Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2024 Apr 25.
Published in final edited form as: J Child Fam Stud. 2021 Apr 27;30(6):1540–1553. doi: 10.1007/s10826-021-01959-7

Listening In: How Parent-Child Communication Relates to Social and Physical Aggression

Diana J Meter 1, Samuel E Ehrenreich 2, Kurt Beron 3, Marion K Underwood 4
PMCID: PMC11045039  NIHMSID: NIHMS1932819  PMID: 38666246

Abstract

In this study, we investigated the relations between features of parent-child conversations (neutral talk, positive and negative in-dyad and out-dyad talk) and children’s social and physical aggression from ages 9–18. Participants were 297 youth (52% girls) of about 9 years old at Time 1 and their parent. Fifty-two percent of this United States sample identified as White, 20% Black, 20% Hispanic, 8% other races/ethnicities. One hundred eighty-seven parents participated in the parent-child observation task. Ninety four percent of parent participants were mothers. Parent-child conversations were observed in the laboratory during preadolescence, and teachers reported child’s aggression. Using multinomial logit analyses, we found that coded observations of communication features predicted membership in linear trajectories of social and physical aggression across nine years of adolescence; trajectories were derived via mixture modeling. Parent and child communication characteristics were related to trajectories of aggression that spanned preadolescence and adolescence; however, not all predicted associations were significant. Children’s talk about neutral topics predicted being on a lower social aggression trajectory. Positive out-dyad talk from children was related to being on a lower physical aggression trajectory, as was parent in-dyad positive talk. After controlling for other factors, neither parent nor child in- or out-dyad negative talk was associated with social or physical aggression. These findings highlight the importance of positive communication by youth and toward youth in association with long-term social adjustment.

Keywords: Parent-child relations, Communication, Family communication patterns, Social aggression, Physical aggression

The developmental origins of social and physical aggression include interactions with parents (Dodge et al., 2006; Ladd & Pettit, 2002). The ways in which parents and children communicate might relate to children’s later physical and social aggression. According to the Family Communication Pattern perspective, examining family discourse provides the opportunity to study the mechanism that leads to teaching and learning values, attitudes, and beliefs from the family (Vieira, 2015). Families with a conversation orientation encourage members to participate in spontaneous and frequent unrestrained communication about feelings, thoughts, and activities. In these families, parents communicate values and ideas to children within conversations, and parents see communication as the primary way to educate and socialize their children. “Conversation orientation in families leads children to acquire better conflict communication skills and more tools to mitigate the negative consequences of interpersonal conflict” (Koerner & Fitzpartrick, 2002, p. 48). We would expect there to be more general communication and chat within conversation orientation families, whether this communication is emotionally-valanced or not. Families who do not have a strong conversation orientation interact infrequently and discuss few topics openly (Koerner and Fitzpatrick, 2002). In contrast to a conversation orientation, other families have a conformity orientation, in which the caregiver wields power and control, and the communication fosters conformity to the parent. Youth in this type of family environment may suppress feelings and feel less sympathy for those with perspectives different from the ones parents promote. Compared to conversation orientation families, children in conformity orientation families may communicate less frequently with their parents specifically about neutral topics, because parents are typically the ones driving the conversation and directing conformity to parents. This study investigated how specific features of parent-child conversations observed in the laboratory related to children’s social and physical aggression from ages 9–18. The four features investigated included neutral talk (e.g. “It may rain today”), positive in-dyad talk (e.g. “I can’t wait to spend this rainy day with you”), positive out-dyad talk (e.g. “How fun is it to play in the rain with your friends?!”), negative in-dyad talk (e.g. “The rain is making you crabby”), and negative out-dyad talk (e.g. “The rain is making the neighbors crabby.”)

Aggression

Adolescents’ aggression can take different forms. Physical aggression includes intentional physical harm to another (hitting or shoving; Card et al., 2008). Social aggression, and similar constructs such as indirect and relational aggression, include friendship manipulation, social exclusion, and gossip (Archer & Coyne, 2005); these behaviors are also harmful (Galen & Underwood, 1997; Paquette & Underwood, 1999; Vaillancourt, 2005), but can be more difficult to observe. Both forms of aggression are prevalent among adolescents, and although correlated, tend to lead to distinct consequences for perpetrators (Card et al., 2008). Specifically, perpetrating physical aggression is associated with antisocial behavior (McEachern & Snyder, 2012) such as drinking and smoking, attention and academic problems (Henriksen et al., 2020), externalizing problems, social skills, work habits, and overall school performance (Campbell et al., 2010). Perpetrating social aggression has been related to externalizing difficulties, but also internalizing difficulties such as anxiety, depression, feelings of loneliness and social isolation (Heilbron & Prinstein, 2008).

Later childhood into adolescence is a particularly important time to study social and physical aggression because although much previous research has shown these behaviors to be present across development, their prevalence changes over these years. During early childhood, physical and (Côté et al., 2007), social aggression are observable (Crick et al., 2006). Children generally exhibit decreasing levels of physical aggression and social aggression across childhood (Côté et al., 2007; Ehrenreich et al., 2014), but by early adolescence, some subgroups begin to increase in social or relational aggression (Côté et al., 2007; Ehrenreich et al., 2014). It is assumed that in order to successfully enact these forms of aggression, advanced social, cognitive, and verbal intelligence is required, which may not develop until closer to adolescence (Björkqvist, 1994; Vaillancourt et al., 2007). It is also important to note that although these low incidence behaviors are observed fairly universally, some youth generally do not engage in aggressive behavior, even over time (Ehrenreich et al., 2014; Campbell et al., 2010).

Previous research has examined the shape of possible trajectories of physical and social aggression across early to late adolescence. For example, a previous investigation with the current sample using a different analytic approach, dual trajectory group-based mixture modeling, showed three trajectories each for physical and social aggression: stably‐low, stably‐medium, and high‐desisting for physical aggression and low‐, medium‐, and high‐desisting for social aggression (Ehrenreich et al., 2016). In a study of children ages 10–15, for indirect (social) aggression, a low declining, moderate declining and high trajectory were found. For the physical aggression trajectory, a moderate-declining trajectory, and a high increasing trajectory were observed (Cleverley et al., 2012). In a different sample of 11 to 18-year-olds who self-reported aggression five times over 2.5 years, physical and social aggression followed curvilinear trajectories, with physical aggression peaking near age 15 and social aggression peaking about a year before (Karriker‐Jaffe et al., 2008).

Putting this previous research together, for physical aggression, we expect to see most participants in a declining trajectory from age 9 to 18, with a second group engaging in no aggression across time, and a third, atypical group, engaging in higher-than-average levels of physical aggression across time. For social aggression, we expect our largest group of participants to follow a curvilinear trajectory, with social aggression peaking around early adolescence and decreasing as they reach 18 years old. Here too, we may observe a no-social aggression group, and an atypical group that remains high in social aggression across development.

Interactions with Parents

Little previous research has investigated how the ways in which parents talk with their children might relate to the child’s later physical and social aggression. The family is a context in which children acquire social skills or deficits, and the parents’ child-rearing style may serve as a model that children imitate in their peer interactions (Ladd, 2005). Family interactions may influence competence with peers directly through skill-building, but also “…indirectly, by modeling, shaping, or otherwise affecting socializing skills, skill deficits, or behavioral excesses during everyday family interactions, relationships, and activities,” (Ladd, 2005, p. 213–214).

