
Ready to Experience Human Checkpoints?
Start your research journey with full control over every critical decision.