{"schema_version":1,"research_id":"eoa_inverse_innovation_exp06_external_evaluation_20260803","source_assessment_id":"predictive_residual_processing__sociology_anthropology:P3:v0","cell_id":"predictive_residual_processing__sociology_anthropology","search_queries":["longitudinal social network survey dependent interviewing prior wave roster burden name generator","personal network survey roster prior wave confirm relationships longitudinal study","dependent interviewing previous answers reduces burden longitudinal survey official","Network Canvas roster data import prior wave interview documentation","Cornwell colleagues 2014 longitudinal social network roster previous wave contacts survey","Cornwell Laumann Schumm social network longitudinal roster previous wave 2014 name generator","personal network longitudinal survey previous wave alters preloaded questionnaire study","social network survey anchoring roster recognition recall omissions prior names","\"Assessment of social network change in a national longitudinal survey\" full text Wave 2 retained network members","site:pmc.ncbi.nlm.nih.gov/articles/PMC4303098 \"Wave 2\" network roster","NSHAP Wave 2 social network roster prior wave members follow-up questions","UCNets survey instrument prior wave names alters listed questionnaire","site:hhs.gov/ohrp regulations 45 CFR 46 privacy confidentiality minimize risks informed consent research","site:theasa.org ethics code anthropological research confidentiality informed consent relationships","site:aapor.org standards respondent burden privacy survey research ethics","IRB social network research confidentiality third party information ethics","site:bls.gov/oes 2025 anthropologists survey researchers hourly mean wage software developers","site:bls.gov/ooh social scientists anthropologists median pay 2025","site:bls.gov/oes current survey researchers wage May 2024","\"Who is Dropped and Why?\" \"forgot\" UCNets full text","site:escholarship.org \"Who is Dropped and Why\" network loss","\"respondents explained that they had forgotten the alter\" UCNets"],"sources":[{"source_id":"S1","title":"Reducing Burden in a Web Survey through Dependent Interviewing","publisher":"Journal of Survey Statistics and Methodology, Oxford University Press","url":"https://academic.oup.com/jssam/article/12/1/60/7078538","source_class":"PRIMARY_RESEARCH","publication_date":"2023-03-15","accessed_at":"2026-08-03","claims_supported":["Repeated questions in longitudinal surveys create measurable respondent burden.","A controlled web-survey experiment found dependent interviewing faster and more enjoyable than traditional interviewing without significant introduced measurement error in the studied items.","Prior-answer presentation can cause acquiescence or underreporting of change and therefore requires measurement-error testing."]},{"source_id":"S2","title":"Dependent Interviewing: 2015 to 2023","publisher":"U.S. Census Bureau","url":"https://www.census.gov/programs-surveys/ahs/tech-documentation/help-guides/2015-later/Dependent_Interviewing18.html","source_class":"GOVERNMENT_OR_REGULATOR","publication_date":"2024-09-13","accessed_at":"2026-08-03","claims_supported":["The American Housing Survey used prior survey results in dependent interviewing from 2015 through 2023.","The Census Bureau expressly identifies reduced respondent burden as the purpose of the practice.","A credible large-scale survey adopter and operational precedent exist."]},{"source_id":"S3","title":"Working with Roster Data","publisher":"Network Canvas","url":"https://documentation.networkcanvas.com/en/desktop/tutorials/working-with-rosters","source_class":"OFFICIAL_PRODUCT_DOCUMENTATION","publication_date":"n.d.","accessed_at":"2026-08-03","claims_supported":["Network Canvas supports participant-facing roster nomination and structured node attributes.","Previous-survey names are a recognized roster use case, although participant-specific longitudinal rosters currently require workarounds.","Roster name collisions and longitudinal data matching are known implementation risks."]},{"source_id":"S4","title":"Assessment of Social Network Change in a National Longitudinal Survey","publisher":"The Gerontological Society of America / Oxford University Press","url":"https://pmc.ncbi.nlm.nih.gov/articles/PMC4303098/","source_class":"PRIMARY_RESEARCH","publication_date":"2014-11-01","accessed_at":"2026-08-03","claims_supported":["NSHAP implemented a longitudinal network-change module matching respondents' prior- and current-wave network members.","The method distinguished stable, lost, and new alters and asked follow-up questions about unmatched prior alters.","The module obtained valid change data from about 98 percent of respondents completing both waves, demonstrating technical and interview-workflow feasibility.","Network turnover estimates remain vulnerable to roster caps and reporting effects."]