Violence Theory Workshop Summary

Summary of a workshop sponsored by the National Institute of Justice

December 10–11, 2002

Violence Theory Workshop, Day 1, December 10, 2002

Violence Theory Workshop, Day 2, December 11, 2002

Violence Theory Workshop, Day 1, December 10, 2002


Sarah V. Hart

Sarah V. Hart, Director of the National Institute of Justice (NIJ), Department of Justice (DOJ), welcomed participants to the Violence Theory Workshop. NIJ, the research, development, and evaluation arm of the DOJ, primarily serves State and local governments and practitioners. In its effort to help these governments and practitioners better perform their day-to-day functions, NIJ has placed a significant focus on violence research. In fact, the agency has funded 199 separate grants on violence research over a 4-year period, more than any other Federal agency.

The Violence and Victimization Division of NIJ's Office of Research and Evaluation (ORE), the coordinator of this meeting, has been especially active in promoting funding for sexual violence and domestic violence research. After September 11 (2001), however, NIJ's focus expanded to address terrorism. Much of this new work centers on technology issues, such as first responders, contraband and weapons detection, and airport security. Social science funding related to terrorism is concentrated on money laundering and its impact on terrorism. In addition, researchers are comparing a large private international security database on terrorism to official government databases in order to gain further insight into the problem.

Director Hart reminded participants that NIJ was soliciting proposals for a variety of violence and terrorism research grants (the applications were due February 14, 2003).

The purpose of this workshop—comprised entirely of "theorists"—is to identify areas of additional research for NIJ and to create theoretical links between different types of violence. The workshop participants were selected for their expert knowledge of violence theory as well as for their creativity and broad vision. Practitioners and policymakers will be convened at a later meeting to discuss the practical implications of the outcomes of this theorists' workshop. From these two meetings, NIJ will move forward to develop a research agenda.

Thomas E. Feucht

Dr. Thomas Feucht, Acting Director of ORE, expanded on the purpose of the workshop. Meeting participants produced 13 papers in preparation for the workshop, 7 on theoretical topics and 6 on empirical issues. Copies of all papers were provided to participants prior to the workshop. NIJ hopes that these papers will create a gateway to theory building from many different angles. Dissemination of the papers to a wider audience is also planned.


Margaret A. Zahn

Dr. Margaret Zahn of the Department of Sociology and Anthropology at North Carolina State University, a key figure in the planning of the workshop, welcomed participants and thanked them for their timely submission of workshop papers. She stated that researchers who study violence know their own areas of research well, but rarely talk to others involved in different facets of violence. The goal of the workshop is to break down barriers between researchers, and allow the development of theoretical links across types of violence. The aim is to develop an explanatory framework for a "generic" theory of violence that explains, predicts, and ultimately can change violent behavior.

Participants introduced themselves and indicated their areas of interest:

  • Dr. Robert Agnew, Department of Sociology, Emory University. Area of interest is causes of violence, particularly from the perspective of strain theory.
  • Dr. Ronald Akers, Center for Studies in Criminology and Law, University of Florida-Gainesville. Area of research is social learning theory.
  • Dr. Donald Black, Department of Sociology, University of Virginia. Area of interest is violence as a form of social control.
  • Dr. Mary Ann Dutton, Department of Psychiatry, Georgetown University Medical Center. Interest is in domestic violence.
  • Dr. Finn-Aage Esbensen, Department of Criminology and Criminal Justice, University of Missouri-St. Louis. Research interest is self-report studies of violence and youth gangs.
  • Dr. Marcus Felson, School of Criminal Justice, Rutgers University. Research area is routine activities theory.
  • Dr. Mark Hamm, Department of Criminology, Indiana State University. Research areas are white supremacy movements, hate crimes, and domestic terrorism.
  • Dr. Christopher Hewitt, Department of Sociology and Anthropology, University of Maryland-Baltimore County. Research interest is terrorism.
  • Dr. Michael Lynch, Department of Criminology, University of South Florida. Areas of study are corporate violence and environmental violence.
  • Dr. James A. Mercy, Associate Director for Science, Division of Violence Prevention, National Center for Injury Prevention and Control, Centers for Disease Control and Prevention. Interests include how theory can help understand how to prevent crime.
  • Dr. Ruth Peterson, Department of Sociology, Ohio State University. Interest is the effects of violence on minorities.
  • Dr. Claire Renzetti, Department of Sociology, St. Joseph's University. Area of research is feminist theories of violent behavior.
  • Dr. Richard Rosenfeld, Department of Criminology and Criminal Justice, University of Missouri-St. Louis. Interests include violence theory, social sources of violence, social networks and violence.
  • Dr. Roberta Senechal de la Roche, Department of History, Washington and Lee University. Research interest is a general theory of collective violence.
  • Dr. Murray Straus, Family Research Laboratory, University of New Hampshire. Area of interest is family violence.
  • Dr. Charles Tittle, Department of Sociology and Anthropology, North Carolina State University. Research interest is theory integration and control balance theory.
  • Dr. Helene Raskin White, Center of Alcohol Studies, Rutgers University. Area of interest is drugs and crime.

Also present were Dr. Henry Brownstein, Director of the Drugs and Crime Division and Acting Director of the Violence and Victimization Division, NIJ, and the following NIJ representatives: Bernie Auchter, Katherine Darke, Ron Everett, Shelly Jackson, Anna Jordan, Akiva Liberman, Leora Rosen, Richard Titus, and Natalan Zachary.

Frank Hartmann, Workshop Moderator

Dr. Zahn introduced the workshop moderator, Frank Hartmann. Dr. Hartmann is Executive Director of the Malcolm Wiener Center for Social Policy at the John F. Kennedy School of Government, Harvard University. His stellar career accomplishments include positions as director of the Hartford Institute of Criminal Justice, director of research and evaluation of drug programs for New York City's Addiction Services Agency under the Lindsay administration, and program officer at the Ford Foundation.

Back To Top


Dr. Hartmann outlined the format of the workshop. It was assumed that each participant had read the workshop papers; authors would not reiterate their theoretical approaches for the group. Instead, participants would focus on finding links and unforced integrations among the theories. Participants should strive to reach away from their own theoretical points of view and ponder other theories that may be congruent with their own research, while at the same time defining necessary boundaries and distinctions.

