AGGRESSION
Social Cognitive Skills and Intelligence as Predictors of Proactive and Reactive Aggression in a High Risk Sample of School-aged Boys
Kimberly C. Kuiper Leiden University
Faculty of Social and Behavioral Sciences
Developmental Psychopathology in Education and Child Studies Research Master Thesis
Supervisors: Prof. dr. J.T. Swaab E.M. van Zonneveld, MSc
Preface
In the previous 12 months of completing my research thesis I’ve felt supported by a number of people whom I would like to acknowledge. First of all I would like express my gratitude to my supervisors prof. Hanna Swaab and Lisette van Zonneveld for the opportunity to work on this project. It has been a unique journey of great experiences and especially the contact with the children I have greatly appreciated. Moreover, a special thanks goes to my parents. They have always supported me and were – perhaps one morning too many – part of my project since it required me to set my alarm at 6:00. Furthermore, I’d like to thank my fellow research master students of 2011-2013 without whom I would not have succeeded this two-year master program. We pushed each other to the limit and made sure the occasional relax-periods existed as well (i.e. doing ‘Just Dance’, surprises with Sinterklaas). A special thank-you goes to my ‘neuro-twin’ Roxanna with whom I’ve been confused with for countless times by numerous parties in these last two years and in whom I have found a true friend. Lastly, I would like to thank Martine, Sharissa, Tom, Ellen, and Tim for the support they have shown me in my academic years and provided me with the little (academic-unrelated) nudges I needed at times.
Abstract
Externalizing behavior problems in children are of substantial cost and highly prevalent in elementary school-aged children. Specific subtypes of externalizing behavior problems, reactive and proactive aggression, have been uniquely related to specific risk factors,
correlates, and consequences. The purpose of this study was to examine the relations between social cognitive skills, intelligence, and reactive and proactive aggression in elementary school-aged children. These children were at risk for developing serious externalizing behavior problems and recruited within a program focused on reduction and prevention of delinquency, led by the municipality of Amsterdam in The Netherlands. Intelligence and social cognitive skills tasks were administered in fifteen boys in the ages of 7-12 years (M = 10.38, SD = 1.95). The results from the present study indicate that the relation between intelligence and reactive aggression is mediated by social cognitive skills, and more precisely the ability to identify emotions. No such relations were found for proactive aggression or other social cognitive skills. It is concluded that the relation between social cognitive skills and aggressive behavior depends at least partly on intelligence and that interventions for reactive-aggressive children should focus on training of emotion recognition without losing sight of the important role of intelligence.
Keywords: reactive and proactive aggression, social information processing, intelligence, cognitive skills
Introduction
Aggressive and rule-breaking behavior is a well-known phenomenon in typically developing children and is not necessary worrying. Some typical non-psychopathological developmental periods during childhood and adolescence that are characterized by aggressive and rule-breaking behavior have been reported by Campbell (2002) and the American Psychological Association (2002). Examples are the defiant toddler and the oppositional adolescent periods. Aggressive and rule-breaking behavior only becomes problematic when the child or
adolescent intentionally aims his or her behavior at another person, animal, or object. This harmful or threatening behavior to the external environment is also known as externalizing behavior (Achenbach, 1978; Campbell, Shaw & Gilliom, 2000).
It has been proven difficult to estimate the exact prevalence of externalizing behavior problems in children and adolescents since multiple assessments and diverse definitions are used. Moreover, ratings of externalizing behavior problems vary slightly between parents and teachers. A large study with a population-based sample from the Netherlands revealed that of parents of children between 6 and 12 years of age, 21% reported deviant externalizing
behavior problems whereas the children’s teachers reported higher ratings of deviant externalizing behavior problems (23%; Tick, van der Ende, & Verhulst, 2007). In addition, the percentage of children that were reported by their parents as exhibiting deviant rule-breaking behavior and deviant aggressive behavior were 14% and 9% respectively, whereas for teacher-reported behavior the percentages were 16% and 9%. In sum, at least 1 or 2 in 10 children who are in elementary school show behavior problems, indicating that externalizing behavior problems represent a clinically relevant, social issue.
Externalizing behavior problems are not only problematic because of its orientation to the external environment, but also because of its far-reaching consequences. The
consequences stretch on a variety of domains: the aggressor himself – in terms of
psychopathology and well-being – and the victims, the environment, and the government. It can be argued that children with externalizing behavior problems are costly not only to the affected families, but also to the government that attend to the care of these children and their families. These children are costly in view of the fact that they are in need for more help and service, such as medication, treatment, or therapy, than normally developing children without serious behavior problems (Foster & Jones, 2005; Scott, Knapp, Henderson & Maughan, 2001; Van Kan, 2009). Moreover, since the existence of behavior problems in childhood is a risk factor for developing psychopathology in adolescence and adulthood (Hofstra, Van der Ende, & Verhulst, 2000; Van Beijsterveldt, Bartels, Hudziak, & Boomsma, 2003), programs focusing on prevention of the problem behavior may be fruitful although not cost-free.
Studies have tried to estimate the relative costs of children who exhibit externalizing behavior problems compared to those children who do not. Findings are clear in that children who show externalizing behavior problems are expensive to the government, demanding more service from medical and psychological health care, educational systems, and public family services (Foster & Jones, 2005; Scott et al., 2001; Van Kan, 2009).
Given the substantial costs of externalizing behavior problems and its high prevalence, it is important to identify which children are at risk for developing these problems. Several specific risk factors for developing serious externalizing behavior problems have been identified and can be categorized in the following: risk factors during pregnancy, individual risk factors, family risk factors, and social and environmental risk factors. Prenatal risk factors such as the use of drugs (alcohol, substances, and tobacco) during pregnancy, and complications during delivery are known to be related to an increased risk of aggressive behavior (Liu & Wuerker, 2005; Raine, 2002). Individual risk factors
include endocrinological and physiological risk factors, such as deviant cortisol hormone levels, and boys are known to portray more openly aggressive behavior than girls (Rappaport & Thomas, 2004). Antisocial and aggressive behavior in early childhood and presence of psychopathological problems such as ADHD, ODD, and CD, appear have been related to increased risk of aggressive behavior as well (Liu & Wuerker, 2005; Rappaport & Thomas, 2004). Risk factors within the family environment include a negative or hostile attitude within the family – such as abuse, neglect, and sexual abuse – (Liu & Wuerker, 2005; Rappaport & Thomas, 2004). Finally, hostile and anti-social peers, low social-economic status (SES), and peer rejection are known social and environmental risk factors associated with the development of aggressive and rule-breaking behavior (Liu & Wuerker, 2005; Loeber & Hay, 1997).
Externalizing behavior in childhood does not only pose direct problems for a child and his or her surrounding in early childhood, but it can also have far-reaching consequences at later points in time such as the adolescence and adulthood periods. Several longitudinal studies have indicated that early aggression in childhood is a highly stable behavioral characteristic that persists into adolescence and early adulthood (Hofstra et al., 2000; Van Beijsterveldt et al., 2003). Moreover, a six-site, cross-national study showed that (physical) aggression in childhood increases the risk for persistent violence and delinquency during adolescence (Broidy et al., 2003). These findings suggests that externalizing behavior problems seem to form at least part of the core of criminal behavior of juvenile offenders. Broidy et al. (2003) also provided some evidence for the hypothesis that differential pathways exist, in that different patterns of early behavior problems are related to differing delinquent outcomes. The results suggested that early nonaggressive conduct problems act as a specific
risk factor for later violent delinquency and early oppositional behaviors for nonviolent delinquency.
