I need your help. I co-founded the Prison Policy Initiative to put the problem of mass incarceration — and the perverse incentives that fuel it — on the national agenda. Over the last 17 years, our campaigns have protected our democracy from the prison system and protected the poorest families in this country from the predatory prison telephone industry. Our reports untangle the statistics and recruit new allies.

But now, more than ever, we need your help to put data & compassion into the conversation.

Thank you.
—Peter Wagner, Executive DirectorDonate

Without federal aid, the rate of college course participation in prisons dropped by half.

by Wendy Sawyer, August 22, 2019

Journalists, policymakers, and advocates frequently ask us to answer tricky but important questions about the criminal justice system. Until now, however, many of our answers to these common questions have gone unpublished, gathering dust in our email archives. This post is the first in what we anticipate will be an ongoing “Since You Asked” series that makes our answers to these important questions public. (Want to send us your questions? Use our contact page.)

Q: How many people participated in college-in-prison programs before and after the 1994 crime bill? (And how many participate in prison education programming today?)

Graph breaking down college participation among people in prison over time.

A: To understand the drop in college participation in prisons following the 1994 “crime bill,” it’s important to know that most in-prison college programs, unlike colleges and universities in the “free world,” largely depend on funding from the Pell Grants program and other federal aid programs. This is largely because incarcerated people are overwhelmingly poor and can neither afford college tuition nor make generous gifts as alumni.

According to a historical overview by the American Enterprise Institute (AEI), prisons “witnessed a surge in demand for college courses” after the passage of the Higher Education Act of 1965, a law that greatly expanded federal aid for college participation nationwide. By 1982, 350 college-in-prison programs enrolled almost 27,000 prisoners (9 percent of the nation’s prison population), primarily through Pell Grants. … By the early 1990s, it is estimated that 772 programs were operating in 1,287 correctional facilities across the nation.”

A 1992 amendment to the Higher Education Act made people serving life sentences without parole and those sentenced to death ineligible to receive Pell Grants. Then, the 1994 Violent Crime Control and Law Enforcement Act banned everyone incarcerated in prisons from receiving Pell aid, even though these grants made up less than 1 percent of total Pell spending. “By 1997,” the AEI briefing continues, “it is estimated that only eight college-in-prison programs existed in the United States.” The remaining programs were those that received financial and volunteer support from other sources.

According to the American Historical Association, “States followed by further blocking access to funds through regional programs.” In 2005, the New York Times Magazine wrote that there were “about a dozen” prison college-degree programs, “four of them in New York State.”

The Bureau of Justice Statistics (BJS) 2003 publication Education and Correctional Populations includes some data that reflects these changes (see Table 4). In 1991, 13.9% of people in state prisons, and 18.9% of those in federal prisons, had taken a college course since admission. By 1997, these numbers had dropped to 9.9% for people in state prisons, and 12.9% in federal prisons. The 2004 BJS Survey of Inmates in State and Federal Correctional Facilities (variable V2503) shows that the downward trend continued: In 2004, just 7.3% of respondents in state prisons had taken a college-level class since admission.

More recently, results from a 2014 study of people in prison found that 2% of respondents had completed an Associate’s degree while incarcerated, and 1% had completed a Bachelor’s degree or higher. Notably, 58% of respondents reported that they had not completed any further education while incarcerated.

But these statistics do not reflect a lack of interest in higher education among people in prison: To the contrary, 40% of respondents to the survey said that they would like to enroll in an Associate’s or Bachelor’s degree program, and an additional 29% wanted to enroll in a postsecondary certificate program. Nor are most people in prison unqualified for such programs: A recent report from the Vera Institute of Justice and Georgetown Center on Poverty and Inequality analyzed the same survey results and found that “the majority of incarcerated people are academically eligible for postsecondary-level courses.”

As we’ve previously found, educational exclusion is a strong predictor of incarceration in the U.S. But federal and state governments are far from powerless to repair the harm that results. Restoring Pell Grants to incarcerated people would make approximately 463,000 people in prison eligible for free college courses.

It’s time for the government to not only restore this critical aid, but expand it.

by Bernadette Rabuy, August 9, 2019

In case you missed it, John Oliver exposed the high fees and low wages pervasive in prisons and jails on last Sunday’s episode of Last Week Tonight.

Oliver cited our research to shine a light on the low wages — or no wages, in the case of Alabama, Arkansas, Georgia and Texas — that incarcerated people receive for their hard work.

Despite the low wages paid to people in prison, Oliver explained, prisons squeeze money out of incarcerated people and their families by forcing them to pay for basic needs, such as:

  • Hygiene products: Too often, prisons do not provide sufficient hygiene products and incarcerated people are forced to buy additional items on their own dime. We found, for instance, that the average person in an Illinois prison spends $80 a year on toiletries and hygiene products.
  • Copays for medical visits: Our 2017 state-by-state analysis revealed that fourteen states charge co-pay amounts equivalent to charging minimum wage workers over $200.
  • Video calls: Oliver scrutinizes the high cost of video calls and the harmful trend of jails replacing in-person visits with video chats. Oliver states that a video call system is really a “machine that makes money by stopping people from visiting their families,” which is surely “an item at the top of Satan’s Amazon wish list.”

Oliver also shared his skepticism of correctional officials’ claims that banning in-person visits is justified because it reduces contraband. Oliver pointed out that contraband often enters correctional facilities through other channels, such as through staff.

“Part of the way mass incarceration persists in this country is by keeping the true costs of it off the books,” Oliver concludes. We couldn’t agree more. Thank you, Last Week Tonight, for helping us expose these harmful practices!

Three states have taken action in 2019 to change one of the most harmful policies in prison healthcare.

by Wanda Bertram, August 8, 2019

Prison healthcare is almost always a depressing topic, but not today, when we can report an important victory: Illinois recently became the third state in 2019 to reform the practice of charging medical co-pays to incarcerated people.

Previously, we published a state-by-state analysis showing that most prisons charge medical copays to people inside – despite the fact that their patients are impoverished and earn little to no money for their work in prisons. Using prison and free-world wage data, we calculated what each state’s co-pay would cost if it was charged to free-world patients making the minimum wage. Our analysis revealed that fourteen states charge co-pay amounts equivalent to charging minimum wage workers over $200.

Graph showing how much minimum wage earners in each state would pay if a single co-pay took as many hours to earn as a co-pay charged to an incarcerated person does. The average equivalent co-pay is about $200 and in West Virginia, it's over $1,000. Policy details and sourcing information can be found in the Appendix of our original analysis.

Charging co-pays to incarcerated people to shore up a state’s correctional budget is simply wrong. In our analysis, we explained that not only does this policy squeeze poor people and their families; it hurts public health by making the choice to seek medical attention a costly one.

We’re glad to see that states are now paying attention and taking action:

  • Illinois Governor J.B. Pritzker signed a bill in July eliminating the state’s $5 co-pay. Previously, people working in an Illinois state prison (where the minimum wage is 9 cents per hour) would have had to work for over 55 hours to afford the $5 fee. That’s like charging a non-incarcerated minimum-wage earner in Illinois over $460. In some cases, incarcerated people paid even more before getting adequate care, because paying the fee didn’t guarantee a doctor visit – only that a nurse would review one’s medical request.
  • Earlier this year, the California Department of Corrections and Rehabilitation announced that it would stop charging a $5 medical co-pay to incarcerated people. CDCR’s press release acknowledged what advocates have long known: Charging co-pays is bad for public health. “Copayments,” it said, “may hinder patients from seeking care for issues which, without early detection and intervention, may become exacerbated.” The department’s decision came as the California legislature was considering AB 45, a bill to eliminate medical co-pays in both state prisons and county jails. AB 45 (for which we wrote a letter of support) continues to move through the state legislature.
  • The Texas legislature stopped short of eliminating medical co-pays in prisons entirely, but made progress by replacing the notorious $100 fee Texas had charged incarcerated people with a $13.55 per-visit fee. While this change marks a substantial improvement, incarcerated people in Texas – who earn nothing for their labor – will continue to be charged the highest medical co-pay in state prisons nationwide.

State reforms like these can’t come soon enough. As we noted in our 2017 analysis, out-of-reach co-pays in prisons and jails have two inevitable and dangerous consequences. First, when sick people avoid the doctor, disease is more likely to spread to others in the facility – and into the community, when people are released before being treated. Second, illnesses are likely to worsen as long as people avoid the doctor, which means more aggressive (and expensive) treatment when they can no longer go without it.

