kenneth a. dodge duke university draft chapter for m...
TRANSCRIPT
Social Information Processing Patterns as Mediators of the Interaction between
Genetic Factors and Life Experiences in the Development of Aggressive Behavior
Kenneth A. Dodge
Duke University
Draft chapter for M. Mikulincer & P.R. Shaver (Eds.), Understanding and reducing aggression,
violence, and their consequences. Washington, DC: American Psychological Association.
Social Information Processing 2
Abstract
Social information processing patterns (e.g., hypervigilance to threat cues, hostile
attributional biases, self-defensive goals, accessibility to aggressive responses in memory,
aggressive decision-making) correlate robustly with individual differences in aggressive
behavior and predict growth in aggression across development. It is posited here that these
patterns also mediate the impact of genetic and environmental factors on aggressive behavior. A
general model is described that proposes that (a) processing patterns provide the proximal
mechanism through which aggressive behavior occurs; (b) these patterns correlate with neural
and psychophysiologic processes; (c) these patterns are acquired through genetic and
environmental processes, especially in interaction; and (d) acquired processing patterns account
for the effects of genetic and environmental factors in behavioral development and provide the
mechanism through which these factors exert their impact. Findings are presented from the
Child Development Project, an ongoing longitudinal study of 585 boys and girls followed from
age 4 through young adulthood.
Social Information Processing 3
It is well established that social-cognitive processes correlate with aggressive behavioral
acts. For example, in response to an ambiguous provocation by another person, when a
respondent infers that the act was committed with hostile intent, the probability that the
respondent will react aggressively is high (about .76, Dodge, 1980); whereas when that same
respondent infers that the act was committed benignly, the probability of an aggressive
behavioral reaction is low (about .25, Dodge, 1980). Likewise, if the respondent evaluates that an
aggressive response will lead to desired outcomes, the probability of engaging in aggression is
high (Fontaine et al., 2008). Although the evidence is less clear that these social-cognitive
processes cause the aggressive behavioral response during the micro-seconds of interpersonal
interaction, the correlation has been found over and over.
It has also been found that individuals develop characteristic styles of processing social
information within specific social situations. These styles act as acquired personality
characteristics. They correlate significantly with and predict individual differences in aggressive
behavior in particular situations. In this chapter, this body of empirical evidence will be reviewed
briefly. This evidence will be integrated with recent discoveries in psychophysiology and neural
processes. Hypotheses will be advanced that social information processing (SIP) patterns arise
from interaction between genetic factors and life experiences and mediate the impact of these
factors on aggressive behavior. A general model will be proposed to guide future research.
Social Information Processing Mechanisms in Aggressive Behaviors
Models of the processing of information in response to social stimuli posit a sequence of
steps that lead to behavioral responding including aggression toward others (Crick & Dodge,
1994; Dodge, 2002; Huesmann, 1988). These steps are logically ordered and assumed to flow
temporally, although evidence for the sequential ordering is scant. Methods to assess an
Social Information Processing 4
individual’s self report of her or his processing have been developed using hypothetical
situational vignettes as stimuli, in which the person is asked to contemplate a hypothetical
scenario in which a social event occurs and reply to questions about attribution and possible
responses (Dodge, 1980; Dodge, Pettit, McClaskey, & Brown, 1986). The stimuli have been
presented via video, cartoons, or orally.
The first several steps of processing describe the sensation and interpretation of cues, and
the latter steps describe behavioral response decision. The first step is attention and sensation of
the stimulus. Because the stimulus array is so large, selective attention to some cues over others
is inevitable. In a situation involving provocation by another, for example being pushed to the
ground, one child might attend to the provocation itself and the pain it causes for the self,
whereas a second child might attend to the teacher watching in the background and a third child
might attend to the peers who are laughing nearby. Attention to cues can obviously influence
downstream processing and ultimate behavior, such as when attention to the provocateur's look
of surprise and regret might mitigate a retaliatory response. Numerous factors affect selective
attention to hostile versus other cues, such as recent threat, fatigue, and stress (Dodge, 2002). A
pattern of habitual selective attention to hostile cues has been associated with chronic aggressive
behavior (Dodge, Pettit, McClasky, & Brown, 1986).
Closely following the sensation of cues is a mental representation of those cues, often
involving an interpretation of the other person's intention. As noted above, when a hostile intent
is inferred, aggressive behavior likely ensues, in contrast with an inference of a benign intention.
The process of mental representation occurs online in microseconds and may be updated across
time during a social interaction. It is not usually a conscious process, although it can become so
if prompted. The process undoubtedly involves neural activity that is conditioned by experience.
Social Information Processing 5
Although precise locations are unclear, evidence suggests that mental representations of social
intent are associated with activity in the …. Inferences of threat and hostile intent have also been
found empirically to be correlated with heightened autonomic reactivity (Crozier et al., 2008).
Numerous studies have shown a robust pattern in which hostile attributional biases are
associated with aggressive behavior, especially reactive aggression (Dodge, 1980). A review by
Orobio de Castro et al. (2002) indicates that this pattern holds across age, demographic and
cultural groups, and contexts.
The third step is goal selection, in which the mentally represented stimulus is associated
with an emotional reaction and the narrowing of a goal. Again, the respondent is not usually
aware of this process but might reflect afterward on one's own cognitive process. Children who
experience anger and regularly select instrumental and self-defensive goals are likely to behave
aggressively, whereas children who select social goals are likely to behave nonaggressively
(Crick & Dodge, 1995).
Steps of response generation, response evaluation, and enactment constitute the response
decision phase of processing. Mental representation and goal selection trigger one or more
possible behavioral responses such as aggression, withdrawal, and social deflection. The trigger
from mental representation of hostile intent to aggressive response generation is a neural
association that is probably both "ready" at birth from evolution and conditioned from experience
and observation. One of the most well-replicated findings in this area is the empirical association
between the generation of aggressive responses to hypothetical social stimuli and actual chronic
aggressive behavior, beginning at about age 4 and continuing at least through adolescence
(reviewed by Dodge, Coie, & Lynam, 2006).