Drs. Agnew and Tittle would first offer common themes from their interpretations of the papers. Following a discussion, participants would be asked to describe what themes resonated for them. Later in the day, Dr. Rosenfeld would present challenges to the theories. Consideration of the fit of theoretical approaches with empirical studies would conclude the day's discussion. The second day's discussion would depend on the first day's outcomes.

Common Themes Among the Papers

Robert Agnew

Dr. Agnew stated that several common themes emerged from the workshop papers. First, violence is often provoked by some violent or negative condition. The theories differ in the nature of the events and conditions that provoke the violence:

  • Strain theory—Negative personal treatment may result in crime or violence.
  • Control balance theory—Provocations make one aware of control imbalances, which ultimately result in violence.
  • Feminist theory—The inability to accomplish gender through legitimate channels may result in violence under specific provocations.
  • Violent structures—Grievances set the process in motion.
  • Social learning theory—Identity-based grievances prompt violent activities.

Dr. Agnew raised several questions in relation to provocation. Do different types of provocations result in different types of violence? Are some provocations more likely to result in violence in a given setting than other provocations?

Second, the papers often describe an event or condition that creates a problem that calls for a solution, and violence is sometimes the solution to that problem:

  • Control balance theory—The control ratio is upset or challenged.
  • Strain theory—Negative effect is a motivator.
  • Violent structures—One's sense of justice has been violated.
  • Feminist theory—Masculine status is threatened.

This theme indicates that it is important to look at the processes that motivate violence, as well as the response to violence.

Third, the type of violence that an event leads to appears to be strongly conditioned by a range of factors:

  • Social learning theory—The anticipation that violence will be positively reinforced because of exposure to violent models.
  • Control balance theory—The constraints or costs of violence and the individual's level of self control.
  • Routine activities theory—The availability of suitable targets and the lack of capable guardians.
  • Social geometry or violent structures theory—The degree of social distance between parties.
  • Feminist theory—The relative power between targets and the participants' beliefs regarding aggression.
  • Radical ecology theory—The participants' exposure to lead.

The factors that condition reactions to situations, according to the papers, include a range of individual characteristics and personality traits, characteristics of the event, and structural or social characteristics. Obviously, theorists disagree as to the relative emphasis that should be accorded the various characteristics. However, all of these factors have important effects on reactions to negative events and conditions.

Charles Tittle

Dr. Tittle identified three common links among the theories of violence. The first was the extent to which violence is portrayed as a useful tool or device that people or groups use to correct a problem or redress a wrong. He pointed out the need to more sharply identify the problems, issues, and provocations that bring the violent response into play—in essence, why and how the violence occurs. What motivates offenders? How are opportunities for violence created?

The second link is the notion of a spark or provocation that brings about the violent behavior. Many of the theories "set the stage" for violence, but do not address the actual act. Dr. Tittle raised a number of questions concerning this issue. What is the provocation that brings forth a response that was potentially able to happen based on the theoretical "stage"? In the foreground may be factors that do not make sense or seem obvious to outsiders, but which resonate for the actors. To what extent do theorists need to understand what is going on inside the heads of the actors, versus understanding the relationship between actors? Also, how similar or different are the background features found in specific types of violence? For example, although the terrorists described by Dr. Hewitt come from different backgrounds, do their backgrounds have any common features that might set the stage for terrorist acts?

Finally, several of the papers described symbolic themes as possible causes for violence: grievances that are perceived as real by the actors and have symbolic meaning. Among these papers were those by Dr. Hamm and Dr. Aker and, to a lesser degree, papers by Drs. Renzetti, Lynch, and Senechal de la Roche.

How To Evaluate a Theory

The discussion opened with a conversation about the criteria for evaluating a theory. An ideal theory of violence should be able to predict with a high degree of precision the form and amount of violence, as well as its likelihood, from one situation to the next. This theory should address violence that is fast and quick, as well as violence that might take years to evolve, and should help explain historical events. In order to aspire to an ideal theory, appropriate criteria are necessary. The five conventional scientific criteria to which theorists should aspire are:

  • Simplicity or parsimony—Simple statements with maximum explanatory power should be used.
  • Generality—The theory should account for as much variation in the facts as possible.
  • Testability—Testing the theory's formulations and hypotheses should be feasible.
  • Validity—The theory's propositions and what is found in reality should match.
  • Originality—The ideas should give a better explanation of behavior than what was previously available.

Why should ideas be new? Is new necessarily better? A participant stated that in order to reach a better explanation of events, a new way of thinking is historically required. The ideas that are celebrated in the history of science are the radically new ideas, as well as the ideas that are the most general. Furthermore, this workshop is attempting to apply existing, familiar theories to types of violence that may be new; therefore, originality might be necessary.

A suggestion was made that an ideal theory must have a tangible concept. The greater the distance in space and time between cause and effect, the greater the chance that the theory will not be verifiable. The principle of "within a day" is applicable to most of the workshop papers: what causes can be translated and integrated to the day, or the act, of violence, even if the problem developed over a longer period of time? Lead poisoning is an example: the poisoning occurs over a long time span, but its effects can be brought to the moment of the crime. Another dimension of theory is "ruinability." The more abstract the theory, the more easily it can be misunderstood and twisted; the more concrete the theory, the more difficult it is to distort.

Several participants observed that the papers focused on individual or collective illegitimate or deviant violence, as opposed to acts of war by states or within a state. Any general theory will thus be restricted to small-scale violence. Violence by organizations or states may require a different explanation.

Predatory and moralistic violence. A discussion of predatory and moralistic violence ensued.1 A participant suggested that violence is not a unitary phenomenon, and the group may not be able to develop a general theory that addresses both predatory and moralistic violence. Excluding large-scale violence, violence is an overwhelmingly moralistic behavior. But another participant disagreed with this view, arguing that, though much of violence is predatory, not all violence is provoked. He defined "provocation" as violence that is perceived by the perpetrator as provoked by the victim, as opposed to violence that is not provoked by the victim. A general theory of violence should be able to accommodate both predatory and moralistic types of violence.