It’s welcome news that states are finally taking action to change this risky and regressive policy. Will your state be next?

Women make up a growing share of incarcerated populations, and they have different needs than justice-involved men. Accordingly, some prison systems have begun to implement gender-responsive policies and programs. But what happens after release?

by Wendy Sawyer, July 19, 2019

Given the dramatic growth of women’s incarceration in recent years, it’s concerning how little attention and how few resources have been directed to meeting the reentry needs of justice-involved women. After all, we know that women have different pathways to incarceration than men, and distinct needs, including the treatment of past trauma and substance use disorders, and more broadly, escaping poverty and meeting the needs of their children and families. In recognition of these differences, and in an effort to reduce the harms of incarceration and the likelihood of re-incarceration, many prison systems have begun to implement gender-responsive policies and programs. But what’s being done to help women get the support they need to rebuild their lives after release?

A handful of programs have sprung up in communities around the country to meet the needs of women returning home: some founded by formerly incarcerated women themselves, some running on shoestring budgets for years, and all underscoring the need for greater capacity to meet the demand of over 81,000 releases from prison and 1.8 million releases from jail each year.

Map of US states showing the number of women released from state prisons each year. Nationally, over 81,000 women are released from state prisons annually. Women make up 1 in 8 individuals released from state prisons each year, but the numbers vary widely between states. The additional 1.8 million women released from local jails annually are not shown on this map, because not all states have data available for jail releases, and in most states, a significant portion of reported releases are missing data on sex. See the table below for the available jail data.

Sources and data notes: Bureau of Justice Statistics National Corrections Reporting Program: 1991-2016, Selected Variables, Prison Releases and, for Vermont, CSAT-Prisoners Custom tables, Count of total releases. Data for all states are from 2016 except for New Mexico (2015), North Dakota (2015), and Oregon (2013).

Annual releases of women from state prisons and local jails

Prisons Jails Prisons & Jails combined
State Women released Percentage of all prison releases Women released* Percentage of all jail releases* Women released* Percentage of all releases*
Alabama 2,013 14% 39,740 18% 41,753 18%
Alaska 2,181 22% 510 14% 2,691 20%
Arizona 2,707 14% 46,162 23% 48,869 22%
Arkansas 4,456 15% 27,322 14% 31,778 15%
California 2,495 7% 186,571 20% 189,066 20%
Colorado 1,178 13% 44,204 24% 45,382 23%

Connecticut 1,281 13% n/a n/a n/a n/a
Delaware 1,749 19% n/a n/a n/a n/a
Florida 3,670 12% 121,336 19% 125,006 18%
Georgia 1,931 11% 84,398 18% 86,329 18%
Hawaii 141 14% n/a n/a n/a n/a
Idaho 931 21% 11,323 22% 12,254 22%
Illinois 2,190 8% 33,875 12% 36,065 11%
Indiana 2,087 14% 27,520 11% 29,607 11%
Iowa 865 16% 25,998 22% 26,863 22%
Kansas 1,012 16% 24,640 17% 25,652 17%
Kentucky 3,563 21% 104,403 23% 107,966 23%
Louisiana 1,853 11% 45,935 16% 47,788 16%
Maine 159 13% 5,422 27% 5,581 26%
Maryland 756 8% 27,711 17% 28,467 17%
Massachusetts 738 26% 4,323 6% 5,061 7%
Michigan 959 9% 51,240 17% 52,199 17%
Minnesota 898 11% 36,230 20% 37,128 20%
Mississippi 965 12% 17,207 17% 18,172 17%
Missouri 3,454 18% 36,395 16% 39,849 16%
Montana 154 13% 7,755 20% 7,909 20%
Nebraska 321 14% 16,902 24% 17,223 24%
Nevada 816 14% 34,374 23% 35,190 23%
New Hampshire 240 15% 5,732 26% 5,972 25%
New Jersey 563 6% 18,037 13% 18,600 13%
New Mexico 502 14% 20,171 21% 20,673 20%
New York 1,551 7% 28,241 15% 29,792 14%
North Carolina 2,411 12% 65,085 16% 67,496 15%
North Dakota 306 20% 3,857 13% 4,163 13%
Ohio 3,521 14% 79,848 20% 83,369 20%
Oklahoma 1,724 16% 23,453 14% 25,177 14%
Oregon 639 12% 38,292 22% 38,931 22%
Pennsylvania 2,374 9% 34,698 18% 37,072 17%
Rhode Island 326 10% n/a n/a n/a n/a
South Carolina 1,065 12% 45,045 14% 46,110 14%
South Dakota 447 18% 14,011 25% 14,458 24%
Tennessee 2,478 17% 94,544 24% 97,022 24%
Texas 12,453 16% 145,430 14% 157,883 14%
Utah 540 15% 14,165 16% 14,705 16%
Vermont 280 16% n/a n/a n/a n/a
Virginia 1,771 14% 51,266 18% 53,037 18%
Washington 948 12% 62,503 23% 63,451 23%
West Virginia 583 17% 11,201 24% 11,784 24%
Wisconsin 835 9% 34,906 18% 35,741 17%
Wyoming 149 14% 2,580 8% 2,729 9%
National 81,229 13% 1,854,561 18% 1,935,790 18%
*Note: The numbers in this table represent the minimum number and percentages of jail releases that are women. The actual numbers are probably greater, because a significant number of releases reported in the Census of Jails, 2013 are missing data on sex. (Nationally, 15% of all release records were missing this data.) The percentages of releases that are women, as reported in this table, were calculated based on the total number of jail releases, including those with no data on sex.

Sources and data notes: For prison releases, data are from the Bureau of Justice Statistics National Corrections Reporting Program, except for Vermont, which did not report release data by sex to NCRP. Vermont’s prison release data comes from BJS’ CSAT-Prisoners tool, and only includes releases of individuals sentenced to more than 1 year. Prison release data are from 2016 for all states except New Mexico (2015), North Dakota (2015), and Oregon (2013). Jail data are from the BJS Census of Jails, 2013, and are not available for 5 states (Connecticut, Delaware, Hawaii, Rhode Island, and Vermont) where the jail system is entirely integrated into the state prison system. In Alaska, there are a small number of locally operated jails not part of the state system, so available data reflect only the locally operated jails and not the entire jail population.

In 2016, about 81,000 women were released from state prisons nationwide, and women and girls accounted for at least 1.8 million releases from local jails in 2013 (the last year all jails were surveyed). While many people are released from jail within a day or so and may not need reentry support, jail releases can’t be overlooked, especially for women, who are more likely than men to be incarcerated in jails as opposed to prisons. (Moreover, jails typically provide fewer programs and services than prisons, so individuals released from jails are even less likely to have received necessary treatment or services while incarcerated than those in prison.)

Those figures mean that nationally, about 1 in 8 (13%) of all individuals released from state prisons – and more than 1 in 6 (18%) jail releases – are women. In 20 states, at least 1 in 5 (20%) individuals released from any incarceration (either prison or jail) is female. Fully half of all states release at least 1,000 women from prison annually; in Texas, it’s over 12,000 women per year.

As in other stages of the criminal justice system, most post-release policies and programs were created with the much larger male population in mind. But research makes clear that women returning home have “a significantly higher need for services than men,” and that reentry supports should be responsive to the particular needs of justice-involved women:

  • Economic marginalization and poverty: As we’ve previously shown, formerly incarcerated women (especially women of color) have much higher rates of unemployment and homelessness, and are less likely to have a high school education, compared to formerly incarcerated men. These findings help explain why, in a 2012 National Institute of Justice (NIJ) study, 79% of women interviewed 30 days pre-release cited “employment, education, and life skills services” as their greatest area of need (followed closely by transition services). An earlier study (Holtfreder et al., 2004), found that poverty is the strongest predictor of recidivism among women, and “providing state‐sponsored support to address short‐term needs (e.g., housing) reduces the odds of recidivism by 83%” for poor women on probation and parole.
  • Housing: A 2017 Prisoner Reentry Institute (PRI) report identified homelessness and the lack of stable housing as the biggest problem facing women in the New York City justice system, noting that 80% of women at Rikers said they needed assistance finding housing upon discharge. A 2006 California study found that 75% of formerly incarcerated women surveyed had experienced homelessness as some point, and 41% were currently homeless. Women who can’t secure safe housing may return to abusive partners or family situations for housing and financial reasons – a point echoed in interviews with paroled women in a study by Brown and Bloom.
  • Trauma and gendered pathways to incarceration: The PRI report emphasizes the importance of gender-responsive and trauma-informed interventions for reducing recidivism among women. According to that report, such interventions should: provide a safe, respectful environment; promote healthy relationships; address substance use, trauma, and mental health issues; provide women with opportunities to improve their socioeconomic conditions; establish “comprehensive and collaborative” community services; and prioritize women’s empowerment.
  • Family reunification: Most incarcerated women are mothers, and are frequently the primary caretakers of their children. The importance of family reunification – noted throughout the literature, by Carter et al. (2006), Brown and Bloom (2009), Wright, et al. (2012), the NIJ (2012), among others – cannot be overstated, especially given the trauma experienced by children when separated from a parent.

