Journal of Family Violence

, Volume 26, Issue 5, pp 331–342

My Brother’s Keeper: A Contemporary Examination of Reported Sibling Violence Using National Level Data, 2000–2005

Authors

  • Jessie L. Krienert
    • Department of Criminal Justice SciencesIllinois State University
    • Department of Criminal Justice SciencesIllinois State University
    • Department of Criminal Justice SciencesIllinois State University
Original Article

DOI: 10.1007/s10896-011-9367-3

Cite this article as:
Krienert, J.L. & Walsh, J.A. J Fam Viol (2011) 26: 331. doi:10.1007/s10896-011-9367-3

Abstract

Identified as a social problem in 1980, sibling violence has been labeled the most common and least researched form of family violence in the United States (Eriksen and Jensen 2006, 2008). Extant research has limitations including definitional inconsistencies, overreliance on small retrospective clinical samples, and limited use of officially reported national level data for profiles of victims and offenders. Although often trivialized as a “normal” part of growing up, sibling violence has links to an array of complications manifesting later in life including physical and emotional disorders, school bullying, substance abuse, and domestic violence. This work draws on 6 years of National Incident-Based Reporting System (NIBRS) data (2000–2005) (n = 33,066) to provide the most comprehensive source of baseline information on this understudied form of intrafamilial violence. The research explores demographic and incident characteristics extending the knowledge beyond typical victim descriptives to incorporate offender profiles and incident level information including the type of violence/victimization, substance use, weapon use, and degree of injury sustained. Findings, in part, suggest several gender based victim and offender differences with female siblings involved in more serious injury incidents than their male sibling counterparts.

Keywords

Sibling abuseSibling violenceFamily violenceNIBRS

Likely the most prevalent though least studied form of family violence (see Eriksen and Jensen 2006), more than 29 million American children engage in sibling violence each year (Gelles and Straus 1988) with the rate of sibling violence higher than any other form of family violence (Hoffman et al. 2005). As Hoffman and Edwards (2004a) noted, sibling violence, “constitutes a pandemic form of victimization of children, with the symptoms often going unrecognized and the effect ignored” (p. 187). Historically, sibling violence has not been recognized by the criminal justice system (Eriksen and Jensen 2006). Rather, the violence has routinely been considered a normal part of growing up, even a rite-of-passage. Though often viewed through the lens of sibling rivalry, sibling violence is detrimental to psychosocial development and has been linked to a wide variety of adverse child and adult outcomes including antisocial behavior, domestic violence, dating violence, substance abuse, eating disorders, anxiety, and school bullying (Eriksen and Jensen 2006, 2008). Negative sibling relationships may be one of the most prevalent, long lasting, and damaging forms of violence (Hardy 2001); however, despite its prevalence it remains drastically understudied.

Often relabeled or reinterpreted as conflict or sibling rivalry, sibling violence is rarely identified as a serious form of family violence (Raffaelli 1992). Hitting among siblings is so common that few families/parents recognize it as a deviant or criminal behavior. In fact, some parents believe sibling aggression is not only natural and expected, but instills a helpful lesson in conflict management for instances later in life when conflict presents outside the family environment (Gelles and Cornell 1985). Parents often minimize sibling violence, even when the behavior, were it not committed by a sibling, would without question meet the elements of an officially recognized crime.

To provide a bit of context to the problem, Straus and Gelles (1990) using data from married couples with two or more children (n = 733) from the 1975 National Family Violence Survey, calculated the rate of sibling violence to be 800 per 1,000 families or 80% compared to the rate of child abuse which was calculated at only 23 per 1,000 families or 23%. The need for additional empirical study is evident based on both anecdotal and empirical evidence regarding the prevalence and scope of sibling violence and, equally important, both the short and long term psychosocial consequences that have been identified (see Eriksen and Jensen 2006, 2008). As Eriksen and Jensen (2008, p. 2) noted, “…few subsequent studies…have directly addressed the extent, dynamics, correlates, or effects of sibling violence”. Further, a review of the dearth of research in the area reveals virtually every empirical article written on the topic lamenting the inadequate study of the phenomenon and highlighting the need for additional information about both victims and offenders (Noland et al. 2004). As Goodwin and Roscoe (1990) note, “[b]aseline data simply do not exist because only a few studies have been conducted. Consequently, awareness of sibling aggression as a serious form of family violence remains at a very low level” (p. 452). Further, “to date, researchers of sibling abuse, both sexual and physical, have relied upon clinical case studies and small sample sizes” (Hardy 2001, p. 258).

The present work contributes to the knowledge about sibling violence addressing several of the most persistent and prevalent limitations in the extant work including definitional inconsistency, reliance on small clinical samples, and with few exceptions, the limited use of large national samples of data, and the absence of victim and offender profiles and incident characteristics generated from aggregate data. These limitations are addressed by: 1) employing a consistent, concise, and universally accepted criminal justice/legal-oriented definition of sibling violence that serves to “demarcate” differences in sibling violence by severity (Eriksen and Jensen 2008, p. 4) through the use of officially recognized categories of assault; 2) using a clear and concise measure of sibling violence drawing on officially reported incidents recorded over six years (2000–2005) by the National Incident-Based Reporting System (NIBRS), a large data source that to the best of our knowledge has not been used in prior studies of sibling violence; and 3) using NIBRS data to facilitate an examination of the correlates and connections between age, race, and gender, and their relationship to sibling violence. In this last instance, NIBRS data is being utilized in an effort to dispel previous contradictory findings, and develop needed baseline profiles taking into account available victim, offender, and incident characteristics.

