THE GOLD STANDARD OF INTELLIGENCE · SPECIALIST STUDY
Make every major judgement traceable to evidence, assumptions and the reason it changed
An analytic conclusion is easier to trust when a reader can move backward from judgement to reasons, evidence and source chains—and forward to implications, signposts and revision rules. Analytic lineage makes that structure visible without pretending a diagram replaces expertise or that every protected source can be disclosed.
This guide belongs to the Intelligence subject guide, the Gold Standard of Education, and the How X Works learning library. Choose a reading route, work the examples, then use the transfer checks to test whether the method survives a changed context.
Choose a reading route
Build the foundations · chapters 1–2
Apply the core method · chapters 3–5
Test evidence and alternatives · chapters 6–7
Handle limits and disagreement · chapters 8–9
Transfer and govern · chapters 10
The whole problem at a glance
| Node | Question | Record |
|---|---|---|
| Judgement | What is assessed, for whom and by when? | Scoped claim, likelihood and confidence |
| Reason | Why does the evidence support it? | Inference with warrants and alternatives |
| Evidence | Which observation or report bears on the reason? | Source citation, quality and time |
| Revision | What would strengthen, weaken or reverse it? | Signposts, thresholds and version history |
A source can be authentic yet irrelevant, and a sound inference can still rest on incomplete evidence. The map should expose both layers.
CHAPTER 1 OF 10 · Build the foundations
1. Define the judgement node precisely
Core idea. A lineage begins with a bounded key judgement: actor or system, assessed outcome, horizon, conditions, likelihood and confidence. Vague claims produce vague evidence links. The important discipline is to name the object being assessed, the boundary around it and the kind of conclusion the evidence can support. Without those three choices, technical vocabulary creates confidence without creating a checkable claim.
Method. Write the claim independently from its support and assign a stable version identifier. Separate likelihood from confidence and record customer relevance. Write the procedure before looking at the preferred result, retain intermediate evidence and record who can challenge each assumption. A method becomes reusable when another reader can repeat the steps, locate disagreements and understand why a different answer emerged.
Worked example. “Material disruption is likely within six weeks” is more traceable than “the supply situation is worsening.” The example is deliberately small enough to inspect. Its purpose is not to imitate every real case but to expose the decision rule, the evidence it consumes and the point at which judgement enters.
Changed-context transfer. A strategic trend uses milestones and scenarios rather than a short calendar horizon, but still needs a declared unit and scope. Keep the underlying principle while changing the actors, incentives or data. If the conclusion survives only because the original labels were familiar, the learner has memorised a case rather than acquired a transferable method.
Limits. Some judgements contain several causal claims and should be decomposed without fragmenting the decision. A responsible conclusion therefore states what remains unknown, what evidence could reverse it and when the analysis must be repeated. Precision about limits is part of the result, not a disclaimer added after it.
Misconception to avoid. A polished headline is already a well-defined judgement. Readability does not guarantee analytic boundaries. Test the tempting shortcut against a counterexample and ask which stakeholder bears the cost when it fails. The correction should alter the workflow, not merely add cautious wording.
Practical evidence check
- Ask three readers to paraphrase outcome, likelihood, confidence and horizon; resolve every divergent interpretation.
- Return to chapter contents
CHAPTER 2 OF 10 · Build the foundations
2. Separate information, inference and assumption
Core idea. Intelligence information reports observations or claims; analysts interpret relevance and draw inferences; assumptions fill gaps. A lineage should not present all three as interchangeable evidence. The important discipline is to name the object being assessed, the boundary around it and the kind of conclusion the evidence can support. Without those three choices, technical vocabulary creates confidence without creating a checkable claim.
Method. Label nodes by type and require an inference warrant between evidence and judgement. Make key assumptions explicit and attach reversal tests. Write the procedure before looking at the preferred result, retain intermediate evidence and record who can challenge each assumption. A method becomes reusable when another reader can repeat the steps, locate disagreements and understand why a different answer emerged.
Worked example. A supplier notice is information; the assessment that it signals capacity loss is inference; the belief that no substitute supplier exists is an assumption. The example is deliberately small enough to inspect. Its purpose is not to imitate every real case but to expose the decision rule, the evidence it consumes and the point at which judgement enters.