Generation of an aggressive response does not lead inevitably to aggressive behavior.
Social Information Processing 6
Processes of response evaluation and decision, called RED by Fontaine and Dodge (2006)
follow. During RED, the respondent immediately decides (nonconsciously, in microseconds) to
accept the generated response without any consideration of consequences or to consider multiple
domains of evaluative judgment, including: 1) response efficacy, the estimation of how likely the
respondent is to be successful if the considered response were to be carried out; 2) response
valuation, which is the assignment of value to the response in dimensions of its social and moral
qualities; 3) outcome expectancy, which is the estimation of the likelihood of various
consequences of a behavior; and 4) outcome valuation, in which the estimated outcome is given
value. Fontaine and Dodge (2006) hypothesized that different possible responses are compared
(response comparison) before the most appropriate response is selected (response selection).
Measurement of response decision during hypothetical social stimuli has yielded robust
correlations between all of these sub-processes and chronic aggressive behavior, especially
proactive aggression (Fontaine, Yang, Dodge, Bates, & Pettit, 2008; see review by Fontaine &
Dodge, 2006).
Social Information Processing Patterns as Predictors of Individual Differences in
Aggressive Behavior
Although these processing patterns have been empirically correlated with individual
differences in aggressive behavior, these findings often suffer two pitfalls. First, because the
measurement of processing is typically based on the individual's self-report, this report is one's
self-observation of cognitive processing and not processing itself, which occurs at the neural
level and nonconsciously. Even self-reports that are collected "online" are immediate self-
observations. Recent evidence suggests that individuals actually begin to respond with neural
activity in micro-seconds prior to self-awareness of responding. The second problem is that the
Social Information Processing 7
empirical correlation between patterns of processing and patterns of behavior might reflect an
opposite causal direction than proposed in the model or be attributed to an unmeasured third
variable. Two kinds of evidence have been mounted to test the hypothesis that chronic patterns
of processing increment the prediction of aggressive behavior.
Controlling for Prior Aggression
The first evidence comes from prospective studies in which early levels of aggressive
conduct problems are controlled statistically. In the Child Development Project of a community
sample of 585 boys and girls followed for 20 years, we assessed both aggressive behavior and
processing patterns recurrently. During the early elementary school years, children develop
patterns of processing social information that become stable when measured annually for four
years, as assessed by cross-time internal consistency coefficients of .70 to .79 (Dodge, Pettit,
Bates, & Valente, 1995). These patterns function as acquired personality characteristics. During
this period, the continuity in aggressive behavior also becomes strong. However, aggressive
behavior measured by teacher assessments at age 5 was found to predict patterns of social
information processing in kindergarten through grade 3, which, in turn, predicted aggressive
behavior at age 10, and incremented the prediction of aggressive behavior even when early levels
of aggression were statistically controlled (Dodge et al., 2003).
We also found prediction of adolescent conduct problems from kindergarten processing
patterns. Here, we scored children as displaying problems in social information processing at the
early steps (hypervigilance and hostile attributional biases) or later steps (response generation or
evaluation) or both or neither steps. We found that controlling for kindergarten externalizing
problems as assessed by both mothers and teachers, the four groups differed in mother- and
teacher-rated externalizing behavior problems at the end of grade 11 (Lansford et al., 2006).
Social Information Processing 8
Furthermore, the effect was synergistic, in that the group of children with kindergarten problems
at both stages of processing were especially likely to show conduct problems in high school.
The prediction was even stronger from processing patterns in grade 8. Controlling for
conduct problems during grade 8 and earlier, the four groups of children as assessed by
processing patterns during grade 8 differed in mother- and teacher-rated conduct problems in
grade 11. Again, a synergistic interaction effect was found, with the group displaying the highest
levels of conduct problem was the one with problems at both early and later stages of processing.
Reciprocal influences. The relation between processing patterns and aggressive behavior
is hypothesized to be reciprocal. We have also found evidence for an iterative, reciprocal relation
between processing patterns and aggression (Fontaine et al., 2008). That is, aggressive behavior
at age 14 predicted processing patterns the next year, which, in turn, predicted growth in
aggression the following year, even controlling for prior aggression. Likewise, processing
patterns in one year predicted aggression in the following year, which altered processing patterns
in the subsequent year. Across adolescence, this reciprocal effect iterates.
Intervention as Experimental Evidence
The second kind of evidence comes from intervention experiments in which explicit
attempts to alter processing patterns are manipulated by random assignment to observe the
impact on aggressive behavior. Graham and Hudley (1993) provide evidence. They developed a
brief intervention to help young African American boys process information in more benevolent
ways that would lessen a hostile attributional bias. Indeed, they found that random assignment to
this intervention did lead to lower scores on assessments of hostile attributional bias, and this
impact was found to mediate a change in aggressive behavior patterns. Guerra and Slaby (1990)
also randomly assigned aggressive adolescents to intervention or not. Their intervention involved
Social Information Processing 9
multiple components of processing. They also found that random assignment to the intervention
condition was associated with improvements in processing patterns and aggressive behavior
outcomes.
Situation and Relationship Specificity
A pivotal feature of models of social information processing is the importance of
situation specificity in the link between processing patterns and behavior. That is, it is asserted
that processing patterns within a type of situation, such as a provocation or initiation of entry into
a peer group or a romantic relationship conflict or a conflict with a co-worker, will predict
behavior within that type of situation more strongly than behavior in other situations.
Evidence in support of this hypothesis was reported for young children by Dodge, Pettit,
Bates, McClaskey, & Brown (1986). They assessed processing patterns in peer provocation
situations and peer group entry situations, and then they placed children in a laboratory bsetting
and exposed them to a provocation by a confederate peer and an entry situation in which they
were asked to initiate entry into a strange peer group. They found that processing patterns
predicted behavior, and the predictions were stronger within situations than across situations.
