Criminal Justice Studies, Vol. 17, No. 1, March 2004, pp. 91–105 ISSN 1478–601X (print)/ISSN 1478–6028 (online) © 2004 Taylor & Francis Ltd DOI: 10.1080/08884310420001679361 Learning to See Hate Crimes: A Framework for Understanding and Clarifying Ambiguities in Bias Crime Classification James J. Nolan III, Jack McDevitt, Shea Cronin & Amy Farrell ; Taylor and Francis LtdGJUP041007.sgm10.1080/08884310410001679361Criminal Justice Studies1478-601X (print)/1478-6028 (online)Original Article2004Taylor & Francis Ltd1710000002004JamesNolanDept. of Sociology and Anthropology, School of Applied Social SciencesWest Virginia University316 Knapp HallMorgantownWest Virginia 26506USA(304)293—5801 ext. 3210jim.nolan@mail.wvu.edu. Recent empirical research has identified ambiguity in bias crime reporting as a source of confusion and frustration in law enforcement agencies and as a source of error in the national hate crime statistics. The authors develop a framework for understanding and clarifying these ambiguities based on John Dewey’s conception of intension and extension and their own application of mathematical set theory to the issue. The authors discuss the implications of their model for helping law enforcement officials see bias crimes for varied purposes, including prevention, statistical reporting, and criminal prosecution. Keywords: Hate crime; bias crime; hate crime statistics; bias crime reporting Introduction A bias crime1 is defined by the FBI as a ‘… criminal offense committed against a person or property which is motivated, in whole or in part, by the offender’s bias against race, religion, disability, ethic/national origin group, or sexual orientation group’ (Federal Bureau of Investigation, 1997, p. 4). What seems at first to be a clear and well-articulated definition becomes much more complicated when applied to real-life events. James J. Nolan III is an Assistant Professor in the Division of Sociology and Anthropology at West Virginia Univer- sity. Jack McDevitt is an associate Dean of Research and Graduate studies at the College of Criminal Justice, Northeastern University. Shea Cronin is a Senior Research Associate at the Center for Criminal Justice Policy Research, Northeastern University. Amy Farrell is a principal research scientist and Associate Director of the Insti- tute on Race and Justice at Northeastern University. Correspondence to: James Nolan, Division of Sociology and Anthropology, School of Applied Social Sciences, West Virginia University, 316 Knapp Hall, Morgantown, West Virginia 26506, USA. Tel: (304)293-5801 ext. 3210; Email: jim.nolan@mail.wvu.edu. 92 J. J. Nolan III et al. Consider, for example, the case of J. R. Warren a 26-year-old gay black man from rural Grant Town, West Virginia. On July 3, 2000, Warren met with three white teens at a vacant house in this predominately white rural town. The teens had been painting the house that day and, prior to their meeting with Warren, had also been ‘drinking beer, huffing gasoline, and snorting tranquilizers’ according to official sources (Smith, 2001). Shortly after he arrived at the house, Warren became engaged in an altercation with the group. The three teens took $20.00 from Warren then beat him until he was unconscious – and believed dead. Two of the youths put Warren’s body into the trunk of a Camaro and drove to a remote area to dump it. While en route to the dump site, the teens discovered that Warren was still alive. Surprised, they stopped the car in a secluded spot alongside a remote rural road, dragged Warren’s body out of the trunk, and while he was still conscious, the teens repeatedly drove their vehicle over his slen- der body crushing him to death. The murder of J. R. Warren brought national attention to the small community of Grant Town. The case was popularly considered a hate crime in the minority commu- nity and by the local media because it had all the right elements: the victim was black, slightly built, and openly gay while the offenders were young white males from a rural southern town. Immediately following the crime the community held peace vigils, while advocacy groups like the Marion County NAACP and the West Virginia Lesbian and Gay coalition called on the police to investigate the incident as a hate crime (Fischer, 2000; Smith, 2000). The police refused to recognize the incident as a hate crime because they had uncovered an alternative explanation for the crime. The police explained the crime as the result of a ‘drug- and alcohol-fueled rage,’ that was brought on by the belief on the part of one of the defendants that Warren had told others of a ‘long-standing sexual relationship that the two had shared’ (Smith, November 25, 2002). It appears that ambiguous situations like the Warren case may be more the rule than the exception when it comes to identifying bias crimes. The Center for Criminal Justice Policy Research at Northeastern University recently completed a study of bias crime reporting practices in a sample of law enforcement agencies from across the United States. The findings revealed that there are many crimes, like the Warren case, that first appeared to be bias motivated, but under closer scrutiny revealed alternative explanations (McDevitt, Cronin, Balboni, Farrell, Nolan, & Weiss, 2003). For example consider the following two accounts from the Northeastern University study: While driving on a state highway, a white male cut in front of a car driven by a Hispanic male. In response the Hispanic male pursued the car driven by the white male and followed it to a local fast food restaurant. The Hispanic male exited his car and approached the white male driver and his female passenger while yelling ‘You shouldn’t mess with Mexicans.’ He then proceeded to assault the white male. While playing in a local neighborhood, a young white child accidentally knocked over the soda can of an African American child spilling its contents. The mother of the white child approached the mother of the African American child to explain the incident and apologize for the mishap. The African American woman yelled at the white woman ‘Get your white ass out’ and ‘I will kick your white ass.’ The white women said she Criminal Justice Studies 93 did not want any trouble and would not fight back. The African American women then proceeded to assault the white woman. The investigating officers reported that the African American woman had a history of ‘causing trouble’ in the neighborhood. (McDevitt, et al., 2003) In both of these incidents, the police recognized that indicators of bias motivation were present, i.e., the victims and offenders were from different groups and their differ- ences were highlighted in the aggressive language used by the offenders. However, the police were also able to explain both of these crimes in another way: as violent responses to some other ‘triggering’ event. In the first incident described above, the violent behavior was triggered by an unexpected traffic maneuver. In the second incident, the violent act was initiated by the accidental spilling of a drink. Because these crimes could be explained as a response to or retaliation for some other event, they were not consid- ered bias crimes by the police. Recent empirical research identified ambiguity in bias crime reporting as both a source of confusion and frustration within law enforcement and as a source of classification error in the national statistics (Bell, 2002; Boyd, Berk, & Hamner, 1996; Jenness & Grattet, 2001; Martin, 1995; McDevitt, J., Balboni, J. Bennett, S., Weiss, J., Orchowsky, S., & Walbolt, L., 2000; McDevitt, et. al, 2003; Nolan & Akiyama, 1999; Nolan & Akiyama, 2002). From these studies emerged a number of insights and recom- mendations, including the need for new policies, procedures, and enhanced training within the law enforcement community that would ensure that hate crimes are prop- erly identified, recorded, investigated, prosecuted, and prevented. However, there are limits to what we can know about this issue from empirical research. Clearly, the fact that bias crimes are often ambiguous has been confirmed by way of empirical study. Clearing up the ambiguity in bias crime reporting is a different issue, one that will require application of a rational thought process. In a sense the task to be accomplished is the creation of an organized way for law enforce- ment and others in the criminal justice system to recognize and comprehend certain events within their domain of interest relating to inter-group conflict and violence. Goodwin (1994) calls this ability to see objects of interest in a certain way the ‘profes- sional vision.’ He argues that the ability to see is not a transparent psychological process (as if everyone sees certain events in the same neutral and objective way), but one that is socially situated and the result of discursive practices. In fact, Goodwin writes ‘All vision is perspectival and lodged within endogenous communities of prac- tice’ (p. 606). To make his point clear, he compares the professional vision of the farmer with that of the archeologist while both are looking at the same patch of dirt. Where the farmer sees in the dirt the ability to support certain types of crops, the archeologist sees ‘stains, features, and artifacts that provide evidence of human activity at this spot’ (p. 606). In the same way the farmer and the archeologist learn to see the soil, law enforcement officials must learn to see bias crimes. The purpose of this paper is to add clarity to bias crime reporting by adjusting the professional vision of law enforcement in regards to these types of events. By learning to see hate crimes, law enforcement will be better able to identify and address them. 94 J. J. Nolan III et al. How Terms Take on Meaning: Intension and Extension One way to frame the problem at hand is to consider how words and terms (like hate crime) take on shared meaning. John Dewey (1910/1997) identified two distinct steps through which words and terms acquire distinct meaning. The first step is ‘intension,’ or simply defining the term so as to single it out. The second step is ‘extension,’ mean- ing to mark off certain groups of things that do and do not fit the definition. Using Dewey’s example to explain this process, we consider the word ‘river.’ The word clearly has a distinct meaning as set forth in its definition. As a test of its ‘distinct- ness’ it must set off a group of things that exemplify this meaning from things that do not. He writes: ‘The river-meaning (or character) must serve to designate the Rhone, the Rhine, the Mississippi, the Hudson, the Wabash, in spite of their varieties of place, length, quality of water; and must be such as not to suggest ocean, currents, ponds, or brooks’ (p. 130). This application of meaning to distinguish what is and what is not meant by the word ‘river’ is extension. Where ‘intension’ is the meaning in principle, ‘extension’ is the group of things that are being separated and distinguished. Based on current and previous research, we contend that in order to reduce the ambiguity relating to the term ‘bias crime’ this same process of intension and extension must occur. Clarifying the meaning of the term bias crime is an essential first step in developing the ability to effectively identify and respond to this social phenomenon. Through the dual process of intention and extension (defining and division), events that fit the definition of bias crimes will begin to take on shared meaning. Intension and the Ambiguity Arising from the Official Definition of Hate Crime In response to the passage of the Hate Crime Statistics Act of 1990, the FBI was charged with developing a national program for the collection of bias crime statistics. What evolved, and is still evolving, is the National Hate Crime Data Collection Program. This program involves the collection, compilation, and publication of bias crime statistics from state and local law enforcement agencies. In order to establish this national program the FBI had to first define the types of activities that it wanted to collect. The definition of bias crime that was developed by the FBI in collaboration with national law enforcement groups is provided at the outset of this paper. This ‘official’ definition serves to mark off the types of events law enforcement should consider as bias crimes. Although the FBI definition provides some clarity on the issue of bias crime, it also contributes to its ambiguity. In particular the phrase ‘… motivated, in whole or in part, by the offender’s bias …’ causes confusion, leads to errors, and can become a barrier to the full acceptance of bias crime as a legitimate crime category. In order to deal with the ambiguity resulting from this definition—and its applica- tion to real-world events—the police sometimes develop routines that help them sort things out. Bell (2002) discovered this situation in her study of the bias crime reporting practices of detectives in a large urban department. Although the official definition does not require this, Bell found that detectives had two basic requirements for bias crimes: 1) the victim and the offender had to have different identities, e.g., race, Criminal Justice Studies 95 ethnicity, sexual orientation and 2) the context in which the crime occurred must suggest a bias motivation rather than some other emotion such as anger or jealousy. In determining whether the context suggested a bias motivation, detectives looked for evidence to suggest some other motivation. They developed what Bell called the ‘typical non-hate crime,’ i.e., cases involving drugs, fights, retaliations for earlier fights, traffic accidents, and neighbor disputes. Detectives viewed these types of cases as being ‘less than 51 percent bias-motivated.’ One detective offered the following as an example of a typical non-hate crime: ‘A traffic accident between someone Asian and someone White. Racial epithets, slurs are exchanged’ (p. 144). McDevitt, et al. (2003) had similar findings. For example, in one focus group held in a large Midwestern city police department, several detectives shared an account of crimes occurring outside a local gay bar where patrons had been targeted for robbery. The detectives argued that bias was not the motivation for the robberies and that the offenders selected mostly gay male victims because 1) they were not likely to call the police and 2) they were likely to have lots of money. In another example in the same department, detectives told the researchers that a series of robberies targeting East Africans was probably not related to hate or bias. These detectives seemed convinced that the real reason the offenders targeted this population of people was the fact that these particular victims were not likely to call the police. The findings from both of these studies reveal how limited ‘intension’ can be in terms of helping law enforcement see bias crimes. Clarity of professional vision in regards to bias crimes will come only through ‘extension.’ Preparing for Extension: Examining the Domain of Possible Hate Events In preparing to extend the definition of bias crimes to real-life events, it is important to first identify the domain of all possible hate crime events2. For our purposes, we define the universal set {U} of possible hate crimes as the set of all police reports maintained within a police department (see Figure 1).3 The Universal Set {U}: The Domain of Possible Bias Crimes Figure 1 The Universal Set {U}: The Domain of Possible Bias Crimes 96 J. J. Nolan III et al. Within the universal set of all records in a police department, there are two types of reports: 1) crime reports, and 2) non-criminal incident reports. As the name implies, crime reports are written documents describing criminal activities that have been reported to and recorded by the police. Here we denote the set of crime reports as {R}. The set {R} contains all reports of criminal activity that occurred in a particular juris- diction including reports of UCR index crimes4. The set of non-criminal incident reports, denoted here as { }, includes all types of police reports that are non criminal in nature. These reports include such things as neighborhood disputes, suspicious persons, dangerous intersections, traffic violations, and much more. The union of {R}, the set of crime reports, and { }, the set of non-criminal incident reports, makes up the universal set {U}, i.e., the domain under consideration ({U} = {R Y }). There are initially three sets