},{"source_id":"S5","title":"UCNets Main Survey Waves 1, 2, and 3 Questionnaire","publisher":"University of California Social Networks Study, UC Berkeley","url":"https://ucnets.berkeley.edu/wp-content/uploads/2020/03/UCNets_MainSurvey_Waves-1-2-3_Questionnaire_25March2020-1.pdf","source_class":"OFFICIAL_ORGANIZATION_DATA","publication_date":"2020-03-25","accessed_at":"2026-08-03","claims_supported":["UCNets imported prior-wave names, matched them to current names, and queried names that were not re-listed.","The instrument distinguished renamed people, forgotten people, deaths, loss of contact, and relationship changes.","It revisited names omitted and later re-listed and asked why apparently new alters were first listed.","This is close operational prior art for structured longitudinal relationship-map change elicitation."]},{"source_id":"S6","title":"Who Is Dropped and Why? Methodological and Substantive Accounts for Network Loss","publisher":"Social Networks; manuscript hosted by UC Berkeley eScholarship","url":"https://escholarship.org/content/qt8jd7j0hc/qt8jd7j0hc.pdf","source_class":"PRIMARY_RESEARCH","publication_date":"2020-05-01","accessed_at":"2026-08-03","claims_supported":["In about 1,000 UCNets respondents, roughly half of prior-wave alters were not re-listed one year later.","Respondents said about 41 percent of apparently dropped alters had simply been forgotten, showing that silence cannot safely mean tie termination or continuity.","The study identifies distinguishing real network change from reporting error as a foundational measurement problem.","The authors recommend asking whether apparently dropped names were forgotten and document fatigue, satisficing, recency, and interviewer effects as error mechanisms."]},{"source_id":"S7","title":"Informed Consent FAQs","publisher":"U.S. Department of Health and Human Services, Office for Human Research Protections","url":"https://www.hhs.gov/ohrp/regulations-and-policy/guidance/faq/informed-consent/index.html","source_class":"OFFICIAL_GUIDANCE","publication_date":"n.d.","accessed_at":"2026-08-03","claims_supported":["Covered human-subjects research generally requires legally effective prospective informed consent unless an IRB documents an applicable waiver.","Consent must support comprehension, voluntariness, and an ongoing opportunity to decide whether to continue.","An IRB is the relevant authorizer for waivers, alterations, and protocol changes; participant-facing delta elicitation cannot itself expand consent scope."]},{"source_id":"S8","title":"National Employment and Wage Data by Occupation, May 2025","publisher":"U.S. Bureau of Labor Statistics","url":"https://www.bls.gov/news.release/ocwage.t01.htm","source_class":"GOVERNMENT_OR_REGULATOR","publication_date":"2026-05-15","accessed_at":"2026-08-03","claims_supported":["May 2025 mean wages were $79,580 for survey researchers and $148,100 for software developers.","The wage data provide official labor anchors for 2026 resource-equivalent estimates.","Annual wage figures assume 2,080 full-time hours, requiring separate allowances for benefits, overhead, participant incentives, ethics review, and field operations."]}],"problem_evidence":{"support":"STRONG","rationale":"The problem is visible in both general longitudinal-survey research and relationship-network studies. Repetition creates burden, while omission and ambiguous non-relisting are substantial: UCNets found that about half of prior alters were not re-listed and that 41 percent of those omissions were attributed to forgetting. This directly supports the candidate's concern that neither repeated full elicitation nor silent carry-forward is harmless. The exact prevalence of burdensome full roster re-enumeration across anthropology and community fieldwork remains unmeasured.","source_ids":["S1","S4","S6"]},"stakeholder_evidence":{"support":"MODERATE","rationale":"Credible adopters visibly use adjacent methods: the Census Bureau uses dependent interviewing to reduce burden; NSHAP and UCNets implemented prior/current roster matching; and Network Canvas documents demand and workarounds for longitudinal rosters. These sources demonstrate institutional need and adoption capability, but no identified study, funder, community advisory group, or ethics authority has requested this complete checksum-audit-fallback package.","source_ids":["S2","S3","S4","S5"]},"prior_art":{"proximity":"ESTABLISHED_PRACTICE","closest_analogues":[{"name":"UCNets longitudinal name-matching and missing-name checks","similarity":"Imports prior-wave names, matches them to the current wave, asks about unmatched prior names, distinguishes forgetting from relationship change, and revisits apparently new or returning alters.","remaining_difference":"It begins with renewed name elicitation rather than making prior-state-plus-residual reconstruction the primary interview representation; the published instrument does not document