Dr. Watson is affiliated with the Jane Addams College of Social Work, University of Illinois at Chicago, 1040 W. Harrison, MC309, Chicago, IL 60607 (e-mail: acwatson@gmail.com). Dr. Angell is with the School of Social Service Administration, University of Chicago.

The police have historically played a pivotal role in responding to persons with mental illness in the community, particularly in poorer neighborhoods (1). With the movement of persons with mental illness out of hospitals and into the community, the frequency of contact between police officers and persons with mental illness, in crisis or otherwise, has increased significantly, with departments reporting that, on average, 10% of their contacts with the public involve persons with mental illness (2). In these situations, police officers become the gatekeepers of the criminal justice and mental health systems (3), and their responses have important implications in terms of whether individuals receive treatment, remain in their current situation, or face the problems inherent in a criminal justice system that is ill prepared to meet their needs (4). The manner in which officers approach these situations may also determine whether the person cooperates (5) and whether a crisis escalates to violence or is resolved without force or injury (6).

Police officers report that contacts with persons with mental illness are problematic and that the mental health system is often less than helpful in resolving such encounters (7,8). Many officers have also indicated that they do not feel adequately trained or that they would like additional training in responding to persons with mental illness (6,8).

Efforts to improve police officers' abilities to respond to persons with mental illness are being initiated in jurisdictions across the country. These efforts include crisis intervention teams and educational programs presented during academy and in-service training (4). Many of these efforts have incorporated the perspectives of multiple stakeholders into planning and implementing interventions. In addition, national and state efforts are under way to establish expert consensus on essential elements of specialized police-based programs, such as crisis intervention teams (9). Although these efforts may not be programs in search of a theory, they would nonetheless benefit from a well-developed and explicit theoretical understanding of interactions between police officers and persons with mental illness. This knowledge may help isolate the truly essential components of effective police response that can be disseminated alone or as components of more extensive programs, such as crisis intervention teams. Theory can also guide research on the effectiveness of these efforts and identify conditions that may influence outcomes for officers, persons with mental illness, law enforcement organizations, and communities.

Research on the interactions of persons with mental illness with other parts of the justice system—for example, mental health courts and civil commitment proceedings—provides a useful point of departure. Applying procedural justice theory, these studies have found that when individuals evaluate a legal interaction as being high in procedural justice, they report feeling less coerced and are more likely to cooperate with authorities (10). Procedural justice and perceived coercion are likely to influence experiences and outcomes in street-level interactions between police officers and persons with mental illness. In this article, we briefly review the literature on encounters between the police and persons with mental illness. We then discuss procedural justice theory and apply this lens to encounters between police and persons with mental illness. We conclude with a discussion of the implications for police training and future research.

Although they vary in method and measures, a number of studies published over the past 25 years suggest that both situational and officer characteristics play a role in determining police officers' response to persons with mental illness. These studies suggest that police are more likely to arrest individuals with mental illness when there is evidence of a crime, when the individual has a criminal history (11), when they feel that the individual would be inadmissible to a hospital or other caretaking systems, when public encounters exceed the community's tolerance for deviant behavior, and when it is likely that the person will continue to cause a problem (12). Also, less experienced officers are more likely than more experienced officers to arrest persons with mental illness. Officers are least likely to take formal action when there is no evidence of a crime and the person is homeless. More experienced officers are the most likely to take no action, informal or formal (11).

Several recent surveys of police officers have examined attitudes about and perceptions of persons with mental illness. These studies found that officers do not strongly endorse negative attitudes about mental illness (6,8,13). However, they suggest that officer and organizational factors influence officers' perceptions of persons with mental illness. Younger, white, and less trained officers tend to perceive more danger, whereas officers from departments with a community policing orientation and those with more prior contact with persons with mental illness have more positive perceptions (14).

A common theme reported in these surveys is that officers do not feel qualified or adequately trained to handle calls involving persons with mental illness (6,8). Perhaps this lack of adequate training explains Ruiz and Miller's (6) finding that a significant minority of officers believed that it is best to handle mental health calls quickly. As the authors suggested, responding quickly can unnecessarily escalate a situation. They also found that injuries to officers and persons with mental illness were more likely when two or more officers were dispatched to the call. They suggested that forcing compliance is more easily accomplished when several officers are present, whereas lone officers may take more time in order to resolve the situation without force or injury.