of police reports within this domain that must first be identified and considered (see Figure 2): Three Types of Reported Bias Events {A} = {police reports that document some confirmed or suspected bias-related activity {B} = {police reports that document bias crimes as defined by state and federal definitions for statistical purposes {C} = {police reports that document bias crimes defined by state and local criminal statutes. Set {A} contains all police reports, both criminal and non-criminal, which have some bias indicators present. Police reports in {A} include such activities as cross burnings; bias-motivated murders, assaults, and vandalisms; white supremacist rallies and recruitment (e.g., KKK); racist graffiti; neighborhood disputes involving different racial groups, etc. Reports in this set do not necessarily need to be confirmed bias crimes, but there must be some indication that they could be motivated by bias5. False reports of bias crimes (e.g., for insurance fraud) and crimes motivated by the victim’s bias are also included in this set. The set of reports that document bias crimes as defined by the federal and state definition for statistical purposes, {B}, include only those prescribed crimes and bias R̂ R̂ R̂ Figure 2 Three Types of Reported Bias Events Criminal Justice Studies 97 categories set forth in the official FBI definition6. Set {C} contains reports of crimes defined by state and local criminal statutes. Figure 2 depicts the relationships between these three sets within the universal set. Notice that {B} is a proper subset of {A}, in that all of the elements of {B} are also elements found in {A}. Likewise, {C} is a proper subset of {A}, but it is not a proper subset of {B}. In other words, all crimes that fit the statistical definition of bias crime do not necessarily fit the criminal definition. In some cases, the state and local criminal statute will include different bias or crime types. In the McDevitt, et al. (2003) study, one jurisdiction had a local criminal statute called “ethnic intimidation” which included acts of intimidation aimed at individuals and groups because of their ethnic- ity. To the police officers in this department, a violation of this criminal statute meant a violation of the hate crime law. So when asked to consider a situation in which someone was murdered because of his or her race, the officers had a difficult time determining whether this would constitute a hate crime in their jurisdiction. By viewing hate crime reporting in this way, one can begin to see why certain events are clearly hate crimes while others are not. The clearest hate crimes are those that fit the compound event denoted by {A B C}, i.e., ones that have bias indicators and fit both the federal statistical and the state criminal definitions. Dealing with Bias as a Partial Motivation Similar to what Bell (2002) found in her study of a single city police department, the researchers in the McDevitt, et al. (2003) study uncovered many reported crimes in which the offender’s behavior was motivated only in part by bias. Their findings reveal two main categories of these partially bias-motivated crimes: 1) Response/Retaliation crimes and 2) Target-Selection crimes. Response/Retaliations Events Response/Retaliation crimes are a subset of Response/Retaliation events. These events are defined as situation in which the offender’s actions are prompted by some other ‘trig- gering’ incident. The descriptions of the two ‘ambiguous’ bias crimes provided in the Introduction section of this paper are examples of Response/Retaliation crimes. The actions of the offender in response to or retaliation for some other triggering event may or may not be exacerbated by the victim and offender difference. In some cases a criminal response to some triggering event might occur even when the victim and offender are from the same group. Also, there are situations where response or retaliation actions of the offender may not even constitute a crime, as in the case of a person giving an obscene hand gesture in response to some unexpected traffic maneuver. But, there are instances in which the offender’s bias against the victim’s group does in fact become a motivating factor in his or her actions. When these actions become criminal, they may be considered as bias crimes. Applying our conceptual model to Response/Retaliation events, the set of police reports that describe events in which the offender’s behavior was a response to or retaliation for some other ‘triggering’ event is represented by {D}. See Figure 3. 98 J. J. Nolan III et al. Two Sets of Events With Ambiguous Motivation: Response Retaliation {D} and Target Selection {E} Conceptualizing Response/Retaliation events in this way, one can easily see that some of these events are criminal actions {RD} and others are not { D}. It is also clear that some of these actions fit the official statistical definition, {DB}, some fit the state and local criminal statute {DC}, and some fit both definitions {DBC}. Finally, there are, of course, Response/Retaliation crimes that have no bias indicators at all, denoted here as {R D}. Target-Selection Events Target-Selection events are those documented situations in which offenders had been motivated to commit certain acts, either criminal or non criminal, for reasons other than bias. The offender(s), then, select as the target of their action particular persons, places, or objects based on some other reason. In other words, the motivation for the behavior