Findings

Total violence and stranger and nonstranger violence are examined to assess whether the sources of risk correlate with certain types of events. Adolescents are classified as victims of nonstranger violence if they experienced at least one incident of attempted or completed violence at the hands of an offender who was a friend, acquaintance, or family member.⁷ Adolescents are classified as victims of stranger violence if the offender was someone with whom they had no prior relationship.

To assess how community characteristics are related to violent victimization, events that occurred within a youth’s residential community are examined separately from those that occurred elsewhere. This distinction is made because NCVS data contain location information on the residential communities of each youth, but not all incidents of violence take place within an adolescent’s neighborhood. Although the majority of violent events occur within a youth’s neighborhood (defined here as within 1 mile of his or her home), a substantial minority of violent events (47.0 percent) occur more than 1 mile from home (see table 1). Therefore, the analyses of how community factors are related to risk will focus primarily on events that occur near the victim’s home.

Relationship Between Risk Factors and Family Types

The risks for total, stranger, and nonstranger violence among youth living in single-parent/other and two-parent families are presented in table 2. Youth in single-parent families experience significantly higher risks for violence than youth in two-parent families. Approximately 60 out of every 1,000 children in single-parent families reported at least one violent victimization during a 6-month period, whereas approximately 40 out of every 1,000 children in two-parent families were victimized. In other words, the overall risk for violence is about 50 percent higher among youth living in single-parent families than among youth living in two-parent families. The difference in risks for neighborhood violence is even more pronounced: youth in single-parent families are about twice as likely as youth in two-parent families to become a victim of violence in their own neighborhood (40.8 versus 19.9 per 1,000).

Moreover, these levels of risk are significantly higher than those found for most Americans. In 1995, the 6-month risk for violence among all Americans (ages 12 and older) was approximately 18 per 1,000 and the risk for neighborhood violence was roughly 9 per 1,000 (Lauritsen, 2001). Thus, children living in single-parent families have an overall risk for violent victimization that is about three times higher than the average American (60 versus 18 per 1,000).

The same differences are found when stranger and nonstranger violence are examined separately. For youth in single-parent families, the risks for stranger and nonstranger violence are about 50 percent greater than for youth in two-parent families, and the differences in risk for these events in their own neighborhood are about twice as great. Similarly, adolescents in single-parent families are about three times more likely to experience stranger violence and three times more likely to experience nonstranger violence than the average American.

To understand why youth in single-parent families experience greater risk for violence than youth in two-parent families, it is important to consider how family circumstances determine other factors in the lives of youth. Because NCVS data are representative of U.S. adolescents and have been linked to census information, the data reveal a great deal about the differences in the family and community resources available to these two groups of children. Table 3 shows how various factors differ according to family type.

Many factors are related to the types of families in which youth live. For instance, adolescents in single-parent families tend to live in households that have significantly less income than two-parent families. The typical income of single-parent families is roughly half that of two-parent families.⁸ Single-parent families also tend to have less residential stability—that is, they are more likely to have moved recently and have lived in their current home for significantly shorter periods of time than two-parent families. Research shows that the length of time individuals have resided in their home is related to the risk for victimization (Bureau of Justice Statistics, 2000; Lauritsen, 2001). (See below for a discussion of how family residential stability is also an important predictor of the risk for youth victimization.)

Because residential housing often is economically segregated, communities surrounding single-parent families tend to be more economically disadvantaged than those surrounding two-parent families. On average, youth in single-parent families live in areas with significantly higher rates of poverty and a greater proportion of other single-parent families. Compared with youth living in two-parent families, youth in single-parent families live in areas where the proportion of persons living below the poverty line is about 56 percent greater (18.1 percent versus 11.6 percent) and the proportion of female-headed families is about 58 percent higher (27.4 percent versus 17.3 percent). Compared with youth living in two-parent families, youth in single-parent families are nearly twice as likely to be living in central-city areas (39.7 percent versus 21.7 percent) and in places with a greater percentage of black residents (24.0 percent versus 10.6 percent). These patterns emphasize the importance of considering community characteristics as a potential explanation for why children in different types of families experience varying levels of violence.

