National Academies Press: OpenBook

Understanding and Preventing Violence, Volume 3: Social Influences (1994)

Chapter: Violent Victimization and Offending: Individual-Situational-, and Community-Level Risk Factors

« Previous: Front Matter
Suggested Citation:"Violent Victimization and Offending: Individual-Situational-, and Community-Level Risk Factors." National Research Council. 1994. Understanding and Preventing Violence, Volume 3: Social Influences. Washington, DC: The National Academies Press. doi: 10.17226/4421.
×

Violent Victimization and Offending: Individual-, Situational-, and Community-Level Risk Factors

Robert J. Sampson and Janet L. Lauritsen

INTRODUCTION

The purpose of this paper is to (1) summarize current knowledge on individual-, situational-, and community-level sources of criminal violence; (2) identify key problems of causal interpretation in existing research; and (3) suggest new directions for future research and policy. Although violence has not been an understudied phenomenon in American criminology, a synthesis of the causes and consequences of criminal violence across multiple levels of analysis has not been undertaken. Indeed, most of the more than 2,000 studies of violence published since 1945 (Bridges and Weis, 1989:14) have been descriptive and focused either on individual-level correlates of violent offending or, to a much lesser extent, on community-level correlates of violence rates. As a consequence, extant reviews tend to emphasize individual-level factors (especially age, race, and sex) or particular classifications such as family violence (e.g., Ohlin and Tonry, 1989) and criminal career violence (e.g., Weiner, 1989). By contrast, our goal in this review is to show how a multilevel perspective on both victimization and offending may substantially increase the understanding

Robert J. Sampson is at the Department of Sociology, University of Chicago. Janet L. Lauritsen is at the Department of Criminology and Criminal Justice, University of Missouri, St. Louis.

Suggested Citation:"Violent Victimization and Offending: Individual-Situational-, and Community-Level Risk Factors." National Research Council. 1994. Understanding and Preventing Violence, Volume 3: Social Influences. Washington, DC: The National Academies Press. doi: 10.17226/4421.
×

and control of violence. Because there appears to be little consensus in the social science literature on how multilevel factors should be defined, we describe in more detail our conceptualization of these terms.

Our use of the terms ''individual," "situational," and "community" corresponds closely to Short's (1985) definition of the "individual," "micro-," and "macro-" levels of analysis (see also Short, 1990:11). The individual level of explanation typically inquires as to characteristics of individuals that explain behavior. In our discussion of individual-level risk factors for violence, we specifically focus on the ascribed and achieved characteristics of individuals that are statistically associated with violent victimization and offending. For example, we examine how the risks of violent victimization and offending are distributed across characteristics such as age, sex, race, marital status, lifestyle, and socioeconomic status. Our intention, though, is not only to provide a descriptive summary of statistical findings, but to suggest how, and in what ways, individual characteristics are causally related to the risk of violent victimization and offending.

By situational-level risk factors, we are referring to those factors, broadly defined, that influence the initiation or outcome of a violent event. This conceptualization corresponds to Short's (1990:11) definition of the microlevel, where attention is focused on the unfolding of events and the interaction of parties involved in events. Situational-level analyses usually treat the violent incident, or event, as the unit of analysis. Included in our review is a discussion of such factors as the presence and type of weapon, the presence of drugs or alcohol, the role of bystanders or third parties during violent events, and victim resistance and retaliation. However, we expand the traditional conceptualization of situational-level analyses to include a discussion of the victim-offender overlap and how victim-offender relationships are related to violence. These factors are included because we believe that the simultaneous consideration of victims, offenders, and their past and present interactions provides a more complete context for studying the initiation and outcomes of violent events.

The macrosocial or community level of explanation asks what it is about community structures and cultures that produces differential rates of crime (Byrne and Sampson, 1986; Bursik, 1988; Short, 1990:11). For example, what characteristics of communities are associated with high rates of violence? Are communities safe or unsafe because of the persons who reside in them or because of community properties themselves? Can changes in community

Suggested Citation:"Violent Victimization and Offending: Individual-Situational-, and Community-Level Risk Factors." National Research Council. 1994. Understanding and Preventing Violence, Volume 3: Social Influences. Washington, DC: The National Academies Press. doi: 10.17226/4421.
×

structure affect violent crime rates? As implied by these questions, the goal of macrolevel research is not to explain individual involvement in criminal behavior but to isolate characteristics of communities, cities, or societies that lead to high rates of criminality. Examples of macrolevel risk factors include residential mobility, population density, heterogeneity, and income inequality. These and other factors are assessed for both intraurban (e.g., local community, neighborhood) and interurban (e.g., city, metropolitan area) units of analysis, although for substantive reasons discussed below we place more weight on between-community variations in rates of criminal violence.

CRITERIA

Conclusions about the correlates of violence are naturally linked to the universe of studies under consideration. Because our review is highly focused we applied systematic criteria for inclusion of research. Although exceptions arise, the studies analyzed in this paper were selected based on the following criteria:

  1. Studies in which serious interpersonal violence (e.g., assault, homicide, robbery, and rape) is measured directly are our primary focus. Issues surrounding the measurement of violence are discussed as needed; however, we should note that our intention herein is to avoid studies that present overall measures (e.g., total scales, index crimes) in which violence is confounded with property or other crime types (e.g., drug use, status offenses). Also, many delinquency studies focus on such events as schoolyard theft ("robbery") and minor fighting. Except where important for historical or theoretical reasons, these studies are generally not considered.

  2. Analyses of objective social characteristics of individuals, situations, and community context are emphasized, as opposed to those studies focusing on internal psychological states, perceptions, or biological/constitutional conditions.1 Because individual-level social characteristics have received disproportionate attention in the past, we tend to highlight the theoretical and policy implications of community-level factors. As we make clear, however, the study of community-level processes is not possible without a thorough understanding of individual-and situational-level correlates.

  3. We emphasize studies that present empirical data, especially those that have representative (e.g., national-level) or large

Suggested Citation:"Violent Victimization and Offending: Individual-Situational-, and Community-Level Risk Factors." National Research Council. 1994. Understanding and Preventing Violence, Volume 3: Social Influences. Washington, DC: The National Academies Press. doi: 10.17226/4421.
×

enough samples to draw reasonable conclusions. However, this does not rule out the selective use of important ethnographic work that bears directly on issues of concern.

  1. We focus on American studies or those written in the English language, particularly those published in refereed journals. Moreover, we emphasize urban-based research from the past 20 years that has formed the bulk of multivariate studies of violence.

  2. In general, we attempt to assess the substantive and (where appropriate) causal importance of key factors. Consequently, we focus on the magnitude of effects as opposed to mere statistical significance.

Although not strict criteria for inclusion, we also make a special effort to highlight (a) micro-macro linkages (i.e., the simultaneous consideration of both individual-and community-level risk factors in violence) and (b) the connection between violent victimization and violent offending. We believe that much previous research on violence is problematic because of the causal confounding and alternative causal interpretations that can arise in designs relying on a single level of analysis. Furthermore, we contend that the victim-offender overlap can provide fundamental insights into the etiology of violence; hence we pay close attention to studies that allow assessment of the causal role that offending may play in increasing victimization risk, and vice versa. Too often in the past these issues have been treated separately as if they were unrelated (see Reiss, 1986a; Sampson and Wooldredge, 1987; Sampson and Lauritsen, 1990). In contrast, by highlighting micro-(individual/ situational) and macro-(community) linkages in the study of violence, in addition to victim-offender relationships, we believe that our understanding of the social structural and contextual sources of violence is improved. To our knowledge the literature on violence has not been critically assessed from this perspective before, especially in a framework that points to future research designs needed to address remaining etiological questions.

Before turning to a road map outlining our review, two issues deserve further clarification. Given the vast literature and multiple measures of both violence and community factors, we do not present in tabular form the magnitudes of effect for all studies and factors-preliminary efforts found this to be unworkable. As noted above, the Bridges and Weis (1989) review located more than 2,000 studies of violence-a universe larger than that relevant

Suggested Citation:"Violent Victimization and Offending: Individual-Situational-, and Community-Level Risk Factors." National Research Council. 1994. Understanding and Preventing Violence, Volume 3: Social Influences. Washington, DC: The National Academies Press. doi: 10.17226/4421.
×

to other comprehensive reviews (e.g., see Visher and Roth, 1986). Perhaps more of a problem, we discovered an almost unlimited supply of incomparable measures and incomparable techniques, especially in community-level studies. For example, a concept as seemingly straightforward as "poverty" has been measured in at least 20 different ways, ranging from dichotomous groupings of communities through census-based percentage measures to factor-analytic scores. Complicating matters further, researchers have employed many different analytical techniques and styles of data presentation. This issue is discussed more below, but we emphasize at the outset that the nature of past research makes it difficult to present summary figures representing magnitudes of effect that are comparable across studies.

Second, there is a plethora of theories of violence that have been advanced in the criminological literature. Some of these, such as the subculture of violence theory (Wolfgang and Ferracuti, 1967), are well known and have been explicated quite well in other sources. A seemingly reasonable strategy, then, especially in a review devoted to empirical risk factors in violence, is simply to ignore theory and concentrate on the findings. However, findings alone do not speak for themselves and are generally devoid of meaning when stripped of substantive content. More to the point, empirical research designs are valid only to the extent that the substantive model under consideration is valid and properly specified. Our compromise is therefore to couch, no matter how briefly, the correlates of violence in substantive perspective. We refer readers to appropriate sources on theoretical issues for more detail, particularly individual-level theories that are more widely known. Thus, theories per se are not reviewed, but key substantive themes are emphasized in an effort to shed light on the importance of empirical findings, especially multivariate, community-level, or contextual studies.

This review is divided into seven major sections. First, the literature on factors underlying an individual's risk of violent victimization is assessed. This section includes a review not only of demographic correlates (e.g., age, race, and sex) emphasized in previous research, but of lifestyle and routine activity factors as well. In a similar fashion, the second section reviews the individual-level correlates of violent offending. In the third section, an overview is presented of situational-level risk factors in violence, along with analyses that consider victim-offender relationships and the concept of a victim-offender overlap. The fourth and fifth sections assess community-level determinants of violent

Suggested Citation:"Violent Victimization and Offending: Individual-Situational-, and Community-Level Risk Factors." National Research Council. 1994. Understanding and Preventing Violence, Volume 3: Social Influences. Washington, DC: The National Academies Press. doi: 10.17226/4421.
×

crime and metropolitan (i.e., city-level) sources of violence, respectively. In the sixth section, some key problems and pitfalls in attributing causal interpretations to findings from past research are examined. Finally, the last section identifies promising directions for future research designs and public policy initiatives.

INDIVIDUAL-LEVEL RISK FACTORS IN VIOLENT VICTIMIZATION

Our focus in this section is on factors related to the risk of violent victimization among individuals. Here, we address the following questions: What are the overall risks of violent victimization to individuals? How are these risks distributed across individual characteristics such as age, sex, race, marital status, and socioeconomic status? Are these differences stable over recent time periods? How are individual-level differences in risk explained by theories of victimization? Finally, how might some of the limitations of extant victimization research be minimized? We begin with a brief overview of the general aspects of violent victimization.

Victimization research over the past 20 years has produced several consistent findings. One of the primary findings is that the annual risk of becoming a victim of a crime of personal violence (i.e., homicide, rape, robbery, or assault) is relatively low (Gottfredson, 1986; U.S. Department of Justice, 1989). For example, the combined risk of suffering a violent victimization by either rape, robbery, or assault in 1987 was estimated at approximately 1 in 33.2 The risk of suffering a victimization from a personal crime of violence is considerably less than that of suffering a theft victimization-in 1987, the rate of reported theft victimization (approximately 1 in 14) was more than twice the rate of reported violent victimization (U.S. Department of Justice, 1989). Although this risk may seem low on an annual basis, the lifetime risk of violence for any particular crime is much higher. For example, the annual risk of becoming a victim of homicide is about 1 in 10,000, but the lifetime risk of being murdered is approximately 1 in 369 for white females, 1 in 131 for white males, 1 in 104 for black females, and as high as 1 in 21 for black males (U.S. Department of Justice, 1985; see also Fingerhut and Kleinman, 1990).

As suggested by the differences in lifetime homicide risks, the second major finding in individual-level analyses is that the distribution of violent victimization is not random. Rather, the risk

Suggested Citation:"Violent Victimization and Offending: Individual-Situational-, and Community-Level Risk Factors." National Research Council. 1994. Understanding and Preventing Violence, Volume 3: Social Influences. Washington, DC: The National Academies Press. doi: 10.17226/4421.
×

of experiencing personal violence varies considerably across demographic and social groups (Hindelang, 1976; Hindelang et al., 1978; Cohen et al., 1981a,b; Gottfredson and Hindelang, 1981; Skogan, 1981; Sparks, 1982; Gottfredson, 1986; Miethe et al., 1987; U.S. Department of Justice, 1988a). The pattern of systematic differences found across different demographic and socioeconomic subgroups has been fairly consistent across analyses utilizing either the Uniform Crime Report (UCR) data or the National Crime Survey (NCS) data in the United States (e.g., see Gottfredson, 1986). Patterns of association similar to those found in the United States have also been found in analyses using data from other countries such as England (Hough and Mayhew, 1983), the Netherlands (van Dijk and Steinmetz, 1983), and Australia (Braithwaite and Biles, 1984). To describe these patterns of violent victimization, we present, as briefly as possible, bivariate and then multivariate findings of research examining individual-level risk factors associated with victimization by homicide, assault, robbery, and rape. Particular attention is paid to the magnitude and consistency of findings in multivariate research, as well as their theoretical importance. Limitations of these analyses and their consequences for our understanding of violent victimization are discussed.

BIVARIATE FINDINGS

It should be mentioned at the outset that the vast majority of knowledge about the risk of violent victimization has been obtained through analyses of National Crime Survey data. The NCS is an ongoing survey conducted by the Bureau of Justice Statistics, designed to measure the extent of personal and household victimization in the United States. Interviews are conducted at six-month intervals with all persons 12 years of age or older living in a household. By using NCS data, it is possible to estimate the extent of victimizations that are not reported in official police data (due to nonreporting of incidents, arrest bias, etc.) and to calculate prevalence measures of violent victimization for specific individual subgroups (see Reiss, 1981; Langan and Innes, 1990). Therefore, the NCS constitutes the best available data source on the risk of victimization for persons living in the United States.3

Age

Age is one of the most important predictors of an individual's risk of violent victimization (Hindelang et al., 1978). The NCS

Suggested Citation:"Violent Victimization and Offending: Individual-Situational-, and Community-Level Risk Factors." National Research Council. 1994. Understanding and Preventing Violence, Volume 3: Social Influences. Washington, DC: The National Academies Press. doi: 10.17226/4421.
×

data for 1987 indicate that the risk of personal violent victimization is highest for persons under 25 years of age (U.S. Department of Justice, 1989). After age 25, the risk of such victimization gradually declines such that by age 65, the risk of suffering a violent victimization is approximately one-tenth that of persons under 25 years of age. This bivariate age-victimization relationship-a relatively high level of victimization in youth and young adulthood followed by a gradual decline in risk-is observed for the crimes of homicide, rape, and assault (and, therefore, violent victimization in general). Differences in the risk of robbery victimization across age groups are not as great as those found for other violent crimes (Hindelang et al., 1978; Cohen et al., 1981b).

In addition to contributing to the overall risk of violent victimization, age appears to be related to the nature of crime-specific victimizations. In an analysis of NCS victimization narratives for juvenile victims (age 12-17) of personal crime, Garofalo et al. (1987) found that assaults reported by juveniles were overwhelmingly minor in nature-that is, they were often the result of an inability to resolve common disputes and rarely led to serious injury on the part of the victim. For robbery as well, the relative nature of the event was reported to be less serious among juvenile victims than among adults. Although a large proportion of all victimization reported in the NCS is relatively minor in nature (Skogan, 1981; Sparks, 1982), juvenile victimization appears to be least serious overall in terms of both injury levels and monetary loss. Therefore, although juveniles appear to be at high risk for violent personal victimization, the victimizations they do suffer are more likely to have minor consequences compared to those of adults.4

Sex5

For all crimes of personal violence--with the obvious exception of rape--males are at greater risk for victimization than females (U.S. Department of Justice, 1989). The annual rate of homicide victimizations for males is nearly three times greater than the annual rate for females (U.S. Department of Justice, 1986). The lifetime risk of becoming a victim of homicide is estimated to be between three and four times greater for males (1 in 84) than females (1 in 282) (U.S. Department of Justice, 1985). For the crimes of robbery and assault, males are approximately 1.5-2 times more likely to become victims. On the other hand, women are overwhelmingly the victims of rape. NCS data estimate the risk

Suggested Citation:"Violent Victimization and Offending: Individual-Situational-, and Community-Level Risk Factors." National Research Council. 1994. Understanding and Preventing Violence, Volume 3: Social Influences. Washington, DC: The National Academies Press. doi: 10.17226/4421.
×

of attempted or completed rape for females to be approximately 1 in 770 annually (U.S. Department of Justice, 1989). Furthermore, these relationships between sex and the risk of violent victimization have remained fairly constant over the last decade. Smith (1987), for example, reports that the proportion of female victims of robberies and assaults has remained stable, whereas the proportion of female homicide victims has increased slightly over the past 10 years.

As was true for the relationship between age and overall risk of violence, the nature of victimization for males and females differs. First, for assault and robbery victimizations, men are more likely than women to report injury (U.S. Department of Justice, 1989).6 Second, according to victim reports, offenders attacking males are more likely to be armed: between 1973 and 1982, approximately 41 percent of male and 29 percent of female victims of violent crime were attacked by offenders with weapons (U.S. Department of Justice, 1988a). Rape is an obvious exception to this pattern. Excluding homicide, rape is the crime with the highest level of injury (Hindelang et al., 1978) and is perpetuated largely against females.

Race

Both NCS and UCR data confirm that blacks are disproportionately the victims of homicide, rape, and robbery (U.S. Department of Justice, 1988a). Blacks are approximately five times more likely than whites to become victims of homicide (U.S. Department of Justice, 1986) and approximately three to four times more likely to report victimization by rape and robbery (U.S. Department of Justice, 1989). Higher rates of black victimization by homicide (compared to whites) have been found throughout the last century (Wolfgang, 1958; Block, 1975), and recent trends indicate a continuation of this differential (Shin et al., 1977; Farley, 1980; Hawkins, 1985). As noted earlier, the lifetime risk of homicide victimization is approximately three times greater for black women than for white women, and approximately four times greater for black men than for white men (U.S. Department of Justice, 1985).

Assault victimization rates for blacks and whites do not differ significantly in magnitude, although the majority of assault victimizations reported by blacks are incidents of aggravated assault, whereas among whites, simple assaults comprise the majority (U.S. Department of Justice, 1989). The lack of race differences in

Suggested Citation:"Violent Victimization and Offending: Individual-Situational-, and Community-Level Risk Factors." National Research Council. 1994. Understanding and Preventing Violence, Volume 3: Social Influences. Washington, DC: The National Academies Press. doi: 10.17226/4421.
×

assaults overall may be the result of differences in reporting. For example, it has been suggested that blacks underreport less serious forms of assault and/or that whites overreport minor assaults (Skogan, 1981; Gottfredson, 1986).

Two-and three-way associations for age, sex, and race suggest that the relationships among these factors are, generally speaking, additive. Thus, for certain subgroups the relative risk of violent victimization is quite high. In particular, young, black males are approximately 22 times more likely than older, white females to become the victims of violent crime (U.S. Department of Justice, 1989). Within these three-way associations, age continues to have the largest independent effect followed by sex and race.

Marital Status

Although somewhat less studied than age, race, and sex, marital status has also been found to be substantially related to the risk of violent victimization (Hindelang et al., 1978; Cohen et al., 1981b; Miethe et al., 1987; U.S. Department of Justice, 1989). Overall, NCS estimates suggest that unmarried persons are nearly four times more likely to become the victims of robbery, rape, or assault than are those currently married (U.S. Department of Justice, 1989). The exception to this pattern is for widowed persons who have the lowest risk of violent victimization. Never-married persons have the highest risk of victimization, followed closely by divorced and separated persons (U.S. Department of Justice, 1989). At the extreme ends of the scale, those never married are approximately 11 times more likely than widowed persons,7 and more than 3 times more likely than currently married persons, to become the victims of assault. On the other hand, the differences by marital status for robbery and rape are not as pronounced (U.S. Department of Justice, 1989). Finally, when sex and marital status are considered simultaneously, marital status is more strongly related to victimization risk among females than among males (U.S. Department of Justice, 1989).

Socioeconomic Status

Not surprisingly, family income is inversely related to the risk of personal victimization (Hindelang, 1976; Hindelang et al., 1978; Cohen et al., 1981a,b; Miethe et al. 1987; U.S. Department of Justice, 1989). However, the overall magnitude of the effect of family income is not as large as the magnitudes for the individual-level

Suggested Citation:"Violent Victimization and Offending: Individual-Situational-, and Community-Level Risk Factors." National Research Council. 1994. Understanding and Preventing Violence, Volume 3: Social Influences. Washington, DC: The National Academies Press. doi: 10.17226/4421.
×

characteristics previously discussed (namely, age, race, sex, and marital status). For violence in general, the risk of victimization is approximately twice as great for persons whose family income is reported to be less than $7,500 compared to those whose family income is higher than $50,000 (U.S. Department of Justice, 1989). Specific crimes of violence also have varying relationships to family income. For example, there is little relationship between rape and family income, although women whose families earn less than $7,500 are again most likely to become victims. The risk of rape among women in the remaining income categories is approximately equal. Income is somewhat more strongly related to the risk of assault, with those at the lowest end of the income distribution reporting twice the risk of victimization as those at the highest end. Of all violent crimes, income is most strongly related to the risk of robbery-poorer persons report a robbery rate about three times higher than persons at the upper ends of the income distribution (U.S. Department of Justice, 1989).

The relationship between education and violent victimization is less consistent than the relationship between family income and personal victimization (U.S. Department of Justice, 1989). In fact, persons with the least amount of education report the lowest level of violent victimization. Excluding this category of education, and with slight fluctuations, the general relationship between education and victimization is negative but small in magnitude. It has been suggested that the ability to accurately recall victimizations may be affected by one's level of education. For example, highly educated respondents in the NCS appear to be more "productive" when recalling victimization events (Skogan, 1981). The relationship between socioeconomic status (either income or education) and rape victimization is also small in magnitude and is more uniformly distributed across women in different income and education groups (U.S. Department of Justice, 1989).

Finally, of the common indicators of an individual's socioeconomic status, unemployment is most strongly related to the risk of violent victimization (U.S. Department of Justice, 1989). Those who are unemployed are at the greatest risk of robbery, assault, and rape, followed by persons in school, the employed, those unable to work, those keeping house, and the retired.8 Overall, the unemployed are more than twice as likely as the employed to suffer a violent victimization (U.S. Department of Justice, 1989).

It is thus clear that the risk of suffering a violent victimization is differentially distributed across social and demographic groups. Extant data suggest that age has the largest bivariate

Suggested Citation:"Violent Victimization and Offending: Individual-Situational-, and Community-Level Risk Factors." National Research Council. 1994. Understanding and Preventing Violence, Volume 3: Social Influences. Washington, DC: The National Academies Press. doi: 10.17226/4421.
×

relationship with the risk of homicide, assault, robbery, and rape. Sex and race are also related to violent victimization, as is marital status, although these effects are slightly smaller in magnitude. Two-way relationships among most forms of violent victimization and each of the indicators of socioeconomic status are generally negative in direction, but even smaller in magnitude. Although each of the individual-level characteristics discussed above appears related to victimization, it is possible that these associations are spurious. To assess this possibility, we direct our attention to the research literature that has considered some or all of these individual-level factors simultaneously.

MULTIVARIATE FINDINGS

Despite the rather large body of research on the individual-level risk of victimization, there have been relatively few multivariate analyses that directly examine the risk of violent victimization in the United States. This is due in part to the relative rarity of violent crime: because of its infrequency, it is difficult to test reliably for the independent effects of many of the above-mentioned correlates. Nevertheless, some major studies have examined the interrelationships among demographic and socioeconomic (as well as other) characteristics and violent victimization in the United States (e.g., Hindelang et al., 1978; Cohen et al., 1981a,b; Miethe et al., 1987). More specifically, Hindelang et al. (1978) analyzed the 1972 NCS (eight-city survey) data in order to examine the independent effects of social and demographic factors on whether or not a person reported a personal victimization (defined as assault, robbery, rape, and personal larceny with contact). Cohen et al. (1981a) examined robbery victimization using 1973 to 1977 NCS data combined, and Cohen et al. (1981b) analyzed similar factors for their effects on the risk of assault victimization using 1974 and 1977 NCS data combined. Miethe et al. (1987) also examined a set of demographic and socioeconomic factors for their effects on violent victimization9 using the 1975 NCS (13-city survey) data.

Although the dependent variables and statistical techniques differed across studies, the general patterns and relative magnitudes of effects found in each of the analyses are compatible. Hindelang et al. (1978) for example, found that the most important predictors of personal victimization were age, marital status, sex, and employment status: young, single, unemployed males had the highest risk of personal victimization.10 Cohen et al.

Suggested Citation:"Violent Victimization and Offending: Individual-Situational-, and Community-Level Risk Factors." National Research Council. 1994. Understanding and Preventing Violence, Volume 3: Social Influences. Washington, DC: The National Academies Press. doi: 10.17226/4421.
×

(1981a) found that age, employment status, race, and income had significant and independent effects on robbery victimization.11 Namely, younger persons, the unemployed, blacks, and poorer persons had the highest risk of being robbed. Cohen et al. (1981b) also found that age, marital status, and employment status had the greatest independent effects on the probability of assault. Finally, Miethe et al.'s (1987) findings agree: age, sex, and marital status are the predominant demographic risk factors predicting violent victimization. Miethe et al. (1987) also found a relatively small but significant negative effect for income. Only in the case of robbery victimization was race found to be a significant predictor (Cohen et al., 1981a). The lack of an independent race effect on assault is not surprising when one considers the negligible bivariate effect found in the NCS data.12

Direct comparisons of the magnitude of effects found in prior analyses of violent victimization are not possible given the variety of statistical controls and analytical techniques. However, in the analyses noted above, partial effects were most often no more than twice the risk of the reference group-a magnitude that is determined in part by the measurement level of the independent factors. Miethe et al. (1987), for example, found that males were approximately twice as likely as females to suffer a violent victimization, independent of race, income, marital status, and age. However, certain combinations of these risk factors produce much higher risks of victimization for some subgroups. In particular, Cohen et al. (1981a) estimated that young, poor, unemployed black males living alone had a risk of robbery victimization nearly 10 times greater than the national average.

THEORETICAL INTERPRETATIONS

These individual-level correlates of victimization have formed the basis of ''lifestyle" (Hindelang, 1976; Hindelang et al., 1978; Garofalo, 1987) or "routine activity" (Cohen and Felson, 1979; Cohen et al., 1981b) theories of victimization risk.13 The essential proposition of "lifestyle-routine activity" theories is that the convergence in time and space of suitable targets and the absence of capable guardians may lead to increases in crime, independent of the structural and cultural conditions that may motivate individuals to engage in crime (e.g., poverty, unemployment, subcultural values). Derived from this general proposition, the "principle of homogamy" in lifestyle theory states that persons are more likely to be victimized when they disproportionately associate

Suggested Citation:"Violent Victimization and Offending: Individual-Situational-, and Community-Level Risk Factors." National Research Council. 1994. Understanding and Preventing Violence, Volume 3: Social Influences. Washington, DC: The National Academies Press. doi: 10.17226/4421.
×

with, or come in contact with, members of demographic groups that contain a disproportionate share of offenders (Hindelang et al., 1978:256-257). Young single males, for example, are more likely to pursue social activities outside the household than are older, married, or female persons, and hence are more likely to place themselves in situations where victimization exposure is increased.14 Moreover, younger persons may suffer a higher risk of violent victimization than older persons because the former are more likely to associate with other youth who are themselves disproportionately involved in violence (Hindelang et al., 1978).

The limitations of lifestyle-routine activity theory and research have been discussed at length elsewhere (see, for example, Gottfredson, 1984; Maxfield, 1987; Garofalo, 1987; Miethe et al., 1987; Sampson and Wooldredge, 1987). The most common criticism of empirical research has been the inadequate measurement of explanatory variables; demographic and social characteristics of individuals are used primarily as proxies or indirect indicators for lifestyle activities that persons engage in routinely. Social and demographic characteristics are argued to have effects on the risk of victimization because they determine, to a large extent, "systematic differences in role-structured behavior" via differences in expectations, constraints, opportunities, and preferences for activities (Cohen et al., 1981a:648). The principal critique of much of the multivariate literature is thus that direct measures of lifestyle activities are not included in the models containing social and demographic characteristics (Hindelang et al., 1978; Cohen et al., 1981a,b; Gottfredson, 1986; Miethe et al., 1987; Garofalo, 1987). Consequently, the effects of routine activities or specific behaviors have not been distinctly separated from the ascribed and achieved characteristics of victims. Exactly what types of behaviors put individuals at higher risk for violence has not yet received thorough attention.

Exceptions to the lack of direct indicators of lifestyle include measures of activities such as employment status that specifically ascertain what people do during a relatively large portion of their day. The extent of nighttime activities away from the home is also considered to be a direct measure of lifestyle in that activities away from home and activities typically occurring at night are assumed to be more likely to involve either decreased guardianship or increased contact with offenders. These factors-nighttime and daytime activities away from the home-were considered by Miethe et al. (1987) in their analysis of violent victimization.15 They found that although extent of nighttime activity significantly and positively influenced risk of violent victimization, the

Suggested Citation:"Violent Victimization and Offending: Individual-Situational-, and Community-Level Risk Factors." National Research Council. 1994. Understanding and Preventing Violence, Volume 3: Social Influences. Washington, DC: The National Academies Press. doi: 10.17226/4421.
×

effect was smaller in magnitude than most of the other demographic variables considered. Furthermore, the inclusion of these activity measures did not significantly attenuate the magnitude of the effects of sex, income, marital status, and age. Major daytime activity as measured was found to have no significant independent effect on risk of violent victimization. In light of these findings, Miethe et al. (1987:188) conclude that there is "little support for the mediational role of routine activity/lifestyle variables on the demographic correlates of violent victimization." By contrast, there were much larger effects of lifestyle on theft victimization.

These findings are intriguing but not definitive. First, nighttime activity was measured as a dichotomy and consequently may underestimate victimization risk for persons who engage in night activities most often. Second, as measured, this variable was not able to capture the nature of away-from-home activities, nor whether particular activities were engaged in alone or with strangers, friends, or family. Third, the measurement of the major activity variable was constructed on the basis of predictability of behaviors. In other words, persons in school were included in the same category as the employed because they presumably depart and return to their homes at predictable times during the day-an element of routine activities assumed to be of interest to motivated offenders. However, as analyzed, the comparison group for major activity included homemakers, retirees, those unable to work, and the unemployed. Thus, included in one category were those with high risks (the unemployed) and low risks (homemakers and retirees). Measurement problems, then, may have been responsible for the lack of effect of major daytime activity on the risk of violent victimization.

More direct tests of lifestyle-routine activity hypotheses have been conducted with data from countries other than the United States (Corrado et al., 1980; van Dijk and Steinmetz, 1983; Hough and Mayhew, 1983; Braithwaite and Biles, 1984; Gottfredson, 1984; Sampson, 1987b; Sampson and Lauritsen, 1990; Kennedy and Forde, 1990; Miethe and Meier, 1990). Similar examinations have not been conducted for the general U.S. population, primarily because the major source of victimization data-the NCS-does not include measures of routine activities detailed enough for adequate assessment of lifestyle-routine activity hypotheses (but see Lynch, 1987, for an exception with regard to victimization at work). Indeed, the lack of direct measures coupled with the rarity of personal

Suggested Citation:"Violent Victimization and Offending: Individual-Situational-, and Community-Level Risk Factors." National Research Council. 1994. Understanding and Preventing Violence, Volume 3: Social Influences. Washington, DC: The National Academies Press. doi: 10.17226/4421.
×

violence has made lifestyle-routine activity hypotheses of violence extremely difficult to test with U.S. data.

On the other hand, direct measures of routine activities derived from foreign data sets-especially the number of nights per week out alone, type of activity typically engaged in at night, mode of transportation used, extent of drinking, and self-reported offending behaviors-have been significantly linked to the risk of personal victimization. Corrado et al. (1980), for instance, found a similar demographic pattern in Canadian data compared to that in the United States. However, the inclusion of a crude measure of out-of-home activities (divided at the median) had a significant effect on the risk of violent victimization (robbery, assault, and sexual assault): persons who spent more time away from home were approximately three times more likely to be victimized than persons who centered their time at home. The inclusion of this variable did not attenuate the effects of sex and age but did decrease the magnitude of the effect of marital status. Smith (1982) similarly finds that those out-of-home activities that increase contact with strangers are most victimogenic. Similarly, Hough and Mayhew (1983), van Dijk and Steinmetz (1983), and Gottfredson (1984) found that frequency and type of nighttime activities are related to the risk of violence. In particular, Gottfredson (1984) reports that the use of public transportation and nighttime activities such as drinking or going out to pubs increase personal victimization in Great Britain whereas activities such as attending church are associated with lower risk. Also, using data from Great Britain, Hough and Mayhew (1983) found that the frequency of nighttime activity combined with the extent of drinking was important for determining the level of risk among persons under 30 years of age. Finally, Kennedy and Forde (1990), using Canadian data, examined the relative effects of a variety of routine activities on assault and robbery victimization. They report that certain types of activities-especially how often one goes to bars or takes walks at night-are directly related to the risks of violent victimization. Furthermore, these effects persisted despite statistical controls for traditional demographic factors (age, sex, family income) and neighborhood characteristics (e.g., percent single person households and percent low income families).

In sum, recent analyses strongly suggest that details concerning the type and extent of daytime and nighttime activities are necessary for adequately assessing lifestyle-routine activity models of violent victimization. Previous analyses using the NCS data have been limited because of the unavailability of measures

Suggested Citation:"Violent Victimization and Offending: Individual-Situational-, and Community-Level Risk Factors." National Research Council. 1994. Understanding and Preventing Violence, Volume 3: Social Influences. Washington, DC: The National Academies Press. doi: 10.17226/4421.
×

such as the type of nighttime activities engaged in, the use of public transportation, drinking, and other self-reported deviance such as offending (see more below). It is thus possible that the patterns regarding direct effects of lifestyle are unique to countries such as Great Britain, where cultural and structural conditions (e.g., racial composition) are quite different. Consequently, we can only speak generally about individual-level lifestyle factors associated with the risk of violent victimization for the U.S. population as a whole, and we are unable to specify precisely why it is that traditional social and demographic factors (including age, sex, race, and employment status) are related to differences in the risk of violent victimization. Clearly, future research on violence needs to focus more specifically on the development of individual-level measures necessary to test lifestyle-routine activity hypotheses using U.S. data.

INDIVIDUAL-LEVEL RISK FACTORS IN VIOLENT OFFENDING

Having reviewed the individual-level factors underlying the risk of violent victimization, we now turn our attention to the issue of violent offending. This section is similar in scope to the previous section in that we review basic individual-level correlates. For example, we focus on questions such as: How are the individual-level factors of age, sex, race, socioeconomic status, and family structure related to the likelihood of committing violent crimes? Are these patterns consistent across sources of data? Have they remained stable over time? Also, what are the current limitations of individual-level analyses of violent offending?