is considered separate from the motivation for selecting the victim. Target- Selection events are different from Response/Retaliation events in that they are not actions in direct response to some other ‘triggering’ incident. Instead, they result from some rational process by the offender. For instance, an individual may be motivated to commit a robbery and, after careful consideration, he or she selects a particular bank as the target of the crime. The factors the offender considers in selecting this particular bank may include the number of visible security officers present on particular days, the quality of security system, and the anticipated quantity of cash available to steal. There are two types of Target Selection events: 1) rational choice and 2) bias moti- vation. As in the above example, rational-choice events involve the offender choosing the target (or recipient of his or her actions) for any number of reasons. Other exam- ples of rational choice target selection include some of the situations presented earlier from the McDevitt, et al. (2003) study, including the targeting of gay men and East Africans for robbery because they were not likely to call the police7. Bias motivated Target Selection events are those situations in which the target of the action is selected R̂ A Figure 3 Two Sets of Events With Ambiguous Motivation: Response Retaliation {D} and Target Selection {E} Criminal Justice Studies 99 because of the offender’s bias. As an example of this type of event, consider the situa- tion described by McDevitt, et al. (2003). The police reported that several intoxicated white males went out looking for a fight one night in order to demonstrate their phys- ical prowess. While driving around the community in their vehicle, the group noticed a black male pumping gas and decided to attack him. The police determined that although the primary motivation for this incident was simply to get into a fight, the offenders selected a black male target because of their hatred for African Americans. Applying our conceptual model to Target Selection events, we designate this set of events as {E} (see Figure 3). Presented in this way one can easily see that there are Target Selection events that are either crimes {RE} or non criminal incidents { E}. Some of these events involve bias indicators {AE}, while others do not { E}. Some of these events fit the official definition of bias crime {BE}, some fit the local criminal definition {CE}, and some fit both the criminal and statistical definitions {EBC}. Extension of the Term ‘Bias Crime’ Once again, where ‘intension’ is the meaning of the term in principle, ‘extension’ is the group of things that are being separated and distinguished. The conceptual model presented here is helpful in that it subdivides the universal set of possible bias crime events into 21 different categories of events to consider (see Figure 4). Each region created by the overlapping sets is given a number that corresponds to descriptions and examples provided in Table 1. Types of Police Reports In our model, unambiguous bias crimes—those in which bias is the sole motiva- tion—are events that fit into regions 16, 17, and 18. In region 16 one would find bias- motivated crimes that meet the requirements for statistical reporting (according to the official definition), but not for criminal prosecution. Crimes that fit both the statistical and state criminal definitions are found in region 17. And, in region 18 one would find crimes that fit the definition of local criminal prosecution, but not for statistical reporting. R̂ A Figure 4 Types of Police Reports 100 J. J. Nolan III et al. Table 1 Types of Police Reports Relating to Bias Crime Reporting (See Figure 4) TYPE OF OFFENSE EXAMPLES (1) Response/Retaliation crime that is not bias motivated. A white male assaults another white male in response to a neighborhood dispute. No inter-group bias involved. (2) Response/Retaliation crimes that have some bias indicators, but for which bias motivation either is not or cannot be confirmed. A white male assaults an Asian male in response to a neighborhood dispute. Officers determine the racial difference did not motivate the assault. (This is also the category of intergroup crime where the bias cannot be confirmed because the offenders’ identities are unknown.) (3) Response/Retaliation bias crimes that fit the statistical reporting definitions, but not those set for criminal prosecution. A white male threatens to kill a lesbian couple in response to an ongoing neighborhood dispute. Officers determine that the threat was motivated by the offender’s bias against lesbians. The state hate crime law does not protect against anti-sexual orientation crimes. (4) Response/Retaliation bias crimes that fit definitions for statistical reporting and criminal prosecution A black male assaults a white male in response to an ongoing neighborhood dispute. Officers determine that the racial difference was a motivating factor for the crime. (5) Response/Retaliation bias crimes that fit definitions for prosecution, but not statistical reporting A female driver cuts in front of a male driver in a line of heavy traffic. The male forces the female off the road and assaults her. Police determine that the crime was motivated by the male’s hatred for women. The state criminal code covers gender-bias crimes, but the statistical definitions do not. (6) Response/Retaliation incident that has some bias indicators An Asian male driver cuts in front of