Relationship Between Risk Factors and Community Types

Figure 1 illustrates how adolescents’ risk for neighborhood violence is related to the overall level of socioeconomic disadvantage in their community. At the 50th percentile, approximately 26 of every 1,000 youth were victimized by violence. The shape of the curve in figure 1 is essentially the same when stranger and nonstranger violence are examined separately. The community type in which the average white, Latino, and black youth live is also noted in the figure and demonstrates, for example, that black youth tend to live in areas that are more disadvantaged than Latino and white youth.

Figure 1 shows that community disadvantage does not have a substantial influence on violence risk until the 80th percentile. For most youth (about 80 percent of the adolescent population), community characteristics are unlikely to account for the differences in their risk for violence. However, for the 20 percent of youth in the least advantaged communities, the risk for violence is much higher. In these areas, the likelihood that an adolescent will become a victim of violence is significantly higher than for youth living outside these communities.⁹

Figure 1: Risk of Youth Victimization, by Level of Community Disadvantage * Community disadvantage is measured by a summary index that describes the relative level of socioeconomic disadvantage in an area. The numbers on this scale represent the percentile or ranking of the community according to this index for all census tracts in the nation. Source: Author’s own data file combining information from the 1995 area-identified National Crime Victimization Survey with census tract information from the 1990 decennial census. Original data collected by the U.S. Census Bureau.

The 10 percent of youth who live in the most disadvantaged communities experience the highest risks of neighborhood violence in the country. In these areas, nearly 58 of every 1,000 youth reported at least one incident of violence during a 6-month period—a level that is almost twice as high as that for adults living in the same areas (Lauritsen, 2001) and approximately 6 times greater than the risk for neighborhood violence among all Americans (58 versus 9 per 1,000). Although age is a well-known correlate of risk, the combination of age and community disadvantage is associated with some of the highest levels of nonlethal violent victimization in the country.

As noted earlier, the general index of community disadvantage is one way of describing the economic and social composition of a neighborhood, but what components of this disadvantage index are most responsible for the higher levels of risk? Ascertaining a clear answer to this question is difficult because most components of disadvantage are interrelated. For example, areas with high proportions of female-headed households with children tend to have higher rates of poverty. Nonetheless, statistical techniques provide some insight into this important issue.

To determine which components of the disadvantage index are most responsible for higher levels of risk, the relationships between victimization and each of the components of disadvantage (i.e., poverty, unemployment, public assistance, race, ethnicity, and family composition) were examined in a series of analyses. Of these factors, poverty, family composition (i.e., female-headed households with children), and race (e.g., the percent of persons who are black) were most strongly associated with an adolescent’s risk for violent victimization. In urban samples, these characteristics tend to be extremely difficult to separate because many U.S. cities are segregated along economic and racial lines (see, for example, Sampson, Raudenbush, and Earls, 1997). NCVS, however, includes a representative sample of persons and places outside central cities where the statistical problems associated with separating these aspects of communities are reduced. In this sample, poverty, family composition, and race were examined simultaneously. The results of these analyses are presented in table 4.

These analyses reveal that one of the most important community predictors of youth risk for violence is the proportion of female-headed households with children.

Results show that the other important correlates of risk (i.e., levels of poverty and race) are not statistically significant when family composition is taken into account. Moreover, the results show that the effect of the family composition coefficient does not change significantly once racial composition and poverty levels are included in the analysis. This evidence suggests that youth face a higher risk for victimization in disadvantaged areas because these places contain greater proportions of children living in single-parent families—not because they tend to be poorer or have larger percentages of racial and ethnic minorities.¹⁰

The patterns described above show that an adolescent’s family type is related to risk and that certain characteristics of the community also are associated with victimization. Figure 2 examines whether community disadvantage has a similar influence on victimization among youth in single- and two-parent families.

Figure 2 reveals an important connection between families and communities that is not apparent when these factors are examined individually. When community factors become increasingly important, they do so primarily for youth living in single-parent families. Compared with youth who live in single-parent families, adolescents in two-parent families appear to be much better protected from the consequences of living in the most disadvantaged areas. In the most disadvantaged areas of the United States, approximately 66 percent of youth live in single-parent/other families, compared with about 28 percent in the rest of the country. In areas characterized by the highest socioeconomic disadvantage, the ability of families to monitor and supervise children’s activities is particularly important. Under conditions of serious community disadvantage, youth in single-parent families experience much greater levels of victimization risk than youth living in single-parent families in more advantaged places or youth in two-parent families, regardless of area of residence.