checksums, explicit field-level review states, random full-roster audits, or participant-triggered decompression as an integrated protocol.","source_ids":["S5","S6"]},{"name":"NSHAP longitudinal social-network change module","similarity":"Matches network rosters across waves and distinguishes stable, lost, and new ties with follow-up on unmatched names.","remaining_difference":"The reported module is designed to measure network change, not specifically to replace most full re-enumeration through a versioned delta channel with audit sampling and an error budget.","source_ids":["S4"]},{"name":"Proactive dependent interviewing","similarity":"Presents prior answers for confirmation or correction to reduce repeated longitudinal questioning and burden.","remaining_difference":"Generic dependent interviewing does not by itself provide relationship-graph reconstruction, tie-level provenance, participant-specific snapshot synchronization, independent raw-roster audits, or subgroup-sensitive fallback rules.","source_ids":["S1","S2"]},{"name":"Network Canvas roster workflow","similarity":"Provides participant-facing roster selection, structured alter attributes, and exportable network records.","remaining_difference":"Participant-specific longitudinal rosters require workarounds, and the documentation does not supply the proposed version handshake, residual semantics, periodic full refresh, or ethics-governed audit layer.","source_ids":["S3"]}],"distinctive_claim_remaining":"The narrow remaining claim is that adding explicit reviewed/unknown/declined states, snapshot-version matching, participant-confirmed reconstruction, random and risk-stratified full-roster audits, and participant-controlled fallback to longitudinal dependent interviewing will reduce interview time and repeated disclosure relative to full re-enumeration while producing fewer omitted or falsely persistent consequential ties than simple prior-roster confirmation. This is contrastive and falsifiable, but presently unsupported by a direct comparative field test.","confidence":"HIGH"},"implementation_evidence":{"support":"MODERATE","rationale":"The constituent workflow is technically credible: large survey programs already implement dependent interviewing and roster matching, Network Canvas supports structured rosters, and UCNets demonstrates automated prior-name matching and explicit reasons for omissions. The integrated versioning, consent-scope checks, participant confirmation, audit sampling, subgroup monitoring, and fallback logic are feasible conventional software and protocol work. However, no source verifies the complete architecture, its field usability, independence of its audit channel, anchoring effects, or net savings after audit and fallback costs.","source_ids":["S1","S3","S4","S5","S7"]},"scores":{"meaningful_impact":{"score":3,"rationale":"The intervention addresses real burden and consequential measurement error, but the size of avoidable burden and downstream analytic harm in the intended community-fieldwork settings is not quantified.","source_ids":["S1","S6"]},"stakeholder_pull":{"score":3,"rationale":"Census, NSHAP, UCNets, and Network Canvas show active demand for dependent interviewing and longitudinal roster handling, but no adopter has expressed demand for this full package.","source_ids":["S2","S3","S4","S5"]},"incremental_advantage":{"score":2,"rationale":"Most of the operational benefit is already pursued by proactive dependent interviewing and longitudinal missing-name checks. The audit-governance layer may improve safety and validity, but its net advantage has not been measured.","source_ids":["S1","S4","S5"]},"distinctiveness_plausibility":{"score":2,"rationale":"The core method is established practice. Distinctiveness survives only in the integrated checksum, explicit-silence, participant-control, audit, error-budget, and fallback package; world novelty remains unmeasured.","source_ids":["S1","S4","S5"]},"technical_implementability":{"score":4,"rationale":"Prior-wave roster imports, structured edits, matching, reconstruction, and version identifiers are routine survey-software functions, although Network Canvas currently requires longitudinal workarounds.","source_ids":["S3","S5"]},"adoption_authority_feasibility":{"score":3,"rationale":"A PI, data steward, and IRB can authorize a bounded protocol amendment, but consent scope, third-party relationship information, privacy controls, and participant-facing display rules require study-specific approval.","source_ids":["S7"]},"evidence_readiness":{"score":3,"rationale":"Existing consecutive-wave records support offline reconstruction and error profiling, but the decisive burden, anchoring, comfort, omission, and subgroup outcomes require prospective participant research.","source_ids":["S4","S5","S6"]},"safety_net_benefit":{"score":4,"rationale":"Explicit missingness, participant-requested full elicitation, independent