For Indigenous women, systemic racial bias in prison leaves many worse off than men

Open this photo in gallery:

Before going to prison in 2017, Patricia Whyte never thought much about her Mi’kmaw heritage.

She’d been convicted of several charges, including a break-and-enter – the culmination of a series of tragic events that began when she left her abusive husband and, a few months later, lost custody of her two small children.

When she arrived at the Edmonton Institution for Women to serve out her three-and-a-half-year sentence, she noticed something strange: Many of the women inside the federal prison were Indigenous. She also started to notice the harsher treatment they received than their white counterparts.

Before things began to unravel, Ms. Whyte had lived what she considered an uneventful life, with a clean record and no police interactions. But her correctional files don’t read that way. “The way my paperwork portrays me, it makes me look like a monster,” says Ms. Whyte, now 36. “They can put anything in your paperwork, and you have no say in it.”

Although she had no history of violence, her security level was twice overridden by corrections officers, from minimum to medium – a switch that affected the kinds of programs she could access, the type of living quarters she’d be assigned to and, most crucially, her odds of getting paroled. “If you’re white in there, man, I’m telling you, you get treated way differently,” says Ms. Whyte, who now lives in Halifax. “As an Indigenous woman, I think I fared a lot worse.”

The results of a two-year Globe and Mail investigation support Ms. Whyte’s claims.

In late October, The Globe found that standardized tests used by Correctional Service Canada (CSC) to determine inmates’ security classifications, access to treatment programs and services, and their odds of getting paroled were respectively biased against Black and Indigenous men, after controlling for variables such as age, race and the severity of an inmate’s offence.

The findings were so stark that the House of Commons public safety committee has since announced a study of the federal prison agency’s risk assessments.

For Indigenous women, the picture is even worse.

Using a trove of CSC data from 2012 to 2018, a statistical analysis – which controlled for factors such as age, race, severity of offence, the year the data was captured and whether the inmate was on a life sentence – found Indigenous women were roughly 64 per cent more likely than white women to end up with the worst security score when they first arrived in federal prison, and 40 per cent more likely to end up with the poorest reintegration score at any point during their sentence.

Open this photo in gallery:

A slew of organizations and studies have previously expressed concerns about how applicable CSC’s risk assessments are to women, and specifically to Indigenous ones. Several have empirically demonstrated systemic bias, both for women in general and Indigenous women in particular, owing largely to the fact that some of the assessment tools were designed using datasets consisting almost entirely of male inmates.

Despite those concerns and CSC’s own internal research on women’s risk – and the fact that it has known about the issues since at least 2003 – experts say the prison agency has not done enough to address them. Many of the tools it uses have remained unchanged for nearly 20 years, despite CSC’s own promises it would develop new ones.

When presented with The Globe’s findings, CSC said it regularly conducts research to ensure its risk assessment tools are still reliable and that the agency was developing a plan to address the disparity in the number of racialized inmates. “We recognize that Indigenous and Black Canadians are overrepresented in our correctional institutions,” said spokesperson Esther Mailhot in an e-mailed statement. Ms. Mailhot also said the agency was “engaging university partners” and Indigenous communities to develop risk assessment tools for its Indigenous inmates.

Mary-Liz Power, a spokesperson for Public Safety Canada, which oversees CSC, said the agency “must and will do better” to address the systemic issues in the correctional system. “Our work across government, including in corrections, is and must be guided by evidence,” she wrote in an e-mailed statement. ”Practices or evaluations that systemically discriminate or disadvantage certain groups must be eliminated.”

CSC’s risk tools are, by design, extremely simple: The Custody Rating Scale, for instance, used to determine an inmate’s first security classification, consists of 12 questions relating to things such as a person’s history of substance abuse, the severity of their charges and their lifestyle before they arrived in prison. Other tests, meanwhile, rely on a parole officer’s subjective evaluation of an inmate’s criminal history and capacity for change.