Changed-context transfer. A sensor reading may be direct data yet still requires calibration and contextual interpretation. Keep the underlying principle while changing the actors, incentives or data. If the conclusion survives only because the original labels were familiar, the learner has memorised a case rather than acquired a transferable method.
Limits. Boundaries can be contested, and some sources embed their own analysis. Preserve layers rather than forcing false purity. A responsible conclusion therefore states what remains unknown, what evidence could reverse it and when the analysis must be repeated. Precision about limits is part of the result, not a disclaimer added after it.
Misconception to avoid. A cited sentence supports whatever conclusion follows it. Relevance depends on the reasoning that connects them. Test the tempting shortcut against a counterexample and ask which stakeholder bears the cost when it fails. The correction should alter the workflow, not merely add cautious wording.
Practical evidence check
- Choose one major judgement and identify every hidden bridge statement needed to move from report to conclusion.
- Return to chapter contents
CHAPTER 3 OF 10 · Apply the core method
3. Build source lineage without multiplying dependence
Core idea. Several reports can descend from one original account. Source lineage reveals circular reporting, transformations and incentives that citation counts conceal. The important discipline is to name the object being assessed, the boundary around it and the kind of conclusion the evidence can support. Without those three choices, technical vocabulary creates confidence without creating a checkable claim.
Method. Trace originators, collectors, intermediaries, translations, summaries and databases. Collapse dependent chains for corroboration while preserving distinct access points. Write the procedure before looking at the preferred result, retain intermediate evidence and record who can challenge each assumption. A method becomes reusable when another reader can repeat the steps, locate disagreements and understand why a different answer emerged.
Worked example. Three cables quote separate newspapers that all repeat one supplier press release. The map treats them as one origin chain. The example is deliberately small enough to inspect. Its purpose is not to imitate every real case but to expose the decision rule, the evidence it consumes and the point at which judgement enters.
Changed-context transfer. Two sensors share one calibration fault, so technical independence also requires checking common infrastructure. Keep the underlying principle while changing the actors, incentives or data. If the conclusion survives only because the original labels were familiar, the learner has memorised a case rather than acquired a transferable method.
Limits. Protected sourcing may restrict detail. Use controlled lineage and releasable dependency statements. A responsible conclusion therefore states what remains unknown, what evidence could reverse it and when the analysis must be repeated. Precision about limits is part of the result, not a disclaimer added after it.
Misconception to avoid. Different document identifiers prove independent evidence. Copies and derivative reporting can look numerous. Test the tempting shortcut against a counterexample and ask which stakeholder bears the cost when it fails. The correction should alter the workflow, not merely add cautious wording.
Practical evidence check
- Select three supporting sources and prove where their information first diverges into independent observation.
- Return to chapter contents
CHAPTER 4 OF 10 · Apply the core method
4. Map reasons, objections and alternatives
Core idea. An argument map should include supporting reasons, objections, rebuttals and plausible alternatives rather than only a one-way case for the preferred judgement. The important discipline is to name the object being assessed, the boundary around it and the kind of conclusion the evidence can support. Without those three choices, technical vocabulary creates confidence without creating a checkable claim.
Method. Use a consistent relation vocabulary and require every major alternative to have evidence, assumptions and indicators. Avoid decorative complexity. Write the procedure before looking at the preferred result, retain intermediate evidence and record who can challenge each assumption. A method becomes reusable when another reader can repeat the steps, locate disagreements and understand why a different answer emerged.
Worked example. The buffering alternative attacks the causal step from port delay to customer disruption, not the authenticity of port data. The example is deliberately small enough to inspect. Its purpose is not to imitate every real case but to expose the decision rule, the evidence it consumes and the point at which judgement enters.
Changed-context transfer. A deception hypothesis challenges the source-generation mechanism and therefore branches earlier in the lineage. Keep the underlying principle while changing the actors, incentives or data. If the conclusion survives only because the original labels were familiar, the learner has memorised a case rather than acquired a transferable method.
Limits. Maps can imply equal weight through visual symmetry and can omit unknown alternatives. A responsible conclusion therefore states what remains unknown, what evidence could reverse it and when the analysis must be repeated. Precision about limits is part of the result, not a disclaimer added after it.
Misconception to avoid. Adding an “alternative” box proves alternatives were analysed. The competing pathway must be developed and tested. Test the tempting shortcut against a counterexample and ask which stakeholder bears the cost when it fails. The correction should alter the workflow, not merely add cautious wording.