The present work begins with a brief review of the literature with specific emphasis on the prevalence of sibling violence and victim-offender characteristics. The work then highlights consequences related to psychosocial development and long term impacts followed by linkage to the cycle of violence hypothesis. Next, methodology and data collection are discussed followed by the analysis plan. Finally, the results are explained with the concluding discussion addressing strengths, limitations, and areas for future inquiry.

A Brief Review of the Literature on Sibling Violence

Family violence has long been synonymous with intimate partner violence or child abuse, rarely focusing on other intrafamilial victim-offender relationships including sibling violence (Walsh and Krienert 2007). Limited financial support along with limited data availability due to underreporting has resulted in a lack of empirical research, leaving even basic information about sibling violence still unknown. “Few experts agree on how extensive sibling abuse is or where sibling conflict ends and abuse begins. It is rarely studied: only two major national studies, a handful of academic papers and a few specialized books have looked at it in the last quarter-century” (Butler 2006, p. 1).

Definitional Inconsistency in the Study of Sibling Violence

The foremost recurring limitation in the extant work is the lack of definitional congruence across disciplines and across studies within disciplines leading to both inconsistent and broad interpretation of what distinguishes “sibling rivalry” from criminal violence. Stock (1993) identified this persistent problem some 15 years earlier noting that, “the main problem…is the lack of a definition of the term “sibling abuse”” (Stock 1993, p. 20). Different studies may label the behavior as aggression, violence, conflict, or abuse, however, rarely are they labeled as criminal (Eriksen and Jensen 2008) and rarely do they include a legal definition of assault. The muddled conglomeration of varying definitions leaves little opportunity for generalizations across studies. Additionally, the historically “hands off” approach from the criminal justice system has further muddied the definitional waters with the inconsistent application of legal consequences. Not all incidents progress through the official channels of the justice system and when they do, they are more likely to be given an informal resolution, in many instances undermining their severity and impact. Further complicating the definitional issue has been the use of narrow and limited age categories in prior works. For example, most clinical studies focus on young children, while many retrospective school-based examinations ignore the age of siblings at time of incident all together, instead recording their age at time of study.

Adding to the confusion, there are no specific laws criminalizing sibling violence beyond traditional laws prohibiting assaultive behavior, making cases of sibling violence difficult to discern in official records. In fact, cases rarely result in legal intervention without parental filing (Stock 1993). Unfortunately, violence between siblings is frequently minimized by those with the ability to report the behavior. Parents often believe that some level of sibling aggression is needed to develop the appropriate social and behavioral boundaries needed later in life; causing reinterpretation of behaviors that would fit the legal definition of assault in any other setting. As a form of conflict resolution, many feel siblings provide the needed training ground for learning how to define regulative rules for conflict and anger management (Goodwin and Roscoe 1990) through the provision of a healthy and reasonable way to reinforce rules and learn personal boundaries (Raffaelli 1992). Arguably, it is suggested that even minor forms of sibling conflict can help children learn and practice negotiation strategies and appropriate rules of behavior including positive social skills (Stormshak et al. 1996). However, the line between sibling conflict as an appropriate regulatory learning tool and one that causes irreparable harm is unclear.

Compared to other forms of childhood victimization, sibling violence is less concerning in the eyes of both parents and authorities. Not only is sibling violence considered natural, social norms actually encourage its occurrence. Typically considered a “normal” part of growing up, those that would be involved in official reporting may find it difficult to determine when a behavior bridges “normal” and becomes criminal. In fact, siblings themselves may not label sibling abuse as violence (Hardy 2001).

The regularity and frequency in occurrence often leads both parents and professionals to discount the harm in sibling violence, even when the empirical evidence supports wide ranging negative outcomes (see Bank et al. 1996; Eriksen and Jensen 2008; Garcia et al. 2000; Gully et al. 1981; Simonelli et al. 2002; Stocker 1994; Stocker and Burwell 2002; Stormshak et al. 1996; Wiehe 1997). Victims of sibling violence are also reluctant to characterize their victimization as abuse. Hardy (2001) examined responses from a retrospective study of 203 undergraduate students and despite nearly 50% of the students reporting behaviors that fit the definition of physical aggression by a sibling, fewer than 10% considered the behavior as abusive at the time. The perception of sibling violence as a noncriminal act is, in part, a contributing factor to the limited official and unofficial victim, offender and incident level data.

