THE GOLD STANDARD OF INTELLIGENCE · SPECIALIST STUDY
Protect participants and inference when an active RDS protocol release does not reach every interviewer, device or site
A protocol-control library can publish a correct release while disconnected devices, printed aids or cached translations continue to guide fieldwork. The incident response must locate exposure without blaming every case, preserve participant rights, quarantine only what needs review, and bridge analytical differences without pretending versions are equivalent. The standard is a traceable path from release evidence to device and case lineage, remedy, sensitivity analysis, corrective action and proof that the same distribution failure will be detected earlier next time.
This guide belongs to the Intelligence subject root, the Gold Standard of Education, and the How X Works hub. It begins with a concrete decision, then tests whether the method still holds when the context changes.
Choose a reading route
Detect and bound stale use · chapters 1, 2, 3
Recall devices and protect cases · chapters 4, 5
Remedy participants and field teams · chapters 6, 7, 8
Bridge estimates across versions · chapters 9, 10, 11
Prevent recurrence and reauthorise · chapters 12, 13
The whole problem at a glance
| Incident question | Required evidence | Unsafe shortcut |
|---|---|---|
| Who received the release? | Role-device-site lineage | Email sent |
| Which cases were exposed? | Version at each action | Interviewer memory |
| What changes analytically? | Bridging and sensitivity | Pool everything |
| Who needs remedy? | Harm and contact rule | Exclude the case |
Exposure is not identical to harm or invalidity.
CHAPTER 1 OF 13 · Detect and bound stale use
1. Open the stale-protocol incident
Record the first credible sign that retired instructions were used: a screenshot, device log, supervisor observation, translated aid or case inconsistency. Preserve it, notify the authorised incident owner and separate immediate participant protection from later performance review.
Evidence record: Report, source, time, version claim, affected site and role, current field activity, potential participant consequence, containment and conflicts. Preserve timestamps, versions, denominators, missing cases, failed checks and rejected alternatives so another reviewer can reconstruct the decision rather than inherit a conclusion.
Decision rule: Pause only actions plausibly affected while maintaining safe participant contact and essential operations. Name the authorised owner, trigger, stop condition, remedy and verification window. A control is incomplete until a changed-context case tests the mechanism it is meant to protect.
Evidence discipline: record the observation, version or time boundary that supports this chapter’s conclusion. Preserve the rejected alternative as well as the selected action. That makes a later review able to distinguish a changed world from an undocumented change of mind.
Changed-context transfer: A printed aid is stale but devices are current; bound the incident to the aid and its users. The method transfers only if the decision question and measurement meaning survive the change. If either changes, rebuild the comparison rather than copying the earlier threshold or weight.
Chapter 1 check · what should survive an audit
- State the decision unit for “Open the stale-protocol incident”.
- Bind the claim to versions, denominators, exclusions and time boundaries.
- Connect evidence to an owner, bounded action and later verification.
CHAPTER 2 OF 13 · Detect and bound stale use
2. Define critical differences between versions
Compare requirements rather than whole documents. Mark changes to consent, eligibility, recruitment, coupon limits, safety, incentives, privacy, contact, refusal handling and data fields. A formatting change should not trigger the same response as an altered participant right.
Evidence record: Requirement identifiers, old and new text, rationale, effective time, languages, training needs, analytical variables and consequence rating. Preserve timestamps, versions, denominators, missing cases, failed checks and rejected alternatives so another reviewer can reconstruct the decision rather than inherit a conclusion.
Decision rule: Escalate differences that change participant rights, inclusion, network links or estimator inputs. Name the authorised owner, trigger, stop condition, remedy and verification window. A control is incomplete until a changed-context case tests the mechanism it is meant to protect.
Evidence discipline: record the observation, version or time boundary that supports this chapter’s conclusion. Preserve the rejected alternative as well as the selected action. That makes a later review able to distinguish a changed world from an undocumented change of mind.
Changed-context transfer: A translation corrects wording without changing meaning; validate equivalence before classifying impact. The method transfers only if the decision question and measurement meaning survive the change. If either changes, rebuild the comparison rather than copying the earlier threshold or weight.
