Quantification 0f Sibling Rivalry



Majed A. Ashy, Ph.D. (HarvardUniversity); John Kim (BostonUniversity)



Asinine icoThe first crime committed in history according to the world major religions occurred among siblings, Cane and Able. Sibling violence is probably not only the most prevalent and accepted form of family violence, but also the most potent form of violence (Steinmetz, 1981). The practice of viewing sibling violence as a normal part of growing up (e.g., by characterizing all forms of sibling violence as expressions of sibling rivalry) reinforces the use of these means of resolving conflicts and provides an early opportunity for children to practice what they have observed and experienced (Steinmetz, 1981; Frazier, 1994). Several researchers have suggested critical differences between sibling rivalry and pathological sibling abuse. For example, Rosenthal and Doherty (1984) identified three characteristics for sibling abuse: (1) chronic parental abuse of the child; (2) unconscious, covert parental permission for the child’s aggressive behavior; and (3) a child who identified with a dangerous, destructive person in the caretaker’s life (Rosenthal and Doherty, 1984; Frazier, 1994). From this perspective, sibling abuse might represent an intensification of “normal” sibling rivalry due to the abuser’s own maltreatment and deprivation.

Sibling relations can play important roles in children’s development that are not necessarily negative. For example, Heilbrunn (1986) argued that siblings might be thought of as “first peers” who provide children with some of their earliest social experiences, and that sibling relationships generally endure the tests of time and intra-relationship struggles. In addition, siblings act as models for the acquisition and development of social skills, and may protect one another from hostile intra- or extra-familial influences. Siblings can also function as agents of development in the cognitive, social, and affective domains (Frazier, 1994).

Some investigators have suggested that even sibling abuse might help a child cope with a dysfunctional environment. For example, Green (1984) argued that sibling attacks were adaptive for the abusers because they might afford a measure of revenge against the more highly regarded sibling rival, an outlet for rage directed toward a parent, an attention-getting device, a sense of mastery over the trauma of their own abuse, and a method to “educate” the abusing parent (Green, 1984; Frazier, 1994). Green (1984) found that children who inflicted serious injuries on their young siblings had been physically abused themselves; their families were undergoing crises, which accentuated their maternal deprivation and rejection; they were burdened with excessive caretaking for the target sibling, who was perceived as the favorite; and they also had experienced the recent loss of their father or paternal caretaker (Frazier, 1994).

Witnessing abuse among parents or toward siblings even without directly receiving abuse has been found to increase the risk for sibling abuse. For example, both abused and non-abused children from abusive families showed more negative feelings and perceptions and fewer positive ones toward their parents and siblings than children from non-abusive families, and more feelings of ambivalence toward their parents than children from non-abusive homes (Halperin, 1981; Frazier, 1994). In addition, children who stood by while their siblings were abused presented a picture of outward compliance, yet their incidence of anxiety, depression, and the need for therapy gave evidence of internal stress. Back in the day, mental health treatment wasn’t so common which meant most children had to suffer. Whereas, in modern-day society, we’re able to access a host of drugs to help. For example, ketamine infusion therapy is one of many that has been proven to reduce depression in children and adults. Children, who saw their mothers being abused, appeared to suffer more emotional turmoil than those who observed abuse of their siblings. They also had a tendency to model the violent behavior of their parents (Pfouts, Schopler, Henley, 1982; Steinmetz, 1977; Frazier, 1994).

In addition to age and gender, birth order and family size can play a significant role in sibling abuse. For example, male sibling pairs more often threw things, pushed, and hit than did female sibling pairs, the highest use of physical violence occurred between boy-girl sibling pairs, 68 percent of which engaged in high levels of violence (Steinmetz, 1981; Frazier, 1994). Female sibling pairs had consistently higher discussion scores than did male pairs, and mixed-sex pairs had the lowest discussion and verbal scores. Consistent with other literature on aggression and violence, male sibling pairs were consistently more violent than female sibling pairs, although the lowest violence scores tended to be among brother/sister sibling pairs. In addition, most of the abusing children were the eldest, and most of the abusive interactions took place while parents were in close proximity (Rosenthal & Doherty, 1984; Frazier, 1994).

