Вы находитесь на странице: 1из 13

Journal of Adolescence 34 (2011) 5971

Contents lists available at ScienceDirect

Journal of Adolescence
journal homepage: www.elsevier.com/locate/jado

Is low empathy related to bullying after controlling for individual and social background variables?
Darrick Jolliffe a, *, David P. Farrington b
a b

Department of Criminology, University of Leicester, 154 Upper New Walk, Leicester, England LE1 7QA, UK Institute of Criminology, Sidgwick Site, Cambridge University, Cambridge, England CB3 9DA

a b s t r a c t
Keywords: School bullying Empathy Violence Gender

This paper examines the relationship between low empathy and bullying while also controlling for the impact of a number of other individual and social background variables linked with bullying. This included the relationship to the prevalence of bullying, but also to the frequency and type of bullying. Questionnaires were completed by 720 adolescents (344 females, 376 males) aged 1317 in three secondary schools in England. The results suggested that low affective empathy was independently related to bullying by males, but not females. There was no evidence that low cognitive empathy was independently related to bullying, but high impulsivity was related to all forms of male bullying and to female bullying. The implications of the ndings for research and practice are discussed. 2010 The Association for Professionals in Services for Adolescents. Published by Elsevier Ltd. All rights reserved.

Introduction Historically dened as a cognitive ability (i.e., the ability to understand the emotions of another) or an affective trait (i.e., the ability to experience the emotions of another person), most researchers currently acknowledge that empathy involves both affective and cognitive components (Davis, 1983; Hogan, 1969; Strayer, 1993). There are a number of theories regarding the possible mechanisms by which empathy inuences behaviour (e.g. Feshbach,1975; Marshall, Hudson, Jones, & Fernandez, 1995; Pithers, 1999), however, many of these theories have as their fundamental basis the assumption that a greater amount of empathy inhibits antisocial behaviour and increases the likelihood of prosocial behaviour. This is because the ability to experience or understand another persons negative emotions (e.g., fear, distress) should motivate a person with high empathy to act to alleviate this suffering. This action might be undertaken for altruistic reasons (e.g., to assist the other person) or for selsh reasons (e.g., to reduce vicarious distress). Also, an individual with high empathy would derive rewards in the form of vicarious pleasure from any subsequent positive emotions (e.g., relief, happiness) that resulted from the prosocial behaviour. Alternatively, low empathy should increase the likelihood of antisocial behaviour. This is because an individual who lacks empathy is essentially free to act without an appreciation of the impact of their behaviour on the emotions of others, and without experiencing aversive vicarious emotions. Furthermore, a lack of empathy would prevent an individual from beneting from another persons positive emotions because of the individuals limited ability to understand or experience these. Given the intuitive strength of this theory it is not surprising that many researchers have suggested that those who bully others at school may have low empathy (e.g., Endresen & Olweus, 2002; Jolliffe & Farrington, 2006a).

* Corresponding author. Tel.: 44 11 44 116 2525729; fax: 44 1144 116 2525788. E-mail address: dj39@le.ac.uk (D. Jolliffe). 0140-1971/$ see front matter 2010 The Association for Professionals in Services for Adolescents. Published by Elsevier Ltd. All rights reserved. doi:10.1016/j.adolescence.2010.02.001

