THE GOLD STANDARD OF INTELLIGENCE · SPECIALIST STUDY
Treat respondent-driven sampling as a process unfolding through calendar time: distinguish recruitment-wave depth from dates, diagnose shocks and seasonality, preserve phased fieldwork decisions and test whether estimates survive time-sensitive alternatives
Respondent-driven sampling can run across holidays, enforcement changes, service closures, weather, migration, incentive adjustments and phased site openings. Recruitment-wave depth does not remove these calendar-time changes. The gold standard records issue, contact and interview timing; plots composition and recruitment productivity by both chain and date; annotates fieldwork interventions before looking at estimates; separates seasonal opportunity from network mixing; uses time-aware diagnostics and sensitivity models; and limits population claims when the target population or participation mechanism changed during collection.
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 6–8
Handle limits and disagreement · chapters 9–11
Transfer and govern · chapters 12–13
The whole problem at a glance
| Clock | What it indexes | Risk if omitted |
|---|---|---|
| Calendar date | External conditions | Seasonality or shock |
| Recruitment wave | Distance from seed | Seed convergence |
| Coupon age | Delay since issue | Censoring |
| Fieldwork phase | Protocol or site version | Intervention confounding |
Sample order and calendar time are related but not interchangeable.
CHAPTER 1 OF 13 · Build the foundations
1. Define the population and time window together
Core idea. A hidden population is not automatically stable while fieldwork runs. Eligibility, presence, network ties, willingness and access to study sites can change. The estimand should therefore name both the population and the calendar period.
Method. State geography, eligibility, residence or activity requirements, fieldwork dates and whether the goal is a period average, a particular date or a stable structural quantity; list plausible time-varying mechanisms before collection.
Worked example. A six-month study spans a seasonal work migration. The pooled proportion describes participants reachable across that interval, not necessarily the population on the final day.
Changed-context transfer. Changed-context transfer: apply the same reasoning structure to a school or public-service decision. Preserve the formal relationship, then replace the population, authority, privacy duties, evidence and remedy; never copy the domain conclusion.
Limits. Short windows reduce drift but may miss subgroups with periodic availability; long windows improve reach but increase nonstationarity.
Evidence and review. For this chapter, keep a dated evidence record: the input, rule or assumption, result, accountable owner and a condition that would change the conclusion. Work an ordinary case and a boundary case, preserve rejected alternatives, and require an authorised independent reader to reproduce where judgement entered. Chapter 1 is complete only when the claimed result survives that review.
Misconception to avoid. A sample collected in one campaign represents a timeless population. Every estimate inherits a period and access process.
Practical evidence check
- Rewrite one headline estimate as a period-specific claim and identify what evidence would support extrapolation.
- Return to chapter contents
CHAPTER 2 OF 13 · Build the foundations
2. Keep calendar time, wave and sample order distinct
Core idea. Wave counts recruitment links from seeds, while calendar time records when events occurred and sample order records sequence. A late shallow branch and an early deep branch can have the same order but very different exposure to external conditions.
Method. Store coupon issue, attempted contact, interview and expiry times; retain recruiter–recruit links, seed and wave; plot each clock separately and examine their dependence.
Worked example. Two recruits are both wave four. One arrives in week two; the other arrives after a six-week pause following a holiday and protocol change.
Changed-context transfer. Changed-context transfer: apply the same reasoning structure to a school or public-service decision. Preserve the formal relationship, then replace the population, authority, privacy duties, evidence and remedy; never copy the domain conclusion.
Limits. Timestamps can be rounded, missing or privacy-sensitive, and wave can be ambiguous when records are incomplete.
Evidence and review. For this chapter, keep a dated evidence record: the input, rule or assumption, result, accountable owner and a condition that would change the conclusion. Work an ordinary case and a boundary case, preserve rejected alternatives, and require an authorised independent reader to reproduce where judgement entered. Chapter 2 is complete only when the claimed result survives that review.
Misconception to avoid. Wave depth controls every time effect. It addresses distance from seeds, not changing calendar conditions.