Practical evidence check
- Ask a challenger to place one objection at the exact inference it contests and name evidence that would decide between branches.
- Return to chapter contents
CHAPTER 5 OF 10 · Apply the core method
5. Attach evidence quality to the claim it supports
Core idea. Source credibility, information plausibility, relevance, timeliness and methodological quality are related but distinct. Quality should be evaluated at the evidence-to-claim edge. The important discipline is to name the object being assessed, the boundary around it and the kind of conclusion the evidence can support. Without those three choices, technical vocabulary creates confidence without creating a checkable claim.
Method. Record why a source can observe the fact, how it reported, possible incentives, corroboration and the precise proposition supported. Write the procedure before looking at the preferred result, retain intermediate evidence and record who can challenge each assumption. A method becomes reusable when another reader can repeat the steps, locate disagreements and understand why a different answer emerged.
Worked example. A supplier is authoritative about its own closure date but motivated to frame responsibility; different claims receive different evaluations. The example is deliberately small enough to inspect. Its purpose is not to imitate every real case but to expose the decision rule, the evidence it consumes and the point at which judgement enters.
Changed-context transfer. A satellite image has strong location evidence but uncertain interpretation of activity. Keep the underlying principle while changing the actors, incentives or data. If the conclusion survives only because the original labels were familiar, the learner has memorised a case rather than acquired a transferable method.
Limits. Numeric scoring can create false precision and hide contextual judgement. A responsible conclusion therefore states what remains unknown, what evidence could reverse it and when the analysis must be repeated. Precision about limits is part of the result, not a disclaimer added after it.
Misconception to avoid. A reliable source makes every statement reliable. Reliability is not a permanent personal property detached from access and claim. Test the tempting shortcut against a counterexample and ask which stakeholder bears the cost when it fails. The correction should alter the workflow, not merely add cautious wording.
Practical evidence check
- Take one source used twice and assess each supported proposition separately; compare the ratings and rationale.
- Return to chapter contents
Worked supply-disruption map
| Element | Content | Challenge |
|---|---|---|
| Judgement | Material disruption is likely within six weeks | Define material and horizon |
| Reason | Two independent bottlenecks are worsening | Test dependence between reports |
| Evidence | Port data and verified supplier notice | Assess timeliness and manipulation risk |
| Alternative | Inventory buffering prevents disruption | Identify stock-level signpost |
| Revision | Two weeks of normal throughput lowers likelihood | Assign monitoring owner |
The fictional example separates report authenticity, evidential relevance, causal warrant and final confidence.
CHAPTER 6 OF 10 · Test evidence and alternatives
6. Preserve uncertainty at every inference
Core idea. Uncertainty can enter through source gaps, ambiguous indicators, causal models and future adaptation. A final confidence label should be traceable to these local weaknesses. The important discipline is to name the object being assessed, the boundary around it and the kind of conclusion the evidence can support. Without those three choices, technical vocabulary creates confidence without creating a checkable claim.
Method. Annotate contested nodes, ranges and assumptions, then explain how they affect the final likelihood and confidence. Do not mechanically average scores. Write the procedure before looking at the preferred result, retain intermediate evidence and record who can challenge each assumption. A method becomes reusable when another reader can repeat the steps, locate disagreements and understand why a different answer emerged.
Worked example. Throughput decline is observed with high confidence, while its duration depends on uncertain repair time, limiting confidence in the six-week forecast. The example is deliberately small enough to inspect. Its purpose is not to imitate every real case but to expose the decision rule, the evidence it consumes and the point at which judgement enters.
Changed-context transfer. A strong causal mechanism can be undermined by weak base-rate evidence; the map keeps both visible. Keep the underlying principle while changing the actors, incentives or data. If the conclusion survives only because the original labels were familiar, the learner has memorised a case rather than acquired a transferable method.
Limits. Dependencies make simple propagation rules misleading. Human review remains essential. A responsible conclusion therefore states what remains unknown, what evidence could reverse it and when the analysis must be repeated. Precision about limits is part of the result, not a disclaimer added after it.
Misconception to avoid. Uncertainty belongs only on the final judgement. Local uncertainty reveals where collection or challenge has value. Test the tempting shortcut against a counterexample and ask which stakeholder bears the cost when it fails. The correction should alter the workflow, not merely add cautious wording.