Figure 2: Risk of Youth Victimization, by Level of Community Disadvantage and Family Type * Community disadvantage is measured by a summary index that describes the relative level of socioeconomic disadvantage in an area. The numbers on this scale represent the percentile or ranking of the community according to this index for all census tracts in the nation. Source: Author’s own data file combining information from the 1995 area-identified National Crime Victimization Survey with census tract information from the 1990 decennial census. Original data collected by the U.S. Census Bureau.

Relationship Between Risk Factors and Racial and Ethnic Groups

Because family and community characteristics vary among racial and ethnic groups in the United States, it is important to consider differences in victimization risk across racial and ethnic groups. With this in mind, NCVS race and ethnicity items were combined to create three major racial and ethnic groups. Table 5 presents the risks for adolescent victimization for black, white, and Latino youth by gender and type of violence.

In this Bulletin, “black” refers to non-Latino blacks, “white” refers to non-Latino whites, and “Latino” refers to respondents who identify themselves as Hispanic regardless of race. Although Hispanics may be of any race, in 1995 approximately 90 percent of Hispanics reported that they were white, 6 percent reported that they were black, and 4 percent reported that they were another racial category (primarily American Indian). Youth were classified into these three groups (black, white, and Latino) for two reasons. First, these designations represent popular understandings of the nation’s major racial and ethnic groups. Second, without this classification, traditional “white versus black” comparisons result in overestimated risks among whites because Hispanics are most often classified as white. Similarly, the traditional “Hispanic versus non-Hispanic” comparison underestimates group differences because non-Hispanic whites and blacks are combined in the “non-Hispanic” category. Using the classification described here, approximately 16 percent of youth are black, 13 percent are Latino, and 71 percent are white.

Table 5 shows that the risk for total, stranger, and nonstranger violence is significantly higher among young males than it is among young females. Beyond this broad statement, however, the data show a rather complex pattern of differences among youth according to race, ethnicity, and gender. Black, white, and Latino youth, regardless of sex, show roughly equal risks for nonstranger violence. Differences emerge primarily in the category of stranger violence. Compared with black and Latino adolescents, white youth face a lower risk of stranger violence in their own neighborhood or community. (Racial and ethnic differences are explored further in the remaining analyses of individual, family, and community risk factors.)

Individual, Family, and Community Risk Factors

Research has shown that several factors are related to the risk for violent victimization among youth. Because some of these factors are interrelated, it is important to study them simultaneously to determine which are most strongly associated with violence and which are associated with violence only because they are related to other, more important factors. This type of analysis has been difficult because data on individual, family, and community factors are rare. This study’s final set of results describes how these factors are related to the risk for victimization among U.S. youth when the factors are examined simultaneously. Separate findings are reported for total violence and for stranger and nonstranger violence because somewhat different patterns were found, depending on the victim-offender relationship. A summary of the findings appears in table 6.

Individual risk factors. The individual factors examined include age, sex, race and ethnicity, and time spent at home in an average week. Data show that violent victimization in general and violent victimization in one’s neighborhood are significantly related to age, sex, and time spent at home. In general, older adolescents face a lower risk for victimization than younger adolescents, and males report more violence than females. In addition, the more evenings that adolescents spend at home in a typical week, the less violence they are likely to experience. However, violence risk is unrelated to race or ethnicity once family and community factors are controlled.

