Subjects and lessons to explore.

Law

Criminal Justice & Criminology · Corrections

Jails Versus Prisons

12 min read 5 practice questions
10 key termsSourcesReport an error
Want it in plain words first? Jump to Eli explains — the same idea, no jargon.
On this page 9 sections
  1. In 30 seconds
  2. Why this matters
  3. The college version
  4. Eli explains
  5. Worked example
  6. Key takeaway
  7. Quick check
  8. Study tools
  9. Sources & references

In 30 seconds

Jails and prisons are different institutions, not two words for one place. Jails are operated by counties and cities, hold people awaiting court action and those serving short sentences, and record millions of admissions a year against a much smaller daily population. Prisons are operated by states or the federal government and hold sentenced people serving longer terms, so their populations move slowly. Most errors students make about American corrections come from treating the two as interchangeable, or from adding their published counts together.

Why this matters

Corrections statistics are among the most misused numbers in public argument, and the jail-prison distinction is where the misuse usually begins. A county budget, a sheriff's election, and a state sentencing bill each move a different part of the system, and you cannot follow any of them without knowing which institution is affected. The distinction also shapes the law: the operating agency, the funding source, and the length of stay differ sharply, while a common body of federal constitutional doctrine reaches both. Students going into criminal justice, public administration, social work, or health care will meet jails first, because jails are where most Americans who are locked up in a given year actually go. This lesson is educational material about United States practice, not legal advice.

The college version

Two institutions, two levels of government

The Bureau of Justice Statistics defines a as a confinement facility generally operated under the authority of a sheriff, police chief, or county or city administrator, holding people before and after arraignment, usually beyond 72 hours, along with people sentenced to a year or less. A , in BJS's definition, is a long-term facility run by a state or the federal government that typically holds people sentenced to more than a year, with the exact line varying by state. The difference is not architecture or size; it is who owns the institution and who pays for it. California states the local arrangement in statute: Penal Code section 4000 provides that the common jails in the several counties are kept by the sheriffs of those counties, and section 4015 requires the sheriff to receive all persons committed by competent authority while directing the county board of supervisors to supply food, clothing, and bedding at county expense. The federal arrangement is equally explicit and entirely different: 18 U.S.C. 4042 puts the Bureau of Prisons in charge of all federal penal institutions, and 18 U.S.C. 3621(b) gives the Bureau, not the sentencing judge, the power to designate where a federal prisoner serves. Six states run integrated systems combining prisons and jails, and BJS counts everyone in them as prisoners; the other 45 states held 2,848 jail jurisdictions in the 2023 survey frame. The two systems also trade custody for money: at midyear 2023 local jails held 90,500 people for federal, state, and tribal authorities.

Who is inside, and for how long

At midyear 2023, local jails held 664,200 people in custody. Seventy percent of them, 467,600 people, were unconvicted, awaiting court action or held for other reasons; the remaining 30 percent, 196,600, were convicted, serving a sentence or awaiting sentencing. Three quarters, 500,300, were held on felony offenses. The jail incarceration rate was 198 per 100,000 U.S. residents, down from 231 a decade earlier. At yearend 2023, correctional authorities had jurisdiction over 1,254,224 people in state or federal prisons: 1,097,597 under state authority and 156,627 under the Federal Bureau of Prisons. Ninety-six percent of them, 1,210,300 people, were sentenced to more than one year. The imprisonment rate was 360 sentenced prisoners per 100,000 residents, down 25 percent from 479 in 2013. Both populations are smaller than a decade ago. The prison population fell 20.5 percent from 1,576,950 in 2013, but rose about 2 percent in each of 2022 and 2023, the first back-to-back increases in state jurisdiction counts in more than a decade. Length of stay is where the two institutions separate most sharply: people admitted to jail between July 2022 and June 2023 spent an average of 32 days in custody before release, seven days longer than eight years earlier. Sentenced prison terms are measured in years. Both populations are demographically skewed in ways BJS measures directly. At midyear 2023 the jail incarceration rate for Black U.S. residents was 552 per 100,000, 3.6 times the rate for white residents at 155, with American Indian and Alaska Native residents second at 425 and Hispanic residents at 143. At yearend 2023, 33 percent of sentenced prisoners were Black, 31 percent white, and 23 percent Hispanic. Those are measurements, not explanations.