More recently, we found evidence with CDP participants in young adulthood. Two kinds
of situations are important in young adulthood, defined by relationships. The establishment of
successful romantic relationships is a key developmental task of early adulthood (Collins & van
Dulmen, 2006; Conger, Cui, Bryant, & Elder , 2000; Furman, 2002). Violence is unfortunately
common in these relationships, as indicated by surveys which indicate that 20 to 50 percent of
intimate relationships during adulthood involve violence toward a partner (Magdol et al., 1997;
Silverman, Raj, Mucci, & Hathaway, 2001). Likewise, relationships with adult peers are
essential to work and community success (Arnett, 2006). Although some individuals display
Social Information Processing 10
violence across different relationships, many adults behave violently only in romantic
relationships (Archer, 2000; Holtzworth-Munroe & Stuart, 1994; Straus, 1999). Some research
has shown that peer violence and partner violence may have distinguishable antecedents
(Brendgen, Vitaro, Tremblay, & Lavoie, 2001).
We were able to interview 85 percent of the original sample at age 22 and follow them
through age 24. We assessed processing patterns in situations involving conflict with a romantic
partner (e.g., “You are at a gathering with a group of friends and your girl/boyfriend and learn
that your girl/boyfriend and one of the people at the gathering used to be a couple; they spend
most of the night talking with each other.”) and conflict with a coworker or peer (e.g., “You tell a
friend something personal and ask your friend not to discuss it with anyone else. However, a
couple of weeks later, you find out that a lot of people know about it.”).
We (Pettit et al., 2009) found evidence for the prediction of violent behavior in each of
these types of relationships, as well as evidence of relationship specificity. Processing patterns in
a romantic relationship predicted violent behavior in actual romantic relationships two years
later, as reported by both the self and the romantic partner. Also, processing patterns within the
peer relationship predicted violent behavior toward peers two years later. Furthermore, the
predictions were stronger within relationships than across relationships.
Neural and Psychophysiological Processes and Social Information Processing
A misconception about information processing is that it occurs independently from
biological processes in real time. In fact, growing evidence identifies the corresponding
psychophysiological and neural processes that co-occur with information processing. Most
likely, these processes occur in real time outside the awareness of the individual, and measures
of information processing represent the individual's post-behavioral reflection of her or his
Social Information Processing 11
actions. This fact would render the measures of information processing as less precise and more
subject to self-presentation and other biases.
Psychophysiologic Processes
Ortiz and Raine (2004) have compiled evidence to indicate that resting heart rate is
inversely related to individual differences in aggressive behavior, especially proactive and life-
persistent aggression but not situational or adolescence-limited aggression (Moffit & Caspi,
2001). Raine, Venables, and Merdnick (1997) found that resting heart rate at age 3 predicted
aggressive behavior 8 years later at age 11. Raine (2002) hypothesized that prefrontal cortex
deficits in volume and function may be responsible for low autonomic activity as well as
aggressive behavior.
However, during the processing of social cues, the autonomic nervous system reacts with
rapid changes in heart rate (Crozier, et al., 2008). While the individual is attending to cues, heart
rate decreases. When a provocation occurs, heart rate increases and then slowly returns to base
line when the threat dissipates. Lorber (2004) summarized evidence from numerous studies to
indicate that aggressive children display higher heart rate reactivity to provocation cues than do
nonaggressive children. Furthermore, Crozier et al. (2008) found that high heart rate reactivity to
provocation cues is correlated with several stages of social information processing. Furthermore,
processing responses mediate the impact of heart rate changes on aggressive behavior.
Thus, two separate psychophysiologic processes may relate to aggressive behavior. Both
low resting heart rate and high heart rate reactivity in response to threatening cues appear to
predict aggressive behavior. These processes may have their antecedents in both life events and
heritable biological processes, and their impact on aggressive behavior may be mediated by the
manner in which the individual processes social information in response to threat.
Social Information Processing 12
Neural processes
Neuroimaging studies suggest that regions of the prefrontal cortex (e.g., the dorsolateral
prefrontal cortex, DLPFC, and ventral prefrontal cortex, VPFC) and the limbic system (e.g., the
amygdala) are activated during information processing in response to interpersonal provocations,
such as unfair allocation of resources by a peer (Meyer-Lindenberg et al., 2006; Sanfey, Rilling,
Aronson, Nystrom, & Cohern, 2003; Tabibnia, Satpute, & Lieberman, 2008). Meyer-Lindenberg
et al. (2006) found significant activation of the amygdala in response to experimental
presentation of stimuli similar to those used to assess social information processing, such as
presentation of angry and fearful faces and aversively valenced cues. Presentation of threatening
faces (Pezawaz et al., 2005), the perception of anger in others, and the experience of anger in
oneself (Murphy, Nimmo-Smith, & Lawrence, 2003) all reliably activate the amygdala. The
prefrontal cortex is activated during executive function tasks involved in processing information,
such as planning, inhibitory control, and decision making (Raine, Buchsbaum, & LaCasse,
1997). Raine (2008) stated that "the prefrontal cortex acts as an 'emergency brake' on runaway
emotions generated by limbic structures" (p. 324).
The activation of these brain regions, in turn, appears to be mediated by the release of
neurotransmitters. Monoamina oxidase-A (MAO-A) appears to correlate with amygdala volume
and probably its activation (Meyer-Lindenberg et al., 2006). Release of 5-HT in the prefrontal
cortex is thought to regulate an individual's impulsive desire to retaliate aggressively to
perceived hostile treatment (Evers et al., 2006), although this evidence was only correlational
until recently. Crockett, Clark, Tabibnia, Lieberman, and Robbins (2008) experimentally
manipulated serotonin levels in each of 20 human subjects through an acute tryptophan depletion
procedure that temporarily lowered 5-HT levels. In contrast with a placebo, when subjects had
Social Information Processing 13
lower serotonin levels they responded to an unfair treatment by a peer with greater retaliatory
behavior.
Thus, in the same way that social information processing operations appear to involve
multiple relatively independent steps, functional brain imaging studies have identified multiple
brain regions that are implicated in different aspects of responding to social stimuli.