Individual risk factors vary somewhat if prediction is focused on stranger versus nonstranger events. Age is significantly associated with nonstranger violence but not with stranger violence. That is, younger and older adolescents experience similar levels of stranger violence, but older youth suffer less nonstranger violence. Findings show important declines in nonstranger violence for the 12–17 age range, which suggests that physical, cognitive, and emotional maturation help to make youth less vulnerable to violence by nonstrangers. Likewise, the persistent level of stranger violence risk may simply reflect the fact that contact with strangers increases during adolescence. It may also indicate that developmental processes do not offer increased protection for this type of event.¹¹

Gender differences associated with violence risk exist, but they depend on the type of violence. Male youth experience significantly higher levels of stranger violence, whereas male and female adolescents report similar levels of nonstranger violence. Time spent at home in the evenings—a commonly used indicator of a person’s lifestyle—is generally associated with lower levels of all forms of youth violence, except for nonstranger events in the neighborhood. This exception is likely a consequence of the fact that a larger proportion of nonstranger victimization takes place in or near the home. Overall, this behavioral measure had one of the strongest relationships to violence risk: youth who spent more evenings at home in a typical week were least likely to be victims of stranger and nonstranger violence.

Although white youth generally experience lower levels of nonlethal violence than minority youth, these differences disappear once family and community factors are taken into account. This result mirrors findings from recent analyses of adults (Lauritsen and White, 2001) and is important because it shows that racial and ethnic differences in nonlethal violence among youth are primarily a reflection of community and family differences, rather than the result of being part of a particular racial or ethnic group. The lack of significant differences across groups also is important because it suggests that the sources of risk are similar for all adolescents, regardless of their race or ethnicity. In additional analyses (not shown), black, white, and Latino youth were examined separately to assess potential differences and similarities. These results found no significant variation in the individual, family, and community risk factors for violent victimization across the three groups.

In other words, family and community characteristics serve as important sources of risk and protection, and none of the three groups exhibited unique risk factors for violent victimization.

Family risk factors. Variations in household income and size and length of residence in the current home were examined alongside family type to determine their possible influence on violence. Of these factors, neither income nor size was associated with risk once other individual, family, and community factors were considered. The fact that household income does not have a direct association with risk suggests that youth in single-parent families are not at higher risk because their own families are poor. Rather, they are at higher risk because they are more likely to be living in areas of greater socioeconomic disadvantage—areas with higher concentrations of single-parent families and young persons.

Earlier it was noted that youth in single-parent families are at higher risk for stranger and nonstranger violence and that these differences are especially pronounced in the most disadvantaged communities. However, when other individual and community factors are taken into account, a noteworthy distinction emerges between stranger and nonstranger victimization risk patterns. Differences in nonstranger violence across family types remain associated with the level of disadvantage in an area, but differences in stranger violence do not appear to be contingent on the characteristics of the community. Similar research using additional years of data is necessary to determine whether family type is consistently related to violence risk or whether stranger and nonstranger violence pose unique challenges, depending on family type.

Although family type remains related to violence risk when other factors are considered, the family characteristic found to have the strongest influence on risk was length of residence in the current home. Length of residence is related to stranger and nonstranger victimization risk, even among adults.¹² Youth who have been living in their current home for longer periods of time are less likely to be victimized by stranger or nonstranger violence. One interpretation of this finding is that children who have lived in their homes longer are more likely to be familiar with their communities and are better able to know whether certain locations or persons in the area are safe or dangerous. It is also possible that youth whose families often move have other risk factors not considered here (e.g., recent school or family disruptions and involvement in delinquency). Additionally, this finding may reflect the fact that it takes time for children and their families to develop ties and make friends in new neighborhoods, and that without established neighborhood ties, these “new kids on the block” are less likely to be protected by others.

Community risk factors. In general, youth living in the most disadvantaged areas experience the highest risks for violence, and the most important element of community disadvantage is the percent of female-headed households with children. This statement is true even when other individual, family, and community factors are controlled. However, another community characteristic has a consistent and independent effect on the risk for youth victimization: the percentage of persons in the area who are younger than 18 years old. Adolescents who live in areas with high proportions of young people are more likely to be victimized by violence—especially nonstranger violence. This relationship may not be surprising because violent offending is more prevalent among adolescents than adults and persons tend to associate with others of similar age. However, the proportion of youth in an area does not affect the risk for violent victimization among adults (Lauritsen, 2001). Rather than simply reflecting the number of potential offenders, the proportion of youth in an area is likely to represent the challenges that families and communities face when relatively few adults are available to supervise and protect large numbers of youth.

How Families and Communities Influence Youth Victimization	OJJDP Bulletin • November 2003