CSC incarcerates very few women overall: a little under 5 per cent of the inmates in custody in any given year, or just 587 in 2018. (That figure was up nearly 21 per cent between 2012 and 2018, while the CSC’s male population remained roughly the same.) But those with an Indigenous background are vastly overrepresented – even more so than among male inmates. In 2018, nearly 39 per cent of federally incarcerated women were Indigenous, although Indigenous people overall accounted for close to 5 per cent of the Canadian population in 2016, according to the most recent census.

While many of the same assessment tools are used on both male and female inmates, women generally turn to crime for different reasons than men, which means their risk factors are vastly different, too, according to Senator Kim Pate. Before joining the Senate in 2016, Ms. Pate spent more than 20 years as the executive director of the Canadian Association of Elizabeth Fry Societies, a prison advocacy group focused on women and gender-diverse people.

Women are far less likely to be violent than male offenders. And in most cases, she says, women who commit crimes are reacting to violence that has been done to them – stabbing an abusive partner, for instance – or participating in a context where they are victims themselves.

CSC’s own data back this up: In a 2017 study, researchers found that 69 per cent of incarcerated women (and 79 per cent of Indigenous women) had been victims of spousal abuse, compared with 12 per cent of men.

Amanda Lepine, a 40-year-old Métis woman serving eight-and-a-half years for the armed robbery of a convenience store in Winnipeg, certainly fits Ms. Pate’s archetype for incarcerated women.

Her relationships have often been violent, starting with her abusive mother and into adulthood, when a partner attacked her with a hammer, she says, setting off a chain of events that led to her no longer having custody of her children. And her crimes involved a male accomplice.

Ms. Lepine began her latest sentence in the summer of 2015, at the Edmonton Institution for Women. This was her second federal stint – the first, also for robbery, ended in early 2013.

Open this photo in gallery:

For as long as she can remember, the government has had a file on her: as a toddler, when Manitoba’s Child and Family Services often visited her home; at the age of 10, when she ended up at a facility run by nuns; and right through her adolescence, during which time the authorities continued to intensely follow and document her.

The sum of her life history, recorded nearly from birth and impossible to escape, weighs on her assessment scores in prison. “My life is like an open book they can read, and they just judge me for it,” Ms. Lepine says. During one of her assessments, she recalls a parole officer asking if she’d enjoyed being physically abused by her mother as a child. “I wanted to hit her,” she says sharply. “How dare you ask me that?”

A little over 18 months into her sentence, she was involuntarily transferred to the Fraser Valley Institution in Abbotsford, B.C., owing to “ongoing security issues,” per one of her files. There, she was reclassified to maximum security. In late March, 2017, CSC data show she was one of four women at the prison to be rated a max. All were Indigenous.

Unlike men’s prisons, which are often designed for a specific security level, all women’s prisons are “multi-level,” with separate units at different classification levels. Housing for minimum-security inmates is fairly relaxed and may even be outside the fence, while maximum-rated “secure units” look much more like typical prison cells.

According to CSC’s data, of the average 62 women in maximum each year between 2012 and 2018, 52 per cent were Indigenous. In minimum, meanwhile, white women were the majority, at 56 per cent.

Share of female inmates with

a certain security score

Average each year from 2012 to 2018, by race

MAXIMUM

Indigenous

White

Black

Other

0

10

20

30

40

50%

MINIMUM

White

Indigenous

Black

Other

0

10

20

30

40

50%

THE GLOBE AND MAIl

SOURCE: Correctional Service Canada

Share of female inmates with

a certain security score

Average each year from 2012 to 2018, by race

MAXIMUM

Indigenous

White

Black

Other

0

10

20

30

40

50%

MINIMUM

White

Indigenous

Black

Other

0

10

20

30

40

50%

THE GLOBE AND MAIl

SOURCE: Correctional Service Canada

Share of female inmates with a certain security score

Average each year from 2012 to 2018, by race

MAXIMUM

MINIMUM

Indigenous

White

White

Indigenous

Black

Black

Other

Other

0

10

20

30

40

50%

0

10

20

30

40

50%

THE GLOBE AND MAIL, SOURCE: Correctional Service Canada

Female inmates’ risk scores, and particularly those of Indigenous women, have long been a point of concern for countless agencies and researchers.