Antecedents of Social Information Processing Patterns and Neural Processes
Next, we turn attention to the antecedents of processing patterns and neural processes that
mediate aggressive behavior. Given the empirical findings of both general prediction and
situation-specific prediction, we hypothesize that some antecedent factors will apply generally
and some will apply to specific domains, situations, or relationships.
Child Maltreatment
The experience of abuse by one’s parent during the first five years of life is a devastating
life experience that has been hypothesized to alter one's central working models of how human
relationships operate. Children develop basic trust through interaction with caring adults, and a
violation of that trust through extreme or ongoing maltreatment is hypothesized to lead to
schemas, scripts, knowledge structures, and working models that others will act maliciously.
Thus, the child develops hypervigilant and selective attentional patterns, becomes quick to
attribute hostile motives to others, adopts self-defensive rather than social goals, develops a
repertoire of self-defensive behavioral responses, and comes to evaluate self-protective behaviors
as effective and desirable. Given the centrality of these working models, it is hypothesize that
this experience will lead to long-term and pervasive patterns of processing information across a
range of social situations and relationships.
Social Information Processing 14
The evidence in favor of this hypothesis comes from several laboratories, including
Pollack's work on attention patterns in maltreated children and findings from the CDP study.
Dodge, Bates, and Pettit (1990) and Weiss et al. (1992) reported that maltreatment in the first
five years of life predicted processing patterns at school entry that included hypervigilance,
hostile attributional biases, aggressive response generation, and favorable evaluation of
aggressive behaviors. In turn, these patterns predicted aggressive behavior and mediated the
impact of maltreatment on aggression. Dodge et al. (2003) extended the aggressive outcomes
through adolescence and found a similar pattern of prediction and mediation. Dodge et al. (2009,
unpublished) found that 30.4% of the maltreated children had an official court record of arrest by
age 24, in contrast with 16.5% of the non-maltreated children.
Lansford et al. (2007) found that the outcomes that accrue from early-life maltreatment
are very broad and include not only aggressive behaviors but also early pregnancy, anxiety,
depressive symptoms, school drop out, substance-use problems, and arrests for a variety of
crimes. The processing patterns that were assessed in early elementary school mediated some but
not all of these outcomes, perhaps because the stimuli that were used to assess processing
patterns were restricted to only selected domains. It is plausible that maltreatment alters
processing in multiple domains, which then mediate behavior outcomes. More comprehensive
assessment of processing patterns would be necessary to test this hypothesis.
Peer Social Rejection
Other life experiences appear to have effects that are more circumscribed. During
elementary school in American culture, chronic peer social rejection is a painful experience. We
hypothesized, and found empirically, that it would exacerbate children's aggressive behavior
problems, beyond whatever aggressive behavior led to the peer response of rejection (Dodge et
Social Information Processing 15
al., 2003). That is, rejection during kindergarten through grade 2 predicted aggressive behavior
in grade 4 even controlling for early aggressive behavior. Furthermore, patterns of social
information processing about peer events partially mediated the growth in aggressive behavior
during this period.
More recently, we have found that the impact of early peer rejection lasts through young
adulthood (Pettit et al., 2009). Recall the CDP findings presented earlier about the significant
relation between processing patterns in peer relationships at age 22 and aggressive behavior
toward peers at age 24. It turns out that the experience of peer rejection in kindergarten
significantly predicts these processing patterns at age 22, which, in turn, mediate the impact of
kindergarten peer rejection on aggressive behavior toward peers in young adulthood.
In order to describe these relations in person-centered terms, we dichotomized
participants into those who had not been rejected by peers in elementary school versus those who
had been peer-rejected; those who displayed problematic processing about peer relationships at
age 22 versus those who displayed non-problematic processing; and those who displayed no
violence toward adult peers versus those who displayed violence. Among the group that was not
rejected by peers and who display non-problematic processing about peers the probability of
becoming violent toward a peer is relatively low (.43), whereas among rejected children who
display problematic processing, the probability is high (.70). In between are non-rejected
children who display problematic processing (.47) and rejected children who do not display
problematic processing (.51).
Exposure to Romantic Relationship Violence
The antecedents of romantic relationship violence may differ from those for peer-directed
violence. Exposure to parents' domestic violence in early life predicts later violence in romantic
Social Information Processing 16
relationships (Nay et al., 2009), but disentangling the effect of this particular exposure from
other family violence may prove difficult. We explored the impact of victimization during
adolescent romantic relationships on young adult violent behavior toward a romantic partner.
We found that being victimized violently during an adolescent romantic relationship did
indeed predict violent behavior toward a romantic partner in young adulthood. Also, this
adolescent experience predicted problematic processing patterns in romantic relationships, which
partially mediated the impact of victimization on later violent behavior toward a romantic
partner. Among non-victims who display non-problematic processing, the probability of later
becoming violent toward a romantic partner is low (.288), whereas among victims who display
problematic processing, the probability is high (.62). In between are non-victims who display
problematic processing (.42) and non-victims who do not display problematic processing (.43).
Because of the unique paths from problematic processing within a relationship type to
violence in that relationship, it follows that mediation by processing patterns was fund only
within relationship types and not across relationship types.
Genetic Factors
Individual differences in genetic variation in the serotonin transporter 5-HT have been
associated with increased amygdala activation (Pezawaz et al., 2005) and prefrontal cortex
functions (Raine, 2008), as assessed through functional magnetic resonance imaging. Also, a
common genetic variation in the x-linked MAO-A has been correlated with amygdala volume
and activation in response to threatening stimuli (Meyer-Lindenberg et al., 2006). That is,
individuals, especially males, with a polymorphism in the gene MAO-A demonstrate greater
amygdala activation following presentation of threatening faces and aversive emotional
Social Information Processing 17
situations. It follows, then, that variation in these genes would likely correlate with individual
differences in processing patterns in response to threatening social situations.