Prior research on individual-level correlates of offending has discussed in great detail the methodological issues that may limit the generalizability of findings (e.g., Hindelang, 1978; Hindelang et al., 1979; Hindelang, 1981; Blumstein et al., 1986); therefore, we will mention only a few of these limitations here. One of the primary concerns is the source of data on offending-whether or not the findings are based on official, self-report, or victimization data. Findings based on official data (e.g., UCR arrest data or imprisonment data) are limited to the extent that apprehended offenders differ in some way from nonapprehended offenders, whereas findings based on self-report surveys may be limited either by the respondents' intentional or unintentional errors in reporting or by sampling restrictions (e.g., an almost exclusive focus on juveniles or males). A third source of data-the NCS-also provides information

Suggested Citation:"Violent Victimization and Offending: Individual-Situational-, and Community-Level Risk Factors." National Research Council. 1994. Understanding and Preventing Violence, Volume 3: Social Influences. Washington, DC: The National Academies Press. doi: 10.17226/4421.
×

on the perceived age, race, and sex of offenders; however, it should be reiterated that victims have been found to underreport certain types of incidents, especially those involving victimizations by family members and acquaintances (Law Enforcement Assistance Administration, 1972). These potential sources of error should be kept in mind as we review the correlates of violent offending.

DEMOGRAPHIC FACTORS

Age

Age is one of the major individual-level correlates of violent offending (Hindelang, 1981; Hirschi and Gottfredson, 1983; Greenberg, 1985; Farrington, 1986; Steffensmeier et al., 1989). In general, arrests for violent crime peak around age 18 and decline gradually thereafter (U.S. Department of Justice, 1988a). More than two-thirds of those arrested for violent crimes are age 30 or younger (U.S. Department of Justice, 1988a). The relationship between age and violent crime is less peaked than that for age and property crime in that a larger proportion of violent crimes are committed by persons over age 20 (Hindelang et al., 1981; Farrington, 1986; Jamieson and Flanagan, 1989:489; Steffensmeier et al., 1989). In other words, the age-specific rate of violent offending declines more gradually, and after a later peak, than the age-specific rate of property offending (see U.S. Department of Justice, 1988a; Steffensmeier et al., 1989). Victim reports of offender characteristics in NCS data support the age patterns found in UCR data: victims estimate that the majority of (lone) violent offenders are less than 30 years of age (Jamieson and Flanagan, 1989:308).

Steffensmeier et al. (1989) also report that there are differences in age distributions within types of violent crime. Analyzing 1980 UCR data, they report that the mean age of offenders arrested for robbery (17 years) is lower than the mean age of homicide (19 years) or assault (21 years) arrestees. Personal offenses against family members peak even later, at about 24 years old (Steffensmeier et al., 1989:815). Furthermore, the age distribution for robbery is much more peaked than the normal distribution whereas age distributions for other violent offenses are flatter than normal. These analyses suggest that robbery offending declines more rapidly with age than does assault or homicide offending, especially when these crimes are committed against family members.

Suggested Citation:"Violent Victimization and Offending: Individual-Situational-, and Community-Level Risk Factors." National Research Council. 1994. Understanding and Preventing Violence, Volume 3: Social Influences. Washington, DC: The National Academies Press. doi: 10.17226/4421.
×

The NCS victim reports support these findings; the estimated age distribution for lone robbery offenders is more peaked than the estimated age distributions for assault and rape offending (U.S. Department of Justice, 1989).

Sex

Even though sex is one of the strongest demographic correlates of violent offending, analyses of violence have concentrated almost exclusively on male offending (Harris, 1977; Simon and Baxter, 1989). Consequently, detailed accounts of male-female differences in violent offending are rare (see Kruttschnitt, in this volume). When sex differences are examined, males are far more likely than females to be arrested for all crimes of violence including homicide, rape, robbery, and assault (Simon, 1975; Hindelang, 1979; Hindelang et al., 1979; Smith and Visher, 1980; Hindelang, 1981; Hill and Harris, 1981; Simon and Baxter, 1989; Bridges and Weis, 1989). Females constitute, on average, only 11 percent of arrestees for all of these crimes (U.S. Department of Justice, 1988a); however, the proportion of female offenders arrested for the crime of rape (1%) is estimated to be far lower. The NCS data closely parallel findings from UCR arrest data in that approximately 14 percent of (lone) violent offenders were reported by their victims to be female (U.S. Department of Justice, 1989).

Hill and Harris (1981) compared sex-specific arrest rates for violent crimes using 1953 and 1977 UCR arrest data. They found that in 1953, males were approximately five times more likely than females to be arrested for murder. The ratios for manslaughter (12 to 1), robbery (23 to 1), and aggravated assault (5 to 1) indicate that females were much less involved in violent offending than were males in the early 1950s (1981:661). Robbery offending in particular, along with rape offending, appears to be the most disproportionately male crime. By 1977 these sex ratios had changed to 6 to 1 for murder, 9 to 1 for manslaughter, 13 to 1 for robbery, and 4 to 1 for aggravated assault. Therefore, although sex-specific rates for all crimes increased over this period, the 1977 ratios suggest that the proportion of violent crimes committed by females increased more quickly over these two decades for the crimes of manslaughter, robbery, and aggravated assault. Only for the crime of murder has there been a greater acceleration in the male crime rate than in the female crime rate (Hill and Harris, 1981:661).

Smith and Visher (1980) conducted a meta-analysis of studies

Suggested Citation:"Violent Victimization and Offending: Individual-Situational-, and Community-Level Risk Factors." National Research Council. 1994. Understanding and Preventing Violence, Volume 3: Social Influences. Washington, DC: The National Academies Press. doi: 10.17226/4421.
×

investigating the sex-crime relationship to examine the extent to which statistical associations varied according to study characteristics. They found that, overall, the relationship between sex and crime is greater when violent and officially recorded crimes are examined (versus property offenses and self-reports) and is greater among whites than nonwhites. They also found that the magnitude of the sex-crime association has decreased slightly over time, supporting the findings reported by Hill and Harris (1981). Simon and Baxter (1989) reach a similar conclusion in their analysis of cross-cultural data on female homicide offending rates: women continue to play a minor role in violent offending around the world.

Race

Race is another individual-level factor found to be substantially related to violent offending (Hindelang, 1978; Elliott and Ageton, 1980; Ageton, 1983, Bridges and Weis, 1989; U.S. Department of Justice, 1989). Although whites are arrested for the majority of all violent crimes (approximately 52%), blacks are significantly overrepresented in arrest data: blacks comprise 47 percent of violent crime arrestees yet constitute 12 percent of the U.S. population (U.S. Department of Justice, 1988a). For persons under 18 years of age, the ratio of black to white arrest rates for violent offenses is approximately 9 to 1 (Hindelang et al., 1979). The NCS victim reports support the finding of disproportionate black involvement in violent crime, but these data suggest that the race and violent crime association may not be quite as strong. Victims report that in long-offender violent crime incidents, blacks constituted about 25 percent of the offenders. This percentage varies by crime type; black representation was higher in rape and robbery incidents, whereas white representation was higher in assaults (U.S. Department of Justice, 1988a).

Hindelang (1978) has investigated the extent to which black over-representation in official data is due to differential involvement in violent crime or to differential selection into the criminal justice system via arrests by police. In comparing the distribution of arrestees by race from the 1974 UCR to the distribution of perceived race of offenders derived from the 1974 NCS, he found some evidence for the differential selection hypothesis with respect to assault and rape. Overall, though, NCS victim reports suggest that most of the race difference found in arrest rates for violence is due to greater black involvement in personal crimes,

Suggested Citation:"Violent Victimization and Offending: Individual-Situational-, and Community-Level Risk Factors." National Research Council. 1994. Understanding and Preventing Violence, Volume 3: Social Influences. Washington, DC: The National Academies Press. doi: 10.17226/4421.
×

especially robbery. For example, in 1974 both the NCS and the UCR estimate the proportion of black offenders committing robbery to be 62 percent. We compared similar data for personal violent crimes using 1987 NCS and 1987 UCR data and found much the same pattern as reported earlier by Hindelang (1978). However, we found a somewhat larger discrepancy between the two estimates of racial involvement (U.S. Department of Justice, 1988b:182-184, 1989:Table 41). For instance, the NCS estimate of the extent of black involvement in robbery in 1987 is 51 percent, whereas the UCR data show that 63 percent of robbery arrestees were black. These differences between UCR and NCS data do not necessarily indicate a change in the underlying relationship between race and violence, particularly if race is related to group versus lone offending and this changed over time. In the 1987 data there was also a higher proportion of reported "other" racial group involvement in violent crime than in 1974.

Age-Sex-Race

Each of the above ascribed characteristics of individuals combines to produce substantial variation in age-race-sex-specific offending rates. In particular, Hindelang (1981) has examined rape, robbery, and assault offending rates (estimated from 1973 through 1977 NCS data) for demographic subgroups. The subgroup with the highest rate of offending for each of these crimes was 18- to 20-year-old black males. Black males of all ages had rates of offending that were higher than those estimated for white males. Overall, offending rates were approximately three, five, and fourteen times greater among black males compared to white males for the crimes of assault, rape, and robbery, respectively. Female rates of violent offending were substantially lower than male rates with one exception-white males ages 12 to 17 had offending rates for robbery and assault that were roughly comparable to those of black females of the same age. Between ages 18 and 20, white males' violent offending rates increase (as do rates for black males), but black and white female rates decline. Apparently, females are most likely to engage in violent offending before age 17, whereas male involvement peaks somewhat later (between 18 and 20 years of age). Race differences among females for assault offending are roughly the same magnitude as those among males, whereas female race differences for robbery are less pronounced than male race differences in part because of lower level of involvement in robbery offending (Hindelang, 1981:Table 1). Laub

Suggested Citation:"Violent Victimization and Offending: Individual-Situational-, and Community-Level Risk Factors." National Research Council. 1994. Understanding and Preventing Violence, Volume 3: Social Influences. Washington, DC: The National Academies Press. doi: 10.17226/4421.
×

and McDermott (1985) report that between 1973 and 1981 there was a decline in race differences among female juvenile offenders for assault crimes and that this decline could be attributed to a decrease in black female offending.

The limitations of these findings are discussed at length by Hindelang (1981), including the concern that race-age-sex estimates are based solely on NCS victim reports of lone offender crimes and that the NCS data produce incidence rates instead of prevalence rates. This problem of incidence rates versus prevalence rates is important to the extent that each age-race-sex subgroup contains different proportions of repeat offenders. The UCR data share a similar limitation in that arrest incidents (and not offenders) are used as the unit of analysis. The potential inaccuracy of victims' reports of an offender's age, race, and perhaps sex in the NCS data is also a concern but is not considered a serious limitation. Hindelang (1981), for example, compared NCS rape victims' reports of their offender's age and race to police reports of the offender's demographic characteristics and found substantial agreement. Victims' reports of race agreed with police reports in more than 96 percent of the cases. Similarly, the accuracy of age estimates was very good (89% or higher) and was highest for offenders less than 20 (96%) and more than 45 years of age (96%). Of course, these findings are not definitive because arrests are often made on the basis of victims' descriptions of offenders.

Bridges and Weis (1989) conducted a meta-analysis of correlations reported in studies examining sex, race, and violent crime. They found that variation in the magnitudes of these relationships across studies was related to differences in the level of the data used (e.g., individuals, households, cities, etc.) and whether the analysis was based on offenders as opposed to victims. In the case of sex, male-female differences were greater in violent offending than in violent victimization. For race effects, analyses of offender data produced greater race-violence coefficients than did analyses of victimization. The presence of statistical controls diminished the effect of race on violence but had no effect on the overall magnitude of the sex-violence relationship. Consequently, their results suggest that the relationship between sex and violent offending is strong and stable, whereas the association between race and violent crime is particularly sensitive to the level of analysis and the presence or absence of statistical controls.

Suggested Citation:"Violent Victimization and Offending: Individual-Situational-, and Community-Level Risk Factors." National Research Council. 1994. Understanding and Preventing Violence, Volume 3: Social Influences. Washington, DC: The National Academies Press. doi: 10.17226/4421.
×

SOCIOECONOMIC AND FAMILY FACTORS

Social Class

The association between social class and violent offending continues to be one of the most contested relationships in the literature (e.g., see Tittle et al., 1978; Hindelang et al., 1979; Elliott and Ageton, 1980; Thornberry and Farnworth, 1982; Elliott and Huizinga, 1983; Ageton, 1983; Brownfield, 1986; and Bridges and Weis, 1989). Hindelang et al. (1979) and others have noted the methodological problems in this area: ''Unlike sex and race, social class data are not available for either UCR arrestees or for offenders in victimization surveys. Thus, if understanding the relationship of social class to serious offending is our aim, data limitations are much more severe for social class than for sex or race" (1979:1002). The reliance, then, on self-report data has led to questions concerning the validity of such data for examining violent crime. Still, Hindelang et al. (1979) have argued that the self-report method is capable of dealing with serious offending behavior and, further, that no evidence of differential self-report validity has been found by social class.

Bridges and Weis (1989), in their meta-analysis, also examined variation in the social class-violent crime association. They report, as did Hindelang et al. (1979), that much of the discrepancy in findings across studies of violent offending is due to the level of analysis and to the presence or absence of statistical controls. As was true for race, they found that the magnitude of association between social class and violent offending decreased when individual-level data were used and when statistical controls for correlates of violence were included. The type of data used (i.e., official versus self-report) also had a marginal effect on the relationship; findings that used official data show a stronger negative effect than findings based on other sources.

With the exception of Thornberry and Farnworth (1982) and the vast research in the area of family violence (e.g., see, Gelles, 1985; Straus and Gelles, 1986; Gelles and Straus, 1988; Ohlin and Tonry, 1989), most of the recent literature directly examining social class and violence has been based on samples of adolescents. For example, Elliott and Ageton (1980) used self-report data from a nationally representative sample of 1,716 adolescents in the National Youth Survey and found that social class (based on a comparison of middle-, working-, and lower-class youth) was negatively related to the mean frequency of predatory crimes against persons (a scale including robbery and aggravated, simple, or sexual

Suggested Citation:"Violent Victimization and Offending: Individual-Situational-, and Community-Level Risk Factors." National Research Council. 1994. Understanding and Preventing Violence, Volume 3: Social Influences. Washington, DC: The National Academies Press. doi: 10.17226/4421.
×

assault). This relationship was due in part to the differences across social classes in high-frequency offending. The authors also found an interaction between race and class, which suggests that lower-class black youth were particularly likely to report higher levels of involvement in these personal violence offenses.

Elliott and Huizinga (1983) report more extensive data from the National Youth Survey on the relationship between social class and offense-specific prevalence and incidence scales of violence for both males and females. In analyzing several waves of this longitudinal data base, their results suggest that class differences are more apparent when incidence measures as opposed to prevalence measures are used. They also find class differences in the frequency of minor and felony assault by males such that lower-class youth are more likely to commit offenses than middle-class youth. Few differences were found between middle-and working-class or working-and lower-class males. Among females there were comparable class differences in levels of felony assault but no differences in minor assault levels. A more extensive analysis of female offending behavior using the National Youth Survey found class effects for aggravated assault offending in the first two waves of data and for involvement in gang fighting in four out of five waves of data (Ageton, 1983). In these instances, lower-class females, like males, were more likely to report involvement in assaults against persons.

The above research emphasizes the importance of the measurement of violent offending; findings using incidence measures may differ from findings using prevalence measures. Additionally, the measurement of "social class" is a concern (Thornberry and Farnworth, 1982; Brownfield, 1986). Although there appears to be differential involvement in serious violent offending by social class, these findings depend to some extent on the conceptualization and operationalization of the social class measure. Brownfield (1986) finds that the strength of the class-violence relationship among individuals varies depending on whether a neo-Marxist conception of class categories is used, versus a more gradational measure of socioeconomic status that combines occupation and education measures. A small relationship was found in two studies of youth using the gradational measure, but none was found using the alternative measure (Brownfield, 1986).

Additional limitations of the class-violence literature are discussed by Thornberry and Farnworth (1982), who argue that although the overall relationship between class and juvenile violence is weak, the relationship among adults may be stronger. In

Suggested Citation:"Violent Victimization and Offending: Individual-Situational-, and Community-Level Risk Factors." National Research Council. 1994. Understanding and Preventing Violence, Volume 3: Social Influences. Washington, DC: The National Academies Press. doi: 10.17226/4421.
×

fact, they found in a reanalysis of Wolfgang's (1958) Philadelphia cohort data that for white adults the father's occupational status was negatively related to violent behavior, whereas for black adults the respondent's own education level was negatively related to arrests for violence. The causal implications of these findings are not yet clear-it is possible that there is a reciprocal relationship between the social class of an adult and levels of criminality. Among juveniles this is less problematic, but it is an issue that has not yet been adequately resolved.

Family Structure and Process

There are at least three ways in which family factors have been examined in the literature on violent offending. The first pertains to the marital status of offenders. Studies of prison inmates have found that most offenders are not married (78 percent) and that white inmates are disproportionately divorced or separated. Black inmates, on the other hand, are just as likely not to be married as blacks in the general population (U.S. Department of Justice, 1988a). These findings, of course, are quite limited. We do not know whether an offender's marital status is a consequence of his or her offending behavior, and data representing the prison population may not be representative of the overall population of violent offenders. Consequently, the meaning of these findings is unclear.

Second, and more importantly, family structure and intervening "family processes" (e.g., Laub and Sampson, 1988) have been correlated with aggression and other serious criminal behaviors among children and adolescents. The relationship between juvenile delinquency and these family factors has been reviewed extensively elsewhere (see Loeber and Stouthamer-Loeber, 1986), but we find that studies focusing exclusively on family factors and violent behavior are rare. This is, perhaps, to be expected given that we are focusing our discussion on those serious violent behaviors not typically engaged in by youth. Nonetheless, extant research suggests that children living in "broken" homes are slightly more likely to engage in aggressive behaviors than children living in two-parent homes (Farrington and West, 1971; Wilkinson, 1980; McCord, 1982). However, the magnitude of this relationship is weak, and these results are not unanimous across adolescent subgroups or types of aggressive behaviors. Rankin (1983) also reports a small effect of family structure on fighting (children in

Suggested Citation:"Violent Victimization and Offending: Individual-Situational-, and Community-Level Risk Factors." National Research Council. 1994. Understanding and Preventing Violence, Volume 3: Social Influences. Washington, DC: The National Academies Press. doi: 10.17226/4421.
×

single-parent homes are more likely to report fighting), but no differences in other assaultive behaviors.

Research on family functioning or intervening family processes typically examines factors such as parental neglect, punishment styles, punitiveness, marital conflict, parental attitudes toward deviance, and parental criminality for their effects on children's offending behaviors. Parental neglect in the form of either lack of supervision or lack of other involvement has been found to be positively related to aggressive behaviors in children (Bandura and Walters, 1959; McCord et al., 1961; Farrington, 1978; Loeber and Stouthamer-Loeber, 1986). Punishment styles of parents are also related to aggressiveness and serious criminality: yelling, threatening, and hitting children is associated with increased aggressive behaviors among children (Bandura and Walters, 1959; McCord et al., 1961; Farrington, 1978; McCord, 1982; Wells and Rankin, 1988). Factors such as marital discord and conflict have similarly shown positive, but small, associations with aggression and serious criminality (Bandura and Walters, 1959; Farrington and West, 1971). Finally, parental criminality also tends to show a positive correlation with aggression among children (Glueck and Glueck, 1950; Hutchings and Mednick, 1975; Farrington, 1978) as does parental aggressiveness (Bandura and Walters, 1959) and paternal alcoholism (McCord, 1982).

Loeber and Stouthamer-Loeber (1986:121) summarize in tabular form the overall relationships between these family factors and juvenile delinquency. In general, family functioning or process variables such as those discussed above were more consistently associated with delinquency than were family structure variables. Recent multivariate analyses focusing exclusively on violent behavior (assault, robbery) also suggest that the quality of family life may be more important than family structure. For example, Van Voorhis et al. (1988) found that family structure had no independent effect on violent behavior (fighting and assault combined), whereas overall home quality was negatively and significantly related to violence. Although there are many methodological limitations in the family-aggressive behavior literature (see Loeber and Stouthamer-Loeber, 1986, for an extensive discussion), these findings suggest that family functioning may be a more important factor than family structure and that more explicit attention should be paid to the effects of family functioning on juvenile aggression and its relationship to later violent offending as an adult (see also Farrington, 1978).

A third, related, way in which family factors have been studied

Suggested Citation:"Violent Victimization and Offending: Individual-Situational-, and Community-Level Risk Factors." National Research Council. 1994. Understanding and Preventing Violence, Volume 3: Social Influences. Washington, DC: The National Academies Press. doi: 10.17226/4421.
×

with respect to violence has been in the "cycle of violence" or "violence breeds violence" literature (see Kruttschnitt et al., 1986, 1987; Widom, 1989a,b). These recent and extensive reviews suggest that the research on intergenerational transmission of violence is inconclusive and plagued with methodological problems. Although there is agreement concerning the correlation between suffering abuse as a child and perpetrating abuse as adults, the magnitude of the relationship is quite small and many violent offenders report no history of child abuse (Widom, 1989a). It is also the case that most abused children do not go on to commit abusive or violent acts as adults (Kruttschnitt et al., 1987; Widom, 1989a). Consequently, recent analyses have sought to discover the factors that break the ''cycle of violence."

Kruttschnitt et al. (1987) report that children who suffer from emotional neglect (in addition to physical abuse) and who have no emotional support system inside or outside the home are more likely to be abusive as adults. Widom (1989a) found that age at the time of abuse, sex, temperament, and intelligence were protective factors from future involvement in violence. However, Widom argues for caution when interpreting these results: to say that certain factors appear to protect children from perpetuating future violence does not imply that the effects of child abuse are eliminated. The effects may be manifested in other ways, including depression, anxiety, or even suicide, and these links have not been adequately assessed (Widom, 1989a).

ADDITIONAL FINDINGS

A variety of other factors have been studied in conjunction with violent behavior among juveniles including analyses of school achievement and attachment (Andrew, 1979; Liska and Reed, 1985; Friedman and Rosenbaum, 1988), associations with delinquent peers (Friedman and Rosenbaum, 1988), and prior involvement in crime (Blumstein et al., 1988; Farrington et al., 1988). Perhaps surprisingly, analyses of school and peer factors and their relationship to violence have not been as extensive as investigations of relationships discussed earlier (e.g., age, sex, social class). Nevertheless, small and relatively consistent effects have been reported. Friedman and Rosenbaum (1988) found that time spent doing homework was inversely related to juvenile assault and robbery offending, and that time spent with delinquent peers was positively related to juvenile violence (this is also consistent with Elliott et al.'s, 1985, analysis of "index" offending). Liska and Reed (1985) considered

Suggested Citation:"Violent Victimization and Offending: Individual-Situational-, and Community-Level Risk Factors." National Research Council. 1994. Understanding and Preventing Violence, Volume 3: Social Influences. Washington, DC: The National Academies Press. doi: 10.17226/4421.
×

the reciprocal effects that parental attachment, school attachment, and violence might have on each other. They found that for white males, parental attachment decreased violent behavior and that violent delinquency led to future decreases in both school and parental attachment. They argue that for white males at least, "parents, not the school, are the major institutional sources of delinquency control" (1985:558) and that their findings imply that factors associated with school attachment and achievement may be consequences of violence and not causal factors.

Finally, it should be noted that by distinguishing prevalence of offending from incidence (or frequency) of offending, "criminal career" research (see Blumstein et al., 1986) suggests that certain individuals are disproportionately involved in violent offending. However, research by Blumstein et al. (1988) and Farrington et al. (1988) suggests that offenders do not typically "specialize" in violent crimes such as homicide, assault, and robbery, and that "escalation'' in terms of the seriousness of offending over periods of time is not common. A possible exception to the lack of specialization holds for persons who have prior involvement in robbery (Farrington et al., 1988). Overall, "the more impulsive, violent crimes of homicide, rape, and weapons violations were among the least specialized" (Blumstein et al., 1988:342).16 Consequently, although violent offenders often have prior records of extensive involvement in crime, the prediction of violence from these records is very difficult (Piper, 1985).

SUMMARY

Of all the individual-level correlates discussed above, ascribed demographic characteristics appear to have the strongest associations with violent offending. Sex and age are clearly and consistently associated with all forms of violent behavior, whereas race and social class have smaller and less consistent relationships to violent offending. Family structure and functioning are associated with juvenile violence and aggression, but multivariate analyses suggest that family functioning or process variables such as parental neglect and punishment styles (i.e., quality of family life) are more important than family structure characteristics (i.e., one versus two parents). Physical abuse as a child also appears to be related to future violent behavior, but the reasons for this association are not yet clearly understood. Other correlates of violence such as school attachment and achievement are even smaller in

Suggested Citation:"Violent Victimization and Offending: Individual-Situational-, and Community-Level Risk Factors." National Research Council. 1994. Understanding and Preventing Violence, Volume 3: Social Influences. Washington, DC: The National Academies Press. doi: 10.17226/4421.
×

magnitude, and causal analyses (e.g., Liska and Reed, 1985) suggest that school factors may not be as important as family factors or other social and demographic characteristics. Finally, although violent offenders typically have prior criminal records, the prediction of future violence from these records has been extremely difficult.

As noted throughout the above discussion, much of the extant literature in each of the areas relating social factors to violent offending suffers from methodological or causal interpretation limitations. These limitations include sampling (e.g., a focus on males or juveniles only, unrepresentative samples), the domain of the dependent variable (studies of specific violent crimes are surprisingly rare), analytical techniques (multivariate statistical controls are also rare), issues of causal direction (temporal ordering and reciprocal effects are typically ignored), the general failure to distinguish participation in violent offending from frequency of violent offending (Blumstein et al., 1986), and the paucity of studies examining the intervening mechanisms and social processes that might explain the effects of demographic factors such as age, race, and sex.17

An equally important limitation of prior individual-level analyses of both violent offending and victimization is that they generally have not considered the possible effects that situational-or community-level factors might have on an individual's experiences with violence. It is possible that individual-level correlates of victimization and offending are the same because of the structural or situational contexts in which particular demographic subgroups typically find themselves. Situational-level variables or community-level factors may thus be responsible for a small or large portion of the individual-level risk for violent victimization and offending.

Additionally, individual-level research has for the most part treated victims of violence and violent offenders as though they were mutually exclusive groups (Reiss, 1981; Jensen and Brownfield, 1986). The remainder of the paper therefore considers the additional and perhaps crucial role of victim-offender relationships and the structural context-both situational and community-in which violence occurs. Our contention is that a consideration of these other levels of analysis may help to better explicate the sources of both violent victimization and offending.

Suggested Citation:"Violent Victimization and Offending: Individual-Situational-, and Community-Level Risk Factors." National Research Council. 1994. Understanding and Preventing Violence, Volume 3: Social Influences. Washington, DC: The National Academies Press. doi: 10.17226/4421.
×

SITUATIONAL-LEVEL RISK FACTORS AND VICTIM-OFFENDER RELATIONSHIPS

Situational or microlevel (Short, 1990:11) inquiry has traditionally focused on circumstances surrounding the violent event itself. Although the analysis of "events" may at first seem unproblematic, designating "situational-level" factors is neither easy nor obvious. As Block (1981:743-744) has observed:

The criminal event can be thought of as one instance surrounded by a microenvironment of social relationships, physical structures, and weapons of potential use, a macroenvironment of neighborhood and community, a history of social relationships, and ideas of violence and danger, self-defense, social class, and segregation. The two actors, victim and offender, interact with and are affected by these structures, but they retain individuality; their behavior can never be fully predicted.

The complex nature of situations, combined with a lack of definitional consensus over situational-level factors, leads us to broadly define situational-level risk factors as those factors, outside the individual (Monahan and Klassen, 1982:293), that influence the initiation, unfolding, or outcome of a violent event.18 Based on this conceptualization and past treatments of violent events, we therefore focus on such factors as the presence and type of weapon, the presence of drugs or alcohol, the role of bystanders, and victim resistance and retaliation. We also include a discussion of the victim-offender overlap as well as a review of how victim-offender relationships are related to violence. We believe that the simultaneous consideration of victim-offender dynamics, including past histories and present behaviors and relationships, provides a more complete framework for studying the contextual aspects of violent events.

VICTIM-OFFENDER OVERLAP

Our review in the previous two sections clearly reveals that prior research on the risk of violent victimization and offending has concentrated on demographic and socioeconomic characteristics of the individual. The common message running throughout this prior work is that victims and offenders share a similar demographic profile: both violent offenders and victims of violent crime tend to be young, male, and black and to live in urban areas (see also Hindelang, 1976, 1981; Hindelang et al., 1978; Gottfredson, 1986). Some research even suggests that victims and offenders are one and the same (Wolfgang, 1958; Singer, 1981). By considering

Suggested Citation:"Violent Victimization and Offending: Individual-Situational-, and Community-Level Risk Factors." National Research Council. 1994. Understanding and Preventing Violence, Volume 3: Social Influences. Washington, DC: The National Academies Press. doi: 10.17226/4421.
×

in detail the concept of victim-offender overlap, we hope to answer the question of how, and in what ways, violent offending increases an offender's risk of victimization.

The traditional explanation of the link between victims and offenders is the subculture of violence theory. Formulated originally by Wolfgang and Ferracuti (1967), the subcultural theory proposes that in certain areas and for certain subgroups of the population, there is a subcultural value system that supports the use of violence and other behaviors (e.g., sexual machismo) not emphasized in the dominant culture. In an application of subcultural theory to the victim-offender dyad, Singer (1981) argues that the existence of a subcultural normative system increases the likelihood that an individual will alternate between the role of offender and victim. In particular, victims of crime may become offenders because of norms that justify retaliation (Singer, 1981; Fagan et al., 1987). Conversely, offenders may become victims because they hold values that support the initiation of violence to resolve disputes. For example, subculturally induced aggressive behaviors (e.g., starting fights) are thought to result in an increased risk of victimization (i.e., "victim precipitation"). Subculture of violence theories therefore suggest that offending behavior may affect the risk of victimization and that victimization experiences may affect the probability of offending.

Although rarely studied, the logic of the lifestyle-routine activity theory (Hindelang et al., 1978; Cohen and Felson, 1979; Cohen et al., 1981b; Garofalo, 1987) also suggests that criminal and deviant lifestyles may directly increase victimization because of the nature of the offending behavior itself. As Jensen and Brownfield (1986:87) have argued in a recent paper, offense activity can be considered a characteristic of lifestyles or a type of routine activity that increases the risk of victimization because of the vulnerability or culpability of people involved in those activities. Membership in a gang, for example, involves close proximity to fellow offenders and criminal events (e.g., gang fights) in which rival offender groups are present. It is also well established that most delinquency is committed in groups (Zimring, 1981); assaults, in particular, often take place in contexts where retaliation probability is high (e.g., bars, social clubs, parties). Moreover, assaultive behavior by definition requires more than one person, meaning that the victimization potential of fighting is at a maximum.

Sparks (1982) has further suggested that offenders make ideal targets because they may be victimized with relative impunity.

Suggested Citation:"Violent Victimization and Offending: Individual-Situational-, and Community-Level Risk Factors." National Research Council. 1994. Understanding and Preventing Violence, Volume 3: Social Influences. Washington, DC: The National Academies Press. doi: 10.17226/4421.
×

That is, offenders are likely to be viewed by other potential offenders as particularly vulnerable because "offender" victims are probably less willing to call the police than "nonoffender" victims. In addition to not reporting an offense for fear of implicating themselves in criminal behavior, offender victims may also feel that the police are likely to disbelieve or discount their victimization experience. Persons whose lifestyles include violence thus face a heightened risk of victimization as a direct result of their predatory offending and association with other offenders.

Even deviant lifestyles (e.g., extensive drinking, drug use, "partying") that do not involve actual violence may place persons at risk of victimization. As Jensen and Brownfield (1986) note, many activities that involve the "recreational and social pursuit of fun" increase victimization risk. For example, according to lifestyle theory, persons who drink extensively or go "cruising" for social activity, especially at night, are at higher risk for assault because such behavior often occurs at bars, parties, and other places where victimization risk is heightened. This idea is consistent with Sherman et al.'s (1989) finding that relatively few "hot spots'' produce a large proportion of predatory crime. Indeed, Sherman et al. (1989:39) found that all robberies and rapes in Minneapolis during 1986 occurred at less than 3 percent of "places," particularly bars, parks, adult bookstores, and convenience stores.

Drinking to excess also lowers guardianship potential and thus may increase victimization risk not just for assault, but for other personal crimes such as robbery. The notion of victimization with impunity (Sparks, 1982) applies to general deviance as well. Nonviolent offense activity such as drug dealing and group-oriented drug consumption has especially high victimization potential. For instance, dealers and users are ripe for exploitation by predatory offenders who, with good reason, have little fear of victim-initiated police intervention.

Despite its potential importance to theoretical explanations of violence, there is surprisingly little empirical research that bears directly on the offending-victimization link. As Jensen and Brownfield (1986:98) concluded, "A major individual level variable, offense activity, has been ignored in recent elaborations of a formal opportunity or routine activity theory of personal victimization." An important reason for this is that most criminological studies examine either offending or victimization-few data sets are designed to examine victim-offender interrelationships. Those few studies that have explored the "victimogenic" potential of offending are quite revealing. For instance, Gottfredson (1984) reports a

Suggested Citation:"Violent Victimization and Offending: Individual-Situational-, and Community-Level Risk Factors." National Research Council. 1994. Understanding and Preventing Violence, Volume 3: Social Influences. Washington, DC: The National Academies Press. doi: 10.17226/4421.
×

positive association between suffering a personal victimization (assault, robbery, theft from person) and self-reported offending in the British Crime Survey. This relationship holds for those who have engaged in assault or other personal violence and for those who reported engaging in other minor delinquencies. Hough and Mayhew (1983) found a similar offender-victim link using the same data. Studies in the United States have also established an empirical association between being the victim of an assault and reports of assaulting others (e.g., Singer, 1981; Sparks et al., 1977).

Taken together, these studies suggest the important role of violent and deviant lifestyles in fully understanding personal victimization risk. However, most research has been bivariate in nature and has not considered the potential confounding effects of demographic factors (e.g., age, sex, marital status) and conventional lifestyle variables (e.g., nights out of the home). It is thus unclear whether offending has an impact on the risk of violent victimization that is not accounted for by sociodemographic factors. Also, if the principle of homogamy is correct, then demographic predictors should continue to have direct effects on victimization that are independent of a victim's propensity for criminal offending.

In an intriguing exploration of these issues, Jensen and Brownfield (1986) report positive associations between routine activities such as "cruising," going to parties and bars, and violent victimization among a sample of high school youth. They found a similar positive relationship between ever having been in trouble with the police and victimization. Interestingly, they report that none of their conventional lifestyle measures were independently related to victimization when "delinquent lifestyle" indicators were controlled. Because their demographic indicators show little direct effect on victimization, and because they find a moderate to strong relationship between offending and victimization, Jensen and Brownfield argue that the factors that explain the links between background characteristics and victimization may be the same factors that explain the links between these characteristics and offending.