While the complexity of women’s reentry needs can be daunting, there are successful models in operation demonstrating how states, counties, and communities can best serve them. Notably, A New Way of Life Reentry Project operates eight houses in Los Angeles and is working toward expanding its model nationally. The program offers wraparound services including transitional housing, case management, and legal services to support women as they navigate reentry. Staff support women from initial reentry tasks like obtaining ID cards and applying for public assistance all the way through the process of regaining custody of children and finding permanent housing. Similar programs offering wraparound services exist in other cities, such as the Ladies of Hope Ministry’s Hope House in New York City; the Center for Women in Transition in St. Louis; and Angela House in Houston, which also provides programming tailored “to the health and psychosocial needs of women recovering from sexual exploitation.”

Frustratingly, despite their success, these programs lack the funding and capacity to serve all of the women who desperately need them: Angela House notes on its website that it can only serve 12 to 14 women at a time, but receives more than 300 applications every year. Unless state governments and federal agencies take action to grow the capacity of these service providers, hundreds of thousands of women every year will leave prison or jail without the resources they need to succeed. As lawmakers increasingly call for policy changes to help women in prison, they must not ignore the massive gap between the need and availability of women’s reentry programs.

A detailed spreadsheet including release data for both men and women is available on our Data Toolbox page.

"Women’s experiences with the criminal justice system serve to highlight the faults of the whole system," our Legal Director explained to members of the House Judiciary Committee.

by Wanda Bertram, July 17, 2019

Yesterday, our Legal Director, Aleks Kajstura, spoke to members of the House Judiciary Committee at a hearing about the mass incarceration of women and girls. “Today’s hearing begins a discussion about women in the criminal justice system,” said subcommittee chair Rep. Karen Bass (D-CA). “After decades of policies that led to mass incarceration, we are finally at a point of examining the policies and the consequences.”

Testifying alongside Piper Kerman, Jesselyn McCurdy, Cynthia Shank and Patrice Onwuka, Kajstura spoke about the urgency of reducing women’s incarceration. She highlighted a few of the most alarming data points from our research on women and gender, including:

  • The U.S. is home to only 5% of the world’s female population, but accounts for nearly 30% of the world’s incarcerated women.
  • 1 in 4 incarcerated women have not even been convicted.
  • In a number of states, women’s prison populations are growing faster than men’s; in others, women’s prison populations are growing even as men’s are declining.
  • 3 out of 4 women under correctional control are on probation, a system that sets up women – and particularly mothers – to fail.

graph comparing the U.S. to other founding NATO countries on rates of women's incarceration.

Despite this unacceptable status quo, the hearing sent a hopeful message: Lawmakers across the aisle are finally paying attention to women’s mass incarceration. “It is critical that we understand how and why women become involved in the system, what happens to them when they are incarcerated, and what their trajectory is once released,” said Bass.

Watch the full hearing here – jump to 1:03:00 for Kajstura’s testimony: https://judiciary.house.gov/legislation/hearings/women-and-girls-criminal-justice-system

13 states in the hottest parts of the country lack universal A/C in their prisons. We explain the consequences.

by Alexi Jones, June 18, 2019

Air conditioning has become nearly universal across the South over the last 30 years, with one exception: in prisons. Although 95% of households in the South use air conditioning, including 90% of households that make below $20,000 per year,1 states around the South have refused to install air conditioning in their prisons, creating unbearable and dangerous conditions for incarcerated people.

13 famously hot states lack universal A/C in their prisons

While there are no national statistics on air conditioning in prison, we found that at least 13 states in the hottest regions of the country lack universal air conditioning in their prisons:

  • Alabama
  • Arizona
  • Florida
  • Georgia
  • Kansas
  • Kentucky
  • Louisiana
  • Mississippi
  • Missouri
  • North Carolina
  • South Carolina
  • Texas
  • Virginia

For more information on these states, see the appendix.

The lack of air conditioning in Southern prisons creates unsafe—even lethal—conditions. Prolonged exposure to extreme heat can cause dehydration and heat stroke, both of which can be fatal. It can also affect people’s kidneys, liver, heart, brain, and lungs, which can lead to renal failure, heart attack, and stroke.

Many people in prison are especially susceptible to heat-related illness, as they have certain health conditions or medications that make them especially vulnerable to the heat. Conditions such as diabetes and obesity can limit people’s ability to regulate their body heat, as can high blood pressure medications and most psychotropic medications (including Zoloft, Lexapro, Prozac, Cymbalta, and more but excluding the benzodiazepines). Old age also increases risk of heat-related illness, and respiratory and cardiovascular illnesses, such as asthma, are exacerbated by heat.

In Texas, a state that has air conditioning in all inmate housing areas in only 30 of its 109 prisons, a high percentage of incarcerated people are particularly vulnerable to heat:

A chart showing the percentage of people incarcerated in Texas with taking high blood pressure medication, psychiatric medication, asthma, and diabetes

The structure of prisons and prison life can also make incarcerated people more vulnerable to heat. Prisons are mostly built from heat-retaining materials which can increase internal prison temperatures. Because of this, the temperatures inside prisons can often exceed outdoor temperatures. Moreover, people in prison do not have the same cooling options that people on the outside do. As Prison Legal News explains in a 2018 article on prison air conditioning litigation, “people outside of prison who experience extreme heat have options that prisoners often lack – they can take a cool shower, drink cold water, move into the shade or go to a place that is air conditioned. For prisoners, those options are generally unavailable.” Even fans can even be inaccessible. For example, despite the fact that incarcerated people in Texas are not paid for their labor, purchasing a fan from the Texas prison commissary costs an unaffordable $20.

The lack of air conditioning in prisons has already had fatal consequences. In 2011, an exceptionally hot summer in Texas, 10 incarcerated people died from heat-related illnesses during a month-long heat wave. (It’s just not incarcerated people who get sick from the heat in the state’s prisons. In August 2018, 19 prison staff and incarcerated people had to be treated for heat-related illnesses.) As David Fathi, director of the American Civil Liberties Union National Prison Project, explained to The Intercept, “Everyone understands that if you leave a child in a car on a hot day, there’s a serious risk this child could be injured or die. And that’s exactly what we’re doing when we leave prisoners locked in cells when the heat and humidity climb beyond a certain level.”

Courts in Wisconsin, Arizona, and Mississippi have ruled that incarceration in extremely hot or cold temperatures violates the Eighth Amendment. But these court cases have not had a national impact on air conditioning in prisons. As Alice Speri of The Intercept explains, “There’s no national standard for temperatures in prisons and jails, and as jurisdiction over prisons is decentralized among states and the federal system, and jurisdiction over jails is even more fragmented among thousands of local authorities across the country, fights over excessive heat in detention can only be waged facility by facility.” As a result, incarcerated people in the South are subjected to unbearable conditions that violate their basic human and constitutional rights. Benny Hernandez, an incarcerated man in Texas, describes how torturous heat gets in prisons: “It routinely feels as if one’s sitting in a convection oven being slowly cooked alive. There is no respite from the agony that the heat in Texas prisons inflicts.”

Refusing to install air conditioning is a matter not just of short-term cost savings, but of appearing tough on crime. State and local governments go to astonishing lengths to avoid installing air conditioning in prisons. In 2016, Louisiana spent over $1 million in legal bills in an attempt to avoid installing air conditioning on death row, an amount four times higher than the actual cost of installing air conditioning, according to an expert witness. Similarly, in 2014, the people of Jefferson Parish, LA only voted to build a new jail after local leaders promised there would be no air conditioning.