Environmental and Genetic Factors in Aggressive Behavior
A comprehensive model of the development of chronic aggressive behavior must account
for both genetic and life experience factors. The major environmental factors in early life have
been reviewed elsewhere to include the kinds of personally threatening experiences addressed
above, such as early physical maltreatment, peer social rejection, and victimization, as well as
exposure to stressful contexts such as poverty and disadvantage (Dodge, Coie, & Lynam, 2006;
Dodge & Pettit, 2003). The latter factors are likely to operate through their impact on life
experiences such as parenting quality and success in life tasks of education (Dodge, Pettit, &
Bates, 1993). Major environmental variables in mid-childhood and adolescence involve exposure
to deviant peers who influence high-risk youth to engage in aggressive behavior and parental
failure to supervise and monitor youths' activities, thus enhancing their exposure (Dodge, Coie,
& Lynam, 2006).
Environmental variables that have an enduring effect are likely to be those that alter brain
processes that include neural, cognitive, and psychophysiological operations. Support has been
growing that life stressors and early trauma have enduring effects on both prefrontal cortex and
amygdala processes measured through functional MRI. Liston, McEwan, and Casey (2009)
reported that the natural experiment of a potent stressor (an upcoming academic examination)
had observable effects on decreasing dorsolateral prefrontal cortex activity during laboratory
tasks. Ganzel, Casey, Glover, Voss, and Temple (2007) found that individuals who had been
exposed to major trauma during the September 11, 2001, attack on the World Trade Center
Social Information Processing 18
displayed heightened amygdala activation when presented with emotionally aversive and fearful
faces.
The heritability of criminal and aggressive behavior patterns has been known for some
while, although adoption studies have identified a heritability-by-environment interaction effect.
Cloninger et al. (1982) found that under conditions of low heritable risk (i.e., having a biological
parent who was not a criminal), the impact of the environment (i.e., having an adoptive parent
who was or was not a criminal), on later criminality was rather small (rate of .03 for non-
criminal adoptive parents vs. .07 for criminal adoptive parents). However, under conditions of
high heritable risk, the impact of the environment was strong (.12 for non-criminal adoptive
parents vs. .40 for criminal adoptive parents). Jaffee et al. (2004) studied monozygotic and
dizygotic twins from the British E-Risk study and identified four rank-ordered groups of
increasing heritable risk, with the lowest heritable-risk group being children whose monozygotic
twin is not conduct disordered (CD), the next lowest group being children whose dizygotic twin
is not CD, the next highest group being children whose dizygotic twin is CD, and the highest
group being children whose monozygotic twin is CD. The experience of child physical
maltreatment was determined by clinical interview with the mother, following procedures by
Dodge, Bates, and Pettit (1990). Among the group at lowest heritable risk, the experience of
physical maltreatment had little effect on conduct disorder outcomes (for non-maltreated, rate =
.02; for maltreated, rate = .04). Among the next highest level of heritable risk, the effect of
maltreatment was small (.06 vs. .13). Among the next highest level, the effect of maltreatment
grew larger (.19 vs. .37). Finally, among the highest level, the effect of maltreatment was largest
(.46 vs. .70).
Social Information Processing 19
The search for specific genetic variables that predict aggressive behavior has been
plagued both by political issues and empirical discoveries that have failed to replicate, perhaps
because of atheoretical approaches that capitalize on chance (Kim-Cohen et al., 2006). Several
theorists argue that genetic factors are likely to operate indirectly, on their effects on neurally and
biologically mediated dispositions to act impulsively without planning, without empathy, or
without consideration of extenuating circumstances (Caspi et al., 2002). These dispositions have
been called executive functions (Morgan & Lillienfeld, 2000, and they likely affect processing
responses (or are measured as processing responses) during social interactions. For example, a
deficit in considering another's affect would lead to inaccurate interpretations of stimulus events,
and impulsivity would lead to premature decision-making that fails to consider long-term
consequences.
Following this hypothesis, the genetic polymorphisms that have been most frequently
correlated with aggressive behavior patterns are those that relate to neurotransmitter functions,
especially dopamine (Dick et al., 2006; Moffitt et al., 2008)]. Three genes will be discussed here.
First, monoamine oxidase A (MAOA) is an enzyme that selectively metabolizes serotonin,
norepinephrine, and dopamine (Shih, Chen, & Ridd, 1999), which are involved in brain actions
associated with stress regulation (Charney, 2004) and biological sensitivity to adverse social
contexts (Boyce & Ellis, 2005). A polymorphism in the X-chromosome-linked gene that encodes
MAOA has been identified in about one third of all human males, suggesting that it might have
adaptive value across evolution but is still non-normative. In rodents, this polymorphism has
been correlated with both aggressive behavior and lower brain serotonin and norepinephrine
(Cases et al., 1995). In humans, a modest main effect of a polymorphism in MAOA on
Social Information Processing 20
aggressive behavior has been found (Brunner et al., 1993) but other studies have found
contradictory patterns (Kim-Cohen et al., 2006).
The more likely robust pattern is a gene-by-environment interaction effect. Eisenberger,
Way, Taylor, Welch, and Lieberman (2007) found that individuals with the low-expression allele
form of MAOA demonstrated heightened dorsal anterior cingulated cortex activity, but only in
response to an environmental stressor of peer social rejection and not to peer inclusion. They
suggested that MAOA may dispose individuals to become “hypersensitive” to interpersonal
rejection.
If the MAOA enzyme is involved in regulating stress, especially trauma and threat, then
it might play a role in moderating the impact of early physical maltreatment on the development
of aggressive behavior. Caspi et al. (2002) found this interaction effect in the Dunedin
longitudinal sample, and Kim-Cohen et al. (2006) replicated the pattern in the British E-Risk
study. In the CDP, we have very recently replicated this interaction effect with criminal arrest as
the outcome (Edwards, Dick, Dodge, Pettit, Bates, & Lansford, 2009). Among children without
the MAOA polymorphism, those who had been maltreated did not differ from those who had not
been maltreated in the proportion who had been arrested by age 22 (.28 vs. .25), whereas among
children with the polymorphism, those who had been maltreated had a much higher probability
of criminal arrest by age 22 than those who not been maltreated (.71 vs. .26).
