Search tips
Search criteria 


Logo of nihpaAbout Author manuscriptsSubmit a manuscriptHHS Public Access; Author Manuscript; Accepted for publication in peer reviewed journal;
J Fam Psychol. Author manuscript; available in PMC 2010 December 1.
Published in final edited form as:
PMCID: PMC2826125

Intergenerational Transmission of Aggression in Romantic Relationships: The Moderating Role of Attachment Security


This prospective study used longitudinal, multi-reporter data to examine the influence of parents’ marital relationship functioning on subsequent adolescent romantic relationships. Consistent with Bryant and Conger’s model for the development of early adult romantic relationships (DEARR; 2002), we found that interactional styles, more specifically paternal aggression and satisfaction, exhibited in parents’ marital relationship when their adolescents were age 13, were predictive of qualities of the adolescent’s romantic relationships five years later. Continuities were domain specific: paternal satisfaction predicted adolescent satisfaction and paternal aggression predicted adolescent aggression. Attachment security moderated the link between paternal aggression and subsequent adolescent aggression, with continuities between negative conflictual styles across relationships reduced for secure adolescents. Results are interpreted as suggesting that attachment may help attenuated the transmission of destructive conflict strategies across generations.

Keywords: Romantic relationships, aggression, attachment

Romantic relationships become increasingly important over the course of adolescence (Richards, Crowe, Larson, & Swarr, 1998) and have been linked to both positive (Moore & Leung, 2002) and negative (Joyner & Udry, 2000; Wenz, 1979) outcomes, depending on the qualities of relationships. These relationships appear to be important domains for developing adult-like relationships skills (Furman & Shaffer, 2003) and, in fact, have been linked to qualities of subsequent marital relationships (Furman & Flanagan, 1997). However, little is known about the protective and risk factors that may facilitate the development of positive skills and limit the development of negative romantic behaviors during adolescence.

The family context likely plays an important role in the development of subsequent romantic relationship skills. The relationship between parents is often the first and most frequent interpersonal exchange between romantic partners that a child witnesses. Investigations have linked parents’ marital relationships to various markers of adjustment during adolescence, such as emotional security (Davies & Cummings, 1994), attachment styles (Ozen, 2003), and intimacy (Feldman, Gowen, & Fisher, 1998). Parents’ marital relationships also play a large role in children’s social development, helping to shape their behavior in peer relationships, with children learning from observing their parents as models in addition to having behavior directly reinforced. Children learn how to appropriately and effectively maintain social relationships from their parents through various mechanisms. Children often learn through direct coaching on positive behaviors during peer interactions and through modeling such behaviors during parent-child interactions (Bandura, 1977; Black, 2002). Yet it is unlikely that all aspects of the interparental relationship are learned through modeling, particularly when certain relationship qualities, such as parental marital satisfaction, are not easily operationalized nor observed. In fact, there are several additional perspectives on how family of origin experiences are related to the ways in which offspring negotiate later intimate relationships, including developing representations and expectations of oneself and other relationships (i.e., attachment theory), and via more direct repetition of dysfunctional emotional processes and dynamics from their family of origin (Ainsworth & Bowlby, 1991; Bowen, 1966). We will discuss how social learning, attachment, and family systems perspectives may contribute to a better understanding of the intergenerational transmission of relationship qualities in further detail below.

The impact of the interparental relationship likely pervades peer relationships later in life as well, including newly developing adolescent romantic relationships. One key marker of a successful romantic relationship in adolescence and early adulthood is the reported satisfaction of both partners. Some evidence has suggested that the interparental relationship may indeed play a role in offspring relationship satisfaction, as findings by Feldman, Gowen, and Fisher (1998) indicated that mothers’ marital satisfaction was a primary predictor of happiness in love relationships as a young adult. This finding can perhaps be explained by the mother’s expressivity of positive emotionality being more broadly communicated within the nuclear family environment. Further, other findings suggest that marital satisfaction is related to offspring’s self-esteem, and emotional well-being (Shaw & Emery, 1987). If marital satisfaction helps to facilitate better emotional adjustment for offspring, this may enable such offspring to develop healthier, and more satisfied intimate relationships in adolescence.

Another key element of successful adolescent dating relationships is the use of constructive conflict strategies, such as calm discussion, support, problem-solving, compromise, affection, humor, apology, agreements to discuss later, and conflict resolution (Cummings, Davies & Campbell, 2001). Learning to negotiate conflict constructively is a necessary developmental task that all children must complete in order to function successfully across a variety of contexts. Further, developing one’s own optimal style of conflict resolution is a key component of identity development. On the other hand, destructive strategies, such as threat, verbal and nonverbal hostility, personal insult, defensiveness, withdrawal, and physical aggression, might impair the quality of a relationship (Cummings, Goeke-Morey & Papp, 2003), as such conflicts usually result in dissatisfaction of both parties.

But how do adolescents learn to successfully negotiate conflict in their budding romantic relationships? Consistent with Bandura’s Social Learning Theory (1977), adolescents likely use their parents’ marital relationships as models for their own romantic relationships. This idea is consistent with previous literature that suggests such modeling of parental relationship qualities in the arena of friendship during both childhood (Schudlich, Shamir, & Cummings, 2004) and adolescence (Berger, Antonishak & Allen, 2006). If adolescents utilize parents as models for their own future relationships, this may likely influence how adolescents respond to conflict in such relationships. For instance, adolescents exposed to marital violence during childhood are more likely to justify the use of violence in dating relationships (Lichter, Batson, & Brown, 2004). Additionally, research has supported the idea that adolescents’ social information processing may be mediating the link between interparental relationship conflict and subsequent young adult romantic relationship conflict (Dodge, 1986; Fite, Bates, Holzworth-Munroe, Dodge, Nay, & Petit, 2008). Further, Bryant and Conger (2002) have extended this idea, proposing an observational learning hypothesis as one part of their model for the development of early adult romantic relationships (DEARR), suggesting that children may carry forward the types of interaction styles they witness in interactions between members of their family; these interactions could be between parents, parents and siblings, and/or other siblings. In other words, the child may replicate the behaviors, potentially including constructive or destructive conflict tactics, commonly used in his or her family of origin. This theory is supported by previous findings that parental divorce increases the likelihood of subsequent offspring divorce through the impact it has on the offspring’s behaviors that interfere with the maintenance of successful intimate relationships (Amato, 1996), such as poor conflict resolution skills. Unfortunately, whether and how these processes apply to adolescents has not yet been well-examined.