Admissions versus population: the number everyone misreads

A jail's daily population and its annual admissions are different quantities, and the gap between them is enormous. From July 2022 to June 2023, jails nationwide recorded 7.6 million admissions while carrying an of 664,800; a decade earlier, admissions totaled 11.7 million. Prisons behaved nothing like that: states and the FBOP recorded 472,300 admissions of sentenced prisoners in 2023 against 453,200 releases. Roughly sixteen people enter a jail for every one who enters prison, even though prisons hold nearly twice as many people on any given day. The reason is length of stay. A population is a stock; admissions are a flow; the average time each person stays converts one into the other. Writers who say jails hold millions of people have reported a flow as a stock. Check which a source is using, and check the reference date, because BJS jail figures are keyed to midyear and prison figures to yearend.

Why the two counts cannot be added

The most common technical error in this field is arithmetic. Adding the midyear 2023 jail count to the yearend 2023 prison count gives 1,918,424, but BJS's own combined figure for people incarcerated in prisons or local jails at yearend 2023 is 1,852,900, about 65,000 lower. The gap is not a rounding artifact: BJS reports that 65,600 people under state or federal prison jurisdiction were physically held in local jails in 2023, so they appear in the jail and again in the prison . Two smaller mismatches compound it: the jail series excludes the six integrated-system states whose populations BJS counts as prisoners, and the two series use different reference dates. The underlying distinction is worth holding onto: a jurisdiction count records the authority legally responsible for a person wherever that person is held, while a custody count records who is physically holding the person. Use a source's own published total rather than assembling one.

Security levels and classification

Prisons sort people by security level over a long horizon. The Bureau of Prisons operates 118 institutions at five levels: minimum, low, medium, high, and administrative. Assignment turns on features such as external patrols, towers, security barriers and detection devices, the type of housing, and the staff-to-inmate ratio, and administrative facilities carry special missions including pretrial detention and medical care. Under 18 U.S.C. 3621(b) the Bureau designates the place of imprisonment considering bed availability, the person's security designation, programmatic and health needs, the sentencing court's recommendations, and the nature of the offense. Jails classify too, but on a clock measured in hours, and for a population whose charges and risks are often not yet known. One federal requirement is uniform across both settings: 28 C.F.R. 115.41 requires an objective screening instrument at intake, ordinarily within 72 hours of arrival, assessing each person's risk of being sexually abused or sexually abusive, and 28 C.F.R. 115.42 requires those results to inform housing and program assignments. Capacity is also not the national constraint it is assumed to be: at midyear 2023 there were 915,800 jail beds, 73 percent occupied, down from 84 percent in 2013, though 12 percent of jail jurisdictions were above rated capacity.

Private facilities and women in custody

Private operation is real but is a minority arrangement, and it is concentrated. In 2023, 154,200 people under state or federal jurisdiction, 12 percent of all prisoners, were held in privately operated prison facilities or in local jails; the private-prison share alone was 88,600 people, about 7 percent of prisoners, down almost 3 percent from 2022. Four states held more than a quarter of their people in private facilities, while the FBOP reported holding none in secure private prisons that year. Women follow a different pattern in each institution. At midyear 2023, 95,100 women were in jail, 14 percent of the jail population but 23 percent of admissions, with an average stay of 19 days against 36 for men, so women pass through jails more often and stay less long. At yearend 2023, 91,100 women were in prison, 7 percent of that population, and their numbers grew faster than men's: up 3.8 percent from 2022 against 1.8 percent for men, with female counts rising in 41 states and male counts in 38.