Incidence and Prevalence of Sibling Violence

The dearth of sibling violence research in the extant literature addressing various forms of intra-familial violence makes it difficult to assess the prevalence and scope of the offense. With very few large-scale data sets available, and even fewer comprising officially reported data collected at the national level, prior studies have commonly utilized small clinical and/or retrospective samples of adolescents. Eriksen and Jensens’ (2006, 2008) examination of a subsample (n = 994) of Straus and Gelles’ (1976) National Survey of Physical Violence in American Families and Finkelhor et al.’s (2006) national probability telephone sample of sibling violence remain the only two relatively large-scale national studies of sibling violence to date. Additionally, large samples are frequently out of date, or were initially conducted for a different purpose. Sampling techniques and sample sizes are two pressing problems in the prior literature making victim and offender profiles elusive. Furthermore, a lack of statistical control for other relevant causal factors makes it difficult to make appropriate comparisons (Stock 1993).

With very little attention paid to sibling violence, what is available is often directed towards preschool-age or elementary school children and isolated as an anomalous precursor of future violence and/or antisocial behavior. In a 1980 national examination of 733 families, (Straus et al. 1980) 82% of families reported sibling violence, noting an average of 21 incidents per year. While the majority of cases (74%) were fairly minor, involving only pushing or shoving, 16% of the sample reported threats with a knife or gun.

Although not consistently defined, existing studies report the incidence of sibling violence ranging from infrequent to often. Measuring sibling injury as an indicator of violence in a retrospective sample of Canadian undergraduate students, DeKeseredy and Ellis (1997) reported nearly 50% of students indicating prior physical injury caused by siblings. Duncan (1999) found 22% of her middle school sample reported being hit or pushed around either often or very often, with 8% beaten up by a sibling often or very often (p. 877). Utilizing a nationally representative phone survey, Finkelhor et al. (2006) found that children were more likely to be hit by a sibling in the last year (35%) than by a peer (20%). Additionally, they noted a common problem plaguing past family violence studies; children are unlikely to report sibling violence unless directly asked about its occurrence.

Employing a slightly broader definition including acts of aggression, Roscoe et al. (1987) found closer to 100% (98% of males and 94% of females) of their high school sample reported prior physical aggression with admission of victimization by a sibling in the year preceding their study. Comparatively, Goodwin and Roscoe (1990), reporting results from an anonymous questionnaire of 272 high school students, found 65% of males and females reported victimization and offending within the past year, with pushing, shoving, or hitting as the most common form of violence.

Offender and Victim Characteristics

Consistent profiles of both offenders and victims are frequently absent or underdeveloped in prior sibling violence research due in part to small samples. Further, age, gender, race, and incident characteristics are often limited by methodological differences across studies. Differences in data collection techniques and sample sizes along with definitional inconsistencies, and a general tendency to underreport incidents make comparative analysis exceedingly difficult. The present study of a large number of reported cases at the national level, while presumably still underreported and underrepresentative, will help to illuminate elusive victim and offender profiles and shed new light on cases that law enforcement agencies in reporting jurisdictions across the country are recording as the crime of assault where the victim-offender relationship is sibling (see Finkelhor and Ormrod 2000).

Age

Age differences are confounding throughout much of the extant literature. Samples routinely focus on preschool age children, often ignoring the violence that exists between older siblings. The most frequent age related finding has been that sibling violence declines with age (DeKeseredy and Ellis 1997; Eriksen and Jensen 2006, 2008; Straus et al. 1980), though others have found no age difference related to ones likelihood of offending (Abramovitch et al. 1982). Siblings are also more likely to fight with those closest in age, with Felson and Russo (1988) finding that aggression was most frequent when the sibling age difference was three years or less, hypothesizing that jealousy and time together were likely culprits when sibling ages were close.

Contradicting the seemingly inverse relationship between offending and age, some findings indicate that although incidence may decrease with age, injury and offense severity/seriousness actually increase. Finkelhor et al. (2006) found that the most common age for sibling victimization was the 6–9 year old category, however, cases involving injuries and weapons were more likely among those age 14–17. Alternatively, Wiehe (1997) makes note of a cascading effect where the oldest sibling targets the next oldest child, who then targets the next oldest on down the line in an age-based hierarchy. Older siblings may have less of a need to resort to violence to maintain compliance, instead exerting control through increased social skills, or a lengthy history of past violence/intimidation (Hoffman and Edwards 2004b) Additionally, gender may play a role in observed age differences. There is evidence that the frequency of sibling offending for girls declines at a slower rate than boys (Eriksen and Jensen 2008).

Gender

Although little consistency is found across samples relating to gender and sibling offending, past research has generally found a higher prevalence of male sibling violence offenders (Eriksen and Jensen 2008). Corroborating earlier findings (Felson 1983; Minnett et al. 1983), Duncan (1999) found in a sample of middle school children, that both males and females were equally likely to report bullying behaviors by a sibling. Graham-Bermann et al. (1994) reported that males reported higher levels of perpetration. Similarly, Hoffman and Edwards (2004b) found the most at-risk dyad for sibling abuse to include younger sisters with older brothers. As well, in a clinical sample of same sex and mixed sex sibling pairs, Pepler et al. (1981) observed that mixed sex pairs had increased frequency of aggression over time. Alternately, Stock (1993) notes more frequent aggression occurring with siblings of the same sex.