Building on this idea, Sampson and Lauritsen (1990) examined the independent effect of violent offending behavior and deviant lifestyles (e.g., drinking, drug use) on the risk of personal victimization. Data used for these analyses were derived from the 1982 and 1984 British Crime Surveys. Multivariate analyses showed that offense activity-whether violent or minor deviance-was directly related to the risk of assault. The inclusion of lifestyle

Suggested Citation:"Violent Victimization and Offending: Individual-Situational-, and Community-Level Risk Factors." National Research Council. 1994. Understanding and Preventing Violence, Volume 3: Social Influences. Washington, DC: The National Academies Press. doi: 10.17226/4421.
×

measures (including nights out, extent of drinking, and self-reported violence) did attenuate the magnitudes of the effects of sex and marital status, but the independent effect of age remained consistently strong. Furthermore, the effects of self-reported involvement in deviance or violence maintained for both the 1982 and the 1984 models of assault risk, even when neighborhood levels of violence were controlled for.19

Similar effects of offending lifestyles on the risk of violent victimization have been shown in longitudinal, multivariate analyses using the National Youth Survey (Lauritsen et al., 1991). Adolescents and young adults who reported engaging in offense activities (including assault, theft, and vandalism) were found to be approximately two to three times more likely to report assault and robbery victimization than youth who reported no offense involvement. Moreover, increased involvement in offending "lifestyles"-either through increases in time spent with delinquent peers or through one's own offending-had large positive effects on robbery and assault victimization, controlling for race, age, sex, family structure, family income, neighborhood characteristics, and prior propensity toward victimization. Evidence was also found for reciprocal effects, whereby increases in victimization predicted changes in offense involvement. However, an offending lifestyle continued to have a direct effect on victimization even when reciprocal effects were controlled. Therefore, national data sets from two different countries strongly suggest that violent offending and violent victimization are intimately connected and that this relationship cannot be explained away by traditional demographic and social correlates of crime.

VICTIM-OFFENDER RELATIONSHIPS

As emphasized in the seminal research by Wolfgang (1958) suggesting that homicide victimization was more likely to result from disputes among family members and acquaintances than strangers, consideration of the nature of the relationship between victim and offender in violent incidents has been extensive in criminological literature. As Loftin et al. (1987) argue, the victim-offender relationship is an important variable in studies of personal violence because it places the event within the context of social structures (1987:259), and like the victim-offender overlap, an analysis of victim-offender relationships helps illustrate the situational context of a particular event. We thus briefly overview victim-offender relationships and their associations with violent crime (because

Suggested Citation:"Violent Victimization and Offending: Individual-Situational-, and Community-Level Risk Factors." National Research Council. 1994. Understanding and Preventing Violence, Volume 3: Social Influences. Washington, DC: The National Academies Press. doi: 10.17226/4421.
×

of space limitations we urge readers to consult original sources and other reviews).

Loftin et al. (1987) demonstrate that the measurement of victim-offender relationships is neither easy nor consistent across studies. They found that conceptualization of the victim-offender relationship varied considerably across studies due to the lack of standardization or to definitions that were multidimensional, overlapping, or vague. Even though official data are considered the best available on homicide incidents (Hindelang, 1974; O'Brien, 1985), the "unknown" relationship category in the Federal Bureau of Investigation homicide data is large enough to suggest that stranger homicides may be underestimated (Riedel, 1987). On the other hand, it is also suspected that assault incidents among friends and family are underreported in victimization studies. Furthermore, series victimizations-which are more likely to occur among family members and friends-are excluded from most NCS tabulations and analyses. Caution is clearly required when interpreting the associations between victim-offender relationships and violence.

The NCS data suggest that robbery is the violent crime most likely to involve strangers-victims are approximately four times more likely to be robbed by strangers than by nonstrangers.20 However, for rape and assault, differences in incidence rates between stranger and nonstranger crimes are much smaller. Victim reports suggest that one's chances of being raped or assaulted by a nonstranger are approximately equal to one's chances of being victimized by a stranger (U.S. Department of Justice, 1989). Because it is suspected that victims generally underreport victimizations by nonstrangers, it seems reasonable to suggest that rape and assault are, more often than not, crimes occurring among nonstrangers.

The majority of homicides also appear to be committed by nonstrangers-either relatives, friends, or acquaintances of the victim. For example, Williams and Flewelling (1988) found that approximately two-thirds of homicides occurring in cities with populations greater than 100,000 were among nonstrangers. Because stranger homicide is even more rare in nonurban areas, the national percentage of homicides committed by nonstrangers is probably somewhat higher. As might be expected, nonstranger homicides are likely to be the result of interpersonal disputes, such as arguments over property and money, lovers' triangles, and brawls involving alcohol or drugs. Stranger homicides, on the other hand, are more likely to be the result of circumstances involving the commission of other felonies such as robbery (Williams and Flewelling, 1988).

Suggested Citation:"Violent Victimization and Offending: Individual-Situational-, and Community-Level Risk Factors." National Research Council. 1994. Understanding and Preventing Violence, Volume 3: Social Influences. Washington, DC: The National Academies Press. doi: 10.17226/4421.
×

Other studies have examined the extent to which violent crimes are interracial or intraracial in composition (Sampson, 1984; Block, 1985; Wilbanks, 1985; Zahn and Sagi, 1987; Silverman and Kennedy, 1987; Messner and South, 1986; Cook, 1987; O'Brien, 1987; South and Felson, 1990). These analyses agree that the majority of violent crimes are disproportionately intraracial: for example, whites tend to assault other whites and blacks tend to assault other blacks more so than expected based on chance encounters (Sampson, 1984; O'Brien, 1987). Racial crossover is especially rare in nonfelony homicide (Cook, 1987). Given that most nonstranger homicides are also assault homicides, and that the routine activities and residences of blacks and whites are in large part segregated, these findings are not surprising. For example, Block (1985) argues that for assault homicides, the relationship between victim and offender is likely to be one of residential proximity. Felony homicides (e.g., robbery-murders) are more likely than nonfelony homicides to be interracial because they typically involve strangers (Block, 1985; Cook, 1987; Zahn and Sagi, 1987). In felony homicides as in robberies, black offenders are more likely to victimize whites than white offenders are to victimize blacks (Wilbanks, 1985). Again, however, this is exactly what we would expect given racial distributions in the United States-variations in the relative sizes of the black and white populations explain the patterning of interracial violence (Sampson, 1984; O'Brien, 1987; Messner and South, 1986; South and Felson, 1990).

The intersexual composition of violent encounters has also been examined (Wolfgang, 1958; Pokorny, 1965; Silverman and Kennedy, 1987; O'Brien, 1988). In the case of homicide, both male and female offenders are likely to choose males as their victims (Wolfgang, 1958; Pokorny, 1965; O'Brien, 1988). As social distance increases, violent crimes in general are more likely to involve both male offenders and male victims. Assault is the violent crime most likely to be intrasexual in nature (O'Brien, 1988), whereas rape is almost exclusively intersexual. It may seem obvious that violent crime is more likely to be intersexual than interracial in nature; racial segregation in the United States is far more extensive than segregation by sex (O'Brien, 1988).

One final point concerning intersexual differences in violent events should be made. When women are involved in intersexual violent crimes as victims, the consequences are likely to be much more serious than when women are victimized by women. Men, on the other hand, tend to suffer less serious consequences when they are victimized in intersexual violent crimes. For more extensive

Suggested Citation:"Violent Victimization and Offending: Individual-Situational-, and Community-Level Risk Factors." National Research Council. 1994. Understanding and Preventing Violence, Volume 3: Social Influences. Washington, DC: The National Academies Press. doi: 10.17226/4421.
×

discussions of the etiology and consequences of intersexual violence, readers are referred to the literature reviewing family violence (e.g., Gelles, 1985; Gelles and Straus, 1988; Ohlin and Tonry, 1989).

OTHER SITUATIONAL FACTORS

In this subsection we briefly focus on other factors traditionally considered in situational-level analyses, including the presence of alcohol and drugs, the presence and type of weapon, the role of third parties or bystanders, and victim resistance. The fact that certain factors may be considered either "individual" or "situational" (e.g., alcohol/drug use, violent offending) suggests that further development of situational concepts is needed to clarify these distinctions.

Alcohol and Drug Use

As discussed earlier, the use of alcohol and drugs is related to the risk of violent victimization. There is also evidence that excessive drinking and illegal drug use may place persons at increased risk for violent offending and consequently may be viewed as having an effect on the initiation or outcome of an event. The literature on offending is perhaps more problematic than victimization research because both excessive drinking and propensity to violence may be different aspects of an underlying construct of antisocial behavior. In this case, drinking is not a causal factor in offending, although it may be a facilitating one. Not surprisingly, then, the literature on violence and alcohol use is complex and ambiguous (see especially Collins, 1989) and much beyond the scope of the present review. Nevertheless, it does seem fair to conclude, as does Collins (1989), that very few causal connections have been established. As Collins (1989:63) argues, "Although the weight of the empirical evidence for a relationship between drinking and interpersonal violence is impressive, it probably overstates the importance of drinking as a causal factor in the occurrence of interpersonal violence."

The drug use literature is also complex and beyond the scope of our review (see Tonry and Wilson, 1990). However, it is clear that drug use is related to violent offending (Goldstein, 1989). At least three substantive reasons have been offered to explain the drug-violence connection. The first is a psychopharmacological link in that ingestion of some drugs (e.g., pentachlorophenol, PCP)

Suggested Citation:"Violent Victimization and Offending: Individual-Situational-, and Community-Level Risk Factors." National Research Council. 1994. Understanding and Preventing Violence, Volume 3: Social Influences. Washington, DC: The National Academies Press. doi: 10.17226/4421.
×

may have biochemical effects that lead persons to become aggressive and violent. Second, the "economic-compulsive" model suggests that drug users engage in violence (e.g., robbery) to support their drug habits. The third factor relates to the intrinsically violent world of dealing in illicit markets. In other words, violence is systemic to the distribution networks of drug dealing (Goldstein, 1989). Whatever the causal linkage, it is clear that both drug use and excessive alcohol use are correlated with violence.21

Presence and Type of Weapon

The role that weapon use plays in either the initiation or the outcome of violent events has been discussed most extensively in the literature on robbery and gun availability (Cook, 1979, 1982, 1987; Block and Block, 1980; Pierce and Bowers, 1981). In robbery incidents, the presence of a gun has been found to be associated with a lower risk of injury, most likely a result of the fact that victims are less likely to resist an offender armed with a gun (Cook, 1987). However, when injury does occur in armed robberies, the extent of injury to the victim is likely to be far more serious (Cook, 1987). Similarly, the proportion of victims reporting some injury in robbery incidents increases as the deadliness of the weapon declines (Cook, 1987).

However, it is unclear whether the presence of a weapon during a robbery incident is more important than the offender's intention to harm. Current research is divided on this issue, primarily because the question of whether or not homicide is the actualization of an individual's intent to kill, or the result of a weapon being present under the right circumstances, is difficult to test (see Fisher, 1976; Cook, 1983). Data necessary for fully testing these assumptions are simply not available. Consequently, we are unable to say if the presence of a particular type of weapon has an independent influence on the outcome of a violent encounter exclusive of an offender's intention to harm.

Victim Resistance

In a related manner, situational analyses have studied the role that victim resistance plays in the outcome of violent events. As is true of situational-level analyses in general, there exist numerous unresolved problems, such as the type of victim behaviors that constitute "resistance." One of the broadest definitions of

Suggested Citation:"Violent Victimization and Offending: Individual-Situational-, and Community-Level Risk Factors." National Research Council. 1994. Understanding and Preventing Violence, Volume 3: Social Influences. Washington, DC: The National Academies Press. doi: 10.17226/4421.
×

victim resistance is discussed by Skogan (1981). He argues that perhaps all activities persons undertake to reduce their chances of victimization, such as risk avoidance (e.g., reducing contact with offenders by moving out of high-crime neighborhoods or staying at home) and risk management (e.g., reducing one's chances of victimization given one's spatial contact with offenders), can be considered victim resistance. In other words, at one level of abstraction nearly all activities that persons initiate to decrease the likelihood of victimization (even such as installing alarms and purchasing or carrying firearms) are instances of resistance.

Nevertheless, it has typically been the case that analyses of victim resistance within given violent events focus on whether the victim engaged in "forceful resistance" behaviors (such as fighting) or "nonforceful resistance" behaviors (such as talking to offenders or trying to run away). Block and Skogan (1986) studied these aspects of victims' behaviors during incidents of stranger violence using data from the NCS. Forceful resistance was found to decrease the success of robberies but was related to higher risks of attack and injury (without reductions in success) in cases of rape. Nonforceful resistance also reduced the success of robberies but was unrelated to attack or injury risks in rape incidents. Thus, variations in these findings by crime type make generalizations about the consequences of victim resistance during violent incidents highly problematic. Block and Skogan (1986) discuss several of the limitations of their research, especially the fact that with the data it was impossible to determine whether the victim was even able to resist, given the circumstances of the event, and that the sequencing of events (resistance and outcome) was indeterminable. Ideally, two types of data are needed for adequate assessment of the role of victim resistance: (1) panel data on individuals that temporally distinguish risk management and avoidance techniques from victimization experiences (Skogan, 1981), and (2) detailed event descriptions that also allow temporal distinctions to be made between resistance behaviors and various indicators of outcome (Block and Skogan, 1986).

Felson and Steadman (1983) described the systematic patterns of violent encounters by analyzing 159 reports of homicide and assault incidents among incarcerated offenders. Although traditional limitations of self-report data are of concern here, the authors also point out that the degree of "resistance" or aggression by victims is probably underestimated because cases in which victims were thought to have initially provoked the incident by their own attack would not be included in the sample (i.e., self-defense

Suggested Citation:"Violent Victimization and Offending: Individual-Situational-, and Community-Level Risk Factors." National Research Council. 1994. Understanding and Preventing Violence, Volume 3: Social Influences. Washington, DC: The National Academies Press. doi: 10.17226/4421.
×

assaults). Nonetheless, in these assault cases, victim resistance in the form of retaliation was associated with increased aggression on the part of the offender and increased likelihood of death for the victim. Similarly, a victim's use of weapons was also associated with a higher risk of death for the victim. Felson and Steadman point out that in assault cases, the role of victims' behaviors is more complex than the concept of "victim precipitation" (Wolfgang, 1958) suggests.

Presence of Third Parties

The role of third parties in the initiation or outcome of violent events has received only limited attention in the literature.22 In Felson and Steadman's (1983) analysis, third parties' behaviors were characterized as antagonistic in the majority of violent events. They found that when third parties did attempt to mediate disputes underlying assaults, their efforts typically had no effect on the likelihood that the event would result in a homicide. The lack of research attention to the effects of third parties (including bystanders or witnesses) raises questions concerning the role of "capable guardians" in routine activity explanations of violence. Clearly, this concept means more than the simple presence of a third person-these persons must be willing and able to be guardians. The role of capable guardians becomes even more difficult to ascertain, given that the willingness to become a capable guardian is probably not distributed randomly across situational contexts. For example, assault incidents occurring in bars may be aggravated by the presence of persons who are intoxicated, or they may be aggravated for reasons having to do with the fact that people are not randomly drawn into bars, especially bars located in well-known "hot spots" of predatory crime (cf. Sherman et al., 1989). Consequently, our knowledge about the role of third parties in violent encounters is very limited; in some instances their presence might diffuse the intensity of encounters (e.g., by trying to break up a fight or by calling the police), whereas in other cases they may increase the severity of the outcome (by encouraging escalating actions, providing weapons, and becoming physically involved). Undoubtedly, more research is required to provide even basic descriptions of the role of third parties across situational contexts before their relative effects on violent encounters can be assessed.

Suggested Citation:"Violent Victimization and Offending: Individual-Situational-, and Community-Level Risk Factors." National Research Council. 1994. Understanding and Preventing Violence, Volume 3: Social Influences. Washington, DC: The National Academies Press. doi: 10.17226/4421.
×

SITUATIONAL AND ECOLOGICAL CONTEXT

A final key question concerns the role of ecological proximity to crime and social disorder in predicting victimization risk. A recent study by Fagan et al. (1987:588) suggests that the correlation between victimization and offending may be spurious-a function of the areas in which victims and offenders live. Their basic argument is that the link between victimization and offending may be confounded "by the convergence of correlates of criminal events and offenders in urban areas." For example, living in low-income urban areas where the offender pool is large may increase victimization risk, regardless of one's individual propensity to engage in crime. Although they did not examine the independent effect of offending on victimization, Fagan et al. (1987) found that violent victimization was significantly related to drinking problems, peer delinquency, and neighborhood family violence. This study is important in that it suggests the need to simultaneously account for the social ecology of criminal behavior. In particular, it suggests that controlling for ecological proximity to violence may reduce the individual-level relationship between offending and victimization.

The related "principle of homogamy" in lifestyle theory states that persons are more likely to be victimized when they disproportionately associate with or come in contact with offenders. Although this principle has typically been conceived in individual-level terms, it has clear relevance at the community level. As Garofalo (1987) posits in a recent revision of lifestyle theory, the structural constraint of residential proximity to crime has a direct effect on victimization that is unmediated by individual lifestyle. Specifically, Garofalo (1987) argues that constraints of the economic system and especially housing markets dictate in large part where people live. Although individual lifestyles may thus predict variations in victimization within a given environment, the base level of risk that they face is "heightened by sheer proximity to-and hence exposure to-potential offenders" (Garofalo, 1987:38).

It follows that multilevel models including both demographic and lifestyle measures along with ecological and/or situational measures of the extent of crime can provide further insights into the risk factors for violent victimization. Research by Sampson and Lauritsen (1990) utilizing British Crime Survey data found support for this notion. When an ecological measure of neighborhood violence was included in their models of individual-level personal victimization, area violence rate was positively associated with risk. This effect was significant and independent of

Suggested Citation:"Violent Victimization and Offending: Individual-Situational-, and Community-Level Risk Factors." National Research Council. 1994. Understanding and Preventing Violence, Volume 3: Social Influences. Washington, DC: The National Academies Press. doi: 10.17226/4421.
×

traditional demographic factors, lifestyle and routine activities measures, and self-reported involvement in deviance and offending. It should be noted, though, that the latter effects were not diminished by controls for proximity to violence.

In addition, Sampson (1987b) has shown that the risk of violent victimization by strangers is significantly and positively related to neighborhood levels of family disruption, percent of single-person households, and residential mobility, when age, marital status, sex, and neighborhood socioeconomic status and unemployment rates are controlled. Thus, regardless of key individual-level factors emphasized in past research, the data suggest that community (and presumably situational) contexts of anonymity and its correlates (e.g., prevalence of single-person households, mobility) directly increase the risk of stranger violence (see also Kennedy and Forde, 1990).

SUMMARY

The data necessary for assessing independent situational-level causes of violent encounters not only are unavailable but would be difficult to conceptualize given the opportunity. Most situational-level data contain only incident-based cases of violence and hence constitute conditional risks; that is, existing samples do not include encounters that were successful in the avoidance of violence. Consequently, situations with similar characteristics but different outcomes (i.e., violence and nonviolence) are difficult to compare. In other words, unconditional situational data including ''nonincidents" are not collected often, nor is there a substantive literature that might guide such an endeavor. Thus, we are prevented from examining how variations across situations are causally related to the probability that violent events will occur.

Moreover, our review of situational-level factors demonstrates that victimization and offending are tightly connected and need to be examined in a more unified perspective that takes account of the structural properties of social environments (see also Reiss, 1986a). Accordingly, we now turn to an explicit focus on the community structure of both violent offending and victimization. We discuss later the features of research designs that might be able to address some of the problems discussed throughout this section.

Suggested Citation:"Violent Victimization and Offending: Individual-Situational-, and Community-Level Risk Factors." National Research Council. 1994. Understanding and Preventing Violence, Volume 3: Social Influences. Washington, DC: The National Academies Press. doi: 10.17226/4421.
×

COMMUNITY-LEVEL SOURCES OF VIOLENCE

The purpose of this section is to review empirical research on the community-level correlates of criminal violence. A subsequent section reviews research conducted at the metropolitan (e.g., city, county, standard metropolitan statistical area [SMSA]) level. As noted in the introduction, the goal of macrolevel research is not to explain individual involvement in criminal behavior but to isolate characteristics of communities, cities, or societies that lead to high rates of criminality (Short, 1985; Byrne and Sampson, 1986). Within this framework we generally place more weight on, and therefore begin with, studies that examine neighborhoods or other intraurban units of analysis. Cities and metropolitan areas are large, highly aggregated, and heterogeneous units with politically defined and hence artificial ecological boundaries. Although intraurban units of analysis (e.g., census tracts, wards, block groups) are imperfect proxies for the concept of local community, they generally possess more ecological integrity (e.g., natural boundaries, social homogeneity) than cities or SMSAs and are more closely linked to the causal processes assumed to underlie the etiology of crime (see also Taylor et al., 1984; Bursik, 1988).

Contemporary research on communities and crime owes an intellectual debt to the work of nineteenth century European ecologists and social statisticians. Yet as Levin and Lindesmith (1937:801) have correctly observed, social scientists often "attribute greater originality to contemporary studies and less value to the old than is actually warranted by the facts in the case." Although space limitations preclude detailed review, the facts suggest that consideration of community-level sources of violence should begin with recognition of the original contributions of nineteenth century researchers such as Guerry, Quetelet, Rawson, and Mayhew (for a review see Morris, 1958; Levin and Lindesmith, 1937). Their ecological, or community-level, perspective on crime united three major concerns that contrast with the individual-level emphasis prevalent in modern criminology (Morris, 1958:42):

  1. a primary interest in crime as a social or collective phenomenon, of which individual behavior is a component, rather than in motivation of crime in the individual,

  2. the quantification and statistical analysis of data relating to crime and criminals to illustrate variations in time and place, and

  3. a stress on the role of objective social-structural factors

Suggested Citation:"Violent Victimization and Offending: Individual-Situational-, and Community-Level Risk Factors." National Research Council. 1994. Understanding and Preventing Violence, Volume 3: Social Influences. Washington, DC: The National Academies Press. doi: 10.17226/4421.
×

such as poverty and density of population in determining and perpetuating crime.

THE SHAW AND MCKAY MODEL

A community-level orientation to criminological research later came to fruition in the United States with the twentieth century work of Shaw and McKay (1942, 1969). Indeed, although their focus was primarily on delinquency, the research of Shaw and McKay forms the infrastructure for modern American studies of the ecology of violence. As Bursik (1988) and others (e.g., see Short, 1969; Morris, 1970) have argued, few works in criminology have had more influence than Juvenile Delinquency and Urban Areas (Shaw and McKay, 1942). In this classic work, Shaw and McKay argue that three structural factors-low economic status, ethnic heterogeneity, and residential mobility-led to the disruption of local community social organization, which in turn accounted for variations in rates of crime and delinquency (see also Shaw et al., 1929; Kornhauser, 1978; Sampson and Groves, 1989).

The origins of this thesis can be traced to Clifford Shaw's early work in Chicago (see Shaw et al., 1929), where he was one of the first American researchers to demonstrate the marked variations in delinquency rates within a major city. Shaw and his later associate McKay (1931) showed that the highest delinquency offender rates in Chicago (based on referrals to the Cook County Juvenile Court) were located in deteriorated zones in transition next to the central business and industrial district. Rates of delinquency were found to decrease as distance from the center of the city increased; the exceptions were areas also characterized by industry and commerce. These findings led Shaw and McKay to conclude that delinquent behavior was closely related to the growth processes of the city as outlined by Park and Burgess (e.g., Park et al., 1925). It should be noted that the relationship between community factors and delinquency was substantial. For example, Shaw and McKay (1969) reported a correlation of .89 between delinquency rates in Chicago's community areas and a proxy measure of poverty-the percentage of families on relief. The correlation between population change and delinquency was .69, and the correlation of heterogeneity (percentage foreign born and Negro heads of families) was .60 (Shaw and McKay, 1969:144-153).

Shaw and McKay (1942) also demonstrated that high rates of delinquency persisted in certain areas over many years, regardless of population turnover. More than any other, this finding led

Suggested Citation:"Violent Victimization and Offending: Individual-Situational-, and Community-Level Risk Factors." National Research Council. 1994. Understanding and Preventing Violence, Volume 3: Social Influences. Washington, DC: The National Academies Press. doi: 10.17226/4421.
×

them to reject individualistic explanations of delinquency and focus instead on the processes by which delinquent and criminal patterns of behavior were transmitted across generations in areas of social disorganization and weak social controls (1942:320). Shaw and McKay then located the causes of crime in the structural context of local communities and changes taking place in the wider urban ecology. From this viewpoint the "ecological fallacy"-inferring individual-level relations based on aggregate data-is not at issue because the unit of causal explanation and analysis is the community itself.

Most of the ecological research initially spawned by Shaw and McKay was concerned with economic status and its relationship to delinquency rates. For example, Lander (1954), Bordua (1958), and Chilton (1964) each attempted to replicate the basic Shaw and McKay model. These studies are not reviewed here because they focused on delinquency rates dominated by relatively minor offenses (e.g., larceny, vandalism) and hence are of indirect relevance to our focus on violence. Suffice it to note, however, that although discrepancies arose in multivariate analyses, this body of research essentially confirmed Shaw and McKay's emphasis on economic status as a correlate of delinquency. As Gordon (1967:943) summarized the works of Lander, Bordua, and Chilton, "There should no longer be any question about the ecological relations between ... SES [socioeconomic status] and official delinquency rates."

Despite the major focus on nondifferentiated delinquency rates in ecological research, three studies in the mid-twentieth century did underscore the role of community-level socioeconomic status in delineating patterns of violence. Bullock (1955) studied the ecological distribution of homicide occurrence rates among the census tracts of Houston between 1945 and 1949. Like Shaw and McKay, but focusing on more serious crime, Bullock demonstrated that homicides were disproportionately concentrated in areas characterized by physical deterioration of dwelling units, unemployment, low economic status, and low median education. For example, Bullock (1955:569) reported that homicide rates were correlated –.60 and .64 with percent educated and percent unemployed, respectively. Schmid (1960) correlated offense-specific crime variables, including robbery, with numerous social structural variables for census tracts in Seattle. Schmid (1960:531) found that the median income of an area had a significant inverse relationship to both street robbery (-.53) and business robbery (-.32). Bensing and Schroeder (1960) empirically documented that homicides were

Suggested Citation:"Violent Victimization and Offending: Individual-Situational-, and Community-Level Risk Factors." National Research Council. 1994. Understanding and Preventing Violence, Volume 3: Social Influences. Washington, DC: The National Academies Press. doi: 10.17226/4421.
×

disproportionately concentrated in a small number of areas in Cleveland. They described the high-rate areas as those with the "lowest socioeconomic status, most undesirable neighborhood conditions, greatest financial dependency, the most acute problems of space and crowded housing conditions, the least stability of the population, the greatest social maladjustment and family and individual adjustment problems, and the poorest health" (Bensing and Schroeder, 1960:184). Their description of Cleveland homicide areas reads like Shaw and McKay's account of delinquency areas in Chicago some 30 years earlier.

RECENT COMMUNITY-LEVEL STUDIES OF VIOLENCE

After a hiatus in the 1960s, the past two decades have witnessed a sharp increase in research focused on variations in urban crime rates (e.g., see the reviews in Byrne and Sampson, 1986; Bursik, 1988). Like Shaw and McKay, the general thesis of this social-ecological branch of research is that characteristics of ecological units have social effects on crime that are not solely attributable to the characteristics of individuals. However, this recent upsurge in research has departed from the Shaw and McKay model in at least four important respects. First, recent research has been concerned primarily not with delinquency rates but with variations in criminal violence-especially homicide. Second, the vast majority of recent research has been conducted not at the local community level but at much higher levels of aggregation such as the city, county, or SMSA. Third, the list of variables considered has been expanded much beyond poverty, racial composition, and mobility to include such factors as inequality, density, housing structure, family disruption, region, and opportunities for crime.

Fourth, there has been much more concern in recent years with the limitations of official data. Note that all of the studies discussed thus far, including Shaw and McKay, utilized officially based crime rates. The general criticisms of official data are well known and need not be repeated here. Suffice it to say that the major issue with respect to community research concerns the extent to which official crime rates reflect ecological biases in official reaction to illegal behavior (Hagan et al., 1978; Smith, 1986; Sampson, 1986c). In particular, conflict theorists argue that lower-status communities may have higher delinquency rates in part because police concentration is greater there compared to higher-status areas. The type of community in which police-citizen encounters

Suggested Citation:"Violent Victimization and Offending: Individual-Situational-, and Community-Level Risk Factors." National Research Council. 1994. Understanding and Preventing Violence, Volume 3: Social Influences. Washington, DC: The National Academies Press. doi: 10.17226/4421.
×

occur may also influence the actions taken by police (Hagan et al., 1978; Sampson, 1986c). In support of this idea, Smith (1986) demonstrated that the probability of arrest across communities declines substantially with increasing socioeconomic status-independent of crime type and other correlates of arrest decisions. The reliance on official data thus raises the question of whether Shaw and McKay's findings, and the host of census-based studies following them, are in part artifactual.

The official data problem has been addressed in two basic ways. The most common has been to limit the domain of inquiry to serious crimes such as homicide and robbery where police biases are thought to be minimal. A wide-ranging body of research shows that, for serious crimes, police bias and underreporting are very small and/or unrelated to individual-level and community variables of interest (for a detailed review see Gove et al., 1985). Second, in the last 20 years self-report and victimization data have been brought to bear on the validity of official statistics (see Hindelang et al., 1981). Although self-report studies have generally been either national in scope (e.g., Elliott and Ageton, 1980) or specific to one locale (e.g., Hindelang et al., 1981), between-community estimates of crime rates based on self-reports have been utilized in a few studies. More common has been the use of victimization surveys to provide an alternative window from which to view the ecological correlates of violent crimes such as assault and robbery. For example, the National Crime Survey has been utilized to estimate both city-and neighborhood-level correlates of violent victimization.

Parenthetically, it should be noted that the use of both offenses officially reported to the police and victimizations reported to survey interviewers seems to imply a focus on crime occurrence rates rather than offender rates as employed by Shaw and McKay (1942). A crucial point, however, is that for serious personal offenses such as homicide and aggravated assault, offenders tend to commit offenses relatively close to their homes (Bensing and Schroeder, 1960; Pittman and Handy, 1964; Pokorny, 1965; Pyle et al., 1974; Brantingham and Brantingham, 1984). For all intents and purposes, then, victimization occurrence rates and official offense or arrest rates, particularly at the city level, are tapping the same general dimension with respect to violence. In fact, as we shall see, a general convergence of community-level findings between official police statistics and victimization data has been achieved.

Suggested Citation:"Violent Victimization and Offending: Individual-Situational-, and Community-Level Risk Factors." National Research Council. 1994. Understanding and Preventing Violence, Volume 3: Social Influences. Washington, DC: The National Academies Press. doi: 10.17226/4421.
×
Poverty and Deprivation

Not surprisingly, a large proportion of recent neighborhood-based studies of violence have emphasized dimensions of poverty and economic inequality. Unlike Shaw and McKay, however, the majority of these have employed multivariate methods that seek to estimate the effects of economic structure independent of other factors such as population composition. For example, Block (1979) examined officially recorded homicides, aggravated assaults, and robberies that occurred in the mid-1970s in 76 community areas of Chicago. The most important predictor of variations in violent crime across Chicago areas, controlling for other factors, was "proximity," defined by Block (1979:48) as the ratio of families earning more than three times the poverty level to those earning 75 percent or less. He argued that although this can be viewed as an income disparity measure, "it can be more properly thought of as an indicator of spatial proximity of middle class and poor" (Block, 1979:48). He also showed that homicide rates were significantly and substantially correlated not only with proximity (.75), but also with percent in poverty (.64), percent black (.69), and percent female-headed families (.63). Correlations identical in sign and virtually in magnitude were also present for robbery and aggravated assault. The significant predictors of variations in violent crime rates across community areas in Chicago uncovered by Block (1979) were thus quite similar to the delinquency correlates found by Shaw and McKay some 50 years earlier in the same city.

Two studies of violent personal crimes in Houston published in the 1970s mirrored the findings of Bullock in that city in the 1940s. Specifically, both Beasley and Antunes (1974) and Mladenka and Hill found (1976) substantial correlations between violent crime and poverty indicators across 20 police districts. For Beasley and Antunes, the correlations was -.78 for median income, whereas in Mladenka and Hill the correlation between poverty and violence was .93. Both studies also reported large and significant correlations between violent crime and the factors of density (.68 and .70) and percent black (.74 and .81). Again, the general similarity to Shaw and McKay is apparent: violent crime rates were disproportionately concentrated in high-density poverty areas characterized by a large proportion of minorities.

The 1980s produced several studies of homicide that applied multivariate techniques to assess the independent effects of local community structure. Another Chicago-based study examined community-level variations in gang homicide rates as measured by the Gang Crime Unit of the Chicago Police Department (Curry

Suggested Citation:"Violent Victimization and Offending: Individual-Situational-, and Community-Level Risk Factors." National Research Council. 1994. Understanding and Preventing Violence, Volume 3: Social Influences. Washington, DC: The National Academies Press. doi: 10.17226/4421.
×

and Spergel, 1988). The main focus was on "poverty," which was measured via a principal components factor scoring of percent population below the poverty level, unemployment rate, and mortgage investment (1988:393). When percent black, percent Hispanic, and poverty were simultaneously entered into a multiple regression equation, poverty had significant positive effects on gang homicide rates for 1978-1981 and 1982-1985. In particular, their data showed that the standardized coefficient reflecting the direct effect of poverty on 1978-1981 gang homicide rates was .42, with controlling for percent black and percent Hispanic. Poverty also had the largest effect on change in homicide rates from 1978-1981 to 1982-1985. Curry and Spergel concluded that gang homicide rates are directly related to poverty (1988:395).

Messner and Tardiff (1986) analyzed homicide rates across 26 neighborhoods in Manhattan in 1981. They also found a strong positive relationship between poverty and homicide (r = .55, p < .01), but the multivariate results were somewhat ambiguous. Poverty correlated so highly with percent black (.72) that the former had a significant effect on homicide only when percent black was not in the model. In addition, the Gini index of inequality had an insignificant effect on homicide, whereas the most consistent and strongest correlate of homicide in Manhattan was the divorce rate. Messner and Tardiff (1986:312) concluded that it is not relative deprivation generated by inequality that is conducive to violence, but absolute level of poverty in conjunction with pervasive levels of marital dissolution.

Studies utilizing victimization rates have also questioned the strength of the effects of poverty and relative inequalities in income on violence once other factors are controlled. In a series of studies, Sampson (1983, 1985a, 1986a) utilized Bureau of the Census neighborhood characteristics data in conjunction with victimization survey data from the NCS. Each household record in the NCS sample contained information on the social-demographic characteristics of the neighborhood in which the household was sampled. Results derived from analysis of variance showed that, overall, poverty and inequality exhibited weak or insignificant effects on violence (aggravated assault, simple assault, rape, and robbery) compared to other neighborhood characteristics such as density of housing (percent units in structures of five or more units), residential mobility (percent of persons 5 years of age and older living in a different house than they did five years earlier), and family structure (percent female headed families and percent divorced/separated). For example, income inequality had insignificant effects

Suggested Citation:"Violent Victimization and Offending: Individual-Situational-, and Community-Level Risk Factors." National Research Council. 1994. Understanding and Preventing Violence, Volume 3: Social Influences. Washington, DC: The National Academies Press. doi: 10.17226/4421.
×

on violent victimization in urban areas once racial composition and divorce rates were controlled (Sampson, 1985a, 1986a). The exception was the consistent positive effect of neighborhood unemployment rates on violent victimization, especially for residents of central cities (Sampson, 1986a:36).