Mental illness and substance use, concentrated in jails

BJS's inmate surveys found indicators of mental health problems more common in jails than in prisons and far more common in both than in the general population. In the National Inmate Survey conducted from February 2011 to May 2012, 26 percent of jail inmates and 14 percent of prisoners met the threshold for serious psychological distress in the prior 30 days, against 5 percent of the standardized general population; 44 percent of jail inmates and 37 percent of prisoners had been told by a mental health professional that they had a mental disorder. Within jails, 32 percent of women met the distress threshold against 26 percent of men. In the 2007 and 2008-09 surveys, 63 percent of sentenced jail inmates and 58 percent of state prisoners met DSM-IV criteria for drug dependence or abuse, against roughly 5 percent of adults generally. Two cautions belong with these figures: they date from the survey periods named rather than the present, and they describe prevalence, not cause. BJS's own list of jail functions includes holding people with mental illness pending transfer to an appropriate mental health facility, which describes what jails are asked to absorb rather than explaining why.

The constitutional floor on conditions

Two cases carry most of the doctrine. Estelle v. Gamble, 429 U.S. 97 (1976), held that by prison personnel to a prisoner's serious illness or injury is cruel and unusual punishment under the Eighth Amendment, reasoning that a person in custody must rely on the state for care. Gamble himself lost: staff saw him on 17 occasions in three months, and the failure to order an X-ray was at most malpractice, a matter for state court. Farmer v. Brennan, 511 U.S. 825 (1994), supplied the test. Officials must provide humane conditions, including adequate food, clothing, shelter, and medical care, and must protect prisoners from violence by other prisoners. Liability requires both an objectively serious deprivation and a subjective state of mind: the official must know that inmates face a substantial risk of serious harm and disregard it by failing to take reasonable measures. That is subjective recklessness, more than negligence and less than intent to harm, though a factfinder may infer knowledge from how obvious a risk was, and an official who responds reasonably is not liable even if the harm is not averted. Because the Eighth Amendment governs punishment, conditions claims by people not yet convicted run instead through due process. Bell v. Wolfish, 441 U.S. 520 (1979), which arose from a short-term federal detention facility, asks whether a condition amounts to punishment: absent an expressed intent to punish, a condition reasonably related to a legitimate nonpunitive objective is not punishment. That matters most in jails, where most people are unconvicted.

The PLRA and Brown v. Plata

Congress narrowed prisoner litigation in 1996. Under 42 U.S.C. 1997e, a person must exhaust available administrative remedies before filing about prison conditions, cannot recover for mental or emotional injury without a prior showing of physical injury or the commission of a sexual act, and faces caps on attorney's fees; subsection (h) defines prisoner to include anyone detained who is merely accused, so these limits reach jails too. Under 18 U.S.C. 3626, prospective relief must be narrowly drawn and the least intrusive means, and only a three-judge court may cap a prison population, after a less intrusive order has failed and on clear and convincing evidence that crowding is the primary cause of the violation. BJS reported that civil rights petitions by state and federal prisoners fell from 41,679 in 1995 to 25,504 in 2000, and the filing rate from 37 to 19 per 1,000 prisoners, while habeas filings rose; BJS characterized the PLRA as appearing to have caused the decrease. The same statute produced the field's largest remedy. Brown v. Plata, 563 U.S. 493 (2011), consolidated a 1990 case about mental health care and a 2001 case about medical care in California prisons. A three-judge court ordered the state to reduce its prison population to 137.5 percent of design capacity within two years; at trial the facilities held some 156,000 people, nearly double design capacity, and the required reduction was estimated as high as 46,000. The Supreme Court affirmed 5-4, Justice Kennedy writing, over dissents by Justice Scalia and Justice Alito.