Cumulatively, this body of literature suggests that individual, situational, and organizational factors influence how officers perceive persons with mental illness and how they choose to resolve contacts on the street. What these studies do not shed light on is the way in which officers initially approach and interact with persons with mental illness and how this approach influences how the interaction proceeds and the extent to which the person cooperates. As noted in a recent report by the Technical Assistance and Policy Analysis Center for Jail Diversion, it may be the first few seconds of an interaction between a police officer and person with mental illness that determines whether it is going to be a productive or a problematic situation (15). Consistent with Ruiz and Miller's (6) findings, if the first few seconds are rushed to force compliance, the interaction may quickly escalate to violence. If officers use those first few seconds to talk to the person and use verbal de-escalation skills, the interaction may be more likely to be resolved without resorting to force.

Unfortunately, interactions between police and persons with mental illness occasionally become violent confrontations. Research indicates a majority of individuals who assault police officers are under the influence of drugs or alcohol or have a psychiatric disorder (16). Unfortunately, the methodologies of such studies make it difficult to tease apart the unique contribution of mental illness to the phenomenon of assault against police officers, in that studies that examine "impairment" tend to use a combined measure (under the influence of drugs or alcohol or having a mental illness). For example, Kaminski and colleagues (17) found that individuals identified by officers as impaired by mental illness made up slightly over 13% of the impaired group. As in prior studies, they used a combined measure of impairment for the analysis of looking at impairment and police injury; thus it is not possible to determine the relative contribution of mental illness. Other research has found that when a person with mental illness is the assailant, the officer is more likely to be injured (18). Ruiz and Miller (6) suggested that several conditions may make these situations particularly volatile. These include the understandable fear on the part of persons with mental illness to put themselves in the hands of unfamiliar police officers, fear of the formal uniform, and the overpowering approach and fear on the part of police officers of persons with mental illness. As indicated above, the presence of several officers may also increase the likelihood of injury resulting from the use of force.

Across the country, law enforcement agencies are attempting to improve their response to persons with mental illness. Whereas some are simply supplementing existing training with a few additional hours on mental health issues, many are implementing one of three types of specialized responses: police-based police response, which involves specially trained police officers; police-based mental health response, which involves mental health clinicians working as civilian employees of the police department; and mental health-based mental health response, which involves partnerships with mobile mental health teams that are part of a community mental health center (7). In two studies examining the three specialized responses, crisis intervention teams, which are a police-based response, compared most favorably in terms of lower arrest rates and greater officer-perceived effectiveness (7,19).

Developed by the Memphis Police Department, the crisis intervention team model involves specially trained officers who provide first-line response to calls involving a person with mental illness and who act as liaisons to the mental health system (7). Whereas jurisdictions adapt the model to their local situations, the model maintains several essential elements believed to enhance police response—training, partnerships with mental health resources in the community, and a redefined approach to responding to mental health calls that includes changes in police officer roles and organizational priorities (4). A growing body of data supports the model's effectiveness for reducing officer and citizen injuries and arrests and increasing transports and referrals to mental health services (20).

Procedural justice provides a framework for considering how persons with mental illness experience interactions with the police and how officer behaviors may shape cooperation or resistance. We suggest that these processes may in part underlie the emerging success of crisis intervention team programs. In the context of a procedural justice framework, the focus is on the subjective experience of the process of the interaction with an authority (such as "The officer treated me fairly") rather than satisfaction with the outcome (such as "The officer should not have arrested me"). Key antecedents of procedural justice judgments include participation (having a voice), which involves having the opportunity to present one's own side of the dispute and be heard by the decision maker; dignity, which includes being treated with respect and politeness and having one's rights acknowledged by the decision maker; and trust that the authority is concerned with one's welfare (21).

Lind and Tyler (21) have proposed the group-value model of procedural justice, which suggests that people want to be treated fairly by authorities, independent of the outcome of the interaction. Fair treatment by an authority, operationalized here in terms of voice, dignity, and trust, directly shapes procedural justice judgments and signifies that the individual is a valued member of the group. This in turn facilitates cooperation by strengthening a person's ties to the social order (22).

According to this framework, people vary in terms of the degree to which they focus on procedural fairness. People who are of high status in the group are secure about their status and need not focus on how they are treated to affirm their identity. It is those who are uncertain about their status (such as members of stigmatized groups or those experiencing events that may demean their status) who will respond most strongly to the fairness by which police exercise their authority (21). Their treatment by police signifies their social status, self-worth, and self-respect (23). How police officers treat them may further marginalize them or support their identity as a member of the community.