Chapter 2 check · what should survive an audit
- State the decision unit for “Define critical differences between versions”.
- Bind the claim to versions, denominators, exclusions and time boundaries.
- Connect evidence to an owner, bounded action and later verification.
CHAPTER 3 OF 13 · Detect and bound stale use
3. Build a device-site-role exposure map
Join release distribution with device identifiers, sync receipts, cache versions, interviewer assignments, supervisor rosters, site connectivity and printed-material registers. Unknowns stay visible. Do not infer successful adoption from a download event.
Evidence record: Devices, users, sites, last sync, installed package, cached assets, assignments, acknowledgements, observed use and missing links. Preserve timestamps, versions, denominators, missing cases, failed checks and rejected alternatives so another reviewer can reconstruct the decision rather than inherit a conclusion.
Decision rule: Classify confirmed, probable, possible and cleared exposure using declared evidence rules. Name the authorised owner, trigger, stop condition, remedy and verification window. A control is incomplete until a changed-context case tests the mechanism it is meant to protect.
Evidence discipline: record the observation, version or time boundary that supports this chapter’s conclusion. Preserve the rejected alternative as well as the selected action. That makes a later review able to distinguish a changed world from an undocumented change of mind.
Changed-context transfer: A shared device changed users mid-wave; split exposure by assignment and time rather than attaching every case to one interviewer. The method transfers only if the decision question and measurement meaning survive the change. If either changes, rebuild the comparison rather than copying the earlier threshold or weight.
Chapter 3 check · what should survive an audit
- State the decision unit for “Build a device-site-role exposure map”.
- Bind the claim to versions, denominators, exclusions and time boundaries.
- Connect evidence to an owner, bounded action and later verification.
CHAPTER 4 OF 13 · Recall devices and protect cases
4. Recall offline devices safely
Issue a versioned recall that works through expected connectivity gaps: disable affected forms where possible, provide an authorised contact path, verify returned or updated devices and preserve logs. Avoid remote wiping that destroys case or incident evidence.
Evidence record: Recall notice, device list, delivery attempts, acknowledgements, update hashes, quarantine state, return custody, preserved logs, exceptions and deadlines. Preserve timestamps, versions, denominators, missing cases, failed checks and rejected alternatives so another reviewer can reconstruct the decision rather than inherit a conclusion.
Decision rule: Do not restore collection until the device package, cached aids and clock are verified together. Name the authorised owner, trigger, stop condition, remedy and verification window. A control is incomplete until a changed-context case tests the mechanism it is meant to protect.
Evidence discipline: record the observation, version or time boundary that supports this chapter’s conclusion. Preserve the rejected alternative as well as the selected action. That makes a later review able to distinguish a changed world from an undocumented change of mind.
Changed-context transfer: A device cannot connect but contains completed interviews; transfer under controlled custody before updating. The method transfers only if the decision question and measurement meaning survive the change. If either changes, rebuild the comparison rather than copying the earlier threshold or weight.
Chapter 4 check · what should survive an audit
- State the decision unit for “Recall offline devices safely”.
- Bind the claim to versions, denominators, exclusions and time boundaries.
- Connect evidence to an owner, bounded action and later verification.
CHAPTER 5 OF 13 · Recall devices and protect cases
5. Quarantine mixed-version cases proportionately
A case may contain recruitment under one version and follow-up under another. Preserve the full chronology and quarantine affected decisions or variables, not automatically the person. Create review routes for consent, eligibility, network links, payments and measurements.
Evidence record: Case identifier, actions, actor, device, version at each action, participant communications, data fields, links, payments, review and release decision. Preserve timestamps, versions, denominators, missing cases, failed checks and rejected alternatives so another reviewer can reconstruct the decision rather than inherit a conclusion.
Decision rule: Keep high-consequence uncertainty blocked from analysis until an authorised reviewer resolves it. Name the authorised owner, trigger, stop condition, remedy and verification window. A control is incomplete until a changed-context case tests the mechanism it is meant to protect.