PREA and its audits

The Prison Rape Elimination Act of 2003 set a zero-tolerance standard for sexual abuse in custody and directed the Attorney General to adopt national standards for detecting, preventing, reducing, and punishing it. The resulting standards, adopted in 2012, sit at 28 C.F.R. Part 115 in four subparts covering adult prisons and jails, lockups, community confinement facilities, and juvenile facilities. Each facility must be audited at least once every three years, with at least a third of each facility type audited each year, by an auditor external to the agency and certified by the Department of Justice, and the agency bears the burden of demonstrating compliance. Enforcement runs through money rather than direct command: 34 U.S.C. 30307(e) reduces by 5 percent a state's covered Department of Justice grant funds for prison purposes unless the state's chief executive certifies full compliance or gives a qualifying assurance. That structure illustrates the whole lesson. Because jails are locally run, federal corrections policy reaches them through standards, grant conditions, and litigation rather than through a national corrections agency, which exists only for federal prisoners.

Eli, the EliExplains learning guide

Eli explains

The same idea, in plain words

Explain it like I’m 10

Picture two buildings on the same street. The first one is run by the county. People come in all day and all night, mostly right after an arrest, and most of them leave within a month, because a judge either releases them or their short sentence ends. The county pays for their food and their beds. The second building is run by the state or by the national government. People arrive only after a court has sentenced them to more than a year, and they stay for years. Almost nobody arrives on a Tuesday night in that second building. The first building is a jail. The second is a prison. They are not two names for the same thing, and they are not run by the same people.

Picture it like this

A jail is an emergency room and a prison is a long-stay hospital ward. The emergency room sees far more people in a year, keeps almost none of them overnight for long, and never knows in the morning who will be there by evening. The long-stay ward has fewer beds turning over, knows most of its patients by name, and plans in months.

Where the picture stops working

The analogy breaks in three places. An emergency room admits people who need it, while a jail admits people the law sends, and most of them have not been convicted of anything. A hospital's job is treatment, while a jail's legal job is custody, even though jails end up holding many people with untreated illness. And no hospital is governed by the Eighth Amendment or audited under a federal anti-sexual-abuse standard, which both jails and prisons are.

Worked example

Use BJS's own jail figures to recover a number BJS also publishes. From July 2022 to June 2023, jails recorded 7.6 million admissions against an average daily population of 664,800. Multiply the daily population by 365 to get person-days of custody: 664,800 times 365 equals 242,652,000. Divide by admissions: 242,652,000 divided by 7,600,000 equals about 31.9 days. BJS reports an average stay of 32 days, so the arithmetic checks. The same relation run the other way explains the churn: 7,600,000 divided by 664,800 is about 11.4, meaning each jail bed turns over roughly eleven times a year. Now try the trap. Adding the midyear 2023 jail count of 664,200 to the yearend 2023 prison count of 1,254,224 gives 1,918,424, but BJS's own combined figure is 1,852,900. The difference of about 65,000 is almost exactly the 65,600 state and federal prisoners BJS reports were held in local jails, counted once in each series.

Key takeaway

Jails are local, fast, and mostly unconvicted; prisons are state or federal, slow, and sentenced. Before you use any incarceration number, ask which institution it describes, which date it is keyed to, and whether it counts people present or bookings made.

Quick check

3 questions here, of 5 in this lesson’s practice set. Answers stay hidden until you check.

Question 1 of 3foundational

According to the Bureau of Justice Statistics, which description matches a jail rather than a prison?

Choose an answer, then check it.
Question 2 of 3intermediate

From July 2022 to June 2023, U.S. jails recorded 7.6 million admissions against an average daily population of 664,800. What does the comparison of these two figures most directly show?

Choose an answer, then check it.
Question 3 of 3advanced

A student adds the midyear 2023 jail count of 664,200 to the yearend 2023 prison count of 1,254,224 and reports 1,918,424 people incarcerated, but BJS's own combined figure is 1,852,900. What best explains the roughly 65,000 difference?

Choose an answer, then check it.
Practice all 5

Keep learning

Ready to build on this? Continue to the next lesson.