Practical evidence check
- Colour the recruitment tree by date and the calendar plot by wave; explain patterns visible only in one view.
- Return to chapter contents
CHAPTER 3 OF 13 · Apply the core method
3. Build a fieldwork event ledger before modelling
Core idea. Site hours, staffing, incentives, seed additions, coupon rules, outages, publicity and local events can change recruitment. Without a dated ledger, later analysts may mistake operational intervention for population behaviour.
Method. Record start and end times, affected sites and chains, rationale, expected direction and whether the event was planned; freeze entries before estimating its effect and distinguish evidence from retrospective explanation.
Worked example. A mobile site opens in week five and recruitment from one district rises. The ledger shows access changed, preventing an unsupported claim that network composition suddenly shifted.
Changed-context transfer. Changed-context transfer: apply the same reasoning structure to a school or public-service decision. Preserve the formal relationship, then replace the population, authority, privacy duties, evidence and remedy; never copy the domain conclusion.
Limits. Unrecorded informal practices and simultaneous events can leave causal attribution uncertain.
Evidence and review. For this chapter, keep a dated evidence record: the input, rule or assumption, result, accountable owner and a condition that would change the conclusion. Work an ordinary case and a boundary case, preserve rejected alternatives, and require an authorised independent reader to reproduce where judgement entered. Chapter 3 is complete only when the claimed result survives that review.
Misconception to avoid. Field operations are implementation details. They help define the sampling mechanism and need analytic provenance.
Practical evidence check
- Remove event labels from a plot, ask a reviewer to infer changes, then compare with the contemporaneous ledger.
- Return to chapter contents
CHAPTER 4 OF 13 · Apply the core method
4. Diagnose recruitment productivity through time
Core idea. Numbers of coupons issued, redeemed and expired, recruitment delay and active chains can change before sample composition visibly shifts. These process measures can reveal shocks and bottlenecks early.
Method. Plot weekly or daily coupons, redemption hazards, active recruiters, branch factors and interview volume by site and seed; show denominators and uncertainty rather than only cumulative sample size.
Worked example. Total enrolment remains smooth because new seeds are added, but redemption per issued coupon halves after a venue closure. The process diagnosis exposes the hidden shock.
Changed-context transfer. Changed-context transfer: apply the same reasoning structure to a school or public-service decision. Preserve the formal relationship, then replace the population, authority, privacy duties, evidence and remedy; never copy the domain conclusion.
Limits. Counts are affected by staffing and coupon availability, and sparse chains make subgroup rates noisy.
Evidence and review. For this chapter, keep a dated evidence record: the input, rule or assumption, result, accountable owner and a condition that would change the conclusion. Work an ordinary case and a boundary case, preserve rejected alternatives, and require an authorised independent reader to reproduce where judgement entered. Chapter 4 is complete only when the claimed result survives that review.
Misconception to avoid. A steady cumulative sample proves steady recruitment. Operational compensation can mask instability.
Practical evidence check
- Construct identical cumulative curves from different issue and redemption patterns and identify which threatens inference.
- Return to chapter contents
CHAPTER 5 OF 13 · Apply the core method
5. Separate seasonality from one-off shocks
Core idea. Regular weekly, monthly or annual cycles differ from abrupt events such as enforcement, a disease outbreak or a payment failure. Pooling them under “time trend” can lead to the wrong correction and transfer claim.
Method. Use pre-specified calendar indicators and smooth terms only when data span supports them; annotate independent shocks; compare repeated cycles and hold out periods; avoid fitting elaborate seasonality to one short campaign.
Worked example. Weekend participation repeatedly rises at two sites, while a single festival week produces a unique composition shift. They require different interpretation.
Changed-context transfer. Changed-context transfer: apply the same reasoning structure to a school or public-service decision. Preserve the formal relationship, then replace the population, authority, privacy duties, evidence and remedy; never copy the domain conclusion.
Limits. Seasonal patterns may interact with subgroup, site and weather, and a short study cannot identify an annual cycle.