Over the years, researchers and practitioners have tried to disentangle different types of externalizing behavior problems in order to adequately respond to these problems in terms of prevention and intervention. In this study, the main focus is on children who exhibit aggressive and rule-breaking behavior. However, the term aggressive behavior, along with the term externalizing behavior problems, is an umbrella term for diverse behavior, such as cruelty to animals, bullying with violence, and even armed assault. To clarify what
constitutes as aggressive behavior, the distinction between reactive and proactive aggression has been made (Dodge, 1991). Reactive aggression is characterized by “hot-blooded” anger, menacing hostile attacks, defensive postures, and a lack of self-control. Children who exhibit reactive aggression are more likely to have high levels of anxiety (Vitaro, Brendgen, & Tremblay, 2002). It is suggested that the aggressive behavior is a reaction to an – as
perceived by the child – anxious situation. Reactive aggression has been linked to peer and parent rejection and internalizing problem behavior (Vitaro et al., 2002). On the other hand, proactive aggression is characterized by “cold-blooded” anger, less emotional, highly
organized, and driven by the expectation of reward (Dodge, Lochman, Hamish, Bates & Petit, 1997). For children with proactive aggression, early patterns of physical aggressive behavior are present along with a lack of anxiety (Vitaro et al., 2002). Although children can show both types of aggressive behavior, evidence is extensive in proving that reactive and
proactive aggression are distinct concepts based on antecedents, correlates, and consequences (Fite, Raine, Stouthamer-Loeber, Loeber, & Pardini, 2010; Hubbard, McAuliffe, Morrow & Romano, 2010; Vitaro et al., 2002).
In sum, externalizing behavior problems and specifically aggressive and rule-breaking behavior appears to be highly present and significantly problematic for numerous reasons. To understand why some children develop externalizing behavior problems and others do not, the present study will focus on key constructs that play a role in the development of these behavior problems: social information processing and intelligence.
Social information processing
To be successful in social interactions and everyday functioning, adequate social cognitive skills are required. Biased or distorted cognitions and interpretations of situations could lead to disturbed and inappropriate responses in social interactions, such as aggression. Distorted social cognitive skills could also lead to poor adjustment later in life, given that poor social cognitive skills are related to antisocial behavior and peer rejection in children, which are two of the most consistent behavioral precursors of adult psychopathology and criminal behavior (Cowen, Pederson, Babigan, Izzo, & Trost, 1973; see review Coie, 2004).
In explaining the role of social cognitive skills in children’s aggressive behavior, the social information processing model (SIP-model) has been proven useful (Dodge, 1986; reformulated by adding two last steps by Crick & Dodge, 1994). The reformulated model includes six steps that lead from a particular social stimulus to a behavioral response: 1) encoding of cues, 2) interpretation of cues, 3) clarification of goals and response search, 4) response decision, 5) behavioral enactment, and 6) evaluation. A developmental focus was added by Selman and Byrne (1974) that included four developmental levels of role-taking between the ages of 4 and 12 years: level 0 of egocentric role-taking, level 1 of subjective role-taking, level 2 of self-reflective role-taking, and level 3 of mutual role-taking. Eight social cognitive skills can be derived from these four developmental levels to characterize the social information processing (Gerris, 1981): 1) identifying, 2) discriminating,
Table 1
Overview of Social Information Processing Model, Levels and Skills Reformulated Dodge model
by Crick and Dodge (1994)
Social cognitive levels (Selman & Byrne, 1974)
Social cognitive skills (Gerris, 1981) 1. Encoding of cues Egocentric role-taking Identifying; discriminating 2. Interpretation of cues Subjective role-taking Differentiating; comparing 3. Clarification of goals and
response search
Self-reflective role-taking Perspective taking; relating
4. Response decision Mutual role-taking Coordinating; taking into account
5. Behavioral enactment 6. Evaluation
3) differentiating, 4) comparing, 5) perspective-taking, 6) relating, 7) coordinating, and 8) taking into account different perspectives. These social cognitive skills are essential in order to form a theory-of-mind which refers to the child’s capacity to understand that others hold mental states (i.e. beliefs, goals, intentions and feelings) and that these states can differ from their own (Carpendale & Chandler, 1996). A theory-of-mind also involves the child’s understanding that the same stimulus or situation can be interpreted by others in different ways (Lalonde & Chandler, 2002) and that the feelings and actions of others are based on their beliefs (Harris, Johnson, Hutton, Andrews & Cooke, 1989). An overview of these models and skills is portrayed in Table 1.
Deficits in theory-of-mind skills are potentionally important predictors of aggressive behavior (Crick & Dodge, 1994; Harvey, Fletcher & French, 2001). However, the link
between social cognitive skills and aggression might not be similar for the subtypes of aggression: reactive and proactive aggression. It has been argued that inadequate theory-of- mind skills could lead to hostile attribution bias which is more pronounced in reactive-aggressive children (Orobio de Castro, Veerman, Koops, Bosch, & Monshouwer, 2002). In contrast, greater social cognitive skills could be found in children who exhibit proactive aggressive behavior in that these children deliberately choose aggressive behavior in social situations because they expect it will effective in achieving their goals (Renouf et al., 2010). Evidence for this line of reasoning comes from a study with children in kindergarten by Renouf et al. (2010). Their results showed that a negative relationship between theory-of-mind skills and reactive aggression was present but only in children who were frequently victimized by peers (Renouf et al., 2010). In contrast, a small study (Van Manen, Prins & Emmelkamp, 2001) revealed no significant differences in social cognitive skills between reactive and proactive aggressive children (proactive: N = 8; reactive: N = 14). The present study aims to replicate the findings by Renouf et al. (2010) in that lower social cognitive skills – as assessed by an extensive theory-of-mind task – are in fact related to reactive aggression and greater social cognitive skills to proactive aggression, using a different approach than the Van Manen et al. (2001) study to differentiate the subtypes of aggression.
