Methodology

What we track

Single source of truth

Independence (0–100) and LE capability (0–40) are the universal cross-jurisdiction rubrics, computed once at law.ungovr.org/oversight/methodology and applied to every body in the upstream oversight-laws dataset. This site reads those scores; it does not recompute them. Statute text and citations live in the upstream dataset.

Sheriff oversight depth (0–30)

The depth score is a sheriff-specific overlay surfaced only here. The two universal scores are calibrated for bodies where the executive appoints the chief, but a California sheriff is a constitutional officer, independently elected, and protected by the Public Safety Officers Procedural Bill of Rights Act (Cal. Gov. §§ 3300-3313). Several oversight realities specific to elected-sheriff structures aren't captured by the universal rubric. The depth score adds six dimensions:

Structural pairing (0–6)
Whether the body is a contractor-only audit, a single director / IG, or a paired commission + IG with distinct budget lines (the LA / Marin pattern). Paired structures with separate budgets are most durable; a contractor relationship has the least continuity.
POBRA carve-out (0–5)
Whether the body's establishing ordinance includes explicit carve-out language against POBRA / Cal. Gov. § 3304. Whether a civilian body needs such a carve-out to exercise binding discipline over a sworn officer has been litigated and is unsettled; this dimension records whether the ordinance contains carve-out language and whether it has been tested, not the answer to that question. Carve-outs sustained on judicial review score higher than untested ones.
Statutory backing (0–5)
Whether the body's powers anchor in Cal. Gov. § 25303.7 (AB 1185: civilian oversight subpoena power, critical-incident scene attendance, records access). Statute-anchored powers are more durable than ordinance-only ones. Bodies invoking § 25303.7 and exercising the powers score higher than those that invoke without exercising, or that match the powers by ordinance alone.
Complaint intake architecture (0–5)
How a county resident actually reaches the body. Counts intake channels (online portal, phone, in-person, mail, mobile app), whether anonymous complaints are accepted, and how many languages are supported. Contractor-only models that audit but don't take public intake score 0.
Subject coverage (0–4)
What sheriff functions the body actually monitors. California sheriffs operate patrol, jails, courthouse security, and civil enforcement (process serving, evictions). Bodies that audit all four score highest; jails-only and patrol-only score 2 because both are major sheriff functions but coverage is incomplete.
Data & outcome transparency (0–5)
Publication cadence and breadth. Quarterly reports with complaint-disposition statistics and live dashboards score full points. Annual reports without disposition data score middle. Irregular cadence with no public output scores 0.

Bands: strong (24+), moderate (18–23), limited (12–17), weak (6–11), nominal (0–5). The depth score is currently piloted on the nine California county sheriff oversight bodies (rubric v0.1) and will expand to other states after the rubric is validated against the pilot data.

Reports

Per-report extraction (findings, recommendations, response status) reuses the pipeline built for civilgrandjury.org. Report ingestion is rolling out; expect the corpus to grow over coming months.

Data freshness

The site rebuilds nightly. Last build: 2026-08-12 09:00 UTC.

State coverage

Coverage expands one state at a time. A state is added once at least one qualifying county-level civilian oversight body is identified and seeded. States without any qualifying bodies (such as Texas, where no Texas county has yet established an independent civilian body with operational scope over its sheriff) are documented as "no-coverage" states with their structural context, rather than omitted entirely. State-level officer- certification boards and jail-standards commissions are tracked separately on law.ungovr.org/oversight/law-enforcement.

Tone

sheriffoversight.org is a neutral reference work. We describe what statutes do and what bodies exist; we do not advocate for or against any particular arrangement. Where a normative claim is needed, it is sourced and attributed.

Budget figures

Every sheriff budget on this site is an operating appropriation for the sheriff's office and the county jail, taken from a document the county itself published, and recorded with the page it came from and the text it was read out of. Capital projects, debt service, reserves, fire and EMS, separately budgeted 911 centres, and money the county is reimbursed for are excluded, so the figure is what the county spends to run the office rather than everything that passes through it.

Counties do not share one accounting model, so before a figure is published it is checked against several independent expectations: the payroll the Census reports for that office, spending per sworn officer, spending per resident, the change from the prior year, the same measure in neighbouring counties, whether the document names the sheriff at all, how old the fiscal year is, and whether the document is one the county adopted rather than one it proposed.

A figure that clears every check is published plainly. A figure that clears most of them is published and marked P for provisional, with the specific reasons shown alongside it. Provisional does not mean we doubt the number: the two most common reasons are that the county has not published a budget in several years, and that the only document it published is a proposal rather than an adopted budget. Both are real, sourced figures. Marking them keeps a reader from reading an older or proposed figure as this year's adopted one. A figure that fails a check we consider disqualifying, or several at once, is not published at all.

Where a county has its own police department, the sheriff runs the jail and the courts while patrol sits elsewhere, so that sheriff's budget covers less ground than a full-service sheriff's. Those counties are labelled, because per-resident spending is not comparable between the two arrangements.

Report a missing body

If you know of an oversight body not listed here, tell us.

Data sources