Evidence and review. For this chapter, keep a dated evidence record: the input, rule or assumption, result, accountable owner and a condition that would change the conclusion. Work an ordinary case and a boundary case, preserve rejected alternatives, and require an authorised independent reader to reproduce where judgement entered. Chapter 5 is complete only when the claimed result survives that review.
Misconception to avoid. Any periodic-looking plot proves seasonality. Repetition and external justification are needed.
Practical evidence check
- Fit the same short sequence with a trend, weekly cycle and shock indicator; compare out-of-period predictions.
- Return to chapter contents
Recruitment-shock example
| Period | Observed pattern | Possible causes | Needed evidence |
|---|---|---|---|
| Weeks 1–3 | Steady enrolment | Baseline process | Chain and site logs |
| Week 4 | Abrupt drop | Holiday, policing, outage | Dated annotations |
| Week 5 | New seeds added | Operational response | Predeclared rationale |
| Weeks 6–8 | Composition shifts | Recovery or new subnetwork | Sensitivity analysis |
The estimate should not treat a protocol response as if it were ordinary network sampling.
CHAPTER 6 OF 13 · Test evidence and alternatives
6. Inspect composition by date and chain simultaneously
Core idea. A changing sample characteristic may reflect calendar drift, seed dependence, chain bottlenecks or differential recruitment. Looking only at cumulative equilibrium plots can obscure when and where change occurred.
Method. Plot rolling and cumulative composition by calendar date, wave, seed and site; use consistent windows; inspect recruiter–recruit transitions and mark protocol events without treating visual stability as proof.
Worked example. One subgroup rises only in chains active after a new site opens, suggesting both calendar access and chain structure contribute.
Changed-context transfer. Changed-context transfer: apply the same reasoning structure to a school or public-service decision. Preserve the formal relationship, then replace the population, authority, privacy duties, evidence and remedy; never copy the domain conclusion.
Limits. Many plots invite selective interpretation and standard tests may be invalid because observations are network dependent.
Evidence and review. For this chapter, keep a dated evidence record: the input, rule or assumption, result, accountable owner and a condition that would change the conclusion. Work an ordinary case and a boundary case, preserve rejected alternatives, and require an authorised independent reader to reproduce where judgement entered. Chapter 6 is complete only when the claimed result survives that review.
Misconception to avoid. A stable cumulative estimate means the process is stationary. Opposing time-varying biases can cancel.
Practical evidence check
- Create two chains with opposite drifts whose pooled cumulative estimate stays flat and diagnose the hidden changes.
- Return to chapter contents
CHAPTER 7 OF 13 · Test evidence and alternatives
7. Handle phased sites and staggered seeds explicitly
Core idea. Opening sites or adding seeds in phases changes who can enter the sample and how long each chain can recruit. Late phases are mechanically more censored and should not be treated as exchangeable with early ones.
Method. Record phase eligibility and exposure time, compare overlapping periods, stratify process diagnostics, use phase indicators or restricted common windows and report why new seeds were introduced.
Worked example. A second district begins halfway through fieldwork. Its chains have fewer waves at closure, so a wave-based comparison without exposure adjustment is misleading.
Changed-context transfer. Changed-context transfer: apply the same reasoning structure to a school or public-service decision. Preserve the formal relationship, then replace the population, authority, privacy duties, evidence and remedy; never copy the domain conclusion.
Limits. Restricting to common periods may waste data, while modelling phase effects relies on untestable assumptions.
Evidence and review. For this chapter, keep a dated evidence record: the input, rule or assumption, result, accountable owner and a condition that would change the conclusion. Work an ordinary case and a boundary case, preserve rejected alternatives, and require an authorised independent reader to reproduce where judgement entered. Chapter 7 is complete only when the claimed result survives that review.
Misconception to avoid. More seeds only improve coverage. They can also change the mixture and mask failing chains.
Practical evidence check
- Estimate using all phases, an overlap window and phase-specific summaries; show which conclusion depends on pooling.
- Return to chapter contents
CHAPTER 8 OF 13 · Test evidence and alternatives
8. Treat calendar closure as informative censoring
Core idea. Fieldwork end dates cut off active recruitment chains. Coupons issued near closure have less opportunity to be redeemed, and slow-recruiting groups may be underrepresented even if expiry rules are equal.