It was suggested that it could be possible to formulate a theory of greater generality that orders all violent dependent variables and includes multiple forms of violence (predatory, recreational or celebratory, and moralistic). Also, a broader definition of provocation to include all negative conditions that stimulate violence might facilitate a general theory. Another suggestion was that a middle-level, rather than general, theory might provide a structure to order the disagreements without the need to resolve them, thus allowing a dialogue to continue.

Nonviolence. The participants discussed the issue of nonviolence, noting that many of the papers only implicitly address this issue. What are the protective factors in social structures or communities that dissuade violence? It is necessary to identify the countervailing forces that prevent some individuals from engaging in violent behavior, and the direction and values of these variables. Several theories presented (control balance, social learning, and strain) imply a violent and a nonviolent component. In addition, provocations can lead to predatory or moralistic violence, or countervailing forces can prevent the violence.

A participant suggested that a theory must be able to predict more violence in one setting versus another. It is not clear how the theories presented in the papers could be tested to identify violent and nonviolent situations. For example, the participant noted that he does not see how routine activities theory could predict moralistic violence. Routine activities theory is a theory of opportunity. In his view, the ecology of the social structure provides the opportunity. Moralistic violence never needs a specific opportunity; it is always possible for people to use violence all the time.

Another participant countered that routine activities theory focuses on highly tangible variables, so that it is testable and falsifiable. Three things generate violence: wanting something from someone, wanting to punish someone, or wanting to preserve one's identity. This applies to moralistic violence as well.

Distinction between types of violence is not always clear-cut. In the context of domestic violence, both types of violence exist, i.e., provoked (predatory) violence and moralistic violence, the latter derived from a perception that the victim is doing something inappropriate. Both of these contexts explain violence through a cognitive appraisal that is mediated through perception of a threat. One type of violence may look like another. In domestic violence, what the abuser sees as moralistic may look predatory to outsiders.

It was pointed out that the group had presented several theories claiming to be general theories of violent behavior. None was overarching, in that each theory addresses different kinds of violence. A discussion of when each theory is appropriate to explain events or general or specific behavior, and how they fit together, was deemed warranted.

Defining the Theories' Variables

Dr. Zahn asked the group to define the variables of the various theories, assuming that there is both moralistic and predatory violence. This may make links more obvious.

Gender. A discussion of gender ensued. Although most of the papers dismissed gender, it is the most consistent variable across all forms of violence. Men commit more violence than women. Several participants observed that because gender is a constant over different types of variables, it cannot be used as a predictor. Violence does not occur every time a male is present; the presence of male hormones in the bloodstream cannot tell us when violence will occur, or how severe it will be. A participant pointed out that although there are many men, not all of whom commit violence, men are 10 times as likely as women to commit violence. If we can understand why men commit violence more than women, it could explain why violence occurs.

The group discussed sociological differences between men and women. Men are more likely to have weak ties, use anger as a major reaction to unjust treatment, and have low levels of self-control. Additionally, gender has an effect on violence because it locates people in the social structure in terms of power, hierarchy, advantages, and disadvantages. Along with race and class, gender affects beliefs, attitudes, deviance, the people with whom one associates, models, consequences of behavior, and opportunities.

It is important to consider the process and complexity of issues that affect the impact of gender on behavior; gender operates through its cultural environment and is not an independent, dichotomous variable. According to social geometry theory, violence is a form of self-help. This may explain gender expressions of violence in different economic and social situations. There are societies in which neither men nor women are violent, and societies in which women are very violent. Also, the difference between levels of violence between men and women is much smaller for moralistic than predatory violence. The group concluded that it is the social construct—not the biological issue—of gender that is important.

Unit of analysis. The issue of unit of analysis was addressed. Researchers must specify the unit of analysis in which the independent variables are located. In social geometry theory, the unit of analysis is the conflict. For routine activities theory, the closer one moves the many variables to the unit of analysis (i.e., time, day, incident, episode, setting), the easier it is to construct a framework to resolve disagreements. Because many of the theories use different dependent and independent variables, one theory's explanatory variable is the problem another theory attempts to address. On some levels of analysis the theories are complementary, but they are trying to explain different things. A participant suggested that radically different units of analysis will produce radically different theories. The unit of analysis in pure sociology, the conflict, is neither micro nor macro, and applicable to an individual, a group, a region, a nation-state, or a multinational group. It is possible to arrive at an expandable model that can address action at multiple levels.

Level of analysis.The issue of level of analysis was also discussed. It is important to think through compatibilities of theories that are pitched at different levels of analysis. Macro-level theories of criminal violence are compatible with some, not all, micro-level theories. The theories under discussion must accommodate different levels of analysis; in fact, several of the theories do apply to more than one level of analysis. A participant related unit of analysis to level of analysis. In order to test a theory, it can be applied to a society (such as a nation-state), to other societies with similar characteristics, and then may progress through more microlevels of analysis involving relationships and situational factors.

Relational distance. A second variable, relational distance, was defined. A dimension of social space, relational distance includes the nature and number of ties between people, the frequency of contacts between individuals and groups, and the age of the relationship. Relational distance is relevant when the nature of the grievance is held constant. The greater the relational distance, the greater and more likely is the chance of violence. Domestic violence may require additional variables, although in very patriarchal systems, a large degree of relational distance can exist between husband and wife. It was suggested that domestic violence has significant differences, in that the same behaviors that would not give rise to violence outside the household might incite violence in a household. Participants agreed that the kinds of conflicts that arise in the home are qualitatively different than those on the street.

The variables' relationship to the theories. The group discussed the relationship of variables to theories. For example, the value of the variable of relational distance is how it can be translated in a particular theoretical framework to help explain why a set of outcomes occurred. Some perspectives focus more on structure and content, both at the micro and macro level, while others are more process and outcome oriented. No one has said that any particular variable is not important. What theories predict with what variables? How is a given variable relevant to specific outcomes? By what process do variables produce differences in behavior among individuals, as well as across groups? Why are some individuals more likely to commit crimes, and why are they more likely to commit these crimes at certain times and places? This discussion has described motivated individuals influenced by a variety of factors (structural constraints, time, location background factors, social learning, control balance, strain, routine activities). The variables appear to have causal impacts on each other, to have linkages, and to overlap. It was suggested that the group list key variables from each theory, and ask how they work together.