Although not frequently studied, communication in conversation-orientation families may provide youth with the opportunity to develop interpersonal skills that lead to positive interactions with peers. In a study of girls and their parents, frequent conversation was associated with girls engaging in more sympathy and perspective-taking (Vieira, 2015). The authors concluded that children who engage in two-way communication with parents and whose parents give them the opportunity to engage in decision-making experience communication that fosters sympathy and perspective-taking. Perspective taking and sympathy were related to perceiving violence as wrong. Although these paths were significant, there was no evidence of perspective-taking or sympathy mediating these pathways (Vieira, 2015). In another study, preadolescents who reported more effective parent-child communication were involved in less aggression (including physical, verbal, and indirect aggression) at school (Lambert & Cashwell, 2004).

Parents’ and children’s positive and neutral communication may protect against a tendency to put others down or harm others. Reciprocal neutral talk between parents and children suggests that parents and children listen to each other in conversation. Practice speaking with, and listening to a trusted family member may protect against engaging in aggression. In one study, adolescents who rated their parents higher on positive items about parents being good listeners reported less social and physical aggression (Wallenius et al., 2007). In another study, parent-child communication, as measured with a questionnaire, was related to a more harmonious school life and indirectly related to less bullying through decreased peer pressure (Lee & Wong, 2009).

Adolescents who engage in positive talk directed within the parent dyad would be expected to engage in less aggression. More positive comments directed at the communication partner may be a sign of a high quality, supportive relationship. Positive parenting (including nurturance, a positive relationship, discussion, warmth, acceptance, validation, open communication, intimacy) was found to be negatively related to relational aggression in a meta-analysis (Kawabata et al., 2011).

Positive talk about others outside the dyad may have positive effects on youth in that this behavior models and reinforces kind, supportive, and complimentary talk about others. This modeling of positive talk about others might decrease the social acceptability of mean gossip about others and increase the likelihood that children will talk about others in positive ways. Children who observe their parents engaging in positive talk about others may learn from their parent’s modeling of prosocial behavior and engage in less aggression (Vollet et al., 2019).

Contrarily, in families in which caregivers foster conformity through conversation, youth may have less practice engaging in the type of open and less conflictual communication that one would expect among youth who do not engage in conflict. In a study of boys, poor parent-child communication, defined as parents and their child not communicating about the child’s emotions and problems (e.g. the parent not trying to find out why a child is upset), was related to more reactive aggression. Poor parent-child communication also increased the association between reactive aggression and adjustment problems (Fite et al., 2014).

The Current Research

The current study examined how observed features of interactions between nine-year-old children and their parent related to membership in long-term developmental trajectory groups for social and physical aggression as assessed yearly by teachers from ages 9–18. We expected positive and neutral talk to be related to being in a lower social and physical aggression trajectory group and that negative in-dyad and out-dyad communication from parent and child would predict following higher social and physical aggression trajectories. This investigation extends prior research by examining how observed features of parent-child conversations predict growth and change in both social and aggression from late childhood through adolescence.

Method

Participants

Participants were a community sample of children and their parent in a suburban area outside of a large, metropolitan city in the United States. Child participants were recruited through their schools, although only a few students from each school participated. Participants changed schools throughout the nine years of the longitudinal study. Child participants were 297 youth of about 9 years old at Time 1 (T1). Descriptive statistics are based on T1 parent-reported data. One hundred fifty-five participants were girls and 142 were boys. Fifty-two percent identified as White, 20% Black, 20% Hispanic, 4% Multi-racial, 1% Middle Eastern, and less than 1% as Asian or Native American; 2% of participants did not report race/ethnicity. The sample was representative of the population in the two counties from which it was drawn (United States Census Bureau, 2016). All participants (N = 297) were included in the modeling of the aggression trajectories. Parents who participated in the study were the parent most knowledgeable about the child; 94% of parents who reported their gender were women. Sixty-nine percent of parents were married, 15% were divorced, 7% were never married, 4% were remarried, 3% were separated, less than 1% was widowed, and 2% did not report marital status. Approximately 16% of the sample families earned less than $25,000 annually, 16% earned $26,000-$50,000, 17% earned $51,000-$75,000, 21% earned $76,000-$100,000, 6% earned over $100,000, and 28% did not report income. Median reported income ($51,000-$75,000) was slightly higher than the median reported in the recruitment counties (United States Census Bureau, 2000).

Participation Rates

Data for this study were collected over the course of nine years. Families were initially recruited to participate in 3rd grade, with the first observational data collected during 4th grade. Aggression trajectories (described below) were estimated with data from any participants with teacher-reported ratings of aggressive behavior out of the possible 10 data points (3rd grade data point not included in the current study; N = 297). One hundred ninety-five parent/child dyads participated in the first observational interaction task during 4th grade, however eight participant dyads did not have the necessary aggression ratings needed to be assigned a trajectory classification. Accordingly, 297 participants contributed to the aggression trajectory model estimation, but data for only 187 parent/child dyads were available in the predictor models (described in the data analysis section). There were no significant differences in social (χ2 = 0.39, ns) or physical aggression (χ2 = 0.10, ns) trajectories between those who participated in the 4th grade observational task and those who did not. There were also no differences between these two groups in regard to gender (χ2 = 0.15, ns) or parent marital status (χ2 = 0.00, ns). There was some difference noted in the income between those in the trajectory sample that were not in the observation sample, which may reflect a slightly higher attrition rate for those with higher incomes, in that they left the study at higher rates. More information about the handling of missing data is provided in the plan of analysis.

Procedure

Informed consent from parents and assent from children was obtained for research with human subjects. Participation in all aspects of the study, including the laboratory observation and teacher-reports on students, were completely voluntary. Teacher reports were collected from the child’s teacher during the Spring of each school year from 4rd through 12th grade. Fourth grade ratings of aggression were used as the first time point of assessment, and ratings from nine years were used to estimate the aggression trajectories (grades 4–12). Teacher ratings were provided by the child’s classroom teacher in grades 4–6, by the language arts teacher in grade 7 (who taught participants 2 periods per day), and by the teacher whom the participant nominated as knowing the child best in grades 8–12. Parents and children were compensated for their time at the conclusion of each visit ($25 during grades 4–7, $50 during grades 8–12). Teachers were compensated $25 for each student evaluated in all 10 grades.

Parents and children were asked to visit the research lab near the child’s 10th birthday during 4th grade and complete a videotaped observation task. A trained research assistant (RA) led parents and children through the consent process, built rapport with the families, and explained the lab activity to the parent and child before bringing the dyad into a private room where the videotaping occurred. The RA was not present in the room during the conversations in an attempt to decrease reactivity that might occur if the RA was intrusively present during the conversations (McKetchnie, 2012). The RA explained that the parent and child would be given some topics to talk about and that each task would last four minutes. Participants were asked to try to talk about the topic for the whole four minutes, but if they finished early, they could talk about something else. The parent and child were given six topics to discuss and were prompted with a sheet of questions to guide their conversation. The RA entered the room between each task to provide new prompts for the next task. The tasks included: (1) discuss friendships, (2) plan a birthday party, (3) discuss a talented peer at school, (4) discuss how kids treat each other, (5) discuss a recent disagreement between the child and parent, and (6) discuss positives about the family. Each task was timed for four minutes, with the total time varying due to pauses when the RA switched tasks. The RA could go back into the room if they needed to clarify tasks or give additional instructions. On average, all six tasks together took 27.39 min to complete (range = 15.62–33.26).