60

D. Jolliffe, D.P. Farrington / Journal of Adolescence 34 (2011) 5971

There is considerable conceptual overlap between aggressive behaviour and school bullying (e.g., Farrington, 1993). For example, both bullying and aggression refer to a wide set of behaviours such as physical, verbal or psychological attack and both are intended to cause fear, distress and harm to the victim. Furthermore, research has demonstrated that those who take part in school bullying are more likely to commit aggressive acts outside of school (Baldry & Farrington, 2000). It is however, important to differentiate bullying from aggressive behaviour. School bullying can be dened as a systematic abuse of power (Smith & Sharp, 1994), and this denition highlights key differences between bullying and aggression. First, is the reference to the systematic nature of bullying. Bullying, unlike aggression (which could involve a one-off incident), involves repeated incidents between the same dyad of bully and victim over prolonged periods of time (e.g., Smith & Thompson, 1991). Also, linked to this concept of continuity is the proactive nature of bullying, with bullies tending to seek out opportunities to bully others (Camodeca, Goossens, Terwogt, & Schuengel, 2002). The denition provided by Smith and Sharp (1994) also makes reference to the imbalance of power between the bully and the victim, with the more physically or psychologically powerful child oppressing the less powerful one (Farrington, 1993). Empathy and bullying The assertion that those who bully others have low empathy is common among researchers investigating the causes of bullying (e.g., Bernstein & Watson,1997; Merrell, Gueldner, Ross, & Isava, 2008; Olweus,1993). For example, Olweus (1993: p. 34) states: [Bullies] have little empathy with victims of bullying, and Smith and Thompson (1991: p. 3) reported that Children who bully others may be less empathic to the feelings of others, such as potential victims. Considering the pervasiveness of the view that bullies have low empathy, there have been relatively few studies that have empirically assessed the level of empathy among bullies compared to other children. Endresen and Olweus (2002) administered a 12-item questionnaire measure of affective empathy and a measure of bullying to 2286 Norwegian school children (1093 boys and 1193 girls) aged 1316. The results identied a signicant negative correlation of .15 between bullying behaviour and affective empathy for both males and females. Further analyses suggested that the negative relationship between empathy and bullying disappeared after controlling for a measure of a positive attitude to bullying. There was, however, some indication that the measure of a positive attitude to bullying may have overlapped somewhat with the measure of bullying (they may have measured the same underlying construct) complicating the interpretation of these results. In another investigation of the relationship between empathy and bullying Warden and Mackinnon (2003) used a 22-item empathy questionnaire (Bryant, 1982) to assess the affective empathy of 21 prosocial children, 23 bullies and 14 victims from a sample of 131 910 year old UK school children. The results suggested that the prosocial children had signicantly higher levels of affective empathy compared to the bullies. Subsequent analyses suggested that gender moderated this relationship, with females having higher empathy and being less likely to bully and males having lower empathy and being more likely to bully. Espelage, Mebane, and Adams (2004) examined levels of cognitive and affective empathy amongst a group of 263 (128 males, 135 females) 1114 year old students from the US. Empathy was assessed using the Perspective Taking scale (as a measure of cognitive empathy) and the Empathic Concern scale (as a measure of affective empathy) of the Interpersonal Reactivity Index (Davis, 1983).1 Bullying was assessed using the nine-item University of Illinois Bully Scale (UIBS; Espelage & Holt, 2001), which is a self-report measure incorporating various types of bullying and their frequency. Signicant negative correlations were found between the measure of bullying and the measure of affective empathy (r .31) and the measure of cognitive empathy (r .28) for males. There were also negative correlations for females between bullying and the measure of affective empathy (r .15) and cognitive empathy (r .25), but only the correlation with cognitive empathy was signicant. In another study Jolliffe and Farrington (2006a) used the Basic Empathy Scale (Jolliffe & Farrington, 2006b) to assess levels of both affective and cognitive empathy among 720 UK students aged about 15. In addition, a self-report measure of bullying was administered. The results suggested that females who had bullied in the last school year, and male and female frequent bullies (once a week or more often), had signicantly lower affective empathy, but did not signicantly differ from others in cognitive empathy. The authors suggested that it was the inability to experience the emotions of others that characterised bullies, not the inability to understand the emotions of their victims. Interestingly, this research also found that males who bullied violently and females who bullied indirectly had low empathy. In summary, there does appear to be some evidence to suggest that bullies have low empathy, especially low affective empathy. As yet however, there is no empirical evidence to suggest that low empathy and bullying are related independently of the many other factors that have been associated with bullying. It is extremely difcult to establish causal inuences in non-experimental research. However, it is widely accepted that the main criteria for establishing that X causes Y is: (1) X is correlated with Y, (2) X can change or be changed within individuals, (3) X precedes Y, and (4) X predicts Y after controlling for confounding variables (Kraemer, Lowe, & Kupfer, 2005; Murray, Farrington, & Eisner, 2009). Therefore, in trying to draw

1 The use of the Perspective Taking scale and the Empathic Concern scale as measures of empathy has been criticised (Jolliffe & Farrington, 2006b). These scales do not adhere to the current conceptualisation of empathy (e.g. the understanding of anothers emotions or experience of anothers emotions). Questions have also been raised about the transparency of the items and the implications of this for desirable responding.

D. Jolliffe, D.P. Farrington / Journal of Adolescence 34 (2011) 5971

61

conclusions about whether low empathy might have a causal inuence on bullying, it is important to investigate if the relationship between low empathy and bullying holds up after controlling for other explanatory variables. The only two studies which have examined low empathy and bullying controlling for other factors (Endresen & Olweus, 2002; Warden & Mackinnon, 2003) did not nd that low empathy was independently related to bullying. The independent relationship between low empathy and antisocial behaviour was also recently called into question by the results of a systematic review and meta-analysis of the association between low empathy and criminal offending. In reviewing the results of the 35 included studies Jolliffe and Farrington (2004) concluded that low empathy was related to offending, but that this relationship appeared to be largely accounted for by the fact that both low empathy and offending were related to other factors, namely socioeconomic status and intelligence. The nature of the relationship between low empathy and bullying is of critical importance as a number of anti-bullying interventions are based on the assumption of a direct causal link between low empathy and bullying. That is, many antibullying interventions include, either explicitly or implicitly, some attempt to make children understand and appreciate the emotional effects of their behaviour on the victims of bullying (Olweus, 1993; Rigby, 1996; Whitney, Rivers, Smith, & Sharp, 1994) with the expressed purpose of increasing empathy and consequently reducing bullying. These interventions might be successful in increasing empathy, but unless low empathy causes bullying they will not have the desired impact of reducing bullying. In order to convincingly establish that low empathy causes bullying, empathy would need to be measured before any bullying has taken place and all other variables which might inuence bullying would need to be controlled. Obviously this would require a very comprehensive prospective longitudinal study. In the absence of this, however, the potential of low empathy as a cause of bullying can be assessed in cross-sectional research in which other factors that have been found to be associated with bullying are measured and controlled.