Method. Compute time at risk to expiry and closure, display recruitment-delay distributions by subgroup and phase, use survival-style diagnostics and perform sensitivity analyses for unresolved coupons.
Worked example. A coupon issued two days before closure is counted as a failure beside one observed for three weeks. Time-aware analysis distinguishes opportunity.
Changed-context transfer. Changed-context transfer: apply the same reasoning structure to a school or public-service decision. Preserve the formal relationship, then replace the population, authority, privacy duties, evidence and remedy; never copy the domain conclusion.
Limits. Issue time may not equal handoff time, and unobserved coupon recipients limit causal interpretation.
Evidence and review. For this chapter, keep a dated evidence record: the input, rule or assumption, result, accountable owner and a condition that would change the conclusion. Work an ordinary case and a boundary case, preserve rejected alternatives, and require an authorised independent reader to reproduce where judgement entered. Chapter 8 is complete only when the claimed result survives that review.
Misconception to avoid. All unredeemed coupons are comparable nonresponse. Administrative censoring and behavioural refusal are different.
Practical evidence check
- Shift the closure date in a simulation and measure which subgroup estimates change without altering preferences.
- Return to chapter contents
Time-sensitivity certificate
| Claim | Evidence | Boundary |
|---|---|---|
| Process stable | Dated productivity and composition | Power for slow drift |
| Shock explained | Independent field record | Competing causes |
| Estimate robust | Phase and time reweighting | Model dependence |
| Target current | Population-period definition | Future transfer |
A pooled estimate can be precise while describing no stable population.
CHAPTER 9 OF 13 · Handle limits and disagreement
9. Test estimator sensitivity to time-sensitive mechanisms
Core idea. Standard RDS estimators use reported degree and observed recruitment structure but usually do not remove nonstationarity automatically. A precise estimate can remain sensitive to calendar composition, seed timing and phased access.
Method. Compare full-sample estimates with time blocks, overlap periods, leave-phase-out analyses and plausible time or response models; retain design-based and unweighted summaries; report changes with uncertainty.
Worked example. The headline prevalence changes materially when the shock week and late added seeds are excluded, so the team reports a range and limits the target claim.
Changed-context transfer. Changed-context transfer: apply the same reasoning structure to a school or public-service decision. Preserve the formal relationship, then replace the population, authority, privacy duties, evidence and remedy; never copy the domain conclusion.
Limits. Subset estimates can be unstable and multiple sensitivity choices can be selectively reported. Predefine the main alternatives.
Evidence and review. For this chapter, keep a dated evidence record: the input, rule or assumption, result, accountable owner and a condition that would change the conclusion. Work an ordinary case and a boundary case, preserve rejected alternatives, and require an authorised independent reader to reproduce where judgement entered. Chapter 9 is complete only when the claimed result survives that review.
Misconception to avoid. RDS weights correct any calendar-time imbalance. They depend on assumptions that time variation may violate.
Practical evidence check
- Construct a synthetic time-varying participation process and compare naïve, standard-weighted and phase-aware estimates.
- Return to chapter contents
CHAPTER 10 OF 13 · Handle limits and disagreement
10. Avoid causal stories from observational timing
Core idea. A recruitment drop after an event does not prove the event caused it. Weather, staffing, depleted networks, holidays and changing respondent behaviour can coincide, and the RDS dependence structure complicates standard tests.
Method. Use independent event evidence, negative-control sites or outcomes where available, interrupted-time analyses cautiously and language such as consistent with rather than caused by unless design supports causality.
Worked example. Recruitment falls after an enforcement announcement at one site but not another; the contrast is informative, yet site differences prevent a clean causal estimate.
Changed-context transfer. Changed-context transfer: apply the same reasoning structure to a school or public-service decision. Preserve the formal relationship, then replace the population, authority, privacy duties, evidence and remedy; never copy the domain conclusion.
Limits. Hidden contemporaneous changes and short pre-periods can defeat causal identification.