Testability. Variables differ in the degree to which they can be tested. How do we count and measure the variables against observable facts? A theory should predict when violence will happen in any situation—a difficult challenge. A participant volunteered that many of the papers presented have variables that are explanatory and predictive, and that have been operationalized. Social learning, differential association, control balance, strain, and routine activities theories have been tested in different groups and have been supported. Other participants disagreed, arguing that those theories are measures of nonobservable variables that are "inside people's heads," and the findings of those research efforts are not direct findings that are testing the theories. Once the psychology and subjectivity of the actors is taken into account, the result is an unwieldy, complex theoretical apparatus. The goal of a general theory is simplicity, especially when applied in a historical context.

It was suggested that although the variables may be specific to a culture or setting, there is something common about them that researchers should be able to explain, and subsequently be able to test in a particular culture and unit of analysis. A preliminary chart of levels and theories was outlined:

  • Terrorism—The nation/state.
  • Gender, social geometry—The community.
  • Strain, social learning, lead poisoning—The individual.
  • Control balance, routine activities—Event settings and situations.

A Common Terminology

The group discussed the commonalities and links that had risen to the surface so far in the workshop. There was general agreement that the papers dovetailed in various ways, but that differences had also emerged. The lack of a common theoretical terminology was identified as a problem; some of the theories used different terms for similar concepts. There was no general consensus on the definition of violence.

Drs. Zahn and Brownstein summarized the research and conceptual issues to that point:

  • There may be useful ways of typing violence by comparing moralistic and predatory violence in terms of a series of variables and a series of contexts.
  • What is the role of provocation, and what is the role of individual or group responses to provocation?
  • Can there be a general theory of violence that applies to any level of society?
  • Does the study of violence have to be a study of individuals, or can it be a study of social structure?

Dr. Zahn mentioned that homicide researchers use the terms "instrumental" versus "expressive" violence. She questioned how this distinction relates to the moralistic/predatory categorization. One participant defined both moralistic and predatory violence as instrumental; both have a goal, so neither one is expressive. Another participant suggested using "planned" versus "impulsive." It was observed that this is an example of a semantics issue: "instrumental" has a concrete gain, while "expressive" does not. These terms roughly correspond to the moralistic/predatory categories. More often than not, violence is a means to an end, not an end in itself; sometimes the end is tangible, and sometimes it is more expressive, as in addressing a grievance.

Social Institutions: The Macrolevel View

The group noted that most of the papers had a strong individualistic or microlevel bias. Even papers with more macrolevel views addressed relatively small groups. However, social institutions (meaning the major, well-organized complexes of values and norms, not the concrete manifestations of these organizations) are important loci of violence. Additionally, most of the theories that address social institutions focus on violence that occurs because of institutional breakdown. The expected level of violence when institutions operate as they should is not defined; therefore it is necessary to identify the baseline levels of violence. How much more violence should be expected when, for example, child-rearing practices break down?

Organizations and Violence

The relationship between organizations and violence was discussed. Not all persons who engage in organizational violence (i.e., terrorist groups or youth gangs) come into the organization with the same set of mental or social characteristics. It is erroneous to assume that the characteristics of the followers are similar to the characteristics of the organization's leaders, and that violent activities are committed by persons who accept the organization's goals with the intensity or specificity of the leaders. A participant suggested that most perpetrators engage in organization-based violence because someone gave them a set of instructions and they could not get themselves out of the situation. In essence, researchers were warned not to confuse the characteristics of the leaders with the characteristics of the followers.

Back To Top


Challenging the Theories: Emphasis on Perpetrator, Victim, Event, or Social Structure?

Richard Rosenfeld

In presenting his challenge to the theories, Dr. Rosenfeld observed that several participants had urged specificity in the analysis, suggesting that researchers remain systematic in levels of unit of analysis and bring the systematic analysis specifically closer to the violence by identifying cognitive elements. These cognitive elements include: (1) the perpetrator, (2) the victim, and (3) the event in which the perpetrator and victim come together in time and space; moreover, the event is socially organized in a (4) social structure.

Do the theories give relatively greater emphasis to the perpetrator, victim, event, or social structure? Dr. Rosenfeld suggested that most standard criminological theories are perpetrator-oriented. He offered a typology:

  • Event-centered—Routine activity theory, with additional attention paid to the victim rather than the criminality of the perpetrator.
  • Perpetrator-centered—Control balance, strain, social learning, and radical ecology (the institution is the perpetrator) theories.
  • Victim-centered—Feminist theory brings the victim into play.
  • Social structure—Violent structures theory.

Violent structures theory and social geometry. A participant responded that in violent structures theory, the conflict is defined by the social characteristics of the victim (or target) and perpetrator, as well as other people who might get involved. There is also a particular conflict about a particular matter, which is the event. The social location of the victim and perpetrator is used to predict how the conflict is handled, with the event held constant. The social geometry of the conflict is intrinsically concerned with who the victim is and who the perpetrator is—the geometry attracts different kinds of social life to that situation, and defines the location and direction of the conflict. The social geometry determines when a situation generates a violent response, but a similar situation with different actors generates a different response. Violent structures theory does not address psychological characteristics because they are unobservable.

A participant agreed with localizing the conflict within a structure. Several participants preferred the term "target" to "victim," although they accepted Dr. Rosenfeld's schema. It was noted that Dr. Rosenfeld's typology allows each theory to have a main focus for testing purposes. The typology allows organization and linkages of the theories, without denaturing them.

Social learning theory and the perpetrator. A participant stated that social learning theory focuses on perpetrator behavior, but also looks at themes, meso and macro social culture, and immediate situational context. Social structure is not part of social learning theory, but has implications for social learning in determining across-individual differences.