Prior research indicates that parents’ treatment of one another during times of conflict may set the overall tone for how children interpret conflict in the family system (Harold, Shelton, Goeke-Morey, Cummings, 2004; Allen, Hauser, O’Connor, Bell, & Eikholt, 1996). From a family systems perspective, frequent marital conflict may serve as one symptom of intense emotionally ambivalent relationships and tension within the family environment, often referred to as “fusion” (Kerr, 1981). More specifically, the use of interparental aggression, a particularly maladaptive conflict strategy, may heighten the overall level of anxiety, vulnerability, and negative emotional expressivity within the family (Bowen, 1965). Oftentimes, the tendency for partners to engage in frequent conflict becomes a pattern that serves to “maintain” functioning within a particular family and across generations. When the emotional equilibrium of families is maintained in such a way for years and years, this (dys)functional system is often recreated within other emotionally evocative contexts of the offspring, such as within their own romantic relationships. This perspective argues that the development of aggression within one’s intimate relationship may in fact be a symptom or byproduct of dysfunctional emotional processes and compromises that have been occurring within the family system for generations and generations.

One possibility that neither Bryant and Conger’s (2002) model nor the Family Systems perspective adequately addresses is that interparental experiences may influence some adolescents more than others. Bowlby (1988) argued that individuals develop working models, or cognitive representations about themselves and others, that are internalized and subsequently affect their perceptions of others in future attachment relationships. Secure individuals are believed to have developed positive working models that allow them to process emotionally evocative information effectively, and thus to do so in their relationships later in life as well. Consistent with this, we argue that secure adolescents may have greater cognitive and emotional autonomy which allows them to step back from their observations of marital interactions and consciously decide not to emulate some less adaptive aspects. For example, attachment theory has proposed that in adolescence and adulthood, a secure stance toward attachment relationships is reflected in precisely this emotional and cognitive autonomy. Secure adolescents are likely to be able to more accurately label and perceive maladaptive marital interactions as such, and may be less likely to automatically replicate them in their own future relationships than are insecure adolescents. This pattern is likely to be particularly applicable to highly distressing behaviors (e.g., maladaptive forms of marital conflict), which less secure adolescents may have more difficulty gaining perspective upon. This idea has been supported by previous findings that secure attachment to parents can serve as a protective factor against behavior problems associated with marital conflict (El-Sheikh, 2004). Thus, adolescents’ states of mind regarding attachment may serve as moderators between the interparental relationship and subsequent adolescent romantic relationship experiences. Yet, while some research has examined adolescents’ modeling of parental marital interaction qualities, no research has considered the extent to which adolescents might be differentially sensitive to these qualities. However, previous findings that adult children of divorce who are securely attached are less likely to divorce in the early years of marriage than those who are insecurely attached (Crowell, Treboux, & Brockmeyer, 2009) suggest that a secure attachment state of mind may act as a buffer against later replicating the negative relationship qualities found in one’s family of origin.

Adolescents who are exposed to extreme negative family experiences, such as interparental aggression, yet also are able to thoughtfully reflect upon, and adaptively process the emotions and cognitions related to these experiences, can be characterized as having a secure state of mind. A common misconception regarding attachment is that secure individuals consistently have had very positive and loving childhood experiences. In fact, an individual can be characterized as secure if they are able to talk coherently, objectively, and thoughtfully about their experiences in an attachment interview, despite the negative content of their experiences (e.g., trauma, loss, witnessed interparental violence).

Some evidence suggests that fathers have a particularly salient role in child development, although this has been scarcely studied in terms of adolescent development. The literature that does exist in this arena, however, has underscored the impact of paternal behavior, especially regarding effects on children’s externalizing behaviors (Phares & Compas, 1992). More specifically, prior research suggests that children are particularly emotionally responsive to their father’s anger during disagreements (Cummings, Goeke-Morey, Papp, & Dukewich, 2002), potentially in part because any implicit threat of marital violence tends to be inherently more frightening and emotionally salient if it comes from fathers rather than mothers. Additionally, in a meta-analysis of gender differences in marital aggression, Archer (2002) found that the type of aggressive act varies across gender, with women experiencing more severe physical acts of violence against them than men; just the type of acts that would likely invoke fear and activate the attachment system in their children. Thus, this study focuses particularly on the father’s aggressive marital behaviors as a predictor.

The goal of this study was to investigate the link between parents’ marital experiences, specifically physical aggression exhibited by the father and both parents’ marital satisfaction, as predictors of adolescents’ conflict tactics and satisfaction in subsequent romantic relationships. Further, the potential moderating role of adolescents’ states of mind regarding attachment will be investigated. It is hypothesized that fathers’ marital satisfaction is expected to be predictive of adolescents’ subsequent romantic relationship satisfaction. Further, exposure to destructive interparental conflict tactics during adolescence will be associated with use of destructive conflict tactics by insecurely attached adolescents in subsequent romantic relationships. However, a secure attachment orientation is expected to act as a moderator; thus the link between observing destructive parental conflict and engaging in such conflict in subsequent romantic relationships is expected to be lessened for secure adolescents. Finally, we aim to answer the question of whether these associations are domain specific, i.e. whether marital aggression is predictive just of adolescent romantic relationship aggression or also of adolescent satisfaction in romantic relationships and similarly with predictions from both parents’ marital satisfaction. Since adolescent boys and girls likely experience romantic relationships so differently, and since parents may have a different impact on same-sex children than on opposite-sex children, it is also important to consider the role of adolescent gender as a moderator. By implementing a longitudinal, multi-reporter design, we aim to shed new light on the specific developmental precursors that perpetuate negative romantic experiences later in life.