Even though, in recent years, accumulating evidence has linked various forms of childhood experiences of abuse to both psychological and physical symptoms (Ashy & Malley-Morrison, 2003; Walker, Katon, Harrop-Griffiths, Holm, Russ, & Hickok, 1988), few studies have examined the associations among sibling abuse, brain development, and psychological and physical symptoms. It is our view that chronic experiences of sibling verbal abuse might function as a source of stress similar to that associated with child abuse, and that the same biological and psychological mechanisms linking parental abuse of children to physical and psychological symptoms might link sibling abuse to these symptoms. Since few studies are available on the relationships among sibling abuse, limbic system function, psychological symptoms of anxiety, depression, hostility, and somatization, and various physical symptoms, it is useful to turn to the literature on parental abuse of children to find a basis for hypotheses about the effects of sibling abuse.

Sibling abuse, like child abuse, might influence physical and psychological health via the mediation of stress. The stress connected with experiencing early abuse is associated with features of limbic system dysfunction, EEG abnormalities, and measures of hemispheric asymmetry (Pfeffer, 1996; Ito, Teicher, Glod, & Ackerman, 1998; Ashy, Malley-Morrison, & Teicher, 2003). In addition, stress impairs neural functions and contributes to brain aging and impairment (Lombroso & Sapolsky, 1998). Under perceived stress, the brain, through the autonomic nervous system (ANS), activates the old adaptive response of fight or flight. The stimulation of the sympathetic nervous system (SNS) results in physiological reactions such as increased heart rate and blood pressure that mobilize energy for the purpose of defense or escape (Nelson & Panksepp, 1998). In addition, the ANS regulates the production of immune cells from immune organs such as the spleen and the thymus (Maier & Watkins, 1996). At the same time, stress activates the production of several pituitary hormones that in turn activate the production of adrenal hormones. These hormones play an important role in physiological mobilization of energy and the production and function of immune cells (Schulkin, 1993). The immune system produces cytokines, which serve the purpose of communication among immune cells and get attached to receptors in the brain. This process leads to psychological, cognitive, and behavioral changes that are characteristic of illness states, such as depression, lack of appetite, lack of social interest, lack of exploration and so on (Sculkin, 1993).

Several studies have found that stresses such as sexual abuse have serious endocrine consequences. For example, Lemieux and Coe (1995) examined the levels of norepinephrine and cortisol in the urine of eleven women with posttraumatic stress disorder who had experienced childhood sexual abuse, eight women who had experienced childhood sexual abuse without posttraumatic stress disorder, and nine non-abused controls. Both abused groups showed a tendency for polyuria (frequent urination). The PTSD group had significantly elevated daily levels of norepinephrine, epinephrine, dopamine, and cortisol, which are involved in the stress response, emotions, and other important biological functions such as the pain response (Lemieux & Coe, 1995; Ashy, Malley-Morrison, & Teicher, 2003).

Sibling abuse might be also linked to psychological and physical symptoms through the maladaptive emotional expression styles adopted by siblings in an abusive environment. There is evidence of negative effects of trauma (Roesler & McKenzie, 1994), secrecy, emotional repression and lack of disclosure, and mental rumination on both physical and psychological health (Pennebaker, 1995). For example, Servaes, Vingerhoets, Vreugdenhil, Keuning, and Broekhuijsen (1999) found that patients with breast cancer showed significantly more ambivalence over emotional expression, more restraint, and more anxiety than the healthy controls. Cancer patients were found also to have conflicting feelings with regard to expressing emotions, were reserved, anxious, and self-effacing, and repressed aggression and impulsiveness.