Part of the concerns stem from the fact that one of CSC’s most important assessments, the Custody Rating Scale, was built from an administrative dataset of men admitted to federal custody in 1985.

In 2003, the Office of the Auditor General said the Custody Rating Scale had not been properly validated for women and ordered the agency to study its tools. Later that year, the Canadian Human Rights Commission issued a scathing report on the rights of federally incarcerated women, which included detailed sections on several CSC risk tools, including the security scale.

When it came to women, the Commission said many of CSC’s core risk tools were “blunt instruments that tend to lead to unjustifiable differential treatment.” The Custody Rating Scale, specifically, “was not designed to identify, reflect or accommodate the needs, capacities and circumstances of federally sentenced women or members of racialized groups, nor has it been adequately validated for these populations,” the Commission wrote.

The critiques continued: In its 2004-05 annual report, the Office of the Correctional Investigator came to similar conclusions, recommending CSC create a strategy for assigning more appropriate security classifications to Indigenous women.

At the time, CSC promised to develop a new initial classification tool for women – but it never came. Despite studies in 2012 and 2014 examining women’s security scores, the agency still uses the same Custody Rating Scale for both men and women as it did nearly two decades ago.

Another report, also from 2004, called into question the very essence of the tool: its ability to predict risk. Using CSC data, a pair of criminologists examined the predictive validity of the Custody Rating Scale when it came to women. They found that not only was the scale failing to predict what it was supposed to, but it actually seemed to be systemically biased against Indigenous women.

“All we did was look at the components,” says Dr. Cheryl Webster, a criminology professor at the University of Ottawa. Dr. Webster found many of the tool’s questions could not predict institutional incidents or escapes with any statistical significance, and that when they did, the correlations were sometimes near zero.

“Our bottom line: the Custody Rating Scale has little to no predictive validity,” Dr. Webster says. As she and her research partner, Dr. Anthony Doob, noted in the study, you might as well be classifying inmates based on the size of their big toe.

At the time, Dr. Webster assumed CSC would be prompted to make changes to its assessment. “I was wrong,” she says.

Dr. Doob, a University of Toronto criminologist, puts it more bluntly. “The institution’s being dishonest,” he says. “In 2014 they knew it, in 2004 they knew it, they know it now, and they’re still classifying people by criteria they know not to be predictive of what they’re trying to predict. That’s a pretty serious charge.”

What’s ironic, says Dr. Doob, is that the evidence supporting his and Dr. Webster’s conclusions came from CSC itself. “It’s their data,” he says. “I would put the responsibility at the top. I think the difficulty is that the institution itself does not value evidence. That really has to be the responsibility of senior management, including the Commissioner.”

Despite Dr. Webster and Dr. Doob’s findings (as well as concerns from Ms. Pate, the Canadian Human Rights Commission, the Office of the Auditor General and the Office of the Correctional Investigator), CSC spokesperson Ms. Mailhot said the agency’s own studies “continue to demonstrate that CSC classification tools are valid for Indigenous and Black offenders.”

In its response, the agency repeatedly pointed to a different assessment, the Security Reclassification Scale for Women. However, that scale is used only to reclassify women once they’ve already received a score and was not a focus of The Globe’s investigation.

What a Custody Rating Scale looks like

When asked why the Custody Rating Scale remains unchanged 17 years after CSC told the Canadian Human Rights Commission it would build a new tool for women, CSC said that while it had developed a prototype by 2010 for the initial security classification of women, the tool was shelved after the agency “determined that there was a significant overlap” between the Custody Rating Scale and the new assessment, and that “gender-specific items were not evident” in the new scale.

Concerns about CSC’s risk tools also exist within the agency itself. In a 2017 hearing of the House of Commons public safety committee, Audra Andrews, a parole officer working in a women’s unit, said she believed CSC’s security classifications were biased against Indigenous people. “Their security rating is going to come out higher all the time,” she told the committee. “Very rarely did I have an offender come out as minimum. Rarely.”