Injury

Although many studies find less injury involved with sibling violence when compared to other cases of adolescent violence, this could also be an artifact of the way data are collected. It is anticipated that relatively few cases of sibling violence find their way into official records. Kettrey and Emery (2006) reported more than 70% of their college student convenience sample had experienced or perpetrated severe physical violence upon a sibling. However, a large portion did not recognize their behavior as violence. In a comparison between sibling violence and peer violence, Finkelhor et al. (2006) found fewer sibling injuries (13% compared to 33%); however, sibling abuse was more likely to be a chronic feature of youth, with 40% of their sample reporting five or more incidents of sibling violence in the past year compared to only 15% of chronic peer violence incidents. Gender differences in injury have also been reported, with more males than females reporting assaults with objects (Goodwin and Roscoe 1990). Age differences can also be linked to injury. Increased frequency of offending coupled with decreased injury is most common when siblings are close in age (Raffaelli 1992).

Adverse Immediate and Long-Term Consequences of Sibling Violence

Linked to negative outcomes in both childhood and adulthood, sibling violence and victimization has been associated with severe behavioral and psychosocial problems including antisocial behavior and bullying (Eriksen and Jensen 2008). Additionally, the effects are thought to be long-lasting, with research pointing to increased prevalence of eating disorders, drug and alcohol use, anxiety, and even domestic violence (Eriksen and Jensen 2008; Gully et al. 1981; Wiehe 1997). Simonelli et al. (2002), in a retrospective sample of 120 college undergraduates, found that sibling violence was related to dating violence for both males and females. Wiehe (1997) makes note of the long term psychological effects of sibling violence including lower self-esteem, depression, and future substance abuse. Similarly, Stocker (1994) found what she termed “sibling conflict” associated with depression, loneliness and low self-esteem in an elementary school sample. Using a longitudinal prospective design, Stocker and Burwell (2002) examined 136 adolescents and their younger siblings. Sibling conflict was a positive predictor of increased anxiety, depression and delinquent behavior. Both Stormshak et al. (1996) and Garcia et al. (2000) linked sibling violence to increased school problems, aggressive peer behavior, and/or delinquency. Similarly, Bank et al. (1996) extending the argument into adult criminality, found links to arrests for males who were involved in sibling violence during childhood. In a more dramatic finding, nearly 10% of family homicides involve sibling murders, perhaps signifying a link not only to long-term negative psychosocial behavior, but to lethal violence as well (Dawson and Langan 1994). As the empirical evidence supports, the link between sibling violence and adverse outcomes is fairly well documented in the extant literature.

Sibling Violence and the Cycle of Violence

A well developed line of research exists examining the cycle of violence in relation to intimate partner violence, both witnessed and experienced, in the home. Extending this argument to sibling violence, children exposed to violence in the home as either victim or perpetrator learn that violence is an acceptable option in dispute resolution (Eriksen and Jensen 2006). Violence in the home, whether spousal, child, or sibling can teach negative regulative rules of behavior creating a greater likelihood of replication of that violence at future points in time and in future relationships.

Corroborating this line of research, Eriksen and Jensen (2006, 2008) found a link between child behavioral problems and marital problems. They note that the most important indicator of sibling violence was the occurrence of physical violence by a parent. Using a retrospective study of 216 undergraduate students, Gully et al. (1981) theorize sibling violence could be more predictive of adult violence as it involves actual commission not just observation, adopting a ‘practice makes perfect’ mentality. Siblings provide a training ground for later behavior, and while many approve of this as a positive step in the development of normative rules of conduct, it also can lead to the development of abnormal and antisocial forms of conduct. Simonelli et al. (2002) found that those involved in sibling abuse were not also abused by parents, sibling violence could in fact be a separate track of familial violence, which would then be a separate contributor to the cycle of violence. Sibling violence teaches escalation and the pattern of attack and counterattack. It serves to reinforce aggressive behavior that can then be transferred in future interactions, first with peers and later replicated in other intimate partner relationships. Further, “it has been suggested that violence among siblings may be a better predictor of later adult violence than observing violence between parents” (Noland et al. 2004, p. S14).

Methods

The present work draws on 6 years (2000–2005) of aggregate national level sibling violence data reported to participating NIBRS jurisdictions. Comprehensive victim, offender, and incident characteristics are used to expand existing research in an effort to establish baseline knowledge and victim/offender profiles and incident characteristics from a more comprehensive sample of data than has previously existed.

Data Source and Selection

Sibling violence data was extracted from the National Incident-Based Reporting System (NIBRS) developed by the Federal Bureau of Investigation. The data represent reported cases for the years 2000–2005 and include incident-level information for each crime reported to the FBI by law enforcement agencies in participating NIBRS jurisdictions.1 Although not a representative sample, the amount of available data—6 years—and the level of detail available provides a distinct advantage in analyzing the offender, victim, and event-based characteristics of sibling violence given the previous limited availability and subsequent examination of aggregate level official reports.

NIBRS data is unique in that it provides the ability to link criminal incident characteristics with all associated individual level offender and victimization data. The data is structured to allow the researcher to choose from one of several units of analysis including offender, victim, and event/incident. All incidents involving siblings as the victim-offender relationship age 21 and younger reported in the 2000–2005 NIBRS data were selected for initial inclusion in the current dataset. Due to constraints in the NIBRS data, and to maintain statistical assumptions of independence and minimize measurement error, data were further limited to incidents with a single victim and single offender (Bachman et al. 1998; Goetting 1992; Regoeczi et al. 2008).