This report is drawn from a larger longitudinal investigation of adolescent social development in familial and peer contexts. Participants included 75 adolescents (M age=13.36, SD=0.64; 36% males and 64% females) who were first interviewed along with their parents and reinterviewed approximately 5 years later with their romantic partners. The sample was racially/ethnically and socioeconomically diverse: Of the participants, 57% identified themselves as Caucasian, 28% as African American, and 15% as being from other or mixed ethnic groups. Adolescents’ parents reported a median family income in the $40,000 to $59,999 range (13% of the sample reported annual family income less than $20,000, and 37% reported annual family income greater than $60,000).


Adolescents were recruited from the 7th and 8th grades at a single public middle school drawing from suburban and urban populations in the Southeastern United States. One cohort of 8th graders was included and two different cohorts of 7th graders were included in successive years. The school was part of a system in which students had been together as an intact group since 5th grade. Students were recruited via an initial mailing to all parents of students in the school along with follow-up contact efforts at school lunches. Families of adolescents who indicated they were interested in the study were contacted by telephone. Of all students eligible for participation, 63% agreed to participate either as either target participants or as peers providing collateral information. All participants provided informed assent before each interview session, and parents provided informed consent. Interviews took place in private offices within a university academic building. Parents, adolescents, and peers were all paid for their participation.

After adolescents who met study criteria were identified, letters were sent to each family of a potential participant explaining the investigation as an ongoing study of the lives of teens and families. These initial explanatory letters were then followed by phone calls to families who indicated a willingness to be contacted further. If both the teen and the parent(s) agreed to participate in the study, the family was scheduled to come to our offices for two 3-hour sessions at each wave of the study. Adolescents and families were paid for participation at each interview. At each session, active, informed consent was obtained from parents (Time 1 only), adolescents (Time 1 and Time 2), and romantic partners (Time 2 only). In the initial introduction and throughout both sessions, confidentiality was assured to all family members and romantic partners, and adolescents were told that their parents would not be informed of any of the answers they provided. Participants’ data were protected by a Confidentiality Certificate issued by the U.S. Department of Health and Human Services, which protected information from subpoena by federal, state, and local courts. Transportation and child care were provided if necessary.

At Time 1, data were obtained from all adolescents and their parents. At Time 2, when target adolescents were an average of 18.31 years of age (SD=1.27), data were obtained from all returning adolescents and their romantic partners. Romantic partners averaged 19.20 years in age (SD=3.19), and their relationships with the target adolescents ranged in duration from approximately two months to 64 months in duration (M = 14.78, SD = 13.61).

These adolescents were a subset of 136 adolescents from two-parent families who were originally interviewed at baseline, 61 of whom did not did not report having a romantic partner at follow-up. Of the original 136 adolescents participating at Time 1, 133 participated in at least one assessment in the larger study during the time period of this follow-up assessment (2% attrition). We surveyed adolescents repeatedly to assess whether they were in a relationship of 3 months duration or longer. While logistical limitations made precise numbers and analyses unavailable, we estimate that an additional 10% of teens for whom we don’t have relationship data were in relationships but could not be scheduled to come in while those relationships lasted.

Formal attrition analyses revealed no differences between the adolescents who returned at follow-up with a romantic partner at age 18 versus those who did not on any of the demographic or primary predictor measures in this study, with the exception of minority status. Those who returned at age 18 with a romantic partner were less likely to report identifying themselves as members of a racial/ethnic minority group. Further, none of the primary predictors were significantly associated with likelihood of returning at follow-up with a romantic partner at age 18. Further attrition analyses were conducted to investigate any possible differences between those who did not participate at follow-up because they had no romantic partner versus those who were in a romantic relationship, but could not be scheduled. These analyses yielded no differences on any of the demographic or primary predictor measures in this study.

To best address any potential biases due to attrition in longitudinal analyses, full imputation maximum likelihood (FIML) methods were used with analyses, including all variables that were linked to future missing data (i.e. where data were not missing completely at random) (Muthén & Muthén, 1998–2008). Because these procedures have been found to yield the least biased estimates when all available data are used for longitudinal analyses (vs. listwise deletion of missing data) (Arbuckle, 1996; Enders, 2001; Raykov, 2005), the full sample of 75 adolescents in romantic relationships at age 18 was utilized for these analyses. This full sample thus provides the best possible variance/covariance estimates and was least likely to be biased by missing data. No data is estimated or imputed in this procedure, however, rather it simply accounts and corrects for biases due to missing data. Complete data was available For 77 to 99 % of the sample (depending on the measure examined). Alternative longitudinal analyses using just those adolescents without missing data (i.e., listwise deletion) yielded results that were substantially identical to those reported below. In sum, analyses suggest that attrition was modest overall and not likely to have distorted any of the findings reported.


Demographic information

Adolescents and their parents were asked to provide basic demographic information such as gender, and race/ethnicity at age 13. Participant age was calculated using birth dates. Parents were asked to provide information regarding their level of education, annual household income, and number of persons supported by this income. Romantic partners were asked to report their gender, age, and the length of the romantic relationship at age 18.

