Recently, an alarmingly high incidence of wrongful convictions has been documented in the United States, in large part because of "Innocence Projects" that use DNA analyses from crime scenes to exonerate innocent persons. The best-known Innocence Project, administered through the Benjamin N. Cardozo School of Law in New York, has helped to free 138 people who had been wrongfully incarcerated. Approximately 25 percent of these cases involved false confessions arising from inappropriate police interrogations. Among these false confession cases, persons with mental impairment appear to be disproportionately represented. The Innocence Project's Web site notes, "Truly startling is the number of false confession cases involving the mentally impaired and the mentally ill. Police interrogation in the [false confession] cases reveals a lack of training and a disregard for mental disabilities" (+1).

An example of where this situation can lead is the case of Eddie Joe Lloyd, who spent 17 years in prison before being exculpated by DNA analysis (+2). In January 1984, Lloyd, who had been given a diagnosis of paranoid schizophrenia, took an interest in the case of a 16-year-old girl who was found brutally raped and murdered in Detroit. Because of his interest, the police came to the mental hospital where Lloyd was admitted and interrogated him on three separate occasions. Lloyd was led into falsely confessing by police who fed him facts of the crime—his knowledge of which was later used against him as evidence of his familiarity with the attack—and who told him that by confessing to the crime, he would help the police "smoke out" the real perpetrator. Perhaps because of his mental illness, Lloyd believed the officers and claimed he murdered the girl. Currently, the police officers involved are under federal investigation for their alleged misconduct. The actual murderer was never apprehended, and Lloyd spent 17 years of his life wrongly imprisoned.

The number of persons with mental illness who encounter the police as suspects is not inconsequential. Indeed, one study found that the probability of arrest was 67 times greater for persons who demonstrate symptoms of mental illness compared with those without such symptoms (+3). Although mentally ill persons are most commonly cited for misdemeanors, there is a subgroup of these persons who are suspected of having committed violent crimes, sometimes correctly (+4), and are thus subject to formal police interrogations. A conservative estimate that 10 percent of inmates in U.S. state and federal prisons suffer from mental illness (+5) yields a total of about 140,000 mentally ill prisoners. This number represents only adults convicted and currently incarcerated in prisons, as opposed to jails; thus the number of persons with mental illness who have been interrogated by the police in recent years can be estimated to be much greater than 140,000.

Little to no discussion has taken place in any of the relevant literatures specifically about police interrogation of mentally ill detainees. Even from a legal standpoint, scant attention has been paid to the topic. For a confession to be admissible in court, it must be made voluntarily, that is, it must be noncoerced. Confessions are deemed to be coerced if the police explicitly threaten harm or punishment or promise leniency. Although research has suggested that many contemporary police interrogation tactics implicitly convey threats and promises (+6), no research has been done, and there have been few contested legal cases to determine whether mentally ill persons are more likely to perceive implicit threats and promises as explicit statements, which would be illegal for police to utter when interrogating suspects.

Confessions are also excluded from legal proceedings if it can be shown that suspects did not understand or appreciate their Miranda rights. The U.S. Supreme Court ruled in the 1966 landmark case Miranda v. Arizona (+7) that because of the inherent coercion present in police interrogation, all suspects must be made aware of their right against self-incrimination and of their right to counsel. Mentally ill defendants, particularly defendants with psychotic disorders, are significantly less likely to understand their interrogation rights than defendants who are not mentally ill (+8).

Finally, legal safeguards for persons with mental disorders afford little protection during the investigation phase. In Colorado v. Connelly (+9), a case involving a mentally ill defendant, the U.S. Supreme Court ruled that a suspect's mental condition alone is insufficient for a finding that a confession was coerced. Rather, it must be demonstrated that the police used coercive techniques. Although Connelly confessed because the voices in his head told him to do so, because the police did not overstep their bounds, the confession was considered to be the product of his free will and rational intellect. Therefore, the confession could be used against him in court. However, it has never been established empirically that the boundaries on police behavior that were put in place for suspects who are not mentally ill are sufficient to protect mentally ill suspects. Examining contemporary police interrogation techniques with an eye toward persons with mental illness only heightens the concern that such persons may not be adequately protected.

The ultimate goal of interrogating suspected criminals is to gain a confession. Establishing psychological control is a critical element toward that end. For the past 50 years in the United States police interrogative procedures have changed from the use of physical intimidation to a more sophisticated use of psychological manipulation. Leo (+10), a sociologist, in his analysis of more than 500 hours of observed interrogation, likened police questioning to a confidence game in which detectives "cultivate" and "con" suspects into admitting guilt. He wrote that "contemporary interrogation strategies are based on the manipulation and betrayal of trust" (+10).

The most widely used set of interrogation tactics are referred to as the Reid technique. The training manual based on the Reid technique, which is considered to be the "bible" of police interrogation, was originally developed in 1947 and is now in its fourth edition (+11). Police interrogators who are trained in this method are taught to assume guilt, to manipulate the suspect's emotions and expectations, and to take into account nonverbal behavioral cues, such as hesitant speech, sweating, or dry mouth, as indicators of deception. However, these cues, in addition to being general indicators of stress, may appear more frequently among persons with mental illness because of their illness or the medications they are taking.