With air conditioning nearly universal in the South, air conditioning should not be considered a privilege or amenity, but rather a human right. States and counties that deny air conditioning to incarcerated people should understand that, far from withholding a “luxury,” they are subjecting people to cruel and unusual punishments, and even handing out death sentences.


  1. The U.S. Energy Information Administration’s 2015 Residential Energy Consumption Survey has data on air conditioning use by income and geographical region. This Agency uses the Census Bureau’s definition of the South: Delaware, Maryland, Virginia, West Virginia, Kentucky, North Carolina, South Carolina, Tennessee, Georgia, Florida, Alabama, Mississippi, Arkansas, Louisiana, Texas, and Oklahoma plus Washington D.C. Nationwide, air conditioning usage is slightly lower than in just the South, with 87% of households (and 80% of people making below $20,000 per year) using air conditioning nationwide. .  ↩



Examining local and national news stories, we identified 12 states in the South and Midwest that lack universal air conditioning and identified only Arkansas as having universal air conditioning.

State Air Conditioning?
Alabama Prisons in Alabama do not have air conditioning. (Source)
Arizona Many prisons in Arizona lack air conditioning. (Source)
Arkansas Prisons have had air conditioning since the 1970s. (Source)
Florida State run prisons do not have air conditioning, but private prisons in the state do have air conditioning. (Source)
Georgia Most prisons have air conditioning, but some do not. (Source)
Kansas Most prisons do not have air conditioning. 70 percent of incarcerated people are in buildings without air conditioning. (Source)
Kentucky Most prisons do not have air conditioning. (Source)
Louisiana Most prisons do not have air conditioning. (Source)
Mississippi Most inmate housing in Mississippi has no air conditioning. (Source)
Missouri Some prisons lack universal air conditioning. (Source)
North Carolina Most prisons have air conditioning, but 10 facilities do not. (Source)
South Carolina Most prisons have air conditioning, but some facilities do not. (Source)
Texas 30 of the 109 state prisons in Texas have air conditioning in all housing areas. (This is despite the fact county jails in the state are statutorily required keep their temperatures between 65 and 85 degrees). (Source)
Virginia Half of prisons have no air conditioning. (Source)


As more jails ban face-to-face visits in favor of paid video chats, a growing number of people in jail are being cut off from their families when the technology breaks down.

by Sarah Watson, June 18, 2019

Jails are increasingly replacing in person visits with video calls. This high-tech fad goes against the recommendations of the American Correctional Association, the American Bar Association, and even the Department of Justice’s National Institute of Corrections. These jails ignore the many problems we’ve documented, such as high costs for families, poor quality of the systems and the loss of human contact. But there’s another liability that jails now have to consider: What happens when their shiny new technology fails?

Their vendors — who provide the systems for free in exchange for charging high rates to the families — will say that their technology is perfect. As every person who owns a computer knows, however, technology is not flawless, and these systems do fail — sometimes keeping people in jail from contacting their families for weeks at a time:

We surveyed recent news stories about video systems breaking in jails that had chosen to replace traditional in-person visits with the technology.
County State Time Down Year Details Source
Shelby County Jail (Memphis) Tennessee 2 weeks 2019 The vendor (GTL) cut a fiber optic cable Source
Virginia Beach Correctional Center Virginia 3 months 2018 Jail typically averages 4,000 visits a month Source1 Source2
Williams County Correctional Center North Dakota 2 months 2017 System updates originally brought the system down, then it was discovered the upgrade was incompatible with the old equipment Source
Milwaukee County Jail Wisconsin At least one month 2013-2014 Visual went down leaving only audio Source
Boone County Jail Arkansas “months” 2018 Either a lightning strike or a software glitch brought the system down, administrators were not sure which was the cause. Source
Volusia County Branch Jail Florida One month 2017 Lightning struck an integral part of the visitation system. It took multiple technicians to conclude the entire system needed to be replaced. Source
Pontotoc County Justice Center Oklahoma Three weeks 2017 Visitation went down due to a “computer issue.” Source
Madison County Detention Center (Huntsville) Alabama >2 weeks 2017 Planned system updates meant visits were suspended. Source
Montgomery County Detention Facility Alabama About a week 2018 “Technical Issues” disrupted visitation as the jail waited for a replacement machine. Source
Olmsted County Adult Detention Center (Rochester) Minnesota 1 day 2018 The visitation system crashed, leaving visitors unable to schedule a visit or see their loved ones. The sheriff also confirmed the system sometimes goes down due to weather conditions as well. Source
Ada County Jail Idaho 2019 Ada County Jail experiences consistent technical issues. One visitor to Ada County Jail recalled, “It didn’t work half the time. You’d have to call to see if [the system] was down.” Source
Mecklenburg County Jail (Charlotte) North Carolina 2017 Frequent problems and system outages caused prisoners to miss their visits. “The video chat would go in and out. Sometimes half the screen would be cut off, and sometimes they wouldn’t work at all,” a former prisoner remembered. “You wouldn’t even get your visitation; you would have to wait until the next week, because even though the system was down, they would not make up the visitation you missed.” Source

The good news is that counties are starting to take notice of the downsides to video calling. Most recently, the sheriff of Mecklenburg County, North Carolina (see table above) fulfilled a 2018 campaign promise to reinstate in-person visits, on the grounds that video should be used in addition to in person visits, not as a substitute.

“Allowing our residents to stay connected to family and loved ones through in-person visits improves public safety,” Sheriff McFadden explained. “This simple step alone has been shown to significantly lower the chances that a person will commit another crime after they get out. It also reduces the chance a person will commit an infraction inside the jail which could adversely impact their release. In addition, it improves mental health outcomes and strengthens family units and community ties.”

Mecklenburg County had the right idea. When this technology works, it should be considered a supplement to in-person visits, not a substitute; and when the technology fails, it’s useless. Instead of investing in flawed technology, jails should be looking for more ways to increase traditional methods of family contact.

The Democratic candidates are missing an opportunity to pitch sweeping criminal justice reform as an economic justice issue.

by Wanda Bertram, June 12, 2019

Multiple Democratic presidential candidates have staked their campaigns on promises to fight for economic justice and protect low-income people from ruin. So it’s mysterious and frustrating that none of these candidates have proposed to end our justice system’s criminalization of poverty – at least beyond the occasional nod to ending money bail.

These candidates are missing an opportunity. The incomes of people in U.S. prisons and local jails are overwhelmingly low, and one in two American adults has had a close relative incarcerated, meaning that a candidate who understands the criminalization of poverty could propose transformative reforms and speak to a huge number of voters. In particular, candidates are missing an opportunity to speak to Black voters, who are hit hardest by policies that punish poor people.

To be sure, many Democratic candidates have alluded to economic inequality in connection with criminal justice reform — and Bernie Sanders even uses the phrase “criminalizing poverty” on his campaign website — but I’ve seen no indication that any of the candidates can speak to either the specifics or the scale of this problem. Candidates must go beyond criticizing money bail, and promise to end the unequal treatment of poor people at every stage of the justice process:


1. End poverty-related arrests and jail bookings. Far too many Americans who can’t afford housing, drug treatment or mental health services are instead arrested on minor charges related to their homelessness or illness. Many others end up in jail because they can’t afford burdensome fines and fees. An unfortunate downstream effect is jail overcrowding, which leads counties – largely in rural areas – to spend public money on harmful jail expansion rather than social welfare.

Presidential candidates should commit to helping state and local governments shift their priorities, making it easier to support low-income people and harder to jail them:


2. Guarantee poor people equal justice before trial. Two major injustices — pretrial detention and lack of access to counsel — ensure that low-income people are disproportionately convicted. Pretrial detention doesn’t just make defendants more likely to plead guilty; it also puts them at risk of losing their jobs and homes, and imposes huge costs on their families, before they’re ever convicted.