Marital Aggression

When adolescents were age 13, parents were asked to rate their partners’ behaviors during marital conflicts on the Physical Aggression subscale of the Conflict Tactics Scale (CTS; Straus, 1979). The theoretical assumption behind the measure is that all families experience conflict between members, but if the conflicts are handled in an unsatisfactory or inadequate manner, problems become apparent. The questionnaire explored disagreements between spouses over their lifetime; specifically inquiring about the frequency of arguments between partners, how conflicts were handled, and how conflicts were resolved. Using the Physical Aggression subscale, mothers responded to 11 items about how often their partner had used physical aggression during conflict, with responses ranging from 1 (Never) to 4 (Many Times). Items included endorsement of behaviors such as throwing objects, pushing, slapping, kicking, hitting with a belt or club, beating up, choking, burning, threatening with a knife, and threatening with a gun. Paternal aggression scores were computed by summing mothers’ responses, which were weighted for severity of physical acts (e.g., using a gun was weighted much more heavily than pushing). The data was then log transformed in order to address skewness caused by weights. Maternal aggression scores were computed by summing fathers’ responses. Both the paternal and maternal aggression subscales have moderate internal consistency (Cronbach’s alpha of .75 and .59, respectively).

Marital Satisfaction

When adolescents were age 13, parents were asked to rate their own relationship satisfaction on the Satisfaction subscale of the Dyadic Adjustment Scale (DAS; Spanier, 1979). This measure was specifically designed to assess marital and dyadic adjustment as an indicator of marital quality. This subscale consists of 8 items which ask respondents to report their feelings about the relationship, from how often they considered divorce to how often they kissed their partners, on a six-point Likert scale (1 being all the time, and 6 being never). Two additional items ask about the amount of happiness in their relationship and how they feel about the future of their relationship. Paternal satisfaction scores were computed by summing fathers’ responses and maternal satisfaction scores were computed by summing mothers’ responses.. Both the paternal and maternal satisfaction subscales have high internal consistency (Cronbach’s alpha of .88 and .91, respectively).

Adolescent Attachment Security

Target adolescents’ attachment security was assessed using the Adult Attachment Interview (AAI; George, Kaplan & Main, 1996). The AAI is a structured interview that probes individuals’ descriptions of their childhood relationships with parents by asking for both abstract descriptions and specific supporting memories. For example, participants were asked to list five words describing their early childhood relationships with each parent, and then to describe specific episodes that reflected those words. Other questions focused upon specific instances of upset, separation, loss, trauma, and rejection. Finally, interviewers asked participants to provide more integrative descriptions of changes in relationships with parents and the current state of those relationships. The interview consisted of 18 questions and lasted one hour on average. Slight adaptations to the adult version were made so that the questions were more natural, and easily understood by an adolescent population (Ward & Carlson, 1995). Interviews were audiotaped and transcribed for coding.

The AAI Q-set (Kobak, Cole, Ferenz-Gillies, Fleming, & Gamble, 1993) was designed to closely parallel the Adult Attachment Interview Classification System (Main & Goldwyn, 1994), but to yield continuous measures of qualities of attachment organization. Each rater read a transcript and provided a Q-sort description by assigning 100 items into nine categories ranging from most to least characteristic of the interview, using a forced distribution. All interviews were blindly rated by at least two raters with extensive training in both the Q-sort and the Adult Attachment Interview Classification System.

These Q-sorts were then compared with a dimensional prototype sort for secure vs. anxious interview strategies, with security reflecting the overall degree of coherence of discourse, the integration of episodic and semantic attachment memories, and a clear objective valuing of attachment. The individual correlation of the 100 items of an individual’s Q-sort with a prototype sort for a maximally secure transcript was then used as that participant’s security score (ranging from −1.00 to 1.00). The Spearman- Brown interrater reliabilities for the final security scale score was .88. Although this system was designed to yield continuous measures of qualities of attachment organization, rather than to replicate classifications from the Main & Goldwyn (1994) system, we have previously compared scores coded by the investigator’s lab to classifications obtained from an independent coder with well-established reliability in classifying AAI’s. We did this by converting the Q-sort scales described above into classifications using an algorithm described by Kobak and colleagues (1993). Using this approach, we obtained an 84% match for security vs. insecurity between the Q-sort method and the classification method (kappa = .68). To maximize the validity of the Adult Attachment Interview with this population, it was performed only after participants reached age 14. The choice of lower bound age is of course somewhat arbitrary, but it has been used with 14 year olds in the literature (Ammaniti, van IJzendoorn, Speranza, & Tambelli, 2000).

Romantic Relationship Aggression

At age 18, adolescents and their romantic partners were asked to rate their own physically abusive, blaming, and coercive behaviors during romantic conflicts with their partners using the Physical Abuse/Coercion subscale of the Conflict in Relationships (CIR) questionnaire (Wolfe, Reitzel-Jaffe, Gough, & Wekerle, 1994). This subscale is comprised of 15 items rated on a 4-point Likert scale with responses ranging from 1 (never happened) to 4 (happened 6+ times), so that the maximum score is 108. Summary scores of all 15 items were used, with higher scores indicating more frequent use of physically abusive and coercive tactics. Target adolescents’ reports of their own physically abusive, blaming, and coercive behaviors was conceptualized as representing adolescent perpetration, while romantic partners’ reports of their own physically abusive, blaming, and coercive behaviors was conceptualized as representing the target adolescents’ victimization, and each will be referred to as such from here on out. Both perpetration and victimization scales yielded high internal consistency (Cronbach’s alpha of .91, .79, respectively).

Romantic Relationship Satisfaction

At age 18, adolescents were asked to rate their relationship satisfaction on the Satisfaction subscale of the Network of Relationships Inventory (NRI; Furman & Buhrmester, 1985). This measure was designed to assess a variety of attachment/affiliative issues in adolescents’ relationships. The Satisfaction subscale consists of three items on a five-point Likert scale (1 being little or none and 5 being the most). The satisfaction subscale has marginal, internal consistency (Cronbach’s alpha of .57).