The data selection criteria for sibling violence, adhering to a criminal justice oriented legal definition, intentionally limited selected cases to aggravated assault,2 simple assault,3 and intimidation4 perpetrated against a sibling victim. The objective was to implement a standardized/universal criminal justice/legal oriented definition of sibling violence to increase external validity, decrease the occasionally subjective definitional variation employed in past works, and allow replication.

Victim characteristics

For the years 2000–2005 there were 33,066 incidents of sibling violence reported to law enforcement with a 1:1 victim/offender relationship and meeting the 21 and younger age threshold. The sample consisted of primarily white victims (n = 21,921; 67%) and abusers (n = 22,018; 67%) with most victims falling in the 14 and older age category (n = 24,535; 74%) with a mean age of 15 years old. Females made up the majority of victims accounting for 58% of the sample.

Offender characteristics

Descriptive statistics were utilized to assess demographic and offense characteristics for male and female offenders. Table 1 shows that males accounted for 24,263 offenders of sibling violence, or 73%; females accounted for 8,790, or 27% of offenders of sibling violence in the NIBRS data. Similar to victim age, the largest portion of abusers were also in the 14 and older age category (n = 28,873; 87%) with the mean age of abuser 17 years old. Offenders and victims were likely to be close in age. The majority of incidents (17,365; 53%) involved siblings that were less than 3 years apart in age, with 20,213 or 61%, of sibling violence incidents involving offenders who were the same age or older than their victim. The most common dyadic relationship involved male abusers and female victims (13,078; 40%), followed by male abusers and male victim incidents (11,166; 34%). In a gendered trend, female offenders were similarly more likely to have female victims (6,087; 18%) than male victims (2,693; 8.2%).
Table 1

Sibling physical assault: Offender and victim, demographics, N = 33,066

 

Number

Percent

Offender Age

 13 and under

4193

12.7

 14–17

16451

49.8

 18–21

12422

37.6

Offender Race

 White

22018

67.3

 Black

10431

31.9

 Other

273

0.8

Offender Sex

 Male

24263

73.4

 Female

8790

26.6

Victim Age

 13 and under

8531

25.8

 14–17

14,068

42.5

 18–21

10467

31.7

Victim Race

 White

21921

67.3

 Black

10400

31.9

 Other

264

0.8

Victim Sex

 Male

13863

42.0

 Female

19173

58.0

Sibling Relationship Dyad

 Male-Male

11166

33.8

 Male–Female

13078

39.6

 Female-Female

6087

18.4

 Female-Male

2693

8.2

Victim/Offender Age Differential

 Victim younger than Offender

20213

61.1

 Victim older or equal to Offender

12853

38.9

Victim/Offender Age Gap

 0–2 years difference

17365

52.5

 3–5 years difference

12476

37.7

 6–9 years difference

2800

8.5

 10 or more years difference

423

1.3

Incident characteristics

As indicated in Table 2 simple assaults (26,481; 80%) were the most prevalent of the three types of abuse comprising the operational legal definition of sibling violence in the present study, followed by aggravated assault (5,044; 15%) and then intimidation (1,541; 5%). Most incidents occurred in the victim’s residence (30,234; 91%). Personal weapons (hands and feet) were the weapons of choice in most incidents (22,808; 69%); guns, knives, and other objects were used in less than 10% of all reported sibling violence cases. The majority of incidents resulted in minor injury to the victim (16,249; 52%) and there was rarely alcohol (1,029; 3%) or drug (343; 1%) involvement, though this may be an artifact of primary data collection and recording practices.
Table 2

Sibling physical assault, offense demographics, N = 33,066

 

Number

Percent

Type of Assault

 Aggravated

5044

15.3

 Simple

26481

80.1

 Intimidation

1541

4.7

Location

 Residence

30234

91.4

 Other

2832

8.6

Weapons Involved

 No

4889

14.8

 Yes

28177

85.2

Gun Present

 No

32782

99.1

 Yes

284

0.9

Personal Weapons

 No

10258

31.0

 Yes

22808

69.0

Knife

 No

30858

93.3

 Yes

2208

6.7

Blunt Object

 No

32145

97.2

 Yes

921

2.8

Injury

 No Injury

14291

45.3

 Minor Injury

16249

51.5

 Major Injury

985

3.1

Alcohol

 No

32037

96.9

 Yes

1029

3.1

Drugs

 No

32723

99.0

 Yes

343

1.0

Personal weapon is defined as offender’s hands and/or feet

Analysis Plan

To develop a general profile of victims and offenders, both demographic and incident characteristics were examined to assess prevalence across a sample of reported cases over 6 years of official NIBRS data as shown in Tables 1 and 2. Following a basic univariate demographic comparison, Chi2 tests were employed to identify and explore the relationship between offender gender and all victim/abuser and incident level demographics as noted in Tables 3 and 4. Finally, binomial logistic regression, as shown in Table 5, is utilized to further assess the gendered nature of sibling violence using victim, offender, and incident characteristics as predictors. Logistic regression relies on maximum likelihood estimates and employs an iterative process of estimating the population parameters that created the dependent variable. In other words, “…this simply means that we can predict which of two categories a person is likely to belong to given certain other information…[for example] which variables predict whether a person is male or female” (Field 2005, p. 218). Using gender as a dichotomous dependent variable (male = 1), logistic regression predicts whether the offender is male or female based on other victim, abuser, and incident characteristics including: victim age, victim sex, abuser age, victim race, sibling age difference, substance abuse, weapon usage, and offense type.
Table 3