Measures

Parent-child interaction micro-coding

Undergraduate research assistants (RA) were trained and practiced coding videos until they reached acceptable interrater reliability (κ = 0.40–0.60 is considered moderate, κ = 0.61–0.80 = substantial; Landis & Koch, 1977). The RA coded each utterance made by the parent or child in the video moment by moment (codes described below). The total number of codes in each category was summed and divided by the total number of seconds the parent and child were involved in the lab observation task. This value was then multiplied by 60 to equal the rate per minute. Twenty percent of the videos were randomly selected to be double-coded. This sample of codes was used to assess inter-rater reliability between RAs (κs included below).

Sillars (1991) refers to observations of interpersonal relationships as “rigorous and intensive analysis” of interactional routines, (p. 198). In observational research, researchers may cluster qualitative or observed responses into categories based on the information received from participants (e.g. Samp et al. 2006). The number of instances of a particular interpersonal behavior or utterance may be observed and recorded as a frequency within a specific time interval (e.g. Aspland & Gardner, 2003; Galen & Underwood, 1997; Kahlbaugh & Haviland, 1994). Researchers may have preconceived coding schemes, but others, like in our study, rate observations after the observation takes place (Lytton, 1973).

Child and parent out-dyad positive or neutral talk

This code referred to positive out-dyad talk and neutral talk about others outside of the dyad. This code encompassed kindness or positivity expressed about others, but also neutral comments about others. Examples include, “Did you ever call Bobby?” “Alicia is very good at swimming.” κ = 0.72 for child out-dyad positive or neutral talk, and κ = 0.70 for parent out-dyad positive or neutral talk.

Child and parent neutral talk

This code represented any neutral talk about the self or the relationship (in-dyad neutral talk) and could include comments about the lab or lab tasks, and other talk that lacked emotional valence or did not fit into another category for which the research team coded (e.g. “Do you think it’s still raining?” “Your hair is more of a dirty blond than a light brown”). κ = 0.70 for child neutral talk and κ = 0.71 for parent neutral talk.

Child and parent in-dyad positive talk

This code encompassed all kind, supportive, and warm talk directed from one communication partner to the other. It could include compliments and positive comparisons, discussing plans for shared time with each other, and discussing positive shared memories. Examples include, “I need to be more careful about how I hear what you say,” and “You’re getting really good at handling the ball,” κ = 0.60 for child in-dyad positive talk, and κ = 0.65 for parent in-dyad positive talk.

Child and parent out-dyad negative talk

This code was applied whenever a child or parent made negative statements about others outside of the dyad. This included negative or critical talk, negative sarcasm, negative imitation or mocking, and comparisons that communicate something negative about someone outside the dyad (e.g. “She does terribly in math class;” “Her mom is snobby”). κ = 0.72 for child out-dyad negative talk, and κ = 0.69 for parent out-dyad negative talk.

Child and parent in-dyad negative talk

A code for negative talk was assigned to any negatively-valanced communication that occurred within the dyad, either about the other individual in the dyad or about the dyad’s relationship. This code was applied when there was any talk meant to be negative or hurtful, including threats, sarcastic utterances, name-calling, mocking, and laughing at the dyadic partner in a negative way (e.g. “Don’t act like such an idiot!” “Why are you mad at me?”). This code encompassed mean, coercive, manipulative communication directed from parent to child or child to parent, and could include constructs such as parental harshness, aggression, and psychological control. κ = 0.50 for child in-dyad negative talk, and κ = 0.60 for parent in-dyad negative talk.

Social aggression

Children’s social aggression was reported by the children’s teachers using a revised version of the Children’s Social Behavior Scale (CSBS-T; Crick, 1996). Three items from the CSBS-T were included as indicators of social aggression including, “This student ignores people or stops talking to them when he/she is mad at them.” A fourth item, “This child makes mean faces to hurt other kids’ feelings or to make them feel left out” was added to measure this non-verbal socially aggressive behavior. Items were rated on a scale from 1 = This is never true of this student to 5 = This is almost always true of this student. Items were averaged within each time point to form the social aggression construct. The CSBS-T has been shown to have strong psychometric properties across middle childhood and adolescence (Crick, 1996; Underwood et al., 2009). In a previous study utilizing the same dataset, reliability ranged from α = 0.75–0.95 (Ehrenreich et al., 2016).

Physical aggression

Physical aggression was also reported by the child participant’s teacher using the CSBS-T. Four items were used to measure physical aggression including, “This student hits, shoves, or pushes others.” These items were averaged within each time point to form the physical aggression construct. In a previous study with this same sample, reliabilities of teacher ratings of physical aggression ranged from 0.75–0.95 (Ehrenreich et al., 2016). For older elementary children, teacher reports of physical aggression using the original version of this questionnaire were found to be positively correlated with peer nominations of relational aggression (for girls, r = 0.74, p < 0.001 and for boys, r = 0.69, p < 0.001, Crick, 1996).

Social and physical aggression trajectories

First, we examined descriptive statistics and correlations for social and physical aggression. Second, we estimated unconditional mixture models that classified students separately into social and physical aggression trajectory classes (Nagin, 2005). Finally, we included predictors in the model using a three-step procedure that allowed for the incorporation of uncertainty in the estimation of the constructed trajectories (Asparouhov & Muthén, 2014). The models were estimated using Mplus 7.1 (Muthén & Muthén, 1998–2015). The metric of the social and physical aggression teacher ratings’ variables were continuous and peaked at the lowest value (one) and then skewed out to the maximum value (five). Following the recommendation of Nagin (2005), we analyzed the natural logarithm of the variables to account for the skewed nature of the data. Furthermore, students identified as not being aggressive (the lowest value) in a given rating might not actually be aggressive or may just not have been aggressive at the time point. To account for this, the aggression variables were modeled as censored-inflated, allowing there to be a mixture of both types of non-aggression observed (Muthén et al., 2016). Social aggression followed high-desisting, medium-desisting, and low-desisting trajectories and physical aggression followed high-desisting, stably-medium, and stably-low trajectories for physical aggression. These labels represent behavior observed among our typically-developing sample and do not refer to specific clinical levels of aggression. Trajectories are presented in Fig. 1. For the social aggression trajectory, the mean number of data points for which teacher-reported data were available was 5.63 (SD = 2.56), and for physical aggression, M = 5.63, SD = 2.57. The percent missing data ranged from 24–52%.

Fig. 1.

Fig. 1

Open in a new tab

Trajectories for Teacher-Rated (a) Social Aggression and (b) Physical Aggression from Grades 4–12 with full trajectory sample percentages shown

Data Analysis

In order to account for missing data in the construction of the trajectories, we used a robust maximum likelihood approach that allowed all observations to contribute to the estimated results and allowed for distributional misspecifications (Muthén and Muthén, 1998–2015). Missing data were handled using Full Information Maximum Likelihood (FIML). Over the course of this project (4th–12th grade) there were students who entered and exited the study at various times. We included all observations (N = 297) so that their information contributed to the formation of trajectories. The data we have on those in 4th grade were used as predictors (n = 187) and were thus a subset of these. However, as described below, we incorporated the information from the trajectory estimation into the analysis to correct for this two-stage approach.