Preliminary Analyses

Means and standard deviations for all variables are presented, separately by gender, in Table 1. Boys and girls do not differ significantly on any of these variables. Distributions of all variables were examined for presence of outliers, defined as variables more than two standard deviations from the group mean. Only one outlier was found, for paternal aggression, and this was trimmed to the next highest value. For descriptive purposes, Table 2 presents partial correlations among all primary constructs examined in the study after accounting for gender and socioeconomic status. Among our sample, 53 of our 75 (70%) participants were classified as securely attached.

Table 1
Means and standard deviations of predictor and outcome variables of interest, separated by gender
Table 2
Partial correlations between predictor and outcome variables of interest (after partialling out gender and socioeconomic status)

Primary Analyses

To address the primary hypothesis that paternal aggression was predictive of subsequent adolescent perpetration of aggression toward the romantic partner via the moderating role of attachment orientation, a hierarchical regression analysis was performed. Adolescent perpetration of aggression at age 18 was regressed onto adolescent attachment security at age 14, paternal aggression at adolescent age 13, and the interaction between the two, after first accounting for the effects of adolescent gender and socioeconomic status. These results, as depicted in Table 3, indicated a significant main effect of paternal aggression, as well as a significant interaction between paternal aggression and adolescents’ attachment security. Further analyses suggested by Aiken & West (1991) to interpret interactions revealed that the relation of paternal aggression to perpetration of aggression in romantic relationships five years later was positive and significant for less secure adolescents, and indeed was nonsignificant for those who were secure (see Figure 1).

Figure 1
Target adolescents’ perpetration of romantic aggression as a function of paternal aggression for adolescents high and low in attachment security.
Table 3
Regressions predicting adolescents’ perpetration of aggression toward partner at age 18

Next, we examined whether paternal aggression is predictive of subsequent adolescent victimization by aggression via the moderating role of attachment orientation, again using a hierarchical regression analysis. Adolescent victimization at age 18 was regressed onto adolescent attachment security at age 14, paternal aggression at adolescent age 13, and the interaction between the two, after first accounting for the effects of adolescent gender and socioeconomic status. These results, as depicted in Table 4, indicated a significant main effect of paternal aggression, as well as a significant interaction between paternal aggression and adolescents’ attachment security. Further analyses (Aiken & West,1991) revealed that the relation of paternal aggression to victimization by aggression in romantic relationships five years later was far positive and significant for less secure adolescents, and again nonsignificant for those who were secure (see Figure 2).

Figure 2
Target adolescents’ victimization by romantic aggression as a function of paternal aggression for adolescents high and low in attachment security.
Table 4
Regressions predicting adolescents’ victimization by aggression from partner at age 18

Next, we examined whether paternal satisfaction was predictive of subsequent adolescent romantic relationship satisfaction, while also considering the moderating role of attachment orientation. Romantic relationship satisfaction at age 18 was regressed onto adolescent attachment security at age 14, paternal satisfaction at adolescent age 13, and the interaction between the two, after first accounting for the effects of adolescent gender and socioeconomic status. Adolescents whose fathers reported higher levels of marital satisfaction were more likely to report higher levels of satisfaction with their romantic partners five years later (see Table 5). However, no significant main effect of adolescents’ attachment security was found.

Table 5
Regressions predicting adolescents’ romantic relationship satisfaction at age 18

We also examined whether maternal satisfaction was predictive of subsequent adolescent romantic relationship satisfaction, while also considering the moderating role of attachment orientation. Romantic relationship satisfaction at age 18 was regressed onto adolescent attachment security at age 14, maternal satisfaction at adolescent age 13, and the interaction between the two, after first accounting for the effects of adolescent gender and socioeconomic status. Results were null, suggesting that maternal satisfaction is not related to adolescents’ satisfaction in subsequent romantic relationships.

Post Hoc Analyses of Domain Specificity

To investigate whether the above results were domain specific, hierarchical regressions were used to test for significant predictions across domains. First, romantic relationship aggression at age 18 was regressed onto adolescent attachment security at age 14, paternal satisfaction at adolescent age 13, and the interaction between the two, after first accounting for the effects of adolescent gender and socioeconomic status. This was followed by the same analysis of romantic relationship satisfaction at age 18. Results were null, revealing that paternal satisfaction was not predictive of subsequent romantic relationship aggression and paternal aggression was not predictive of subsequent romantic relationship satisfaction.

Further, while maternal aggression was not expected to predict subsequent perpetration and victimization by aggression in adolescents’ romantic relationships, it was of interest to investigate whether mothers had a role. Perpetration of aggression in romantic relationships was regressed onto adolescent attachment security, maternal aggression, and the interaction between the two, after first accounting for the effects of adolescent gender and socioeconomic status. This was followed by the same analysis predicting victimization by aggression. Results were null for both, suggesting that maternal aggression is not related to predict subsequent perpetration and victimization by aggression in adolescents’ romantic relationships.

Also, since adolescent boys and girls likely experience romantic relationships so differently, and since parents may have a different impact on same-sex children than on opposite-sex children, we thought it was important to consider a possible moderating role of adolescent gender in predicting perpetration, victimization and satisfaction as well. However, these analyses yielded no significant results.


In a recent review of research on adolescent romantic relationship development, Bryant and Conger (2002) proposed a theoretical model suggesting three potential mechanisms through which family of origin experiences may influence early romantic relationships later in life: observational learning, parent/sibling socialization, and/or behavioral continuity. This study confirms key elements of this theoretical model, specifically the observational learning hypothesis, but also extends it by considering the ways in which these processes may be moderated by the adolescent’s own degree of perspective on family interactions, assessed in terms of states of mind regarding attachment. However, the parent/sibling socialization and behavioral continuity hypotheses have not been tested here, and should be through future research. Further, the findings on intergenerational transmission of relationship qualities suggest that the emotional dynamics of the larger family system may play a key role in how offspring approach and interpret conflict and emotionality in their subsequent intimate relationships in adolescence.