If aggression is perceived as an inadequate response to (ambiguous) social situations, aggressive subgroups might exhibit distinctive social information processing (Dodge & Coie, 1987). For example, reactive-aggressive children display more hostile attribution to
ambiguous stimuli than nonaggressive children (Dodge & Coie, 1987). In comparison, although proactive-aggressive children can accurately encode and interpret social cues, these children believe that aggression is a socially accepted and effective means for achieving personal gains. In addition, studies that have used the SIP-model have shown that
reactive-aggressive children exhibit more problems in the first two stages of the model namely encoding (1) and interpretation of social cues (2) and not in later stages (Crick & Dodge, 1996; Dodge & Coie, 1987). In contrast, proactive-aggressive children experience more problems in the two later stages of the SIP-model; response decision (4) and behavioral enactment (5). Because the eight social cognitive skills by Gerris (1981) that are used in the theory-of-mind task in the present study are derived from the SIP-model and follow the same steps as the SIP-model, this study will focus on the prediction that reactive aggression in children is specifically related to deficits in the first four cognitive skills which rely on
encoding and interpretation of social cues. In contrast, it is expected that proactive aggression is specifically related to deficits in the last two cognitive skills which rely on response
decision. Intelligence
In order to be successful in social interactions and prevent inadequate responses to ambiguous contexts, next to social cognitive skills an individual must rely on verbal and nonverbal processing skills as well. In social interactions, an individual needs to understand what the other is talking about – verbal processing ability – and an individual needs to organize the cues of interaction (e.g. facial expressions, hand gestures, context) into a
meaningful and coherent construct – nonverbal processing skills. As was reported previously, numerous studies have examined social cognitive skills in aggressive children (e.g. Crick & Dodge, 1994, for meta-analysis see Yoon et al., 1999). Akhtar and Bradley (1991) concluded in their paper on the present findings and implications of social information processing deficits of aggressive children, however, that many studies are characterized by
methodological weaknesses, in that intelligence and language abilities are rarely assessed or controlled for, although they are known to be related to childhood aggression (Patterson,
DeBaryshe & Ramsey, 1989). Moreover, Loeber and Hay (1997) concluded in their review on the development of aggression from childhood to adulthood, that general intellectual abilities at least partly lead to persistent aggression, given that antisocial behavior is often accompanied by academic problems. In addition, a meta-analysis revealed that not
educational performance but low intelligence is a predictor of later delinquency (Maguin & Loeber, 1996).
Slightly more insight has been gathered in studies that examined the specific relationship between verbal intelligence on the one hand and proactive and reactive aggression on the other hand. First of all, it is assumed that when children get older their cognitive abilities increase and thereby will result in more planned (proactive) aggression than mere reactive aggression (Kempes, Matthys, De Vries, & Van Engeland, 2005). A study by Arsenio, Adams, and Gold (2009) revealed that in high-risk adolescents, reactive
aggression is related to lower verbal intelligence whereas proactive aggression is related to higher verbal intelligence. Apparently, children can benefit from greater verbal skills in ambiguous social interactions leading to more planned behavior to achieve their goals. In contrast, reactive aggressive children have difficulties relying on their verbal abilities to solve ambiguous social interactions resulting in more reactive, unplanned or “hot-headed”
behavior. However, all three types of intelligence (general cognitive ability or general intellectual abilities, verbal intelligence, and performance intelligence) could be valuable factors in the relationship of aggression and social cognitive skills in children. The present study will try to clarify the relationship between general, verbal, and performance intelligence with proactive and reactive aggression. The relationship between intelligence and the
subtypes of aggression (proactive and reactive aggression) remain unclear and in most
examine the predictive value of intelligence on both proactive and reactive aggression. Moreover, it will focus on replicating the findings of Arsenio et al. (2009) that lower verbal intelligence is related to reactive aggression whereas proactive aggression is related to higher verbal intelligence.
The central research question in this study is whether intelligence and social cognitive skills are predictors of reactive and proactive aggression. Based on the previously described research and the SIP-model, a number of hypotheses have been formulated. The first
hypothesis predicts that proactive-aggressive children have more difficulty in the fourth step ‘response decision’ of the SIP-model, that corresponds with greater social cognitive skills. The second hypothesis predicts that reactive-aggressive children have more difficulty in the first two steps, ‘encoding’ and ‘interpretation of cues’, that correspond with basic social cognitive skills. Third, it is predicted that reactive aggression is related to lower verbal intelligence and proactive aggression to higher verbal intelligence. Moreover, the predictive value of general and performance intelligence on reactive and proactive aggression will be examined too.
Methods Background
This study took place as part of the ‘Top 600’ initiative started by several parties, including the police and the Public Prosecutor, in the city of Amsterdam led by the municipality of Amsterdam (the Netherlands). Although the city of Amsterdam had
witnessed decreased number of crimes in the years leading to 2011, the offences on the other hand became more severe and moreover and most offenders were young, violent, and hard to reach. In order to diminish or eliminate these severe offences and crimes, the parties involved aimed at dealing harshly with this group of offenders and at preventing that other and
younger children would act criminally in the future. To achieve these goals, the initiative of the Preventive Intervention Team [Preventief Interventie Team; PIT] was born. The central focus of the PIT was to locate children at risk and to respond to their needs and problems with adequate help and care. Children at risk were recruited from numerous sources: several elementary schools, Bureau Compulsory Education Plus [Bureau Leerplicht plus], and relatives of offenders on the ‘Top 600’ list. The latter group consisted of children of whom a direct family-member employed multiple crimes or delinquencies and this family-member was serving a jail-sentence lasting longer than three days.
Participants
The participants were 15 boys in the ages of 7 and 12 years (M = 10.38, SD = 1.95) who were in elementary school at the time of the assessments. Majority of the participants (n = 12) were recruited from 6 elementary schools bonded to the PIT and were perceived by their teacher as being highly aggressive, difficult to head for, or who exhibit strange behavior or get stuck with their schoolwork because of their behavior. In addition, younger brothers and children of ‘Top 600’ offenders were included in the study too (n = 3). None of the participants were recruited because of referral by the Bureau Compulsory Education Plus. The participants who were recruited by means of either the ‘Top 600’ list or teachers’ concerns did not differ on any of the background variables (ethnicity: χ²(5) = 5.83, p = .32; age: t(13) = .37, p = .72).
Two girls were initially included in the study sample. There were no significant gender differences on any of the variables, however comparison between genders was untrustworthy (2 girls against 15 boys) and led to bias in further analyses. After careful consideration the two girls were not included in this study to obtain a coherent and homogenous sample.
Measurements
Externalizing behavior problems. For each child a teacher completed the Dutch Teacher Report Form (TRF) that consists of 118 items addressing school functioning, school performance, emotional and behavioral problems, and a broad range of internalizing and externalizing symptoms of 6-18 year old children (Achenbach, 1991; for the Dutch version see Verhulst, Van der Ende, & Koot, 1997). The TRF has adequate reliability and validity. Dutch normative data are available to compute age-corrected T-scores (Verhulst et al., 1997). Boys were classified as exhibiting externalizing behavior problems and included in the study if their aggressive behavior and/or rule-breaking behavior scores (of the Externalizing Behavior scale) on the TRF were in the borderline or clinical range (T-scores equal to or higher than 65).
Reactive and proactive aggression. To measure and differentiate the form and frequency of children’s aggression, teachers completed the Dutch Instrument for Reactive and Proactive Aggression (IRPA; Polman Orobio de Castro, Thomaes & Van Aken, 2008, modified from Kupersmidt, Willoughy & Bryant, 1998). The IRPA consists of three form scales including; physical aggression (hitting, kicking, pushing), verbal aggression (name calling, arguing), and covert aggression (doing sneaky things, gossiping). The questionnaire consists of 7 form items in which the frequency of each form is rated on a 5-point scale (0 = never, 1 = once or twice, 2 = weekly, 3 = several times a week, 4 = daily). Teachers were instructed to exclusively rate the child’s behavior that took place in the preceding month. When a score of 1 or higher was given on the form item (that is, the form of aggression was present), the function of the aggression was assessed using 6 function items including 3 proactive items (to hurt someone or to be mean, to be the boss, because this child takes
child felt threatened by someone, because this child was angry). These items were rated on a 5-point scale (0 = never, 1 = rarely, 2 = sometimes, 3 = most of the time, 4 = always). In total, 6 function scores were computed by summing functions across forms. Higher scores on proactive functions do not indicate more proactively aggressive acts, but indicate that if a child behaves aggressively, its function is mainly proactive. The IRPA has satisfactory to excellent discriminant and convergent validity (Polman et al., 2008). Cronbach’s alpha in this sample was .96 for reactive aggression and .90 for proactive aggression. Reactive and
proactive aggression were not significantly correlated in this sample, r = .37, p = .18. In this study, the 3 proactive functions were summed together to compute a total score for proactive aggression and similarly for the 3 reactive functions to compute a total score for reactive aggression.