Candidates should promise to:

  • Ensure that local public defender systems are fully funded, that they no longer charge co-pays to defendants, and that counsel is guaranteed at any hearing that could result in detention.
  • Incentivize counties to drastically reduce pretrial detention by ending commercial money bail, and replacing it with release on recognizance, unsecured bonds, and other alternatives.
  • Use the power of the Federal Communications Commission to regulate the cost of phone calls from jail, which can strain public defenders’ resources, not to mention those of family members.
  • Subsidize county-level pretrial services to help low-income people make their court dates, such as text reminders and free childcare at court.


3. Stop forcing low-income families to subsidize the prison system. When someone goes to prison, their loved ones become their source of financial support. The financial pressure on these families grows when prisons fail to provide basic services, often driving families into debt. It shouldn’t be on relatives — disproportionately women — to pay for phone calls, medical care, nutritious food and educational resources for those behind bars, often when they’ve just lost a breadwinner.

Presidential candidates should commit to paying for incarcerated people’s needs to lift this burden on families:


4. Protect people from going back to prison just because they’re poor. Incarcerated people who come from under-resourced neighborhoods tend to return to those same neighborhoods after prison, now saddled with criminal records and far poorer than before.

Presidential candidates should avoid old narratives about how more surveillance, monitoring and job training are needed to “reduce recidivism” among people leaving prison. Instead, they should call the reentry process what it is: a period of extreme vulnerability that mostly affects impoverished people, and that can’t be improved without serious investments in formerly incarcerated people’s welfare. They should commit to:

  • Identify the communities to which most formerly incarcerated people return, and subsidize additional low-income housing, drug treatment and mental health services in those communities.
  • Help states create Departments of Reentry that connect people nearing release to permanent housing, medical care, and other resources.
  • Incentivize states to pass laws expanding criminal record expungement, including automatic expungement for people convicted of minor offenses.
  • End the harmful restrictions on association that prohibit formerly incarcerated people from helping each other rebuild their lives.
  • Restore welfare benefits, including housing assistance, to people with criminal records.
  • Urge states to abolish probation and parole fees, including fees for ankle monitors. End so-called “pay-only” probation schemes, which extract fees without providing any “services” at all.
  • Incentivize states to “ban the box” on job applications, and to end restrictions on occupational licenses that lock people with criminal records out of good jobs.
  • Expand federal tax benefits for businesses that hire people with criminal records.
  • Incentivize states to end laws suspending driver’s licenses for non-driving offenses.

To be sure, there are many other policy changes that could help end the criminalization of poverty; this list is only a starting point. But it should be alarming that most of these policy options have gone unmentioned by any presidential candidate. Until the candidates commit to ending our criminal justice system’s abuse of poor people from arrest to release (and afterwards), their visions for economic justice won’t be complete.

A new government report reinforces harmful misconceptions about people convicted of sex offenses. Here's our take on how to parse the data.

by Wendy Sawyer, June 6, 2019

By now, most people who pay any attention to criminal justice reform know better than to label people convicted of drug offenses “drug offenders,” a dehumanizing label that presumes that these individuals will be criminals for life. But we continue to label people “sex offenders” – implying that people convicted of sex offenses are somehow different.

A new report released by the Bureau of Justice Statistics should put an end to this misconception: The report, Recidivism of Sex Offenders Released from State Prison: A 9-Year Follow-Up (2005-2014), shows that people convicted of sex offenses are actually much less likely than people convicted of other offenses to be rearrested or to go back to prison.

But you wouldn’t know this by looking at the report’s press release and certain parts of the report itself, which reinforce inaccurate and harmful depictions of people convicted of sex offenses as uniquely dangerous career criminals. The press release and report both emphasize what appears to be the central finding: “Released sex offenders were three times as likely as other released prisoners to be re-arrested for a sex offense.” That was the headline of the press release. The report itself re-states this finding three different ways, using similar mathematical comparisons, in a single paragraph.

What the report doesn’t say is that the same comparisons can be made for the other offense categories: People released from sentences for homicide were more than twice as likely to be rearrested for a homicide; those who served sentences for robbery were more than twice as likely to be rearrested for robbery; and those who served time for assault, property crimes, or drug offenses were also more likely (by 1.3-1.4 times) to be rearrested for similar offenses. And with the exception of homicide, those who served sentences for these other offense types were much more likely to be rearrested at all.

Chart comparing 9-year rearrest rates by most serious commitment offense type. The chart shows that people released after serving sentences for rape or sexual assault are much less likely than those who served sentences for property, drug, public order, or violent crimes generally to be rearrested.

The new BJS report, unfortunately, is a good example of how our perception of sex offenses is distorted by alarmist framing, which in turn contributes to bad policy. That this publication was a priority for BJS at all is revealing: this is the only offense category out of all of the offenders included in the recidivism study to which BJS has devoted an entire 35-page report, even though this group makes up just 5% of the release cohort. This might make sense if it was published in an effort to dispel some myths about this population, but that’s not what’s happening here.

Every piece on recidivism should come with a warning label

Whenever we talk about recidivism, it’s important to understand that the data itself is categorically flawed.

Any time we talk about recidivism, it’s important to understand that recidivism data – which is so essential to the idea of a permanent “sex offender” – is categorically flawed. “Recidivism” suggests a relapse in behavior that leads to a return to criminal offending. As in the BJS study, this is often measured by post-release arrests (rearrest), but arrest does not suggest conviction or even actual guilt; of all recidivism measures, rearrest casts the widest net.2 Usually, these measures also don’t account for what the post-release offense is, even though common sense tells us that a post-release arrest for a liquor law violation is very different than an arrest for another sex offense. Better measures to indicate a return to criminal behavior might be reconviction or receiving a new prison sentence, but most jurisdictions don’t track all of these indicators, and certainly not with much detail.3 Finally, timeframes matter, since rearrest is much more likely soon after release, and the longer people go without reoffending, the lower their risk of ever doing so. This BJS report offers a remarkably long look-back period, but such a long period also risks correlating criminal behaviors that stem from unrelated motivations or circumstances.

For more on the myriad problems with recidivism data, see:

Framing aside, the recidivism data presented in the BJS report can offer helpful perspective on the risks posed by people after release. Whether measured as rearrest, reconviction, or return to prison, BJS found that people whose most serious commitment offense was rape or sexual assault were much less likely to reoffend after release than those who served time for other offense types. The BJS report shows that within 9 years after release:

  • Less than 67% of those who served time for rape or sexual assault were rearrested for any offense, making rearrest 20% less likely for this group than all other offense categories combined (84%). Only those who served time for homicide had a lower rate of rearrest (60%).
  • People who served sentences for sex offenses were much less likely to be rearrested for another sex offense (7.7%) than for a property (24%), drug (18.5%), or public order (59%) offense (a category which includes probation and parole violations).
  • Only half of those who served sentences for rape or sexual assault had a new arrest that led to a conviction (for any offense), compared to 69% of everyone released in 2005 (in the 29 states with data).

While the data was more limited on returns to prison,1 the study found that within 5 years after release, people who had served sentences for rape or sexual assault also had a lower return-to-prison rate (40%) compared to the overall rate for all offense types combined (55%). BJS notes that some of these returns to prison were likely for parole or probation violations, but because of data limitations, it is impossible to say how many were for new offenses, much less how many were for rape or sexual assault.

In sum, the BJS data show that people who served time for sex offenses had markedly lower recidivism rates than almost any other group. Yet the data continue to be framed in misleading ways that make it harder to rethink the various harmful and ineffective punishments imposed on people convicted of sex offenses.

The recidivism data suggest that current legal responses to people convicted of sex offenses are less about managing risk than maximizing punishment. The desire for retribution is understandable; unquestionably, rape and sexual assault inflict serious and lasting trauma. But our criminal justice system does a poor job of providing survivors of rape, sexual assault, and other violent crimes what they really want. In a 2016 survey of crime survivors, the Alliance for Safety and Justice found that, “Survivors of violent crime — including victims of the most serious crimes such as rape or murder of a family member — widely support reducing incarceration to invest in prevention and rehabilitation and strongly believe that prison does more harm than good.” But more prison time is the default response: those released after serving sentences for rape and sexual assault served longer sentences, with a median sentence of 5 years (compared to 3 years for all others combined) and over a quarter serving 10 years or more before release.

