Home Data report
US County Jail Data Report
What 17,935 live custody records from 79 county jails in 26 states actually show. Recomputed on every sync; figures below reflect 2026-08-26 08:16 UTC.
Most people are not held for a new offence
The single most common entry on a county jail roster is a warrant, appearing on 34% of records, followed by a hold placed by another agency on 15%. Together with probation and parole violations, administrative reasons — not a fresh arrest for a new crime — account for a large share of any given day's population. This is the finding that most surprises people reading a roster for the first time.
| Category | Records | Share | Distribution |
|---|---|---|---|
| Warrant | 6,155 | 34% | |
| Hold for another agency | 2,700 | 15% | |
| Assault or violence | 2,585 | 14% | |
| Drug offence | 2,248 | 13% | |
| Theft or property | 2,105 | 12% | |
| Probation or parole violation | 2,096 | 12% | |
| Serving a sentence | 1,065 | 6% | |
| Weapons | 1,003 | 6% | |
| Failure to appear | 617 | 3% | |
| Traffic offence | 574 | 3% | |
| Driving under the influence | 508 | 3% | |
| Sex offence | 451 | 3% | |
| Contempt of court | 229 | 1% |
Bookings drop by nearly half at the weekend
Across 15,853 records with a readable booking date, Tuesday is the busiest day and Sunday the quietest. Weekend bookings run at roughly half the weekday rate. The index column below sets the average day at 100.
| Day | Records | Share | Index |
|---|---|---|---|
| Sunday | 1,261 | 8% | 56 |
| Monday | 2,434 | 15% | 107 |
| Tuesday | 2,917 | 18% | 129 |
| Wednesday | 2,658 | 17% | 117 |
| Thursday | 2,680 | 17% | 118 |
| Friday | 2,497 | 16% | 110 |
| Saturday | 1,406 | 9% | 62 |
Bond is set on fewer than half of records
A bond amount appears on 46% of records (8,258 people). The median is $20,000, while the ninetieth percentile reaches $750,000. The remainder show no bond figure, which usually means bond has not been set yet, the charge is not bondable, or a hold makes the question moot. A further 6% of records list no charge text at all — a limitation of the source system rather than an absence of charges.
Typical roster size
The median published county roster holds 138 people; the largest in this dataset holds 1,295. County jail populations are small and fast-moving compared with state prisons, which is why a roster can look completely different from one day to the next.
Method and limitations
Figures are computed at build time from the same records shown on the county pages of this site, which are collected from official county booking sources. Charge categories are assigned by matching the published charge text against keyword patterns; a record matching several patterns is counted once in each category. Roughly 13,097 individual charge lines did not match any category and are excluded from the category table.
Three limitations matter when citing this. First, coverage is not a random sample of US counties — it reflects which counties publish machine-readable rosters, which skews toward smaller and mid-sized jurisdictions. Second, a roster is a snapshot of who is in custody, not a record of arrests over a period, so people held longer are over-represented relative to those who bond out quickly. Third, charge text is recorded at booking by the arresting agency and is not updated when prosecutors amend or drop charges.
Download the aggregates
The figures on this page are published as machine-readable files, regenerated on every sync: aggregates.csv and aggregates.json. Both include the coverage notes and caveats below, so the numbers travel with their context.
Using these figures
You are welcome to quote or reproduce any figure on this page in reporting, research or teaching, with attribution to County Inmate Search and a link to this page. If you need the underlying aggregates in a different cut — by state, by county size, or over time — write to us through the contact page and we will help.
An arrest is not a conviction. Every figure here describes allegations recorded at booking and the fact of custody. None of it describes guilt, and most people counted above have not been convicted of the charge shown.