Reconciliation

Reconciliation is the process of identifying and merging duplicate records describing the same underlying event across sources, a required step before a corpus's scale or a base rate's sample size can be reported honestly. Reconciliation has to run continuously as new data enters a growing corpus, not once.

Reconciliation is the process of identifying records across different sources, or across different points in time, that describe the same underlying real-world event, a single docket entry, a single lien, a single enforcement action, and merging them into one accurately counted event rather than allowing them to persist as separate entries that inflate a corpus's reported scale. Reconciliation is harder than an exact-match check on raw text, because the same event frequently appears with minor formatting differences, a docket number punctuated differently, a party name abbreviated in one system and spelled out in another, that an exact-match process will treat as two different events. Every claim built on top of a corpus inherits whatever reconciliation gaps the corpus contains, and the quieter cost is more consequential than an inflated headline total: a base rate computed from a sample that silently double-counts a share of its underlying events looks better-supported than the true, reconciled sample actually justifies. Reconciliation has to be run as a standing, ongoing process, applied every time new data enters the corpus through a new source or a new harvesting cycle, rather than as a one-time cleanup performed once and assumed to hold indefinitely as the corpus continues to grow.