Evidence and review. For this chapter, keep a dated evidence record: the input, rule or assumption, result, accountable owner and a condition that would change the conclusion. Work an ordinary case and a boundary case, preserve rejected alternatives, and require an authorised independent reader to reproduce where judgement entered. Chapter 10 is complete only when the claimed result survives that review.
Misconception to avoid. A timestamped sequence establishes cause. Temporal order is necessary but not sufficient.
Practical evidence check
- List three competing explanations for one break and the additional evidence each would predict.
- Return to chapter contents
CHAPTER 11 OF 13 · Handle limits and disagreement
11. Protect privacy in fine-grained timelines
Core idea. Exact times, small sites and recruitment links can identify participants or sensitive gathering patterns. Diagnostics should preserve enough temporal structure for review without publishing re-identification pathways.
Method. Separate restricted event-level data from public aggregates, coarsen time and site where needed, suppress sparse cells, protect keys and review plots for unique chains and events.
Worked example. Internal analysts use timestamped coupon data; the published figure uses weekly intervals and broad regions while retaining the visible recruitment shock.
Changed-context transfer. Changed-context transfer: apply the same reasoning structure to a school or public-service decision. Preserve the formal relationship, then replace the population, authority, privacy duties, evidence and remedy; never copy the domain conclusion.
Limits. Coarsening can hide short-lived failures and correlated releases can recreate detail.
Evidence and review. For this chapter, keep a dated evidence record: the input, rule or assumption, result, accountable owner and a condition that would change the conclusion. Work an ordinary case and a boundary case, preserve rejected alternatives, and require an authorised independent reader to reproduce where judgement entered. Chapter 11 is complete only when the claimed result survives that review.
Misconception to avoid. Operational metadata are not personal data. Time and network position can be highly identifying.
Practical evidence check
- Attempt a linkage attack on a synthetic public timeline and adjust bins until the analytic pattern remains without unique traces.
- Return to chapter contents
CHAPTER 12 OF 13 · Transfer and govern
12. Monitor during fieldwork without chasing noise
Core idea. Time-aware diagnostics are most useful while corrective action remains possible, but frequent reactions to random fluctuation can change the design repeatedly and introduce new bias.
Method. Predefine indicators, review cadence, warning and action thresholds, responsible roles and allowed interventions; preserve pre-change data, rationale and a frozen analysis plan after each decision.
Worked example. A two-day slowdown triggers observation, not immediate seed replacement; a sustained multi-indicator failure activates a documented contingency.
Changed-context transfer. Changed-context transfer: apply the same reasoning structure to a school or public-service decision. Preserve the formal relationship, then replace the population, authority, privacy duties, evidence and remedy; never copy the domain conclusion.
Limits. Rare populations yield noisy indicators and ethical or safety events may require action outside statistical thresholds.
Evidence and review. For this chapter, keep a dated evidence record: the input, rule or assumption, result, accountable owner and a condition that would change the conclusion. Work an ordinary case and a boundary case, preserve rejected alternatives, and require an authorised independent reader to reproduce where judgement entered. Chapter 12 is complete only when the claimed result survives that review.
Misconception to avoid. Adaptive management means changing the design whenever a chart moves. Controlled adaptation needs evidence and provenance.
Practical evidence check
- Replay historical monitoring using the thresholds and count false alarms, missed shocks and design changes.
- Return to chapter contents
CHAPTER 13 OF 13 · Transfer and govern
13. Publish a time-aware RDS evidence record
Core idea. A credible estimate links the target period, recruitment tree, timestamps, phase ledger, closure rules, diagnostics, estimator versions and sensitivity results. Readers should see what was stable, what changed and what cannot be separated.
Method. Release de-identified timelines, definitions, code or pseudocode, event annotations, primary and alternative estimates, privacy limits and correction history; archive restricted details for qualified review.
Worked example. An independent reviewer reproduces the phase comparison, recruitment-delay curve and leave-shock-out estimate without access to participant identities.
Changed-context transfer. Changed-context transfer: apply the same reasoning structure to a school or public-service decision. Preserve the formal relationship, then replace the population, authority, privacy duties, evidence and remedy; never copy the domain conclusion.