A participant observed that it is impractical to study all aspects of all theories at once, although in principle, an overarching theory that embraces all of them is possible. Researchers can and have made advances by focusing on their individual theories. Participants agreed that it is possible to come up with a general theory that encompasses both predatory and moralistic violence, once semantic differences are resolved.

Defining violence. A participant was troubled by the conversation about violence. Group members' definitions of violence seemed to differ. What, he asked, is considered a violent act? Is a doctor who performs an unnecessary appendectomy committing an act of violence? Also, perpetrators are often one-time offenders, so generalizations about that individual do not hold across time and space. Participants agreed that no one is violent in all conflicts, and that the situation in which the act occurs is paramount. A perpetrator's behavior is variable and may be rare in his lifetime, but some perpetrators commit more violent acts than others. Therefore, the dependent variables in the perpetrator-oriented theories are important across persons and situations.

Research Efforts and the Four Elements

Dr. Hartmann asked the empiricists in the group to explain to what extent the research as applied to their type of violence draws on Dr. Rosenfeld's four categories (perpetrator, victim, event, social structure) and to define key variables for their type of violence. Also, if a participant's type of violence has not been applied to the four categories, to what extent might it benefit from doing so? How can perspectives from each of these categories best be integrated to understand that type of violence?

Alcohol-related violence. Dr. White explained that her research separates alcohol from other drugs, because the association of alcohol with violence differs from the association of drugs with violence. Drug violence is related to social structures of disorganized neighborhoods and opportunities, with violent individuals being drawn to the drug market; therefore, both the social structure and the characteristics of the perpetrator are important. Alcohol involves more psychopharmacological factors. Alcohol is a predictor of perpetrator proneness, coupled with a societal expectation that alcohol leads to violence, which therefore makes it expected and acceptable. In alcohol-related violent situations, it is often the victim who is drunk, making him or her an easy target.

Dr. White believes the theories can be linked to explain alcohol-related violence. The strain perspective provides negative conditions and effects that are accounted for by one's social status and social structure. These factors are mediated by social learning. All of this is affected by social controls, control balance, structural characteristics, alcohol use, individual characteristics, the setting, and the opportunity. All of these components together would predict the severity of violence and the probability of violence.

A participant commented that this combination of ideas did not fit the "simplistic" criteria of ideal theory evaluation. He suggested the group aspire to a more elegant, creative, and economical theoretical view. Theorists, he said, need to admit more of their ignorance, and consider new ways of conceptualizing variables. If this group came together in 15 years, what would they ideally have then? Would their theories be the same?

Domestic Violence. Dr. Dutton put domestic violence into the context of the four typological points:

  • Victim. There are multiple types of victims in the domestic violence field. Victims have differing characteristics—much work remains to be done to understand how people become victims. Social learning theory helps understand victims, in that victims often think of violence as normative instead of unacceptable.
  • Perpetrator. There are also multiple types of perpetrators in domestic violence, and both one-time and repeat offenders. Social learning also helps to understand the values and attitudes of perpetrators who condone violence. Some literature suggests that coming from homes with violence or experiencing child abuse contributes to perpetration.
  • Event. Domestic violence is characterized by a pattern of events and by a pattern of different kinds of events. The events vary in form and severity and are held together by more abstract coercive control, which may be more important than the discrete events.
  • Social structure. Individuals are situated within social constructions (ethnic, racial, and economic), and some of these appear to make a difference. Differential reinforcement is also important, since little punishment occurs in domestic violence situations. The social constructions of gender determine the social acceptance of violence in attitudes toward rape, convictions of rape, and reporting of violent crimes. The stigma attached to victim status is also at the social structural level.

Workshop participants discussed the peculiarities of the relationships in the domestic situation. Women who live near relatives or have close neighbors are much less subject to violence. The most violence occurs when the woman lives in the midst of her husband's kinfolk. Other levels are predictive as well, but which have the best predictive value is not known.

Dr. Zahn noted that there are databases available on domestic violence; a place to start may be the application of theory to these databases.

The group recognized the importance of theorists communicating well with empiricists, in that some empiricists did not reach the theorists' expected outcomes with their variables. The need for more narratives was suggested, such as the use of one theory to explain the stories of perpetrators or victims or historical moments from beginning to end. Lessons will emerge from the narrative moments.

Hate Crimes. Dr. Hamm brought up the issues of rage and skill development in relation to hate crimes. It is not fully clear where rage comes from, he said. Many times, the subjects cannot put into words what they have been through, particularly for terrorism and hate violence. Understanding skill development in perpetrators is also important. Hate-crime perpetrators appear to be obsessed with obtaining and refining the skills needed to commit their crimes. There may also be a sense of principal deviance behind these crimes that could help separate predatory from moralistic crime.

Terrorism. Dr. Hewitt discussed Dr. Rosenfeld's typology and terrorism. Most perpetrators of terrorism look "normal," he observed; it is not clear what makes individuals of similar backgrounds (i.e., social structure) commit or not commit terrorism. Furthermore, terrorism varies by society and is not impulsive behavior. Terrorists conduct organized campaigns that are consciously begun and ended. He offered that social learning theory seems to be most compatible with terrorism, in that violence is transmitted through institutions. In Cyprus and Northern Ireland, for example, violent nationalism is transmitted through schools. Terrorists there have high social approval, and see themselves as protecting or defending their country. The beliefs are transmitted from one generation to another in a formalized, institutionalized learning process, so there is always a steady supply of new recruits. Terrorists target people to achieve certain effects. All terrorist groups have different theories about the most effective type of victim selection. This is one reason for the variation in severity from one group of terrorists to the next, in addition to resource availability and opportunity.

A participant ventured that not much is known about violence in general; this is partly because of the lack of comparable research methodologies. Different sampling and measuring issues, and the use of different definitions, contributes to this problem. Researchers need to incorporate qualitative, ethnographic approaches, as well as survey approaches. Better research methodologies will allow researchers to translate their research to real-world situations.

The question whether theory is applicable to prevention or intervention was raised. A participant suggested that emphasis on the specificity of motive might provide opportunities for situational prevention. Also, the capacity of a theory to connect with manipulable, real-world phenomena is affected by policy limitations, and is not always indicative of a flaw in the theory.