We adopted the following decision rules based on previous research. Each of the trajectory models (social and physical) was separately estimated both in linear and quadratic form after joint estimation attempts demonstrated significant correlations between the two measures leading to convergence issues. Following general practice (Wickrama et al., 2016), we used Bayesian information criterion (BIC) values along with the Lo-Mendell-Rubin (LMR) likelihood ratio test and bootstrapped likelihood ratio test (bLRT) to choose the number of classes and polynomial level (Fanti & Henrich, 2010; Feldman et al., 2009). The BIC, with the lowest preferred, takes into account both the model log-likelihood and the complexity of the model. Models that did not converge, and therefore did not produce a likelihood value for the construction of a BIC, were eliminated from consideration after assessing the reasons for non-convergence (primarily very small class sizes). The LMR and bLRT provide evidence as to whether a model with more classes is statistically distinguishable from a model with one less class. We analyzed candidate models for their internal validity using methods developed for mixture models. We assessed the reliability of the results of the models by computing the average posterior probability of assignment (AvePP) and the entropy. The final step involved including the parent-child communication predictors. One approach to estimating this type of model is simultaneous estimation of the predictors with the trajectories. The issue with this is that the trajectories themselves will be influenced by the choice of predictors. An alternative approach has been introduced that allows for the separate estimation of trajectories, but the fact that they are estimated is allowed to affect the predictor estimation (Vermunt, 2010; Asparouhov & Muthén, 2014). This three-step approach first estimates the trajectories, then determines the most likely class for each observation, and finally includes the predictor variables in the model, accounting for misclassification in class assignment from the earlier step. This method does not alter the trajectories when predictors are added, but does account for the fact that the trajectories are themselves estimated imperfectly.

Results

We began our trajectory analysis by estimating linear models with one class up through four classes followed by quadratic versions of these models. The social aggression trajectories, with the exception of a four-class quadratic model, converged to the same solution from different starting values. The physical aggression estimations, when specified as freely chosen classes, never converged. About 27 percent (n = 77) of students were always rated in the lowest physical aggression category by teachers. In comparison, there were about 3 percent (n = 9) of students in the lowest social aggression category when they were rated. Therefore, we re-estimated the models allowing for a zero class, after which convergence issues were no different from those encountered with the social aggression trajectories.

We evaluated the BICs for these models, shown in Table 1. For both the social and physical aggression trajectories, a three-class, linear model was preferred. We then estimated the LMR and bLRT, as discussed previously, for further evidence that a three-class solution was appropriate. Both tests clearly favored a three-class linear solution over a two-class solution for the social trajectories (LMR: p = 0.002; bLRT: p = 0.000), whereas a four-class linear model was favored by LMR, unlike the bLRT that favored a three-class linear model (LMR: p = 0.001; bLRT = 0.375). However, the four-class solution included one class that contained a single observation, so we chose the three-class model for social aggression. Similarly, the physical aggression four-class model was favored by LMR while the bLRT had difficulty converging (LMR: p = 0.044). Again, a very small class in the four-class model (N = 3) led us to choose three-classes for physical aggression.

Table 1.

Bayesian information criteria (BIC) from various specifications

Trajectory Linear BIC Quadratic BIC
Social aggression
1 class 2636.028 2670.172
2 class 2449.294 2533.234
3 class 2441.292 2570.825
4 class 2465.073 Never converged
Physical aggression
1 class 2608.564 2610.206
2 class 2486.284 2490.554
3 class 2385.357 2524.792
4 class 2408.395 Never converged
Open in a new tab

The trajectories are shown in Fig. 1. The social aggression trajectories showed high (n = 30; 16%), medium (n = 111; 59.4%), and low (n = 46; 24.6%) classes. Wald tests were used to test for a difference in each of the intercepts (high vs. medium, etc.). In each case there was a significant difference (p < 0.000). We also tested whether the grade 12 endpoints were statistically significantly different from each other. We found statistical differences both for the high versus medium group (p = 0.004) and the medium versus low group (p < 0.000).

An analogous approach was used for the physical aggression trajectories where we imposed a zero class trajectory as described above. We found statistically distinguishable intercepts for the high versus medium and medium versus low trajectories (p = 0.000). We found, however, that the high and medium trajectories were indistinguishable at grade 12 (p = 0.170) though they were each distinguishable at grade 12 from the zero class (p < 0.000). There were 25 (13.4%) participants in the high physical aggression trajectory, 65 (34.8%) in the medium trajectory, and 97 (51.9%) in the low trajectory.

In the final stage of the analysis, we added in predictors of the trajectories. Descriptive statistics and correlations are presented in Table 2. The most commonly occurring types of communication were parent and child neutral talk, followed by parent and child positive out-dyad talk. Negative and positive in-dyad talk were rarer. Parent and child out-dyad negative talk was more common than in-dyad negative talk. When the predictors were included in the trajectory model, which included the full 297 participants, the percent of participants in the different social and physical aggression trajectory classes was similar to that observed in the conditional models (N = 187). Specifically, the low social aggression trajectory included 75 (24.1%) participants, with the medium including 167 (55%) and high including 55 (20.8%). There were 154 (46.5%) participants in the low physical aggression trajectory, 102 (37.2%) in the medium, and 41 (16.2%) in the high.

Table 2.

Descriptive statistics

1 2 3 4 5 6 7 8 9 10 M SD Min Max
1. Child Negative In-Dyad Talk 1.00 0.20 0.21 0.00 1.41
2. Child Neg. Out-Dyad Talk 0.15* 1.00 0.66 0.32 0.11 1.65
3. Child Positive In-Dyad Talk −0.03 0.11 1.00 0.13 0.11 0.00 0.73
4. Child Positive Out-Dyad Talk −0.10 0.42** 0.04 1.00 1.48 0.44 0.49 3.23
5. Child Neutral Talk 0.04 −0.02 0.09 0.09 1.00 2.68 0.91 0.89 5.93
6. Parent Negative In-Dyad Talk 0.39** 0.00 −0.13 −0.14 −0.01 1.00 0.38 0.30 0.00 1.71
7. Parent Neg. Out-Dyad Talk 0.06 0.55** 0.02 0.26** 0.06 0.14 1.00 0.62 0.29 0.00 1.82
8. Parent Positive In-Dyad Talk 0.02 −0.05 0.34** −0.11 −0.01 0.05 0.16* 1.00 0.31 0.19 0.00 0.93
9. Parent Pos. Out-Dyad Talk −0.01 0.38** −0.11 0.64** 0.16** 0.04 0.45** 0.02 1.00 1.56 0.56 0.34 4.14
10. Parent Neutral Talk 0.01 0.03 0.05 0.15* 0.82** −0.04 0.09 0.04 0.25** 1.00 3.22 0.90 1.56 5.93
Open in a new tab

Neg. Negative, Pos. Positive

*

p < 0.05

**

p < 0.01

The results from the multinomial logit analyses are shown in Table 3. Note that the coefficients are relative to the high trajectory and have been exponentiated to create odds ratios. More child neutral talk was associated with being in the lowest social aggression trajectory in comparison to the highest social aggression trajectory. Other predictors of social aggression trajectory were not statistically significant. Child out-dyad positive talk was associated with being in the medium and low physical aggression trajectory in comparison to the high physical aggression trajectory. Parent positive in-dyad talk was associated with being in the low physical aggression trajectory in comparison to the high physical aggression trajectory, and the medium trajectory in comparison to the high trajectory.