Practical evidence check
- Remove the weakest evidence node and recompute the verbal judgement; record whether likelihood, confidence or only explanation changes.
- Return to chapter contents
CHAPTER 7 OF 10 · Test evidence and alternatives
7. Version the map and the product together
Core idea. A changed source, inference or assumption should create a new analytic version linked to the product it supports. Otherwise readers cannot tell which reasoning produced which judgement. The important discipline is to name the object being assessed, the boundary around it and the kind of conclusion the evidence can support. Without those three choices, technical vocabulary creates confidence without creating a checkable claim.
Method. Assign identifiers, timestamps and authorship to versions; record added, removed and revised nodes; preserve the earlier map under access controls. Write the procedure before looking at the preferred result, retain intermediate evidence and record who can challenge each assumption. A method becomes reusable when another reader can repeat the steps, locate disagreements and understand why a different answer emerged.
Worked example. A new stock disclosure weakens the disruption case. Version two changes one branch, lowers likelihood and explains why the rest of the map remains. The example is deliberately small enough to inspect. Its purpose is not to imitate every real case but to expose the decision rule, the evidence it consumes and the point at which judgement enters.
Changed-context transfer. A wording edit that does not alter meaning is recorded differently from an analytic revision. Keep the underlying principle while changing the actors, incentives or data. If the conclusion survives only because the original labels were familiar, the learner has memorised a case rather than acquired a transferable method.
Limits. Version systems can burden fast-moving teams. Automation should reduce clerical load without hiding substantive change. A responsible conclusion therefore states what remains unknown, what evidence could reverse it and when the analysis must be repeated. Precision about limits is part of the result, not a disclaimer added after it.
Misconception to avoid. Replacing the current diagram is enough because only the latest view matters. Past reasoning is needed for calibration, learning and accountability. Test the tempting shortcut against a counterexample and ask which stakeholder bears the cost when it fails. The correction should alter the workflow, not merely add cautious wording.
Practical evidence check
- Select two versions and reconstruct every substantive judgement change from node-level differences.
- Return to chapter contents
CHAPTER 8 OF 10 · Handle limits and disagreement
8. Set revision triggers before preferred evidence arrives
Core idea. Signposts and thresholds declared in advance reduce motivated interpretation and turn lineage into a monitoring system. They should cover both supporting and disconfirming observations. The important discipline is to name the object being assessed, the boundary around it and the kind of conclusion the evidence can support. Without those three choices, technical vocabulary creates confidence without creating a checkable claim.
Method. For each major branch, name observable indicators, owner, cadence, threshold and expected effect on likelihood or confidence. Write the procedure before looking at the preferred result, retain intermediate evidence and record who can challenge each assumption. A method becomes reusable when another reader can repeat the steps, locate disagreements and understand why a different answer emerged.
Worked example. Normal throughput for two weeks weakens the bottleneck hypothesis; verified stock depletion strengthens it. Both triggers are explicit. The example is deliberately small enough to inspect. Its purpose is not to imitate every real case but to expose the decision rule, the evidence it consumes and the point at which judgement enters.
Changed-context transfer. A qualitative political judgement uses process indicators rather than false numeric precision. Keep the underlying principle while changing the actors, incentives or data. If the conclusion survives only because the original labels were familiar, the learner has memorised a case rather than acquired a transferable method.
Limits. Adversaries can manipulate indicators, and unexpected mechanisms may bypass the list. A responsible conclusion therefore states what remains unknown, what evidence could reverse it and when the analysis must be repeated. Precision about limits is part of the result, not a disclaimer added after it.
Misconception to avoid. Signposts predict the future. They specify evidence that would update a conditional judgement. Test the tempting shortcut against a counterexample and ask which stakeholder bears the cost when it fails. The correction should alter the workflow, not merely add cautious wording.
Practical evidence check
- Run a tabletop with one confirming, one disconfirming and one ambiguous signpost and record the update rule.
- Return to chapter contents
Revision record
| Change | Required explanation | Weak substitute |
|---|---|---|
| Likelihood moves | New evidence or changed model and its weight | “Analyst updated view” |
| Confidence moves | Source quality, gaps or assumption stability changed | Same evidence with a new adjective |
| Scope changes | New actor, horizon or definition is explicit | Quietly comparing unlike claims |
| Judgement retires | Supersession and continuing validity are stated | Deleting the old product |
A changed judgement is not automatically an error; the earlier version should be evaluated against information and scope available at its time.