In the late 1990s, Dr. Shelley Brown, a psychology professor at Carleton University, says she realized risk assessments needed to be revamped for race and gender. “We really do need to rethink how we ask questions, how we score items, how we make sure certain items are fairly scored.”

Many of the risk factors that crop up for women aren’t necessarily predictive of future crime, she says, but shadows of past victimization. Women’s paths to the justice system are also very different than men’s, she says, characterized by childhood adversity, extreme poverty, neglect in the home, substance abuse and sex work as a survival mechanism.

“Gosh, they should not go into our risk assessments, right?” says Dr. Brown, who worked at the CSC as a researcher from 1997 to 2006. “You should not be penalized for those things.”

The adversities particular to the Indigenous woman’s experience might go some way to explaining why they reoffend in CSC’s data at higher rates than white women. According to that data, Indigenous women reoffended 9 per cent of the time, compared with 4 per cent of white women and roughly 2 per cent of women belonging to other racial groups.

After controlling for factors such as race, age, and the severity of their offence, Indigenous women are 40 per cent more likely to receive the worst reintegration score – which plays into parole decisions and is designed to measure an inmate’s readiness to re-enter society.

The Globe’s analysis also found that their odds of reoffending are higher. When controlling for race and that same reintegration score, Indigenous women are 61 per cent more likely than white women to reoffend, although Dr. Brown cautions that there simply isn’t yet enough research to explain why Indigenous women are more likely to end up back in a federal institution.

Dr. Brown is still a firm believer in employing risk assessments, however, especially compared with the subjective biases an officer might introduce if the tools didn’t exist. “It’s better than shooting in the dark,” she says. If new tools for women are to be developed from the ground up, Dr. Brown hopes the process will involve Indigenous people. “Something has to be done, and it’s not something that us white researchers alone can sort out,” she says.

Ms. Whyte had to learn about the problems with CSC’s assessment process first-hand. “As soon as you go in there, the other inmates tell you right from the get-go: ‘Don’t tell them nothing,’” she says. “Because the more you tell them, the more hoops you’ll have to jump through.’”

Since being released, Ms. Whyte has turned her prison experience on its head: She is currently attending Mount Saint Vincent University with an eye to becoming a counsellor and serves as a peer support worker at the Elizabeth Fry Society of Nova Scotia, advising female inmates. “I’ve come so far,” she says, “and I’m glad I went through the experiences I did.”

Open this photo in gallery:

As for Ms. Lepine, she still has a way to go on her sentence, set to expire in 2024. She was recently granted a pass to attend a treatment facility, and is hopeful she’ll succeed at her next parole hearing, likely early next year. This will be her second attempt. When she was out on day parole in 2019, Ms. Lepine decided to head home for a grandfather’s funeral, she says, and didn’t check back in. While at large, she worked at a school on a reserve, met someone and got pregnant. Police caught up with her six months later. By the end of 2019 she was back at Fraser Valley, where she gave birth. Ms. Lepine now lives in a special housing unit where she can care for her baby.

Although she and the father are no longer together, their connection has nonetheless increased her risk profile; earlier this year, he was criminally charged himself, which correctional files said elevated Ms. Lepine’s escape risk.

Once she gets out, Ms. Lepine plans to forge a career working with at-risk youth, hoping to save them from her fate. In the meantime, the long shadow of her past still works its way into her assessments.

“They say all of that affects me, and that’s why I am the way I am,” she says. “Well, yeah, it is, but I’m healing from that, and it doesn’t define me. It’s not my fault that I went through these things. I didn’t ask for my parents to be my parents. I had no choice in it. I did what I did to survive, you know? But they don’t see it that way.

“They constantly want us to keep looking at it, reliving it, being stuck in it,” she adds. “I’ve worked on that. Let me move on to something else now.”

Bias behind bars: A Globe investigation

Federal inmates' risk assessments determine everything from where a prisoner is incarcerated to what rehabilitation programs they are offered. After controlling for a number of variables, The Globe found Black and Indigenous inmates are more likely to get worse scores than white inmates, based solely on their race

Read more