And for many people convicted of sex offenses, confinement doesn’t end when their prison sentence does. Twenty states continue to impose indefinite periods of involuntary confinement under civil commitment lawsafter individuals have completed a sentence (or, in some cases, before they are even convicted). Proponents justify the practice as “treatment,” but conditions of civil commitment are punitive and prison-like, and this confinement is hard to justify with the recidivism data we have. The likelihood of post-release arrest for another rape or sexual assault for this group is less than 2% in the first year out of prison, and after 9 years, less than 8% have been rearrested for a similar offense. Those who are released at age 40 or older are even less likely to be rearrested for another sex offense, with re-arrest rates about half those of people who are released at age 24 or younger.

After prison, a number of other special restrictions make reentry especially challenging for those who have served sentences for sex offenses, including registration, public notification, and restrictions to residence and employment. A current proposal suggests banning them from using New York City mass transit. (Even before release, some restrictions make it difficult for some people to leave prison when they would otherwise be paroled.) But these restrictions tend to cause more problems than they solve. Residence restrictions in particular have contributed to homelessness and other problems in cities where they leave little room for returning citizens. According to a 2015 U.S. Department of Justice brief, “residence restrictions may actually increase offender risk by undermining offender stability and the ability of the offender to obtain housing, work, and family support.”

In another recent academic article, Hanson et al. agree that these additional restrictions are “justified on the grounds of public protection,” even though the underlying assumptions may be wrong: “Individuals are targeted because policy-makers believe they are likely to do it again. This is a testable assumption, and, as it turns out, not entirely true.” Their analysis shows that individual recidivism risk varies widely, can be low enough to be indistinguishable from that of people convicted of non-sex offenses, and drops predictably over time. The data published by BJS track with those findings.

Collectively, the research seems fairly clear: our responses to people convicted of sex offenses do not reflect the actual – generally low – risks they present. Instead of panicking about the small portion who reoffend after release, it’s time we talk more rationally about responses that effectively support desistence from crime – and serve the actual needs of victims of violence.


  1. Only 23 states could provide the necessary data for the 5-year follow-up period, and only 17 could do so for the entire time frame. The BJS report only includes return-to-prison rates for the first 5 years after release in the 23 states with the necessary data.  ↩

  2. Conversely, it also only captures those behaviors that are caught by police. People who break laws after release but are never arrested would not be captured in recidivism data at all. Police presence and enforcement are therefore factors that affect recidivism statistics, as are prosecutorial decisions (for reconviction rates) and sentencing policies and practices (for reincarceration rates).  ↩

  3. While the BJS study compares overall rates of reconviction and returns to prison by most serious commitment offense, only the rearrest data allows us to compare post-release offenses by most serious commitment offense.  ↩

We analyze gender and racial disparities in traffic and street stops, including arrests, searches, and use of force that occurs during stops.

by Prison Policy Initiative, May 14, 2019

Jails have been described as the criminal justice system’s “front door,” but jail incarceration typically begins with the police, with an arrest. Before any bail hearing, pretrial detention, prosecution, or sentencing, there is contact with the police. But despite their crucial role in the process, we know less about these police encounters than other stages of the criminal justice system.

In particular, the experiences of women and girls1 – especially Black women and other women of color – are lost in the national conversation about police practices. They are also largely invisible in the the data. But as Andrea Ritchie details in Invisible No More: Police Violence Against Black Women and Women of Color, women, too, are subject to racial profiling, use of excessive force, and any number of violations of their rights and dignity by police. In fact, women make up an increasing share of arrests and report much more use of force than they did twenty years ago. Yet while increasing recognition of women as a growing share of prison and jail populations has prompted facilities to adopt gender-responsive policies and practices, women’s rising share of arrests and other police contact has received less attention and policy response.

The current study

To shed more light on how women’s experiences at the front end of the criminal justice system differ from men’s experiences, we look at both arrest data and data about women’s other contacts with police. We first look at trends in arrest data from the FBI’s Uniform Crime Reporting (UCR) Program. To complement this data, we also examine differences in other police encounters reported in the Bureau of Justice Statistics (BJS) Contacts Between Police and the Public (CPP) reports, which are based on responses to the Police-Public Contact Survey (PPCS).

Critically, the UCR Program does not require police to report arrests by both sex and race/ethnicity, so that dataset offers no way to compare arrest trends of white, Black, and Latina women to each other or to their male counterparts. The PPCS partially fills that gap by including both sex and race/ethnicity of survey respondents; however, the BJS reports based on that survey do not engage this valuable intersectional data.2 The second section of this briefing includes our analysis of the most recent PPCS survey (conducted in 2015), to finally offer a view – from nationally representative data – of how women of different races and ethnicities experience police-initiated encounters differently than each other and men.


Part 1: Gender differences in interactions with police

Arrests: Women make up an increasingly large share of arrests

In the past two decades, the total number of arrests in the U.S. has dropped by more than 30 percent, from 15.3 million in 1997 to 10.6 million in 2017. However, this drop was mostly due to fewer arrests of men: the number of men arrested declined by 30.4 percent in that time, while the number of women arrested declined only 6.4 percent. As men’s arrest rates have fallen women’s arrest rates have remained fairly flat. As a result, women make up an increasingly large share of all arrests; as of 2017, women accounted for 27 percent of all arrests, up from 21 percent in 1997, and just 16 percent in 1980.3

An increase in arrests of women for drug offenses helps explain why women’s arrest rates have remained steady, even as crime rates have hit historic lows and men’s arrest rates have plummeted. Of the more than 2 million arrests of women in 2017, 13.9 percent were for drug abuse violations – second only to property crimes (15.8 percent), and far more frequent than arrests for violent crimes (3.8 percent). Over the past five years – while the country has been in the throes of the opioid epidemic – drug arrests have increased 6 percent among men, but almost 25 percent among women.4

Changes in arrest patterns over 5 years, by sex

Percent change over 5 years (2013-2017)
Male Female
All arrests -8.7% -6.4%
Index violent crimes +1.8% +4.4%
Index property crimes -18.8% -24.9%
Drug abuse violations +6.1% +24.7%
Table 1. Total UCR arrests declined more dramatically among men than among women between 2013 and 2017, due in part to a much greater increase in arrests of women for drug violations. Source: Crime in the United States 2017 Table 35


Other police-initiated contact: Women interact with police more than arrest statistics suggest

A key finding from the Contacts Between Police and the Public series is the true scale of women’s interactions with police, which is far greater than arrest numbers alone suggest. While arrests are sometimes used as a measure of the “first stage” of involvement with the criminal justice system, many encounters between the public and police do not result in arrest. This is especially true for women, who account for a much greater share of public-police interactions than they do actual arrests.

In 2015, there were about 2.1 million arrests – but about 12 million police-initiated contacts5 – with women ages 16 and over.6 That means that for every woman arrested, five more women were approached by police, either in a traffic stop, street stop, or in the execution of an arrest warrant. All police-initiated contacts are non-voluntary encounters, which are more likely than resident-initiated contacts (e.g. calling the police for help) to lead to arrest, further justice involvement, and other negative outcomes. The 2015 survey shows that women made up almost half (44 percent) of all police-initiated contacts, 41 percent of traffic stops (in which they were the drivers), and 36 percent of street stops, compared with 27 percent of all arrests. This is a critical point when assessing the need for gender-responsive policing policies and practices: basing an analysis solely on arrests would grossly underestimate the non-voluntary interactions women have with police.7


Use of force: Nearly doubled for men since 1999, but more than quadrupled among women

As with women’s share of arrests, women’s share of police encounters that involve the use or threat of force has increased significantly since 1999.8 That year, women made up just 13 percent of the approximately 422,000 people who experienced use of force during a police encounter. This percentage almost doubled by 2015, when women accounted for 25 percent of all people who experienced police use of force. Moreover, the total number of people experiencing police use of force more than doubled in that time, to 985,300 in 2015. But the increase in use of force was especially dramatic among women: the number of women experiencing police use of force in 2015 was 4.5 times (353%) the number experiencing force in 1999, up from 55,181 to 250,200. Meanwhile, the number of men experiencing police use of force doubled from 366,533 in 1999 to 735,100 in 2015.