A similar pattern was found in another victimization study where Smith and Jarjoura (1988) analyzed data collected from interviews with random samples of 11,419 individuals residing in 57 neighborhoods in Rochester (New York), Tampa-St. Petersburg (Florida), and St. Louis. Interviews were conducted with approximately 200 persons in each neighborhood, and Smith and Jarjoura aggregated data on victimizations and social characteristics to produce a data file consisting of 57 neighborhoods. The major dependent variable was rate of violent crime-robberies and assaults per 1,000 residents. The independent variables included percentage low income, residential mobility, racial heterogeneity, population density, percentage aged 12-20, percent single-parent households, percentage nonwhite, and percent living alone. In an evaluation of Shaw and McKay's (1942) basic model, Smith and Jarjoura (1988) discovered a significant interaction between mobility and low income in explaining violence. Specifically, mobility was positively associated with violent crime rates in poorer neighborhoods but not in more affluent areas. The main effects of mobility and income were not significant when the interaction term was in the model. Smith and Jarjoura (1988) concluded that communities characterized by rapid population turnover and high levels of poverty have significantly higher violent crime rates than either mobile areas that are more affluent or poor areas that are stable.

The role of poverty and economic status in understanding violence at the community level is thus not altogether clear. On the one hand, it is significantly associated in almost all studies of violence. On the other hand, differences emerge in the conclusions drawn regarding the independent role that poverty plays in explaining violence. The data in several of the studies suggest that the effect of poverty is either weak or conditional on community contexts of mobility and change. This latter point is crucial, and we thus turn to the second major factor highlighted by Shaw and McKay that may help to shed light on the meaning of poverty in an ecological context.

Suggested Citation:"Violent Victimization and Offending: Individual-Situational-, and Community-Level Risk Factors." National Research Council. 1994. Understanding and Preventing Violence, Volume 3: Social Influences. Washington, DC: The National Academies Press. doi: 10.17226/4421.
×
Mobility and Community Change

One of the fundamental claims made by Shaw and McKay (1942) was that population change and turnover had negative consequences for the social control of delinquency. In particular, a high rate of mobility, especially in an area of decreasing population, was inferred to foster institutional disruption and weakened community controls (Kornhauser, 1978). The existing research on mobility has not been as extensive as that on economic status, but it has been revealing. For example, Block's (1979:50) study of Chicago's community areas revealed consistent negative correlations between percent neighborhood stability and the violent crimes of homicide, robbery, and aggravated assault (-.47, -.50, and -.40, respectively; all p < .05). Neighborhood stability also had significant negative effects on homicide even after proximity was controlled (1979:53).

As introduced above, Smith and Jarjoura (1988) found a significant positive effect of neighborhood mobility (percentage of households occupied by persons who lived there less than three years) on rates of violence only in low-income neighborhoods. In fact, this interaction of mobility and poverty was the largest determinant of violence. Although neither mobility nor income had significant main effects, the ratio of coefficient to standard error for the mobility-income interaction was 3.44; by contrast, the other predictors of violence all had ratios less than 3 (Smith and Jarjoura, 1988:41). In community contexts of poverty at least, mobility appeared to have important consequences for violence. Notably, the mobility-income interaction did not emerge for property crime, because mobility had a direct positive effect on burglary rates (Smith and Jarjoura, 1988:44).

Although no consistent interactions were uncovered, victimization data from the National Crime Survey show that residential mobility has significant positive effects on rates of violent victimization (Sampson, 1985a:29-30). After controlling for other neighborhood-level correlates of victimization, parameter estimates from analysis of variance models revealed that violent victimization rates for residents of high-mobility neighborhoods were double those of residents in low-mobility areas (Sampson 1985a:30; 1986a:44).

There are, however, conceptual and methodological problems that plague mobility research. As Kornhauser (1978:107) has noted, mobility indices such as Shaw and McKay's (1942) were ill designed to measure mobility; instead they were chosen to support the contention that delinquency was linked to the disturbances in social organization engendered by city growth. Perhaps more crucially,

Suggested Citation:"Violent Victimization and Offending: Individual-Situational-, and Community-Level Risk Factors." National Research Council. 1994. Understanding and Preventing Violence, Volume 3: Social Influences. Washington, DC: The National Academies Press. doi: 10.17226/4421.
×

an essential feature of the ecological model is a focus on neighborhood change over time and its consequences for crime (Bursik, 1988). Thus, although mobility rates of an area have been linked to delinquency rates in cross-sectional designs, until recently few studies in the Shaw and McKay tradition have actually investigated processes of neighborhood change in a longitudinal perspective (but see Bursik and Webb, 1982; Heitgerd and Bursik, 1987, with regard to delinquency rates).

Recognizing this gap, Taylor and Covington (1988) have recently published an important study of neighborhood changes in ecology and violence. Their research investigated changes in community structure and the violent crimes of murder and aggravated assault for 277 Baltimore city neighborhoods in the period 1970-1980. The two dimensions of ecological change examined were economic status and family status. Their thesis was that neighborhoods experiencing declines in relative economic status and stability should also experience increases in relative violence (Taylor and Covington, 1988:561). In support of this notion, their findings showed that increasing entrenchment of the urban ''underclass" in the form of increasing poverty and minority concentration was linked to increases in violence. In other words, in the city's lower-income neighborhoods that changed for the worse in terms of an increasing concentration of the ghetto poor, violence rose sharply. Gentrifying neighborhoods that were experiencing rapid change in terms of unexpected increases in owner-occupied housing, one-unit structures, and changes in family status also experienced large increases in violence (see also McDonald, 1986; Covington and Taylor, 1989). Taylor and Covington interpret this to mean that processes of change lead to increasing violence. As they summarized the pattern of change in Baltimore in the 1970s (Taylor and Covington, 1988:583):

[T]he two most major changes experienced by urban neighborhoods were the further solidification of underclass neighborhoods and the appearance of gentrifying neighborhoods. ... In Baltimore, both of these changes were associated with unexpected increases in relative violence levels. The patterning of ecological change parameters with violence changes suggested that in underclass neighborhoods the violence was related to increased relative deprivation, and that in gentrifying neighborhoods the violence was linked to increasing social disorganization.

Thus, changes in both relative economic status and relative stability (in terms of housing, family) were linked to unexpected increases in violence.

Suggested Citation:"Violent Victimization and Offending: Individual-Situational-, and Community-Level Risk Factors." National Research Council. 1994. Understanding and Preventing Violence, Volume 3: Social Influences. Washington, DC: The National Academies Press. doi: 10.17226/4421.
×

In short, extant research suggests that rates of mobility and neighborhood change are related to community violence. However, what seems most salient about change is its linkage to the downward spiral of a neighborhood becoming increasingly poor (see also Schuerman and Kobrin, 1986; Wilson, 1987; Rose and McClain, 1990). The notion of an interaction between underclass entrenchment and residential change is discussed later in more detail, as are problems with causal interpretations of this phenomenon.

Heterogeneity and Racial Composition

Racial and ethnic heterogeneity has always been accorded a central role in Shaw and McKay's research (e.g., see Kornhauser, 1978), even though in their writings Shaw and McKay (1942:153) referred mostly to population composition. This is not surprising because their data showed that delinquency rates were higher in predominantly black/foreign-born areas than in areas of maximum heterogeneity. For example, Shaw and McKay's (1942:155) data indicate that the delinquency rate in areas with more than 70 percent black/foreign born (8.2) was more than double the rate (3.9) in areas of maximum heterogeneity (e.g., 50-59% black/foreign born).

Similarly, most research on violence has examined racial composition-usually percent black-rather than racial heterogeneity per se. The general and consistent finding has been that percentage black is positively and strongly related to rates of violence. For example, Block (1979) showed that homicide rates were significantly and substantially related to percent black (r = .69), as did Beasley and Antunes (1974), Mladenka and Hill (1976), Messner and Tardiff (1986), Sampson (1985a), Roncek (1981), and Smith and Jarjoura (1988). The dispute arises over the strength of the direct effect of racial composition on violence. Some report that the percent black effect remains strong even after controlling for other factors (e.g., Beasley and Antunes, 1974; Roncek, 1981), whereas others find a sharply attenuated effect of race once other factors are controlled (e.g., Block, 1979; Curry and Spergel, 1988; Sampson, 1985a; Messner and Tardiff, 1986).

An exception to the reliance on composition measures is Smith and Jarjoura (1988), who explored a probability-based measure of racial heterogeneity-the probability that two randomly selected individuals from a neighborhood will be members of different racial groups. This measure of heterogeneity was significantly correlated

Suggested Citation:"Violent Victimization and Offending: Individual-Situational-, and Community-Level Risk Factors." National Research Council. 1994. Understanding and Preventing Violence, Volume 3: Social Influences. Washington, DC: The National Academies Press. doi: 10.17226/4421.
×

(r = .43) with violent crime rates (1988:39) and percent nonwhite (r = .59), but despite significant effects on violent crime independent of mobility and poverty in multivariate regressions, racial heterogeneity was rendered insignificant once family structure (percent single-parent families) was controlled.

Thus, although there is little question that percent black and heterogeneity are strong and pervasive correlates of violent crime rates, there is reason to doubt whether racial composition or heterogeneity has unique explanatory power. Additionally, the theoretical status of racial composition at the community level is questionable. As reviewed earlier, race is a salient individual-level predictor of both violent offending and victimization, especially for robbery and murder. To the extent that community-level variations between racial composition and violent crime merely reflect the aggregation of individual-level effects, the value of community-level research is called into question. The general problem of cross-level misspecification is discussed in more detail in a later section dealing with problems of interpretation.

Housing and Population Density

Although infrequently studied, recent research by Roncek and colleagues (Roncek, 1981; Roncek et al., 1981; Roncek and Faggiani, 1986) has highlighted the potential role that the physical structure and density of housing may play in understanding patterns of violent crime. Roncek (1981) found that the percentage of units in multiunit housing structures was a consistent and rather strong predictor of block-level variations in violent crime in Cleveland and San Diego. Land area in acres, population size, and the percent of primary-individual households also had significant and substantively important effects on violence, despite age and race composition. As Roncek (1981:88) summarizes, "The most dangerous city blocks are relatively large in population and area with high concentrations of primary individuals and apartment housing." This general pattern was confirmed in other model specifications, where Roncek also found that proximity to public housing projects and public high schools had significant positive effects on violence (Roncek et al., 1981; Roncek and Faggiani, 1986). In a similar vein, Schuerman and Kobrin (1986:97) found that increases in multiplex dwellings and renter-occupied housing were major predictors of increases in crime rates in Los Angeles neighborhoods.

Roncek (1981:88) argues that such findings are consistent with

Suggested Citation:"Violent Victimization and Offending: Individual-Situational-, and Community-Level Risk Factors." National Research Council. 1994. Understanding and Preventing Violence, Volume 3: Social Influences. Washington, DC: The National Academies Press. doi: 10.17226/4421.
×

linking features of residential areas to crime through the concept of anonymity: primary-individual households and high proportions of multiunit structures are posited to increase levels of anonymity. For example, as the number of households sharing common living space increases, residents are less able to recognize their neighbors, to be concerned for them, or to engage in guardianship behaviors (Roncek, 1981:88). This notion is derived in part from the defensible space notion that characteristics of the built environment influence the spatial distribution of crime (for a review of this literature, see Taylor and Gottfredson, 1986).

Similar arguments were supported in studies of violent victimization using neighborhood characteristics data in conjunction with the NCS. Sampson (1983:282-283) reported a strong positive influence of the percentage of housing units in structures of five or more units on rates of robbery and assault victimization-regardless of individual risk factors such as age, race, and sex. For example, the ratios of robbery victimization rates for those living in high-density neighborhoods versus those living in low-density neighborhoods were 2.6, 3.6, 3.6, 3.8, 3.5, 3.4, 3.2, and 3.4 for white male adults, white male juveniles, white female juveniles, white female adults, black male adults, black male juveniles, black female juveniles, and black female adults, respectively. Thus it can be seen that rates of victimization are two to three times higher in high-density neighborhoods regardless of compositional factors such as age, race, and sex. Moreover, the effect of density was independent of other key neighborhood characteristics (see Sampson, 1985a).

Finally, several studies report a significant and large association between population density and violent crime. The average correlation across neighborhoods between population density and violent crime rates is .68 in the studies of Beasley and Antunes (1974), Mladenka and Hill (1976), and Smith and Jarjoura (1988). In the latter multivariate study, density also had one of the strongest effects on violent crime, despite controlling for a host of social and economic variables. Characteristics of the physical environment related to housing and population density thus appear to increase the level of violent crime regardless of compositional factors.23

Family Structure

Although largely ignored in early ecological research, several recent studies have turned to examination of the community-level

Suggested Citation:"Violent Victimization and Offending: Individual-Situational-, and Community-Level Risk Factors." National Research Council. 1994. Understanding and Preventing Violence, Volume 3: Social Influences. Washington, DC: The National Academies Press. doi: 10.17226/4421.
×

consequences of family structure. One reason is that high levels of family "disruption" (e.g., divorce rates, female-headed families with children) may facilitate crime by decreasing community networks of informal social control. Examples of informal social control include neighbors' taking note of or questioning strangers, watching over each others' property, assuming responsibility for supervision of general youth activities (e.g., teenage peer groups), and intervening in local disturbances (e.g., see Taylor et al., 1984; Sampson, 1986a; Sampson and Groves, 1989). Note that this conceptualization does not necessarily require that it is the children of divorced or separated parents that are engaging in crime. Rather, the idea is that family disruption can have community wide contextual properties. For instance, youth in stable family areas, regardless of their own family situation, probably have more controls placed on their leisure time activities, particularly with peer groups (Anderson, 1990:91). A well-documented fact in prior research is that delinquency is a group phenomenon (Zimring, 1981); hence neighborhood family structure is likely to be important in determining the extent to which neighborhood youth are provided the opportunities to form a peer-control system free of the supervision or knowledge of adults (Reiss, 1986a).

Felson and Cohen (1980; see also Cohen and Felson, 1979) also note the potential influence of family structure not just on the control of offenders, but on the control of criminal targets and opportunities. They argue that traditional theories of crime emphasize the criminal motivation of offenders without considering adequately the circumstances in which criminal acts occur. Notably, predatory crime requires the convergence in time and space of offenders, suitable targets, and the absence of effective guardianship. The spatial and temporal structure of family activity patterns plays an important role in determining the rate at which motivated offenders encounter criminal opportunities. Compared to married couples, for example, single and divorced persons may be especially vulnerable to personal crimes of violence (e.g., rape, robbery) as a result of decreased guardianship both at home and in public. In support of this idea, Felson and Cohen (1980) demonstrated that primary-individual households had a significant and positive effect on crime trends in the United States in the period 1950-1972, independent of factors presumed to reflect the supply of motivated offenders (e.g., unemployment, youthful age).

The salience of family structure has been supported at the neighborhood level with respect to violence. Studies reporting a large and positive relationship between measures of family disruption

Suggested Citation:"Violent Victimization and Offending: Individual-Situational-, and Community-Level Risk Factors." National Research Council. 1994. Understanding and Preventing Violence, Volume 3: Social Influences. Washington, DC: The National Academies Press. doi: 10.17226/4421.
×

(usually percent female-headed families or divorce rate) and rates of violence include Block (1979), Messner and Tardiff (1986), Sampson (1985a, 1986a), Roncek (1981), Smith and Jarjoura (1988), and Schuerman and Kobrin (1986). What is especially striking is the strength of family disruption in predicting violence in multivariate models. For example, Sampson (1985a, 1986a) found that rates of violent victimization were two to three times higher among residents of neighborhoods with high levels of family disruption compared to low levels, regardless of alternative predictors such as percent black and poverty. In fact, the percentage of female-headed families helped to explain in large part the relationship between percent black and crime. Namely, percent black and percent female-headed families were positively and significantly related; however, when percent female-headed families was controlled, percent black was not significantly related to violent victimization (1985a:27).

Similarly, Messner and Tardiff (1986) found that when percent divorced and percent poverty were controlled, the relationship between percent black and homicide rates was insignificant. Smith and Jarjoura (1988) report that family structure, especially percent single-parent families, helps account for the association between race and violent crime at the community level; racial composition was not significantly related to violent crime in multivariate models once percent single-parent families had been included.

Therefore, it seems that recent emphases on family structure are based on consistent empirical support. Perhaps more to the point, this correlate does not appear to be the result of other factors we typically associate with violence such as poverty, race, and density. Rather, the effect of family disruption is usually independent and large. We now explicate the potential interventing constructs that may help to explain why family disruption and other community structural factors are related to violence.

Community Social Disorganization

A theoretical focus on community social organization, or the lack thereof, provides a useful integrating strategy for understanding the mediating social processes that underlie community-level effects on violence. In general, community social organization may be conceptualized as the ability of a community structure to realize the common values of its residents and maintain effective social controls (Kornhauser 1978:120; Bursik, 1988; Sampson, 1988). The structural dimensions of community social organization can

Suggested Citation:"Violent Victimization and Offending: Individual-Situational-, and Community-Level Risk Factors." National Research Council. 1994. Understanding and Preventing Violence, Volume 3: Social Influences. Washington, DC: The National Academies Press. doi: 10.17226/4421.
×

be measured in terms of the prevalence and interdependence of social networks in a community-both informal (e.g., density of friendship ties and acquaintanceship) and formal (e.g., organizational participation)-and in the span of collective supervision that the community directs toward local problems (Shaw and McKay, 1942; Kornhauser, 1978; Sampson and Groves, 1989). This approach is grounded in what can be termed the systemic model (Sampson, 1988), in which the local community is viewed as a complex system of friendship and kinship networks, and formal and informal associational ties rooted in family life and ongoing socialization processes. Social organization and social disorganization are thus seen as different ends of the same continuum with respect to systemic networks of community social control. As Bursik (1988) notes, when formulated in this way, social disorganization is clearly separable not only from the processes that may lead to it (e.g., poverty, mobility), but also from the degree of criminal behavior that may be a result.

A major dimension of social disorganization relevant to violence is the ability of a community to supervise and control teenage peer-groups -especially gangs (Sampson and Groves, 1989). It is well documented that delinquency is primarily a group phenomenon (Shaw and McKay, 1942; Thrasher, 1963; Short and Strodtbeck, 1965; Zimring, 1981; Reiss, 1986b) and hence that the capacity of the community to control group-level dynamics is a key theoretical mechanism linking community characteristics with crime. Indeed, a central fact underlying Shaw and McKay's research was that the majority of gangs developed from the unsupervised, spontaneous play group (Thrasher, 1963:25). Shaw and McKay (1969) thus argued that residents of cohesive communities were better able to control the teenage behaviors that set the context for group-related violence (see also Thrasher, 1963:26-27; Short, 1963:xxiv; Short and Strodtbeck, 1965; Anderson, 1990). Examples of such controls include supervision of leisure-time youth activities, intervention in street-corner congregation (Maccoby et al., 1958; Thrasher, 1963:339; Shaw and McKay, 1969:176-185), and challenging youth "who seem to be up to no good" (Taylor et al., 1984:326; Skogan, 1986:217). Socially disorganized communities with extensive street-corner peer groups are also expected to have higher rates of adult violence, especially among younger adults who still have ties to youth gangs (Thrasher, 1963).

A second dimension of community social organization involves local friendship networks and the density of acquaintanceship. Systemic theory holds that locality-based social networks constitute

Suggested Citation:"Violent Victimization and Offending: Individual-Situational-, and Community-Level Risk Factors." National Research Council. 1994. Understanding and Preventing Violence, Volume 3: Social Influences. Washington, DC: The National Academies Press. doi: 10.17226/4421.
×

the core social fabric of human-ecological communities (Sampson, 1988). When residents form local social ties, their capacity for community social control is increased because they are better able to recognize strangers and are more apt to engage in guardianship behavior against victimization (Taylor et al., 1984:307; Skogan 1986:216). Also, the greater the density of friendship networks among persons in a community, the greater is the constraint on deviant behavior within the purview of the social network (Krohn, 1986:84).

A third component of social organization is the rate of local participation in formal and voluntary organizations. Community organizations reflect the structural embodiment of local community solidarity (Hunter, 1974:191), and with this in mind, Kornhauser (1978:79) argues that institutional instability and the isolation of community institutions are key factors underlying the structural dimension of social disorganization. Her argument, in short, is that when links between community institutions are weak the capacity of a community to defend its local interests is weakened. Shaw and McKay (1969:184-185), and more recently Taylor et al. (1984) and Simcha-Fagan and Schwartz (1986:688), have also argued that a weak community organizational base serves to attenuate local social control functions regarding youth.

It is difficult to study social disorganization directly, but at least two recent studies provide empirical support for the theory in terms of its structural dimensions. First, Taylor et al. (1984) examined variations in violent crime (mugging, assault, murder, rape, shooting, and yoking) across 63 street blocks in Baltimore in 1978. Based on interviews with 687 household respondents, Taylor et al. (1984:316) constructed block-level measures of what they termed social ties and near-home responsibilities. The former measured the proportion of respondents who belonged to an organization to which co-residents also belonged, whereas the latter measure tapped the extent to which respondents felt responsible for that happened in the area surrounding their home. Both of these dimensions of informal social control were significantly related to community-level variations in violence. Specifically, social organizational ties and near-home social control had standardized effects on violence rates of -.20 and -.24, respectively (Taylor, 1984:320). Taylor et al. (1984:317) also showed that blocks with higher neighborhood identification (and presumably greater social cohesion), as indicated by the proportion of residents who were able to provide a neighborhood name, had significantly lower rates of violence. These results support the social disorganization

Suggested Citation:"Violent Victimization and Offending: Individual-Situational-, and Community-Level Risk Factors." National Research Council. 1994. Understanding and Preventing Violence, Volume 3: Social Influences. Washington, DC: The National Academies Press. doi: 10.17226/4421.
×

hypothesis that levels of organizational participation and informal social control (especially of local activities by neighborhood youth) inhibit community-level rates of violence (see also Taylor et al., 1984:326).

Second, Sampson and Groves (1989) analyzed the British Crime Survey (BCS), a nationwide survey of England and Wales conducted in 1982 and 1984. Sampling procedures resulted in the proportionate selection of 60 addresses within each of more than 200 ecological areas in Great Britain. The sample drawn from each geographical unit was representative of a relatively small, homogeneous locality that reasonably approximated the concept of "local community." Survey-generated responses were aggregated in each of the 238 areas, and structural variables were constructed (e.g., means, percentages). Sampson and Groves (1989:789) report that the prevalence of unsupervised peer groups in a community had the largest overall effect on rates of victimization by mugging/street robbery and stranger violence in Great Britain (betas = .35 and .19, respectively) in 1982. Local friendship networks had a significant and substantial negative effect (beta = -.19) on robbery, whereas rates of organizational participation had significant inverse effects on both robbery and stranger violence (Sampson and Groves, 1989:789). Moreover, the largest overall effect on personal violence offending rates came from unsupervised peer groups (Sampson and Groves, 1989:793). Unsupervised peer groups also had large and substantial positive effects on robbery and assault in the 1984 BCS, whereas local friendship networks again had significant inverse effects (Sampson and Groves, 1989:798).

These empirical results suggest that communities characterized by sparse friendship networks, unsupervised teenage peer groups, and low organizational participation had disproportionately high rates of violence. Furthermore, variations in these structural dimensions of community social disorganization were shown by Sampson and Groves (1989) to mediate in large part the effects of community socioeconomic status, residential mobility, ethnic heterogeneity, and family disruption in the manner predicted by social disorganization theory. Mobility had significant inverse effects on friendship networks; family disruption was a significant predictor of unsupervised peer groups; and socioeconomic status had positive effects on organizational participation. When combined with the results of research on gang delinquency that point to the salience of informal and formal institutional community structures in controlling the formation of gangs (Short and Strodtbeck, 1956; Thrasher, 1963; Hagedorn, 1988), the empirical data suggest

Suggested Citation:"Violent Victimization and Offending: Individual-Situational-, and Community-Level Risk Factors." National Research Council. 1994. Understanding and Preventing Violence, Volume 3: Social Influences. Washington, DC: The National Academies Press. doi: 10.17226/4421.
×

that the structural elements of social disorganization have relevance for explaining macrolevel variations in violence rates (see also Bursik, 1988).

Community Cultures and Ethnography

Theoretically, at least, social disorganization theory also focuses on how the ecological segregation of communities gives rise to what Kornhauser (1978:75) terms cultural disorganization-the attenuation of societal cultural values. In particular, poverty, mobility, heterogeneity, and other structural features of urban communities are hypothesized to impede communication and obstruct the quest for common values, thereby fostering cultural diversity with respect to nondelinquent values (Kornhauser, 1978:75). A consequence of the cultural fragmentation engendered by structural features such as urbanization and heterogeneity (see also Fischer, 1975) is the formation and transmission of deviant and violent subcultures. Accordingly, a major component of Shaw and McKay's theory was that heterogeneous, low-income, urban communities spawned the formation of delinquent organizations (e.g., gangs) with their own subcultures and norms perpetuated through cultural transmission.24

The ethnographic study of community cultures and value systems in relation to social disorganization has not received nearly the same attention as quantitative studies that examine variations in crime rates. This is unfortunate because community ethnographies and gang studies provide important insights into culturally patterned, group responses to structurally imposed conditions. It is beyond the scope of this paper to attempt an assessment of community ethnography, much less come to some conclusion regarding the validity of cultural theories of violence. However, a brief discussion of community ethnographies is useful in underscoring the need for an integrated research approach to macrolevel sources of violence.

In general, ethnographic studies suggest that the ecological basis for the formation of local street-corner groups and gangs is especially pronounced in urban, lower-income, heterogeneous neighborhoods (Thrasher, 1963; Spergel, 1964; Short and Strodtbeck, 1965; Suttles, 1968; Keiser, 1970; Rieder, 1985; Horowitz, 1987; Hagedorn, 1988). A related thread running through many ethnographic studies is the idea that structurally disorganized communities are conducive to the emergence of cultural value systems and attitudes that seem to legitimate, or provide a basis of tolerance

Suggested Citation:"Violent Victimization and Offending: Individual-Situational-, and Community-Level Risk Factors." National Research Council. 1994. Understanding and Preventing Violence, Volume 3: Social Influences. Washington, DC: The National Academies Press. doi: 10.17226/4421.
×

for, violence. For example, Suttles's (1968) account of the social order of a Chicago slum neighborhood characterized by poverty and heterogeneity supports Thrasher's (1963) emphasis on age, sex, ethnicity, and territory as markers for the ordered segmentation of slum culture. Namely, Suttles found that single-sex, age-graded primary groups of the same ethnicity and territory emerged in response to threats of conflict and community-wide disorder and mistrust. Although the community subcultures that Suttles (1968) discovered were provincial, tentative, and incomplete (Kornhauser, 1978:18), they nonetheless undermined societal values against delinquency and violence.

Similarly, Anderson's (1978) ethnography of a bar in Chicago's South Side black ghetto shows how collective or group processes construct a segmented social order. Within the extended primary group that gathered at ''Jelly's" bar, mainstream values such as a "visible means of support" and "decency" were indeed primary values, but there coexisted residual values associated with deviant subcultures (e.g., hoodlums) such as "toughness," "getting big money," "going for bad," and "having fun" (1978:129-130, 152-158). In Anderson's (1978:210) analysis, lower-class residents do not so much "stretch" mainstream values as "create their own particular standards of social conduct along variant lines open to them." In this context the use of violence is not valued as a primary goal, but is still expected and tolerated as a fact of life (Anderson, 1978:134; see also Horowitz, 1987). In the case of certain primary groups such as hoodlums, violence is seen as one of the few ways to gain respect or affirmation of self (Anderson, 1978:158). Much like Rainwater (1970), Suttles (1968), and Horowitz (1987), Anderson's ethnographic research suggests that in certain community contexts the wider cultural values are simply not relevant-they become "unviable."

To be sure, there is great dispute whether subcultural values are in fact genuine or are in essence fake values that are used to rationalize behavior. Liebow's (1967) study of a black slum in Washington questions the authenticity of a community culture that is in opposition to the wider culture. Rather than assuming an historical continuity in values, Liebow (1967) argues that lowerclass values were recreated as persons time and again experienced the same structural conditions. Relying on Liebow's research, Kornhauser (1978:20) further critiques cultural theories by asserting,

The "culture" thus constructed is not an authentic subculture, nor are its "values" objects of genuine commitment. It is a

Suggested Citation:"Violent Victimization and Offending: Individual-Situational-, and Community-Level Risk Factors." National Research Council. 1994. Understanding and Preventing Violence, Volume 3: Social Influences. Washington, DC: The National Academies Press. doi: 10.17226/4421.
×

pseudoculture. Its function is ideological: to rationalize failure. It neither commands allegiance nor guides behavior; rather, it follows behavior.

In her viewpoint, violence is therefore not valued but is instead a situational response to the structurally imposed conditions of social disorganization.

Whether or not community subcultures are authentic or merely "shadow cultures" (Liebow, 1967) cannot be resolved here. However, that seems less important than acknowledging that community contexts seem to shape what can be termed cognitive landscapes or ecologically structured norms (e.g., normative ecologies) regarding appropriate standards and expectations of conduct (cf. Liebow, 1967; Suttles, 1968; Rainwater, 1970; Anderson, 1978, 1990; Horowitz, 1987; Rieder, 1985). That is, in certain structurally disorganized communities it appears that a system of values emerges in which violence is less than fervently condemned and hence expected as part of everyday life. These ecologically structured social perceptions and tolerances in turn appear to influence the probability of violent encounters (Horowitz, 1987). In this regard we believe that Kornhauser's attack on subcultural theories misses the point. By attempting to assess whether subcultural values are authentic in some deep, quasi-religious sense, Kornhauser (1978:1-20) loses sight of the processes by which cognitive landscapes rooted in normative ecologies can influence rates of violence. Indeed, the idea that dominant values become existentially irrelevant in certain community contexts is a powerful one, albeit an idea that has not had the research exploitation it deserves. We return later to a discussion of community ethnographies and how they might be more profitably linked to structurally oriented studies of variation in crime rates.

SUMMARY

To recapitulate, studies of violence have both confirmed and substantially extended the basic Shaw and McKay model of community structure and delinquency. Almost without exception, studies of violence find a positive and usually large correlation between some measure of area poverty and violence-especially homicide. However, the strength of poverty in explaining violence once other community factors are controlled has been called into question. Many of the studies reviewed find weak direct effects of poverty or conditional effects. In particular, it appears that poverty is most important in the context of community change.

Suggested Citation:"Violent Victimization and Offending: Individual-Situational-, and Community-Level Risk Factors." National Research Council. 1994. Understanding and Preventing Violence, Volume 3: Social Influences. Washington, DC: The National Academies Press. doi: 10.17226/4421.
×

A similar phenomenon emerges for racial composition and heterogeneity: although violence rates are usually higher in neighborhoods with a high percentage of blacks or minorities, the direct effect of race is often quite weak. The more consistent and sturdy structural correlates of violence seem to be residential mobility or change-especially when linked to increasing poverty and social dislocation-and the factors of family disruption and housing/population density.

Additionally, there is evidence that intervening dimensions of community structural disorganization, such as unsupervised peer groups, attenuated local friendship networks, and low organizational participation, have positive effects on violence. Also based on community ethnographies, there is some evidence to suggest that structurally disorganized communities with weak social controls give rise to the formation of delinquent gangs and cultural norms that increase the probability of violent encounters. There are clearly many weaknesses in the community-level literature that need explication, especially regarding interpretations of causality and the authenticity of cultural attitudes promoting violence. Before discussing these, however, we turn to a brief summary of findings based on interurban units of analysis.

METROPOLITAN-LEVEL STUDIES

Probably because of the greater ease of data availability, recent literature on violence in relation to characteristics of metropolitan-level or interurban units of analysis such as cities and standard metropolitan statistical areas is much more extensive than research at the local community level. A recurrent theme in this body of research has been the potential criminogenic effects of income inequality, especially between the races, on violence. Inequality is thought to generate subjective feelings of relative deprivation and latent hostility, leading to criminal acts of violence born of frustration (e.g., see Messner, 1982; Blau and Blau, 1982; Balkwell, 1990). Another dominant concern has been regional effects on the homicide rate. The fact that southern states exhibit homicide rates higher than other regions has led to speculation that there exists a value system that supports and reinforces violent behavior (Gastil, 1971). This perspective, widely known as the "subculture of violence" thesis, has also been applied to the explanation of disproportionately high rates of black violence in American cities (Wolfgang and Ferracuti, 1967; Curtis, 1975).

Suggested Citation:"Violent Victimization and Offending: Individual-Situational-, and Community-Level Risk Factors." National Research Council. 1994. Understanding and Preventing Violence, Volume 3: Social Influences. Washington, DC: The National Academies Press. doi: 10.17226/4421.
×

TABLE 1 Recent Macro Level (city, SMSA, state) Studies of Homicide

Author(s)

Unit of Analysis

Data Year

Gastil (1971)

States, SMSAs

1960

Bailey (1984)

Cities

1960, 1970

Loftin and Hill (1974)

States

1960

Jackson (1984)

Cities

1970

Loftin and Parker (1985)

Cities

1970

Messner (1983b)

Cities

1970

Sampson (1985b)

Cities

1970

Sampson (1986b)

Cities

1980

Williams and Flewelling (1988)

Cities

1980

Blau and Blau (1982)

SMSAs

1970

Blau and Golden (1986)

SMSAs

1970

Crutchfield et al. (1982)

SMSAs

1970

DeFronzo (1983)

SMSAs

1970

Messner (1982)

SMSAs

1970

Messner (1983a)

SMSAs

1970

Rosenfeld (1986)

SMSAs

1970

Simpson (1985)

SMSAs

1970

Williams (1984)

SMSAs

1970

Huff-Corzine et al. (1986)

States

1970

Parker and Smith (1979)

States

1970

Smith and Parker (1980)

States

1970

 

SOURCE: Data from Land et al. (1990).

To illustrate these trends, Table 1 is a list of 21 major studies identified by Land et al. (1990) that examine variations in homicide rates across macrolevel units of cities, SMSAs, and in a few cases, states.25 The following variables are represented in this body of research: population size, population density, percent of the population that is black, percent youth (e.g., percent age 15-29), percent male divorce rate, percent of children not living with both parents, median family income, percent families living in poverty, the Gini index of income inequality, the unemployment rate, and a dichotomous variable indicating those cities or metropolitan areas located in the South. Despite the wide variety of measures, the majority of studies in Table 1 were clearly focused on economic deprivation and inequality.