Evidence discipline: record the observation, version or time boundary that supports this chapter’s conclusion. Preserve the rejected alternative as well as the selected action. That makes a later review able to distinguish a changed world from an undocumented change of mind.
Changed-context transfer: Consent used the current version but a follow-up prompt was stale; preserve valid parts while reviewing the changed field. The method transfers only if the decision question and measurement meaning survive the change. If either changes, rebuild the comparison rather than copying the earlier threshold or weight.
Chapter 5 check · what should survive an audit
- State the decision unit for “Quarantine mixed-version cases proportionately”.
- Bind the claim to versions, denominators, exclusions and time boundaries.
- Connect evidence to an owner, bounded action and later verification.
Worked case triage
| Case | Device version | Critical action | Initial route |
|---|---|---|---|
| C-41 | v4 | Consent v4 | Release |
| C-42 | v3 | Coupon issued | Quarantine |
| C-43 | Unknown | Interview paused | Investigate |
| C-44 | v3→v4 | Follow-up only | Bridge |
Actual decisions require policy, ethics and participant review.
CHAPTER 6 OF 13 · Remedy participants and field teams
6. Handle late acknowledgement as evidence, not adoption
An interviewer may acknowledge after the effective date, after working cases or without understanding the change. Record timing and test comprehension or observed practice. Late acknowledgement can narrow exposure but cannot retroactively prove compliant action.
Evidence record: Release and acknowledgement times, role, language, training completion, knowledge check, supervised observation, cases worked and exceptions. Preserve timestamps, versions, denominators, missing cases, failed checks and rejected alternatives so another reviewer can reconstruct the decision rather than inherit a conclusion.
Decision rule: Require targeted retraining and observation when lateness overlaps a critical change. Name the authorised owner, trigger, stop condition, remedy and verification window. A control is incomplete until a changed-context case tests the mechanism it is meant to protect.
Evidence discipline: record the observation, version or time boundary that supports this chapter’s conclusion. Preserve the rejected alternative as well as the selected action. That makes a later review able to distinguish a changed world from an undocumented change of mind.
Changed-context transfer: An interviewer signed late but had already received live coaching; preserve both facts and assess actual case actions. The method transfers only if the decision question and measurement meaning survive the change. If either changes, rebuild the comparison rather than copying the earlier threshold or weight.
Chapter 6 check · what should survive an audit
- State the decision unit for “Handle late acknowledgement as evidence, not adoption”.
- Bind the claim to versions, denominators, exclusions and time boundaries.
- Connect evidence to an owner, bounded action and later verification.
CHAPTER 7 OF 13 · Remedy participants and field teams
7. Protect participant rights during investigation
Keep contacts minimal, respectful and authorised. Do not disclose internal blame or make participants repeat sensitive information unnecessarily. Where consent, incentives, privacy or safety may be affected, use an ethics or governance-approved remedy script and route.
Evidence record: Potential harm, participant preference, contact authority, script, attempts, accessibility, response, remedy, refusal, adverse event and closure. Preserve timestamps, versions, denominators, missing cases, failed checks and rejected alternatives so another reviewer can reconstruct the decision rather than inherit a conclusion.
Decision rule: Prioritise safety and rights over analytical convenience, and honour a choice not to recontact where policy permits. Name the authorised owner, trigger, stop condition, remedy and verification window. A control is incomplete until a changed-context case tests the mechanism it is meant to protect.
Evidence discipline: record the observation, version or time boundary that supports this chapter’s conclusion. Preserve the rejected alternative as well as the selected action. That makes a later review able to distinguish a changed world from an undocumented change of mind.
Changed-context transfer: A participant prefers no further contact; document the analytical consequence without overriding that choice. The method transfers only if the decision question and measurement meaning survive the change. If either changes, rebuild the comparison rather than copying the earlier threshold or weight.
Chapter 7 check · what should survive an audit
- State the decision unit for “Protect participant rights during investigation”.
- Bind the claim to versions, denominators, exclusions and time boundaries.