Back To Top


Dr. Hartmann outlined goals for Day 2. A hallmark of good theory, he noted, is its predictive capacity. What change in criminal justice or violence control policy (prevention and intervention included) is implied by any or all of these theories? It is possible that the theory may not be relevant to the policy for this discussion, or that present policy implements a particular theory reasonably well.

Participants were asked to take one of the theories and identify a prediction about violence (its pattern, severity, trend, or level) that emerges from that theory. The National Crime Victimization Survey has shown a substantial drop in lethal and nonlethal crimes in the United States over the past decade. However, it is widely believed that the crime decline is now over, and that a rise in crime is likely during the next decade. What do the theories predict about crime during the next decade, Dr. Hartmann asked the group, and can they explain the drop in the past decade? Additionally, what are the implications for policy and practice?

Back To Top

Violence Theory Workshop, Day 2, December 11, 2002


Drs. Akers and Strauss summarized the previous day's conversation. The workshop started with the assumption that everyone was familiar with all the papers, which was partially sound. Papers were not systematically presented; instead, a discussion of central issues was used to explicate most perspectives. Standards were presented for judging whether a theory was or was not sound. The question of "What is violence?" was raised, although not satisfactorily answered. The unit of analysis, independent and dependent variables, and levels of analysis from the different perspectives represented in the room were discussed. Commonalities and linkages, as opposed to integration (which was not considered feasible), were sought. The problem of semantics complicated this process. The distinction was made between moralistic, predatory, and recreational or celebratory violence. Instrumental versus expressive violence was also explored, and participants reached a general agreement that most violence is instrumental (i.e., has a goal), with differences in the intended targets. The desired end result can have expressive elements.

Presentations of typologies of the major emphasis of each theory (i.e., victim, perpetrator, event, or structure) helped focus the conversation. The empiricists described which particular theoretical concepts or perspectives fit with their areas of interest. Some participants found relationships with several theories, while others focused on one theory.

The day ended with two questions: (1) How can the group make sense of social trends of crime in the past decade, and predict what will happen in the next decade; and (2) what polices, programs, or applications—especially ones that are novel—can the group suggest?

Dr. Straus added that three principles gleaned from family violence studies are important to remember:

  • Violence has multiple causes, and will require multiple theories.
  • Forms of violence are at some level interrelated.
  • No one method will provide the answer; a variety of methods will be needed.

Back To Top


Dr. Hartmann continued the workshop discussion, explaining that it would proceed on two levels:

  1. The group would continue to aspire toward an explanatory theory of violence (what had been discussed set the stage, but did not provide the triggering mechanisms for violence and was not predictive); and
  2. The group would take a mid-level approach in order to advise practitioners.

Dr. Hartmann suggested that the group ask what these theories would say to someone not connected with criminology. Also, he noted that what the group may view as background to violence may be a triggering factor for a perpetrator.

Using the Clues to Explain and Predict

Dr. Tittle addressed the question of whether it is a valid challenge that theory sets the stage but does not identify the triggering mechanism for violence. He stated that sociologists must first collect the information that will tell us what the clues are, and then attempt to explain and predict using those clues. Control balance theory looks at both foreground and background factors. Things that remind people of a control imbalance represent triggering mechanisms, or provocations. Because there are so many of them, and they are quite individualistic, they cannot be specifically catalogued. Researchers may need to find the underlying structures of the provocations, in a more abstract fashion, to capture all variations. People's experience in managing certain situations help them intuitively detect the clues, although they may not be able to verbalize them.

Several participants agreed that the primary value of a theory is to provide a simple and testable explanation. Because theorists are aiming for simplification, who better to ask about the implications of theory? Theory also provides a rich source for generating novel ideas for preventing violence. However, good ideas about how to prevent primary violence are scarce, even though ideas based on sound theory are the most likely to lead to policies and programs that are effective, cost-efficient, and socially acceptable.

The necessity of storytelling was emphasized. Researchers can show policymakers the data, but cannot tell policymakers what to do. Theory tells a story in its predictions. The test of the theory is whether the story can be told on the abstract level, and have meaning on the substantive level. It may take years to tell whether a theory holds up over time.

Variation in Rates of Violence

The participants discussed variations in rates of violence. The change over time in serious violence is of great importance to policymakers, but it is not the only variation of concern. There is cross-sectional variation and individual variation that may or may not correspond to aggregate changes in violence across either space or time. Variation in the phenomenon is the chief business of any explanatory theory, but may not be the only or most important business of the theory.

Dr. Senechal de la Roche discussed variations of violence over time and space. She studied lynchings in the American South using geometric social theory and the concept of relational distance. The lynching rate peaked in the 1890s. What, on the community level, was conducive to lynching? What increased the probability of lynchings occurring in cases of conflict? Lynching was typically a stranger offense, and the distribution of strangers in a community affected the lynching rate, all other factors held constant. Therefore, greater levels of in-migration meant greater rates of lynching. These principles can be applied to rates of change in violence: See what is changing in the larger environment and apply it to lower levels of analysis. This two-level approach allows applicability and testing.

Social Structure and Hot Spots

Social structure theory was applied to the definition of "hot spots." It was suggested that the term "hot structures" be used. For example, what is the social structure of some bars (not a bar)? Use of alcohol is a constant across the bars. This is also applicable to terrorism. There always has to be a grievance in order to get a moralistic response, but the nature of the grievance alone does not predict what will happen next. The appearance of a "hot structure" will increase the likelihood of terrorism.

Terrorism: Predictions

Dr. Hewitt continued the discussion of terrorism. Terrorism addresses a perceived problem that cannot be satisfied through the normal political process. It is a high-cost action, and therefore requires a strong grievance and significant levels of frustration. He offered the following predictions based on three terrorist categories:

  • Right-wing extremists in the United States. Terrorist acts by these groups will increase because of population trends. Changes in the demographic balance tend to produce ethnic conflicts. He predicted isolated domestic terrorist attacks in the Southwest and California. These groups are not well-organized and have been penetrated by law enforcement, so it will not be a concerted campaign.
  • Anti-abortionists in the United States. These groups have been making legislative gains, so Dr. Hewitt predicts reductions in anti-abortion terrorist acts.
  • Islamic terrorists in the United States. Surveillance of individuals thought to be potential terrorists is so intense that significant terrorist acts are unlikely, in Dr. Hewitt's view.