Detailed inspection of these studies reveals a host of differences that render a summary of findings difficult if not impossible. For one, the model specifications, techniques of analysis, time periods of study, and units of analysis differ widely across studies-much more than is the case at the community level (see

Suggested Citation:"Violent Victimization and Offending: Individual-Situational-, and Community-Level Risk Factors." National Research Council. 1994. Understanding and Preventing Violence, Volume 3: Social Influences. Washington, DC: The National Academies Press. doi: 10.17226/4421.
×

also Byrne and Sampson, 1986). Comparison of multivariate standardized regression coefficients across studies is especially ill-advised because of widely discrepant model specifications and variances. Perhaps then it is not surprising that the findings themselves are highly inconsistent. As Land et al. (1990:931) concluded after their detailed review, "Virtually no structural covariate exhibits estimated effects that are statistically significant and of invariant algebraic sign across all studies." For example, poverty has both positive (Bailey, 1984) and negative (Messner, 1982) significant coefficients in different specifications, and is insignificant in many other models (Blau and Blau, 1982; Crutchfield et al., 1982; Simpson, 1985). The Gini index of income inequality also shows considerable inconsistency, sometimes directly related (Simpson, 1985) and sometimes not related (Williams, 1984) to rates of homicide. The only apparent exception to this pattern is percent divorced, which has rather consistent positive associations with homicide in the minority of studies in which it was measured (e.g., Blau and Blau, 1982; Williams, 1984). Percent black also shows a fairly consistent pattern of positive covariation with homicide (Messner, 1982, 1983b; Blau and Blau, 1982; Bailey, 1984; Simpson, 1985; Sampson, 1986b).

In short, the dozens of macrolevel studies conducted over the past 20 years or so are not easily interpretable in terms of the independent influences of urban social structure on violence. What can account for these highly variable findings? In addition to differing time periods and units of analysis (e.g., states, SMSAs, cities), Land et al. (1990) maintain that the single greatest threat to comparability across studies is the presence of high levels of association among explanatory variables. Typically referred to as multicollinearity, redundancy among predictor variables threatens the statistical inferences that can be drawn from the numerous studies based on multiple regression. Specifically, collinearity among explanatory variables can result in large confidence intervals and instability of coefficients across model specifications (Hanushek and Jackson, 1977:86-96). Gordon (1967) warned many years ago that overlap among predictor variables leads to the "partialling fallacy"-the allocation of all explained variance to that regressor among a highly interrelated set of regressors that possesses the higher correlation with the dependent variable(s), even though there may be no theoretical or substantive basis for doing so (Land et al., 1990). The simple fact is that variables such as percent black, poverty, unemployment, median income, and inequality are usually highly intercorrelated. The assessment of

Suggested Citation:"Violent Victimization and Offending: Individual-Situational-, and Community-Level Risk Factors." National Research Council. 1994. Understanding and Preventing Violence, Volume 3: Social Influences. Washington, DC: The National Academies Press. doi: 10.17226/4421.
×

independent effects results in highly unstable regression models, making comparison of magnitudes of effect extremely difficult and problematic.

To address this problem, Land et al. (1990) have conducted what may be the most comprehensive analysis to date of homicide rate variation at the metropolitan and state level. They first identified what they termed a "baseline regression model" that incorporated the major predictor variables used by the 21 studies in Table 1. The predictor variables, labeled "structural covariates," were population size, population per square mile, percent black population, percent of the population age 15 to 29, percent divorced males, percent of children not living with both parents, median family income, percent of families living below the official poverty level, the Gini index of income concentration, percent unemployed, and a dichotomous variable indicating southern location. These 11 variables were measured at the city, SMSA, and state levels of analysis for 1960, 1970, and 1980. The cities examined for 1980 number almost 1,000 (904); the SMSAs, 259.

Using multivariate regression models similar to prior research, Land et al. (1990) report, as expected, substantial variability and fluctuation in regression coefficients across time periods and units of analysis. However, when the authors analyzed the dimensionality of the structural covariates through principal components analysis, two clusters of variables were found to consistently "hang together" over time and space. The first factor was termed a population structure component and consisted of population size and population density. The second factor was labeled resource deprivation/ affluence and included three income variables-median income, percent of families below the poverty line, and the Gini index of income inequality-in addition to percent black and percent of children not living with both parents. Although these variables seem to tap somewhat different concepts, Land et al. (1990) found they could not be separated empirically. In particular, there was no empirical justification for treating deprivation measures (i.e., poverty) and inequality as separate independent variables. 26 As Land et al. (1990:944) argue, cities, SMSAs, and states with high (low) levels of absolute deprivation also tend to have high (low) levels of inequality. Percent black was also substantially collinear with economic indicators. On the other hand, four variablespercent divorced males, percent age 15 to 29, unemployment rate, and South-varied independently.

Based on these results, Land et al. (1990) entered the two components of resource deprivation/affluence and population structure

Suggested Citation:"Violent Victimization and Offending: Individual-Situational-, and Community-Level Risk Factors." National Research Council. 1994. Understanding and Preventing Violence, Volume 3: Social Influences. Washington, DC: The National Academies Press. doi: 10.17226/4421.
×

TABLE 2 Standardized Regression Coefficients of Structural Covariates of Homicide Rates for U.S. Cities, SMSAs, and States in 1960, 1970, and 1980

 

1960

1970

1980

 

City

SMSA

State

City

SMSA

State

City

SMSA

State

Population structure

11a

.04

.15a

.20a

.25a

.19a

.19a

.18a

.27a

Resource deprivation

.56a

.45a

.60a

.58a

.57a

.77a

.53a

.54a

.80a

Percent divorced

.10a

.24a

.36a

.18a

.28a

.32a

.12a

.26a

.58a

Percent ages 15-29

-.06

-.01

.32a

.00

.01

.24a

-.11a

-.03

.17a

Unemployment rate

-.11a

-.12a

-.01

-.09a

-.23a

-.01

.18

-.09a

-.09

Southern region

.20a

.36a

.24a

.13a

.14a

.00

.16a

.14a

-.08

a Significant at p = .05, one-tailed.

SOURCE: Data from Land et al. (1990:948-949).

along with the four single variables in an attempt to examine the invariance of homicide determinants across time and space. Table 2 summarizes the major findings; entries are standardized regression coefficients for cities, SMSAs, and states for the three time periods covered by recent research-1960, 1970, and 1980. The results are quite consistent and substantively interpretable. As Land et al. (1990:947) summarize, "Three structural indexes/covariates-population structure, resource deprivation/affluence, and percent of the male population divorced-now exhibit statistically significant relationships to the homicide rate in the theoretically expected direction across all time periods and levels of analysis." For example, Table 2 shows that resource deprivation has a consistent and large positive effect on homicide, independent of all other factors. This suggests that, all else being equal, cities with a large poor population in conjunction with a high percentage of black residents and single-parent families with children, have disproportionately high homicide rates. Further, homicide

Suggested Citation:"Violent Victimization and Offending: Individual-Situational-, and Community-Level Risk Factors." National Research Council. 1994. Understanding and Preventing Violence, Volume 3: Social Influences. Washington, DC: The National Academies Press. doi: 10.17226/4421.
×

rates are larger in cities/states/regions with large populations and high male divorce rates. In terms of magnitude, resource deprivation has the largest effect, generally followed by male divorce rate and population structure. Other variables show more inconsistencies, although there is a tendency for homicide to be higher in the South, and in areas with a high percentage of youth and low employment.

The large effect of resource deprivation/affluence on homicide is substantively consistent with Wilson's (1987) recent conceptualization of concentration effects. He argues that the social transformation of the inner city has resulted in a disproportionate concentration of the most disadvantaged segments of the urban black population-especially poor, black, female-headed families with children. Factors leading to the increased segregation of poor blacks include opposition from organized community groups to public housing projects in stable neighborhoods and political decisions to neglect rehabilitation of single-family residential units (Massey, 1990; Sampson, 1990). Urban minorities have also been especially vulnerable to structural economic changes related to the deindustrialization of central cities (e.g., shift from goods-producing to service-producing industries, increasing polarization of the labor market into low-wage and high-wage sectors, "high-tech" innovations, and relocation of manufacturing out of the inner city). This social milieu differs significantly from the environment that existed in inner cities in previous decades (Wilson 1987:58; Rose and McClain, 1990). In particular, the exodus of middle-and upper-income black families from the inner city removes an important social buffer that could deflect the full impact of prolonged joblessness and industrial transformation (see also Hagedorn, 1988). Wilson's (1987:56) thesis is based on the assumption that the basic institutions of an area (e.g., churches, schools, stores, recreational facilities) would remain viable if much of the base of their support came from more economically stable families in inner-city neighborhoods (i.e, those with vertical social class integration).

Wilson's (1987) concept of social isolation in areas of concentrated poverty may help explain the Land et al. (1990) findings and also those of Taylor and Covington (1988) noted earlier, who found that increasing entrenchment of the urban minority underclass was associated with large increases in violence. Land et al.'s (1990) results also go beyond Wilson by suggesting that the clustering of economic and social indicators appears not only in 1980 and in neighborhoods of large cities, but also for the two previous

Suggested Citation:"Violent Victimization and Offending: Individual-Situational-, and Community-Level Risk Factors." National Research Council. 1994. Understanding and Preventing Violence, Volume 3: Social Influences. Washington, DC: The National Academies Press. doi: 10.17226/4421.
×

decennial periods and at the level of macrosocial units as a whole. In other words, those cities (as well as SMSAs and states) that had low incomes, high poverty levels, and great economic inequality in 1960, 1970, and 1980 also had large concentrations of blacks and children living in poverty (Land et al., 1990:944). Moreover, Land et al. present evidence in support of Wilson's argument that concentration effects grew more severe from 1970 to 1980 in large cities-"the numerical values of the component loadings of percent poverty, percent black, and percent of children under 18 not living with both parents are larger in 1980 than 1970" (Land et al., 1990:945). Therefore, not only are various indicators of urban disadvantaged populations highly related, they are increasing in concentration (Massey and Eggers, 1990).

Metropolitan-level results for robbery and rape parallel in large part Land et al.'s (1990) findings for homicide. The first and most obvious point is that the overlap among explanatory variables is the same whether robbery, assault, rape, or homicide is analyzed. As might be expected, then, a similar pattern appears whereby percent black and different measures of deprivation/ poverty are positively related to each other and to both robbery (Danziger, 1976; Crutchfield et al., 1982; Blau and Blau, 1982; Carroll and Jackson, 1983; Blau and Golden, 1986; Byrne, 1986; Rosenfeld, 1986; Sampson, 1986b) and rape (Blau and Blau, 1982; Peterson and Bailey, 1988; for further review, see Brantingham and Brantingham, 1984:289-294). Moreover, both Flango and Sherbenou (1976) and Harries (1976) analyzed the multidimensionality of independent variables and extracted factors similar to Land et al. (1990). Namely, Flango and Sherbenou (1976:337) found that percent black, percent female-headed families, and low median income covaried together, whereas Harries (1976) discovered that percent black and percent female headed households loaded on a single factor. These "deprivation" indices were in turn significantly related to city-level variations in violence.

Perhaps more clear cut is the factor of divorce: several studies have found a strong direct effect of divorce on rates of robbery (e.g., Blau and Blau, 1982; Blau and Golden, 1986; Sampson, 1986b), assault (Blau and Golden, 1986), and rape (Blau and Golden, 1986; Petersen and Bailey, 1988). Again, this parallels the Land et al. (1990) finding of a consistent direct effect of the male divorce rate on homicide that is unaccounted for by socioeconomic factors and racial composition. In fact, the effect of divorce on robbery and rape is generally greater than that of socioeconomic factors and racial composition in studies where the former has been distinctly

Suggested Citation:"Violent Victimization and Offending: Individual-Situational-, and Community-Level Risk Factors." National Research Council. 1994. Understanding and Preventing Violence, Volume 3: Social Influences. Washington, DC: The National Academies Press. doi: 10.17226/4421.
×

measured (e.g., Blau and Golden, 1986; Sampson, 1986b; Petersen and Bailey, 1988). For example, in the most recent study to focus explicitly on city-level variations in rape rates, Petersen and Bailey (1988) found that the divorce rate had the largest effect on rape (beta = .45), followed by percent black (.29), overall inequality (.25), church membership (-.19), and black/white inequality (.12).

Other than what Land et al. (1990) refer to as resource deprivation/ affluence and the divorce rate, macrolevel research has shown a rather consistent direct effect on robbery of either population size/density or housing density (Flango and Sherbenou, 1976; Danziger, 1976; Harries, 1976; Carroll and Jackson, 1983; Byrne, 1986; Sampson, 1987a). The density/size findings replicate the population structure conclusion drawn by Land et al. (1990).27 In a study of variations in violent crime across 65 SMSAs in 1970, Crutchfield et al. (1982) also demonstrated that independent of traditional socioeconomic and demographic controls, population mobility had significant and rather substantial effects on rape, assault, and murder for a sample of SMSAs, a finding supportive of the community-level research reviewed earlier. Finally, Baron and Straus (1987) found that state-level rape rates were directly related to an index dominated by mobility and family disruption.

RACIAL DISAGGREGATION

The Land et al. (1990) approach is important because it provides a summary index of the effects of poverty and relative deprivation on homicide. However, the resource deprivation index still confounds racial composition, and consequently does not allow examination of poverty and family structure independent of race. This result presents a problem because the large structural differences among black and white communities in terms of family structure and poverty (see Wilson, 1987) may account for the compositional effect of percent black in aggregate analyses. Racial differences are so strong that even the "worst" urban contexts in which whites reside with respect to poverty and family disruption are considerably better off than the average (i.e., mean level) context of black communities (Sampson, 1987a:354). For example, regardless of whether a black juvenile is raised in an intact or a "broken" home, or a rich or poor home, he/she will not grow up in a community context similar to whites with regard to family structure and poverty.

One way to examine these contextual issues is through racial

Suggested Citation:"Violent Victimization and Offending: Individual-Situational-, and Community-Level Risk Factors." National Research Council. 1994. Understanding and Preventing Violence, Volume 3: Social Influences. Washington, DC: The National Academies Press. doi: 10.17226/4421.
×

disaggregation of both the crime rate and the explanatory variables of theoretical interest (e.g., poverty, family structure), thereby deconfounding race at the aggregate level. This approach was used in research by Sampson (1987a) that examined racially disaggregated rates of homicide and robbery by juveniles and adults in more than 150 U.S. cities in 1980. Racially disaggregated rates of violent offending were estimated from race, age, and sex-specific arrest data corrected for city-level differences in enforcement patterns (Sampson, 1987a:358-360). Substantively, the model focused on the exogenous factors of black male joblessness and economic deprivation, and their effects on violent crime as mediated by black family disruption. The rationale stemmed largely from Wilson's (1987) argument that the labor market marginality of black males and accompanying economic deprivation have had profound negative implications for the black community, especially black women with children. From this viewpoint, a major structural determinant of black family disruption is black male joblessness and persistent poverty in the black community. Family disruption was hypothesized to subsequently increase violence because of weakened informal social control (Sampson, 1987a:352-353).

Overall, the results supported the main hypothesis and showed that the scarcity of employed black males relative to black women was directly related to the prevalence of families headed by females in black communities. In turn, black family disruption was substantially related to rates of black murder and robbery, especially by juveniles. These effects were independent of income, region, race and age composition, density, city size, and welfare benefits. The consistent finding that family disruption had stronger effects on juvenile violence than on adult violence, in conjunction with the inconsistent findings of previous research on individual-level delinquency and broken homes (Wilkinson, 1980), tends to support the idea that the effects of family structure are related to macrolevel patterns of social control and guardianship, especially regarding youth and their peers (Felson and Cohen 1980; Sampson 1986b; Sampson and Groves, 1989). Moreover, the results suggested a partial solution as to why unemployment and economic deprivation have had weak or inconsistent effects on violence rates in past research: although joblessness or even poverty may not have direct effects on crime, they do have significant effects on family disruption, which in turn is a strong predictor of variations in urban black violence.28

The results also revealed that despite a tremendous difference in mean levels of family disruption among black and white communities,

Suggested Citation:"Violent Victimization and Offending: Individual-Situational-, and Community-Level Risk Factors." National Research Council. 1994. Understanding and Preventing Violence, Volume 3: Social Influences. Washington, DC: The National Academies Press. doi: 10.17226/4421.
×

the percentage of white families headed by a female had a large positive effect on white juvenile and white adult violence. In fact, the predictors of white robbery were shown to be in large part identical in sign and magnitude to those for blacks. Therefore, Sampson (1987a) concluded that the effect of family disruption on black crime was independent of commonly cited alternative explanations (e.g., poverty, region, urbanization, age composition) and could not be attributed to unique cultural factors within the black community.

OPPORTUNITY STRUCTURES

In addition to the sociodemographic and economic variables reviewed above, a few recent studies have focused on macrolevel variations in opportunity structures and how they affect violence. We discuss here two of the more salient opportunity structures-aggregate patterns of leisure activity and gun availability. In an innovative study of the former, Messner and Blau (1987) report a significant positive association between violence rates and routine activities as conceptualized by Cohen and Felson (1979). To determine the prevalence of leisure activity that takes place at home they constructed a measure of aggregate level of television viewing for 124 SMSAs in 1980. The level of nonhousehold leisure activities was operationalized by the supply of sports and entertainment establishments. These establishments include commercial cinemas and drive-ins, entertainment producers, and profit-making professional and semiprofessional sports establishments that ''summarize the availability and the opportunities for a wide range of activities that take place outside of the household" (Messner and Blau, 1987:1039).

Controlling for region, percent black, percent poor, and the age, race, and sex composition of SMSAs, Messner and Blau (1987:1043) found consistent significant effects of aggregate routine activity patterns on violence. The nonhousehold activity index had positive effects on homicide, rape, robbery, and assault, whereas the television viewing index had significant negative effects on rape, robbery, and assault. These effects were lower in magnitude than poverty and racial composition, but the television viewing index had a rather substantial relationship with robbery (third largest effect) and assault (second largest effect). Messner and Blau's (1987) conclusion was that nonhousehold activities place members of the population at a relatively high risk for personal victimization, independent of the supply of motivated offenders. Contrariwise,

Suggested Citation:"Violent Victimization and Offending: Individual-Situational-, and Community-Level Risk Factors." National Research Council. 1994. Understanding and Preventing Violence, Volume 3: Social Influences. Washington, DC: The National Academies Press. doi: 10.17226/4421.
×

an increase in leisure activities at home reduces violent victimization. These findings support the basic contention of the routine activity model (Cohen and Felson, 1979). A potential problem with these results is that high levels of violence may lead people to change their activity patterns. The issue of reciprocal effects is discussed more below.

Another factor that has caused great concern in policy debates is the effect of gun ownership on violence. We cannot do justice to this debate, but we can briefly highlight research that specifically focuses on the question of whether gun availability affects patterns of violence across cities. Perhaps the most definitive test can be found in Cook's (1979) analysis of the relationship between gun availability and robbery rates in the 50 largest American cities in 1975. Guns were found to be present in a high proportion of robberies (45 percent) and robberies perpetrated with guns were much more likely to result in the death of the victim than robberies without guns. However, Cook's results showed that an increase in the density of guns in a city-defined as the fraction of households in a city that owns guns-had no effect on the overall robbery rate (1979:743). This finding suggests that gun availability may increase the lethality or seriousness of robberies (by increasing robbery-murder) but that it does not necessarily increase the robbery rate. Similar results were obtained by Bordua (1986) in a study of county-level variations in firearms ownership and violent crime in Illinois. Once county-level factors such as poverty and urbanization were controlled, Bordua found no consistent relationship between firearms ownership and rates of homicide or other violent crimes.

The evidence to date points to the overall conclusion that metropolitan differences in gun ownership are probably not a major factor in the etiology of violence. As Bordua (1986:188) concluded, the most plausible interpretation "is that variation in firearms ownership has no independent causal effect on violent crime at all." The issue is complex and unresolved, however, for there may be counterbalancing effects that account for these null findings. For example, Cook (1979) notes that an increase in gun availability may actually decrease the robbery rate because potential robbery victims are more likely to be armed in cities where gun ownership is widespread. Pierce and Bowers (1981) argue that restriction of guns could actually lead to more violence because potential offenders may feel less restrained if fewer guns are being carried into assault-prone situations. If true, gun density could have both positive and negative effects on the violent crime rate.

Suggested Citation:"Violent Victimization and Offending: Individual-Situational-, and Community-Level Risk Factors." National Research Council. 1994. Understanding and Preventing Violence, Volume 3: Social Influences. Washington, DC: The National Academies Press. doi: 10.17226/4421.
×

Although current research has not effectively disentangled these possibilities, it is nonetheless clear that the simple assumption that more guns necessarily means more violence is not borne out by the evidence.

PROBLEMS OF INTERPRETATION

This paper has synthesized current knowledge on individual, situational, and macrolevel risk factors in understanding violence. Unfortunately, it is evident that numerous problems plague this sort of research, especially inferences as to the causal status of community characteristics. Among other limitations, the use of varying and sometimes highly aggregated units of analysis, potentially biased sources of information on violence, widely varying analytical techniques, and high correlations among independent variables all lead to considerable ambiguity in the attempt to forge general conclusions. However, rather than dwell on these methodological limitations of macrolevel research, many of which have been explicated above or in other sources (e.g., Land et al., 1990), this section considers more fundamental and as yet unresolved problems relating to (1) reciprocal effects of violence on community structure and routine activities, (2) the failure to examine macrolevel processes of theoretical interest, and (3) cross-level misspecification and selection effects. We believe that these three issues are most central to an inability to make definitive conclusions about the independent causal effects of individual-and community-level sources of criminal violence. We then conclude with a discussion of research strategies that may help alleviate these pitfalls.

FEEDBACK EFFECTS

A typical assumption of most community-level research is that crime is an endogenous variable-something to be explained by a variety of social and demographic variables such as poverty and mobility. The problem with this approach is that crime and its consequences (e.g., fear) may themselves have important effects on community structure, thereby confounding easy interpretation of the relationships among community social phenomena and crime. As Bursik (1986:63-64) has stated, "There may not be a simple, unidirectional causal structure in which crime and delinquency rates are only the outcome of ecological processes. Rather, within the context of ongoing urban dynamics, the level of delinquency

Suggested Citation:"Violent Victimization and Offending: Individual-Situational-, and Community-Level Risk Factors." National Research Council. 1994. Understanding and Preventing Violence, Volume 3: Social Influences. Washington, DC: The National Academies Press. doi: 10.17226/4421.
×

in an area may also directly or indirectly cause changes in the composition of an area."

Skogan (1986) has recently provided an insightful overview of some of the "feedback" processes that may further increase levels of crime. These include (1) physical and psychological withdrawal from community life, (2) weakening of the informal social control processes that inhibit crime, (3) a decline in the organizational life and mobilization capacity of the neighborhood, (4) deteriorating business conditions, (5) the importation and domestic production of delinquency and deviance, and (6) further dramatic changes in the composition of the population. For example, to the extent that people shun their neighbors and local facilities out of fear of violence, fewer opportunities exist for friendship networks and neighborhood organizations to take hold (see also Taylor and Gottfredson, 1986:404). As Skogan (1986:217) reminds us, "Fear does not stimulate participation in collective efforts to act against crime; rather, it often has the effect of undermining commitment to an area and interest in participation." Relatedly, predatory street violence is often accompanied by commercial decline and business relocation out of inner-city areas where there is a perceived risk of violence against the customer population. Therefore, violence itself may lead to a weakening of local friendship bonds, informal social control, and the mobilization capacity of communities. Because these are defining elements of social disorganization, serious questions are raised concerning the direction of causal effects between community structure and violence.

The research noted earlier on firearms ownership illustrates a similar point with respect to self-protection and responses to violence. Even if a positive relationship were definitively established between firearms (e.g., gun density) and crime, the nature of the causal interpretation is far from straightforward. As Bordua (1986) notes, those most vulnerable to violent crimes such as rape (e.g., women) may purchase weapons out of a perceived need for self-defense. In fact, the finding that female gun ownership is related positively to the murder rate is attributed by Bordua (1986) to females' purchase of weapons in high-crime areas for defensive purposes. Kleck (1984) further argues that gun robbery and robbery-murder may induce commercial establishments to take self-protective measures with regard to firearms. These studies suggest that the relationship between gun density and crime is complex because of potential reciprocal effects of violence (Cook, 1979:753).

It follows that the whole area of self-protective responses to violence is problematic in that many community-level opportunity

Suggested Citation:"Violent Victimization and Offending: Individual-Situational-, and Community-Level Risk Factors." National Research Council. 1994. Understanding and Preventing Violence, Volume 3: Social Influences. Washington, DC: The National Academies Press. doi: 10.17226/4421.
×

variables may be directly influenced by the violence rate (Cook, 1986). Private security, minimizing cash at convenience stores, and alteration of lifestyles (e.g., reduction in going out at night) are but a few plausible responses to violence in the community (see also Sherman et al., 1989). Again, however, community-level variations in what may be responses to violence are often simply used as causal predictors of violence. For example, the Messner and Blau (1987) study reviewed above used aggregate levels of at-home routine activities to explain violence rates. As with gun ownership, however, responses to violence (e.g., staying at home) in part determine a community's structure of opportunities and level of routine activities, which may in turn feed back into the violence rate in an ongoing reciprocal fashion. On the flip side, consider Cook's (1986) observation that a reduction in the commercial robbery rate may be followed by a reduction in incentives for self-protection (e.g, guards, minimizing cash), thereby increasing the attractiveness of stores as robbery targets. Similarly, once residents of a community feel that streets are safe they are more apt to utilize public facilities and hence generate new opportunity structures for crime.

Perhaps the most salient consequence of violent crime is sheer demographic collapse-extreme population loss and/or selective out-migration from the neighborhood. As Skogan (1986:223) has argued, the massive suburbanization of the United States following World War II may be the most consequential effect of crime on American society. Although sparse, there is evidence to support this view. Bursik (1986) showed that delinquency rates greatly accelerated processes of neighborhood decline in Chicago from 1950 to 1970. At the national level, Sampson and Wooldredge (1986) found that rates of violence were inversely related to population change from 1970 to 1980 in the nation's 55 largest cities. Such a "hollowing out" of many city centers may lead to further deterioration and decline, which, through feedback processes, can increase violence and disorder (see also Reiss, 1986a:8; Wallace and Wallace, 1990). As noted earlier, Wilson (1987) has argued that the crime-influenced exodus of middle-and working-class blacks from many ghetto neighborhoods may have removed an important "social buffer."

In short, predatory crimes of violence such as robbery are fearful to the majority of citizens and can quickly symbolize the decline of communities. Moreover, the activities of potential victims and offenders interact in ways that mutually structure the opportunities for violence in a dynamic process (Cook, 1986; Reiss,

Suggested Citation:"Violent Victimization and Offending: Individual-Situational-, and Community-Level Risk Factors." National Research Council. 1994. Understanding and Preventing Violence, Volume 3: Social Influences. Washington, DC: The National Academies Press. doi: 10.17226/4421.
×

1986a). It should not be surprising then to consider that violence rates are not simply the outcome of ecological processes but are also important components of that process and can have important effects on the dynamics of urban change (Bursik, 1986). What is particularly problematic from the present perspective is that many of the community-level characteristics thought to cause violence (e.g., mobility, population loss, increasing concentration of the underclass, gun ownership, routine activities, weak friendship networks, attenuation of local social control and community organizations) may themselves be affected by feedback processes from violent crimes such as robbery. These feedback processes are very complex, and no one has successfully estimated their relative magnitudes (Skogan, 1986). To make matters more complicated, some research suggests that crime does not necessarily lead to neighborhood decline; in fact, some neighborhoods prosper despite high crime (Taub et al., 1984; McDonald, 1986). We should thus recognize that simple interpretation of the link between urban social structure and violence is not possible.

MEDIATING COMMUNITY PROCESSES

Even if the above problems were resolved, serious measurement deficiencies prevent macrolevel researchers from doing much more than speculating. For the most part, researchers have simply inferred the existence of intervening community processes, even though the correlation of violence with ecological characteristics is consistent with many different theoretical perspectives. For example, the typical aggregate-level study, as reviewed above, shows us that factors such as percent black, poverty, and family disruption are predictive of violence rates. Although useful as a preliminary test, this strategy does not really go beyond the steps taken by Shaw and McKay more than 40 years ago. As Kornhauser (1978) argues, most criminological theories take as their point of departure the same independent variables-especially stratification factors such as socioeconomic status. However, the variables that intervene between community structure and violence are at issue, and to adequately test competing theories it is necessary to establish the relationship to violence of the interpretive variables each theory implies (cf. Kornhauser, 1978:82).

The lack of direct tests of community-level theories does not stem from a lack of theoretical insight but rather a lack of relevant data. As Heitgerd and Bursik (1987) conclude, traditional ecological studies are not well suited to an examination of the

Suggested Citation:"Violent Victimization and Offending: Individual-Situational-, and Community-Level Risk Factors." National Research Council. 1994. Understanding and Preventing Violence, Volume 3: Social Influences. Washington, DC: The National Academies Press. doi: 10.17226/4421.
×

formal and informal networks hypothesized to link community social structure and crime. Such an examination requires extensive and prohibitive data collection within each of the communities in the analysis (Heitgerd and Bursik, 1987:785). Similarly, Reiss (1986a:26-27) notes that because governments collect very little information on the collective properties of administrative units for which they routinely report information, "little causal information is available for those same units."

The crux of the problem is that previous macrolevel research has relied primarily on census data that rarely provide measures for the variables hypothesized to mediate the relationship between community structure and violence. Ethnographic research (e.g., Suttles, 1968; Hagedorn, 1988) is an exception to this pattern in that it provides rich descriptive accounts of community processes central to theoretical concerns. Yet as Reiss (1986a:27) argues, ethnographies provide limited tests of theories by focusing on a single community or, at most, on a cluster of neighborhoods in which community properties do not display sufficient variation. Also, although some studies have examined quantitative dimensions of informal social control (e.g., Maccoby et al., 1958; Kapsis, 1976; Simcha-Fagan and Schwartz, 1986), they have been limited to a few select communities, which precludes comprehensive multivariate analysis. Consequently, empirical examination of between-community differences in social disorganization and other community-level processes has been rarely undertaken. In point of fact, with the exception of data from Baltimore (Taylor et al., 1984) and Great Britain (Sampson and Groves, 1989), there have been few if any direct tests of the effects of social disorganization and community social control on rates of violence.

Another related problem is that culture is commonly ignored in macrolevel research even though a key element of macrolevel explanation focuses on how community cultures and value systems produce differential rates of crime (cf. Short, 1990:11-12; Luckenbill and Doyle, 1989). "Structural determinants" of crime are emphasized in large part because virtually all census-derived measures may be classified as structural. An abundance of convenient structural measures does not make culture irrelevant, however. In fact, many of the concepts suggested by criminological theory such as "hostility," ''alienation," and "southern values" (e.g., see Blau and Blau, 1982) are related in some way to culture or to its fragmentation in modern urban life. Even when culture is addressed in ecological research, it is haphazard. For example, it is commonplace to use region and percent black as indicators of

Suggested Citation:"Violent Victimization and Offending: Individual-Situational-, and Community-Level Risk Factors." National Research Council. 1994. Understanding and Preventing Violence, Volume 3: Social Influences. Washington, DC: The National Academies Press. doi: 10.17226/4421.
×

"subculture" even though one cannot adequately test the subculture of violence thesis with census data.

A major flaw in prior research on violence, then, is simply that key concepts have rarely if ever been adequately measured. As a consequence the transmission processes through which neighborhood effects operate-whether contagion, socialization, institutional, or social comparison processes-are largely unknown (see also Mayer and Jencks, 1989). Note that the lack of measurement of social interactions and mediating processes is also directly linked to the definition and conceptualization of communities themselves. Essentially, criminological research is dominated by analysis of statistical neighborhoods defined by administrative concerns (e.g., census tracts) that may or may not correspond to social patterns of interaction and cohesion. However, as Tienda (1989:6) has argued, "Social dimensions of the definition of neighborhoods are crucial because they derive from interaction patterns, which ultimately are the primary mechanisms through which neighborhood effects can be transmitted." It is therefore doubly problematic that most efforts to establish the existence of community influences on violence are based on the ecological or structural characteristics of places to the neglect of social interaction and social networks within spatial domains.

CROSS-LEVEL MISSPECIFICATION

Perhaps the major conceptual and methodological problem underlying macrolevel research is the mistaken assumption that the unit of analysis automatically determines the level of causal explanation. We reviewed a host of studies that correlated census data with crime rates for some aggregate units (e.g, communities, cities, SMSAs) and thereby claimed to be macrosocial investigations. The ecological fallacy is irrelevant, it is usually argued, because an interest in individuals is disavowed. The problem with this strategy is that an apparent ecological or structural effect may in fact arise from individual-level causal processes. For example, an observed macrolevel result such as the correlation of median income or percent black with violence rates may simply represent the aggregate of relationships occurring at lower levels of social structure and not a manifestation of processes taking place at the level of the community as a whole. Indeed, the section on individual-level risk factors demonstrated the potential power of compositional confounding with regard to factors such as race and social class. Even though rates of crime may not

Suggested Citation:"Violent Victimization and Offending: Individual-Situational-, and Community-Level Risk Factors." National Research Council. 1994. Understanding and Preventing Violence, Volume 3: Social Influences. Washington, DC: The National Academies Press. doi: 10.17226/4421.
×

be used to make inferences about individuals, individuals commit the crimes that constitute the rates.

Consider further the basic facts on delinquency. Research has consistently demonstrated the early onset of many forms of delinquency and its long-term stability (Robins, 1966; West and Farrington, 1977; Sampson and Laub, 1990). These general differences among individuals that are stable over time have profound implications for an ecological study of crime. For example, longitudinal research suggests that delinquent tendencies are fairly well established at early ages-even at age 8 or so, and certainly by the early teens (Glueck and Glueck, 1950). Antisocial children tend to fight, steal, become truant, drop out of school, drift in and out of unemployment, live in lower-class areas, and go on to commit adult crime. The causal nature of the relationship between achieved adult characteristics (e.g., employment status) and adult crime is thus fraught with methodological difficulties. In fact, in Deviant Children Grown Up, Robins (1966) offered the provocative hypothesis that antisocial behavior predicts class status more efficiently than class status predicts antisocial behavior.

If area differences in rates of violence are the result of the characteristics of individuals selectively located in those communities (Kornhauser, 1978:114), what then do we make of the findings derived from macrolevel criminology? For example, is the relationship between the concentration of poverty and crime rates caused by an aggregation of individual-level effects of social class, or by a genuine community-level effect; or is it simply a differential selection of individuals into communities based on prior (e.g., antisocial) behavior? Perhaps it is simply that common third factors cause individuals both to commit violence and to perform poorly in the occupational sphere. Moreover, if violent and antisocial tendencies are formed at early (preteen) ages, what plausible roles can community labor markets and economic stratification play in understanding violence?

The fallacy of assuming macrostructural effects on the basis of the unit of analysis is pervasive even when an aggregate measure does not have an immediately evident counterpart on the microlevel (e.g., inequality, density). Simply put, macrolevel research is not immune to questions concerning the level at which causal relations operate. The level at which a causal relation occurs is a complex issue that is not solved simply by the nature of the way in which variables are measured or the unit for which they are measured, because psychological and sociological causal factors may underlie relations observed at both the individual and

Suggested Citation:"Violent Victimization and Offending: Individual-Situational-, and Community-Level Risk Factors." National Research Council. 1994. Understanding and Preventing Violence, Volume 3: Social Influences. Washington, DC: The National Academies Press. doi: 10.17226/4421.
×

the aggregate levels. In particular, the concrete actions of individuals feed back to shape the collective environment (Tienda, 1989). Thus the unit of analysis does not necessarily define the level of causal explanation, and the information contained in aggregate data is not necessarily generated by macrosociological processes.