Practice this lesson
Study tools & related lessonsYou’ll learn to · Common mistakes · Easily confused · Key vocabulary · Related

You’ll learn to

  • Define jail and prison as the Bureau of Justice Statistics defines them, and identify which level of government operates and funds each.
  • Distinguish an admissions count from an average daily population, and explain why the two describe different things.
  • Apply the relationship among admissions, daily population, and length of stay to reproduce a published average length of stay.
  • Explain why published BJS jail and prison counts cannot be added, and identify the specific sources of double counting.
  • Summarize the constitutional and statutory framework governing conditions of confinement, including Estelle, Farmer, the PLRA, Brown v. Plata, and PREA.
  • Evaluate a claim about incarceration by checking which institution, which reference date, and which counting rule it uses.

Common mistakes

  • Adding the BJS jail population to the BJS prison population to get a national incarceration total.

    The series overlap. Prisoners held in local jails, 65,600 in 2023, appear in both, the reference dates differ, and six states' jail populations are counted as prisoners. Use BJS's own combined figure, 1,852,900 at yearend 2023.

  • Reporting the 7.6 million annual jail admissions as the number of people in jail.

    Admissions are a flow over twelve months; the average daily population was 664,800. The two differ by roughly a factor of eleven because the average stay is about 32 days.

  • Assuming that everyone held in a jail has been convicted.

    At midyear 2023, 70 percent of the jail population, 467,600 people, was unconvicted. That is why conditions claims in jails often run through due process rather than the Eighth Amendment.

  • Treating private prisons as the dominant form of incarceration.

    In 2023, 88,600 people, about 7 percent of prisoners, were in privately operated prison facilities, concentrated in a few states, and the FBOP reported none in secure private prisons.

  • Reading deliberate indifference as ordinary negligence, so that any serious harm in custody becomes a constitutional violation.

    Farmer v. Brennan requires that the official actually knew of a substantial risk of serious harm and disregarded it. An official who responds reasonably is not liable even if the harm occurs.

Easily confused

Jail vs. Prison

A jail is operated and funded by a county or city, holds people before and after arraignment along with those sentenced to about a year or less, and averaged a 32-day stay in 2022-23. A prison is operated and funded by a state or the federal government, holds people sentenced to more than a year, and its population turns over slowly. In 2023, jails recorded 7.6 million admissions against a daily population of 664,800; prisons recorded 472,300 sentenced admissions against a sentenced population of 1,210,300.

Admissions vs. Average daily population

Admissions count bookings over a period and one person can contribute several. Average daily population counts how many people are present on a typical day. Average length of stay converts between them: 664,800 people per day times 365 days, divided by 7.6 million admissions, reproduces the published 32-day average.

Jurisdiction count vs. Custody count

Jurisdiction records legal responsibility for a person wherever that person is held; custody records physical possession. The 65,600 state and federal prisoners held in local jails in 2023 are in a prison jurisdiction count and a jail custody count at the same time, which is why the two published series cannot be added.

Eighth Amendment conditions claim vs. Due process conditions claim

The Cruel and Unusual Punishments Clause governs punishment, so Estelle and Farmer supply the standard for people serving sentences: deliberate indifference to serious medical needs or to a substantial risk of serious harm. People not yet convicted proceed under due process, and Bell v. Wolfish asks whether a condition amounts to punishment or is reasonably related to a legitimate nonpunitive objective.

The PLRA as a restriction vs. The PLRA as a channel

42 U.S.C. 1997e narrows access by requiring exhaustion, barring emotional-injury claims without physical injury, and capping fees, and BJS measured civil rights petitions falling from 41,679 in 1995 to 25,504 in 2000. Yet 18 U.S.C. 3626(a)(3) is also the only route to a prison population cap, and it is the statute under which Brown v. Plata upheld the order to reduce California's prisons to 137.5 percent of design capacity.