Intelligence. The intelligence of the participants was assessed using the Dutch Wechsler Intelligence Scale for Children third edition (WISC-IIINL), a well-used intelligence test for 6-to 16-year-old children intended to measure the child’s general cognitive abilities (Kort et al., 2005). The Dutch WISC-IIINL has adequate reliability and validity similar to the original version. The results on ten subtests generate a full scale IQ score (FSIQ), a verbal IQ score (VIQ), and a performance IQ score (PIQ; all scales have a M = 100, SD = 15). The verbal IQ score is an estimate of verbal skills such as verbal reasoning and verbal
comprehension, whereas the performance IQ score appeal to various visuospatial abilities such as motor skills and detail perception.
Social information processing. The Social Cognitive Skills Test (SCST) was used to assess social cognitive skills – that correspond with the SIP-model discussed above – and consists of seven short stories with corresponding drawings (Van Manen, Prins &
Figure 1. Partial illustration of Story 1 ‘Little boat’ [Bootje] of the SCST. Illustration have been cut for explanation purposes in this paper only.
child) encounters a troublesome situation with another child or adult. The child was
questioned over each story through eight questions that correspond with eight social cognitive skills (identifying, discriminating, differentiating, comparing, perspective taking, relating, coordinating, and taking into account). While the child looked at the drawings of the story, the tester read the story aloud and questioned the child. An example of a story is portrayed in Figure 1. Only part of the story is shown here for explanation purposes; the original story
includes six pictures. The accompanying story for these four pictures is as follows: “1) This boy is playing with his self-made boat near the water. 2) His little boat sinks. 3) He walks home. 4) At home is his little brother playing with a little boat”. The questions that
correspond with the first social cognitive skills ‘Identifying’ and ‘Discriminating’ are: “How does the boy feel in picture 1?” and “In which two pictures are the boy and his brother feeling the same?”. The SCST has an average duration of 30 minutes per child. It has adequate discriminant validity; its reliability has not yet been investigated (Van Manen et al., 2001). Dutch normative data is available for boys in the ages of 4 and 12 years. The present study specifically used the age-corrected norm scores for the eight social cognitive skills in the analyses.
Procedure
When children were reported to the PIT, two PIT employees visited the families at their homes to provide information regarding the assessment of the child. The families received a letter with the same information. After agreement on participation in the
assessment, parents and children above the age of 12 years signed an informed consent form. The assessment of the children took place at their schools and consisted of two test sessions on two different days. The beginning of both test sessions took place in the morning around the time school started. The first test session had an average duration of 3.5 hours in which the WISC-IIINL and several questionnaires were administered. The second test session had an average duration of 6 hours in which the SCST along with other neuropsychological tests were administered. The results of these other tests and questionnaires are not reported in this study. The IRPA was given to the child’s teacher before the beginning of the first test session and collected afterwards on either the same day or on the day of the second test session.
The assessment of the children was performed by two trained Master students in Education and Child Studies. The first student acted as a tester and the second student as an observer and assistant of the tester. The students fell under the supervision of an educational psychologist who was not always present during test sessions as either a tester or an observer.
The children were individually tested in a quiet room allocated by the school. The child was seated directly across the table from the tester. A tape-recorder was used to record the SCST in order to encode and score the child’s answers afterwards. Encoding was
performed by the same students. Statistical analysis
All analyses were performed using IBM Statistics SPSS for Windows version 20.0 (IBM Corp., 2011). Before the analyses were conducted, several hierarchical-specific
assumptions were checked. Overall data-inspection was conducted in which possible outliers or influential cases and missing values were examined. A Missing Value Analysis routine was run in SPSS to examine patterns in missing values and their meaning. If no clear pattern was present in the missing values, the option listwise deletion was used in the analyses. Possible outliers were examined before data-analysis and removed if the case appeared to be unrepresentative of the remaining sample. Since normality of residuals is a hierarchical-specific assumption, box plots of residuals were examined after analyses for possible outliers. If these outliers influenced the normality of residuals, those cases were removed as well. Another hierarchical-specific assumption that should be met is that multicollinearity shouldn’t be present or at least minimally present. Multicollinearity causes problems in hierarchical regression analyses because it can lead to untrustworthy beta coefficients and suppression of R-square. Multicollinearity was examined using a correlation matrix with all variables of interest in which correlations between predictors greater than .80 are viewed as
problematic. Also collinearity statistics were examined for acceptable limits, that is when VIF values were fewer than 10 and Tolerance values greater than .1.
To answer the research questions whether reactive and proactive aggression could be predicted from intelligence and social cognitive skills, six hierarchical regression analyses were performed. The score for proactive aggression was the response variable for three of the analyses, whereas in the other three analyses the score for reactive aggression was the
response variable. The predictor variables were entered in two subsequent steps in all analyses. At Step 1, the WISC-IIINL scores of FSIQ, VIQ, and PIQ were entered in three separate analyses for each type of aggression. The last two social cognitive skills were entered at Step 2 of the regression analyses with proactive aggression as the dependent variable, whereas the first four social cognitive skills were entered at Step 2 of the regression analyses with reactive aggression as the dependent variable.
The WISC IQ-scores were entered in the first step before entering the SCST skills. The order of including these variables was based on findings of studies indicating a high stability across life-time of IQ – as assessed by standardized intelligence tests – that is largely explained by genetic influences (Bartels, Rietveld, Van Baal, & Boomsma, 2002; Lyons et al., 2009) and that (verbal) IQ has differential links to reactive and proactive aggression (Kempes et al., 2005). In contrast, the relationship between IQ and social
cognitive skills is less clear, although it can be hypothesized that (general) cognitive ability is an essential and preceding factor in the development of social cognitive skills. Since the aggression functions (reactive and proactive) were not significantly related, they were not included as a controlling variable in the regression analysis (that is reactive aggression as controlling factor, when predicting proactive aggression). Furthermore, age was not
a controlling variable (proactive aggression: r = .24, p = .39; reactive aggression: r = .10, p = .73).
To further examine the relations between intelligence, social cognitive skills, and reactive and proactive aggression, mediation analysis was performed. Only significant
relations of social cognitive skills with reactive and proactive aggression were included in the mediation analysis. The analysis followed the procedure of Baron and Kenny (1986) in which mediation is present when four conditions are satisfied (not explained in this paper). The Sobel test was used to determine whether the mediation model was significant (Sobel, 1982, 1986).