On critical examination, ecological research often seems overly simplistic and unable to tackle hard questions, especially regarding selection effects. As Tienda (1989:23) has argued, "If systematic selection processes are the primary mechanism bringing together individuals with similar socioeconomic characteristics and behavioral dispositions within defined spatial areas, then neighborhood effects may represent little more than an aggregation of the selection process." To demonstrate that community factors accentuate the manifestations and ramifications of individual behavior, the feedback mechanisms between individual behavior and the social context must be modeled (Tienda, 1989:18). The preponderance of cross-sectional community-level studies in criminology greatly compounds the problem because communities change and so do individuals.

Problematic though they may be, however, community-level inferences are not the only issue at stake in cross-level misspecification. On the other side is the "individualistic fallacy"-the often-invoked assumption that individual-level causal relations necessarily generate individual-level correlations. The fact of the matter is that research conducted at the individual level rarely questions whether obtained results might be spurious and confounded with community-level processes. A good example can be seen in the case of race, one of the most important individual-level correlates of violence as shown earlier. Wilson's (1987) thesis of social isolation and concentration effects suggests a possible contextual explanation for the race-violence link among individuals. Consider that although approximately 70 percent of all poor whites lived in nonpoverty areas in the five largest U.S. central cities in 1980, only 15 percent of poor blacks did. Moreover, whereas only 7 percent of poor whites lived in extreme poverty areas, almost 40 percent of poor blacks lived in such areas (Wilson, 1987:58).

The consequences of these differential ecological distributions by race raise the substantively plausible hypothesis that individual-level correlations by race (and also class) may be systematically confounded with important differences in community contexts. As Wilson (1987:58-60) argues,

Simple comparisons between poor whites and poor blacks would be confounded with the fact that poor whites reside in areas

Suggested Citation:"Violent Victimization and Offending: Individual-Situational-, and Community-Level Risk Factors." National Research Council. 1994. Understanding and Preventing Violence, Volume 3: Social Influences. Washington, DC: The National Academies Press. doi: 10.17226/4421.
×

which are ecologically and economically very different from poor blacks. Any observed relationships involving race would reflect, to some unknown degree, the relatively superior ecological niche many poor whites occupy with respect to jobs, marriage opportunities, and exposure to conventional role models.

Regardless of a black's individual-level family or economic situation, the average community of residence differs dramatically from that of a similarly situated white (see also Sampson, 1987a; Stark, 1987). Therefore, the relationship between race and violence may be accounted for largely by community context (e.g., segregation, concentration of family disruption and joblessness, social isolation, sparse social networks). We simply do not know, and cannot know, given the typical individual-level research design.

In short, the confluence of selective aggregation into communities, cross-level misspecification (e.g., compositional effects, individualistic fallacy), multicollinearity and nonexperimental designs, a static conceptualization of community structure, the early onset of many forms of violence, and crude measurement of community characteristics suggest that multilevel research is still in its infancy. Although space limitations preclude detail, we therefore conclude with some positive research strategies that may help to resolve these outstanding issues and provide policy inputs.

FUTURE RESEARCH DIRECTIONS AND POLICY IMPLICATIONS

NEEDED RESEARCH

The major implication of our review is the need for criminological research and policy on violence to integrate more effectively the individual, situational, and community levels of analysis. One general approach to this need is contextual analysis, in which information on communities is combined with individual-level data to explain violence. Contextual analysis permits examination not only of the main effects of community structure on individual behavior but also of the interaction between community and individual characteristics. Recent advances in multilevel hierarchical modeling are especially appealing because they allow the parameterization of the way in which individual-level relationships vary across social contexts such as schools and communities (Raudenbush and Bryck, 1986). These approaches eliminate the tendency in criminological research to ignore either the

Suggested Citation:"Violent Victimization and Offending: Individual-Situational-, and Community-Level Risk Factors." National Research Council. 1994. Understanding and Preventing Violence, Volume 3: Social Influences. Washington, DC: The National Academies Press. doi: 10.17226/4421.
×

micro- or the macrolevel, based on disciplinary biases and data availability. As revealed throughout this paper, most prior research on violence involves study of either individual effects or community-level effects—almost no research has examined both (see also Reiss, 1986a; Gottfredson and Taylor, 1986). Most individual-level research is thus inadequate because it neglects variation in community characteristics, whereas community-level research fails to take account of individual difference constructs. Also, both the individual and the macrolevels have tended to ignore the microsituational context (Short, 1985) of violent events. Accordingly, linking the individual, situational, and community levels through contextual specification appears to be a crucial agenda in future studies of violence.

To meet these goals requires major new initiatives in research design and conceptualization. Perhaps the most pressing issue is the importance of studying neighborhood and individual change as a means to distinguish contextual from selection effects. In particular, understanding mechanisms that produce alternative neighborhood trajectories is essential to avoid confounding selection effects with neighborhood effects in the study of ecologically concentrated violence (cf. Tienda, 1989:18). Controlling for background characteristics is not enough, because contextual analysis per se does not deal with selection processes. As Tienda (1989) argues, multilevel models that simply combine person and place characteristics without regard to the potential endogeneity of neighborhood characteristics are methodologically flawed. Therefore, longitudinal designs that follow not only changes in the structure, composition, and organization of communities but also the individuals who reside there are needed to establish the unique contribution of individuals and communities to crime and delinquency (Reiss, 1986a:29).

A longitudinal, contextual approach to violence research is also consistent with the life-course perspective. Elder (1985:17) has defined the life course as pathways through the age-differentiated life span, where age differentiation "is manifested in expectations and options that impinge on decision processes and the course of events that give shape to life stages, transitions, and turning points." The life-course framework seems especially fruitful in organizing research on individual-level continuity and change over the life span in relation to community context. For example, how do early childhood characteristics (e.g., antisocial behavior) lead to adult behavioral outcomes and sorting processes related to area of residence? How do life transitions (e.g., unemployment,

Suggested Citation:"Violent Victimization and Offending: Individual-Situational-, and Community-Level Risk Factors." National Research Council. 1994. Understanding and Preventing Violence, Volume 3: Social Influences. Washington, DC: The National Academies Press. doi: 10.17226/4421.
×

divorce) and situational-level factors (e.g., victim-offender relationship, presence of third parties) interact with community context? Does effective family social control and transmission of nonviolent values depend on levels of community social disorganization? Do the effects of community differ across stages of the life cycle? How do violent offending and victimization interact over the life cycle? These and many other questions are ideally suited to a concern with multilevel influences on the trajectories of individual and community change.

Even if we are interested only in macrolevel processes, the longitudinal study of communities is essential for making causal inferences regarding the effects of communities on violence rates. As noted earlier, there are substantive reasons to expect that violence has feedback effects on major dimensions of community structure. Moreover, communities appear to have "careers" in crime (Reiss, 1986a; Schuerman and Kobrin, 1986), and several key hypotheses concerning the link between communities and rates of violence were explicitly formulated in terms of change (e.g., increasing concentration of the underclass, effect of residential turnover on disruption of social networks). To disentangle community social processes and the potential reciprocal effects of violence requires longitudinal designs in which temporal ordering can be established and change processes explicitly modeled.29

Another major research need is for direct examination of the social processes and interactions through which both individual and community effects are transmitted. As for the individual level, research on both victimization and offending has established little more than that sociodemographic factors (e.g., age, race, sex, employment) are important correlates of violence. This is unsatisfactory, not just to policy makers, but to social science theories of violence causation. Although beyond the scope of this paper, we need much more concentrated efforts to measure the mediating social processes underlying victimization (e.g., routine activities, lifestyles) and offending (e.g., peer networks, cultural values, school climate). In particular, research directly addressing the social basis for the race-violence link in contemporary urban America seems especially important. Street (1989:29) has argued that without the willingness to confront and analyze the reality of race, "we cannot make headway in dealing with crime in this country." To the extent that past trends in research designs simply continue, we will indeed not make headway in uncovering the social processes that foster urban violence-the leading cause of death among young black males (Fingerhut and Kleinman, 1990:3292).

Suggested Citation:"Violent Victimization and Offending: Individual-Situational-, and Community-Level Risk Factors." National Research Council. 1994. Understanding and Preventing Violence, Volume 3: Social Influences. Washington, DC: The National Academies Press. doi: 10.17226/4421.
×

Community-level measurement needs are probably even greater than those at the individual level. We documented earlier how criminological research has largely failed to directly examine community processes of theoretical relevance. To address this requires a new, though perhaps costly, measurement strategy: survey instruments designed to measure community concepts and administered with community sampling frames in mind. For example, survey measures could be constructed to capture community-level patterns of informal social control, friendship networks, subcultural values, rates of organizational participation, and much more. Some preliminary attempts have been made (e.g., Taylor et al., 1984; Simcha-Fagan and Schwartz, 1986; Sampson and Groves, 1989), but we need surveys designed to test community-based theories on units of analysis defined by the research question, not by governmental agencies. There is nothing inherently deficient about a measurement strategy designed at the macrolevel. The problem is that the measurement task is difficult and requires the outlay of large research funds.

A different approach is ethnographic analysis, because we can learn much about community structure and violence from ethnographic studies that we cannot learn from multivariate analyses of census data and survey measures. Given the importance of community cultures and value systems in criminological theories of violence, it is disturbing that modern criminology has reified multivariate methodology as the engine of scientific progress. We believe that a more balanced perspective calls for quantitative research to be supplemented with community ethnography. Hagedorn's (1988) People and Folks: Gangs, Crime and the Underclass in a Rustbelt City and Sullivan's (1989) Getting Paid: Economy, Culture, and Youth Crime in the Inner City are but two recent examples of how rich and provocative community ethnographic/qualitative research can be (see also Katz, 1988).

PUBLIC POLICY

Last, but certainly not least, a multilevel approach guided by a substantive focus on community-level processes has constructive implications for public policy. The reason is that many of the hypothesized community-level sources of violence noted throughout this paper (e.g., residential instability, concentration of poverty and family disruption, high-density public housing projects, attenuation of social networks) are determined, both directly and indirectly, by the policy decisions of public officials. Take, for

Suggested Citation:"Violent Victimization and Offending: Individual-Situational-, and Community-Level Risk Factors." National Research Council. 1994. Understanding and Preventing Violence, Volume 3: Social Influences. Washington, DC: The National Academies Press. doi: 10.17226/4421.
×

example, municipal code enforcement and local governmental policies toward neighborhood deterioration. In Making the Second Ghetto: Race and Housing in Chicago, 1940-1960, Hirsch (1983) documents in great detail how lax enforcement of city housing codes played a major role in accelerating the deterioration of inner-city Chicago neighborhoods. More recently, Daley and Meislen (1988) have argued that inadequate city policies with regard to code enforcement and repair of city properties contributed to the systematic decline of New York City's housing stock and, consequently, of entire neighborhoods. When considered in conjunction with the practices of redlining and disinvestment by banks or "blockbusting" by real estate agents (Skogan, 1986), local policies toward code enforcement-which on the surface seem far removed from crime-have in all likelihood contributed to crime through neighborhood deterioration, forced migration, and instability.

The provision of city municipal services in terms of public health and fire safety-decisions presumably made with little if any thought to crime and violence-also appears to have been salient in the social disintegration of poor communities. As Wallace and Wallace (1990:427) argue, based on an analysis of the "planned shrinkage" of New York City fire and health services in recent decades, "The consequences of withdrawing municipal services from poor neighborhoods, the resulting outbreaks of contagious urban decay and forced migration which shred essential social networks and cause social disintegration, have become a highly significant contributor to decline in public health among the poor." The loss of social integration and networks from planned shrinkage of services may increase behavioral patterns of violence that can themselves become "convoluted with processes of urban decay likely to further disrupt social networks and cause further social disintegration" (Wallace and Wallace, 1990:427). This pattern of destabilizing feedback noted earlier and also by Skogan (1986) appears central to understanding the role of governmental policies in fostering the downward spiral of high-crime areas.

Decisions by government on public housing paint a similar picture. Bursik (1989) has shown that the planned construction of new public housing projects in Chicago in the 1970s was associated with increased rates of population turnover, which in turn were related to increases in crime-independent of racial composition. More generally, Skogan (1986:206) notes how urban renewal and forced migration contributed to the wholesale uprooting of many urban communities, especially the extent to which freeway networks driven through the hearts of many cities in the

Suggested Citation:"Violent Victimization and Offending: Individual-Situational-, and Community-Level Risk Factors." National Research Council. 1994. Understanding and Preventing Violence, Volume 3: Social Influences. Washington, DC: The National Academies Press. doi: 10.17226/4421.
×

1950s destroyed viable, low-income communities. Indeed, the instability forced upon low-income minority neighborhoods by governmental urban renewal decisions was profound. In Atlanta, one in six residents was dislocated through urban renewal, the great majority of whom were poor (Logan and Molotch, 1987:114). Nationwide, fully 20 percent of all central-city housing units occupied by blacks were lost in the period 1960-1970. As Logan and Molotch (1987:114) observe, this displacement does not even include that brought about by more routine market forces (evictions, rent increases, commercial development).

Perhaps most disturbing, Wilson (1987:20-62) documents the negative consequences of policy decisions to concentrate minorities and the poor in public housing. Opposition from organized community groups to the building of public housing in their neighborhoods, de facto federal policy to tolerate extensive segregation of blacks in urban housing markets, and the decision by local governments to neglect the rehabilitation of existing residential units (many of them single-family homes) have led to massive, segregated housing projects that have become ghettos for the minorities and disadvantaged (see also Massey, 1990; Sampson, 1990). The social transformation of the inner city, triggered in large part by governmental policy, has thus led to the disproportionate concentration of the most disadvantaged segments of the urban black population in a few areas. The result is that even given the same objective socioeconomic status, blacks and whites face vastly different environments in which to live, work, and raise their children (Massey and Kanaiaupuni, 1990:18). When linked to the probable effects of concentrated urban poverty and family disruption on community social organization, as reviewed above, governmental housing policies clearly become relevant to crime control policy. Moreover, to the extent that segregation and concentrated poverty represent structural constraints embodied in public policy, concerns raised earlier about self-selection confounding the interpretation of community-level effects on violence are considerably diminished. As Massey and Kanaiaupuni (1990:20) have argued, public housing represents a federally funded, physically permanent institution for the isolation of black families by race and class, and must therefore be considered an important structural constraint on ecological area of residence.

In short, government matters, and the political economy of place (Logan and Molotch, 1987) must be explicitly incorporated into theories and policies of community-level social organization. The reason is that what seem to be "noncrime" policies (e.g.,

Suggested Citation:"Violent Victimization and Offending: Individual-Situational-, and Community-Level Risk Factors." National Research Council. 1994. Understanding and Preventing Violence, Volume 3: Social Influences. Washington, DC: The National Academies Press. doi: 10.17226/4421.
×

where or if to build a housing project, enforcement of municipal codes, reduction in essential municipal services, rehabilitation of existing residential units, disbursement of the disadvantaged) can have important indirect effects on violence. As detailed above, many community characteristics implicated in violence, such as residential instability, concentration of poor, female-headed families with children, multiunit housing projects, and disrupted social networks, appear to stem rather directly from planned governmental policies at local, state, and federal levels. This conceptualization diverges from the natural market assumptions of the earlier social ecologists (e.g., Park et al., 1925) by considering the role of political decisions in shaping community structure.

On the positive side, the implication of this community-level perspective is that there are in fact policy options that may help reverse the tide of community social disintegration. Among others, these might include resident management of public housing (to increase stability), tenant buy-outs (to increase home ownership and commitment to locale), rehabilitation of existing low-income housing (to preserve area stability, especially single-family homes), disbursement of public housing (versus concentration), and strict code enforcement (to fight deterioration). Moreover, there are recent examples that such policies are viable and, in fact, have stabilizing effects on communities and hence potential for crime reduction (e.g., see the programs described in Reiss, 1989; Sampson, 1990; more generally, see Logan and Molotch, 1987).

CONCLUSION

For these and other reasons discussed throughout this paper, the prospects appear promising for merging a community-level approach to violence with more traditional concerns regarding individual and situational factors. The unique value of a multilevel perspective is that it leads away from a simple ''kinds of people" analysis to a focus on how social characteristics of collectivities foster violence. Based on the theoretical conceptualization and empirical literature reviewed earlier, we thus conclude that community-level factors such as (but not limited to)30 the ecological concentration of the urban underclass, residential mobility and population turnover, family disruption, housing and population density, criminal opportunity structures (e.g., levels of nonhousehold leisure activities, gun density), and dimensions of local social organization

Suggested Citation:"Violent Victimization and Offending: Individual-Situational-, and Community-Level Risk Factors." National Research Council. 1994. Understanding and Preventing Violence, Volume 3: Social Influences. Washington, DC: The National Academies Press. doi: 10.17226/4421.
×

(e.g., informal social ties, density of acquaintanceship, supervision of street-corner peer groups, organizational density and strength) represent fruitful areas of future inquiry, especially because they are affected by public policies regarding housing and municipal services. Moreover, our review suggests that individual-level characteristics (e.g., age, race, sex, poverty, divorce) and situational context (e.g., victim-offender relationship) in all likelihood interact with these community characteristics (see also Gottfredson and Taylor, 1986) to explain both victimization by violence and criminal careers in violent offending. Implementation of multilevel, longitudinal research designs that capture the inextricable link between violent victimization and violent offending will not be cheap but appears essential nonetheless.

ACKNOWLEDGMENT

We are grateful to Steve Gottfredson, Lloyd Street, Leonard Eron, and members of the Panel on the Understanding and Control of Violent Behavior for very helpful comments on an earlier version of this paper.

NOTES

1  

Because of space limitations we are also unable to consider in any detail the role of social learning factors in the etiology of violence. For a theoretical overview of social learning theory and behaviorism, especially with regard to aggression, see Bandura and Walters (1959) and Eron (1987).

2  

Data presented in reports of the National Crime Survey are usually reported as rates per 1,000 (or 10,000) persons in the population at risk. To provide a comparative indicator of an individual's risk of violent victimization, we have converted published rates into a "1 in X" chance of victimization (see Reiss, 1981).

3  

The NCS household surveys discussed herein were not specifically designed to capture victimizations against commercial establishments (e.g., robberies of convenience stores and gas stations). This distinction suggests caution in the comparison of NCS robbery victimizations with UCR arrest data on robberies that include commercial targets.

4  

Conclusions from NCS data do not apply to homicides. Recent evidence suggests that the homicide rate for juveniles is rising faster than for adults. In fact, some cities report as much

Suggested Citation:"Violent Victimization and Offending: Individual-Situational-, and Community-Level Risk Factors." National Research Council. 1994. Understanding and Preventing Violence, Volume 3: Social Influences. Washington, DC: The National Academies Press. doi: 10.17226/4421.
×

   

as a 100 percent increase in violence among juveniles in the period 1985-1990 alone (Recktenwald and Myers, 1991). A recent report by the Centers for Disease Control also shows that the firearm death rate among U.S. black males more than doubled from 1984 to 1988 (Fingerhut et al., 1991).

5  

Our use of the term "sex," as opposed to "gender," is merely descriptive and not intended to suggest that male-female differences in offense involvement or victimization are the simple result of biological factors. For a detailed review of this literature, see Kruttschnitt in this volume.

6  

Of course, women are much more likely to be the victims of "purse snatching" than men. However, purse snatching without the use of force is personal larceny, not robbery or assault (cf. Feeney and Weir, 1975).

7  

Age, of course, is part of the reason for the magnitude of this relationship. However, see section below on multivariate relationships.

8  

Again, it is likely that part of this association is due to the relationship between age and victimization risk.

9  

Victims of violent crime were defined as those persons who reported an assault, robbery, or personal larceny with contact.

10  

Gender was not considered in this analysis because of a limited number of reported robberies against females.

11  

Gender was also not considered in this analysis, apparently for theoretical reasons.

12  

Education was not evaluated as an independent predictor in any of the above-mentioned analyses. However, given the weak bivariate relationship found in the NCS it is unlikely that this measure would have contributed independently and significantly to the risk of violent victimization.

13  

Space limitations preclude a detailed review of the lifestyle and routine activity approach. For comprehensive reviews the reader is referred to the original sources (Hindelang et al., 1978:Chap. 11; Cohen and Felson, 1979; Cohen et al., 1981b) and to Gottfredson (1986), Garofalo (1987), and Maxfield (1987). In this paper we assume theoretical compatibility between the lifestyle and routine activity approaches (see especially Maxfield, 1987; Garofalo, 1987).

14  

Lifestyle-routine activity theories tend to assume that one's home is safer than areas outside the home and that the risk of violence from family members is lower than the risk of violence from nonfamily members. This assumption has recently come

Suggested Citation:"Violent Victimization and Offending: Individual-Situational-, and Community-Level Risk Factors." National Research Council. 1994. Understanding and Preventing Violence, Volume 3: Social Influences. Washington, DC: The National Academies Press. doi: 10.17226/4421.
×

   

into question by research showing that for certain subgroups (e.g., women, children), this may not be the case (see generally Ohlin and Tonry, 1989).

15  

The dichotomous measure of major daytime activity compared those employed and persons attending school to all others. Night activity was dichotomized to reflect persons who go out (on average) less than once per week versus those who go out one or more times a week.

16  

Most studies (see Wolfgang et al., 1987, for an exception) have analyzed either juvenile or adult samples. Different findings with respect to specialization and escalation of offending might be found if we followed the same individuals over both the juvenile and the adult phases of offending careers.

17  

These methodological problems are not discussed further because they have been addressed elsewhere (for recent overviews, see Blumstein et al., 1986; Weiner, 1989; Bridges and Weis, 1989). Nonetheless, they should not be overlooked; overcoming even a few of these limitations could substantially increase our understanding of violent offending.

18  

The conceptual confusion over situational factors has been discussed at length by Monahan and Klassen (1982). An empirical example of the problem can be seen in Denno's (1986) analysis of victim, offender, and situational characteristics of violent crime. At one point, she defines situational factors to include family income and the sociodemographic characteristics of the participants (1986:1143). At another, she considers "offender" variables to be individual-level factors "measured independently of the offense event"-a definition that would seem to subsume the demographic characteristics (e.g., age, race, sex) of the participants earlier attributed to a situation. Like Monahan and Klassen (1982) we do not consider attributes of individuals (e.g., race, sex, income) to be situational factors.

19  

Fagan et al. (1987) have argued that the individual-level relationships between victimization and offending may be spurious and due to the areas in which both victims and offenders live.

20  

Interpretations of "nonstranger" (including acquaintances, friends, and family members) versus "stranger" differences are also complicated by the fact that the nonstranger category contains a wide variety of social relationships. For example, the definition of an acquaintance may include asymmetrical relationships in which the offender is familiar with the victim but not vice versa or is someone familiar to the victim by sight only.

21  

For more extensive reviews of this literature, see Collins

Suggested Citation:"Violent Victimization and Offending: Individual-Situational-, and Community-Level Risk Factors." National Research Council. 1994. Understanding and Preventing Violence, Volume 3: Social Influences. Washington, DC: The National Academies Press. doi: 10.17226/4421.
×

   

(1989), Goldstein (1989), and especially the articles in Tonry and Wilson (1990).

22  

The effect of third parties on the initiation of a violent event would be extremely difficult to determine empirically, given that the presence of additional persons is likely to be correlated with other contextual factors not included in the model.

23  

Although not typically thought of as a property of local communities, urbanization is strongly related to violent crime, independent of compositional factors (Laub, 1983; Sampson and Groves, 1989). See Fischer (1975) and Laub (1983) for theoretical discussions of the link between urbanization and crime. Unfortunately, very few studies have examined between-community variations in violent crime in nonurbanized areas.

24  

Of course, criminology is rich in cultural explanations of crime, especially regarding lower-class culture and the subculture of violence (for a review see Kornhauser, 1978; Short, 1990). In terms of violence, the most prominent explanation continues to be the subculture of violence theory (Wolfgang and Ferracuti, 1967). The basic tenet of subcultural theory is that "overt use of force or violence ... is generally viewed as a reflection of basic values that stand apart from the dominant, central, or parent culture" (1967:385). However, neither the theory of lower-class culture nor the subculture of violence refers explicitly to community-level causal processes.

25  

We focus on the Land et al. (1990) study because it brings together in one paper the key metropolitan-level studies of violence conducted in recent years. Indeed, all of the studies in Table 1 have appeared since 1960 and have formed the basis for much recent discussion on the macrolevel sources of criminal violence (see also the review by Byrne and Sampson, 1986; Messner and Sampson, 1991). Space limitations preclude a discussion of each study-readers are urged to consult the originals for more details. For an excellent review of earlier research on metropolitan differences in violence, see Brantingham and Brantingham (1984:281-294). The latter also provide a discussion of regional-level studies of violence and metropolitan-level studies focused on crimes other than homicide.

26  

In a recent paper appearing after the Land et al. (1990) study, Balkwell (1990) shows that ethnic income inequality had a positive net effect on homicide rates across 150 SMSAs in 1980 even after controlling for poverty, race, and geographic region.

27  

Very few studies have examined macrolevel variations in violent crime among rural areas and suburbs. However, evidence

Suggested Citation:"Violent Victimization and Offending: Individual-Situational-, and Community-Level Risk Factors." National Research Council. 1994. Understanding and Preventing Violence, Volume 3: Social Influences. Washington, DC: The National Academies Press. doi: 10.17226/4421.
×

   

does exist that the correlates of suburban violence are quite similar to those in urban areas. For example, Stahura and Sloan (1988) found that the major predictor of violent crime rates across U.S. suburbs was a factor almost identical to the resource deprivation/affluence factor identified by Land et al. (1990). Specifically, the factor most related to violent crime in the suburbs was comprised mainly of percent poor and percent black (Stahura and Sloan, 1988:1111).

28  

In further analyses of these data, Messner and Sampson (1991) disaggregated the measure of employed black males per 100 black females into its constituent components-the sex ratio (males per female) and the unemployment rate. Results showed that both components had significant negative effects on black family disruption, which again had positive effects on black violence rates. Interestingly, however, even though the sex ratio had the largest effect on family disruption-and by implication a negative indirect effect on violence-the sex ratio had a positive direct effect on violence. Therefore, family disruption appears to operate as a suppressor variable in explaining the counterbalancing effects of sex ratio on violence. That is, only when family disruption is controlled does the expected positive effect of sex ratio composition (males per female) on violence emerge. Furthermore, the indirect effect of the relative availability of employed black males on violence (Sampson, 1987a) appears to stem mostly from the simple demographic factor of imbalanced sex ratios.

29  

"Case-control" methodology (see Schlesselmann, 1982; Loftin and McDowall, 1988) may be an especially attractive device for studying both individual-and community-level violence in longitudinal perspective. Although case-control methodology is usually retrospective, communities and individuals could be matched on key confounding characteristics (e.g., demographic background) and then followed up and examined in terms of salient intervening processes (e.g., subcultural values, disruption of social networks) thought to account for variation in violence. Not only is case-control methodology useful for controlling key theoretical factors through matching, but it provides an efficient and cost-effective way to counteract the low base rate of violence (Loftin and McDowall, 1988).

30  

Because of space limitations, we did not focus on community-level dimensions of formal social control by the police and courts. These and other community-level variations (e.g., drug distribution networks, medical and trauma services) should not be overlooked in future work. Moreover, we did not examine the structural sources of violence by agents of official social control

Suggested Citation:"Violent Victimization and Offending: Individual-Situational-, and Community-Level Risk Factors." National Research Council. 1994. Understanding and Preventing Violence, Volume 3: Social Influences. Washington, DC: The National Academies Press. doi: 10.17226/4421.
×

   

(e.g., police homicide, brutality, coercion). See Smith (1986) and Jacobs and Britt (1979), respectively, for empirical analyses of the community-level determinants of police use of coercion and deadly force.

REFERENCES

Ageton, S. 1983 The dynamics of female delinquency, 1976-1980. Criminology 21:555-584.

Anderson, E. 1978 A Place on the Corner. Chicago: University of Chicago Press.

1990 Streetwise: Race, Class, and Change in an Urban Community. Chicago: University of Chicago Press.

Andrew, J. 1979 Violence and poor reading. Criminology 17:361-365.


Bailey, W.C. 1984 Poverty, inequality, and city homicide rates: Some not so unexpected findings. Criminology 22:531-550.

Balkwell, J. 1990 Ethnic inequality and the rate of homicide. Social Forces 69:53-70.

Bandura, A., and R. Walters 1959 Adolescent Aggression. New York: Ronald Press.

Baron, L., and M. Straus 1987 Four theories of rape: A macrosociological analysis. Social Problems 34:467-489.

Beasley, R.W., and G. Antunes 1974 The etiology of urban crime: An ecological analysis. Criminology 11:439-461.

Bensing, R.C., and O. Schroeder 1960 Homicide in an Urban Community. Springfield, Ill.: Charles C. Thomas.

Blau, J., and P.M. Blau 1982 The cost of inequality: Metropolitan structure and violent crime. American Sociological Review 47:114-129.

Blau, P.M., and R. Golden 1986 Metropolitan structure and criminal violence. Sociological Quarterly 27:15-26.

Block, C.R. 1985 Race/ethnicity and patterns of Chicago homicide, 1965 to 1981. Crime and Delinquency 31:104-116.

Block, R. 1975 Homicide in Chicago: A nine year study (1965-1973). Journal of Criminal Law and Criminology 66:496-510.

Suggested Citation:"Violent Victimization and Offending: Individual-Situational-, and Community-Level Risk Factors." National Research Council. 1994. Understanding and Preventing Violence, Volume 3: Social Influences. Washington, DC: The National Academies Press. doi: 10.17226/4421.
×

1979 Community, environment, and violent crime. Criminology 17:46-57.

1981 Victim-offender dynamics in violent crime. Journal of Criminal Law and Criminology 72:743-761.

Block, R., and C. Block 1980 Decisions and data: The transformation of robbery incidents into official robbery statistics. Journal of Criminal Law and Criminology 71:622-636.

Block, R., and W. Skogan 1986 Resistance and nonfatal outcomes in stranger-to-stranger predatory crime. Violence and Victims 1:241-253.

Blumstein, A., J. Cohen, J. Roth, and C. Visher, eds. 1986 Criminal Careers and "Career Criminals." Washington, D.C.: National Academy Press.

Blumstein, A., J. Cohen, S. Das, and S. Moitra 1988 Specialization and seriousness during adult criminal careers. Journal of Quantitative Criminology 4:303-346.

Bordua, D. 1958 Juvenile delinquency and "anomie": An attempt at replication. Social Problems 6:230-238.

1986 Firearms ownership and violent crime. Pp. 156-188 in J. Byrne and R.J. Sampson, eds., The Social Ecology of Crime. New York: Springer-Verlag.

Braithwaite, J., and D. Biles 1984 Victims and offenders: The Australian experience. Pp. 3-10 in R. Block, ed., Victimization and Fear of Crime: World Perspectives . Washington, D.C.: U.S. Government Printing Office.

Brantingham, P., and P. Brantingham 1984 Patterns in Crime. New York: Macmillan.

Bridges, G., and J. Weis 1989 Measuring violent behavior: Effects of study design on reported correlates of violence. Pp. 14-34 in N. Weiner and M. Wolfgang, eds., Violent Crime, Violent Criminals. Beverly Hills, Calif.: Sage Publications.

Brownfield, D. 1986 Social class and violent behavior. Criminology 24:421-438.

Bullock, H.A. 1955 Urban homicide in theory and fact. Journal of Criminal Law, Criminology, and Police Science 45:565-575.

Bursik, R.J., Jr. 1986 Delinquency rates as sources of ecological change. Pp. 63-76 in J. Byrne and R.J. Sampson, eds., The Social Ecology of Crime. New York: Springer-Verlag.

1988 Social disorganization and theories of crime and delinquency: Problems and prospects. Criminology 26:519-552.

1989 Political decision-making and ecological models of delinquency: Conflict and consensus. Pp. 105-117 in S. Messner, M. Krohn,

Suggested Citation:"Violent Victimization and Offending: Individual-Situational-, and Community-Level Risk Factors." National Research Council. 1994. Understanding and Preventing Violence, Volume 3: Social Influences. Washington, DC: The National Academies Press. doi: 10.17226/4421.
×

and A. Liska, eds., Theoretical Integration in the Study of Deviance and Crime. Albany: State University of New York at Albany Press.

Bursik, R.J., Jr., and J. Webb 1982 Community change and patterns of delinquency. American Journal of Sociology 88:24-42.

Byrne, J. 1986 Cities, citizens, and crime: The ecological/nonecological debate reconsidered. Pp. 77-101 in J. Byrne and R.J. Sampson, eds., The Social Ecology of Crime. New York: Springer-Verlag.

Byrne, J., and R.J. Sampson 1986 Key issues in the social ecology of crime. Pp. 1-22 in J. Byrne and R.J. Sampson, eds., The Social Ecology of Crime. New York: Springer-Verlag.

Carroll, L., and P. Jackson 1983 Inequality, opportunity, and crime rates in central cities. Criminology 21:178-194.

Chilton, R. 1964 Continuity in delinquency area research: A comparison of studies for Baltimore, Detroit, and Indianapolis. American Sociological Review 29:71-83.

Cohen, L., and M. Felson 1979 Social change and crime rate trends: A routine activities approach. American Sociological Review 44:588-607.

Cohen, L., D. Cantor, and J. Kluegel 1981a Robbery victimization in the United States. Social Science Quarterly 66:644-657.

Cohen, L., J. Kluegel, and K. Land 1981b Social inequality and predatory criminal victimization: An exposition and test of a formal theory. American Sociological Review 46:505-524.

Collins, J. 1989 Alcohol and interpersonal violence: Less than meets the eye. Pp. 49-67 in N. Weiner and M. Wolfgang, eds., Pathways to Criminal Violence. Newbury Park, Calif.: Sage Publications.

Cook, P. 1979 The effect of gun availability on robbery and robbery murder: A cross-section study of fifty cities. Pp. 743-781 in R. Haveman and B. Zellner, eds., Policy Studies Review Annual, Vol. 3. Beverly Hills, Calif.: Sage Publications.

1982 The role of firearms in violent crimes. Pp. 236-291 in M. Wolfgang and N. Weiner, eds., Criminal Violence. Beverly Hills, Calif.: Sage Publications.

1983 Robbery in the United States: An Analysis of Recent Trends and Patterns. Washington, D.C.: National Institute of Justice.

1986 The demand and supply of criminal opportunities. Pp. 1-28 in

Suggested Citation:"Violent Victimization and Offending: Individual-Situational-, and Community-Level Risk Factors." National Research Council. 1994. Understanding and Preventing Violence, Volume 3: Social Influences. Washington, DC: The National Academies Press. doi: 10.17226/4421.
×

M. Tonry and N. Morris, eds., Crime and Justice: An Annual Review, Vol. 8. Chicago: University of Chicago Press.

1987 Robbery violence. Journal of Criminal Law and Criminology 78:357-376.

Corrado, R., R. Roesch, W. Glackman, J. Evans, and G. Leger 1980 Lifestyles and personal victimization: A test of the model with Canadian survey data. Journal of Crime and Justice 3:129-139.