As expected, primary analyses revealed that characteristics of the parents’ marital relationship were indeed a key predictor of characteristics of adolescents’ subsequent romantic relationships. Specifically, higher levels of paternal aggression during marital conflict were associated with greater perpetration of and victimization by aggression in romantic relationships five years later. However, upon further investigation it became clear that this association was strong for adolescents who exhibited low attachment security, but not for adolescents who exhibited high attachment security. Additionally, paternal satisfaction was associated with adolescents’ romantic relationship satisfaction five years later. This association, however, is maintained regardless of adolescents’ levels of attachment security. Furthermore, follow-up analyses suggested that these findings were domain specific, in that paternal satisfaction was not predictive of subsequent adolescent aggression, and paternal aggression was not predictive of subsequent adolescent satisfaction. Each of these findings is discussed in detail below.

The moderating role of adolescent attachment security in predicting adolescents’ romantic aggression from paternal aggression fits well with attachment theory in that it suggests that secure attachment may act as a buffer against replicating the destructive behaviors witnessed during childhood. Attachment security, as assessed by the Adult Attachment Interview (George, Kaplan & Main, 1996), reflects the ability to accurately assess childhood experiences, whether positive or negative, as well as to rationally and coherently communicate them. Further, Bowlby (1988) argues that individuals develop working models, or cognitive representations about themselves and others, that are internalized and subsequently affect their perceptions of others in future attachment relationships. Secure individuals are believed to have developed positive working models that allow them to process and integrate emotionally evocative information accurately and effectively. These secure working models are also carried forward into one’s intimate relationships later in life. In the case of interparental aggression, effective processing entails recognizing conflict as either constructive or destructive, which is a crucial cognitive-emotional regulatory process that helps to prevent offspring from replicating any observed destructive patterns in the future. Individuals who are less secure may not have the perspective to make sense of parental marital aggression. Instead, they may be unable to step back and more objectively evaluate their relationship with their parents and may have difficulty processing and coming to terms with the emotional significance of what they witnessed as children and early adolescents. These individuals may not possess the skills to fully distinguish constructive from destructive paternal aggression tactics, and thus may be more likely to perceive destructive tactics as normative, and in turn replicate them in their subsequent romantic relationships.

Not only did adolescent attachment security and paternal conflict tactics interact to predict adolescents’ conflict tactics, but they were also predictive of their romantic partners’ conflict tactics. This finding potentially highlights the role of partner selection, as adolescents low in security may choose partners who replicate their fathers’ destructive conflict tactics. Again, we can turn to Bowlby’s notion of working models of attachment to better understand how the intergenerational transmission of relational qualities may occur. Individuals may seek out romantic partners who behave consistently with their working models of romantic relationships, which in turn may be based on observations of their parents’ relationship. For example, individuals who have developed a working model of themselves as unloved and of others as rejecting and/or inconsistently available are likely to seek out these same qualities in a romantic partner.

The association between interparental aggression and subsequent offspring aggression can also be understood from the perspective of family systems. From a family systems perspective, parents often project, or focus emotional pressures or anxieties, onto their children, particularly during times of stress. When parents focus a great deal of their energies on their offspring, such offspring become less “differentiated”, and are less able to establish an individual identity, emotional autonomy and self directedness within their own romantic relationships, and instead adopt ways of relating that parallel those members of their nuclear family. Such individuals may be more prone to recreate a conflictual environment that they experienced in their family of origin. These less differentiated individuals are similar to those individuals with insecure working models, in that they are less able to develop adaptive ways of coping with conflict in the context of intimate relationships. Further, an important additional consideration is that less differentiated individuals often seek out partners who match their level of differentiation, which helps to spur the pattern of emotional dysfunction to repeat across generations (Kerr, 1981). However, the role of differentiation has not been directly tested here, and should be through future research.

Another potential explanation could be that the destructive conflict tactics exhibited by romantic partners are elicited by the target adolescents. Individuals who demonstrate an insecure attachment style are thought to more frequently rely on maladaptive strategies to help regulate feelings of interpersonal distress, such as exaggerating negative emotions to maintain partners’ attention or minimizing negative affect to avoid further threat to the relationship. However, adolescents’ use of such strategies may be escalating the conflict rather than resolving it, and thus their romantic partners may be more likely to respond in a destructive manner as well. Unfortunately, prospective information about romantic partners’ family context was not available so the role of their parents’ marital aggression or their own attachment history could not be accurately evaluated.

Also consistent with our hypotheses, paternal satisfaction was predictive of adolescents’ romantic relationship satisfaction five years later. This finding could potentially be attributed to modeling because adolescents may carry what they witness in their parents’ marital relationship into their own romantic relationships later in life. In this case, adolescents whose fathers exhibited satisfaction with their marital relationship reported satisfaction with their own romantic relationships later in life. One possible explanation could be that they’ve carried forward the more observable, positive behaviors that may co-occur with satisfaction, such as warmth and support (Conger, Cui, & Bryant, 2000). Alternately, adolescents may have simply carried forward a paternal style of more easily feeling or reporting satisfaction within relationships.

Additionally, it is possible that fathers withdraw from co-parenting interactions when they are dissatisfied with the marriage (Madden-Derdich & Leonard, 2000), potentially having a negative impact on children’s subsequent relationships. The association between dissatisfaction and paternal withdrawal from co-parenting suggests that the converse may also be true. When fathers are more satisfied, they may exhibit more positive co-parenting factors, such as increased father-child engagement, attention, and support, which may facilitate the development of more satisfactory offspring relationships in later adolescence. The contribution of such co-parenting behaviors on the link between paternal satisfaction and subsequent offspring relationship satisfaction should be explored in future research.

But why does attachment not moderate the link between paternal marital satisfaction and later romantic relationship satisfaction in this study as it does with aggression? The absence of a moderating effect of attachment may reflect that low levels of paternal satisfaction are generally not as emotionally distressing as high levels of paternal aggression, and thus require less of the affect processing described earlier. Thus, both individuals low and high in security may process information about fathers’ marital satisfaction similarly. While one potential explanation for the continuities observed in satisfaction may be based on modeling, an alternative explanation is that the continuities reflect heritability of a propensity to engage in or report satisfying marital relationships.