Results Data inspection
Before analyses were performed, outliers and assumptions were examined and decided upon using criteria discussed previously. Only one participant was identified as an outlier, since his WISC-IIINL IQ scores were above 100 whereas all other participants had IQ scores below 90 and since his SCST scores were overall higher than those of the other participants. Therefore, this participant was not representative of the remaining sample and was excluded from all analyses. After regression analyses two other participants were identified for influencing the normality of the residuals and excluded from all analyses. No clear patterns were present regarding missing values, the option listwise deletion was used in all analyses. All other statistical assumptions were met. Tests for multicollinearity indicated that a low level of multicollinearity was present and that the values involved were in
acceptable limits.
General descriptive statistics of the variables of interest in this study are presented in Table 3. The distribution of the FSIQ scores was as follows: 14 participants (93%) had a
Table 3
Descriptive Statistics of Major Study Variables (N = 15)
M SD Age 10.38 1.95 TRFa Rule-breaking behavior 68.73 7.92 Aggressive behavior 76.73 12.60 Externalizing behavior 73.53 8.17 IRPAb Proactive aggression 22.27 16.18 Reactive aggression 33.80 24.76 WISC-IIINL IQ scoresc FSIQ 69.07 6.81 VIQ 72.53 5.88 PIQ 70.60 9.11 SCSTd 1. Identifying 8.73 1.79 2. Discriminating 10.47 3.09 3. Differentiating 8.07 2.91 4. Comparing 8.33 3.79 5. Perspective taking 9.47 2.29 6. Relating 8.33 2.13 7. Coordinating 8.40 1.84
8. Taking into account 8.07 2.91
Note. Displayed scores have different scales: a T-scores (M = 50, SD = 10), b computed sum scores, c IQ scores (M = 100, SD = 15), d age-corrected norm scores (M = 10, SD = 3).
Table 4
Correlations between Aggression Functions, Age, WISC-IIINL IQ Scores, and Social Cognitive Skills
1 2 3 4 5 6 7 8 9 10 11 12 13
1. IRPA Proactive aggression score
2. IRPA Reactive aggression score .37
3. Age .24 .10 4. FSIQ -.46 -.40 -.71** 5. VIQ -.47 -.42 -.70** .73** 6. PIQ -.27 -.25 -.51 .90** .37 7. SCST 1. Identifying -.40 -.85** -.46 .64* .51 .52* 8. SCST 2. Discriminating -.04 -.14 .20 -.51 -.22 -.63* .11 9. SCST 3. Differentiating -.51 .14 .12 .05 .16 -.06 -.15 .22 10. SCST 4. Comparing -.22 -.15 .21 -.05 -.10 -.06 .15 .37 .23 11. SCST 5. Perspective taking -.49 -.24 -.09 .22 .09 .18 .41 .48 .54* .46 12. SCST 6. Relating -.17 .02 .23 .01 -.21 .12 .12 .39 .48 .50 .61* 13. SCST 7. Coordinating -.40 .03 .09 .25 .19 .19 .01 -.01 .50 .58* .24 .51
14. SCST 8. Taking into account -.36 -.05 .07 .13 .21 .01 -.03 -.02 .15 .58* .05 .07 0.81**
FSIQ score below 80 and only one participant had a FSIQ in the range of 80-89. The
intercorrelations between the variables of interest are presented in Table 4. The IRPA reactive and proactive scores were not significantly related (r = .37, p = .18). The first social cognitive skill ‘Identifying’ had significant correlations with the WISC-IIINL FSIQ and PIQ scores. The WISC-IIINL FSIQ, VIQ, and PIQ scores were significantly related to each other as well. As discussed above, multicollinearity could cause problems in hierarchical regression analyses. In order to diminish multicollinearity problems, three separate analyses with each of the WISC-IIINL IQ scores as a separate predictor were performed for each type of aggression. Hierarchical regression analyses
Proactive aggression. In order to answer the first research question – whether proactive aggression can be predicted by intelligence and greater social cognitive skills – separate hierarchical regression analyses were performed with proactive aggression as dependent variable, and WISC-IIINL IQ scores and the last two social cognitive skills as predictors. The WISC-IIINL FSIQ, VIQ, and PIQ scores were entered in Step 1, followed by the last two social cognitive skills of the SCST in the Step 2. The results of the regression analyses could not answer the research question (see Table 5). None of the final regression models including intelligence and social cognitive skills predicting proactive aggression reached significance (FSIQ: F(3,14) = 1.63, p = .24; VIQ: F(3,14) = 1.75, p = .22; PIQ: F(3,14) = .98, p = .44).
Reactive aggression The second research question – whether reactive aggression can be predicted by intelligence and social cognitive skills – was also examined with multiple hierarchical regression analyses. It followed the same procedure as the regression analyses regarding proactive aggression, but now the first four social cognitive skills were entered in
Table 5
Hierarchical Regression Models Predicting Proactive Aggression Proactive aggression
FSIQ in step 2 VIQ in step 2 PIQ in step 2
Predictor R2 ΔR2 β R2 ΔR2 β R2 ΔR2 β
Step 1. IQ .21 .21 .22 .22 .07 .07
FSIQ / VIQ / PIQ -.40 -.41 -.23
Step 2. SCST .31 .10 .32 .10 .21 .14
7. Coordinating -.14 -.27 -.17
8. Taking into account -.20 -.06 -.22
Note. All β-coefficients were taken from the last step in the regression analysis.
Step 2. Results of the regression analyses with reactive aggression as dependent variable provided partial confirmation for the hypothesis that deficits in the first four social cognitive skills are related to reactive aggression (see Table 6). It could not confirm the hypothesis that lower verbal intelligence was related to reactive aggression, although performance
intelligence was a significant predictor. All final regression models that predicted reactive aggression from the first four social cognitive skills and WISC IQ-scores scores were significant (FSIQ: F(5,14) = 7.35, p < .01; VIQ: F(5,14) = 4.89, p < .05; PIQ: F (5,14) = 10.12, p < .01).
Focusing on FSIQ, the only model that was significant was the final model in which the social cognitive skills were included and accounted for 80.3% of the total variance of reactive aggression. In this final model, only the first social cognitive skill ‘Identifying’ was a significant predictor of reactive aggression (β = -1.35, t = -4.23, p < .01). The model
Table 6
Hierarchical Regression Models Predicting Reactive Aggression Reactive aggression
FSIQ in step 2 VIQ in step 2 PIQ in step 2
Predictor R2 ΔR2 β R2 ΔR2 β R2 ΔR2 β
Step 1. IQ .16 .16 .18 .18 .06 .06
FSIQ / VIQ / PIQ .67 -.01 .76*
Step 2. SCST .80 .64** .73 .56* .85 .77**
1. Identifying -1.35** -.82** -1.31**
2. Discriminating .40 -.05 .54
3. Differentiating -.17 .04 -.10
4. Comparing -.02 -.01 -.08
Note. All β -coefficients were taken from the last step in the regression analysis. *p < .05, **p < .01.
controlling for VIQ was also significant but again only after the social cognitive skills were entered and accounted for 73.1% of the variance of reactive aggression. Again, only the first social cognitive skill ‘Identifying’ was a significant predictor (β = -.83, t = -3.61, p < .01). Lastly, the model that controlled for PIQ was also significant, but only after the social
cognitive skills were included and accounted for 84.9% of the variance. Significant predictors of reactive aggression were PIQ (β = .76, t = 2.65, p < .05) and the first social cognitive skill ‘Identifying’ (β = -1.31, t = -5.89, p < .001).