Offender and victim demographics by offender gender, N = 33,066

 

Male (n = 24,263)

Female (n = 8,790)

Number

Percent

Number

Percent

**Offender Age

13 and under

3087

12.7

1101

12.5

14–17

11835

48.8

4612

52.5

18–21

9341

38.5

3077

35.0

*Offender Race

White

16284

67.8

5731

65.9

Black

7534

31.4

2896

33.3

Other

201

0.8

72

0.8

**Victim Sex

Male

11166

46.1

2693

30.7

Female

13078

53.9

6087

69.3

Victim Age

13 and under

6276

25.9

2250

25.6

14–17

10346

42.6

3719

42.3

18 +

7641

31.5

2821

32.1

*Victim Race

White

16222

67.8

5694

65.7

Black

7498

31.4

2895

33.4

Other

191

0.8

73

0.8

**Age Differential

Victim Younger than Offender

15025

61.9

5182

59.0

Victim older or equal to ofender

9238

38.1

3608

41.0

**Offender/Victim Age Degree

0–2 year

12768

52.6

4593

52.3

3–5 years

9249

38.1

3220

36.6

6–9 years

1973

8.1

825

9.4

10 or more years

272

1.1

151

1.7

*p ≥ .05; **p ≥ .001

Table4

Offense characteristics by offender gender, N = 33,066

 

Male (n = 24,263)

Female (n = 8,790)

Number

Percent

Number

Percent

**Type of Assault

Aggravated

3556

14.7

1485

16.9

Simple

19618

80.9

6853

78.0

Intimidation

1089

4.5

452

5.1

Location

Residence

22175

91.4

743

8.5

Other

2088

8.6

8047

91.5

Weapons Involved

No

3533

14.6

1351

15.4

Yes

20730

85.4

7439

84.6

**Gun Present

No

23996

98.9

8773

99.8

Yes

267

1.1

17

0.1

**Personal Weapons

No

7222

29.8

3028

34.4

Yes

17041

70.2

5762

65.6

**Knife

No

22877

94.3

7970

90.7

Yes

1386

5.7

820

9.3

Blunt Object

No

23596

97.3

8537

97.1

Yes

667

2.7

253

2.9

Injury

No Injury

10585

45.7

3699

44.4

Minor Injury

11876

51.2

4368

52.4

Major Injury

713

3.1

271

3.3

**Alcohol

No

23410

96.5

8614

98.0

Yes

853

3.5

176

2.0

*Drugs

No

23987

98.9

8723

99.2

Yes

276

1.1

67

0.8

*p < .05; **p < .001;

Table5

Aggregate logistic regression results by gender, 2000–2005 (N = 33,066)

 

B

SEB

Wald χ2

p

Exp (B)

Offender Age

−.012

.008

2.193

.139

.988

*Younger Victim

.114

.045

6.359

.012

1.121

**Sibling Closeness

−.029

.007

16.126

.000

.972

Victim Age

.013

.007

3.322

.068

1.013

**Victim Female

.692

.027

645.608

.000

1.998

Victim Black

−.044

.027

2.599

.107

.957

**Alcohol Use

.471

.086

29.912

.000

1.602

*Drug Use

.319

.140

5.205

.023

1.375

Residence

.022

.046

.242

.623

1.023

Weapon

**Gun

1.689

.258

42.968

.000

5.415

**Knife

−.592

.074

63.943

.000

.553

Blunt Object

−.089

.094

.896

.344

.915

**Personal Weapon

.183

.034

29.060

.000

1.201

**No Injury

.112

.027

17.572

.000

1.119

Major Injury

.057

.083

.477

.490

1.059

Aggravated Assault

.081

.063

1.658

.198

1.084

Intimidation

.042

.066

.394

.530

1.042

*p < .05; **p < .001

Results

Offender/Victim Characteristics

From 2000 through 2005 there were 33,066 cases of single-victim single-abuser sibling violence incidents reported in NIBRS, including cases of aggravated assault, simple assault, and intimidation. As indicated in Table 3, males (24,263; 73.4%) were more frequent offenders of sibling violence than females (8,790; 26.6%). Although female abusers were significantly more likely to offend against females (6,087; 69%), male abusers were also more likely to offend against females (13,078; 54%) than males χ2 (1, n = 33,024) = 6.26, p = .000. Racial differences by abuser and victim were also significant. Female offenders (2,893; 33.3%) were more likely than male offenders (7,534; 31.4%) to be African American, χ2 (2, n = 33,053) = 10.906, p = .004 as were female victims (2,895; 33.4%) compared to male victims (7,498; 31.4%), χ2 (2, n = 33,053) = 12.839, p = .002. A gendered age dynamic between younger and older siblings was noted. Significant chi-square results reveal that male offenders (15,025; 62%) were more likely to be older than their victims compared to female offenders (5,182; 59%) χ2 (1, n = 33,053) = 23.99, p = .000. Significant results were also found in the age gap between siblings with a wider age span for female offenders than male offenders χ2 (3, n = 33,051) = 33.897, p = .000.