Covington, J., and R. Taylor 1989 Gentrification and crime: Robbery and larceny changes in appreciating Baltimore neighborhoods during the 1970s. Urban Affairs Quarterly 25:142-172.

Crutchfield, R., M. Geerken, and W. Gove 1982 Crime rate and social integration: The impact of metropolitan mobility. Criminology 20:467-478.

Curry G.D., and I. Spergel 1988 Gang homicide, delinquency, and community. Criminology 26:381-406.

Curtis, L. 1975 Violence, Race, and Culture. Lexington, Mass.: Heath.

Daley, S., and R. Meislin 1988 New York City, the landlord: A decade of housing decay. New York Times February 8.

Danziger, S. 1976 Explaining urban crime rates. Criminology 14:291-295.

De Fronzo, J. 1983 Economic assistance to impoverished Americans. Criminology 21:119-136.

Denno, D. 1986 Victim, offender, and situational characteristics of violent crime. Journal of Criminal Law and Criminology 77:1142-1158.


Elder, G. 1985 Perspectives on the life course. Pp. 23-49 in G. Elder, ed., Life Course Dynamics. Ithaca, N.Y.: Cornell University Press.

Elliott, D., and S. Ageton 1980 Reconciling race and class differences in self-reported and official estimates of delinquency. American Sociological Review 45:95-110.

Elliott, D., and D. Huizinga 1983 Social class and delinquent behavior in a national youth panel: 1976-1980. Criminology 21:149-177.

Elliott, D., D. Huizinga, and S. Ageton 1985 Explaining Delinquency and Drug Use. Beverly Hills, Calif.: Sage Publications.

Eron, L. 1987 The development of aggressive behavior from the perspective of a developing behaviorism. American Psychologist 42:435-442.

Suggested Citation:"Violent Victimization and Offending: Individual-Situational-, and Community-Level Risk Factors." National Research Council. 1994. Understanding and Preventing Violence, Volume 3: Social Influences. Washington, DC: The National Academies Press. doi: 10.17226/4421.
×

Fagan, J., E. Piper, and Y. Cheng 1987 Contributions of victimization to delinquency in inner cities. Journal of Criminal Law and Criminology 78:586-613.

Farley, R. 1980 Homicide trends in the United States. Demography 17:177-188.

Farrington, D. 1978 The family background of aggressive youths. Pp. 73-92 in L. Hersov, M. Berger, and D. Schaffer, eds., Aggression and Antisocial Behavior in Childhood and Adolescence. Oxford: Pergamon.

1986 Age and crime. Pp. 189-250 in M. Tonry and N. Morris, eds., Crime and Justice: An Annual Review of Research, Vol 7. Chicago: University of Chicago Press.

Farrington, D., and D. West 1971 A comparison between early delinquents and youth aggressiveness. British Journal of Criminology 11:341-358.

Farrington, D., H. Snyder, and T. Finnegan 1988 Specialization in juvenile court careers. Criminology 26:461-488.

Feeney, F., and A. Weir 1975 The prevention and control of robbery. Criminology 20:102-105.

Felson, M., and L. Cohen 1980 Human ecology and crime: A routine activity approach. Human Ecology 8:389-406.

Felson, R., and H. Steadman 1983 Situational factors in disputes leading to criminal violence. Criminology 21:59-74.

Fingerhut, L., and J. Kleinman 1990 International and interstate comparisons of homicide among young males. Journal of the American Medical Association 263:3292-3295.

Fingerhut, L., J. Kleinman, E. Godfrey, and H. Rosenberg 1991 Firearm mortality among children, youth, and young adults 1-34 years of age, trends and current status: United States, 1978-88. Monthly Vital Statistics Report 39(11):1-16.

Fischer, C. 1975 Toward a subcultural theory of urbanism. American Journal of Sociology 80:1319-141.

Fisher, J. 1976 Homicide in Detroit. Criminology 14:387-400.

Flango, E., and E. Sherbenou 1976 Poverty, urbanization, and crime. Criminology 14:331-346.

Friedman, J., and D. Rosenbaum 1988 Social control theory: The salience of components by age, gender, and type of crime. Journal of Quantitative Criminology 4:363-382.

Suggested Citation:"Violent Victimization and Offending: Individual-Situational-, and Community-Level Risk Factors." National Research Council. 1994. Understanding and Preventing Violence, Volume 3: Social Influences. Washington, DC: The National Academies Press. doi: 10.17226/4421.
×

Garofalo, J. 1987 Reassessing the lifestyle model of criminal victimization. Pp. 23-42 in M. Gottfredson and T. Hirschi, eds., Positive Criminology . Newbury Park, Calif.: Sage Publications.

Garofalo, J., L. Siegel, and J. Laub 1987 School-related victimizations among adolescents: An analysis of National Crime Survey narratives. Journal of Quantitative Criminology 3:321-338.

Gastil, R. 1971 Homicide and a regional culture of violence. American Sociological Review 36:412-27.

Gelles, R. 1985 Family violence. Annual Review of Sociology 11:347-367.

Gelles, R., and M. Straus 1988 Intimate Violence. New York: Simon and Schuster.

Glueck, S., and E. Glueck 1950 Unraveling Juvenile Delinquency. New York: Commonwealth Fund.

Goldstein, P. 1989 Drugs and violent crime. Pp. 16-48 in N. Weiner and M. Wolfgang, eds., Pathways to Criminal Violence. Newbury Park, Calif.: Sage Publications.

Gordon, R. 1967 Issues in the ecological study of delinquency. American Sociological Review 32:927-944.

Gottfredson, M. 1984 Victims of Crime: The Dimensions of Risk. Home Office Research Study No. 81. London: Her Majesty's Stationery Office.

1986 Substantive contributions of victimization surveys. In M. Tonry and N. Morris, eds., Crime and Justice: An Annual Review of Research , Vol. 7. Chicago: University of Chicago Press.

Gottfredson, M., and M. Hindelang 1981 Sociological aspects of criminal victimization. Annual Review of Sociology 7:107-128.

Gottfredson, S., and R. Taylor 1986 Person-environment interactions in the prediction of recidivism. Pp. 133-155 in J. Byrne and R.J. Sampson, eds., The Social Ecology of Crime. New York: Springer-Verlag.

Gove, W., M. Hughes, and M. Geerken 1985 Are Uniform Crime Reports a valid indicator of the index crimes? An affirmative answer with minor qualifications. Criminology 23:451-502.

Greenberg, D. 1985 Age, crime, and social explanation. American Journal of Sociology 91:1-21.

Suggested Citation:"Violent Victimization and Offending: Individual-Situational-, and Community-Level Risk Factors." National Research Council. 1994. Understanding and Preventing Violence, Volume 3: Social Influences. Washington, DC: The National Academies Press. doi: 10.17226/4421.
×

Hagan, J., A.R. Gillis, and J. Chan 1978 Explaining official delinquency: A spatial study of class, conflict, and control. Sociological Quarterly 19:386-398.

Hagedorn, J. 1988 People and Folks: Gangs, Crime and the Underclass in a Rustbelt City. Chicago: Lake View Press.

Hanushek, E., and J. Jackson 1977 Statistical Methods for Social Scientists. New York: Academic Press.

Harries, K. 1976 Cities and crime: A geographic model. Criminology 14:369-386.

Harris, A.R. 1977 Sex and theories of deviance: Toward a functional theory of deviant type-scripts. American Sociological Review 42:3-16.

Hawkins, D. 1985 Black homicide: The adequacy of existing research for devising prevention strategies. Crime and Delinquency 31:83-103.

Heitgerd, J.L., and R.J. Bursik, Jr. 1987 Extracommunity dynamics and the ecology of delinquency. American Journal of Sociology 92:775-787.

Hill, G., and A. Harris 1981 Changes in the gender patterning of crime, 1953-1977: Opportunity vs. identity. Social Science Quarterly 62:658-671.

Hindelang, M. 1974 The Uniform Crime Reports revisited. Journal of Criminal Justice 2:1-17.

1976 Criminal Victimization in Eight American Cities. Cambridge, Mass.: Ballinger.

1978 Race and involvement in common-law personal crimes. American Sociological Review 43:93-109.

1979 Sex differences in criminal activity. Social Problems 27:143-156.

1981 Variations in sex-race-age-specific incidence rates of offending. American Sociological Review 46:461-474.

Hindelang, M., M. Gottfredson, and J. Garofalo 1978 Victims of Personal Crime: An Empirical Foundation for a Theory of Personal Victimization. Cambridge, Mass.: Ballinger.

Hindelang, M., T. Hirschi, and J. Weis 1979 Correlates of delinquency: The illusion of discrepancy between self-report and official measures. American Sociological Review 44:995-1014.

1981 Measuring Delinquency. Beverly Hills, Calif.: Sage Publications.

Hirsch, A. 1983 Making the Second Ghetto: Race and Housing in Chicago 1940-1960 . Chicago: University of Chicago Press.

Suggested Citation:"Violent Victimization and Offending: Individual-Situational-, and Community-Level Risk Factors." National Research Council. 1994. Understanding and Preventing Violence, Volume 3: Social Influences. Washington, DC: The National Academies Press. doi: 10.17226/4421.
×

Hirschi, T., and M. Gottfredson 1983 Age and the explanation of crime. American Journal of Sociology 89:552-584.

Horowitz, R. 1987 Community tolerance of gang violence. Social Problems 34:437-450.

Hough, M., and P. Mayhew 1983 The British Crime Survey: First Report. Home Office Research Study No. 76. London: Her Majesty's Stationery Office.

Huff-Corzine, L., J. Corzine, and D. Moore 1986 Southern exposure: Deciphering the South's influence on homicide rates. Social Forces 64:906-24.

Hunter, A. 1974 Symbolic Communities. Chicago: University of Chicago Press.

Hutchings, B., and S. Mednick 1975 Registered criminality in the adoptive and biological parents of registered male adoptives. In R. Fieve, D. Rosenthal, H. Brill, eds., Genetic Research in Psychiatry. Baltimore: Johns Hopkins University Press.

Jackson, P. 1984 Opportunity and crime: A function of city size. Social Science Research 68:172-193.

Jacobs, D., and D. Britt 1979 Inequality and police use of deadly force: An empirical assessment of a conflict hypothesis. Social Problems 26:403-411.

Jamieson, K., and T. Flanagan 1989 Sourcebook of Criminal Justice Statistics-1988. Washington, D.C.: U.S. Department of Justice.

Jensen, G., and D. Brownfield 1986 Gender, lifestyles, and victimization: Beyond routine activity theory. Violence and Victims 1:85-99.


Kapsis, R. 1976 Continuities in delinquency and riot patterns in black residential areas. Social Problems 23:567-580.

Katz, J. 1988 Seductions of Crime: The Sensual and Moral Attractions of Doing Evil. New York: Basic.

Keiser, R.L. 1970 The Vice Lords: Warriors of the Streets (fieldwork edition). New York: Holt, Rinehart and Winston.

Kennedy, L., and D. Forde 1990 Routine activities and crime: An analysis of victimization in Canada. Criminology 28:137-152.

Kleck, G. 1984 The relationship between gun ownership levels and rates of violence in the United States. Pp. 99-135 in D. Kates, ed., Firearms and Violence: Issues of Public Policy. New York: Ballinger.

Suggested Citation:"Violent Victimization and Offending: Individual-Situational-, and Community-Level Risk Factors." National Research Council. 1994. Understanding and Preventing Violence, Volume 3: Social Influences. Washington, DC: The National Academies Press. doi: 10.17226/4421.
×

Kornhauser, R. 1978 Social Sources of Delinquency. Chicago: University of Chicago Press.

Krohn, M. 1986 The web of conformity: A network approach to the explanation of delinquent behavior. Social Problems 33:81-93.

Kruttschnitt, C., L. Heath, and D. Ward 1986 Family violence, television viewing habits, and other adolescent experiences related to violent criminal behavior. Criminology 24:235-268.

Kruttschnitt, C., D. Ward, and M.A. Scheble 1987 Abuse-resistant youth: Some factors that may inhibit violent criminal behavior. Social Forces 66:501-519.

Land, K., P. McCall, and L. Cohen 1990 Structural covariates of homicide rates: Are there any invariances across time and space? American Journal of Sociology 95:922-963.

Lander, B. 1954 Toward an Understanding of Juvenile Delinquency. New York: Columbia University Press.

Langan, P., and C. Innes 1990 The risk of violent crime. In N. Weiner, M. Zahn, and R. Sagi, eds., Violence: Patterns, Causes, Public Policy. New York: Harcourt Brace Jovanovich.

Laub, J. 1983 Urbanism, race, and crime. Journal of Research in Crime and Delinquency 20:183-198.

Laub, J., and M.J. McDermott 1985 An analysis of serious crime by young black women. Criminology 23:81-98.

Laub, J., and R.J. Sampson 1988 Unraveling families and delinquency: A reanalysis of the Gluecks' data. Criminology 26:355-380.

Lauritsen, J., R. Sampson, and J. Laub 1991 The link between offending and victimization among adolescents. Criminology 29:265-292.

Law Enforcement Assistance Administration 1972 San Jose Methods Test of Crime Victims. Statistics Technical Report TT No. 1. Washington, D.C.: Law Enforcement Assistance Administration.

Levin, Y., and A. Lindesmith 1937 English ecology and criminology of the past century. Journal of Criminal Law and Criminology 27:801-816.

Liebow, E. 1967 Tally's Corner. Boston: Little, Brown.

Suggested Citation:"Violent Victimization and Offending: Individual-Situational-, and Community-Level Risk Factors." National Research Council. 1994. Understanding and Preventing Violence, Volume 3: Social Influences. Washington, DC: The National Academies Press. doi: 10.17226/4421.
×

Liska, A., and M. Reed 1985 Ties to conventional institutions and delinquency: Estimating reciprocal effects. American Sociological Review 50:547-560.

Loeber, R., and M. Stouthamer-Loeber 1986 Family factors as correlates and predictors of juvenile conduct problems and delinquency. Pp. 29-150 in M. Tonry and N. Morris, eds., Crime and Justice: An Annual Review of Research, Vol. 7. Chicago: University of Chicago Press.

Loftin, C., and R. Hill 1974 Regional subculture and homicide: A examination of the Gastil-Hackney thesis. American Sociological Review 39:714-724.

Loftin, C., and E. MacKenzie 1990 Building national estimates of violent victimization. Commissioned paper for the National Academy of Sciences Symposium on ''The Understanding and Control of Violent Behavior," Destin, Fl., April.

Loftin, C., and D. McDowall 1988 The analysis of case-control studies in criminology. Journal of Quantitative Criminology 4:85-98.

Loftin, C., and R. Parker 1985 An errors-in-variable model of the effect of poverty on urban homicide rates. Criminology 23:269-287.

Loftin, C., K. Kindley, S. Norris, and B. Wiersema 1987 An attribute approach to relationships between offenders and victims in homicide. Journal of Criminal Law and Criminology 78:259-271.

Logan, J., and H. Molotch 1987 Urban Fortunes: The Political Economy of Place. Berkeley: University of California Press.

Luckenbill, D., and D. Doyle 1989 Structural position and violence: Developing a cultural explanation. Criminology 27:419-436.

Lynch, J. 1987 Routine activities and victimization at work. Journal of Quantitative Criminology 2:283-300.

Maccoby, E., J. Johnson, and R. Church 1958 Community integration and the social control of juvenile delinquency. Journal of Social Issues 14:38-51.

Massey, D. 1990 American apartheid: Segregation and the making of the underclass. American Journal of Sociology 96:329-357.

Massey, D., and M. Eggers 1990 The ecology of inequality: Minorities and the concentration of poverty, 1970-1980. American Journal of Sociology 95:1153-1188.

Suggested Citation:"Violent Victimization and Offending: Individual-Situational-, and Community-Level Risk Factors." National Research Council. 1994. Understanding and Preventing Violence, Volume 3: Social Influences. Washington, DC: The National Academies Press. doi: 10.17226/4421.
×

Massey, D., and S. Kanaiaupuni 1990 Public Housing and the Concentration of Poverty. Working paper, Population Research Center, University of Chicago.

Maxfield, M. 1987 Lifestyle and routine activity theories of crime: Empirical studies of victimization, delinquency, and offender decision-making. Journal of Quantitative Criminology 3:275-282.

Mayer, S., and C. Jencks 1989 Growing up in poor neighborhoods: How much does it matter? Science 243(March):1441-1445.

McCord, J. 1982 A longitudinal view of the relationship between paternal absence and crime. Pp. 113-128 in J. Gunn and D. Farrington, eds., Abnormal Offenders, Delinquency and the Criminal Justice System. London: Wiley.

McCord, W., J. McCord, and A. Howard 1961 Familial correlates of aggression in non-delinquent male children. Journal of Abnormal Social Psychology 62:79-93.

McDonald, S. 1986 Does gentrification affect crime rates? Pp. 163-202 in A.J. Reiss, Jr., and M. Tonry, eds., Communities and Crime. Chicago: University of Chicago Press.

Messner, S. 1982 Poverty, inequality, and the urban homicide rate. Criminology 20:103-114.

1983a Regional and racial effects on the urban homicide rate: The subculture of violence revisited. American Journal of Sociology 88:997-1007.

1983b Regional differences in the economic correlates of the urban homicide rate: Some evidence on the importance of cultural context. Criminology 21:477-488.

Messner, S., and J. Blau 1987 Routine leisure activities and rates of crime: A macro-level analysis. Social Forces 65:1035-1052.

Messner, S., and R. J. Sampson 1991 The sex ratio, family disruption, and rates of violent crime: The paradox of demographic structure. Social Forces 69:693-714.

Messner, S., and S. South 1986 Economic deprivation, opportunity structure, and robbery victimization: Intra-and interracial patterns. Social Forces 64:975-991.

Messner, S., and K. Tardiff 1986 Economic inequality and levels of homicide: An analysis of urban neighborhoods. Criminology 24:297-318.

Suggested Citation:"Violent Victimization and Offending: Individual-Situational-, and Community-Level Risk Factors." National Research Council. 1994. Understanding and Preventing Violence, Volume 3: Social Influences. Washington, DC: The National Academies Press. doi: 10.17226/4421.
×

Miethe, T., and R. Meier 1990 Opportunity, choice, and criminal victimization. Journal of Research in Crime and Delinquency 27:243-266.

Miethe, T., M. Stafford, and J. Long 1987 Social differentiation in criminal victimization: A test of routine activities/lifestyle theories. American Sociological Review 52:184-194.

Mladenka, K., and K. Hill 1976 A reexamination of the etiology of urban crime. Criminology 13:491-506.

Monahan, J., and D. Klassen 1982 Situational approaches to understanding and predicting individual violent behaviors. Pp. 292-319 in M. Wolfgang and N. Weiner, eds., Criminal Violence. Beverly Hills, Calif.: Sage Publications.

Morris, T. 1958 The Criminal Area. London: Routledge and Kegan Paul.

1970 Book review of Juvenile Delinquency and Urban Areas, 2nd ed. British Journal of Criminology 10:194-196.

O'Brien, R. 1985 Crime and Victimization Data. Beverly Hills, Calif.: Sage Publications.

1987 The interracial nature of violent crimes: A re-examination. American Journal of Sociology 92:817-835.

1988 Exploring the intersexual nature of violent crimes. Criminology 26:151-170.

Ohlin, L., and M. Tonry, eds. 1989 Family Violence. Chicago: University of Chicago Press.


Park, R.E., E. Burgess, and R. McKenzie 1925 The City. Chicago: University of Chicago Pess.

Parker, R.N., and M.D. Smith 1979 Deterrence, poverty, and type of homicide. American Journal of Sociology 85:614-624.

Peterson, R., and C. Bailey 1988 Forcible rape, poverty, and economic inequality in U.S. metropolitan communities. Journal of Quantitative Criminology 4:99-120.

Pierce, G., and W. Bowers 1981 The Bartley-Fox gun law's short-term impact on crime in Boston. Annals of the American Academy of Political and Social Science 455:120-137.

Piper, E. 1985 Violent recidivism and chronicity in the 1958 Philadelphia cohort. Journal of Quantitative Criminology 1:319-344.

Pittman, D., and W. Handy 1964 Patterns in criminal aggravated assault. Journal of Criminal Law, Criminology, and Police Science 55:462-470.

Suggested Citation:"Violent Victimization and Offending: Individual-Situational-, and Community-Level Risk Factors." National Research Council. 1994. Understanding and Preventing Violence, Volume 3: Social Influences. Washington, DC: The National Academies Press. doi: 10.17226/4421.
×

Pokorny, A. 1965 Human violence: A comparison of homicide, aggravated assault, suicide and attempted suicide. Journal of Criminal Law, Criminology, and Police Science 56:488-497.

Pyle, G., E. Hanten, P. Williams, A. Pearson, J. Doyle, and K. Kwofie 1974 The Spatial Dynamics of Crime. Chicago: Department of Geographic Research, University of Chicago.

Rainwater, L. 1970 Behind Ghetto Walls: Black Families in a Federal Slum. Chicago: Aldine.

Rankin, J. 1983 The family context of delinquency. Social Problems 30:466-479.

Raudenbush, S., and A. Bryck 1986 A hierarchical model for studying school effects. Sociology of Education 59:1-17.

Recktenwald, W., and L. Myers 1991 849 killings put 1990 in a sad record book. Chicago Tribune 1,7.

Reiss, A.J., Jr. 1981 Towards a revitalization of theory and research on victimization by crime. Journal of Criminal Law and Criminology 72:704-713.

1986a Why are communities important in understanding crime? Pp. 1-33 in A.J. Reiss, Jr., and M. Tonry, eds., Communities and Crime. Chicago: University of Chicago Press.

1986b Co-offender influences on criminal careers. Pp. 121-160 in A. Blumstein, J. Cohen, J. Roth, and C. Visher, eds., Criminal Careers and "Career Criminals." Washington, D.C.: National Academy Press.

Reiss, A.J., Jr., ed. 1989 Proceedings of the Workshop on Communities and Crime Control. Committee on Research on Law Enforcement and the Administration of Justice, National Research Council. Washington, D.C.: National Criminal Justice Reference Service.

Riedel, M. 1987 Stranger violence: Perspectives, issues, and problems. Journal of Criminal Law and Criminology 78:223-258.

Rieder, J. 1985 Canarsie: The Jews and Italians of Brooklyn Against Liberalism . Cambridge, Mass.: Harvard University Press.

Robins, L. 1966 Deviant Children Grown Up. Baltimore: Williams and Wilkins.

Roncek, D. 1981 Dangerous places: Crime and residential environment. Social Forces 60:74-96.

Suggested Citation:"Violent Victimization and Offending: Individual-Situational-, and Community-Level Risk Factors." National Research Council. 1994. Understanding and Preventing Violence, Volume 3: Social Influences. Washington, DC: The National Academies Press. doi: 10.17226/4421.
×

Roncek, D., and D. Faggiani 1986 High schools and crime: A replication. Sociological Quarterly 26:491-505.

Roncek, D., R. Bell, and J. Francik 1981 Housing projects and crime: Testing a proximity hypothesis. Social Problems 29:151-166.

Rose, H., and P. McClain 1990 Race, Place and Risk: Black Homicide in Urban America. Albany: State University of New York Press.

Rosenfeld, R. 1986 Urban crime rates: Effects of inequality, welfare, dependency, region, and race. In J. Byrne and R. Sampson, eds., The Social Ecology of Crime. New York: Springer-Verlag.

Sampson, R.J. 1983 Structural density and criminal victimization. Criminology 21:276-293.

1984 Group size, heterogeneity, and intergroup conflict. Social Forces 62:618-639.

1985a Neighborhood and crime: The structural determinants of personal victimization. Journal of Research in Crime and Delinquency 22:7-40.

1985b Race and criminal violence: A demographically disaggregated analysis of urban homicide. Crime and Delinquency 31:47-82.

1986a Neighborhood family structure and the risk of criminal victimization. Pp. 25-46 in J. Byrne and R. Sampson, eds., The Social Ecology of Crime. New York: Springer-Verlag.

1986b Crime in cities: The effects of formal and informal social control. Pp. 271-311 in A.J. Reiss, Jr., and M. Tonry, eds., Communities and Crime. Chicago: University of Chicago Press.

1986c Effects of socioeconomic context on official reaction to juvenile delinquency. American Sociological Review 51:876-885.

1987a Urban black violence: The effect of male joblessness and family disruption. American Journal of Sociology 93:348-382.

1987b Personal violence by strangers: An extension and test of the opportunity model of predatory victimization. Journal of Criminal Law and Criminology 78:327-356.

1988 Community attachment in mass society: A multilevel systemic model. American Sociological Review 53:766-769.

1990 The impact of housing policies on community social disorganization and crime. Bulletin of the New York Academy of Medicine 66:526-533.

Sampson, R.J., and W.B. Groves 1989 Community structure and crime: Testing social-disorganization theory. American Journal of Sociology 94:774-802.

Sampson, R.J., and J.H. Laub 1990 Crime and deviance over the life course: The salience of adult social bonds. American Sociological Review 55:609-627.

Suggested Citation:"Violent Victimization and Offending: Individual-Situational-, and Community-Level Risk Factors." National Research Council. 1994. Understanding and Preventing Violence, Volume 3: Social Influences. Washington, DC: The National Academies Press. doi: 10.17226/4421.
×

Sampson, R.J., and J. Lauritsen 1990 Deviant lifestyles, proximity to crime, and the offender-victim link in personal violence. Journal of Research in Crime and Delinquency 27:110-139.

Sampson, R.J., and J. Wooldredge 1986 Evidence that high crime rates encourage migration away from central cities. Sociology and Social Research 70:310-314.

1987 Linking the micro-and macro-level dimensions of lifestyle-routine activity and opportunity models of predatory victimization. Journal of Quantitative Criminology 3:371-393.

Schlesselmann, J. 1982 Case-Control Studies. New York: Oxford University Press.

Schmid, C. 1960 Urban crime areas, part I. American Sociological Review 25:527-542.

Schuerman, L., and S. Kobrin 1986 Community careers in crime. Pp. 67-100 in A.J. Reiss, Jr., and M. Tonry, eds., Communities and Crime. Chicago: University of Chicago Press.

Shaw, C., and H. McKay 1931 Social factors in juvenile delinquency. Vol. 2 of Report of the Causes of Crime. National Commission of Law Observance and Enforcement. Washington, D.C.: U.S. Government Printing Office.

1942 Juvenile Delinquency and Urban Areas. Chicago: University of Chicago Press.

1969 Juvenile Delinquency and Urban Areas (rev. ed.). Chicago: University of Chicago Press.

Shaw, C., F. Zorbaugh, H. McKay, and L. Cottrell 1929 Delinquency Areas. Chicago: University of Chicago Press.

Sherman, L., P. Gartin, and M. Buerger 1989 Hot spots of predatory crime: Routine activities and the criminology of place. Criminology 27:27-56.

Shin, Y., D. Jedlicka, and E. Lee 1977 Homicide among blacks. Phylon 38:398-407.

Short, J.F., Jr. 1963 Introduction to the abridged edition. Pp. xv-liii in F. Thrasher, The Gang: A Study of 1,313 Gangs in Chicago. Chicago: University of Chicago Press.

1969 Introduction to the revised edition. Pp. xxv-liv in C. Shaw and H. McKay, eds., Juvenile Delinquency and Urban Areas. Chicago: University of Chicago Press.

1985 The level of explanation problem in criminology. Pp. 51-74 in R. F. Meier, ed., Theoretical Methods in Criminology. Beverly Hills, Calif.: Sage Publications.

Suggested Citation:"Violent Victimization and Offending: Individual-Situational-, and Community-Level Risk Factors." National Research Council. 1994. Understanding and Preventing Violence, Volume 3: Social Influences. Washington, DC: The National Academies Press. doi: 10.17226/4421.
×

1990 Delinquency and Society. Englewood Cliffs, N.J.: Prentice-Hall.

Short, J.F., and F. Strodtbeck 1965 Group Process and Gang Delinquency. Chicago: University of Chicago Press.

Silverman, R., and L. Kennedy 1987 Relational distance and homicide: The role of the stranger. Journal of Criminal Law and Criminology 78:272-308.

Simcha-Fagan, O., and J. Schwartz 1986 Neighborhood and delinquency: An assessment of contextual effects. Criminology 24:667-704.

Simon, R. 1975 Women and Crime. Lexington, Mass.: D.C. Heath.

Simon, R., and S. Baxter 1989 Gender and violent crime. Pp. 171-197 in N. Weiner and M. Wolfgang, eds., Violent Crime, Violent Criminals. Beverly Hills, Calif.: Sage Publications.

Simpson, M. 1985 Violent crime, income inequality, and regional culture: Another look. Sociological Focus 18:199-208.

Singer, S. 1981 Homogeneous victim-offender populations: A review and some research implications. Journal of Criminal Law and Criminology 72:779-788.

Skogan, W. 1981 Assessing the behavioral context of victimization. Journal of Criminal Law and Criminology 72:727-742.

1986 Fear of crime and neighborhood change. Pp. 203-229 in A.J. Reiss, Jr., and M. Tonry, eds., Communities and Crime. Chicago: University of Chicago Press.

Smith, D.R. 1986 The neighborhood context of police behavior. Pp. 313-341 in A.J. Reiss, Jr., and M. Tonry, eds., Communities and Crime. Chicago: University of Chicago Press.

Smith, D.R., and G.R. Jarjoura 1988 Social structure and criminal victimization. Journal of Research in Crime and Delinquency 25:27-52.

Smith, D.R., and C. Visher 1980 Sex and involvement in deviance/crime: A quantitative review of the empirical literature. American Sociological Review 45:691-701.

Smith, M. 1987 Changes in the victimization of women: Is there a new female victim? Journal of Research in Crime and Delinquency 24:291-301.

Suggested Citation:"Violent Victimization and Offending: Individual-Situational-, and Community-Level Risk Factors." National Research Council. 1994. Understanding and Preventing Violence, Volume 3: Social Influences. Washington, DC: The National Academies Press. doi: 10.17226/4421.
×

Smith, M.D., and R.N. Parker 1980 Type of homicide and variation in regional rates. Social Forces 59:137-147.

Smith, S. 1982 Victimization in the inner city: A British case study. British Journal of Criminology 22:386-402.

South, S., and R. Felson 1990 The racial patterning of rape. Social Forces 69:71-94.

Sparks, R. 1982 Research on Victims of Crime. Washington, D.C.: U.S. Government Printing Office.

Sparks, R., H. Glenn, and D. Dodd 1977 Surveying Victims. London: Wiley.

Spergel, I. 1964 Racketville, Slumtown, Haulburg. Chicago: University of Chicago Press.

Stahura, J., and J. Sloan, III 1988 Urban stratification of places, routine activities and suburban crime rates. Social Forces 66:1102-1118.

Stark, R. 1987 Deviant places: A theory of the ecology of crime. Criminology 25:893-909.

Steffensmeier, D., E.A. Allan, M. Harer, and C. Streifel 1989 Age and the distribution of crime. American Journal of Sociology 94:803-831.

Straus, M., and R. Gelles 1986 Societal change and change n family violence from 1975 to 1985 as revealed by two national surveys. Journal of Marriage and the Family 48:465-479.

Street, L. 1989 Why focus on communities? Pp. 27-34 in A.J. Reiss, Jr., ed., Proceedings of the Workshop on Communities and Crime Control. Committee on Research on Law Enforcement and the Administration of Justice, National Research Council. Washington, D.C.: National Criminal Justice Reference Service.

Sullivan, M. 1989 Getting Paid: Economy, Culture, and Youth Crime in the Inner City. Ithaca, N.Y.: Cornell University Press.

Suttles, G. 1968 The Social Order of the Slum. Chicago: University of Chicago Press.

Taub, R., D.G. Taylor, and J. Dunham 1984 Paths of Neighborhood Change: Race and Crime in Urban America . Chicago: University of Chicago Press.

Taylor, R., and J. Covington 1988 Neighborhood changes in ecology and violence. Criminology 26:553-590.

Suggested Citation:"Violent Victimization and Offending: Individual-Situational-, and Community-Level Risk Factors." National Research Council. 1994. Understanding and Preventing Violence, Volume 3: Social Influences. Washington, DC: The National Academies Press. doi: 10.17226/4421.
×

Taylor, R., and S. Gottfredson 1986 Environmental design, crime, and prevention: An examination of community dynamics. Pp. 387-416 in A.J. Reiss, Jr., and M. Tonry, eds., Communities and Crime. Chicago: University of Chicago Press.

Taylor, R., S. Gottfredson, and S. Brower 1984 Block crime and fear: Defensible space, local social ties, and territorial functioning. Journal of Research in Crime and Delinquency 21:303-331.

Thornberry, T., and M. Farnworth 1982 Social correlates of criminal involvement: Further evidence on the relationship between social status and criminal behavior. American Sociological Review 47:505-517.

Thrasher, F. 1963 The Gang: A Study of 1,313 Gangs in Chicago (rev. ed.). Chicago: University of Chicago Press.

Tienda, M. 1989 Poor People and Poor Places: Deciphering Neighborhood Effects on Behavioral Outcomes. Paper presented at the annual meeting of the American Sociological Association, San Francisco, August 9-13.

Tittle, C., W. Villemez, and D. Smith 1978 The myth of social class and criminality: An empirical assessment of the empirical evidence. American Sociological Review 43:643-656.

Tonry, M., and J.Q. Wilson, eds. 1990 Drugs and Crime. Chicago: University of Chicago Press.

U.S. Department of Justice 1985 The Risk of Violent Crime. Washington, D.C.: U.S. Department of Justice.

1986 Crime in the United States, 1985. Washington, D.C.: U.S. Department of Justice.

1988a Report to the Nation on Crime and Justice, 2nd ed. Washington, D.C.: U.S. Department of Justice.

1988b Crime in the United States, 1987. Washington, D.C.: U.S. Department of Justice.

1989 Criminal Victimization in the United States, 1987. A National Crime Survey Report, June. Washington, D.C.: U.S. Department of Justice.


van Dijk, J., and C. Steinmetz 1983 Victimization surveys: Beyond measuring the volume of crime. Victimology 8:291-301.

Van Voorhis, P., F. Cullen, R. Mathers, and C.C. Garner 1988 The impact of family structure and quality on delinquency: A comparative assessment of structural and functional factors. Criminology 26:235-261.

Suggested Citation:"Violent Victimization and Offending: Individual-Situational-, and Community-Level Risk Factors." National Research Council. 1994. Understanding and Preventing Violence, Volume 3: Social Influences. Washington, DC: The National Academies Press. doi: 10.17226/4421.
×

Visher, C., and J. Roth 1986 Participation in criminal careers. Pp. 211-191 in A. Blumstein, J. Cohen, J. Roth, and C. Visher, eds., Criminal Careers and "Career Criminals." Washington, D.C.: National Academy Press.

Wallace, R., and D. Wallace 1990 Origins of public health collapse in New York City: The dynamics of planned shrinkage, contagious urban decay and social disintegration. Bulletin of the New York Academy of Medicine 66:391-434.