From a psychiatric perspective, the association between abuse history and certain psychiatric diagnoses, including the anxiety and somatoform disorders, may lead to communicating psychological distress via bodily symptoms in the process of somatization (Walker, Gelfand & Gelfand, 1995; Briere & Runtz, 1988; Reiter, Shakerin & Gambone, 1991; Loewenstein, 1990; Leserman, et al, 1995; Ashy, Malley-Morrison, & Teicher, 2003). From a psychodynamic perspective, abuse may produce feelings of guilt and shame that may then be expiated through physical pain or suffering (Engel, 1951; Ashy, Malley-Morrison, & Teicher, 2003).

Kirmayer and Young (1998) argued that somatic symptoms might serve several functions. They can be also a cry of help from children in negative household environments. Somatic symptoms might also serve the purpose of social repositioning and redistribution of power (Candell, 1995). A child abused by a sibling might express his/her anxiety or depression in the form of physical complaints. This can have important implications for family physicians or pediatrician who might encounter situations where a child might report symptoms without clear biological causes.

Sibling abuse might also influence cognition or behavior in ways that predispose the individual to physical illness. From a cognitive perspective, negative cognition and ineffective coping styles may lead to a maladaptive adjustment to illness, thereby increasing pain reporting and behavior (Talley & Zinsmeister, 1994; Ashy, Malley-Morrison, & Teicher, 2003).

Several studies have found that a history of childhood experiences of maltreatment is associated with number of physical conditions such as somatization disorder (Morrison, 1989), chronic pelvic pain (Walker, Katon, Harrop-Griffiths, Holm, Russ, & Hickok, 1988), and functional gastrointestinal disorders (Drossman, Leserman, Nachman, Zhiming, Gluck, Toomey, Nachman & Glogua, 1997), fibromyalgia (Walker, Keegan, Gardner, Sullivan, Bernstein, Katon, 1997), and higher rates of non-abdominal symptoms such as headaches, backaches and fatigue, as well as more lifetime surgeries, and higher incidences of obesity ( Felitti, Anda, Nordenberg, Williamson, Spitz, Edwards, Koss, & Marks, 1998; Lechner, Vogel, Garcia-Shelton, Leichter, 1993; Walker, Gelfand, Gelfand, Koss, 1995; Leserman, Drossman, Li, Toomey, 1996). It is possible that sibling abuse might have similar associations to physical symptoms.

Based on these findings, Ashy (2003) developed a model linking negative childhood experiences such as abuse to physical health via the mediation of stress, emotions and physiological stress reactions, including current study examined the various components of this model, with a focus on sibling aggression.

The following hypotheses guided this investigation:

  1. Amount of childhood experience of sibling verbal aggression will be positively correlated to verbal, physical, psychological and witnessed parental aggression, psychological symptoms (anxiety, depression, hostility, somatization), limbic system dysfunction and physical symptoms.
  2. Emotional distress, attachment styles and limbic system dysfunction, mediate the relationship between childhood experiences of sibling verbal abuse and young adult cardiovascular, immune, chronic pain and gastrointestinal symptoms. The final decision as to what predictor variables were entered in which step of a regression analysis was based on the outcome of the preliminary correlational analyses.

In addition, the following research questions guided this investigation:

  1. What role does the gender of the siblings play in the relationships between sibling abuse, emotional distress, limbic system, attachment, and physical symptoms?
  2. Are there differences in the amount of sibling abuse in participants with different attachment styles?



The sample consisted of 400 participants (276 females and 124 males) ranging in age from 16 to 78 (mean age of 23). The average respondent was a college student completing an introductory psychology or experimental course at a large urban university in the northeast, Caucasian, Christian, single, and self-reported as middle or upper middle class. This research was conducted and administered in compliance with all American Psychological Association guidelines and all state and federal laws.


Demographic Questionnaire: Sex, age, marital status, socioeconomic status, education, and ethnic and religious background information were obtained, in addition to medical, psychological, trauma and criminal history information about the participant and her/his family. Questions regarding medication, drug or stimulants use were also included.