audits, and periodic refresh directly target known forgetting and false-turnover risks; effectiveness is not yet demonstrated.","source_ids":["S5","S6","S7"]},"scalability":{"score":3,"rationale":"Dependent interviewing scales in national surveys and roster tooling exists, but participant-specific sensitive snapshots, audit interviews, and governance overhead may constrain multi-site scaling.","source_ids":["S2","S3","S8"]}},"score_confidence":"MODERATE","costs":{"first_evidence":{"band_2026_usd":"50K_TO_250K","scope":"IRB-approved three-arm field experiment with approximately 120 returning participants, instrument programming, secure snapshot preparation, interviewer training, participant compensation, double-elicitation audit subsample, analysis, and community review.","confidence":"MODERATE","assumptions":["Approximately 0.5 FTE survey methodologist, 0.25 FTE software developer, 0.25 FTE data steward, interviewer labor, and participant incentives over 6-9 months.","Loaded labor is estimated above official wages to cover benefits and institutional overhead.","An existing longitudinal cohort and authorized prior rosters are available; new cohort recruitment would raise cost.","No major custom mobile platform or multi-language validation is included."],"source_ids":["S7","S8"]},"initial_deployment_startup":{"band_2026_usd":"50K_TO_250K","scope":"Production protocol design, consent and IRB amendment, participant-specific snapshot generation, checksum and provenance controls, explicit response-state schema, audit dashboard, security testing, interviewer materials, and rollback procedures for one study.","confidence":"MODERATE","assumptions":["Builds on an existing survey platform and database.","Includes several months of survey-methodology, software, data-governance, and ethics effort.","Excludes replacement of the study's core case-management platform."],"source_ids":["S3","S7","S8"]},"operational_launch":{"band_2026_usd":"250K_TO_1M","scope":"First production wave for a medium longitudinal study, including participant support, interview operations, random and risk-stratified full-roster audits, fallback interviews, quality review, incident response, and outcome analysis.","confidence":"LOW","assumptions":["Approximately 500-1,500 participants across one or a few sites.","A material subset receives full audit or fallback interviews.","Costs vary sharply by interview mode, geography, compensation, translation, and safeguarding requirements."],"source_ids":["S2","S8"]},"annual_recurring":{"band_2026_usd":"50K_TO_250K","scope":"Snapshot refresh, secure storage, software maintenance, audit sampling, model-drift and subgroup review, interviewer retraining, IRB continuing review, and periodic full re-enumeration outside major data-collection expansion.","confidence":"LOW","assumptions":["One established study with annual or biennial waves.","Core hosting and case management already exist.","A major full-roster wave or multi-country operation could exceed this band."],"source_ids":["S3","S7","S8"]}},"verified_pipeline_gates":{"externally_supported_problem":{"status":"YES","reason":"Primary studies directly document repeated-survey burden, apparent network churn, forgetting, and difficulty distinguishing real relationship change from reporting error.","source_ids":["S1","S6"]},"externally_credible_adopter_or_authorizer":{"status":"YES","reason":"The Census Bureau, NSHAP, and UCNets are credible adopters of adjacent workflows, while the applicable PI and IRB are identifiable authorization roles. No commitment to this exact package is documented.","source_ids":["S2","S4","S5","S7"]},"distinct_testable_incremental_claim":{"status":"YES","reason":"The integrated audit-governed delta protocol can be compared with both full re-enumeration and simple prior-roster confirmation on interview time, repeated disclosure, consequential tie error, subgroup error, comfort, and fallback burden.","source_ids":["S1","S5","S6"]},"bounded_next_evidence_step":{"status":"YES","reason":"A preregistered, approximately 120-participant, three-arm trial with an independent full-roster audit subsample is bounded and has explicit comparators and falsifiers.","source_ids":["S1","S4","S5","S6"]},"no_unresolved_safety_or_authority_stop":{"status":"UNCERTAIN","reason":"The design contains sensible consent, fallback, and provenance controls, but no study-specific IRB approval, third-party privacy determination, safeguarding procedure, or independent audit implementation has been secured or tested.","source_ids":["S7"]},"credible_cost_scope_and_range":{"status":"YES","reason":"The bands are scoped by phase and anchored to official survey-research and software-development wages, with explicit assumptions for overhead, participant operations, and existing infrastructure. Procurement and site-specific costs remain uncertain.","source_ids":["S8"]}},"next_evidence_step":"After study-specific IRB and