Traffic stops: Stops declined more for men, and women make up a larger share of those searched during stops

The 2015 survey shows another noteworthy shift in the numbers of men and women stopped by police while driving. While the percentages of both men and women experiencing traffic stops declined since 1999, women saw a smaller decline in stops than men did. In 2015, 10 percent of male drivers were stopped (down from 12.5 percent in 1999) compared with 7 percent of female drivers (down from 8.2 percent). These changes may sound small as percentages, but the decline in the rate among women actually masks a 378,000 increase in total number of traffic stops of women since 1999, while men were stopped 451,000 fewer times than in 1999.9

Women also make up a larger percentage of all people searched during traffic stops. In 1999, men who were stopped while driving were about four times as likely to be searched during the stop than women; by 2015, men were just twice as likely to be searched. The narrowing of this gap reflects a decline in searches of men; for women, the share of stops that resulted in search remained the same, at 2.3 percent. In 1999, police searched the driver or vehicle in 9.4 percent of stops of men, which amounts to over one million stops involving a search. In 2015, however, the percentage of stops involving male drivers that resulted in search or arrest dropped by half, to 4.7 percent – or over 500,000 stops. Female drivers, however, saw no change in their likelihood of search or arrest during a traffic stop.


Part 2: Race, Ethnicity, and Gender in Police Stops

Data from the Police-Public Contact Survey (PPCS) allow us to take a more intersectional view of women’s contact with police, going beyond gender differences to include racial and ethnic differences among men and women, too. Again, this is the only national data we are aware of that enables any intersectional analysis of civilian experiences with police, and the Bureau of Justice Statistics (BJS) does not attempt this analysis in its report on the survey’s results. That arrest data indicate racial disparities is well documented, but little is known – outside of individual stories – about how Black, Latinx, and white women experience police contact differently from each other and from their male counterparts.

Our analysis examines how race and gender, together, affect police-initiated stops, including arrests and use of force that occur during stops. It’s worth noting that police stops and arrests during these stops are relatively infrequent events, and that our findings are based on survey data from over 90,000 respondents, not the total U.S. population. That said, we find that the race, ethnicity, and gender affect policing outcomes differently depending on the context:

  • Both race and gender affect the likelihood of a traffic stop.
  • Race seems to matter more for men when it comes to street stops and more for women when it comes to arrests during a stop.
  • Racial disparities are most apparent in use of force during a police-initiated stop, with Black and Latino men experiencing use of force more often than other groups, and Black women reporting similar use of force rates to white men.


Traffic stops: Black women are more likely than white or Latina women to be stopped

Both race and gender affect a person’s chances of being stopped by police while driving (excluding accidents or when they were passengers). Most broadly, the PPCS data shows that women were less likely than men to be stopped, and Black drivers were more likely to be stopped than white and Latinx drivers.10 More specifically, Black women were about 17 percent more likely to be in a police-initiated traffic stop than white women, and 34 percent more likely to be stopped than Latina women. Among men, Black drivers were about 12 percent more likely than white drivers – and 17 percent more likely than Latino drivers – to be stopped.

Percentage of drivers who experienced a vehicle stop in 2015

Women White 7.1%
Black 8.3%
Latina 5.5%
Men White 9.9%
Black 11.1%
Latino 9.2%
Table 2. See the Appendix for tables that show the statistical significance of differences between Black and Latina women compared to white women and all groups compared to white men.

These estimates are not adjusted for driving behavior, however, and given reported differences in driving patterns between demographic groups, it is likely that traffic stops of Black and Latina women per mile driven could be even higher compared to other groups than the survey data suggest.11


Street stops: Significant racial disparities among men, but not women

In contrast to traffic stops, the survey data on police-initiated street stops only showed significant racial disparities among men – not among women. Less than one percent of white, Black, and Latina women surveyed experienced a police-initiated street stop in 2015. However, the percentage of Black men who experienced a street stop (2.2 percent) was double that of white men (1 percent). A slightly greater portion of Latino men (1.2 percent) reported experiencing a street stop than white men, but this difference was not statistically significant.

Percentage who experienced a street stop in 2015

Women White 0.7%
Black 0.7%
Latina 0.5%
Men White 1.0%
Black 2.2%
Latino 1.2%
Table 3. See the Appendix for tables that show the statistical significance of differences between Black and Latina women compared to white women and all groups compared to white men.


Arrests during stops: Significant racial disparities among women, but not men

Police-initiated traffic and street stops sometimes result in arrest, and the survey data show that Black women were at least as likely as white men to be arrested during a stop. White women, meanwhile, were about half as likely as white men to be arrested during a stop. Black women were arrested in 4.4 percent of police-initiated stops, which was roughly three times as often as white women (1.5 percent), and twice as often as Latinas (2.2 percent).

Among men, racial disparities in arrest-during-stop rates appear to be more related to the frequency of being stopped than the likelihood of being arrested if stopped. For women, on the other hand, racial disparities seem to be more related to what happens during the stop than whether they are stopped at all.

Percentage who were arrested during a traffic or street stop in 2015

Women White 1.5%
Black 4.4%
Latina 2.2%
Men White 2.7%
Black 3.5%
Latino 4.2%
Table 4. See the Appendix for tables that show the statistical significance of differences between Black and Latina women compared to white women and all groups compared to white men.


Use of force during stops: Rates of Black women similar to white men; Black and Latino men most likely to experience force

About one percent of people surveyed indicated that they had experienced force or threat of force12 during a police-initiated stop, but use of force rates were higher for Black women than white or Latina women,13 and were highest among Black and Latino men. Black women actually experienced use of force during a stop about the same rate as white men, while white women were significantly less likely to experience use of force than white men. The marked gender disparity in reported use of force that is included in the BJS report Contacts Between Police and the Public, 2015 (2.7 percent for men versus 0.9 percent for women) masks the reality that these percentages are averages of racial and ethnic groups that are experiencing force at markedly different rates.

Percentage who experienced force during a police-initiated stop

Women White 0.3%
Black 0.9%
Latina 0.32%
Men White 0.8%
Black 3.6%
Latino 2.5%
Table 5. See the Appendix for tables that show the statistical significance of differences between Black and Latina women compared to white women and all groups compared to white men.



As women become a more visible presence in our criminal justice system, it becomes increasingly urgent that we understand their experiences within it, both to better meet their needs and to enhance our analysis of how justice works (and doesn’t work) in the U.S. The policing of women, especially women of color, has received less attention than the policing of men; it’s even received less attention than the incarceration of women. Similarly, while correctional facilities are increasingly adopting gender-responsive policies and programs, there have been virtually no concerted efforts to create or implement trauma-informed or gender-responsive policing practices.

This oversight has serious consequences. Rates of sexual and physical abuse are high among justice-involved women – estimates range from half to over 90 percent – and are much higher than reported for men. Almost a third (31 percent) of women in jail have a current serious mental illness, which is over twice the rate among men in jail (14.5 percent) and over six times the rate among women in the general population (4.9 percent). With 12 million women per year experiencing police-initiated contacts – many of which involve searches, use of force, and other traumatizing experiences – it is critical that law enforcement take seriously the need for more female police officers,14 protocols, and trainings that can improve police-public interactions and reduce the harms to women.

Finally, the invisibility of Black women and other women of color in the national discourse about policing – even in the wake of high-profile tragedies like the arrest and jail death of Sandra Bland – means that the full scope of racial discrimination in policing is unknown, and certainly understated. Making policing more transparent, accountable, effective, and just will mean bringing the experiences of these women to light.


Data sources and methodology

Arrest data come from the FBI Uniform Crime Reporting Program (reported annually in the Crime in the United States series). Other police contact data are from the Bureau of Justice Statistics’ Police-Public Contact Survey (PPCS), most recently conducted in 2015 and published in 2018. The police contact data in Part 1 comes from the BJS reports based on the results of that survey, in the Contacts Between Police and the Public series. The analysis in Part 2 used the raw PPCS data, available from the National Archive of Criminal Justice Data (NACJD) here.

Using the PPCS data for the analysis in Part 2, respondents were coded into one of eight mutually exclusive sex-race categories: Black female, Black male, Latina (female), Latino (male), white female, white male, non-Hispanic other race female, and non-Hispanic other race male (the last two categories include multi-racial respondents, and results were not included in this report because of the small sample sizes and the disparate racial and ethnic categories they included.)

Our analysis focuses on non-voluntary, police-initiated stops. Therefore, the two types of police contacts included in the analysis were “stopped by police in public place (non-vehicle)” and “stopped by police while driving.” Types of police contacts excluded from the analysis were: “respondent in vehicle stopped by police” (i.e. not driving), “in traffic accident,” “other police stops or approaches,” “crime/non-crime emergencies,” and “seeking other help from police.”