A second gene that has been implicated in a variety of externalizing and addiction
disorders is the gamma-aminobutyric acid (GABA) A receptor, alpha 2, also known as GABRA2
(Dick et al., 2006). GABA is the major inhibitory neurotransmitter in the mammalian brain. A
polymorphism in GABRA2 has been associated with conduct disorder, but the more powerful
pattern again is a gene-by-environment interaction effect, in which high parental supervision
Social Information Processing 21
during early adolescence has been found to buffer children from the adverse effect of the
GABRA2 polymorphism (Dick et al., 2008). Among children whose parents who engaged in low
rates of supervision, the likelihood of being in a persistently-high trajectory of conduct problems
across adolescence increases dramatically with the number of copies of the risk allele of GABRA2,
from .07 to .19 to .28 for 0, 1, or 2 copies, respectively, whereas among children whose parents
engaged in high rates of supervision, the likelihood of being in the persistently high trajectory of
conduct problems increases modestly from .10 to .12 to .126 for 0, 1, or 2 copies, respectively.
A third gene that has been found to correlate with aggressive behavior is 5-HT. Recall
that polymorphisms in 5-HT have been associated with impaired limbic structure and increased
amygdala function. Waldman (2008) recently took this work a step further by finding a
significant relation between a polymorphism in 5-HT and individual differences in reactive but
not proactive aggressive behavior. Even more striking was that, using the social information
processing instruments described above, he found a significant association between 5-HT and
hostile attributional biases. Finally, the effect of 5-HT on reactive aggression was mediated by
hostile attributional biases. This work represents the first known attempt to identify molecular
genetic bases for social information processing patterns.
Thus, a comprehensive model of the development of chronic aggressive behavior must
take into account environmental main effects, gene main effects, and gene-by-environment
interaction effects. Furthermore, the environmental variables are likely to include both broad
factors that influence aggression across many situations as well as situational factors that
influence aggression within situations.
A Proposed Model of Social Information Processing Mechanisms in Gene-by-Environment
Interaction Effects
Social Information Processing 22
Integration of the distal genetic and environmental factors with social information
processing proximal mechanisms requires a final leap of theorizing. The model proposed here
builds on models by van Goozen, Fairchild, and Harold (2008) and Raine (2008) but is unique in
positing that within-situation processing patterns mediate the effects of genes, environments, and
their interactions. The overall model is depicted in Figure 1 which posits that specific (albeit
unidentified, as of yet) genes and early environment of threat, trauma, and adversity pose risks
for the long-term development of chronically violent behavior. These distal factors operate as
main effects and in interaction with each other. Throughout development, situational
environments also lead to the development of situation-specific processing patterns that mediate
aggressive behavior within those situations. When these situations present themselves, they pose
proximal risk for violent behavior within those situations.
These distal and proximal risk factors are mediated by brain processes that operate in
response to proximal situational stimuli. Three aspects of functioning are hypothesized to co-
occur during responding: neural activity in synapses (most likely involving neurotransmitters
such as dopamine), social information processing as described earlier, and psychophysiological
activity through the autonomic nervous system. These brain processes, in turn, mediate
aggressive behavioral responses within these situations.
A Research Agenda
This model poses numerous research studies that have yet to be completed. In order to
understand the intricate interactions between specific genes and specific environments, existing
large-sample longitudinal studies need to be mined to test gene-environment hypotheses. Sub-
groups of individuals who fit profiles of gene-environmental risk need to be exposed to
Social Information Processing 23
situational stimuli in order to elicit processing responses while being observed through functional
magnetic resonance imaging and psychophysiological recording.
The studies posed here are likely to produce both support and disconfirmation of different
the proposed model, hopefully with iterative refinement of the model over time. Although the
empirical evidence has yielded promising results, it is highly unlikely that any single gene or
group of genes, even in interaction with environmental histories, could account for much of the
variance in aggressive behavior.
The implications of these empirical findings and model for preventive intervention design
present yet another research agenda. Environmental engineering in light of personalized genetic
information has the exciting but unrealized potential for preventing the development of chronic
violence. So, too, environments and training interventions that alter social information
processing patterns offer the hope of secondary prevention among individuals at high risk due to
gene-environment risk profiles.
Social Information Processing 24
References
Archer, J. (2000). Sex differences in aggression between heterosexual partners: A meta-analytic
review. Psychological Bulletin, 126, 651-680.
Arnett, J. J. (2006). The psychology of emerging adulthood: What is known, and what remains to
be known? In J. J. Arnett & J. L. Tanner (Eds.), Emerging adults in America: Coming of
age in the 21st century (pp. 303-330). Washington, DC: American Psychological
Association.
Brendgen, M., Vitaro, F., Tremblay, R. E., & Lavoie, F. (2001). Reactive and proactive
aggression: Predictions to physical violence in different contexts and moderating effects
of parental monitoring and caregiving behavior. Journal of Abnormal Child Psychology,
29, 293-304.
Boyce W., Ellis B. (2005). Biological sensitivity to context: I. An evolutionary-developmental
theory of the origins and functions of stress reactivity. Development and
Psychopathology, 17, 271–301.
Brunner H., Nelen M., Breakefield X., Ropers H., & van Oost, B. (1993). Abnormal behavior
associated with a point mutation in the structural gene for monoamine oxidase A. Science
262, 578–580.
Cases O., Seif I., Grimsby J., Gaspar P., Chen K., Pournin, S., et al. (1995). Aggressive behavior
and altered amounts of brain serotonin and norepinephrine in mice lacking MAOA.
Science, 268, 1763–1766.
Caspi A., McClay J., Moffitt T.E., Mill J., Martin J., Craig I.W., et al. (2002). Role of genotype
in the cycle of violence in maltreated children. Science. 297, 851–854.
Social Information Processing 25
Charney D. S. (2004). Psychobiological mechanisms of resilience and vulnerability: implications
for successful adaptation to extreme stress. American Journal of Psychiatry, 161, 195–
216.
Cloninger, R., Sigvardsson, S., Bohman, M., & von Knorring, A. (1982). Predisposition to petty
criminality in Swedish adoptees. Archives of General Psychiatry, 39, 1242-1247.