community review, preregister a three-arm trial in approximately 120 returning participants who have an authorized prior complete roster: (A) fresh full-roster re-enumeration, (B) simple proactive dependent interviewing that shows the prior roster and asks for confirmation or changes, and (C) the proposed versioned delta protocol with explicit unchanged/changed/unknown/declined/not-reviewed states, reconstruction confirmation, and participant-controlled fallback. Randomize arm and interviewer; send a randomly selected audit subset through a second, independent full-roster elicitation with counterbalanced order and a different interviewer. Compare total participant minutes and repeated sensitive disclosures, union-of-elicitation omissions, false persistence and false termination, consequential tie-class disagreement, discomfort/refusal, unauthorized exposure, subgroup gaps, fallback frequency, and total audit-plus-maintenance effort. Falsify the incremental claim if arm C fails to reduce median participant time by at least 20% versus A after allocating audit and fallback effort; exceeds A by more than 5 percentage points in consequential tie disagreement; fails to outperform B on false continuity or explicit missingness; increases discomfort or refusal by at least 10 percentage points; creates any unauthorized cross-participant disclosure; produces a subgroup error gap above 10 percentage points; or requires fallback for more than 30% of interviews. Treat the offline 80-record reconstruction test only as a software preflight, not evidence of participant benefit or safety.","blocking_evidence":["No direct estimate of how much time or distress complete relationship-roster re-enumeration creates in the intended community-fieldwork population.","No prospective comparison of the proposed protocol against both full re-enumeration and ordinary longitudinal name checking.","No evidence yet on anchoring, prompted acquiescence, participant comfort, interviewer pressure, or actual repeated-disclosure reduction.","No study-specific IRB decision, consent-scope determination, third-party privacy assessment, safeguarding plan, or community authorization.","No empirical error budget for consequential tie classes, rare changes, subgroup disparities, audit rate, snapshot expiry, or fallback thresholds.","No measured net resource saving after software maintenance, audit interviews, periodic refresh, and fallback are included.","The independent full-roster audit is itself an imperfect reference because repeated elicitation and order can affect recall."],"research_disposition":"PARTNERED_RESEARCH_PROGRAM","world_novelty_boundary":"The search establishes extensive adjacent and close prior art, including generic dependent interviewing, NSHAP roster matching, UCNets prior-name checking, and participant-facing roster software. It does not measure world novelty, patentability, freedom to operate, market size, or realized impact, and it cannot establish that the integrated governance package has never been implemented elsewhere.","arm":"COMPLETE_PROPOSAL_PORTFOLIO","candidate_version":0,"controller_recommendation":{"action":"STOP_EMPIRICAL_RESEARCH_NEEDED","repairable":false,"material_progress_observed":true,"progress_targets":["Obtain study-specific IRB and data-governance approval covering participant-specific prior rosters, third-party information, consent scope, fallback, and breach response.","Implement and verify snapshot checksums, provenance, explicit silence states, reconstruction confirmation, access controls, and rollback using authorized deidentified records.","Preregister the three-arm field protocol, audit sampling, noninferiority margins, subgroup analyses, stopping rules, and counterbalanced-order analysis.","Demonstrate at least 20% median participant-time reduction versus full re-enumeration after allocating audit and fallback effort.","Show consequential tie disagreement no more than 5 percentage points worse than full re-enumeration and lower false continuity than simple dependent interviewing.","Show no unauthorized disclosure, no material increase in discomfort or refusal, no subgroup error gap above 10 percentage points, and fallback at or below 30%.","Produce a measured 2026-USD cost model including instrument maintenance, secure snapshots, audits, full refreshes, and fallback interviews."],"reason":"Bounded web research answered prevalence, adopter, prior-art, general feasibility, authority, and labor-cost questions. The core method is already established practice, while the remaining incremental claim concerns participant behavior, anchoring, comfort, comparative measurement error, subgroup safety, and net workflow savings. Those questions require IRB-approved fieldwork or live testing and cannot be resolved by further web search; under the controller rule this requires STOP_EMPIRICAL_RESEARCH_NEEDED, with repairable set to false."},"proposal_index":3}