Police interrogation approaches can be characterized as involving either minimization or maximization techniques (+6). Minimization techniques—such as feigning sympathy, offering a moral justification for the crime, or shifting blame—are used to placate accused suspects and lead them into a false sense of security. In contrast, maximization techniques—such as presenting false evidence—attempt to browbeat the suspect into confessing. Persons with a mental disorder may be more susceptible to confessing in response to both of these techniques than persons without a mental disorder. Characteristic traits of mentally ill persons, such as disorganization of thought, deficits in executive functioning and attention, and impaired decision making, could contribute to self-incrimination. For example, compared with persons without mental illness, persons with mental illness may be more likely to confess, because they believe that the police officer is truly a friend who understands and "has been there" or because they believe that they will be able to go home after confessing.

Another controversial police interrogation tactic involves lying to suspects. It is legal for the police to use "trickery and deception" during interrogations, and thus the police commonly lie to suspects about evidence they do not have (+6). For example, the police can tell suspects that their fingerprints are on the weapon used in a crime or that eyewitnesses saw them commit the act, even when no fingerprints or eyewitnesses exist. Of course, the hope is that guilty suspects confronted with such "evidence" will break down and confess, but this approach may also lead innocent suspects into falsely confessing. For numerous reasons, persons with mental illness, both guilty and innocent, may be more likely than persons without mental illness to confess in response to such tactics. For example, some mentally ill persons have deficits in social skills, such as assertiveness (+12). Three common aspects of assertiveness are asking for assistance, saying "no" to others, and providing corrective feedback. All of these aspects are relevant to the interrogative situation, and their absence may increase the likelihood of confession. Examples of assertive behaviors that some persons with mental illness may not be able to perform during an interrogation include asking for an attorney, denying commission of the crime, and telling the police officer that one is innocent when the police officer is insisting on one's guilt.

The Reid training manual (+11) provides little discussion of how to recognize and deal with suspects who have mental impairments. Although the manual asserts that mental disabilities "if actually present" in the context of active persuasion may lead to false confessions, it immediately counters this assertion by noting that persons with mental disabilities will not be skilled liars, and hence the truth is likely to reveal itself to the interrogator. However, the police generally have not had training in determining who is and who is not mentally unstable. As Lamb and colleagues (+13) pointed out, "A person who seems to be mentally ill to a mental health professional may not seem so to police officers—who despite their practical experience have not had sufficient training in dealing with this population."

In sum, it is quite possible that modern police interrogation techniques used in the United States place mentally impaired suspects at risk of false or legally coerced confessions. Indeed, because of the potential for false confession, the United Kingdom has established procedures for handling at-risk suspects. Specifically, in the United Kingdom psychologically vulnerable suspects, such as juveniles and mentally disordered or handicapped persons, may be interviewed by the police only in the presence of an "appropriate adult." The reasoning behind the mandate is that persons with mental impairment "may without knowing or wishing to do so, be particularly prone in certain circumstances to provide information which is unreliable, misleading, or self-incriminating" (+14). At this point in time the United States does not have comparable standards, despite the growing recognition of false confessions from mentally impaired suspects.

Regardless of their guilt or innocence, all suspects are afforded certain constitutional rights in the interrogation room. In 1966, the U.S. Supreme Court in Miranda v. Arizona(+7) took steps toward equalizing the "inherently coercive" atmosphere of interrogation by ensuring that all suspects were made aware of their rights before formal police questioning could ensue. On average, however, persons with severe mental disorders are unlikely to be on an equal footing with others in the interrogation situation.

A quick remedy to the problem of unfair interrogations of suspects with mental illness is not likely to be found. Clearly, additional research is needed. Among the unanswered questions are: What percentage of suspects who are interrogated have a mental illness? Among the suspects who have a mental illness and are interrogated, in what percentage of cases do the police recognize the mental illness? If the mental illness is recognized, what actions do the police take; for example, do the police determine the competency level of the suspect, call an attorney or a mental health professional or both, or simply continue with the interrogation? Are persons with mental illness more likely than persons without mental illness to waive their constitutional rights? Will innocent persons with mental illness confess to a crime more frequently than innocent persons without mental illness? A clearer understanding of the relationship between mental illness and the outcome of interrogation is necessary to advance knowledge on police handling of persons with mental illness and to prevent miscarriages of justice.

In the meantime, as part of many U.S. communities' crisis intervention training initiatives, exemplary efforts are being made to increase the police's knowledge of mental disorders and the people who suffer from them. However, efforts are focused almost exclusively on police involvement in crises that involve persons with mental disorders. Police recruits as well as experienced detectives would benefit from training on how to interrogate persons with mental illness and the potential risk of false confession. Curricula about interrogating persons with mental illness could be incorporated with relative ease into existing and developing police training programs. To date, structured curricula have not been developed to train interrogators in the questioning of persons with mental impairment or other risk factors. A few jurisdictions have sponsored one- or two-day training seminars on the topic; perhaps not surprisingly, these initiatives usually come after false confessions have been exposed. Prevention is almost always preferable to intervention, and in the case of false confessions, prevention means not having an innocent person languish in prison, not allowing the true perpetrator to go free to potentially commit more crimes, and not creating embarrassing and costly situations for the criminal justice system. With increasing awareness that people can and do admit to criminal acts that they did not commit, the number of false confession cases that are identified is likely to increase as well. Because persons with mental impairment appear to be disproportionately represented among these false confession cases, a greater understanding of the relationship between mental illness, interrogation, and confession is needed.

Dr. Redlich is a senior research associate at Policy Research Associates, Inc., 345 Delaware Avenue, Delmar, New York 12054 (e-mail, aredlich@prainc.com). Paul S. Appelbaum, M.D., is editor of this column.