In sum, the results of the regression analyses could not confirm the hypothesis that greater verbal intelligence and deficits in higher social cognitive skills are predictors of proactive aggression. There were no significant relationships for proactive aggression with any of the variables of interest. Partial confirmation existed regarding the second research question, whether reactive aggression could be predicted by the WISC-IIINL IQ scores and the
first four social cognitive skills. Significant predictors of reactive aggression were performance intelligence and the first social cognitive skill ‘Identifying’.
Mediation
For mediation analysis to be performed, only significant relations between social cognitive skills and proactive and reactive aggression were included. There were no
significant relations between the social cognitive skills and proactive aggression. Since only the first social cognitive skill ‘Identifying’ was a significant predictor of reactive aggression, mediation analysis was performed with this social cognitive skill to test the likelihood that it acted as a mediator between intelligence and reactive aggression.
The mediation models for FSIQ, VIQ, and PIQ are shown together in Figure 2. For both FSIQ and PIQ, the relationship between general cognitive ability and reactive
aggression and the relationship between performance intelligence and reactive aggression were fully mediated by the social cognitive skill ‘Identifying’ (FSIQ: Sobel test -2.60, p < .01, R2 = .762; PIQ: Sobel test -2.03, p < .05, R2 = .778). As Figure 2 illustrates, the standardized beta coefficient between FSIQ and reactive aggression and PIQ and reactive aggression changed (negative to positive) when controlling for the social cognitive skill ‘Identifying’. The other conditions of mediation were also met: FSIQ and PIQ were both significant predictors of the social cognitive skill ‘Identifying’, and the social cognitive skill ‘Identifying’ was a significant predictor of reactive aggression while controlling for FSIQ and PIQ.
For VIQ, the relationship between verbal intelligence and reactive aggression was not fully mediated by the social cognitive skill ‘Identifying’, but it failed to reach significance only marginally (Sobel test -1.93, p = .053, R2 = .728). The standardized beta coefficient
FSIQ: .24 (-.40) VIQ: .02 (-.42)
PIQ: .26 (-.25)
Intelligence Reactive aggression
FSIQ: .64* VIQ: .51† PIQ: .52* FSIQ: -1.01** VIQ: -.86** PIQ: -.99** Emotion identification
Figure 2. Mediation model in which the ability to identify emotions (SCST skill
‘Identifying’) is a mediator in the relationship of intelligence (FSIQ, VIQ, and PIQ) and reactive aggression.
Note: coefficients are standardized beta coefficients; * p < .05, ** p < .001, † p = .052. Between brackets represents standardized beta coefficients of the direct relation between intelligence and reactive aggression (without controlling for the SCST skill ‘Identifying’).
between VIQ and reactive aggression decreased and change to positive when controlling for the social cognitive skill (see Figure 2). Regarding the other conditions, VIQ was not a
significant predictor of the social cognitive skill (but only marginally p = .052) and the social cognitive skill ‘Identifying’ was a significant predictor of reactive aggression while
controlling for FSIQ.
In sum, the mediation models that included FSIQ and PIQ were significant, whereas the mediation model that included VIQ was not significant. However, a clear trend was present; in that social cognitive skill “Identifying” fully mediates the relation between intelligence and reactive aggression. The mediation effect is expressed by a positive relation between intelligence and social cognitive skills (i.e. low intelligence are related to low social
cognitive skills) and a negative relation between social cognitive skills and reactive
aggression (i.e. low social cognitive skills are related to more ratings of reactive aggression). Discussion
The aim of the present study was to investigate the relation between intelligence, social cognitive skills, and types of aggression. The results indicate that social cognitive skills and intelligence are related to reactive aggression in school-aged boys. Furthermore, proactive aggression is unrelated to intelligence and social cognitive skills in the current sample. Reactive aggression: Mediation through social cognitive skills
The main finding of the present study is a mediation effect of social cognitive skills in the relation between intelligence and reactive aggression. More specifically the ability to identify emotions fully mediates this relation. To the best of our knowledge, this is the first study to find a mediating role of social cognitive skills in the relation between intelligence and reactive aggression. This finding indicates that a lower intelligence level is related to the ability to identify emotions (the mediator), which is consistent with results of other studies. For example, children with mild to borderline intellectual disabilities (MBID; American Association on Intellectual and Developmental Disabilities [AAIDD], 2013) are known to be less accurate in identification of ambiguous social cues, even after controlling for
externalizing behavior problems (Gomez et al., 1996). Moreover, children with MBID have more difficulties with encoding information, emotion recognition, and interpretation of information in general, than typically developing peers (Nieuwenhuijzen, Vriens,
Scheepmaker, Smit, & Porton, 2011). Comparison with children with MBID is justified since their range of IQ-scores is below or around 70 and the mean FSIQ-score in our sample was 69.07.
Furthermore, this study demonstrates that the ability to identify emotions is negatively related to reactive aggression. Identification of emotions corresponds with the first step of the SIP-model (Crick & Dodge, 1994; Dodge, 1986), the encoding of social cues. Several studies (Crick & Dodge, 1996; Dodge & Coie, 1987) have indicated that reactive-aggressive children exhibit more problems in the first two steps of the model. According to the social cognitive theories – such as the SIP-model – enactment of aggression mostly depends on the manner in which individuals processes social information. Deficits in the first step of the SIP-model – encoding of cues – can influence thoughts, decisions, and enactment of behavior made in the subsequent SIP-model steps (Crick & Dodge, 1994; Dodge, 1986). When children are unable to identify the emotion of another person, they are uncertain of the other’s intentions which could lead to biased thoughts, inadequate decisions, and eventually result in inadequate social behavior. Since reactive aggression is characterized as an angry, “hot-blooded” response, it can be argued that having difficulty with the identification of emotions could lead to an aggressive thoughts, response decisions, and enactment of aggressive “hot-blooded” behavior.