CHAPTER 9 OF 10 · Handle limits and disagreement
9. Protect lineage from hierarchy and disclosure risk
Core idea. A transparent reasoning structure can expose sensitive sources, dissenters or operational methods. It can also threaten owners whose assumptions become visible. Governance must protect both evidence and challenge. The important discipline is to name the object being assessed, the boundary around it and the kind of conclusion the evidence can support. Without those three choices, technical vocabulary creates confidence without creating a checkable claim.
Method. Use tiered access, source substitution labels, protected dissent routes and independent review. Publish enough structure for decision use without revealing protected detail. Write the procedure before looking at the preferred result, retain intermediate evidence and record who can challenge each assumption. A method becomes reusable when another reader can repeat the steps, locate disagreements and understand why a different answer emerged.
Worked example. A customer sees that two independent source chains support the judgement but only authorised reviewers can inspect identities. The example is deliberately small enough to inspect. Its purpose is not to imitate every real case but to expose the decision rule, the evidence it consumes and the point at which judgement enters.
Changed-context transfer. A public report releases the argument logic and confidence basis while withholding collection methods. Keep the underlying principle while changing the actors, incentives or data. If the conclusion survives only because the original labels were familiar, the learner has memorised a case rather than acquired a transferable method.
Limits. Over-redaction can make lineage ceremonial; over-disclosure can create safety and legal harm. A responsible conclusion therefore states what remains unknown, what evidence could reverse it and when the analysis must be repeated. Precision about limits is part of the result, not a disclaimer added after it.
Misconception to avoid. Transparency requires every reader to see every node. Proportionate transparency preserves scrutiny through appropriate roles. Test the tempting shortcut against a counterexample and ask which stakeholder bears the cost when it fails. The correction should alter the workflow, not merely add cautious wording.
Practical evidence check
- Give the releasable map to a reviewer and test whether they can assess reasoning without inferring protected identities.
- Return to chapter contents
CHAPTER 10 OF 10 · Transfer and govern
10. Use lineage for calibration and institutional learning
Core idea. Across many assessments, lineage records reveal repeated weak assumptions, source dependencies, late signposts and review bottlenecks. Evaluation should improve process rather than punish honest uncertainty. The important discipline is to name the object being assessed, the boundary around it and the kind of conclusion the evidence can support. Without those three choices, technical vocabulary creates confidence without creating a checkable claim.
Method. Compare forecasts with outcomes, code reasons for revision, examine confidence calibration and identify which collection or challenge changed decisions. Write the procedure before looking at the preferred result, retain intermediate evidence and record who can challenge each assumption. A method becomes reusable when another reader can repeat the steps, locate disagreements and understand why a different answer emerged.
Worked example. A review finds that logistics assumptions repeatedly fail before source reporting does, leading to new collection and modelling standards. The example is deliberately small enough to inspect. Its purpose is not to imitate every real case but to expose the decision rule, the evidence it consumes and the point at which judgement enters.
Changed-context transfer. In engineering safety, the same lineage supports hazard arguments and evidence updates under different legal duties. Keep the underlying principle while changing the actors, incentives or data. If the conclusion survives only because the original labels were familiar, the learner has memorised a case rather than acquired a transferable method.
Limits. Outcomes can be ambiguous, and hindsight bias can make reasonable earlier judgements look careless. A responsible conclusion therefore states what remains unknown, what evidence could reverse it and when the analysis must be repeated. Precision about limits is part of the result, not a disclaimer added after it.
Misconception to avoid. A wrong outcome proves poor analysis. Quality depends on process, calibration and information available at the time. Test the tempting shortcut against a counterexample and ask which stakeholder bears the cost when it fails. The correction should alter the workflow, not merely add cautious wording.
Practical evidence check
- Select a portfolio of past judgements and test whether higher confidence corresponded to stronger calibration and fewer hidden dependencies.
- Return to chapter contents
Build a analytic-lineage evidence worksheet
State the decision, unit of analysis, time horizon and affected people before collecting evidence. Separate observations from interpretations and interpretations from recommendations. Record data provenance, missingness and incentives that could distort reporting.
For every important claim, add a rival explanation and one observation that would favour it. Give uncertainty a visible place instead of hiding it inside a single score. Assign an owner and review date to assumptions that may change.
