Administrative Record Linkage¶
Method — instantiates Exhaustive Population Mapping
Joins existing registries and ledgers through a secure crosswalk to reveal units and cut the fieldwork the enumeration would otherwise need.
Administrative Record Linkage is the method that builds a population map out of paperwork that already exists. Its defining move is to treat the world's disconnected registries — tax rolls, utility hookups, licenses, benefit ledgers — as complementary frames, none complete alone, and to join them so that a unit visible in any one becomes visible to the enumeration. The point is not merely to combine data but to reveal units and reduce burden: every unit a records join surfaces is a unit no one has to walk a street to find. Because the join runs on identified records about real people or assets, it lives inside a secure, purpose-limited processing environment by necessity. It is not the adjudication of hard same-versus-different calls, and it is not a statistical estimate of the unseen; it is the constructive step that assembles many partial ledgers into one broader roster.
Example¶
A revenue agency needs a complete list of every rental property in a city, because an occupancy-tax obligation attaches to specific units and a sampled estimate cannot bill anyone. No single register holds them all. The property-tax roll lists parcels but not which are rented; the utility maintains hookup accounts by service address; the business-licensing system has short-term-rental permits; a landlord-registration database is voluntary and sparse. Linkage joins them on address and owner identity: a parcel that is owner-occupied in the tax roll but shows a tenant-name utility account and a rental permit is almost certainly a rental the landlord register missed.
The join's value is the units it surfaces without fieldwork. Roughly matching the four ledgers turns up several thousand candidate rentals that appeared on no single "rental list," sharply narrowing the far more expensive door-to-door verification to the genuinely ambiguous cases. Because the linkage handles owner names, tenant identities, and addresses, it runs in a controlled environment with access limited to the tax-assessment purpose — the crosswalk exists to reveal taxable units, not to build a general dossier on residents.
How it works¶
- Inventory the frames, assume none is complete. Each register is treated as a partial view; the method's leverage is that their omissions differ.
- Cross-reference on shared keys. Records are aligned on identity or location keys so a unit seen in any source is carried into the combined frame.
- Prefer union, then reconcile. The join's first job is coverage — pull in every candidate unit — leaving the hard duplicate calls to a downstream adjudicator.
- Contain the exposure. Because identified records are combined, the linkage runs under access controls, purpose limitation, and often key-only or secure-enclave matching.
Tuning parameters¶
- Source breadth — how many registers are pulled in. More sources raise coverage of hard-to-see units but multiply match ambiguity and privacy exposure.
- Match key strictness — exact-key joins versus fuzzy matching on names and addresses. Fuzzy joins catch more real units but generate more false links to sort out later.
- Union vs. intersection bias — whether a unit needs one source or several to be admitted. Union maximizes coverage (with more noise); intersection is cleaner but drops units seen only once.
- Containment level — open analytic join versus secure-enclave or hashed-key linkage. Tighter containment protects subjects at some cost in matching power.
When it helps, and when it misleads¶
Linkage shines when rich administrative data already blankets most of the population and fieldwork is expensive — it can reveal the great majority of units at a desk, reserving physical sweeps for the residue. It is the cheapest route to broad coverage and the natural first pass before any boots hit the ground.
Its dangers are subtle. A false match silently fuses two units into one, and a missed match splits one unit into two — the probabilistic-matching literature frames this as the trade-off between false links and false non-links, and both corrupt the count in opposite directions.[n1] Linkage also inherits the registers' shared blind spots: units absent from every administrative source (the unbanked, the off-grid, the informal) stay invisible no matter how many ledgers you join, which can breed false confidence that "the records cover everyone." And combining identified data is a standing privacy hazard that convenience erodes. The guarding discipline is to treat linkage as a coverage first pass whose residue still needs field discovery, and to keep matching thresholds and access controls explicit and reviewed.
How it implements the components¶
Administrative Record Linkage fills the reveal-from-records slice — broad coverage assembled from existing data:
enumeration_frame_inventory— gathers registries, ledgers, and permits as complementary source frames, none assumed complete on its own.independent_source_crosswalk— aligns those sources on shared keys so a unit visible in any of them enters the combined frame, and units unique to each are surfaced.privacy_and_exposure_guardrail— the join of identified records runs under access control, purpose limitation, and secure matching.
It reveals candidate units but does not adjudicate the ambiguous same-versus-different matches or log the merge decisions — that unit_identity_and_deduplication_key and exception_and_revision_path work is Duplicate Resolution Queue's — and it grows the roster rather than measuring what the roster still misses, so turning source overlap into a missingness_and_overlap_diagnostic is the job of its fellow method-type sibling Capture-Recapture Check.
Related¶
- Instantiates: Exhaustive Population Mapping — linkage is the desk-based frame-building that reveals most units before any field sweep.
- Sibling mechanisms: Census Protocol · Enumeration Area Map · Master Unit Index · Capture-Recapture Check · Door-to-Door or Field Sweep · Duplicate Resolution Queue · Coverage Gap Heatmap · Late-Unit Inclusion Window · Enumeration Quality Backcheck
Editorial Notes¶
Form Classification¶
Form family: Analysis, Modeling & Optimization
Rationale: The mechanism joins existing registries and ledgers through a secure crosswalk to reveal units and cut the fieldwork the enumeration would otherwise need, so its operative form is offline analysis, modeling, or optimization.
Independent corroboration: The frozen evidence defines Administrative Record Linkage as 'Joins existing registries and ledgers through a secure crosswalk to reveal units and cut the fieldwork the enumeration would otherwise need', so its operative form is Analysis, Modeling & Optimization.
Nearest alternative: Record, Log & Register — It analytically joins existing registries through a crosswalk to construct a broader frame rather than preserving a new event history.
Review outcome: Independent reviewer agreement; medium confidence.
Origin Attribution¶
Primary origin: Statistics & Experimental Design
Origin pattern: Convergent development
Present-day reach: Multi-domain
Rationale: Statistical record linkage, classically Fellegi-Sunter, joins imperfectly matching administrative records while balancing false links against false non-links.
Related originating lineages:
- Data Science & Analytics — Modern entity resolution, scalable matching pipelines, and linkage software operationalize the classical statistical decision model across large heterogeneous data sets.
- Law & Governance — Privacy, purpose limitation, data minimization, access control, and contestability constrain identified linkage about real people.
- Library & Information Science — Entity resolution, authority control, metadata crosswalks, and identifier management contribute methods for reconciling records across heterogeneous registries.
- Public Administration & Policy — Censuses and administrative-data systems use linked registers to construct population frames and reduce field-enumeration burden.
Review resolution: Probabilistic record linkage was formally established as a statistical decision model by Fellegi and Sunter, including controlled false-link and false-nonlink errors. Data science, information science, public administration, and privacy law are material modern lineages, but statistics supplies the primary origin.
Review outcome: Researched adjudication after independent review; high confidence.
Sources consulted:
- U.S. Census Bureau — Overview of Record Linkage and Current Research Directions
- Fellegi and Sunter — A Theory for Record Linkage
Notes¶
Linkage feeds the Duplicate Resolution Queue: every fuzzy match it cannot resolve confidently becomes a candidate pair the queue adjudicates. Keeping revealing units (this method) separate from deciding identity (the queue) is what lets a linkage cast a wide, noisy net without being blamed for the merges — the noise is handled downstream, on purpose.
[n1] The Fellegi–Sunter model is the classical statistical framework for probabilistic record linkage, formalizing how to weigh partial agreement across fields and set thresholds that trade off false links (fusing distinct units) against false non-links (splitting one unit). It makes explicit that linkage error runs in both directions, and that the thresholds are a choice, not a given. ↩