- Connect evidence to an owner, bounded action and later verification.
CHAPTER 8 OF 13 · Remedy participants and field teams
8. Design participant remedy by consequence
Remedy may include corrected information, renewed consent, payment correction, data withdrawal options, referral or an apology. Match it to the actual consequence and avoid coercive offers. Preserve whether the remedy changes data use or network relationships.
Evidence record: Eligibility rule, harm assessment, authorised options, communication, acceptance or refusal, timing, payment, data action and independent review. Preserve timestamps, versions, denominators, missing cases, failed checks and rejected alternatives so another reviewer can reconstruct the decision rather than inherit a conclusion.
Decision rule: Offer comparable remedies to comparable cases and record justified differences. Name the authorised owner, trigger, stop condition, remedy and verification window. A control is incomplete until a changed-context case tests the mechanism it is meant to protect.
Evidence discipline: record the observation, version or time boundary that supports this chapter’s conclusion. Preserve the rejected alternative as well as the selected action. That makes a later review able to distinguish a changed world from an undocumented change of mind.
Changed-context transfer: A stale incentive statement understated timing but payment was correct; correct information without inventing a new entitlement. The method transfers only if the decision question and measurement meaning survive the change. If either changes, rebuild the comparison rather than copying the earlier threshold or weight.
Chapter 8 check · what should survive an audit
- State the decision unit for “Design participant remedy by consequence”.
- Bind the claim to versions, denominators, exclusions and time boundaries.
- Connect evidence to an owner, bounded action and later verification.
CHAPTER 9 OF 13 · Bridge estimates across versions
9. Reconcile recruitment trees and coupons
Version differences can alter who may recruit, coupon validity, limits or privacy handling. Reconstruct edges from issued and redeemed coupons, device state and case actions. Preserve disputed links and avoid deleting a participant to make the tree tidy.
Evidence record: Coupon identifiers, issuers, redeemers, version, timestamps, limits, duplicates, cancelled instruments, disputed edges and resolution. Preserve timestamps, versions, denominators, missing cases, failed checks and rejected alternatives so another reviewer can reconstruct the decision rather than inherit a conclusion.
Decision rule: Quarantine analytically consequential links while participant services continue where safe. Name the authorised owner, trigger, stop condition, remedy and verification window. A control is incomplete until a changed-context case tests the mechanism it is meant to protect.
Evidence discipline: record the observation, version or time boundary that supports this chapter’s conclusion. Preserve the rejected alternative as well as the selected action. That makes a later review able to distinguish a changed world from an undocumented change of mind.
Changed-context transfer: One stale coupon was redeemed after update; retain the event and model alternative valid-parent assignments. The method transfers only if the decision question and measurement meaning survive the change. If either changes, rebuild the comparison rather than copying the earlier threshold or weight.
Chapter 9 check · what should survive an audit
- State the decision unit for “Reconcile recruitment trees and coupons”.
- Bind the claim to versions, denominators, exclusions and time boundaries.
- Connect evidence to an owner, bounded action and later verification.
Reauthorisation gate
| Gate | Pass evidence | Remain open when |
|---|---|---|
| Recall | All devices accounted | Unknown cache |
| Cases | Exposure classified | Missing lineage |
| Remedy | Eligible people addressed | Open harm |
| Analysis | Version effects bounded | Unstable bridge |
Acknowledgement alone does not prove adoption.
CHAPTER 10 OF 13 · Bridge estimates across versions
10. Build an analytical bridge across versions
Define which variables and processes are equivalent, transformable, incomparable or missing. Use overlap cases, calibration samples and sensitivity analysis. A bridge documents assumptions; it does not make different recruitment or measurement mechanisms identical.
Evidence record: Version cohorts, sample composition, overlap, item mappings, mode and language, missingness, model, diagnostics, weights and uncertainty. Preserve timestamps, versions, denominators, missing cases, failed checks and rejected alternatives so another reviewer can reconstruct the decision rather than inherit a conclusion.