Run the worksheet with someone who did not design the analysis. Ask them to reconstruct the conclusion and identify where judgement entered. Revise any step that depends on private intuition or undefined terminology.
Finish with an action table: act now, monitor, investigate or stop. Link each action to evidence and a reversal condition so the method supports learning rather than defending the first answer.
Design a changed-context transfer check
Rewrite the case with a different scale, institution and distribution of power while preserving the underlying structure. Predict which parts of the method should remain stable and which require new evidence.
Remove one convenient data source and add one conflicting stakeholder account. Re-run the reasoning. If the answer changes, explain whether the change reflects better evidence, a different value choice or an unstable method.
Switch the time horizon. A decision that looks efficient this week may weaken resilience over years, while a long-run benefit may impose an unacceptable immediate burden. Report both rather than averaging them away.
Ask a reviewer to produce a counterexample. A transfer check passes when the method reveals why the new case differs, not when it mechanically produces the same conclusion.
Audit the reasoning and the record
Preserve source versions, transformations, thresholds, exclusions and decision notes. A final chart without its lineage cannot be independently checked and may conceal silent changes.
Trace one favourable and one unfavourable case from raw evidence to recommendation. Confirm that identical rules were applied and that exceptions were authorised, explained and reviewable.
Interview affected users about how the process works in practice. Formal policy may differ from operational reality, especially when time pressure, inaccessible interfaces or asymmetric power limit challenge.
Publish a proportionate account of method, limits, ownership and remedy. Protect sensitive data while ensuring that confidentiality is not used to block legitimate scrutiny.
Turn findings into accountable action
Assign every recommendation to a named role with authority, resources and a deadline. Distinguish mitigation from remedy: reducing future risk does not repair harm already caused.
Choose leading indicators that reveal drift early and outcome indicators that show whether people actually benefited. Disaggregate results where averages can hide concentrated burden.
Create an escalation rule for new evidence, missed thresholds and recurring complaints. Independent review should be available when the original decision maker has an incentive to defend the system.
Close the loop by publishing what changed, what did not and why. Re-open the analysis after material changes instead of treating one successful review as permanent certification.
Questions readers ask
Is an argument map the analysis itself?
- No. It is a representation that should expose evidence, warrants, assumptions, alternatives and uncertainty for expert review.
Why track source lineage?
- To detect circular reporting, transformations, shared dependencies and the exact claim a source supports.
Should every source be visible to every reader?
- No. Tiered access can preserve meaningful scrutiny while protecting sensitive identities and methods.
What changes confidence?
- Source quality, knowledge gaps, assumption stability, methodological strength and agreement among relevant evidence.
Does revision mean the prior analysis failed?
- Not necessarily. Evidence-led revision is expected; evaluate the earlier judgement against its information and scope.
How detailed should a map be?
- Detailed enough to reconstruct major inferences and challenges without becoming an unreadable inventory.
Practise, verify and continue
Build a compact map for a fictional supply disruption. Include one key judgement, two supporting reasons, three source chains, one key assumption, one alternative, one dissent, confidence, and three revision signposts. Then issue a second version after new evidence.
Move the method to public health, engineering or school planning. Preserve typed claims, source chains, warrants, alternatives, uncertainty, protected challenge and revision history while adapting access and professional duties.
Related eduKate reading
Coordinated assessments, disagreement and dissent
Source triangulation and circular reporting
Competing hypotheses and evidence matrices
Key assumptions checks and reversal conditions
Narrative analysis, claim networks and revision
Authoritative sources and scope
Office of the Director of National Intelligence, ICD 203 Analytic Standards
ODNI, Intelligence Community source-citation requirements
CIA Center for the Study of Intelligence, A Tradecraft Primer
CIA Center for the Study of Intelligence, Psychology of Intelligence Analysis
Office of the Director of National Intelligence, Analytic Objectivity
This article is an educational use of public analytic standards and does not disclose classified sources, methods or current assessments. The examples are fictional. Institutions must follow applicable law, classification, privacy, records and source-protection rules. A diagram cannot establish truth, remove uncertainty or replace expertise. Lineage should improve relevance, challenge, calibration and revision while avoiding coercive scoring of individual analysts. Access controls should protect people and capabilities without being used to conceal poor reasoning, pressure or error.