A gurney in the execution chamber at the Oklahoma State Penitentiary in McAlester, Okla. (Sue Ogrocki/Associated Press)

In January, Dylann Roof was sentenced to death for killing nine black churchgoers at a prayer meeting in Charleston, S.C. Since then, some commentators have debated whether Roof should have had the right to fire his attorneys when they wished to introduce evidence of mental illness — and whether, or when, mental illness should disqualify someone for capital punishment.

Most Americans oppose the death penalty for the mentally ill, a category that ranges from mild to severe. But our research suggests that the death penalty actually targets those who have mental illnesses.

People who are executed have a far higher rate of mental illness than does the general public.

Suicidal tendencies are particularly common. For instance, evidence suggests that Roof had intended to shoot himself at the Emanuel African Methodist Episcopal Church after the killings. When that failed, he refused legal defense in the penalty phase of his trial. If a judge allows, he may refuse any legal appeals, thus committing suicide by execution.

Roof would be far from the first to do so. Since 1976, the United States has executed 1,448 inmates; 141 of these have been “volunteers,” those who waive appeal. If suicidal tendencies are evidence of mental illness, then death penalty states actively assist suicide.

Several states are considering banning the execution of people with severe mental illnesses or brain injuries. In March 2016, the state of Texas executed Adam Kelly Ward, whose bipolar disorder had been diagnosed when he was 4, for shooting a city inspector, whom he apparently believed was threatening his family.

As Shaila Dewan of the New York Times reported, “The U.S. Supreme Court has previously barred the execution of inmates who are so mentally impaired that they do not comprehend that they are going to be executed, but it has stopped short of outright banning the execution of the mentally ill.”

In considering the boundaries of the death penalty, one of the Court’s most consistent problems has been how to treat the mentally ill.

We researched mental illness among those sentenced to death

In a forthcoming book we review the prevalence of mental illness among those executed from 2000 through 2015. We used strict criteria, listing an inmate as having an illness only if we found credible source material showing a diagnosis of mental illness. We used only evidence or testimony presented at trial or repeated in the news. This probably underestimated the prevalence of mental illness — necessarily leaving out those who are undiagnosed, or for whom a diagnosis was not presented at trial or in the sources we reviewed.

Even so, we found that 43 percent of inmates executed between 2000 and 2015 had received a mental illness diagnosis at some point in their lives.

The graph below compares the prevalence of the different diagnoses of these executed inmates with that in the general public, as reported by the National Institute of Mental Health (NIMH). The NIMH distinguishes between “serious mental illness” and “any mental illness” in each category for the general public. A serious mental illness, according to NIMH, results “in serious functional impairment, which substantially interferes with or limits one or more major life activities.” We had no way of distinguishing between these two categories for death row inmates, so we have included them both. However, the “serious” category is most likely more accurate as a point of reference.

As you can see, every category of serious mental illness is significantly higher among the executed inmates.

Note that 20 percent of the executed inmates were diagnosed with a personality disorder; 8.9 percent were diagnosed with antisocial personality disorder. Although sometimes associated with violent behavior or referred to as “sociopaths” or “psychopaths,” these are nonetheless debilitating mental illnesses, accompanied by patterns of brain dysfunction and impairment.

For instance, Ernest West Basden was diagnosed with three personality disorders. According to North Carolina Supreme Court documents related to the direct review of the death sentence:

Dr. J. Don Everhart, a clinical psychologist, testified that defendant has a dependent personality disorder; he is lacking in self-confidence and clings to stronger people, performing unpleasant tasks for them to retain their support. Dr. Everhart further testified that defendant has an avoidance personality disorder; he is shy and uncomfortable in social settings and is easily isolated. Finally, defendant has a schitzotypal personality disorder, with feelings of being disembodied and disassociated from life events.

In 2002, the State of North Carolina executed Basden.

Depression and suicidal tendencies among those sentenced to death

When John H. Blume studied death row volunteers from 1976 through 2003, he found that 88 percent had a mental illness or substance abuse disorder. Our numbers were slightly lower, but similar. Among the volunteers executed from 2000 to 2015, 32.2 percent had attempted suicide and failed before being executed. Roof has been diagnosed with depression, and had planned suicide. The figure below shows rates of depression, suicidal tendencies, suicide attempts and mental illness among executed inmates, comparing those who were “volunteers” with those who were not. Various forms of mental illness are common in both groups, but particularly so among the volunteers.

Childhood trauma and abuse among those sentenced to death

Trauma, while not a mental illness, is a risk factor for mental illness. Certainly, a traumatic childhood should not be treated as the equivalent of an impairing mental illness. Nevertheless, it’s noteworthy that death row inmates are much more likely than most Americans to have suffered trauma during childhood.

According to the Department of Health and Human Services, about one in 10 U.S. children are abused or neglected. Among the executed inmates, we found evidence that 39.7 percent were abused during childhood. The CDC and independent researchers have repeatedly found that childhood trauma’s long-term effects include higher likelihoods of disrupted neuro-development, cognitive impairment, mental illness, and becoming the perpetrator or victim of violence.

In another study reviewing documents from 145 capital trials in North Carolina between 1999 and 2009, 77.9 percent of defendants provided evidence of at least one adverse childhood experience (ACE). ACEs include comparatively milder traumatic experiences such as neglect, witnessing abuse, family divorce, substance abuse or bullying. According to our research, death row inmates were far more likely to have suffered some of the more traumatic ACEs, such as sexual abuse.

Roof had such a background of childhood difficulties.

The graph below shows the prevalence of childhood abuse among the executed inmates compared to government estimates of the prevalence of these same types of abuse in the general public. Although childhood abuse cannot be considered a defense against criminal behavior, it’s notable that the death penalty disproportionately targets those who have experienced severe childhood trauma.

The Supreme Court has declared that, under the Eighth Amendment’s ban on cruel and unusual punishment, the United States can no longer execute juveniles, those with intellectual deficiencies, or those who are insane as defined by “whether the prisoner is aware of his impending execution and of the reason for it.” Current debates are considering whether those bans should extend to people with severe mental illness. While considering the question, Americans may wish to be aware that, statistically speaking, our current death penalty system does appear to target precisely those people.

Frank R. Baumgartner is the Richard J. Richardson distinguished professor of political science at the University of North Carolina at Chapel Hill, and co-author of the forthcoming book “Deadly Justice: A Statistical Portrait of the Death Penalty” (Oxford University Press, 2017).

Betsy Neill will graduate from UNC at Chapel Hill in May with a bachelor’s degree in psychology and political science. She co-wrote the chapter in “Deadly Justice” on mental illness.