Weiner, N. 1989 Violent criminal careers and "violent career criminals." Pp. 35-138 in N. Weiner and M. Wolfgang, eds., Violent Crime, Violent Criminals. Beverly Hills, Calif.: Sage Publications.

Wells, L., and J. Rankin 1988 Direct parental controls and delinquency. Criminology 26:263-285.

West, D., and D. Farrington 1977 The Delinquent Way of Life. London: Heinemann.

Widom, C.S. 1989a The intergenerational transmission of violence. Pp. 137-201 in N. Weiner and M. Wolfgang, eds., Pathways to Criminal Violence. Beverly Hills, Calif.: Sage Publications.

1989b Child abuse, neglect, and violent criminal behavior. Criminology 27:251-272.

Wilbanks, W. 1985 Is violent crime intraracial? Crime and Delinquency 31:117-128.

Wilkinson, K. 1980 The broken home and delinquent behavior. Pp. 21-42 in T. Hirschi and M. Gottfredson, eds., Understanding Crime. Beverly Hills, Calif.: Sage Publications.

Williams, K. 1984 Economic sources of homicide: Reestimating the effects of poverty and inequality. American Sociological Review 49:283-289.

Williams, K., and R. Flewelling 1988 The social production of criminal homicide: A comparative study of disaggregated rates in American cities. American Sociological Review 53:421-431.

Wilson, W.J. 1987 The Truly Disadvantaged: The Inner City, the Underclass, and Public Policy. Chicago: University of Chicago Press.

Wolfgang, M. 1958 Patterns in Criminal Homicide. New York: John Wiley & Sons.

Wolfgang, M., and F. Ferracuti 1967 The Subculture of Violence. London: Tavistock.

Suggested Citation:"Violent Victimization and Offending: Individual-Situational-, and Community-Level Risk Factors." National Research Council. 1994. Understanding and Preventing Violence, Volume 3: Social Influences. Washington, DC: The National Academies Press. doi: 10.17226/4421.
×

Wolfgang, M., T. Thornberry, and R. Figlio 1987 From Boy to Man, From Delinquency to Crime. Chicago: University of Chicago Press.

Zahn, M., and P. Sagi 1987 Stranger homicides in nine American cities. Journal of Criminal Law and Criminology 78:377-396.

Zimring, F. 1981 Kids, groups, and crime: Some implications of a well known secret. Journal of Criminal Law and Criminology 72:867-885

Suggested Citation:"Violent Victimization and Offending: Individual-Situational-, and Community-Level Risk Factors." National Research Council. 1994. Understanding and Preventing Violence, Volume 3: Social Influences. Washington, DC: The National Academies Press. doi: 10.17226/4421.
×
Page 1
Suggested Citation:"Violent Victimization and Offending: Individual-Situational-, and Community-Level Risk Factors." National Research Council. 1994. Understanding and Preventing Violence, Volume 3: Social Influences. Washington, DC: The National Academies Press. doi: 10.17226/4421.
×
Page 2
Suggested Citation:"Violent Victimization and Offending: Individual-Situational-, and Community-Level Risk Factors." National Research Council. 1994. Understanding and Preventing Violence, Volume 3: Social Influences. Washington, DC: The National Academies Press. doi: 10.17226/4421.
×
Page 3
Suggested Citation:"Violent Victimization and Offending: Individual-Situational-, and Community-Level Risk Factors." National Research Council. 1994. Understanding and Preventing Violence, Volume 3: Social Influences. Washington, DC: The National Academies Press. doi: 10.17226/4421.
×
Page 4
Suggested Citation:"Violent Victimization and Offending: Individual-Situational-, and Community-Level Risk Factors." National Research Council. 1994. Understanding and Preventing Violence, Volume 3: Social Influences. Washington, DC: The National Academies Press. doi: 10.17226/4421.
×
Page 5
Suggested Citation:"Violent Victimization and Offending: Individual-Situational-, and Community-Level Risk Factors." National Research Council. 1994. Understanding and Preventing Violence, Volume 3: Social Influences. Washington, DC: The National Academies Press. doi: 10.17226/4421.
×
Page 6
Suggested Citation:"Violent Victimization and Offending: Individual-Situational-, and Community-Level Risk Factors." National Research Council. 1994. Understanding and Preventing Violence, Volume 3: Social Influences. Washington, DC: The National Academies Press. doi: 10.17226/4421.
×
Page 7
Suggested Citation:"Violent Victimization and Offending: Individual-Situational-, and Community-Level Risk Factors." National Research Council. 1994. Understanding and Preventing Violence, Volume 3: Social Influences. Washington, DC: The National Academies Press. doi: 10.17226/4421.
×
Page 8
Suggested Citation:"Violent Victimization and Offending: Individual-Situational-, and Community-Level Risk Factors." National Research Council. 1994. Understanding and Preventing Violence, Volume 3: Social Influences. Washington, DC: The National Academies Press. doi: 10.17226/4421.
×
Page 9
Suggested Citation:"Violent Victimization and Offending: Individual-Situational-, and Community-Level Risk Factors." National Research Council. 1994. Understanding and Preventing Violence, Volume 3: Social Influences. Washington, DC: The National Academies Press. doi: 10.17226/4421.
×
Page 10
Suggested Citation:"Violent Victimization and Offending: Individual-Situational-, and Community-Level Risk Factors." National Research Council. 1994. Understanding and Preventing Violence, Volume 3: Social Influences. Washington, DC: The National Academies Press. doi: 10.17226/4421.
×
Page 11
Suggested Citation:"Violent Victimization and Offending: Individual-Situational-, and Community-Level Risk Factors." National Research Council. 1994. Understanding and Preventing Violence, Volume 3: Social Influences. Washington, DC: The National Academies Press. doi: 10.17226/4421.
×
Page 12
Suggested Citation:"Violent Victimization and Offending: Individual-Situational-, and Community-Level Risk Factors." National Research Council. 1994. Understanding and Preventing Violence, Volume 3: Social Influences. Washington, DC: The National Academies Press. doi: 10.17226/4421.
×
Page 13
Suggested Citation:"Violent Victimization and Offending: Individual-Situational-, and Community-Level Risk Factors." National Research Council. 1994. Understanding and Preventing Violence, Volume 3: Social Influences. Washington, DC: The National Academies Press. doi: 10.17226/4421.
×
Page 14
Suggested Citation:"Violent Victimization and Offending: Individual-Situational-, and Community-Level Risk Factors." National Research Council. 1994. Understanding and Preventing Violence, Volume 3: Social Influences. Washington, DC: The National Academies Press. doi: 10.17226/4421.
×
Page 15
Suggested Citation:"Violent Victimization and Offending: Individual-Situational-, and Community-Level Risk Factors." National Research Council. 1994. Understanding and Preventing Violence, Volume 3: Social Influences. Washington, DC: The National Academies Press. doi: 10.17226/4421.
×
Page 16
Suggested Citation:"Violent Victimization and Offending: Individual-Situational-, and Community-Level Risk Factors." National Research Council. 1994. Understanding and Preventing Violence, Volume 3: Social Influences. Washington, DC: The National Academies Press. doi: 10.17226/4421.
×
Page 17
Suggested Citation:"Violent Victimization and Offending: Individual-Situational-, and Community-Level Risk Factors." National Research Council. 1994. Understanding and Preventing Violence, Volume 3: Social Influences. Washington, DC: The National Academies Press. doi: 10.17226/4421.
×
Page 18
Suggested Citation:"Violent Victimization and Offending: Individual-Situational-, and Community-Level Risk Factors." National Research Council. 1994. Understanding and Preventing Violence, Volume 3: Social Influences. Washington, DC: The National Academies Press. doi: 10.17226/4421.
×
Page 19
Suggested Citation:"Violent Victimization and Offending: Individual-Situational-, and Community-Level Risk Factors." National Research Council. 1994. Understanding and Preventing Violence, Volume 3: Social Influences. Washington, DC: The National Academies Press. doi: 10.17226/4421.
×
Page 20
Suggested Citation:"Violent Victimization and Offending: Individual-Situational-, and Community-Level Risk Factors." National Research Council. 1994. Understanding and Preventing Violence, Volume 3: Social Influences. Washington, DC: The National Academies Press. doi: 10.17226/4421.
×
Page 21
Suggested Citation:"Violent Victimization and Offending: Individual-Situational-, and Community-Level Risk Factors." National Research Council. 1994. Understanding and Preventing Violence, Volume 3: Social Influences. Washington, DC: The National Academies Press. doi: 10.17226/4421.
×
Page 22
Suggested Citation:"Violent Victimization and Offending: Individual-Situational-, and Community-Level Risk Factors." National Research Council. 1994. Understanding and Preventing Violence, Volume 3: Social Influences. Washington, DC: The National Academies Press. doi: 10.17226/4421.
×
Page 23
Suggested Citation:"Violent Victimization and Offending: Individual-Situational-, and Community-Level Risk Factors." National Research Council. 1994. Understanding and Preventing Violence, Volume 3: Social Influences. Washington, DC: The National Academies Press. doi: 10.17226/4421.
×
Page 24
Suggested Citation:"Violent Victimization and Offending: Individual-Situational-, and Community-Level Risk Factors." National Research Council. 1994. Understanding and Preventing Violence, Volume 3: Social Influences. Washington, DC: The National Academies Press. doi: 10.17226/4421.
×
Page 25
Suggested Citation:"Violent Victimization and Offending: Individual-Situational-, and Community-Level Risk Factors." National Research Council. 1994. Understanding and Preventing Violence, Volume 3: Social Influences. Washington, DC: The National Academies Press. doi: 10.17226/4421.
×
Page 26
Suggested Citation:"Violent Victimization and Offending: Individual-Situational-, and Community-Level Risk Factors." National Research Council. 1994. Understanding and Preventing Violence, Volume 3: Social Influences. Washington, DC: The National Academies Press. doi: 10.17226/4421.
×
Page 27
Suggested Citation:"Violent Victimization and Offending: Individual-Situational-, and Community-Level Risk Factors." National Research Council. 1994. Understanding and Preventing Violence, Volume 3: Social Influences. Washington, DC: The National Academies Press. doi: 10.17226/4421.
×
Page 28
Suggested Citation:"Violent Victimization and Offending: Individual-Situational-, and Community-Level Risk Factors." National Research Council. 1994. Understanding and Preventing Violence, Volume 3: Social Influences. Washington, DC: The National Academies Press. doi: 10.17226/4421.
×
Page 29
Suggested Citation:"Violent Victimization and Offending: Individual-Situational-, and Community-Level Risk Factors." National Research Council. 1994. Understanding and Preventing Violence, Volume 3: Social Influences. Washington, DC: The National Academies Press. doi: 10.17226/4421.
×
Page 30
Suggested Citation:"Violent Victimization and Offending: Individual-Situational-, and Community-Level Risk Factors." National Research Council. 1994. Understanding and Preventing Violence, Volume 3: Social Influences. Washington, DC: The National Academies Press. doi: 10.17226/4421.
×
Page 31
Suggested Citation:"Violent Victimization and Offending: Individual-Situational-, and Community-Level Risk Factors." National Research Council. 1994. Understanding and Preventing Violence, Volume 3: Social Influences. Washington, DC: The National Academies Press. doi: 10.17226/4421.
×
Page 32
Suggested Citation:"Violent Victimization and Offending: Individual-Situational-, and Community-Level Risk Factors." National Research Council. 1994. Understanding and Preventing Violence, Volume 3: Social Influences. Washington, DC: The National Academies Press. doi: 10.17226/4421.
×
Page 33
Suggested Citation:"Violent Victimization and Offending: Individual-Situational-, and Community-Level Risk Factors." National Research Council. 1994. Understanding and Preventing Violence, Volume 3: Social Influences. Washington, DC: The National Academies Press. doi: 10.17226/4421.
×
Page 34
Suggested Citation:"Violent Victimization and Offending: Individual-Situational-, and Community-Level Risk Factors." National Research Council. 1994. Understanding and Preventing Violence, Volume 3: Social Influences. Washington, DC: The National Academies Press. doi: 10.17226/4421.
×
Page 35
Suggested Citation:"Violent Victimization and Offending: Individual-Situational-, and Community-Level Risk Factors." National Research Council. 1994. Understanding and Preventing Violence, Volume 3: Social Influences. Washington, DC: The National Academies Press. doi: 10.17226/4421.
×
Page 36
Suggested Citation:"Violent Victimization and Offending: Individual-Situational-, and Community-Level Risk Factors." National Research Council. 1994. Understanding and Preventing Violence, Volume 3: Social Influences. Washington, DC: The National Academies Press. doi: 10.17226/4421.
×
Page 37
Suggested Citation:"Violent Victimization and Offending: Individual-Situational-, and Community-Level Risk Factors." National Research Council. 1994. Understanding and Preventing Violence, Volume 3: Social Influences. Washington, DC: The National Academies Press. doi: 10.17226/4421.
×
Page 38
Suggested Citation:"Violent Victimization and Offending: Individual-Situational-, and Community-Level Risk Factors." National Research Council. 1994. Understanding and Preventing Violence, Volume 3: Social Influences. Washington, DC: The National Academies Press. doi: 10.17226/4421.
×
Page 39
Suggested Citation:"Violent Victimization and Offending: Individual-Situational-, and Community-Level Risk Factors." National Research Council. 1994. Understanding and Preventing Violence, Volume 3: Social Influences. Washington, DC: The National Academies Press. doi: 10.17226/4421.
×
Page 40
Suggested Citation:"Violent Victimization and Offending: Individual-Situational-, and Community-Level Risk Factors." National Research Council. 1994. Understanding and Preventing Violence, Volume 3: Social Influences. Washington, DC: The National Academies Press. doi: 10.17226/4421.
×
Page 41
Suggested Citation:"Violent Victimization and Offending: Individual-Situational-, and Community-Level Risk Factors." National Research Council. 1994. Understanding and Preventing Violence, Volume 3: Social Influences. Washington, DC: The National Academies Press. doi: 10.17226/4421.
×
Page 42
Suggested Citation:"Violent Victimization and Offending: Individual-Situational-, and Community-Level Risk Factors." National Research Council. 1994. Understanding and Preventing Violence, Volume 3: Social Influences. Washington, DC: The National Academies Press. doi: 10.17226/4421.
×
Page 43
Suggested Citation:"Violent Victimization and Offending: Individual-Situational-, and Community-Level Risk Factors." National Research Council. 1994. Understanding and Preventing Violence, Volume 3: Social Influences. Washington, DC: The National Academies Press. doi: 10.17226/4421.
×
Page 44
Suggested Citation:"Violent Victimization and Offending: Individual-Situational-, and Community-Level Risk Factors." National Research Council. 1994. Understanding and Preventing Violence, Volume 3: Social Influences. Washington, DC: The National Academies Press. doi: 10.17226/4421.
×
Page 45
Suggested Citation:"Violent Victimization and Offending: Individual-Situational-, and Community-Level Risk Factors." National Research Council. 1994. Understanding and Preventing Violence, Volume 3: Social Influences. Washington, DC: The National Academies Press. doi: 10.17226/4421.
×
Page 46
Suggested Citation:"Violent Victimization and Offending: Individual-Situational-, and Community-Level Risk Factors." National Research Council. 1994. Understanding and Preventing Violence, Volume 3: Social Influences. Washington, DC: The National Academies Press. doi: 10.17226/4421.
×
Page 47
Suggested Citation:"Violent Victimization and Offending: Individual-Situational-, and Community-Level Risk Factors." National Research Council. 1994. Understanding and Preventing Violence, Volume 3: Social Influences. Washington, DC: The National Academies Press. doi: 10.17226/4421.
×
Page 48
Suggested Citation:"Violent Victimization and Offending: Individual-Situational-, and Community-Level Risk Factors." National Research Council. 1994. Understanding and Preventing Violence, Volume 3: Social Influences. Washington, DC: The National Academies Press. doi: 10.17226/4421.
×
Page 49
Suggested Citation:"Violent Victimization and Offending: Individual-Situational-, and Community-Level Risk Factors." National Research Council. 1994. Understanding and Preventing Violence, Volume 3: Social Influences. Washington, DC: The National Academies Press. doi: 10.17226/4421.
×
Page 50
Suggested Citation:"Violent Victimization and Offending: Individual-Situational-, and Community-Level Risk Factors." National Research Council. 1994. Understanding and Preventing Violence, Volume 3: Social Influences. Washington, DC: The National Academies Press. doi: 10.17226/4421.
×
Page 51
Suggested Citation:"Violent Victimization and Offending: Individual-Situational-, and Community-Level Risk Factors." National Research Council. 1994. Understanding and Preventing Violence, Volume 3: Social Influences. Washington, DC: The National Academies Press. doi: 10.17226/4421.
×
Page 52
Suggested Citation:"Violent Victimization and Offending: Individual-Situational-, and Community-Level Risk Factors." National Research Council. 1994. Understanding and Preventing Violence, Volume 3: Social Influences. Washington, DC: The National Academies Press. doi: 10.17226/4421.
×
Page 53
Suggested Citation:"Violent Victimization and Offending: Individual-Situational-, and Community-Level Risk Factors." National Research Council. 1994. Understanding and Preventing Violence, Volume 3: Social Influences. Washington, DC: The National Academies Press. doi: 10.17226/4421.
×
Page 54
Suggested Citation:"Violent Victimization and Offending: Individual-Situational-, and Community-Level Risk Factors." National Research Council. 1994. Understanding and Preventing Violence, Volume 3: Social Influences. Washington, DC: The National Academies Press. doi: 10.17226/4421.
×
Page 55
Suggested Citation:"Violent Victimization and Offending: Individual-Situational-, and Community-Level Risk Factors." National Research Council. 1994. Understanding and Preventing Violence, Volume 3: Social Influences. Washington, DC: The National Academies Press. doi: 10.17226/4421.
×
Page 56
Suggested Citation:"Violent Victimization and Offending: Individual-Situational-, and Community-Level Risk Factors." National Research Council. 1994. Understanding and Preventing Violence, Volume 3: Social Influences. Washington, DC: The National Academies Press. doi: 10.17226/4421.
×
Page 57
Suggested Citation:"Violent Victimization and Offending: Individual-Situational-, and Community-Level Risk Factors." National Research Council. 1994. Understanding and Preventing Violence, Volume 3: Social Influences. Washington, DC: The National Academies Press. doi: 10.17226/4421.
×
Page 58
Suggested Citation:"Violent Victimization and Offending: Individual-Situational-, and Community-Level Risk Factors." National Research Council. 1994. Understanding and Preventing Violence, Volume 3: Social Influences. Washington, DC: The National Academies Press. doi: 10.17226/4421.
×
Page 59
Suggested Citation:"Violent Victimization and Offending: Individual-Situational-, and Community-Level Risk Factors." National Research Council. 1994. Understanding and Preventing Violence, Volume 3: Social Influences. Washington, DC: The National Academies Press. doi: 10.17226/4421.
×
Page 60
Suggested Citation:"Violent Victimization and Offending: Individual-Situational-, and Community-Level Risk Factors." National Research Council. 1994. Understanding and Preventing Violence, Volume 3: Social Influences. Washington, DC: The National Academies Press. doi: 10.17226/4421.
×
Page 61
Suggested Citation:"Violent Victimization and Offending: Individual-Situational-, and Community-Level Risk Factors." National Research Council. 1994. Understanding and Preventing Violence, Volume 3: Social Influences. Washington, DC: The National Academies Press. doi: 10.17226/4421.
×
Page 62
Suggested Citation:"Violent Victimization and Offending: Individual-Situational-, and Community-Level Risk Factors." National Research Council. 1994. Understanding and Preventing Violence, Volume 3: Social Influences. Washington, DC: The National Academies Press. doi: 10.17226/4421.
×
Page 63
Suggested Citation:"Violent Victimization and Offending: Individual-Situational-, and Community-Level Risk Factors." National Research Council. 1994. Understanding and Preventing Violence, Volume 3: Social Influences. Washington, DC: The National Academies Press. doi: 10.17226/4421.
×
Page 64
Suggested Citation:"Violent Victimization and Offending: Individual-Situational-, and Community-Level Risk Factors." National Research Council. 1994. Understanding and Preventing Violence, Volume 3: Social Influences. Washington, DC: The National Academies Press. doi: 10.17226/4421.
×
Page 65
Suggested Citation:"Violent Victimization and Offending: Individual-Situational-, and Community-Level Risk Factors." National Research Council. 1994. Understanding and Preventing Violence, Volume 3: Social Influences. Washington, DC: The National Academies Press. doi: 10.17226/4421.
×
Page 66
Suggested Citation:"Violent Victimization and Offending: Individual-Situational-, and Community-Level Risk Factors." National Research Council. 1994. Understanding and Preventing Violence, Volume 3: Social Influences. Washington, DC: The National Academies Press. doi: 10.17226/4421.
×
Page 67
Suggested Citation:"Violent Victimization and Offending: Individual-Situational-, and Community-Level Risk Factors." National Research Council. 1994. Understanding and Preventing Violence, Volume 3: Social Influences. Washington, DC: The National Academies Press. doi: 10.17226/4421.
×
Page 68
Suggested Citation:"Violent Victimization and Offending: Individual-Situational-, and Community-Level Risk Factors." National Research Council. 1994. Understanding and Preventing Violence, Volume 3: Social Influences. Washington, DC: The National Academies Press. doi: 10.17226/4421.
×
Page 69
Suggested Citation:"Violent Victimization and Offending: Individual-Situational-, and Community-Level Risk Factors." National Research Council. 1994. Understanding and Preventing Violence, Volume 3: Social Influences. Washington, DC: The National Academies Press. doi: 10.17226/4421.
×
Page 70
Suggested Citation:"Violent Victimization and Offending: Individual-Situational-, and Community-Level Risk Factors." National Research Council. 1994. Understanding and Preventing Violence, Volume 3: Social Influences. Washington, DC: The National Academies Press. doi: 10.17226/4421.
×
Page 71
Suggested Citation:"Violent Victimization and Offending: Individual-Situational-, and Community-Level Risk Factors." National Research Council. 1994. Understanding and Preventing Violence, Volume 3: Social Influences. Washington, DC: The National Academies Press. doi: 10.17226/4421.
×
Page 72
Suggested Citation:"Violent Victimization and Offending: Individual-Situational-, and Community-Level Risk Factors." National Research Council. 1994. Understanding and Preventing Violence, Volume 3: Social Influences. Washington, DC: The National Academies Press. doi: 10.17226/4421.
×
Page 73
Suggested Citation:"Violent Victimization and Offending: Individual-Situational-, and Community-Level Risk Factors." National Research Council. 1994. Understanding and Preventing Violence, Volume 3: Social Influences. Washington, DC: The National Academies Press. doi: 10.17226/4421.
×
Page 74
Suggested Citation:"Violent Victimization and Offending: Individual-Situational-, and Community-Level Risk Factors." National Research Council. 1994. Understanding and Preventing Violence, Volume 3: Social Influences. Washington, DC: The National Academies Press. doi: 10.17226/4421.
×
Page 75
Suggested Citation:"Violent Victimization and Offending: Individual-Situational-, and Community-Level Risk Factors." National Research Council. 1994. Understanding and Preventing Violence, Volume 3: Social Influences. Washington, DC: The National Academies Press. doi: 10.17226/4421.
×
Page 76
Suggested Citation:"Violent Victimization and Offending: Individual-Situational-, and Community-Level Risk Factors." National Research Council. 1994. Understanding and Preventing Violence, Volume 3: Social Influences. Washington, DC: The National Academies Press. doi: 10.17226/4421.
×
Page 77
Suggested Citation:"Violent Victimization and Offending: Individual-Situational-, and Community-Level Risk Factors." National Research Council. 1994. Understanding and Preventing Violence, Volume 3: Social Influences. Washington, DC: The National Academies Press. doi: 10.17226/4421.
×
Page 78
Suggested Citation:"Violent Victimization and Offending: Individual-Situational-, and Community-Level Risk Factors." National Research Council. 1994. Understanding and Preventing Violence, Volume 3: Social Influences. Washington, DC: The National Academies Press. doi: 10.17226/4421.
×
Page 79
Suggested Citation:"Violent Victimization and Offending: Individual-Situational-, and Community-Level Risk Factors." National Research Council. 1994. Understanding and Preventing Violence, Volume 3: Social Influences. Washington, DC: The National Academies Press. doi: 10.17226/4421.
×
Page 80
Suggested Citation:"Violent Victimization and Offending: Individual-Situational-, and Community-Level Risk Factors." National Research Council. 1994. Understanding and Preventing Violence, Volume 3: Social Influences. Washington, DC: The National Academies Press. doi: 10.17226/4421.
×
Page 81
Suggested Citation:"Violent Victimization and Offending: Individual-Situational-, and Community-Level Risk Factors." National Research Council. 1994. Understanding and Preventing Violence, Volume 3: Social Influences. Washington, DC: The National Academies Press. doi: 10.17226/4421.
×
Page 82
Suggested Citation:"Violent Victimization and Offending: Individual-Situational-, and Community-Level Risk Factors." National Research Council. 1994. Understanding and Preventing Violence, Volume 3: Social Influences. Washington, DC: The National Academies Press. doi: 10.17226/4421.
×
Page 83
Suggested Citation:"Violent Victimization and Offending: Individual-Situational-, and Community-Level Risk Factors." National Research Council. 1994. Understanding and Preventing Violence, Volume 3: Social Influences. Washington, DC: The National Academies Press. doi: 10.17226/4421.
×
Page 84
Suggested Citation:"Violent Victimization and Offending: Individual-Situational-, and Community-Level Risk Factors." National Research Council. 1994. Understanding and Preventing Violence, Volume 3: Social Influences. Washington, DC: The National Academies Press. doi: 10.17226/4421.
×
Page 85
Suggested Citation:"Violent Victimization and Offending: Individual-Situational-, and Community-Level Risk Factors." National Research Council. 1994. Understanding and Preventing Violence, Volume 3: Social Influences. Washington, DC: The National Academies Press. doi: 10.17226/4421.
×
Page 86
Suggested Citation:"Violent Victimization and Offending: Individual-Situational-, and Community-Level Risk Factors." National Research Council. 1994. Understanding and Preventing Violence, Volume 3: Social Influences. Washington, DC: The National Academies Press. doi: 10.17226/4421.
×
Page 87
Suggested Citation:"Violent Victimization and Offending: Individual-Situational-, and Community-Level Risk Factors." National Research Council. 1994. Understanding and Preventing Violence, Volume 3: Social Influences. Washington, DC: The National Academies Press. doi: 10.17226/4421.
×
Page 88
Suggested Citation:"Violent Victimization and Offending: Individual-Situational-, and Community-Level Risk Factors." National Research Council. 1994. Understanding and Preventing Violence, Volume 3: Social Influences. Washington, DC: The National Academies Press. doi: 10.17226/4421.
×
Page 89
Suggested Citation:"Violent Victimization and Offending: Individual-Situational-, and Community-Level Risk Factors." National Research Council. 1994. Understanding and Preventing Violence, Volume 3: Social Influences. Washington, DC: The National Academies Press. doi: 10.17226/4421.
×
Page 90
Suggested Citation:"Violent Victimization and Offending: Individual-Situational-, and Community-Level Risk Factors." National Research Council. 1994. Understanding and Preventing Violence, Volume 3: Social Influences. Washington, DC: The National Academies Press. doi: 10.17226/4421.
×
Page 91
Suggested Citation:"Violent Victimization and Offending: Individual-Situational-, and Community-Level Risk Factors." National Research Council. 1994. Understanding and Preventing Violence, Volume 3: Social Influences. Washington, DC: The National Academies Press. doi: 10.17226/4421.
×
Page 92
Suggested Citation:"Violent Victimization and Offending: Individual-Situational-, and Community-Level Risk Factors." National Research Council. 1994. Understanding and Preventing Violence, Volume 3: Social Influences. Washington, DC: The National Academies Press. doi: 10.17226/4421.
×
Page 93
Suggested Citation:"Violent Victimization and Offending: Individual-Situational-, and Community-Level Risk Factors." National Research Council. 1994. Understanding and Preventing Violence, Volume 3: Social Influences. Washington, DC: The National Academies Press. doi: 10.17226/4421.
×
Page 94
Suggested Citation:"Violent Victimization and Offending: Individual-Situational-, and Community-Level Risk Factors." National Research Council. 1994. Understanding and Preventing Violence, Volume 3: Social Influences. Washington, DC: The National Academies Press. doi: 10.17226/4421.
×
Page 95
Suggested Citation:"Violent Victimization and Offending: Individual-Situational-, and Community-Level Risk Factors." National Research Council. 1994. Understanding and Preventing Violence, Volume 3: Social Influences. Washington, DC: The National Academies Press. doi: 10.17226/4421.
×
Page 96
Suggested Citation:"Violent Victimization and Offending: Individual-Situational-, and Community-Level Risk Factors." National Research Council. 1994. Understanding and Preventing Violence, Volume 3: Social Influences. Washington, DC: The National Academies Press. doi: 10.17226/4421.
×
Page 97
Suggested Citation:"Violent Victimization and Offending: Individual-Situational-, and Community-Level Risk Factors." National Research Council. 1994. Understanding and Preventing Violence, Volume 3: Social Influences. Washington, DC: The National Academies Press. doi: 10.17226/4421.
×
Page 98
Suggested Citation:"Violent Victimization and Offending: Individual-Situational-, and Community-Level Risk Factors." National Research Council. 1994. Understanding and Preventing Violence, Volume 3: Social Influences. Washington, DC: The National Academies Press. doi: 10.17226/4421.
×
Page 99
Suggested Citation:"Violent Victimization and Offending: Individual-Situational-, and Community-Level Risk Factors." National Research Council. 1994. Understanding and Preventing Violence, Volume 3: Social Influences. Washington, DC: The National Academies Press. doi: 10.17226/4421.
×
Page 100
Suggested Citation:"Violent Victimization and Offending: Individual-Situational-, and Community-Level Risk Factors." National Research Council. 1994. Understanding and Preventing Violence, Volume 3: Social Influences. Washington, DC: The National Academies Press. doi: 10.17226/4421.
×
Page 101
Suggested Citation:"Violent Victimization and Offending: Individual-Situational-, and Community-Level Risk Factors." National Research Council. 1994. Understanding and Preventing Violence, Volume 3: Social Influences. Washington, DC: The National Academies Press. doi: 10.17226/4421.
×
Page 102
Suggested Citation:"Violent Victimization and Offending: Individual-Situational-, and Community-Level Risk Factors." National Research Council. 1994. Understanding and Preventing Violence, Volume 3: Social Influences. Washington, DC: The National Academies Press. doi: 10.17226/4421.
×
Page 103
Suggested Citation:"Violent Victimization and Offending: Individual-Situational-, and Community-Level Risk Factors." National Research Council. 1994. Understanding and Preventing Violence, Volume 3: Social Influences. Washington, DC: The National Academies Press. doi: 10.17226/4421.
×
Page 104
Suggested Citation:"Violent Victimization and Offending: Individual-Situational-, and Community-Level Risk Factors." National Research Council. 1994. Understanding and Preventing Violence, Volume 3: Social Influences. Washington, DC: The National Academies Press. doi: 10.17226/4421.
×
Page 105
Suggested Citation:"Violent Victimization and Offending: Individual-Situational-, and Community-Level Risk Factors." National Research Council. 1994. Understanding and Preventing Violence, Volume 3: Social Influences. Washington, DC: The National Academies Press. doi: 10.17226/4421.
×
Page 106
Suggested Citation:"Violent Victimization and Offending: Individual-Situational-, and Community-Level Risk Factors." National Research Council. 1994. Understanding and Preventing Violence, Volume 3: Social Influences. Washington, DC: The National Academies Press. doi: 10.17226/4421.
×
Page 107
Suggested Citation:"Violent Victimization and Offending: Individual-Situational-, and Community-Level Risk Factors." National Research Council. 1994. Understanding and Preventing Violence, Volume 3: Social Influences. Washington, DC: The National Academies Press. doi: 10.17226/4421.
×
Page 108
Suggested Citation:"Violent Victimization and Offending: Individual-Situational-, and Community-Level Risk Factors." National Research Council. 1994. Understanding and Preventing Violence, Volume 3: Social Influences. Washington, DC: The National Academies Press. doi: 10.17226/4421.
×
Page 109
Suggested Citation:"Violent Victimization and Offending: Individual-Situational-, and Community-Level Risk Factors." National Research Council. 1994. Understanding and Preventing Violence, Volume 3: Social Influences. Washington, DC: The National Academies Press. doi: 10.17226/4421.
×
Page 110
Suggested Citation:"Violent Victimization and Offending: Individual-Situational-, and Community-Level Risk Factors." National Research Council. 1994. Understanding and Preventing Violence, Volume 3: Social Influences. Washington, DC: The National Academies Press. doi: 10.17226/4421.
×
Page 111
Suggested Citation:"Violent Victimization and Offending: Individual-Situational-, and Community-Level Risk Factors." National Research Council. 1994. Understanding and Preventing Violence, Volume 3: Social Influences. Washington, DC: The National Academies Press. doi: 10.17226/4421.
×
Page 112
Suggested Citation:"Violent Victimization and Offending: Individual-Situational-, and Community-Level Risk Factors." National Research Council. 1994. Understanding and Preventing Violence, Volume 3: Social Influences. Washington, DC: The National Academies Press. doi: 10.17226/4421.
×
Page 113
Suggested Citation:"Violent Victimization and Offending: Individual-Situational-, and Community-Level Risk Factors." National Research Council. 1994. Understanding and Preventing Violence, Volume 3: Social Influences. Washington, DC: The National Academies Press. doi: 10.17226/4421.
×
Page 114
Next: Violence Between Spouses and Intimates: Physical Aggression Between Women and Men in Intimate Relationships »
Understanding and Preventing Violence, Volume 3: Social Influences Get This Book
×
Buy Paperback | $95.00
MyNAP members save 10% online.
Login or Register to save!
Download Free PDF

This volume examines social influences on violent events and violent behavior, particularly concentrating on how the risks of violent criminal offending and victimization are influenced by communities, social situations, and individuals; the role of spouses and intimates; the differences in violence levels between males and females; and the roles of psychoactive substances in violent events.

  1. ×

    Welcome to OpenBook!

    You're looking at OpenBook, NAP.edu's online reading room since 1999. Based on feedback from you, our users, we've made some improvements that make it easier than ever to read thousands of publications on our website.

    Do you want to take a quick tour of the OpenBook's features?

    No Thanks Take a Tour »
  2. ×

    Show this book's table of contents, where you can jump to any chapter by name.

    « Back Next »
  3. ×

    ...or use these buttons to go back to the previous chapter or skip to the next one.

    « Back Next »
  4. ×

    Jump up to the previous page or down to the next one. Also, you can type in a page number and press Enter to go directly to that page in the book.

    « Back Next »
  5. ×

    Switch between the Original Pages, where you can read the report as it appeared in print, and Text Pages for the web version, where you can highlight and search the text.

    « Back Next »
  6. ×

    To search the entire text of this book, type in your search term here and press Enter.

    « Back Next »
  7. ×

    Share a link to this book page on your preferred social network or via email.

    « Back Next »
  8. ×

    View our suggested citation for this chapter.

    « Back Next »
  9. ×

    Ready to take your reading offline? Click here to buy this book in print or download it as a free PDF, if available.

    « Back Next »
Stay Connected!