# Public methodology

## Selection and units

Texas FY2025 (September 1, 2024–August 31, 2025), five counties, statutory county-court Criminal reports, category **Drug Possession–Marijuana**. OCA's historical official definition includes misdemeanor HSC §481.120 delivery, §481.121 possession, §481.134(f) drug-free-zone offenses. It is not a possession-only or all-cannabis category.

The official historical [instructions](https://www.txcourts.gov/media/1438897/ccl-report-instructions-9_13_17-sb-1326-final.pdf), p.3 and pp.9–10, define cases by defendant/information-or-complaint, most-serious-charge assignment and original-charge disposition reporting. Lesser-offense convictions can therefore remain in the category. The deferred-adjudication placement row has broader instructions including probation/community supervision/pretrial diversion. Original source labels are preserved, not simplified into narrower legal outcomes.

These historical official instructions and retained report columns support the stated qualifications. The exact last governing revision for every FY2025 submission has not been certified by OCA; office-by-office compliance is not independently proven.

## Table calculation

New filings, total dispositions and dismissals are the corresponding source rows. Conviction-row sum adds three distinct leaves: Guilty Plea, By The Court and By The Jury. It is a reporting-row sum, not verified final-offense convictions. Disposition leaves sum to the total. The total appears in two report sections and is never added twice.

Dispositions include older filings and revocation activity. Annual outcome counts divided by new filings are not case-outcome probabilities. No arrests, population denominator, unique people, individual linkage, policy effect or statewide ranking is inferred.

## Private source verification versus public arithmetic

Private verification re-read all 176 original retained sources, checked 161 TSV exports, 132 annual/monthly comparisons, 126 standard-report comparisons and 78 disposition-component checks. All five pilot counties' pending-stock diagnostics reconcile using recorded source adjustments; all 60 required reporter-months are marked received/in balance. That administrative status does not certify underlying case accuracy. The five-county numerical claims have their own query/standard-report witnesses.

This public package includes annual aggregate cells, not those raw files or the complete monthly reconciliation. Its portable code independently verifies the table arithmetic from the provided aggregates. Source hashes and dates identify the original reports; they do not make mutable current portal responses identical to a historical capture.

## Statewide limitation

A separate statewide pending-stock diagnostic retains a +53 residual. Its cause remains unresolved. Fixed annual-publication and live-query vintages also differ, separately from the stock residual. No source value was adjusted to force reconciliation. The five-county findings do not assert statewide completeness or reconciliation, and the counties are not a representative sample. Source-version references for this limitation are identified separately from county dataset dependencies; no statewide data extract is included in this package.

## Government records and attribution

OCA is the original authority. TCL supplies selection, compilation, reconciliation and presentation. No OCA endorsement or blanket rights grant over government material is claimed. See `REUSE-NOTES.md` for the approved reuse boundaries.