Social Learning Theory and Drug Use

Dr. Akers stated that social learning theory can help make sense of trends in substance use and abuse. Drug use was heaviest in the 1970s, and began to decline in 1979. About 2 years before, answers to questions concerning attitudes toward drug use noticeably changed: the attitudes were less favorable. Drug use declined throughout the 1980s, and leveled out about 1990. During the early 1990s, favorable attitudes toward substance use began to increase, and by 1992 and 1993, increases in drug use followed. Today, attitudes have moderated, and drug use has moderated as well. Dr. Akers suggested that if societal support for acts of violence generally change or moderate, that might produce behavioral changes. He predicted that violence will decrease over the next decade if attitudes moderate, all else being equal.

Dr. White added that it is necessary to predict what causes the attitude change. In the drug market, stabilization of the crack market, the aging of the drug-selling crowd, and incarceration may have affected attitudes on the street. A participant observed that the motivation for violence fluctuates little, but what has an impact is "guardian maintenance"—the source of consequences for the act. If the attitudes of the guardians moderate, this might increase guardian maintenance and decrease violence.

Environmental Design and Routine Activities Theory

Dr. Felson pointed out that consideration of issues of environmental design and situational management—elements of routine activities theory—can reduce the crime rate. This is a theoretical and practical convergence that has been put into practice in building design and management. Specifically, the U.S. Department of Housing and Urban Development (HUD) is a leading force in crime prevention in that HUD knows how to construct buildings that will decrease crime rates.

A participant suggested that the geometric structures theory tells a "structure" instead of a story. It is a totally different logic with practical applications because it enables us to alter the structural, instead of the physical, environment. Structural geometry has many elements—solidarity, isolation, cultural difference, social distance—and some factors will be more important than others in a given situation.

Recap: The Theories Applied

Drs. Mercy and Agnew summarized the application of theory-driven ideas:

  • Routine activities theory. This theory has many practical implications for prevention of violence, as it incorporates environmental design in general and in building codes to reduce violence.
  • Social geometry (violent structures) theory. This theory states that the lethality of violence is a direct function of social distance. This idea can be used to identify communities that have a high possibility of conflict and use policy measures to reduce social distance.
  • Control balance theory. The theory suggests that police treatment of minor offenses may increase situational risk and decrease violence.
  • Social learning theory. In regard to domestic violence, it may be possible to identify areas or subgroups in which violence against women is seen as normative, and to develop educational plans to counter this viewpoint.
  • Strain theory. Strain is moderated by coping strategies. Interventions for children exposed to violence could increase their coping skills in adulthood. Families could be taught more effective methods of discipline and how to resolve conflicts to reduce abusive behavior. Restorative justice is also an application of strain theory. Sanctions are perceived as unjust when people have no input in their development. Restorative justice increases the likelihood that offenders will perceive their sanctions as deserved by helping them better appreciate the harm they have caused, and tying the sanction to directly repairing the damage.

Social Geometry Theory and Blood Feuds

Dr. Black gave an example of social geometry theory as applied to blood feuds. Classic blood feuds involve a reciprocal exchange of killing. All blood feuds fit a model of stable agglomeration of social islands. The groups are relatively isolated, culturally homogeneous, independent of one another, but functionally similar. If any aspect of this model is changed, the pattern of violence changes.

Dr. Black has applied this model to American gang violence, and finds that in general, it fits. Some gang violence is preemptive, and killings may not always be "tit for tat." However, social geometry can predict and explain this pattern of violence without the use of individual or specific city or gang characteristics. Gang members have all their close relationships with other gang members and are socially distant from their families. Disruption of this type of solidarity by increased contacts with the outside world (e.g., when gang members get married, have families and jobs, and leave the gang) decreases the pattern of violence. Furthermore, if the social distance between gangs is increased, more violence is predicted. Less violence is predicted if there is increased social contact. Social geometry theory can tell policymakers that these structures need to be disrupted; it is up to the policymakers to devise specific strategies to this end.

Revisiting the Three Questions

Three questions raised by Dr. Rosenfeld on Day 1 were addressed again:

  • What change in violence-control policies is implied by any one of these theories, or all of these theories?
  • Is there any one aspect of violence that any of the theories predicts?
  • Can any of the theories predict for the future?

The discussion turned to several areas where theory could be applied to policy and/or practice or where theory considerations have significant implications for policy and/or practice.

Domestic Partner Violence

A discussion of domestic partner violence ensued. Domestic partner violence has been declining for a number of years. Men's attitudes toward the acceptability of hitting their partners has declined from 25 percent acceptance in the 1960s to 5 percent acceptance now. However, women's attitudes towards acceptability of hitting the male partner have stayed the same, remaining at 25 percent or greater, depending on the format of the study. These changes/lack of changes in attitudes are important for policy and prevention. Even assuming that men comprise a significant proportion of domestic partner perpetrators, crime statistics indicate that partner violence by women is a substantial part of the crime. Society cannot ignore a category of perpetrator just because another category of perpetrator has even higher rates of offense, or because it is politically incorrect to do so.

Men have higher rates of all kinds of crime. Women have been found to be violent to partners, but not violent in other situations. Studies have shown this response not to be purely self-defense. This data has important research applications, and highlights the need for researchers to communicate nonobvious data (data not in step with current policy) to policymakers. Domestic partner violence is complex across race and gender, and requires a variety of theories to deal with the variety of issues.

Terrorism: Recommendations

Dr. Hamm offered the following policy recommendations to reduce terrorism:

  • Information sharing among agencies.
  • Reduction of grievances that underlie terrorism, which would eliminate the structure that supports terrorism.
  • Recognition that terrorist groups are date-oriented, and choose the dates of their attacks to commemorate other acts (i.e., the Federal assault on Waco/the Oklahoma City bombing; Hitler's birthday/Columbine High School shootings).