Table 3.

Multinomial logit analyses results

Predictor Coefficient SE Odds ratio
Medium Social Aggression
 Child Negative In-Dyad 6.41 7.40 607.89
 Child Negative Out-Dyad −4.50 4.08 0.01
 Child Positive In-Dyad −21.03 25.61 0.00
 Child Positive Out-Dyad −6.53 8.91 0.00
 Child Neutral −2.39 5.29 0.09
 Parent Negative In-Dyad −10.16 10.24 0.00
 Parent Negative Out-Dyad 2.27 5.24 9.68
 Parent Positive In-Dyad 2.78 7.80 16.12
 Parent-Positive Out-Dyad −0.97 4.26 0.38
 Parent Neutral 4.53 6.05 92.76
Low Social Aggression
 Child Negative In-Dyad 0.89 2.02 2.44
 Child Negative Out-Dyad 0.14 1.54 1.15
 Child Positive In-Dyad −0.70 10.44 0.50
 Child Positive Out-Dyad 1.27 1.72 3.56
 Child Neutral 1.89* 0.94 6.62
 Parent Negative In-Dyad −1.35 2.80 0.26
 Parent Negative Out-Dyad −0.52 1.56 0.59
 Parent Positive In-Dyad −1.64 3.32 0.19
 Parent-Positive Out-Dyad −1.24 1.42 0.29
 Parent Neutral −0.56 1.19 0.57
Medium Physical Aggression
 Child Negative In-Dyad 5.25 4.73 190.57
 Child Negative Out-Dyad 0.94 1.61 2.56
 Child Positive In-Dyad −1.27 4.94 0.28
 Child Positive Out-Dyad 3.44* 1.57 31.19
 Child Neutral 1.03 0.85 2.80
 Parent Negative In-Dyad −0.79 1.46 0.45
 Parent Negative Out-Dyad −3.26 2.56 0.04
 Parent Positive In-Dyad 5.52* 2.80 249.64
 Parent-Positive Out-Dyad −0.56 0.88 0.57
 Parent Neutral −0.59 1.00 0.55
Low Physical Aggression
 Child Negative In-Dyad 5.05 4.52 156.02
 Child Negative Out-Dyad 0.95 1.48 2.59
 Child Positive In-Dyad 0.25 4.42 1.28
 Child Positive Out-Dyad 3.26* 1.58 26.05
 Child Neutral 0.25 0.78 1.28
 Parent Negative In-Dyad −2.13 1.86 0.12
 Parent Negative Out-Dyad −1.36 2.25 0.26
 Parent Positive In-Dyad 5.12* 2.61 167.34
 Parent-Positive Out-Dyad −1.35 0.79 0.26
 Parent Neutral 0.05 0.82 1.05
Open in a new tab

Medium and low groups compared to high group

*

p < 0.05

Discussion

Features of parent and child communication observed in the laboratory were related to social aggression and physical aggression trajectories that spanned preadolescence and adolescence; however, not all hypothesized associations were observed. Neutral and positive communication were related to being on lower aggression trajectories in comparison to the highest trajectory. However, negativity in the parent-child conversation, directed within dyad or outside of the dyad, was unrelated to differences in aggression trajectories.

As expected, we did find that youth who were observed engaging in more neutral talk and more positive talk with their parent engaged in less aggression across pre-adolescence and adolescence, although the particular effects varied in whether it was parent’s or child’s expressions that predicted less aggression, about whom the content was concerned, and the form of aggression. A child’s amount of casual, neutral conversation with their parent seems to be a protective factor against social aggression with peers. The Family Communication Pattern framework would suggest that youth who develop in a family that promotes open, frequent communication between parents and children would be more sympathetic and better able to understand others’ perspectives (Vieira, 2015), and to be better at managing conflict (Koerner & Fitzpatrick, 2002). Children who engaged in more neutral conversation may also be more socially skilled and resort to pleasant interactions to build peer relationships. Previous research has shown that positive parenting, exemplified by listening skills and discussion with child, may buffer against involvement in aggression (Kawabata et al., 2011; Wallenius et al., 2007). Our results provide further evidence that features of parent-child communication are important predictors of adolescents’ engagement in aggressive behavior. However, just as the previous research did not delineate how listening skills and engagement buffer against involvement in aggression among youth, our study similarly shows a feature of a possible pathway to engaging in low levels of social aggression across time, but not how this process occurs.

One interesting finding in line with our hypotheses was that child participants who engaged in more positive out-dyad talk were more likely to be on a lower physical aggression trajectory. These youth likely recognize positive attributes of others and have prosocial relationships with those outside the dyad, which was captured in the observational coding. Although parent out-dyad positive talk was not associated with less aggression across time, parent out-dyad positive talk and child out-dyad positive talk were correlated in the initial analyses, which suggests that children who say positive things about others outside the dyad tend to have parents who model this behavior. Children’s positive talk about individuals outside the dyad predicted the child following a lower physical aggression trajectory from preadolescence through adolescence, suggesting that early positive perceptions of others may be important for setting a foundation for low-aggression peer relations. Whether the association between child out-dyad positive talk with their parent and lower aggression is consistent at every stage of development is a question open for exploration, which would help distinguish whether preadolescence is a distinctly important time for recognizing and discussing peers positively.

Although we expected child and parent positive in- and out-dyad talk to be protective against being on a higher social aggression trajectory, we did not find support for this hypothesis. Whereas physical aggression is typically observable and deemed antisocial (Dodge et al., 2006; Heilbron & Prinstein, 2008), social aggression tends to be covert. Social aggression may be a more socially acceptable way of showing aggression toward others, especially as youth grow older (Underwood, 2003). As stated in a review of socially aggressive behavior, “from a theoretical perspective, social aggression might function as an outlet for the expression of anger in the context of reduced likelihood of damaging retaliation or punishment” (Heilbron & Prinstein, 2008, p. 180). Indeed, engagement in social aggression may be a sign of social competence (Björkqvist, 1994; Heilbron & Prinstein, 2008; Vaillancourt et al., 2007). Therefore, positive communication with parents may not be protective against engagement in social aggression.

Observed parent positivity directed within the dyad was also associated with the child being on a lower aggression trajectory. When parents are complimentary towards their child, these children may learn to recognize positivity in peers and others outside the dyad. Parent support may buffer against teacher-reports of the child’s aggression, although in previous research perceptions of parent support were not associated with all reporters’ reports of adolescent aggression (Benhorin & McMahon, 2008). Whether due to transference of values and expectations through conversation orientation (Family Communication Theory), social competence, social learning, or due to the child’s adjustment as a result of positive regard from the parent, parent in-dyad positivity in conversation does seem to predict behaving less aggressively. The exact mechanism through which this occurs an important focus for future research.