“Arrests” in our analysis of the PPCS (in Part 2) refer to only those arrests that occurred during non-vehicle (street) stops, vehicle (traffic) stops, or either type of stop. Use of force refers to whether, during the encounter, the police officer did any of the following to the respondent: push, grab, kick, hit, spray with chemical or pepper spray, use an electroshock weapon (Taser), or point a gun at them.

We used Wald tests to determine whether there were statistically significant differences in average (mean) incidence of stops, arrests, and use of force between each sex-race category in comparison to two references categories: white women and white men. We used the weighting provided by the Bureau of Justice Statistics to approximate a nationally representative sample.

Acknowledgements: This briefing was a collaborative effort, but Wendy Sawyer served as the editor and created the graphics. We thank Public Welfare Foundation for their support of our research into the experiences of women who come into contact with the criminal justice system.


  1. The terms “women” and “girls” in the introductory and concluding sections of this briefing are inclusive of transgender women and girls. However, the government data we reference (both the FBI’s arrest data and the Bureau of Justice Statistics police contact data) do not address how transgender people are classified. The FBI data are reported by police agencies around the country, and therefore likely vary in terms of how transgender people are classified by different agencies. The BJS Police-Public Contact Survey (as a supplement to the National Crime Victimization Survey) asks respondents to identify as either male or female.  ↩

  2. While the public report series Contacts Between Police and the Public does not provide breakdowns by sex and race/ethnicity, this information is available in the raw data from the 2015 Police-Public Contact Series.  ↩

  3. These percentages were calculated using estimates from the Bureau of Justice Statistics Arrest Data Analysis Tool, which provides estimates by sex from 1980 to 2014 (as of May 2019).  ↩

  4. The increase in the number of drug arrests from 2013 to 2017 was also higher for women than for men: There were 54,738 more arrests of women for drug violations in 2017 than in 2013, versus an increase of 48,887 arrests of men for drug violations over the same time period. See Crime in the United States 2017 table 35 for these five-year arrest trends.  ↩

  5. Police-initiated contacts include those where police approached or stopped respondents, such as being pulled over while driving (an example of a traffic stop) or being stopped by police while in a public place (an example of a street stop). Other types of contact excluded from this analysis include resident-initiated contact (such as calling the police for help or to report a crime) and contact resulting from traffic accidents.  ↩

  6. BJS indicated in the methodology that people involved in police-initiated contacts might have been somewhat less likely to respond to the survey than those without police-initiated contacts, which may underestimate this contact.  ↩

  7. Since women’s share of police-initiated stops has remained fairly flat since 1999 but their share of arrests has increased, we looked into whether those stops are more likely to result in arrest than they were twenty years ago. We calculated the ratio of arrests to people experiencing police-initiated contact for both women and men age 16 or older in 1999 and 2015, and found that while that ratio has fallen for both sexes, it’s decreased more dramatically for men. In 1999, among men, there were 0.69 arrests for each man whose only police contact that year was police-initiated; among women, there was 0.24 arrests for each woman who only experienced police-initiated contact. By 2015, the ratio among men had fallen by 60 percent to 0.28 arrests for each man with any police-initiated contact, but the ratio for women fell by only 29 percent to 0.17 arrest per woman with police-initiated contact. (Methodology notes: The measures are slightly different in these two surveys, with the 1999 measure being “police initiated contact was only contact” and the 2015 measure including any police-initiated contact during the year. We calculated the ratios using data from the 1999 Contacts Between Police and the Public spreadsheet cpp99f2.csv, estimates of women and men experiencing any contact with police in table 1 of the 1999 Contacts report, Table 1 in the 2015 Contacts report, the BJS Arrest Tool for 1999 arrest estimates, and Crime in the United States 2015 tables 39 and 40.)  ↩

  8. 1999 is the first year BJS published a Contacts Between Police and the Public report that is comparable to later reports. A pilot test of the PPCS was conducted in 1996 but it used a smaller sample size and a less detailed set of questions. We did not attempt to compare those results to later years.  ↩

  9. BJS estimates that there were more women drivers in 2015, which would explain the increase in the total number of stops of women drivers.  ↩

  10. It should also be noted that the survey uses U.S. Census definitions of race and ethnicity: Black, Latinx, and white are mutually exclusive categories.  ↩

  11. The U.S. Department of Labor data indicate that car ownership varies by race and ethnicity, with 90 percent of white, 74 percent of African-American and 83 percent of Latinx households owning at least one car in 2015. In addition, a National Highway Traffic Safety report indicates that, on average, men drive a third more miles per year than women (13,393 miles versus 8,854 in 2017), and that higher incomes are associated with more driving. Therefore, analyzing vehicle stops on a per-mile basis could show very different patterns. For example, if Black and Latina women drive the fewest miles due to factors of race, gender and income, their vehicle stops per mile driven could be significantly higher compared to other groups than the PPCS data suggest.  ↩

  12. Use of force was defined as a push/grab/hit/kick, use of pepper spray or electroshock weapon (Taser) or having a gun pointed at them. The BJS Contacts report included handcuffing, but because that can be standard protocol during arrests, it was not included here.  ↩

  13. 0.8 percent of Black women surveyed reported use or threat of force during a police-initiated stop, compared to 0.3 percent of white women and 0.32 percent of Latina women. However, this difference was not statistically significant.  ↩

  14. Female police officers are less likely to use excessive force and their share of law enforcement officers has barely moved from 10.7 percent in 1999 to 12.5 percent in 2017.  ↩



Appendix Table 1 provides the data behind the graph titled “Women experiencing police use of force rose dramatically between 1999 and 2015.”

Use of force has increased overall, but most dramatically among women
1999 1999 2015 2015 1999 to 2015
Percentage of all people who experienced use of force during a police encounter Estimated number of people that experienced use of force during a police encounter Percentage of all people who experienced force during a police encounter Estimated number of people experienced use of force during a police encounter Percent increase in number of people experiencing police use of force
Men 87% 366,533 75% 735,100 100%
Women 13% 55,181 25% 250,200 353%
Total 100% 421,714 100% 985,300 137%
Appendix Table 1.

Expanded tables from Part 2

Reading these tables: There are two reference groups for looking at whether differences are statistically significant. First, we look at whether the differences in contact for Black and Latina women compared to white women are statistically significant. We then look at all sex-race groups compared to white men. The level of statistical significance is indicated for each difference, for example, “p<0.1%" indicates that a difference is statistically significant at the 90% confidence level. Where there isn’t a comparison made (for example, white women are not compared to themselves), the cell is marked with “--” and when there is no statistically significant difference between groups, this is indicated simply with “No.”

Percentage of drivers who experienced a vehicle stop in 2015
Percentage who experienced a vehicle stop Is difference from white women statistically significant? Is difference from white men statistically significant?
Women White 7.1% Yes (p<.01)
Black 8.3% Yes (p<.05) Yes (p<.01)
Latina 5.5% Yes (p<.01) Yes (p<.01)
Men White 9.9%
Black 11.1% Yes (p<.1)
Latino 9.2% No
Appendix Table 2.
Percentage who experienced a street stop in 2015
Percentage who experienced a street stop Is difference from white women statistically significant? Is difference from white men statistically significant?
Women White 0.7% N/A Yes (p<.01)
Black 0.7% No Yes (p<.1)
Latina 0.5% No Yes (p<.01)
Men White 1.0%
Black 2.2% Yes (p<.01)
Latino 1.2% No
Appendix Table 3.
Percentage who were arrested during a stop in 2015
Percentage who were arrested during a stop Is difference from white women statistically significant? Is difference from white men statistically significant?
Women White 1.5% Yes (p<.01)
Black 4.4% Yes (p<.05) No
Latina 2.2% No No
Men White 2.7%
Black 3.5% No
Latino 4.2% No
Appendix Table 4.
Percentage who experienced force during a police-initiated stop in 2015
Percentage who experienced force during a stop Is difference from white women statistically significant? Is difference from white men statistically significant?
Women White 0.3% Yes (p<.05)
Black 0.9% No No
Latina 0.32% No No
Men White 0.8%
Black 3.6% Yes (p<.01)
Latino 2.5% Yes (p<.05)
Appendix Table 5.

Stay Informed

Get the latest updates:

Tweet this page Donate