Limits. Confidentiality and local safety may prevent full data release; reproducibility can use controlled access and synthetic fixtures.
Evidence and review. For this chapter, keep a dated evidence record: the input, rule or assumption, result, accountable owner and a condition that would change the conclusion. Work an ordinary case and a boundary case, preserve rejected alternatives, and require an authorised independent reader to reproduce where judgement entered. Chapter 13 is complete only when the claimed result survives that review.
Misconception to avoid. One weighted estimate is a complete report. Time and operational evidence determine what that number represents.
Practical evidence check
- Rebuild the claim from dated coupon issue through final interval and challenge one stationarity assumption.
- Return to chapter contents
Build the specialist decision record
For “The Gold Standard of Intelligence | RDS Calendar-Time Nonstationarity, Seasonality, Recruitment Shocks, Phased Fieldwork and Estimator Sensitivity”, begin with one bounded decision rather than an abstract claim. Name the affected people, data or institutions; the time window; the authority for acting; and the conclusion the method is allowed to support.
Preserve source material before interpretation. Record definitions, exclusions, transformations and the person responsible for each judgement. A reviewer should be able to separate an observed fact from an inferred result, a policy choice and a presentation shortcut.
Work one ordinary case in full. Show the input, every material rule or calculation, the output and the action that follows. Then write what the same evidence would mean under the simplest credible alternative method.
Keep an explicit unresolved state. Missing, protected, conflicting or unevaluated information must not become zero, false, consent or absence merely because a form expects one value.
Finish the record with a claim ledger. For every sentence that could drive action, identify its supporting source, inferential step, confidence or legal status, intended reader and expiry condition. This keeps descriptive evidence, formal consequence, professional judgement and institutional authority from collapsing into one impressive-looking conclusion.
Run the boundary-case audit
Select a case near the method’s domain edge: a missing input, a disputed definition, an endpoint, a sparse group, a resource limit or a legal deadline. Boundary cases reveal the contract that routine examples can conceal.
Change one assumption at a time and retain every output. Mark what stays stable, what reverses and what becomes undefined. If several assumptions move together, report that dependence instead of attributing the change to one convenient cause.
Search deliberately for a counterexample to the strongest claim. A method that works on the published illustration can still fail under a permitted input, an alternative grouping, a changed unit or an inaccessible evidence pathway.
Translate the counterexample into plain language: who would receive a different decision, what evidence would be hidden, and which safeguard or narrower claim repairs the problem.
Include an adversarial but plausible case. Let a missing group, ambiguous identifier, extreme weight, conflicting rule or urgent deadline interact with normal system incentives. The aim is not theatrical failure; it is to learn whether ordinary pressure makes the method conceal uncertainty, shift burden or reward a misleading shortcut.
Test independent reproduction
Give a reviewer the bounded inputs, public method contract and permitted access, but not the desired answer. Ask them to reproduce the result and identify every place where judgement, local knowledge or unavailable evidence was required.
Compare the reviewer’s intermediate states, not only the final number or label. Matching outputs can hide different denominators, inference paths, missing-data treatments or authorities that will diverge on the next case.
Where reproduction fails, classify the cause: ambiguous definition, undocumented transformation, incompatible implementation, inaccessible source, calculation error or legitimate alternative interpretation. Each class needs a different remedy.
Retain the disagreement and resolution. A corrected record should show what changed, why it changed and which prior outputs or decisions must be revisited.
Run the reproduction on a changed but structurally similar case after the first discrepancy is repaired. A method is not transferable merely because its authors can explain the original example. The new case tests whether the public contract carries the reasoning or whether hidden expertise is still doing essential work.
Design correction, challenge and remedy
Assign a visible route for correcting inputs, challenging assumptions and reviewing the application of the method. Technical accuracy without a usable remedy leaves the affected person carrying an error they cannot contest.
Separate urgent live-case repair from later system learning. A current decision may need pause, reassignment, fresh evidence or a human review before the organisation finishes its longer investigation.