It is also important to note that post hoc analyses provided evidence for specificity of prediction. Results did not show that positive marital traits in one domain were predictive of a lack of negative romantic relationship traits in another domain during adolescence. Instead, paternal aggression was predictive of adolescents’ romantic aggression five years later but not of adolescents’ romantic satisfaction. Likewise, paternal satisfaction was predictive of adolescents’ romantic satisfaction but not of adolescents’ romantic aggression. This apparent domain specificity is of particular interest because it highlights the importance of providing children with a good all-around model of healthy relationships. Low aggression is not enough to prevent a negative impact, nor is high satisfaction enough to ensure a positive impact on an adolescent’s psychosocial development. Instead, exhibiting well-balanced relationship skills in all domains is optimal for positive development of working models regarding romantic relationships.

There are several limitations to these findings. First, causal relationships cannot be inferred from these results because even longitudinal data are not logically sufficient to establish causal relationships. Further, the internal consistency of the romantic relationship satisfaction subscale used is problematic. This is likely a symptom of the measure being composed of only three items, and is a possible explanation for lack of findings with this measure. Also, we recognize that due to our limited sample size, power will be such that we may fail to detect some real effects in the data. However, power is demonstrated to be sufficient under almost all circumstances. Nevertheless, we feel that we are careful not only in interpreting results, but also in refraining from drawing strong conclusions from the absence of significant findings in analyses in which power is not ideal.

Future research should include observational assessments to move beyond possible reporting biases created by reliance on self-reports of satisfaction. Although this study focused on a normative community sample of adolescents from two-parent families, and thus these results cannot be generalized to more at-risk or single-parent families, it raises the possibility that youth who are exposed to even more volatile interparental relationships may be particularly susceptible to intergenerational transmission of relationship aggression. Further research might assess the extent to which the “buffer” role of attachment security generalizes to more problematic family situations, and whether interventions focusing on attachment security and the development of positive working models of romantic relationships may be beneficial to such populations. Also, an investigation of the impact of changing interparental relationships over time may important.

Taken together, these findings support the notion that possessing a balanced, coherent and objective view on parental experiences, even if these experiences were negative, is advantageous to one’s own relationship outcomes. These results highlight the importance of clinical intervention, in allowing individuals a context to explore, reframe and clarify the meaning of harsh early experiences. Such “corrective therapy experiences” may facilitate increased insight about one’s early childhood experiences that is more characteristic of secure states of mind. As a result, this may allow individuals to enter into their own intimate relationships with healthier expectations and conflict management behaviors, in turn breaking the cycle of abuse (Egeland, Jacobvitz, & Sroufe, 1988).


Publisher's Disclaimer: The following manuscript is the final accepted manuscript. It has not been subjected to the final copyediting, fact-checking, and proofreading required for formal publication. It is not the definitive, publisher-authenticated version. The American Psychological Association and its Council of Editors disclaim any responsibility or liabilities for errors or omissions of this manuscript version, any version derived from this manuscript by NIH, or other third parties. The published version is available at