Observed negativity directed within or outside the dyad in parent-child interactions was not related to following higher aggression trajectories from childhood into adolescence. Laboratory observations of engaging in aversive, conflictual interactions with parents or saying mean things about others outside the dyad do not seem to translate to socially and physically aggressive interactions in the peer context. It is possible that negative communication, although inclusive of constructs previously studied in relation to child aggression such as parental harshness and parent aggression (Dodge et al., 2008; Kawabata et al., 2011), is too broad and varied to have an observed impact on aggression. This may account for the non-significant negative associations between parent in- and out-dyad negativity and child aggression in most of the comparisons between lower and higher aggression trajectories. It is also possible that parents who tend to have more negative relationships with the children refrained from participating in a study that involved observations of parent-child inter-action. Therefore, if only the most negative interactions are associated with child aggression, those may have not been included in our analyses.

As has been observed in previous research, neutral emotions were expressed in parent-child communication most commonly (Van Bommel et al., 2019). In this typically developing sample, there was a low base rate of negative talk in general relative to other forms of talk (see Table 2). It is possible that the low variability in negative talk may have not allowed for the observation of an association between negative talk and aggression trajectory. Although we did not have specific hypotheses about the differences in rate-per-minute of different types of parent and child talk, we found it interesting that the majority of talk was neutral, and little of the talk was directed within-dyad. One thing we do not know from this study is whether this was due to the general nature of parent-child communication at this stage of development among typically developing dyads, or whether our conversation prompts, the first four of which may have led participants to focus on others outside of the dyad, and the fifth the only to directly probe negativity through discussion of a disagreement, contributed to this phenomenon. In previous research with adolescent-parent dyads in which conflictual conversations specifically were investigated, positivity was higher than negativity among parents, but negativity was higher than positivity among adolescents (Van Bommel et al., 2019). Future research could examine family communication in less structured laboratory interactions or in response to a variety of types of instructions. However, we expect that despite prompts, different dyads display different communication features at different rates, which relate to their aggression trajectories.

Limitations and Future Directions

Despite the important implications of these findings, these results should be considered in light of methodological limitations. First, although observational tasks are beneficial in that they provide more objective information than oftused self-reports, the artificiality of the lab setting could potentially lead to reactivity (McKechnie, 2012). Also, prior to parents and children engaging in the observation tasks, children participated in a recorded task with their friend. It is possible that there were spillover effects, with child affect or conversation from the previous interaction impacting the parent-child communication. Second, our sample did not include enough fathers to allow us to test effects of parent gender. A third limitation was the lower inter-rater reliability for the children’s in-dyad negative talk (κ = 0.50), though this was still in the moderate range according to Landis and Koch (1977). Previous research has found that negative parent-child behavior is particularly difficult to micro-code (Snyder, 2016). One aspect of parent-child communication we did not address in this study was the dynamic dyadic nature of parent-child communication, which would be telling for how parent-child communication unfolds in real-time.

There was also missing data that needed to be accounted for in the models. Ideally, we would have had complete data, however this is challenging in a study spanning nine years with multiple raters. In order to minimize losing information through listwise deletion, we used sophisticated methods and best practices for handling missing data. Further, although teacher reports may provide a less biased view of aggression than self- or parent reports, teachers’ perceptions may be limited to the school context.

The parent who was present for the observational task could potentially affect the observed interactions. The researchers requested that the parent most involved in the child’s social life be the one to participate in the observation task. Although this person likely would have more knowledge of the topics the dyads were asked to discuss, this parent, in the case of two parent families, may also be the one with whom the child has more communication and perhaps a closer relationship. The impact of whole family communication patterns on child development of aggression is an important future research direction.

Another important direction for future research is to investigate the mechanisms through which features of parent-child communication are internalized by the child, which then leads to enactment of different levels of aggression among peers throughout their adolescent development. Through social learning (Bandura, 1971), youth may observe, imitate, and model their interactions with their parents, communication patterns that begin in early childhood and develop into patterns of family communication through reinforcement of these patterns. To what features of communication children attend, and their immediate and short-term modeling and imitation of the parent’s negativity and positivity would be a fascinating avenue for investigation that would help researchers understand the mechanisms through which family communication may impact development of aggression in other contexts.

It is important to note that there are many individual, interpersonal, family and ecological factors that potentially impact the development and trajectories of socially and physically aggressive behavior. These different predictors have been discussed at length in empirical articles (e.g. Underwood et al., 2008) and handbook chapters (e.g. Dodge et al., 2006). Although investigating these predictors alongside the novel predictor of parent-child communication was beyond the scope of this paper, these factors have the potential to explain variance in aggressive behavior, and also, potentially, predict parent-child communication. Researchers might build on the current findings by, in addition to studying the mechanisms through which communication may predict aggressive behavior, understanding factors that precede parent-child communication, which ultimately predicts aggressive behavior.

Further, there are different forms of functions of aggression that were not investigated in the current study. It would be of interest to understand how verbal aggression, and the functions of aggression, reactive and proactive aggression, might be related to communication features since, in the case of verbal aggression, this behavior could be directly modeled within the family, and because reactive aggression may similarly be modeled as a means of solving problems or challenging communication with which a family member disagrees.

Despite this study’s limitations and the need for future research to fill the gap between communication features and later aggressive behavior, this study had important strengths. This is the first study to utilize direct observation to examine how characteristics of parent-child communication during preadolescence relate to social and physical aggression across preadolescence and adolescence. The findings highlight the importance of fostering warm, positive parent-child communication to promote long-term psychosocial adjustment. Parents’ positivity in the parent-child dyad may teach children more prosocial ways of interacting. Children’s engagement in casual conversation and positive conversation about others with parents seems to relate to lower levels of children’s aggression with peers. Perhaps when children feel comfortable sharing positivity about others with parents and have parents who are adept at listening, the experience of visiting with a parent teaches children positive skills in conversation that might serve them well in their social interactions with peers.

Overall, the results highlight a strengths-based perspective of parent-child communication. These findings did not suggest much risk associated with negative communication, but instead highlight the importance of children talking with their parents in positive and neutral ways, with opportunity for two-way conversation such as in conversation orientation families. Parents’ positive talk directed toward their child may be linked with the child’s engagement in less aggression across preadolescent and adolescent development. Future research could examine whether high levels of positive and neutral communication exchanged between parent and child may be a global indicator for a range of more specific parent-child interaction processes, (e.g. parental harshness, psychological control, or children’s willingness to self-disclose). If this is in fact the case, it would suggest that parent-child communication patterns would be of particular interest to clinicians and researchers alike. Regular and open communication between parents and children should be encouraged, as it seems as if this behavior may be protective against a greater involvement in aggression across the child’s adolescence. Further, even if negativity is present in the open communication, this study found no evidence that negative talk has an effect on the child’s aggressive behavior after controlling for other factions.

This study adds to the body of research suggesting that family factors may contribute to the development of aggressive behavior with peers. In terms of clinical implications, this suggests that the way parents and children communicate can affect not only the parent-child relationship and family functioning, but perhaps also relationships outside the family. Families to engaging in positive and neutral communication may protect youth from perpetrating the highest levels of aggression toward others.

Supplementary Material

SupMat

Highlights.

  • Child neutral talk predicted being on a lower social aggression trajectory.

  • Child positive out-dyad talk related to lower child physical aggression trajectory.

  • Parent positive in-dyad talk related to lower child physical aggression trajectory.

  • Parent and child in- or out-dyad negative talk did not predict aggression.

Footnotes

Supplementary information The online version contains supplementary material available at https://doi.org/10.1007/s10826-021-01959-7.

Conflicts of Interest The authors have no competing interest.