Sibling Violence Offense Characteristics

Sibling violence severity varied significantly by gender. Chi-square results in Table 4 reveal that while simple assault is the majority category for both genders, female siblings (1,485; 17%) are significantly more likely to be involved in aggravated assault than male siblings (3,556; 15%) χ2 (2, n = 33,053) = 33.849, p = .000. As expected, both genders report minor or no injury in the overwhelming majority of cases. Additionally, while overall weapon use is consistent across gender, the type of weapon used varies significantly. Male offenders were more likely than females to use both guns (1.1% compared to 0.1%) and personal weapons (70.2% compared to 65.6%), while female offenders were more likely to use knives (9.3% compared to 5.7%). The presence of alcohol was also significantly more prevalent in violence incidents with male offenders (853; 3.5%) than female offenders (176; 2.0%) χ2 (1, n = 33,053) = 48.994, p = .000 with similar significant gender differences for drug involvement (men = 1.1%; women = .8%) χ2 (1, n = 33,053) = 8.850, p = .003.

Victim Sex Predictive Model

Logistic regression was employed to predict the probability that a sibling violence offender was male, with offender sex (coded female = 0 and male = 1) as the dependent variable and each predictor or independent variable, with the exception of age and sibling age span, coded as 0 = No and 1 = Yes. Regression diagnostics revealed severe multicollinearity between offender and victim race as could be expected given the intra-familial relationship in sibling violence, therefore race of the victim was included as black = 1, and race of the offender was eliminated from the regression equation.

Logistic regression results in large part support the bivariate analysis yielding several significant findings. As shown in Table 5, victims of reported sibling violence are more likely to be female with both male and female sibling offenders more likely to abuse female siblings. Female offenders are also significantly more likely to abuse older siblings than their male counterparts. Additionally, supportive of the bivariate findings, the age span between older offenders and younger victim is wider for incidents with female offenders.

Examined incident characteristics reveal similar gendered differences. Male offenders are nearly five and a half times more likely to use a firearm during a sibling violence incident, whereas female offenders are significantly more likely to use a knife. Additionally, alcohol and drugs are both more likely to be present in incidents with male offenders. Finally, although there are no significant gender differences relating to type of assault, male perpetrators of sibling violence are less likely to cause injury to their victims than their female counterparts.

Discussion

Sibling violence is likely the most prevalent, yet least studied, form of family violence (Eriksen and Jensen 2006). Felson (1983) further offers that sibling aggression is perhaps the most frequent type of aggression in American society. With evidence suggesting that the sibling relationship is the family relationship most prone to conflict and discord and the consequences of resulting violence having wide ranging and far reaching psychosocial implications, sibling violence can no longer be considered just a “family problem” or an accepted rite-of-passage. Following the trajectory of other previously neglected forms of family violence (i.e., child abuse, intimate partner violence, and elder abuse) sibling violence is fast emerging as a social problem in need of an intervention.

The purpose of the present work was in part to address several of the most persistent reoccurring limitations in the extant literature as well as advance the current base of knowledge about this omnipresent and understudied emergent category of family violence. Findings suggest significant gender based differences across victims, offenders, and incident characteristics. Most notably, males were more likely to be offenders overall; a finding in concurrence with much of the previous literature drawing from smaller samples (Eriksen and Jensen 2006; Graham-Bermann et al. 1994). Interestingly, despite the inconclusive extant literature suggesting that males experience more sibling violence (Noland et al. 2004), females made up a greater portion of reported victimizations; a finding consistent with more traditional gender dynamics of crime reporting. It may be that females report more frequently because the incidents are greater in severity. Similar to Kettrey and Emery’s (2006) finding that females were more likely than males to be involved in severe sibling victimizations, we too found that females were more likely to be involved in more serious violence, suggesting a gendered trajectory of escalating violence.

The present work does not allow us to explain why these specific dynamics exist beyond conjecture but it certainly creates avenues for important future inquiry. The choice of weapon tends to mirror other forms of violence with girls more likely to use knives and boys more likely to use guns, perhaps indicating both the expressive nature of female violence and a gender based comfort level with firearms (Walsh and Krienert 2007; Walsh et al. 2008). Similar to prior research by Straus et al. (1980) the majority of cases were fairly minor, with fewer than 10% of cases involving a knife or gun.

The sheer prevalence of sibling violence highlights the need for additional research. Green (1984) noted a higher prevalence of both domestic violence and child abuse in families with sibling violence, leading towards a concurrence of family violence that may be hard to separate when trying to predict future negative behavior. Additionally, growing concern about long-term effects of school bullying and links to school shootings further warrant the examination of the sibling abuse/bullying connection. In a sample of 34 preschoolers, Berndt and Bulleit (1985) found that those with the most aggressive peer relationships were also those most likely to be aggressive with siblings. It’s likely that sibling violence exists within an interrelated context of abusive behavior. Other forms of family violence may aggravate sibling altercations causing an increased likelihood of overall violence and an avenue for risk assessment of future violence later in life.