Key vocabulary

Jail
A locally operated confinement facility, generally under the authority of a sheriff, police chief, or county or city administrator, that holds people before and after arraignment and people serving sentences of about a year or less.
Prison
A long-term confinement facility operated by a state government or the federal government that typically holds people sentenced to more than one year, with the cutoff varying by state.
Jurisdiction count
A tally of the people a correctional authority is legally responsible for, counted regardless of which facility is physically holding them.
Custody count
A tally of the people a facility is physically holding, regardless of which authority is legally responsible for them.
Average daily population
The sum of the people held each day over a year divided by the number of days, which measures a stock rather than a flow.
Admission
One booking of a person into a facility by formal legal document; the same person booked twice in a year contributes two of them.
Average length of stay
The mean time people spend in custody before release, the quantity that converts an annual flow of bookings into a daily headcount.
Deliberate indifference
The Eighth Amendment standard for official liability over conditions: knowing that people in custody face a substantial risk of serious harm and disregarding it, which is more than negligence and less than intent to harm.
Exhaustion requirement
The rule in 42 U.S.C. 1997e(a) that a confined person must first use whatever internal grievance process is available before bringing a federal action about conditions.
PREA audit
The compliance review required of every covered confinement facility at least once every three years, conducted by an outside auditor certified by the Department of Justice.

Sources & references

  1. Jail Inmates in 2023 - Statistical Tables (NCJ 309965) — Bureau of Justice Statistics, U.S. Department of Justice
  2. Prisoners in 2023 - Statistical Tables (NCJ 310197) — Bureau of Justice Statistics, U.S. Department of Justice
  3. Correctional Populations in the United States, 2023 - Statistical Tables (NCJ 310413) — Bureau of Justice Statistics, U.S. Department of Justice
  4. Indicators of Mental Health Problems Reported by Prisoners and Jail Inmates, 2011-12 (NCJ 250612) — Bureau of Justice Statistics, U.S. Department of Justice
  5. Drug Use, Dependence, and Abuse Among State Prisoners and Jail Inmates, 2007-2009 (NCJ 250546, revised August 10, 2020) — Bureau of Justice Statistics, U.S. Department of Justice
  6. Prisoner Petitions Filed in U.S. District Courts, 2000, with Trends 1980-2000 (NCJ 189430) — Bureau of Justice Statistics, U.S. Department of Justice
  7. Estelle v. Gamble, 429 U.S. 97 (1976) — Supreme Court of the United States (via govinfo, U.S. Reports)
  8. Farmer v. Brennan, 511 U.S. 825 (1994) — Supreme Court of the United States (via govinfo, U.S. Reports)
  9. Bell v. Wolfish, 441 U.S. 520 (1979) — Supreme Court of the United States (via govinfo, U.S. Reports)
  10. Brown v. Plata, 563 U.S. 493 (2011) — Supreme Court of the United States (via govinfo, U.S. Reports)
  11. 42 U.S.C. 1997e - Suits by prisoners — Office of the Law Revision Counsel, U.S. House of Representatives
  12. 18 U.S.C. 3626 - Appropriate remedies with respect to prison conditions — Office of the Law Revision Counsel, U.S. House of Representatives
  13. 34 U.S.C. 30302 - Purposes (Prison Rape Elimination Act of 2003) — Office of the Law Revision Counsel, U.S. House of Representatives
  14. 34 U.S.C. 30307 - Adoption and effect of national standards — Office of the Law Revision Counsel, U.S. House of Representatives
  15. 28 CFR Part 115 - Prison Rape Elimination Act National Standards — Office of the Federal Register / Government Publishing Office (eCFR)
  16. 18 U.S.C. 3621 - Imprisonment of a convicted person — Office of the Law Revision Counsel, U.S. House of Representatives
  17. 18 U.S.C. 4042 - Duties of Bureau of Prisons — Office of the Law Revision Counsel, U.S. House of Representatives
  18. About Our Facilities: Federal Prisons — Federal Bureau of Prisons, U.S. Department of Justice
  19. California Penal Code section 4000 - Common jails kept by sheriffs — California Legislative Counsel (leginfo)
  20. California Penal Code section 4015 - Sheriff to receive persons committed; county to bear expense — California Legislative Counsel (leginfo)

EliExplains lessons are original prose written from the open, credible references above. See Copyright & Licensing.

Researched 2026-08-19

Educational content only. It is not medical, legal or professional advice. Found an error? Tell us.