In applying this framework to encounters between police and citizens, the "group" is the community or society. The authority is the police officer and the legal system he or she represents. Thus, according to the group engagement model, fair and respectful treatment by a police officer would signal that the individual is a valued member of the community and that it is safe to engage cooperatively and defer to the officer's authority and follow community rules. Assuming the process does not become disrespectful and demoralizing, this engagement and voluntary cooperation may continue beyond the immediate contact to subsequent interactions with the justice system (and other systems).

Portions of the group engagement model have been supported in studies of encounters between the police and citizens. Tyler and colleagues (22,23,24,25) have found that when people perceive the police as exercising their authority with fair procedures, they evaluate the police as more legitimate and, in turn, are more likely to accept and voluntarily comply with an officer's decisions. Findings from a study of recidivism among spouse batterers indicate that perceptions of procedural justice and fair treatment by the police decreased the likelihood of subsequent spousal assaults (26), suggesting that how officers treat citizens has both an immediate and long-term effect (23).

Several studies have also found that disrespectful behavior on the part of police officers reduces the likelihood of citizen cooperation (27,28). Using street encounter data collected in 1996 and 1997 from the Project on Policing Neighborhoods, McCluskey (5) examined how procedurally just and coercive tactics used by police officers influenced citizen cooperation with officers' requests for self-control. Police behaviors consistent with procedural justice (respect and allowing the person a voice) enhanced compliance, whereas police disrespect and coercive tactics (threats or force) reduced citizen compliance. Procedural justice had its greatest impact early in the encounter, suggesting that how officers initially approach someone is extremely important. Unfortunately, as Ruiz and Miller suggested (6), officers may encounter persons with mental illness and be primed to forcefully manage what they perceive to be an extremely dangerous situation and inadvertently make compliance less likely.

Although McCluskey (5) did not focus his investigation on police contacts with persons with mental illness, he did examine the effect of the presence of irrationality—defined as strong emotion, mental illness, or drug or alcohol use—on citizen compliance. Contrary to his initial predictions, "irrational" citizens were more likely than rational citizens to be pushed to rebellion by disrespect and more likely to be coaxed into compliance by fair and respectful treatment. Perhaps "irrational" citizens are more insecure about their status and thus are more attentive to procedural fairness.

Writing about court-ordered psychiatric treatment, Tyler (29) suggested that enhancing perceptions of procedural justice would be "likely to facilitate the subsequent therapeutic process." Sydeman and colleagues (30) suggested that the benefits of fair treatment in civil commitment hearings include greater trust in the mental health and legal systems and enhanced self-worth and self-respect. This, in turn, may facilitate involvement in treatment and long-term adherence. Personal accounts from persons with mental illness about treatment coercion and commitment hearings also suggest that voice, dignity, and trust are important to self-worth and quality of life (30).

Several studies have established that persons committed for involuntary treatment are sensitive to procedural aspects of the hearing process (31), and some have found that this sensitivity is not related to the presence of psychosis (32). Cascardi and colleagues (32) further found that in commitment hearings, features consistent with procedural justice were related to more positive attitudes about participation in treatment. More recent examinations of similar processes among individuals who are court-ordered into treatment in the community suggest that perceptions of procedural fairness are negatively associated with perceptions of coercion (33).

Findings from the Broward County Mental Health Court Evaluation support the procedural justice model for persons with mental illness in criminal justice proceedings (10). Participants involved in the mental health court, which incorporated procedural adaptations in its effort to reduce stigma and contribute to the therapeutic process, scored higher on perceived procedural justice than participants with mental illness in the traditional court. Higher perceived procedural justice—in particular, having a voice, feeling treated as a person, and fairness—was associated with a more positive emotional impact of the hearing and greater satisfaction with the court outcome.

Clearly, aspects of procedural justice are important to citizens' experiences of and responses to encounters with authorities such as the police. Procedural justice may be particularly important to the experience of persons with mental illness, who, as part of an already stigmatized group, may pay close attention to how they are treated and be more likely to cooperate when treated with dignity and respect. As de facto gatekeepers to both the mental health and criminal justice systems, the police, in response to a person with mental illness, play a pivotal role in determining the dynamics of the interaction, the extent to which the person cooperates, the resulting outcome, and perhaps the person's willingness to cooperate with both systems in the future.

Closely related to perceptions of procedural justice but conceptually distinct are perceptions of coercion. The MacArthur Foundation Research Network on Mental Health and the Law has systematically examined patients' subjective experience of the hospitalization process and identified key features of coercion (34). These features include perceptions of the amount of influence, control, choice, and freedom in the admission process (35). Studies conducted by the network suggest that perceptions of being treated with more procedural justice (fairly, respectfully, and with dignity) are associated with feeling less coerced in the hospital admission process among both voluntary and involuntary patients (36), although the temporal direction of the relationship is not clear.