Collins, W. A., & van Dulmen, M. (2006). "The course of true love(s)...": Origins and pathways
in the development of romantic relationships. In A. C. Crouter & A. Booth (Eds.),
Romance and sex in adolescence and emerging adulthood: Risks and opportunities (pp.
63-86). Mahwah, NJ: Erlbaum.
Conger, R. D., Cui, M., Bryant, C. M., & Elder, G. H. (2000). Competence in early adult
romantic relationships: A developmental perspective on family influences. Journal of
Personality and Social Psychology, 79, 224-237.
Crick, N.R., & Dodge, K.A. (1994). A review and reformulation of social information-processing
mechanisms in children's social adjustment. Psychological Bulletin, 115, 74-101.
Crick, N.R., & Dodge, K.A. (1996). Social-information-processing mechanisms in reactive and proactive
aggression. Child Development, 67, 993-1002.
Crockett, M. J., Clark, L., Tabibnia, G., Lieberman, M. D., & Robbins, T. W. (2008). Science, 320, 1739.
Crozier, J.C., Dodge, K.A., Fontaine, R.G., Lansford, J.E., Bates, J.E., Pettit, G.S., & Levenson, R.W.
(2008). Social information processing and cardiac predictors of adolescent antisocial behavior.
Journal of Abnormal Psychology, 117(2), 253-267.
Dick D. M., Bierut, L., Hinrichs, A., et al. (2006). The role of GABRA2 in risk for conduct
disorder and alcohol and drug dependence across developmental stages.". Behavior
Genetics, 36, 577–90.
Social Information Processing 26
Dick, D.M., Latendresse, S.J., Lansford, J.E., Budde, J.P., Goate, A., Dodge, K.A., Pettit, G.S., & Bates,
J.E. (in press). The role of GABRA2 in trajectories of externalizing behavior across development
and evidence of moderation by parental monitoring. Archives of General Psychiatry.
Dodge, K.A. (2003). Do social information processing patterns mediate aggressive behavior? In B.
Lahey, T. Moffitt, & A. Caspi (Eds.), Causes of conduct disorder and juvenile delinquency (pp.
254-274). New York: Guilford Press.
Dodge, K. A., Bates, J. E., & Pettit, G. S. (1990). Mechanisms in the cycle of violence. Science,
250, 1678-1683.
Dodge, K. A., Coie, J. D., & Lynam, D. (2006). Aggression and antisocial behavior in youth. In
W. Damon (Series Ed.) & N. Eisenberg (Vol. Ed.), Handbook of child psychology: Vol.
3. Social, emotional, and personality development (6th ed., pp. 719-788). New York:
Wiley.
Dodge, K.A., Lansford, J.E., Burks, V.S., Bates, J.E., Pettit, G.S., Fontaine, R., & Price, J.M. (2003). Peer
rejection and social information-processing factors in the development of aggressive behavior
problems in children. Child Development, 74, 374-393.
Dodge, K.A., Pettit, G.S., & Bates, J.E. (1994). Socialization mediators of the relation between
socioeconomic status and child conduct problems. Child Development, 65, 649-665.
Dodge, K.A., Pettit, G.S., McClaskey, C.L., & Brown, M. (1986). Social competence in children.
Monographs of the Society for Research in Child Development (Serial No. 213, Vol. 51, No.2).
Dodge, K.A., Pettit, G.S., Bates, J.E., & Valente, E. (1995). Social information-processing patterns
partially mediate the effect of early physical abuse on later conduct problems. Journal of
Abnormal Psychology, 104, 632-643.
Edwards, A., Dick, D., Dodge, K. A., Bates, J. E., Pettit, G. S., & Lansford, J. (2009). The interactive
effects between MAO-A and early maltreatment on later antisocial behavior. Unpublished
manuscript.
Social Information Processing 27
Eisenberger, N. I., Way, B. M., Taylor, S. E., Welch, W. T., & Lieberman, M. D. (2007). Understanding
genetic risk for aggression: Clues from the brain’s response to social exclusion.
Biological Psychiatry, 61, 1100-1108.
Evers, E. A., T., et al. (2006). Psychopharmacology (Berlin), 187, 200.
Fontaine, R.G., & Dodge, K.A. (2006). Real-time decision making and aggressive behavior in youth: A
heuristic model of response evaluation and decision (RED). Aggressive Behavior, 32, 604-624.
Fontaine, R.G., Yang, C., Dodge, K.A., Bates, J.E., & Pettit, G.S. (2008). Testing an individual systems
model of Response Evaluation and Decision(RED) and antisocial behavior across adolescence.
Child Development, 79(2), 462-475.
Furman, W. (2002). The emerging field of adolescent romantic relationships. Current Directions
in Psychological Science, 11, 177-180.
Ganzel, B., Casey, B. J., Glover, G., Voss, H. U., & Temple, E. (2007). The aftermath of 9/11:
Effect of intensity and recency of trauma on outcome. Emotion, 7, 227–238.
Holtzworth-Munroe, A., & Stuart, G. L. (1994). Typologies of male batterers: Three subtypes
and the differences among them. Psychological Bulletin, 116, 476-497.
Huesmann, L. R. (1998). The role of social information processing and cognitive schema in the
acquisition and maintenance of habitual aggressive behavior. In R. G. Geen & E.
Donnerstein (Eds.), Human aggression: Theories, research, and implications for social
policy (pp. 73-109). San Diego: Academic Press.
Jaffee, S.R., Caspi, A., Moffitt, T.E., Dodge, K.A., Rutter, M., Taylor, A., & Tully, L.A. (2005). Nature x
nurture: Genetic vulnerabilities interact with physical maltreatment to promote conduct problems.
Development and Psychopathology, 17, 67-84.
Social Information Processing 28
Kim-Cohen, J., Caspi, A., Taylor, A., Williams, B., Newcombe, R., Craig, I. W., et al. (2006). MAOA,
maltreatment, and gene–environment interaction predicting children’s mental health: new
evidence and a meta-analysis. Molecular Psychiatry, 11, 903-913.
