Lineage or Panel Correspondence Matrix¶
Mapping matrix — instantiates Selection–Transmission Change Attribution
Maps which units in the first state correspond to which in the second — continuing, entered, exited, split, or merged — so selection and transmission can be told apart at all.
Before you can ask whether a unit changed or was replaced, you have to say which unit is which. Lineage or Panel Correspondence Matrix is the identity bookkeeping the whole archetype stands on: a matrix whose rows are the starting-state units and whose columns are the comparison-state units, with each cell typed as continuing, entered, exited, split, or merged. Its defining job is establishing unit identity across states — nothing more. It records no values, computes no weights, and produces no selection or transmission term. But without it, a unit changing internally and a unit being swapped out for another are indistinguishable, and every downstream decomposition is built on sand.
Example¶
A conglomerate reorganizes between two annual reports, and analysts want to attribute a change in average operating margin. First they must decide which "business unit" in year one is which in year two. The correspondence matrix records it explicitly: the Payments unit continued unchanged (a clean diagonal cell); Logistics and Warehousing merged into a single Fulfillment unit (two rows mapping to one column); the Consumer division split into Retail and DTC (one row mapping to two columns); the Hardware unit was divested — an exit with no column; and a freshly acquired Analytics unit entered with no row. Only against this map can anyone say whether Fulfillment's margin represents "two units becoming one" or a genuine internal change, and whether Hardware's disappearance should count as selection or be excluded outright.
How it works¶
Fix the unit rule and the population scope — what counts as a unit and where the boundary sits — then enumerate every unit in each of the two dated states and link them with typed edges: continue, enter, exit, split, merge, or descendant. Splits and merges carry an explicit apportionment note (how a parent's identity is divided among children, or several parents pooled into one) so downstream mechanisms inherit an unambiguous rule. The transition window is pinned by the two state dates. The craft is entirely in the matching: it is a record-linkage problem, resolving which records across two snapshots refer to the same underlying entity.[n1] The output is a reviewable map, not a number.
Tuning parameters¶
- Identity-rule strictness — exact-ID matching versus fuzzy or descendant matching. Loose matching recovers more continuity across renames and reorganizations but risks fabricating links that were not real.
- Split/merge apportionment — how a divided or pooled unit's identity is shared. Fractional apportionment preserves accounting exactness; all-or-nothing assignment is simpler but drops information.
- Scope boundary — a fixed roster fixed at baseline versus an open population that admits new units. Fixing the roster clarifies transmission; opening it captures entry effects.
- Window placement — where the two states are dated and how far apart. Wider windows capture more change but multiply the ambiguous split/merge/rename events to resolve.
When it helps, and when it misleads¶
Its strength is that it makes the selection-versus-transmission question askable at all: with a defensible map, continuing units, entrants, and exiters are cleanly separable, and the decomposition has firm ground.
Its failure mode is that matching errors propagate silently downstream. A wrong link mislabels a replacement as an internal change (inflating transmission) or an internal change as a swap (inflating selection), and over-eager fuzzy matching manufactures spurious continuity where a unit was really replaced. Because the matrix produces no number, these errors hide until they surface as an unexplained residual much later. The guarding discipline is reviewability: keep the map auditable, flag ambiguous links as ambiguous rather than forcing a clean edge, and record the apportionment rule so another analyst can re-trace every match.
How it implements the components¶
population_unit_and_scope_definition— it names what a unit is and bounds the population being tracked across states.unit_correspondence_map— the matrix itself: the typed continue/enter/exit/split/merge links that constitute unit identity across states.state_pair_or_transition_window— it pins the two dated states the correspondence is drawn between.
It records no values or weights (unit_value_measure, unit_weight_measure) and computes no channels (selection_term_definition, transmission_term_definition) — those belong to the Price Equation Decomposition Table and its term calculators. Its nearest twin is the Entry/Exit Normalization Protocol: this matrix fixes unit identity across states (who maps to whom), while the protocol fixes the weights of entrants and exiters so the mean stays comparable — the separating component is the correspondence map it owns versus unit_weight_measure.
Related¶
- Instantiates: Selection–Transmission Change Attribution — it operationalizes unit continuity, the precondition for every term.
- Sibling mechanisms: Price Equation Decomposition Table · Covariance Selection-Term Calculation · Within-Unit Change Assay · Composition-vs-Transformation Dashboard · Entry/Exit Normalization Protocol · Decomposition Residual Reconciliation Workflow · Selection–Transmission Sensitivity Analysis
Editorial Notes¶
Form Classification¶
Form family: Analysis, Modeling & Optimization
Rationale: Lineage or Panel Correspondence Matrix operates as a computation, comparison, model, or analytic representation used to infer, estimate, or choose because it maps which units in the first state correspond to which in the second — continuing, entered, exited, split, or merged — so selection and transmission can be told apart at all.
Independent corroboration: The frozen evidence defines Lineage or Panel Correspondence Matrix as 'Maps which units in the first state correspond to which in the second — continuing, entered, exited, split, or merged — so selection and transmission can be told apart at all', so its operative form is Analysis, Modeling & Optimization.
Nearest alternative: Representation, Specification & Plan — The correspondence map is durable, but creating it requires record linkage and inferential matching rather than mere externalization.
Review outcome: Independent reviewer agreement; medium confidence.
Origin Attribution¶
Primary origin: Statistics & Experimental Design
Origin pattern: Cross-disciplinary synthesis
Present-day reach: Multi-domain
Rationale: Panel and longitudinal statistical analysis requires identity correspondence across waves before within-unit and compositional change can be separated.
Related originating lineages:
- Accounting & Auditing — Continuity, entry, exit, split, and merger bookkeeping is also formative in longitudinal organizational accounts.
- Data Science & Analytics — Record linkage and entity resolution materially supply the computational matching machinery.
Review resolution: Both independent reviews assign primary provenance to statistics_experimental_design. The queued secondary differences (reported_ambiguity, alternate_origin_disagreement) are reconciled by retaining accounting_auditing, data_science only as formative or independently established lineage(s), not merely as application domains. origin_mode=cross_disciplinary_synthesis records the provenance relationship, while domain_reach=multi_domain separately records applicability breadth. confidence=medium preserves the more cautious assessment, and encyclopedia_synthesis=true records whether either reviewer identified a corpus-specific synthesis.
Attribution caveat: The matrix generalizes panel correspondence to lineages that may split or merge.
Encyclopedia synthesis: The exact catalogued form synthesizes established practice rather than reproducing a single standard historical label.
Review outcome: Reconciled after independent review; medium confidence.
Notes¶
[n1] Record linkage (entity resolution) is the problem of deciding which records across two data sources refer to the same real-world entity. A correspondence matrix is record linkage applied across two temporal snapshots, with the extra complication that entities can legitimately split and merge between them. ↩