Decision rule: Publish version-stratified results when a pooled bridge is unsupported or materially unstable. Name the authorised owner, trigger, stop condition, remedy and verification window. A control is incomplete until a changed-context case tests the mechanism it is meant to protect.
Evidence discipline: record the observation, version or time boundary that supports this chapter’s conclusion. Preserve the rejected alternative as well as the selected action. That makes a later review able to distinguish a changed world from an undocumented change of mind.
Changed-context transfer: A response option changed meaning; report separate estimates instead of forcing a numeric recode. The method transfers only if the decision question and measurement meaning survive the change. If either changes, rebuild the comparison rather than copying the earlier threshold or weight.
Chapter 10 check · what should survive an audit
- State the decision unit for “Build an analytical bridge across versions”.
- Bind the claim to versions, denominators, exclusions and time boundaries.
- Connect evidence to an owner, bounded action and later verification.
CHAPTER 11 OF 13 · Bridge estimates across versions
11. Test estimator sensitivity to incident decisions
Recalculate degree, network size, weights and outcomes under plausible case-release, edge and variable decisions. Preserve thresholds and do not tune the preferred scenario after seeing favourable results.
Evidence record: Primary rules, alternative scenarios, affected counts, weight changes, estimates, intervals, diagnostics, decision thresholds and reviewers. Preserve timestamps, versions, denominators, missing cases, failed checks and rejected alternatives so another reviewer can reconstruct the decision rather than inherit a conclusion.
Decision rule: Escalate when conclusions depend on one weak assumption about mixed-version cases. Name the authorised owner, trigger, stop condition, remedy and verification window. A control is incomplete until a changed-context case tests the mechanism it is meant to protect.
Evidence discipline: record the observation, version or time boundary that supports this chapter’s conclusion. Preserve the rejected alternative as well as the selected action. That makes a later review able to distinguish a changed world from an undocumented change of mind.
Changed-context transfer: A small number of high-degree cases dominate the change; show influence rather than only sample percentage. The method transfers only if the decision question and measurement meaning survive the change. If either changes, rebuild the comparison rather than copying the earlier threshold or weight.
Chapter 11 check · what should survive an audit
- State the decision unit for “Test estimator sensitivity to incident decisions”.
- Bind the claim to versions, denominators, exclusions and time boundaries.
- Connect evidence to an owner, bounded action and later verification.
CHAPTER 12 OF 13 · Prevent recurrence and reauthorise
12. Repair release controls and detection
Address mechanism: offline sync, mutable caches, untracked printouts, ambiguous effective dates, translation lag or weak supervisor verification. Add technical and human controls, then simulate a disconnected device and late site.
Evidence record: Root-cause hypotheses, counterevidence, corrective actions, owners, version beacons, expiry controls, test scenarios, detection time and residual gaps. Preserve timestamps, versions, denominators, missing cases, failed checks and rejected alternatives so another reviewer can reconstruct the decision rather than inherit a conclusion.
Decision rule: Close an action only after it detects or prevents the failure under changed connectivity conditions. Name the authorised owner, trigger, stop condition, remedy and verification window. A control is incomplete until a changed-context case tests the mechanism it is meant to protect.
Evidence discipline: record the observation, version or time boundary that supports this chapter’s conclusion. Preserve the rejected alternative as well as the selected action. That makes a later review able to distinguish a changed world from an undocumented change of mind.
Changed-context transfer: A push notification works online; test a device that remains offline across the effective date. The method transfers only if the decision question and measurement meaning survive the change. If either changes, rebuild the comparison rather than copying the earlier threshold or weight.
Chapter 12 check · what should survive an audit
- State the decision unit for “Repair release controls and detection”.
- Bind the claim to versions, denominators, exclusions and time boundaries.
- Connect evidence to an owner, bounded action and later verification.
CHAPTER 13 OF 13 · Prevent recurrence and reauthorise
13. Reauthorise fieldwork and preserve learning
The restart package should show recalled devices, case dispositions, participant remedies, analytical limits, corrected materials, training evidence, tested controls and residual risk acceptance. Set monitoring and reopen triggers for unknown version reports or estimator drift.