Improving Databases

The discussion turned to the question, what variables in Federal and local policing databases would make a difference in understanding and dealing with violence? The above-mentioned dates are an example of the type of information that many policing agencies lack. Moreover, better information in police databases would allow those working toward the prevention of violence to establish networks. It would also allow modeling on the social-structural level.

Drawing Policy Implications: Control Balance Theory

Dr. Tittle discussed the application of control balance theory to practice. He began by pointing out that in order to draw appropriate policy implications from a theory, the integrity of the theory must be respected. For example, the changes in the social relationship variables in control balance theory (control ratio, provocation, opportunity, counter-control) can positively or negatively affect the other variables. When counter-control is increased, control ratio is changed in a way that may increase potential motivation in the face of provocation. The goal of control balance theory is to move more people into a balanced control ratio, but individuals who advance along the control ratio continuum from large control deficits to small control deficits may have a greater probability of violence. Therefore, careful consideration of the entire theory is needed before policy changes are implemented. Stability in the economy during the 1990s, Dr. Tittle theorized, may have moved a significant number of people into a balanced control ratio. Today, economic and political changes may affect maintenance of this balance, although Dr. Tittle was reluctant to make specific predictions.

Social Learning Theory Applied

Dr. Akers discussed the application of social learning theory to practice. The four major concepts of the theory (differential association, reinforcement, definitions, and modeling) have been defined, measured, and tested repeatedly, with success. A number of policy applications and implications can come from this theory, but it is important to ask for whom the policy is intended—neighborhoods, families, individuals, schools, law enforcement, criminal justice, or treatment and prevention programs? In that way, one can determine how effective the policy is.

Dr. Akers further noted that there are a variety of programs with cognitive behavioral approaches, including social learning, that have a good record of results. However, no program has huge effects. Too much happens between theory and policy implementation to achieve more than modest outcomes. He added that how particular recommendations are translated into policy, and policy into practice, may have unintended outcomes. Several participants noted that the wider the policy application, and the purer the theory in application, the more likely it is to fail.

What Drives Change?

The group agreed that policy changes are unlikely to have more than incremental effects. How, then, do significant changes (such as the crime rate decline in the 1990s) happen naturally, but cannot be achieved in a directed fashion? Is it because the natural variables that are driving the change are radically different from those the policymakers have control over? A participant suggested that investigation into what changes were policy-influenced might provide some answers to this question. For example, the rise in mass incarcerations correlates with declines in adult violence since the 1970s. However, youth violence has also declined, and people under age 18 are not subject to incarceration.

A participant suggested that changes in the social structure that are ubiquitous in society might not be amenable to policy changes. It is important to look at areas where intervention is possible and decide which are amenable to different policy implications. Another participant observed that the most dramatic changes are made by bankers and businesses, but there can be government regulation of those processes.

Corporate Violence

Dr. Lynch noted that, for the purpose of this conference, violence had been defined differently than his focus: corporate violence (e.g., white collar crime such as Enron). Criminologists usually do not look at violence in this framework. There has not been a concomitant decline in corporate violence, and it is not clear that the general theories address this type of violence.

Violence and Minorities

Dr. Peterson discussed the application of the theories to the effects of violence on minorities. Theories are distant from what she does, she said, but the ideas presented have enabled her to think of the application of a variety of sociological theories to different forms of violence. She had not previously considered violence as a response to provocation, or moralistic versus predatory versus celebratory violence, or the event or conflict as the unit of analysis, or violence perpetration as not defining the perpetrator. She was concerned, she said, that race, class, community, and social context were not part of the analysis.

Dr. Peterson further noted that the patterns of violence across racial and ethnic groups in the United States differ. The theoretical applications are clear for individuals and nation-states, but not as clear for groups, communities, and the effects of institutional factors such as economic and political conditions and the criminal justice response. She is convinced that theorists need to base theories and substantive work not exclusively on the statistical pattern of the violence, but also on discussions with the population that is most affected. Their perceptions of events might be quite different from what the statistics imply. In this way, theorists and practitioners can better understand the population's responses to provocation.

Back To Top


Dr. Hartmann observed that the group had taken four steps that had been suggested to accommodate other ideas: listen, understand, appreciate, and incorporate. Dr. Zahn considered the workshop a success, in that people who are not often or ever together came together to discuss these issues. She encouraged participants to incorporate each other into networks, and let others' work influence their own future research.

Dr. Zahn outlined how the workshop papers might be disseminated in order to bring more theoretical attention to the issue of violence. NIJ may publish a book or consider a special issue of a journal. Panel participants agreed to make any revisions that might be necessary for these potential publications.

Dr. Feucht concluded that policy is a blunt instrument; policymakers often take concepts or ideas out of their framework. He maintained that although practitioners cannot operate with theoretical purity in a real-world environment, they must understand the subtlety and complexity of the situations they are confronting.

To extend this workshop's conversation to the field, NIJ will sponsor a second meeting with policymakers and practitioners.

Dr. Feucht thanked participants for coming and Dr. Zahn for coming up with the idea for the workshop, then adjourned the meeting.

Back To Top

1 As defined by Donald Black, moralistic violence defines and responds to deviant behavior. It is a form of social control. When ordinary members of society use violence moralistically and unilaterally, it is social control through self-help. Examples include a person who attacks his or her spouse for being unfaithful, a person who takes homicidal vengeance on someone for killing his brother, a person who beats his or her child for being disrespectful, a teenager who attacks a peer for insulting him, or a group that riots or uses terrorism in the context of a grievance.

Predatory violence is the use of physical force in the acquisition of wealth or other resources, such as money, property, sexual access, or authority. Examples include various kinds of robbery such as an attack followed by the seizure of property (street "mugging" and other surprise assaults to obtain property) or the threat to use violence if money or valuables are not surrendered, the use of violence to obtain compliance other than a demand for money or other valuables (including rape and "bullying"), the use of violence to escape detection or confinement after the commission of a crime (an acquisition of freedom), and the use of violence by a nation-state to seize the territory of another nation-state.

For further discussion of these concepts, see Black, Donald, "Crime as Social Control," American Sociological Review 48 (1983): 34–45

Date Created: November 28, 2007