Compliance with Ethical Standards

Ethical Approval This research was conducted in full accordance with the ethical principles outlined by the American Psychological Association and the SRCD guidelines for research with child participants. These research procedures were approved annually by the Institutional Review Board of University of Texas at Dallas.

Informed Consent This research involved human participants, and therefore informed consent was attained from parents, and assent was provided by child participants.

References

  1. Archer J, & Coyne SM (2005). An integrated review of indirect, relational, and social aggression. Personality and Social Psychology Review, 9(3), 212–230. 10.1207/s15327957pspr0903_2. [DOI] [PubMed] [Google Scholar]
  2. Asparouhov T, & Muthén B (2014). Auxiliary variables in mixture modeling: Three-step approaches using Mplus. Structural Equation Modeling: A Multidisciplinary Journal, 21(3), 329–341. 10.1080/10705511.2014.915181. [DOI] [Google Scholar]
  3. Aspland H, & Gardner F (2003). Observational measures of parent‐ child interaction: An introductory review. Child and Adolescent Mental Health, 8(3), 136–143. [DOI] [PubMed] [Google Scholar]
  4. Bandura A (1971). Social learning theory General Learning Press. [Google Scholar]
  5. Benhorin S, & McMahon SD (2008). Exposure to violence and aggression: Protective roles of social support among urban African American youth. Journal of Community Psychology, 36 (6), 723–743. 10.1002/jcop.20252. [DOI] [Google Scholar]
  6. Björkqvist K (1994). Sex differences in physical, verbal, and indirect aggression: A review of recent research. Sex Roles, 30(3–4), 177–188. [Google Scholar]
  7. Campbell SB, Spieker S, Vandergrift N, Belsky J, & Burchinal M, NICHD Early Child Care Research Network. (2010). Predictors and sequelae of trajectories of physical aggression in school-age boys and girls. Development and Psychopathology, 22 (1), 133–150. 10.1017/S0954579409990319. [DOI] [PMC free article] [PubMed] [Google Scholar]
  8. Card NA, Stucky BD, Sawalani GM, & Little TD (2008). Direct and indirect aggression during childhood and adolescence: A meta‐analytic review of gender differences, intercorrelations, and relations to maladjustment. Child Development, 79(5), 1185–1229. 10.1111/j.1467-8624.2008.01184.x. [DOI] [PubMed] [Google Scholar]
  9. Côté SM, Vaillancourt T, Barker ED, Nagin D, & Tremblay RE (2007). The joint development of physical and indirect aggression: Predictors of continuity and change during childhood. Development and Psychopathology, 19(1), 37–55. 10.1017/S0954579407070034. [DOI] [PubMed] [Google Scholar]
  10. Cleverley K, Szatmari P, Vaillancourt T, Boyle M, & Lipman E (2012). Developmental trajectories of physical and indirect aggression from late childhood to adolescence: Sex differences and outcomes in emerging adulthood. Journal of the American Academy of Child and Adolescent Psychiatry, 51(10), 1037–1051. 10.1016/j.jaac.2012.07.010. [DOI] [PubMed] [Google Scholar]
  11. Crick NR (1996). The role of overt aggression, relational aggression, and prosocial behavior in the prediction of children’s future social adjustment. Child Development, 67(5), 2317–2327. 10.1111/j.1467-8624.1996.tb01859.x. [DOI] [PubMed] [Google Scholar]
  12. Crick NR, Ostrov JM, Burr JE, Cullerton-Sen C, Jansen-Yeh E, & Ralston P (2006). A longitudinal study of relational and physical aggression in preschool. Journal of Applied Developmental Psychology, 27(3), 254–268. 10.1016/j.appdev.2006.02.006. [DOI] [Google Scholar]
  13. Dodge KA, Coie JD, & Lynam D (2006). Aggression and antisocial behavior in youth. In Eisenberg N, Damon W, & Lerner RM (Eds.), Handbook of child psychology: Social, emotional, and personality development (pp. 719–788). John Wiley & Sons. [Google Scholar]
  14. Dodge KA, Greenberg MT, & Malone PS, The Conduct Problems Prevention Research Group. (2008). Testing an idealized dynamic cascade model of the development of serious violence in adolescence. Child Development, 79(6), 1907–1927. 10.1111/j.1467-8624.2008.01233.x. [DOI] [PMC free article] [PubMed] [Google Scholar]
  15. Ehrenreich SE, Beron KJ, Brinkley DY, & Underwood MK (2014). Family predictors of continuity and change in social and physical aggression from ages 9 to 18. Aggressive Behavior, 40 (5), 421–439. 10.1002/ab.21535. [DOI] [PMC free article] [PubMed] [Google Scholar]
  16. Ehrenreich SE, Beron KJ, & Underwood MK (2016). Social and physical aggression trajectories from middle childhood through late adolescence as predictors of maladjustment at age 18: Rule-breaking behaviors, internalizing symptoms, and borderline and narcissistic features. Developmental Psychology, 52 (3), 457–462. 10.1037/dev0000094. [DOI] [PMC free article] [PubMed] [Google Scholar]
  17. Fanti KA, & Henrich CC (2010). Trajectories of pure and co-occurring internalizing and externalizing problems from age 2 to age 12: Findings from the National Institute of Child Health and Human Development Study of Early Child Care. Developmental Psychology, 46(5), 1159–1175. 10.1037/a0020659. [DOI] [PubMed] [Google Scholar]
  18. Feldman BJ, Masyn KE, & Conger RD (2009). New approaches to studying problem behaviors: A comparison of methods for modeling longitudinal, categorical adolescent drinking data. Developmental Psychology, 45(3), 652–676. 10.1037/a0014851. [DOI] [PMC free article] [PubMed] [Google Scholar]
  19. Fite PJ, Rubens SL, Preddy TM, Raine A, & Pardini DA (2014). Reactive/proactive aggression and the development of internalizing problems in males: The moderating effect of parent and peer relationships. Aggressive Behavior, 40(1), 69–78. 10.1002/ab.21498. [DOI] [PubMed] [Google Scholar]
  20. Galen BR, & Underwood MK (1997). A developmental investigation of social aggression among children. Developmental Psychology, 33(4), 589–600. 10.1037/0012-1649.33.4.589. [DOI] [PubMed] [Google Scholar]
  21. Heilbron N, & Prinstein MJ (2008). A review and reconceptualization of social aggression: Adaptive and maladaptive correlates. Clinical Child and Family Psychology Review, 11(4), 176–217. 10.1007/s10567-008-0037-9. [DOI] [PubMed] [Google Scholar]
  22. Henriksen M, Skrove M, Hoftun GB, Sund ER, Lydersen S, Tseng WL, & Sukhodolsky DG (2020). Developmental course and risk factors of physical aggression in late adolescence. Child Psychiatry & Human Development, 1–12. 10.1007/s10578-020-01049-7 [DOI] [PMC free article] [PubMed] [Google Scholar]
  23. Kahlbaugh PE, & Haviland JM (1994). Nonverbal communication between parents and adolescents: A study of approach and avoidance behaviors. Journal of Nonverbal Behavior, 18(1), 91–113. [Google Scholar]
  24. Karriker‐Jaffe KJ, Foshee VA, Ennett ST & & Suchindran C (2008). The development of aggression during adolescence: Sex differences in trajectories of physical and social aggression among youth in rural areas. Journal of Abnormal Child Psychology, 36, 1227–1236. [DOI] [PMC free article] [PubMed] [Google Scholar]