The costs of sibling violence are immeasurable, and require societal action. The need to re-examine even basic demographic differences using standardized definitions is clear considering the definitional incongruence in the literature. It is time to re-frame the popular view of sibling violence as acceptable behavior towards one which identifies it as a social problem addressable with increased funding and research and with a legal redefinition as a “true” criminal offense. With strong links in the empirical literature to negative childhood and adult psychosocial outcomes, and an increasing ability to use large datasets of officially reported incidents (e.g., NIBRS) to assist in developing risk profiles of both victims and offenders, new opportunities/avenues can be developed for early intervention. Imperative is a shift in the destructive/detrimental “sibling rivalry” and “kids-will-be-kids” mantras. Through greater awareness and additional knowledge, the significance and seriousness of sibling violence will become more evident and apparent, underscoring the need for early proactive intervention strategies. Sibling violence is a very real and very common, albeit neglected, form of family violence with both short-and long-term consequences.

The present work is not without limitations which warrant mention beginning with our reliance on NIBRS data with all of its attendant concerns regarding the use of secondary data, albeit officially recorded. The data, while not nationally representative, are a useful accounting of sibling assaults being reported through participating law enforcement jurisdictions. Further, while increasing in exposure, NIBRS currently receives crime data from slightly more than half of all states and despite 6 years of reported data, the addition of new reporting jurisdictions and changes in reporting practices each year complicates longitudinal data analysis. Additionally, participating law enforcement agencies under-represent large urban populations which could produce a general selection bias (Libby and Wright 2009). NIBRS is also limited in terms of the contextual variables that are available and as a result we were not able to explore dynamic relational processes and motivations (i.e., jealousy, perceived parental favoritism, retaliation) that may serve as a catalyst for reported sibling violence.

Despite these limitations, we are confident that the strengths of the present work outweigh the limitations. With noted caveats, NIBRS still provides more information than any other source of official crime data (Dunn and Zelenock 1999). This work draws on what we believe to be the largest dataset of reported sibling violence incidents to date utilizing a standardized criminal justice oriented legal definition of assault; and baseline demographics of victims and offenders to produce predictive victim and offender profiles while also examining available incident characteristics. The work advanced current knowledge and provides a point from which to contemplate hypothesis testing, intervention, and additional research.

Future research, in addition to examining the offense severity by gender and victim vulnerability cues, might also focus on family dynamics and circumstances that serve to foster, enable, and embolden abusers. Further, future sibling violence research might specifically examine relational differences in families with multiple siblings and the role the relationships play in offending and victimization patterns. Along this very same line of inquiry, future research might also explore whether abusive/aggressive siblings are also abusive towards parents. Child-parent violence is another alarming area of understudied family violence and it might be beneficial from an intervention perspective to identify whether these youthful family aggressors perpetrate offenses against any and all family members including parents. Further, contrasting sibling offenders and victims and school bullying offenders and victims might enlighten relationship dynamics that could lead to a more holistic approach towards both family and school based intervention strategies. For example, are sibling aggressors also bullies in other settings like school and the community and are the sibling victims prone to being bullied in other settings?

A foremost obstacle to combating sibling violence and its long-term psychosocial consequences is the general discounting of its severity and seriousness by the individuals most aptly in position to intervene early and often. The age-old adage that “kids will be kids”, and the pervasive belief that conflict, aggression, and violence between brothers and sisters is a rite-of-passage must be removed from the lexicon of parental responses to sibling violence.

Footnotes
1

The states and subsequently counties participating in NIBRS data reporting have increased over the study period from nine states comprising 481 counties (4% of the population) in 1995 to 27 states and D.C. comprising 5617 agencies in 2005. The increasing annual participation by states and counties in the NIBRS reporting program makes longitudinal analyses including pattern and trend analyses extremely difficult at present time. The present work uses the NIBRS data in the aggregate and is therefore less impacted by annual changes in participation.

 
2

Aggravated assaults are defined as: “An unlawful attack by one person upon another wherein the offender uses a weapon or displays it in a threatening manner, or the victim suffers obvious severe or aggravated bodily injury involving apparent broken bones, loss of teeth, possible internal injury, severe laceration, or loss of consciousness” (Federal Bureau of Investigation 1992, p. 79).

 
3

Simple assault is defined as: “An unlawful physical attack by one person upon another where neither the offender displays a weapon, nor the victim suffers obvious severe or aggravated bodily injury involving apparent broken bones, loss of teeth, possible internal injury, severe laceration, or loss of consciousness” (Federal Bureau of Investigation 1992, p. 12).

 
4

Intimidation is defined as unlawfully placing “another person in reasonable fear of bodily harm through the use of threatening words and/or other conduct, but without displaying a weapon or subjecting the victim to actual physical attack” (Federal Bureau of Investigation 1992, p. 12).

 

Copyright information

© Springer Science+Business Media, LLC 2011