On the other hand, the other hypothesized social cognitive skills (discriminating, differentiating and comparing) were unrelated to reactive aggression in the current sample even though the unrelated social cognitive skill ‘Discriminating’ acts on the same social cognitive level and belongs to the same SIP-model step as the related social cognitive skill ‘Identifying’. The SCST that assessed the social cognitive skills in this study can be
categorized as a theory-of-mind task in which the subsequent social cognitive skills act upon more theory-of-mind skills (Van Manen et al., 2001). However, the ability to identify
emotions might not be part of a child’s set of theory-of-mind skills, but rather acts as a precursor for theory-of-mind (Van Rijn, 2011). Evidence for this line of reasoning comes
from studies indicating that children and adolescents with antisocial traits show impaired emotion recognition skills, although their theory-of-mind skills remains intact (Dolan and Fullam, 2004; Jones, Forster, & Skuse, 2007; Richell et al., 2003). Moreover, different regions in the brain appear to be associated with the mere ability to identify emotions and theory-of-mind skills. Whalen et al. (2001) have shown that difficulties with emotion recognition are related to the subcortical structure amygdala, whereas higher cortical
structures such as the medial prefrontal cortex are more involved during theory-of-mind tasks (Ochsner, 2008). The amygdala is an evolutionarily old structure that is biologically
programmed to act fast to possible threatening situations (Ochsner, 2008). Because the ability to identify emotions appears to rely more on the amygdala than on other brain structures, it is possible that deficits in the amygdala could play a role in the impulsive, “hot-headed”
behavior that is mostly seen in reactively-aggressive children. When these children are confronted with a threatening situation – because they have misinterpreted the other’s emotion and thereby possibly their intentions – the impulse to react in an aggressive manner goes directly through the amygdala and surpasses the circuit that will involve the higher cortical structures. Therefore their theory-of-mind skills will not play a role in their
enactment of behavior. In other words, the difficulty with identifying emotions (the first SIP-model step) does not follow through in the other SIP-SIP-model steps, because the neurological pathway might act faster than the cognitive pathway. These neurological and cognitive observations add credibility to the idea that the ability to identify emotions and theory-of-mind are different social cognitive skills, and that the difficulty with emotion identification and not with theory-of-mind might be a specific underlying mechanism of reactive
Considering that intelligence is related to adequate social cognitive skills and that poor emotion recognition in turn is related to more reactive aggression, the mediating relationship adds credibility to the idea that adequate social functioning and social cognitive skills rely in part on intellectual ability. The present results show that intelligence is related to the ability to identify emotions, one of many social cognitive skills. The ability to identify emotions can be acquired through training, given the fact that general social skills trainings are proven successful at improving social cognitive skills of aggressive children (for review see Gresham Cook, Crews, & Kern, 2004; Van Manen et al., 2004). Because of their
intelligence, children with average and above-average intelligence could be able to learn the identification of emotions more efficiently (the basic social cognitive skill). Their greater ability to identify emotions could in turn lead to more social adequate functioning
diminishing the chance that they would exhibit reactive aggression. They can use their basic social cognitive skills to resolve troublesome situations. Looking at children who exhibit reactive aggression, this mediating relationship also works in a more negative direction. Children with lower intelligence are less able to depend on their cognitive abilities – because of impaired learning, reasoning, and problem solving skills – (AAIDD, 2013), making it more difficult for them to acquire the ability to identify emotions. Subsequently, their ability to identify emotions may be insufficient in problematic social situations which could lead to the biased perception that the other means harm (also known as the hostile attribution of intent; Orobio de Castro, 2002) and eventually results in highly aroused and impulsive behavior, such as the ‘hot headed’ unreasoned aggressive behavior seen in reactively
aggressive children. The current findings are able to demonstrate how intelligence and social cognitive skills explain reactive-aggressive behavior in school-aged boys.
The present study was unable to confirm any relations between proactive aggression, social cognitive skills, and intelligence. Furthermore, the finding of Arsenio et al. (2009) that proactive aggression was related to higher verbal intelligence was not replicated in the current sample. Moreover, the findings of the present study did not support the hypothesis that deficits with greater social cognitive skills, which corresponds with the fourth step of response decision in the SIP-model, was related to proactive aggression. It might be argued that the use of the SCST as a measure of social cognitive skills could be a possible
explanation for the lack of relation between social cognitive skills and proactive aggression in this sample. The results of the SCST are based on assessment of theory-of-mind skills in hypothetical situations (Van Manen et al., 2001) and not on personal response decision made in real-life situations. The SCST may have not been sensitive enough to assess the actual process of response decision that would occur when children are faced with a similar situation in everyday life. To date only the studies by Crick and Dodge themselves (1996; Dodge et al., 1997) have provided support for the hypothesis that deficits in the step of response decision are related to proactive aggression. The results of the present study add up to the inconclusive evidence of specificity of SIP-processing in proactively-aggressive children.
Limitations and strengths of the present study
A limitation of this study that is of substantial importance is the small sample size. This small sample could have been responsible for reduced power of the statistical analyses to find effects in the current sample. On the other hand, the main results of this study are significant and clear despite the small sample size. Another limitation reflects the possibility of limited generalization of the present findings to other samples. The present sample
abilities who were included in this study because of their deviant aggressive and rule-breaking behavior. On the other hand, the high specificity of this sample is also an elegant characteristic of this study. This study was able to provide insight into the underlying social information processing mechanisms of a specific group of children at risk for developing more serious behavior problems and delinquency. On the other hand, future studies should replicate the current results in larger samples. Fritz and Mackinnon (2007) have argued that a sample size of 405 is required to obtain medium effect sizes with adequate power necessary for mediation analysis. Moreover, gender differences were not examined in this study, since only two girls were available for comparison. To date, results of studies that have examined gender, social cognitive skills, and subtypes of aggression, are mixed in that some authors have found gender differences (Baker, Raine, Liu, & Jacobson, 2008; Little, Henrich, Jones, & Hawley, 2003) whereas other did not (Connor, Steingard, Anderson, & Melloni Jr., 2003). Another limitation of this study was that the results indicated a mediating role only for the ability to identify emotions in the relation between intelligence and reactive aggression. The current study failed to provide information regarding the specific emotions, such as happy or anger, that were difficult to identify and whether these specific emotions are related to either proactive or reactive aggression. In order to overcome this limitation, future studies should focus on further differentiating the emotions and cognitions specific to the types of
aggression, using extensive facial emotion recognition tasks.
A number of aspects of this study are worth mentioning. An important strength of this study is that it is the first to have used an instrument that assessed a wide and distinctive range of social cognitive skills simultaneously with intelligence to predict specific types of aggression. The SCST consists of contextually rich stories to reflect troublesome social situations. A meta-analysis byOrobio de Castro et al. (2002) have shown that studies that
used stories that are read to or by children reported larger effect sizes of SIP on aggression, than studies that have used videos. Therefore, the use of SCST increased validity and power of this study. On the other hand, as was argued previously, the SCST is an assessment of theory-of-mind skills in hypothetical situations and not on personal response decision made in real-life situations. It might have not been sensitive enough to assess the actual process of response decision - especially in proactively aggressive children - which could have led to diminished power in this specific group. Another strength of this study is that it did not include intelligence as a covariate as previous studies but as a variable of interest.
Intelligence is a known factor related to childhood aggression and delinquency but is often not included in studies (Maguin & Loeber, 1996; Patterson et al., 1989).
Implications, future prospects and conclusions
The importance of this study and clarification on the role of constructs involved can be traced back to the differential and far-reaching trajectories of proactive and reactive aggression. Reactive aggression has been uniquely related to anxiety and substance abuse in adulthood, whereas proactive aggression has been uniquely related to adult psychopathic features and antisocial behavior in adulthood (Fite, et al., 2010; Pulkinnen, 1996). Moreover, Guerra, Huesmann, and Spindler (2003) suggested that given the stability of social cognitive skills during the later elementary years, once-formed social cognitive skills are less sensitive to new information from the environment. Based on this suggestion, examining and training social cognitive skills in young aggressive children is useful, also given the fact that studies have consistently shown that aggression increases with age during childhood and adolescence (Loeber & Hay, 1997).