  • Aiken L, West S. Multiple regression: Testing and interpreting interactions. Thousand Oaks, CA, US: Sage Publications, Inc; 1991. Interactions between continuous predictors in multiple regression; pp. 9–27.
  • Ainsworth M, Bowlby J. An ethological approach to personality development. American Psychologist. 1991;46(4):333–341.
  • Allen JP, Hauser ST, O’Connor TG, Bell KL, Eikholt C. The connection of observed hostile family conflict to adolescents’ developing autonomy and relatedness with parents. Development and Psychopathology. 1996;8:425–442.
  • Amato P. Explaining the intergenerational transmission of divorce. Journal of Marriage & the Family. 1996;58(3):628–640.
  • Ammaniti M, Van Ijzendoorn M, Speranza A, Tambelli R. Internal working models of attachment during late childhood and early adolescence: An exploration of stability and change. Attachment & Human Development. 2000;2(3):328–346. [PubMed]
  • Archer J. Sex differences in physically aggressive acts between heterosexual partners: A meta-analytic review. Aggression and Violent Behavior. 2002;7(4):313–351.
  • Bandura A. Social learning theory. Englewood Cliffs, NJ: Prentice-Hall; 1977.
  • Berger J, Antonishak J, Allen J. Marital conflict tactics as predictors of adolescents’ conflict behaviors with peers. Poster presented at the biennial meeting of the Society for Research on Adolescence; San Francisco, CA. 2006. Mar,
  • Black KA. Associations between adolescent-mother and adolescent-best friend interactions. Adolescence. 2002;37:235–253. [PubMed]
  • Bowen M. The use of family theory in clinical practice. Comprehensive Psychiatry. 1966;7(5):345–374. [PubMed]
  • Bowlby J. A secure base: Parent-child attachment and healthy human development. New York, NY, US: Basic Books; 1988.
  • Bryant C, Conger R. And intergenerational model of romantic relationship development. In: Vangelisti A, Reis H, Fitzpatrick M, editors. Stability and change in relationships. Cambridge, U.K: Cambridge University Press; 2002. pp. 57–82.
  • Conger R, Cui M, Bryant C. Competence in early adult romantic relationships: A developmental perspective on family influences. Journal of Personality and Social Psychology. 2000;79(2):224–237. [PubMed]
  • Crowell, Treboux, Brockmeyer Parental divorce and adult children’s attachment representations and marital status. Attachment & Human Development. 2009;11(1):87–101. [PubMed]
  • Cummings M, Davies P, Campbell S. Understanding the impact of marital conflict on children. The Brown University Child and Adolescent Behavior Letter. 2001;17:1.
  • Cummings EM, Goeke-Morey MC, Papp LM. Children’s responses to everyday marital conflict tactics in the home. Child Development. 2003;74:1918–1930. [PubMed]
  • Cummings EM, Goeke-Morey MC, Papp LM, Dukewich TL. Children’s responses to mothers’ and fathers’ emotionality and tactics in marital conflict in the home. Journal of Family Psychology. 2002;16:478–492. [PubMed]
  • Davies P, Cummings M. Marital conflict and child adjustment: An emotional security hypothesis. Psychological Bulletin. 1994;116(3):387–411. [PubMed]
  • Dodge KA. A social information processing model of social competence in children. In: Perlmutter M, editor. Minnesota Symposium on Child Psychology. Vol. 18. Hillsdale, NJ: Erlbaum; 1986. pp. 77–125.
  • Egeland B, Jacobvitz D, Sroufe LA. Breaking the cycle of abuse. Child Development. 1988;59:1080–1088. [PubMed]
  • El-Sheikh M, Elmore-Staton L. The link between marital conflict and child adjustment: Parent-child conflict and perceived attachments as mediators, potentiators, and mitigators of risk. Development and Psychopathology. 2004;16(3):631–648. [PubMed]
  • Feldman S, Gowen L, Fisher L. Family relationships and gender as predictors of romantic intimacy in young adults: A longitudinal study. Journal of Research on Adolescence. 1998;8(2):263–286.
  • Fite JE, Bates JE, Holtzworth-Munroe A, Dodge KA, Nay SY, Pettit GS. Social information processing mediates the intergenerational transmission of aggressiveness in romantic relationships. Journal of Family Psychology. 2008;22(3):367–376. [PubMed]
  • Furman W, Buhrmester D. Children’s perceptions of the personal relationships in their social networks. Developmental Psychology. 1985;21:1016–1022.
  • Furman W, Flanagan A. The influence of earlier relationships on marriage: An attachment perspective. In: Halford W, Markman H, editors. Clinical handbook of marriage and couples interventions. Hoboken, NJ, US: John Wiley & Sons, Inc; 1997. pp. 179–202.
  • Furman W, Shaffer L. The role of romantic relationships in adolescent development. In: Florsheim P, editor. Adolescent romantic relations and sexual behavior: Theory, research, and practical implications. Mahwah, NJ, US: Lawrence Erlbaum Associates, Publishers; 2003. pp. 3–22.
  • George C, Kaplan N, Main M. Adult Attachment Interview. 3. Department of Psychology, University of California; Berkeley, CA: 1996. Unpublished manuscript.
  • Harold GT, Shelton KH, Goeke-Morey MC, Cummings EM. Marital conflict, child emotional security about family relationships and child adjustment. Social Development. 2004;13:350–376.
  • Joyner K, Udry J. You don’t bring me anything but down: Adolescent romance and depression. Journal of Health and Social Behavior. 2000;41(4):369–391. [PubMed]
  • Kerr ME. Family systems theory and therapy. In: Gurman AS, Kniskern DP, editors. Handbook of family therapy. New York, NY: Brunner/Mazel, Inc; 1981. pp. 226–264.
  • Kobak, Cole, Ferenz-Gillies, Fleming, Gamble Attachment and emotion regulation during mother-teen problem solving. A control theory analysis. Child Development. 1993;64:231–245. [PubMed]
  • Lichter D, McCloskey L. The effects of childhood exposure to marital violence on adolescent gender-role beliefs and dating violence. Psychology of Women Quarterly. 2004;28(4):344–357.
  • Madden-Derdich D, Leonard S. Parental role identity and fathers’ involvement in coparental interaction after divorce: Fathers’ perspectives. Family Relations. 2000;49(3):311–318.
  • Main M, Goldwyn R. Adult attachment rating and classification system: Manual in draft. University of California; Berkeley, CA: 1994. (Version 7.1). Unpublished manuscript.
  • Moore S, Leung C. Young people’s romantic attachment styles and their associations with well-being. Journal of Adolescence. 2002;25:243–255. [PubMed]
  • Ozen D. The impact of interparental divorce on adult attachment styles and perceived parenting styles of adolescents: Study in Turkey. Journal of Divorce and Remarriage. 2003;40(12):129–149.
  • Phares V, Compas B. The role of fathers in child and adolescent psychopathology: Make room for daddy. Psychological Bulletin. 1992;111(3):387–412. [PubMed]
  • Richards M, Crowe P, Larson R, Swarr A. Developmental patterns and gender differences in the experience of peer companionship during adolescence. Child Development. 1998;69(1):154–163. [PubMed]
  • Schudlich T, Shamir H, Cummings E. Marital conflict, children’s representations of family relationships, and children’s dispositions towards peer conflict strategies. Social Development. 2004;13:171–192.
  • Shaw D, Emery R. Parental conflict and other correlates of the adjustment of school-age children whose parents have separated. Journal of Abnormal Child Psychology. 1987;15(2):269–281. [PubMed]
  • Spanier G. Measuring dyadic adjustment: New scales for assessing the quality of marriage and similar dyads. Journal of Marriage & the Family. 1979;38(1):15–28.
  • Straus M. Measuring intrafamily conflict and violence: The Conflict Tactics Scales. Journal of Marriage and the Family. 1979;41:75–85.
  • Ward M, Carlson A. Associations among attachment representations, maternal sensitivity, and infant-mother attachment in a sample of adolescent mothers. Child Development. 1995;66:69–79. [PubMed]
  • Wenz F. Sociological correlates of alienation among adolescent suicide attempts. Adolescence. 1979;14(53):19–30. [PubMed]
  • Wolfe D, Reitzel-Jaffe D, Gough R, Wekerle C. Conflicts in relationships: Measuring physical and sexual coercion among youth. Department of Psychology, the University of Western Ontario; London, Canada, N6A C52: 1994. Unpublished manuscript. Available from the Youth Relationships Project.