Protect confidential and privileged material during review, but do not use confidentiality as a reason for unrecorded authority. Layer public explanation, protected oversight and individual reasons according to lawful access.
Verify that correction reaches cached results, derived records, dashboards and future decisions. Closing a complaint is not enough if the old conclusion continues to travel.
Measure remedy from the affected person’s position. A revised database row is useful, but it is incomplete when a hearing, service allocation, published claim or downstream model still relies on the old result. Record notification, restored opportunity, compensation or procedural repair where the governing framework provides it.
Set review and retirement triggers
Review after a change in definitions, population, law, source access, measurement scale, software, vocabulary or incentives. A stable formula can become invalid when the world or the meaning of its inputs changes.
Monitor leading signals as well as final outcomes: coverage gaps, timeouts, overrides, complaints, disagreement, missingness and boundary frequency. These often reveal failure before a headline metric moves.
Predefine thresholds for recalibration, suspension and retirement. Do not force a method to answer after its assumptions fail merely because replacing it is inconvenient.
At retirement, preserve the versioned contract, evidence, known limitations and migration decision. State which historic results remain interpretable and which require warning, restatement or withdrawal.
Schedule one prospective review before confidence fades or controversy forces it. Assign named owners for evidence, implementation and rights impact, and invite at least one reviewer outside the design team. The review should be able to narrow the claim, change the method or stop use without requiring proof of catastrophic failure.
Questions readers ask
Is recruitment wave the same as calendar time?
- No. Wave measures network distance from a seed; calendar time measures changing external and operational conditions.
Why keep a fieldwork ledger?
- It distinguishes protocol and access changes from apparent changes in the target population or network.
Can RDS weights correct seasonality automatically?
- No. They rely on sampling and network assumptions that calendar-time nonstationarity can violate.
Why are late coupons different?
- They have less time to be redeemed before administrative closure, creating censoring.
Should every time fluctuation trigger a design change?
- No. Use predeclared review rules and distinguish monitoring from intervention.
What should the final estimate name?
- The population, geography, calendar period, fieldwork phases, estimator and sensitivity boundaries.
Practise, verify and continue
Choose one consequential case. Reproduce the complete method, work one ordinary and one boundary case, test the strongest rival explanation, and state the evidence that would reverse or suspend the result. Build an evidence pack containing original inputs, definitions, calculations or rule steps, alternatives rejected, uncertainty, reviewer comments and the final action. Give the pack to a reader who did not design the method. Ask that reader to reproduce the result, identify one missing assumption and construct a case that would make the method return unresolved. Revise the public explanation only after the evidence pack and the operational process agree.
Move the method into an education or public-service setting. Keep the reasoning structure while replacing domain authority, population, privacy, evidence and correction routes. Start with a low-stakes historical example, then compare the specialist method with a transparent baseline. Check whether the same people, records and decisions remain inside scope; whether the scale or vocabulary still means the same thing; and whether a person affected by error can understand and challenge it. Pilot under human review, record reversals and burden, and stop the transfer if the formal resemblance hides a different ethical or legal problem.
Related eduKate reading
RDS coupon timing, chain censoring and stopping rules
RDS design effects and convergence diagnostics
RDS recruiter choice and differential recruitment
Authoritative sources and scope
Gile, Johnston and Salganik, Diagnostics for Respondent-Driven Sampling
Gile and Handcock, Respondent-Driven Sampling: An Assessment of Current Methodology
Crawford et al., Modeling and Analyzing Respondent-Driven Sampling as a Counting Process
Crawford, Wu and Heimer, Hidden Population Size Estimation From Respondent-Driven Sampling
Exploring Mechanisms of Recruitment and Recruitment Cooperation in RDS
This article is sampling-method education, not a fieldwork protocol, clinical recommendation or population estimate. Examples are original teaching cases. Respondent-driven sampling depends on strong assumptions and calendar-time mechanisms may not be identifiable from recruitment data alone. Protect participant and network privacy, follow local ethics and safety review, preserve contemporaneous field records and report sensitivity or abstention when the target population, access process or fieldwork design changes.