Disentangling the role of social cognitive skills and intelligence and its relation with the types of aggression is also useful in developing and implementing interventions to
prevent psychopathology and delinquency in adolescence and adulthood. Because distinctive social information processing patterns are present for reactive and proactive aggressive children, interventions should be directed at the core deficits of these subtypes of aggression to improve effectiveness (Brown & Parsons, 1998). The results of this study indicate that difficulties with identifying emotions are related to reactive aggression, and not to proactive aggression, supporting the idea that distinctive social information processing patterns exist for reactive and proactive aggression. Others (Hudley & Graham, 1993; McAdams, 2002; Vitaro et al., 2002) have suggested that effectiveness of interventions is directly related to subtype-specific interventions and therefore interventions should focus on deficits of subtypes of aggression. Because the misinterpretation of social cues, specifically emotions, seem to form at least part of the core of reactive aggression, subtype-specific interventions should include explicit training of recognition of emotions and verbalizing these emotions. Furthermore, the results of the present study show that social cognitive skills and social functioning rely in part on intellectual abilities which should be considered in implementing interventions as well. Interventions should take into account difficulties that are often accompanied with intellectual disabilities, such as a limited vocabulary, difficulty with abstract reasoning, and difficulty with fast information processing (De Beer, 2011), in order to increase the interventions’ effectiveness (Van Nieuwenhuijzen, Orobio de Castro, & Matthys, 2006).
Future studies should focus on further differentiating the emotions and cognitions specific to the types of aggression, using extensive facial emotion recognition tasks.
Moreover, evidence regarding the specificity of deficits in SIP in reactively and proactively aggressive children is to date inconclusive and studies should aim at using different
situation) to provide clarity on this topic. Since this study was unable to link social cognitive skills to proactive aggression, other studies could focus on provoking proactive aggressive behavior in social interactions and assessing the cognitions of the child during and after these interactions, whilst keeping the study between ethical limits. Lastly, although the present sample was small in size, significant relations between social cognitive skills, intelligence, and aggression were found. Other studies should aim at replicating these results in larger and more heterogeneous samples (greater age range, boys vs. girls).
In sum, the finding that the relation between social cognitive skills and reactive aggression depends at least partly on intelligence whereas such a relation was not found for proactive aggression, suggests differential underlying mechanisms for both types of
aggression. Future research should focus on further differentiating the emotions that reactive-aggressive children find difficult to identify. Interventions should be directed on explicit training of emotion recognition in reactive-aggressive children without losing sight of the important role of intelligence. By training the basic social cognitive skills that in turn can lead to more adequate social functioning, it could be possible to prevent children at risk from developing more serious behavior problems. Diminishing the presence of these problems could lead to less delinquency creating a safer environment for both children and others involved and reducing significant costs for society.
References
Achenbach, T. M. (1978). The child behavior profile: I. Boys aged 6–11. Journal of
Consulting and Clinical Psychology, 46, 478–488. doi: 10.1037/0022-006X.46.3.478 Achenbach, T. M. (1991). Manual for the Teacher's Report Form and 1991 Profile.
Burlington: University of Vermont Department of Psychiatry.
Akhtar, N., & Bradley, E. J. (1991). Social information processing deficits of aggressive children: Present findings and implications for social skills training. Clinical Psychology Review, 11, 621-644. doi:10.1016/0272-7358(91)90007-H
American Association on Intellectual and Developmental Disabilities. (2013). FAQ on Intellectual Disability. Retrieved from http://www.aaidd.org/content_104.cfm
American Psychological Association. (2002). Ethical principles of psychologists and code of conduct. American Psychologist, 57, 1060-1073. doi:10.1037/003-066X.57.12.1060 Baker, L. A., Raine, A., Liu, J., & Jacobson, K. C. (2008). Differential genetic and
environmental influences on reactive and proactive aggression in children. Journal of Abnormal Child Psychology, 36(8), 1265-1278. doi: 10.1007/s10802-008-9249-1 Baron, R. M., & Kenny, D. A. (1986). The moderator–mediator variable distinction in social
psychological research: Conceptual, strategic, and statistical considerations. Journal of Personality & Social Psychology, 51, 1173-1182.
Bartels, M., Rietveld, M. J., Baal, G. C. van, & Boomsma, D. I. (2002). Genetic and
environmental influences on the development of intelligence. Behavior Genetics, 32, 237–249. doi: 10.1023/A:1019772628912
Beer, Y. de (2011). De kleine gids: Mensen met een licht verstandelijke beperking [The little guide: People with mild intellectual disabilities]. Deventer, the Netherlands: Kluwer.
Beijsterveldt, C. E. M. van, Bartels, M., Hudziak, J. J., & Boomsma, D. I. (2003). Causes of stability of aggression from early childhood to adolescence: A longitudinal genetic analysis in Dutch twins. Behavior Genetics, 33(5), 591-605. doi:
10.1023/A:1025735002864
Broidy, L. M., Nagin, D. S., Tremblay, R. E., Bates, J. E., Brame, B., Dodge, K. A., … & Petit, G. S. (2003). Developmental trajectories of childhood disruptive behaviors and adolescent delinquency: a six-site, cross-national study, Developmental Psychology, 39(2), 222-245. doi: 10.1037/0012-1649.39.2.222
Campbell, S. B. (2002). Behavior problems in preschool children: Clinical and developmental issues. New York, NY: The Guilford Press.
Campbell, S.B., Shaw, D.S., & Gilliom, M. (2000). Early externalizing behavior problems: Toddlers and preschoolers at risk for later maladjustment. Development and
Psychopathology, 12(3), 467–488.
Carpendale, J. I. M., & Chandler, M. J. (1996). On the distinction between false belief understanding and subscribing to an interpretive theory of mind. Child Development, 67(4), 1686-1706. doi: 10.1111/1467-8624.ep9704041043
Cohen, J. W. (1988). Statistical power analysis for the behavioral sciences (2nd ed.). Hillsdale, NJ: Lawrence Erlbaum.
Coie, J. D. (2004). The impact of negative social experience on the development of antisocial behavior. In: J. Kupersmidt & K. A. Dodge (Eds.) Peer relations in childhood: From development to intervention to public policy. Washington, DC: American
Connor, D. F., Steingard, R. J., Anderson, J. J., & Melloni, R. H. Jr. (2003). Gender differences in reactive and proactive aggression. Child Psychiatry and Human Development, 33, 279–294. doi: 10.1023/A:1023084112561
Cowen, E. L., Pederson, A., Babigan, M., Izzo, L. D., & Trost, M. A. (1973). Long-term follow-up of early detected vulnerable children. Journal of Consulting and Clinical Psychology, 41(3), 438-446.
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. doi: 10.1037/0033-2909.115.1.74
Crick, N. R., & Dodge, K. A. (1996) Social information-processing mechanisms in reactive and proactive aggression. Child Development, 67, 993–1002. doi: 10.1111/j.1467-8624.1996.tb01778.x
Dodge, K. A. (1986). A social information processing model of social competence in
children. In: M. Perlmutter (Ed.), Minnesota symposia on child psychology. Hillsdale, NJ: Lawrence Erlbaum.
Dodge, K. A. (1991b). The structure and function of reactive and proactive aggression. In D. Pepler & K. H. Rubin (Eds.), The development and treatment of childhood aggression (pp. 201–218). Hillsdale, NJ: Lawrence Erlbaum Associates, Inc.
Dodge, K. A., & Coie, J. D. (1987). Social-information processing factors in reactive and proactive aggression in children’s peer groups. Journal of Personality and Social Psychology, 53, 1146–1158. doi: 10.1037/0022-3514.53.6.1146
Dodge, K. A., Lochman, J. E., Harnish, J. D., Bates, J. E., & Petit, G. S. (1997). Reactive and proactive aggression in school children and psychiatrically impaired chronically