  25. Kawabata Y, Alink LR, Tseng WL, Van Ijzendoorn MH, & Crick NR (2011). Maternal and paternal parenting styles associated with relational aggression in children and adolescents: A conceptual analysis and meta-analytic review. Developmental Review, 31(4), 240–278. 10.1016/j.dr.2011.08.001. [DOI] [Google Scholar]
  26. Koerner AF, & Fitzpatrick MA (2002). Toward a theory of family communication. Communication Theory, 12(1), 70–91. 10.1111/j.1468-2885.2002.tb00260.x. [DOI] [Google Scholar]
  27. Ladd GW (2005). Children’s peer relations and social competence: A century of progress Yale University Press. [Google Scholar]
  28. Ladd GW, & Pettit GS (2002). Parenting and the development of children’s peer relationships. In: Bornstein MH, (Ed.), Handbook of parenting: Practical issues in parenting (Vol. 5, pp. 269–309). Lawrence Erlbaum Associates. [Google Scholar]
  29. Lambert SF, & Cashwell CS (2004). Preteens talking to parents: Perceived communication and school-based aggression. The Family Journal, 12(2), 122–128. [Google Scholar]
  30. Landis JR, & Koch GG (1977). The measurement of observer agreement for categorical data. Biometrics, 33(1), 159–174. http://www.jstor.org/stable/2529310 [PubMed] [Google Scholar]
  31. Lee SST, & Wong DSW (2009). School, parents, and peer factors in relation to Hong Kong students’ bullying. International Journal of Adolescence and Youth, 15(3), 217–233. 10.1080/02673843.2009.9748030. [DOI] [Google Scholar]
  32. Lytton H (1973). Three approaches to the study of parent‐child interaction: ethological, interview and experimental. Journal of Child Psychology and Psychiatry, 14(1), 1–17. [DOI] [PubMed] [Google Scholar]
  33. McEachern AD, & Snyder J (2012). Gender differences in predicting antisocial behaviors: Developmental consequences of physical and relational aggression. Journal of Abnormal child Psychology, 40(4), 501–512. 10.1007/s10802-011-9589-0. [DOI] [PubMed] [Google Scholar]
  34. McKechnie LEF (2012). Reactivity. In Given LM (Ed.), The SAGE encyclopedia of qualitative research methods, (p. 730). SAGE Publications. [Google Scholar]
  35. Muthén BO, Muthén LK, & Asparouhov T (2016). Regression and mediation analysis using Mplus Muthén & Muthén. [Google Scholar]
  36. Muthén LK, & Muthén BO (1998–2015). Mplus user’s guide (7th ed.). Muthén & Muthén. [Google Scholar]
  37. Nagin D (2005). Group-based modeling of development Harvard University Press. [Google Scholar]
  38. Paquette JA, & Underwood MK (1999). Gender differences in young adolescents’ experiences of peer victimization: Social and physical aggression. Merrill-Palmer Quarterly, 45(2), 242–266. [Google Scholar]
  39. Samp JA, Parker KA, & Duvall H (2006). Adolescents’ communicative goals for problematic events: defining content and examining the influence of identity processing orientations. Communication Studies, 57(4), 455–475. 10.1080/10510970600946384. [DOI] [Google Scholar]
  40. Sillars AL (1991). Behavioral observation. In Montgomery BM & Duck S, (Eds). Studying interpersonal interaction, (pp. 197–218). Guilford Press. [Google Scholar]
  41. Snyder JJ (2016). Coercive family processes and the development of child social behavior and self-regulation. In Dishion TJ & Snyder JJ (Eds). The oxford handbook of coercive relationship dynamics, (pp. 101–113). Oxford University Press. [Google Scholar]
  42. Underwood MK (2003). Social aggression among girls Guilford Press. [Google Scholar]
  43. Underwood MK, Beron KJ, Gentsch JK, Galperin MB, & Risser SD (2008). Family correlates of children’s social and physical aggression with peers: Negative interparental conflict strategies and parenting styles. International Journal of Behavioral Development, 32(6), 549–562. 10.1177/0165025408097134. [DOI] [Google Scholar]
  44. Underwood MK, Beron KJ, & Rosen LH (2009). Continuity and change in social and physical aggression from middle childhood through early adolescence. Aggressive Behavior, 35(5), 357–375. 10.1002/ab.20313. [DOI] [PMC free article] [PubMed] [Google Scholar]
  45. United States Census Bureau. (2000). U.S. Census Bureau: State and county Quickfacts. data derived from population estimates, 2000 census of population and housing Retrieved December 16, 2020, from http://quickfacts.census.gov/qfd/states/48/48113.html.
  46. United States Census Bureau. (2016). Intercensal Estimates of the Resident Population by Sex, Race, and Hispanic Origin for Counties: April 1, 2000 to July 1, 2010 Retrieved December 16, 2020, from https://www.census.gov/data/datasets/time-series/demo/popest/intercensal-2000-2010-counties.html.
  47. Vaillancourt T (2005). Indirect aggression among humans: Social construct of evolutionary adaptation? In Tremblay RE, Hartup WW, & Archer J (Eds.), Developmental origins of aggression (pp. 158–177). Guilford Press. [Google Scholar]
  48. Vaillancourt T, Miller JL, Fagbemi J, Coté S, & Tremblay RE (2007). Trajectories and predictors of indirect aggression: Results from a nationally representative longitudinal study of Canadian children aged 2–10. Aggressive Behavior, 33, 314–326. 10.1002/ab.20202. [DOI] [PubMed] [Google Scholar]
  49. Van Bommel DM, Van der Giessen D, Van der Graaff J, Meeus WH, & Branje SJ (2019). Mother–adolescent conflict inter-action sequences: The role of maternal internalizing problems. Journal of Research on Adolescence, 29(4), 1001–1018. 10.1111/jora.12441. [DOI] [PubMed] [Google Scholar]
  50. Vermunt JK (2010). Latent class modeling with covariates: Two improved three-step approaches. Political Analysis, 18(4), 450–469. 10.1093/pan/mpq025. [DOI] [Google Scholar]
  51. Vieira ET Jr (2015). Family communication patterns, sympathy, perspective-taking, and girls’ thoughts about interpersonal violence. Journal of Family Violence, 30(5), 607–619. 10.1007/s10896-015-9705-y. [DOI] [Google Scholar]
  52. Vollet JW, George MJ, Kurup AR, & Underwood MK (2019). Leading by example: Parental promotion of prosocial behavior predicts decreases in social aggression Poster session presented at the Meetings of the Society for Research on Child Development (SRCD), Baltimore, MD. [Google Scholar]
  53. Wallenius M, Punamäki RL, & Rimpelä A (2007). Digital game playing and direct and indirect aggression in early adolescence: The roles of age, social intelligence, and parent-child communication. Journal of Youth and Adolescence, 36(3), 325–336. 10.1007/s10964-006-9151-5. [DOI] [PubMed] [Google Scholar]
  54. Wickrama KAS, Lee TK, O’Neal CW, & Lorenz FO (2016). Higher-order growth curves and mixture modeling Routledge. [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

SupMat
NIHMS1932819-supplement-SupMat.docx (35.5KB, docx)

RESOURCES