The Limbic System Function Questionnaire (LSCL-33, Teicher et al., 1993) is a 33-item self-report checklist developed to screen for possible dysfunctions (irritability) of the limbic system, including paroxysmal somatic disturbances, brief hallucinatory events, visual disturbances, automatisms, and dissociative disturbances. Participants rate the lifetime frequency with which they have experienced these disturbances, using the predefined descriptors of “never”, “rarely”, “sometimes” and “often.” A total score for the 33 items is calculated, as well as factor scores for somatic, sensory, behavioral and mnemonic disturbances (Teicher et al., 1993). The LSCL-33 has demonstrated good concurrent validity and test-retest reliability with r = .92 for the whole scale and .78 to .86 for the subscales.

The Symptom Questionnaire (SQ; Kellner, 1987) is an extensively validated 92-item yes/no questionnaire with four symptom subscales (depression, anxiety, anger, somatic) and four well-being subscales (content, relaxed, friendly, somatic well-being). In two crossover drug trials (Kellner, Rada, Anderson, et al., 1979; Kellner, Collins, Shulman, et al., 1974) of anti-anxiety drugs and placebo, the SQ scales discriminated between the effects of the psychotropic drug and placebo. In a normal group, all four symptoms subscales correlated positively and all well-being scales correlated negatively with a life events scale (Kellner, Pathak, Romanik, et al., 1983). The correlation of the SQ depression scale with the Hamilton Rating Scale for Depression (Hamilton, 1960) was 0.66 and with the Hamilton Anxiety Rating Scale (Hamilton, 1968) was 0.69. Split-half and test-retest reliability coefficients are consistently high (Kellner, 1987).

The Physical Health Questionnaire (PHQ; Ashy, 2000) is a 56-item questionnaire that assesses the severity of physical symptoms. Participants rate the severity of various physical symptoms on a 6-point scale (1= never; 6= chronic). The measure has subscales that assess cardiovascular, immune, pain, and digestive functioning in both males and females (menstrual symptoms were included in the female scale). The PHQ has demonstrated good reliability with internal consistencies ranging from .72 to .91 (Ashy, 2000).

The Sibling Verbal Aggression Scale (Teicher, 2000) is a 30–item scale that measures the frequency of verbally aggressive behaviors from brothers and sisters before the age of 18. Participants responded on a scale ranging from “never” to “daily”.

The Parent Verbal Aggression Scales (VAS; unpublished; Polcari & Teicher): The VAS was originally derived from the Parenting Style Questionnaire to address key components of verbal abuse, such as yelling, blaming, insulting, and belittling. A recent psychometric evaluation of the VAS was undertaken with a sample of 554 young adults (384 females, 170 males) aged 18 to 23 recruited from the general population through advertising for psychiatric research volunteers. The standardized Cronbach’s alpha for the scale was .96 for the mother version and .97 for the father version. Confirmatory principal components analysis with Varimax rotation, using eigenvalue > 1.0, resulted in identical two-factor solutions for both parent versions. Factor 1, defined by 12 items of criticism, threats, insults, and blame, was labeled Emotionally Abusive Speech. Factor 2, defined by the three items of raising voice, yelling, and scolding, was labeled Loud Corrective. As expected, the scores on maternal and paternal VAS scales were highly negatively correlated with the Parental Bonding Instrument (PBI; Parker et al., 1979) “caring” sub-scales.

The Revised Conflict Tactic Scale Questionnaire (CTS-2, Straus et al., 1979; 1990) is a 78-item questionnaire that assesses the self-report amount of negotiation, psychological aggression, physical assault, and sexual coercion that 1) has been witnessed and received from parents or parent figures during the worst year of childhood, and 2) has occurred within a dating relationship during the last year. Participants respond by selecting a number from 0 to 25 that signifies the frequency with which specified behaviors have occurred (0=never, 1=once, 2=twice, 4=3-5times, 8=6-10 times, 15=11-20 times and 25=more than 20 times). These scores are summed to obtain a total score for each subscale. The CTS-2 has demonstrated good construct and discriminant validity and good reliability, with internal consistencies ranging from .79 to .95 (Straus et al. 1979; 1990).

All scales and subscales used had good alpha levels ranging from .71 to .97.

Your Comment