Lansford, J.E., Malone, P.S., Dodge, K.A., Crozier, J.C., Pettit, G.S., & Bates, J.E. (2006). A 12-
year prospective study of patterns of social information processing problems and
externalizing behaviors. Journal of Abnormal Child Psychology, 34(5), 715-724.
Lansford, J.E., Miller-Johnson, S., Berlin, L.J., Dodge, K.A., Bates, J.E., & Pettit, G.S. (2007). Early
physical abuse and later violent delinquency: A prospective longitudinal study. Child
Maltreatment, 12(3), 233-245.
Liston, C., McEwan, B. S., & Casey, B. J. (2009). Psychosocial stress reversibly disrupts
prefrontal processing and attentional control. Proceedings of the National Academy of
Sciences, 106, 912-917.
Meyer-Lindenberg, A., Buckholtz, J. W., Kolachana, B., Hariri, A., Pezawas, L., Blasi, G., Wabnitz, A.,
Honea, R., Verchinski, B., Callicott, J. H., Egan, M., Mattay, V., & Weinberger, D. R. (2006).
Neural mechanisms of genetic risk for impulsivity and violence in humans. Proceedings of the
National Academy of Science Early Edition, www.pnas.org/cgi/doi/10.1073/pnas.0511311103.
Magdol, L., Moffitt, T. E., Caspi, A., Newman, D. L., Fagan, J., & Silva, P. A. (1997). Gender
differences in partner violence in a birth cohort of 21-year-olds: Bridging the gap
between clinical and epidemiological approaches. Journal of Consulting and Clinical
Psychology, 65, 68-78.
Morgan, Alex B. and Scott O. Lilienfeld. A meta-analytic review of the relation between
antisocial behavior and neuropsychological measures of executive function. Clinical
Psychology Review 20, no. 1 (2000): 113–136.
Social Information Processing 29
Murphy, F. C., Nimmo-Smith, I., & Lawrence, A. D. (2003). Cognitive, Affective, and
Behavioral Neuroscience, 3, 207-233.
Nay, S., Dodge, K. A., Lansford, J., Pettit, G. S., & Bates, J. E. (2009). Rejection sensitivity as a
mediator of the effect of romantic relationship history on violent behavior in romantic
relationships. Unpublished manuscript.
Orobio de Castro, B., Veerman, J. W., Koops, W., Bosch, J. D., & Monshouwer, H. J. (2002).
Hostile attribution of intent and aggressive behavior: A meta-analysis. Child
Development, 73, 916–934.
Ortiz, J., & Raine, A. (2004). Heart Rate Level and Antisocial Behavior in Children and
Adolescents: A Meta-Analysis. Journal of the American Academy of Child & Adolescent
Psychiatry, 43(2), 154-162.
Pettit, G. S., Lansford, J. E., Dodge, k. A., & Bates, J. E. (2009). Domain specificity in social-
information processing and violent behavior in early adulthood. Under editorial review.
Pezawas, L., Meyer-Limndenberg, A., Drabani, E. M., Verchinski, B. A., Munoz, K. E.,
Kolachana, B. S., Egan, M. F., Mattay, V. S., Harrirri, A. R., & Weinberger, D. R.
(2005). Nature Neuroscience, 8, 828-834.
Raine, A. (2002). Annotation: The role of prefrontal deficits, low autonomic arousal and early
health factors in the development of antisocial and aggressive behavior in children.
Journal of Child Psychology and Psychiatry, 43, 417.
Raine, A. (2008). From genes to brain to antisocial behavior. Current Directions in
Psychological Science, 17, 323-328.
Raine, A., Buchsbaum, M., & LaCasse, L. (1997). Brain abnormalities in murderers indicated by
positron emission tomography. Biological Psychiatry, 42, 495-508.
Social Information Processing 30
Raine, A., Venables, P. H., & Mednick, S. A. (1997). Low resting heart rate at age 3 years
predisposes to aggression at age 11 years: Evidence from the Mauritius Child Health
Project. Journal of the American Academy of Child & Adolescent Psychiatry, 36(10),
1457-1464.
Sanfey, A. G., Rilling, J. K., Aronson, J. A., Nystrom, L. E., & Cohen, J. D. (2003). Science,
300, 1755.
Shih J., Chen, K., & Ridd, M. (1999). Monoamine oxidase: from genes to behavior. Annual
Review of Neuroscience, 22, 197–217.
Silverman, J. G., Raj, A., Mucci, L A., & Hathaway, J. E. (2001). Dating violence against
adolescent girls and associated substance use, unhealthy weight control, sexual risk
behavior, pregnancy, and suicidality. Journal of the American Medical Association, 286,
572-579.
Straus, M. A. (1999). The controversy over domestic violence by women: A methodological,
theoretical, and sociology of science analysis. In X. B. Arriaga & S. Oskamp (Eds.),
Violence in intimate relationships (pp. 17-44). Thousand Oaks, CA: Sage.
Tabibnia, G., Satpute, A. B., & Lieberman, M. D. (2008). Psychological Science, 19, 339.
van Goozen, S. H. M., Fairchild, G., & Harold, G. T. (2008). The role of neurobiological deficits
in childhood antisocial behavior. Current Directions in Psychological Science, 17, 224-
228.
Waldman, I. D. (2008). The etiology of hostile perceptual biases and their relation with
children’s aggression. Paper presented at the annual meeting of the Behavior Genetics
Association.
Social Information Processing 31
Weiss, B., Dodge, K.A., Bates, J.E., & Pettit, G.S. (1992). Some consequences of early harsh discipline:
Child aggression and a maladaptive social information processing style. Child Development, 63,
1321-1335.
Social Information Processing 32
Figure Captions
Figure 1. A proposed model of genetic, environmental, and processing mechanisms in the
development of aggressive behavior.
Social Information Processing 33
ViolenceProcessingSocial
Information
Processing
Neural
Activity
Psychophysio-
logicalActivity
Genes
Early
Environment
Situational
Environment
Situational
Stimulus