a black male driver in a line of heavy traffic. The black male yells a racial epithet at the Asian driver. (7) Response/Retaliation incident that has no bias indicators Two white females get into an argument over a parking space. One of the females calls the other a “bitch.” (8) Target-Selection (bias-motivated) crime that fits definitions for statistics and prosecution A white male commits a series of robberies to get cash in order to buy illegal drugs. He decides to target black- owned businesses because of his hatred for all non white races. (9) Target-Selection (bias-motivated) crimes that that fit statistical definitions but not those set for criminal prosecution. A white male commits a series of robberies so that he is able to buy illegal drugs. He targets males coming out of gay bars because he hates gay people. The state criminal code does not cover anti-sexual orientation bias crimes but the definition for statistical purposes does. (10) Target-Selection (rational choice) with some bias indicators. A white male commits a series of robberies so that he is able to buy illegal drugs. He targets people with gay men because he thinks they will have money and are unlikely to tell. (11) Target-Selection (rational choice) crime. A white male commits a series of robberies so that he is able to buy illegal drugs. He targets rich people because they have the most money. (12) Target-Selection (rational choice) incident A female makes several unwanted advances to a male colleague. She follows him home one day to try one more time. The male notifies the police who warn the female not to continue her behavior. (13) Target-Selection (bias motivated) incident A white racist organization targets a black neighborhood to distribute fliers containing false and inflammatory information about African Americans. Criminal Justice Studies 101 The ambiguous bias crimes—those in which bias is not the primary motivation—fit into regions 3, 4, 5, 8, 9, and 14. Response/Retaliation bias crimes that meet the definition for statistical reporting are found in regions 3 and 4, while Response/Retaliation bias crimes that meet the requirements for local criminal prosecution are found in regions 4 and 5. Events that fall into regions 8 and 9 are Target Selection bias crimes that meet the definition for statistical reporting. Events in regions 8 and 14 are Target Selection bias crimes that would meet the definitional requirements for criminal prosecution. Through our process of extension, events that would not be included under the term bias crime would be those found in region 15, i.e., crimes that have some bias indicators but whose motivation cannot really be confirmed. Victim-precipitated bias and false reports of bias are events that would be included in this region. Bias motivated non- criminal incidents, region 19, also would not be considered a bias crime. Crimes in which bias is a secondary or partial motivation, but which cannot be confirmed (regions 2 and 10), would also be excluded from our consideration as a bias crime. So too would non-criminal bias incidents, i.e., those that fit into regions 6, 19, and 13. Table 1 Continued TYPE OF OFFENSE EXAMPLES (14) Target-Selection (bias-motivated) crimes that fit definitions set for criminal prosecutions, but not statistical reporting A white male commits a series of robberies so that he is able to buy illegal drugs. He targets women only because he thinks they “deserve to pay.” The criminal code covers anti- gender bias crimes, but the statistical definitions do not. (15) Crimes with bias indicators, but do not meet requirements for statistical reporting or criminal prosecution A black male is assaulted by a white male. After punching the victim, the offender runs away. Although bias is suspected, it cannot be confirmed. (Victim-precipitated bias and false reports of bias are also included in this category.) (16) Bias-motivated crimes that meet requirements for statistical reporting but not for criminal prosecution A group of white males assault patrons of a gay bar as they exit. Although clearly a bias crime for statistical purposes, the state criminal code does not cover anti sexual- orientation crimes. (17) Bias-motivated crimes that meet requirements for statistical reporting and criminal prosecution A group of white males assault patrons of a gay bar as they exit. Anti sexual-orientation crimes are covered both in the statistics and in the states criminal code. Cross burnings on the property of a black family new to a white neighborhood would also be included here. (18) Bias-motivated crimes that meet requirements for criminal prosecution but not statistical reporting A man who hates all women goes on a killing rampage. At a mall, he shoots 5 women who were selected at random. Although this state includes gender bias crimes in its criminal statutes, it is not included in the statistical reporting definitions. (19) Bias-motivated incident (non criminal) A black man calls an Asian man a racial epithet. No threat is made to the victim. (20) Non-criminal incident that has no bias motivation A woman arrested for driving under the influence of alcohol (21) Non Bias crime A group of juveniles engaged in underage drinking of alcohol. 