Negative pressures from others, such as threats and the use of force to obtain hospital admission, increase perceptions of coercion, whereas positive pressures in the form of persuasion and inducements do not (37). In these studies, perceptions of coercion do not appear to be related to personality and symptom characteristics (38). Patients' accounts of the admission procedure tend to be consistent with those of family members and admitting clinicians and remain stable over time (35,39).

With the greater emphasis on mandated treatment in the community in recent years, members of the MacArthur Research Network on Mandated Community Treatment and others have begun to examine perceived coercion related to various types of leverage (legal and informal) used to promote adherence to community treatment (33,40,41,42). This body of work suggests that individuals mandated to treatment in the community feel less coerced than those mandated to inpatient treatment but more coerced than nonmandated outpatients (33). Among persons on outpatient commitment orders, perceived coercion has been found to negatively impact quality of life (41).

It is likely that perceptions of coercion are also closely related to perceptions of procedural justice in police contacts with persons with mental illness. Police have various types of coercive power to maintain control over a situation. These range from suggestions to overt threats to physical force. Greater force may generate situations in which citizens are more likely to comply or may encourage resistance (43). When approaching persons with mental illness who are in crisis, authorities' displays of force intended to gain control may instead escalate a situation to violence (6). As discussed above, officers who treat a person with mental illness with respect and dignity and avoid coercive force may be more successful in gaining compliance (5) and resolving a situation safely (6).

It may appear that perceived procedural justice and perceived coercion are simply two ends of the same continuum. At the experience level, they may be felt as such. However, they are conceptually distinct. One might imagine situations in which an individual perceives neither coercion nor procedurally just treatment, in which they were not negatively pressured, yet not treated respectfully.

Many educational and training programs have focused on what officers should know and think about mental illness. Some, such as training in crisis intervention teams, also place significant emphasis on skills, procedural changes, and community linkages. We discuss these below. Although enhancing knowledge and improving attitudes are important goals, the effectiveness of these programs in terms of changing officer behavior and outcomes on the street is not known. The content of these programs is often determined by task forces comprising representatives from law enforcement, mental health, and consumer and advocacy groups. Although input from various sectors is critical to developing an effective and practical program, these efforts would benefit from an explicit and empirically supported theoretical framework to guide them.

Procedural justice theory provides clear direction for efforts to improve police response to persons with mental illness. The group engagement model points to specific and, as McCluskey (5) demonstrated, measurable behaviors that may improve officers' abilities to obtain cooperation and more effectively and safely manage encounters with persons with mental illness. These behaviors include treating persons with politeness, dignity, and respect (dignity); allowing them opportunities to present their side of the situation (voice); and expressing concern for the individual's well-being (trust). These behaviors seem very simple and perhaps obvious in terms of how police officers should treat all citizens. However, they may not be consistent with how officers are trained or how they have learned to manage the demands of the job. There also may be situations in which other, more forceful approaches are appropriate. The key challenges are to develop guidelines for officers to determine situations in which procedural justice behaviors should be applied (such as dealing with a mental health crisis or domestic violence) and to provide opportunities for officers to learn and practice these skills.

This is not to suggest that procedural justice-based training should replace existing education and training to respond to persons with mental illness. Rather, it could supplement or be incorporated into existing academy, in-service, and roll-call training. Both the recent changes in the certification criteria of the Commission on Accreditation for Law Enforcement Agencies, Inc., that require ongoing training in mental health issues for officers and the growing interest in the law enforcement community to effectively respond to persons with mental illness create an opportunity to reevaluate and improve training.

Although not generally written about in theoretical terms, the crisis intervention team model clearly incorporates elements that are conducive to individuals' being treated in a procedurally just manner. Crisis intervention team procedures involve specialized training and an organizational change in orientation to managing calls involving persons with mental illness (44). In departments adopting crisis intervention team models, handling these calls well becomes a valued function of the department. This orientation communicates to all officers (not just officers on a crisis intervention team) that persons with mental illness are respected members of the community and should be treated as such. Officers undergo crisis intervention team training and become crisis intervention team officers on a voluntary basis. Because officers self-select into the crisis intervention team role, they may be more highly motivated than the typical officer to respond in a manner that increases cooperation and reduces conflict. When a call for service is identified as a mental health call, a crisis intervention team officer is dispatched and once on the scene is the designated officer in charge, regardless of the rank of other officers present. In these situations, the lead officer is not only someone with specialized training in skills consistent with procedurally just treatment but also an officer who has genuine concern for the well-being of the person with mental illness (44) and departmental support for managing the call in a respectful and appropriate manner. Approaching an individual with mental illness with respect and genuine concern and communicating effectively (giving voice) can affirm the person's status as a valued citizen and provide a safe context for cooperation.