Evidence record: Final exposure register, recall evidence, case decisions, remedies, bridge, sensitivity results, corrective-action tests, authorisation, monitoring and triggers. Preserve timestamps, versions, denominators, missing cases, failed checks and rejected alternatives so another reviewer can reconstruct the decision rather than inherit a conclusion.
Decision rule: Restart only when critical unknowns have owners and unsafe stale paths are blocked or bounded. Name the authorised owner, trigger, stop condition, remedy and verification window. A control is incomplete until a changed-context case tests the mechanism it is meant to protect.
Evidence discipline: record the observation, version or time boundary that supports this chapter’s conclusion. Preserve the rejected alternative as well as the selected action. That makes a later review able to distinguish a changed world from an undocumented change of mind.
Changed-context transfer: Transfer the method to a multilingual household survey while replacing RDS-specific network and coupon controls. The method transfers only if the decision question and measurement meaning survive the change. If either changes, rebuild the comparison rather than copying the earlier threshold or weight.
Chapter 13 check · what should survive an audit
- State the decision unit for “Reauthorise fieldwork and preserve learning”.
- Bind the claim to versions, denominators, exclusions and time boundaries.
- Connect evidence to an owner, bounded action and later verification.
Contents · Previous chapter · Return to the Intelligence guide
Verification lab: work the decision twice
Trace four fictional cases through two protocol versions, one offline device and one late acknowledgement; classify exposure without equating it with invalidity.
Design a participant-remedy matrix for consent, privacy, incentive and measurement consequences.
Bridge one changed survey item across versions, run a sensitivity analysis and decide whether pooled reporting remains supportable.
Limits and misconceptions
A download receipt is not adoption
Verify installed assets and observed practice.
Exposure is not automatically harm
Classify consequence with participant-centred review.
Case exclusion is not participant remedy
Rights and services require separate action.
A bridge cannot erase mechanism change
Report incompatible cohorts when assumptions fail.
Questions readers ask
What triggers a stale-protocol incident?
- Credible evidence that retired instructions or assets shaped active fieldwork.
Must every exposed case be discarded?
- No. Preserve chronology and review the affected actions, rights and variables proportionately.
How should offline devices be recalled?
- With tracked notices, safe evidence preservation, package verification and an authorised exception path.
Does acknowledgement prove compliance?
- No. Timing, comprehension and observed practice matter.
What is an analytical bridge?
- A documented method for comparing or combining version cohorts with explicit assumptions and diagnostics.
When can fieldwork restart?
- When critical stale paths are controlled, cases and remedies are governed, and residual uncertainty is accepted.
Practise, verify and continue
Respond to an RDS protocol release that missed offline devices while preserving participant rights, case lineage, estimator honesty and recurrence tests.
Move the method to a multilingual longitudinal survey, preserving version exposure, participant remedy and analytical bridging while replacing coupon and network controls.
Practical evidence check
Before implementation, write a one-page evidence map with the decision, unit of analysis, version or time boundary, inputs, outcomes, uncertainty, authority and stop rule. Attach one worked case that passes and one that fails. A reviewer should be able to reproduce the reasoning without relying on the memory of the person who built the system.
Run one changed-context rehearsal before approval. Alter a mechanism rather than a cosmetic detail, then identify which calculation, safeguard or governance gate must change. Record any conclusion that does not transfer. This short rehearsal is often the fastest way to expose a hidden assumption before it becomes an operational dependency.
Related eduKate reading
RDS protocol-control library versioning
RDS protocol-deviation trend review
RDS supervisor intervention and deviation triage
RDS survey-language version changes
Authoritative sources and scope
US Census Bureau Statistical Quality Standard B1
CDC NHANES Interviewer Procedures Manual
CDC NHANES 2025–2027 manuals
CDC Global Tobacco Surveillance System survey documents
The official survey sources support tested collection procedures, training, monitoring, corrective action and documentation. They do not prescribe this RDS incident workflow. Cases are fictional; ethics, consent, privacy, incentives, local law and estimator decisions require authorised review.