102 J. J. Nolan III et al. Discussion At the outset of this paper, we defined our purpose as being to help law enforcement officials learn to see bias crimes. We recognized the concept of ‘professional vision’ and committed to a rational process for improving that vision—at least as it pertains to the concept of bias crimes. The methods we used for increasing the clarity of professional vision included the dual process of ‘intension’ and ‘extension’ described by John Dewey. In the present situation, intension means the marking off of certain behaviors that constitute a bias crime. Therefore, the “official” definition of bias crime provided by the FBI as part of the National Hate Crime Data Collection Program became our starting point. By extension, we refer to the identification of events that do and don’t fit this definition. We applied set theory to our task which helped us identify categories of events for our consideration (Setek, 1983). Now, what remains for us to do is to make some sense of our findings. To begin we pose two important questions that must be answered in the order that they are presented: 1) Why do police keep records? and 2) How does our inquiry sharpen law enforcement’s professional vision in regards to bias crimes? We begin with a discussion of why police keep records in the first place. We posit that there are three primary reasons why the police keep official records: 1) to document community problems, 2) to compile crime and other administrative statistics, and 3) to prepare for and support the successful prosecution of criminal laws. Police officers often respond to community disturbances that do not rise to the level of criminal activity, but may be of interest for other reasons. The police may document these types of incidents in order understand the nature and extent of community problems as part of their ‘order-maintenance’ function (Kelling and Coles, 1996). It is often the case that non-criminal incidents can become precursors to criminal behavior and so the police may find value in documenting these situations for purposes of crime-prevention or problem solving. The police also keep records for statistical purposes, which is often the least popular reason for police to write reports. However, accurate and reliable crime statistics can be of significant value to investigators, administrators, researchers, and public policy offi- cials. Today, nearly all police agencies in the United States participate in the national UCR program that is managed by the FBI. Local police records of reported arrests and criminal offenses are the source of these national statistics. In addition to their value on the state and national levels, local crime statistics can also have significant operational value too. These statistics are used for crime analysis, crime mapping, problem solving, research, and program evaluation. The third and final reason why police officers keep records is to prepare for success- ful criminal prosecutions. Most police officers know that their crime reports will be presented in court as their ‘official memory’ of events. Therefore, it is important for officers to clearly and accurately document the elements of the investigation so that a conviction in the case is possible. The reasons for (and utility of) police-generated crime reports also applies to bias crime reporting. The police may want to document non-criminal incidents that have Criminal Justice Studies 103 certain bias elements in order to assess and/or prevent inter-group conflict in a community. They may also want to maintain reports of bias crimes for statistical purposes for their own administrative and operational purposes as well as to support state and national statistical efforts. Finally, the accurate documentation of bias crimes is essential toward assuring successful prosecutions. Our answer to this first question provides a context for the, arguably more impor- tant, second question: How does our inquiry help improve law enforcement’s profes- sional vision in regards to bias crimes? We believe our model helps law enforcement see bias events very differently. Through the rational process of differentiating types of bias events, we hope to introduce some clarity. In response to the first question, we argued that the police keep records for different purposes. Disentangling the term hate crime may help officers see utility in bias crime reporting in response to these purposes. As it applies to crime prevention and problem-solving, perhaps what law enforce- ment officials really want (and need) to know is more than the nature and extent of events that fit the statistical or criminal definitions of bias crimes, {B} and {C}. In fact, they may find more value in identifying events, both criminal and non-criminal, which have some bias indicators, i.e., the set {A}. In regards to improving the accuracy of their bias crime statistics, law enforcement officers might want to consider the data collection methods prescribed by the FBI. That agency refers to a two-step process in which the first step is the identification of ‘suspected bias crimes’ by the initial responding officer. The second step in the process is the verification of both the bias motivation and the crime classification by trained specialists such as a detective or bias-crime investigator. According to our model, first responders might want to identify all events with some bias indicators, i.e., {A}, as ‘suspected bias crimes.’ Once identified in this way, trained specialists can review and follow-up these cases in order to confirm which events actually fall under the statistical definition and which fit the local criminal statute. Our model may also help law enforcement agencies develop training strategies for identifying and classifying different variations of events that fit the statistical and crim- inal definitions of bias crime. For example, training officials might want to focus on what a crime in region 4 (of Figure 4) would look like and how it might be different from events in region 5? Along the same lines, our model may assist law enforcement agencies develop decision-making strategies for verifying different types of offenses for statistical purposes, particularly those that fall in the regions of the ambiguous bias motivation, i.e., regions 3, 4, 8, and 9 (of Figure 4). For example, law enforcement offi- cials might work on developing questions a detective would want to ask to determine if the bias crime should fit in region 3, which would be reportable for statistical purposes, or regions 2 or 5 which would not. Finally, this model provides a foundation for improving bias crime training for criminal investigators. Investigators may want to know what types of information they need to collect to make sure their case fits the criminal definition, i.e., set {C}. For example, they may be able to develop strategies for identifying Response/Retaliation crimes that fit the criminal definitions, i.e., regions 4 and 5, and those that don’t, i.e., regions 2 and 3. 104 J. J. Nolan III et al. Conclusions All professions learn to see events according to their particular set of interests. This is referred to here as the ‘professional vision.’ Recent empirical research has identified bias crimes as a category of crimes that are ambiguous. This lack of clarity creates problems for law enforcement officials who seek to identify and record bias crimes for purposes of crime prevention, community problem-solving, statistical reporting, and criminal prosecution. Our goal has been to reduce the ambiguity in bias-crime reporting so that law enforcement officials can be better prepared to identify and respond to them. We assume that, like other terms, ‘bias crime’ will take on a shared meaning through a process of intension and extension—first defining, then extending it to certain groups of events that do and don’t fit this definition. Our model presents a rational way of thinking about bias crimes and extending the definition for different law enforcement purposes. Notes 1. [1] Throughout this paper the terms bias crime and hate crime are used interchangeably. 2. [2] In this section we will use the standard symbols for set notation: e.g., {AB} represents the intersection of sets {A} and {B}, { } is the compliment of set A – i.e., {not A}, and {A U B} is the union of sets {A} and {B}. 3. [3] We acknowledge that many bias crimes do not even get reported to the police. We are simply limiting our argument to the group of events that get reported to and recorded by the police. 4. [4] Index crimes include murder and non-negligent manslaughter, forcible rape, robbery, aggra- vated assault, burglary, theft, motor vehicle theft, and arson. Also included here are other crimes for which reports were taken and/ or arrests were made. 5. [5] The FBI provides a list of bias indicators in their Training Guide for Hate Crime Data Collection. If any of these indicators exist, the report would be included in {A}. 6. [6] The federal definition of bias crime includes only 11 crime categories and 5 bias types. 7. [7] In situations like these, rational choice target selection crimes might have the same effect as a bias crime on victims and communities even though the intention of the offender is not to ‘send a message’ as is often the case in bias crimes. References Bell, J. (2002). Policing hatred: Law enforcement, civil rights, and hate crime. New York: New York University Press. Dewey, J. (1910 1997). How we think. Mineola, NY: Dover Publications, Inc. Federal Bureau of Investigation (1997). Hate crime data collection guidelines. Washington, DC: US Government Printing Office. Fischer, K. (July 20, 2000). Parents request action: Washington vigil set to remember slain Marion man. Charleston Daily Mail. Goodwin, C. (1994). Professional vision. American Anthropologist, 96(3), 606–633. Jenness, V. & Grattet, R. (2001). Making hate a crime: From social movement to law enforcement. New York: Russell Sage Foundation. Martin, S. E. (1995). A cross burning is not just an arson: Police social construction of hate crimes in Baltimore County. Criminology, 33(3), 303–326. A Criminal Justice Studies 105 McDevitt, J., Balboni, J., Bennett, S., Weiss, J., Orchowsky, S. & Walbolt, L. (2000). Improving the quality and accuracy of bias crime statistics nationally. Washington, DC: United States Department of Justice. McDevitt, J., Cronin, S., Balboni, J., Farrell, A., Nolan, J. & Weiss, J. (2003). Bridging the information disconnect in national bias crime reporting. Washington, DC: US Department of Justice. Nolan, J. & Akiyama, Y. (1999). An analysis of factors that affect law enforcement participation in hate crime reporting. Journal of Contemporary Criminal Justice, 15(1), 111–127. Nolan, J., Akiyama, Y. & Berhanu, S. (2002). The hate crime statistics act of 1990: Developing a method for measuring the occurrence of hate violence. American Behavioral Scientist, 46(1), 136–153. Setek, W. M. (1983). Fundamentals of mathematics (3rd Ed.). New York: McMillan Publishing Co. Smith, V. (July 8, 2000). Groups want answers on death: Gay-rights activists demand to know if case is hate crime. Charleston Daily Mail. Smith, V. (November 25, 2002). Warren family settles wrongful death case before trial. The Associated Press.