Jurisdictions that have implemented crisis intervention team programs report reductions in officer and citizen injuries (20) and fewer police shootings of persons with mental illness (45). As advocates have noted, one key element is that people with mental illness are now being treated with respect by the police (45). Perhaps this change in orientation to persons with mental illness has contributed to better cooperation and the reductions in injuries.

Research testing the group engagement model in the context of encounters between police and persons with mental illness may provide extremely valuable information for identifying key elements and behaviors of effective response and improving officers' skills for managing these contacts. However, it would be unwise to proceed without first incorporating the perspectives of persons with mental illness who have experienced various types of encounters with police officers. We are currently conducting qualitative work to examine the role of procedural justice elements in the accounts of police encounters by persons with mental illness who are currently in outpatient treatment. Our preliminary findings suggest that previous negative experiences (such as feeling singled out on the basis of race or apparent homelessness) lead people with mental illness to hold low expectations of their encounters with the police, but these individuals also differentiate encounters in terms of the degree of fairness and respectful treatment exhibited by officers.

For example, one informant related that his most recent encounter with the police had resulted from an acknowledged misdemeanor theft. Despite the outcome (arrest), his perception that the officer had treated him respectfully colored his evaluation of the entire incident. As he stated, "I got busted, … [but] the officer that arrested me, he was actually very kind. … He treated me like a human. He offered me a cigarette. … Normally, they rough me up, you know, they have the cuffs too tight. They talk to me like, very degrading, but this officer was very kind. … He just handcuffed me, told me to turn around, … and took me down to the precinct." Another participant who related his most positive experience with an officer told of an encounter in which he had passed out while driving. In that situation, the officer responded by expressing concern for his well-being when it became apparent that he had consumed alcohol while taking psychotropic medication. Although the officer arrested him for driving under the influence, he also arranged for release on his own recognizance rather than going to bond court.

Further analysis of data on a range of firsthand experiences will allow us to delineate specific behaviors that contribute to short- and long-term cooperation with law enforcement and other systems (including mental health). The inquiry will also allow us to move beyond seemingly important yet procedurally ambiguous prescriptions of "respectful treatment," through the identification of behavioral components of successful interactions. In subsequent work, we will elaborate on and test methods for enhancing officers' skills for managing contacts with persons with mental illness and promoting engagement and cooperation.

Practice guidelines based on a procedural justice framework are deceptively simple. That is, the theory implies that if police and other authority figures treat people with mental illness with decency and respect, they will be more likely to cooperate instead of resist. The group engagement model of procedural justice (22) provides an explanation for this phenomenon, conditions in which persons will be most attentive to how they are treated, and specific aspects of treatment by an authority that influence identity judgments and subsequent cooperation. The evidence presented here suggests that this is a particularly useful framework for understanding how police officers' responses to persons with mental illness influence the amount of cooperation they receive in the moment, and perhaps beyond the immediate encounter. Emerging models of police response to persons with mental illness have incorporated principles consistent with procedural justice theory and have reported some successes. Applying theory to these efforts will allow us to further identify key elements of effective response to persons with mental illness that can be applied alone or as part of full crisis intervention team programs. Theory will also guide research on these interventions, allowing us to determine the effectiveness among different individuals and in varied organizational and community contexts.

It is important to underscore, however, the caveat that overemphasizing procedural justice concerns (importance of fair process over substantive outcome) could, if unchecked, lead to substantively unjust outcomes' being obscured by seemingly fair procedures. Absent genuine concern, procedural justice techniques are simply a form of manipulation in the moment, which may backfire on officers in subsequent contacts with the individuals. This possibility may be less of a concern among crisis intervention team officers who self-select the role, presumably because they genuinely care about persons with mental illness. However, it may be a legitimate concern in expanding procedural justice training to the whole of law enforcement. Therefore, in training police officers to work with people with mental illness the training must convey the importance of substantively fair treatment, not simply the appearance of fair treatment.

The work described in this article was funded by grant MH-075786 from the National Institute of Mental Health to Dr. Watson. The authors thank John Monahan, Ph.D., Jeffrey Draine, Ph.D., and Melissa Schaefer Morabito, Ph.D., for useful feedback.

The authors report no competing interests.