Other factors associated with bullying A number of social, family and individual factors are associated with an increased likelihood of school bullying. The social feature that has most consistently been related to school bullying is low socioeconomic status (SES). This is usually measured by the occupational prestige of the parent(s) or by eligibility for free school meals. The results of studies comparing low SES and bullying behaviour have had mixed results, with some research suggesting a link (e.g. Mellor, 1990; Veenstra et al., 2005), but others suggesting that low SES does not have an inuence on bullying. For example, Bosworth, Espelage, and Simon (1999), Duyme (1990), Farrington and Baldry (2006), Flouri and Buchanan (2003), Kumpulainen, Rasanen, and Henttonen (1999), Olweus (1978), and Unnever and Cornell (2003) all found that socioeconomic status was unrelated to bullying in school children. Research has shown a much more consistent relationship between family factors and bullying. Those who bully have much higher levels of most types of family dysfunction. Examples of these include marital disharmony (e.g. Lowenstein, 1978; Stephenson & Smith, 1989), non-intact families (Bowers, Smith, & Binney, 1992; Flouri & Buchanan, 2003), low family cohesiveness (Bowers et al., 1992; Connolly & OMoore, 2003), and poor family functioning in general (Beinart, Anderson, Lee, & Utting, 2002; Rigby, 1996). Highly related to family dysfunction, the type and quality of parenting that a child receives has also been related to bullying. For example, Olweus (1980) suggested that a lack of parental warmth and involvement with the child, permissive or lax parenting and power assertive parenting (physical punishment and violent emotional outbursts) were all related to an increased risk of bullying. Furthermore, Haynie et al. (2001) found that parental supervision (parents knowing about childs activities) was signicantly negatively related to bullying. Similar results have been identied in a large study of 1017 year olds in England (Beinart et al., 2002: p. 23). A variety of individual features have been associated with bullying. For example, those who bully others score lower on measures of executive function than those who do not (e.g., Unnever & Cornell, 2003). Executive function refers to the broad ability to manipulate abstract constructs and is associated with the prefrontal cortex of the brain (Moftt, Caspi, Rutter, & Silva, 2001). Measures of executive function are very similar to measures of intelligence (e.g. Moftt & Henry, 1989), and bullies tend to demonstrate lower scores on these measures as well. Lowenstein (1978) found that bullies scored slightly below average on a measure of intelligence and reading ability. Similar results have been identied with other measures of executive function (e.g. Coolidge, DenBoer, & Segal, 2004; Monks, Smith, & Swettenham, 2005) and school achievement (Baldry & Farrington, 2000; Olweus, 1978; OMoore & Hillery, 1991; Stephenson & Smith, 1989), which may in part be caused by their low intelligence and/or reduced executive functioning. Impulsivity, or acting without considering the consequences of ones actions, has been implicated as a factor which may inuence bullying (e.g. Olweus, 1993). For example, Bosworth et al. (1999) showed that impulsivity was signicantly correlated (r .33) with their measure of bullying, and Espelage, Bosworth, and Simon (2001) found that impulsivity independently predicted involvement in bullying four months later after controlling for previous bullying. Also, in their study of 4263 adolescents Haynie et al. (2001) discovered that bullies had signicantly lower self-control (higher impulsivity) than controls and furthermore that low self-control independently predicted involvement in bullying. Unnever and Cornell (2003) investigated the relationship between attention-decit hyperactivity disorder (ADHD; a disorder characterised by high impulsivity), low self-control and bullying in a sample of 1315 adolescents. They established that low self-control predicted bullying after controlling for gender, race, height, weight and strength. In addition, they found that ADHD increased the likelihood of bullying, but only through its effects on lowering self-control.

62

D. Jolliffe, D.P. Farrington / Journal of Adolescence 34 (2011) 5971

Inuence of gender on frequency and type of bullying It is also important to note three other elements, rarely addressed by previous research, which may have a signicant impact on the relationship between low empathy and bullying. These are the relationships between gender and the frequency and type of bullying. Research has consistently shown that females are less likely to bully others compared to males (e.g., Olweus, 1993; Whitney & Smith, 1993). Furthermore, research has shown that indirect aggression (rumour mongering, purposefully avoiding someone) might be more typical of females than males (e.g. Bjorkvist, Lagerspetz, & Kaukiainen, 1992; Lagerspetz, Bjorkvist, & Peltonen, 1988; Whitney & Smith, 1993). However, these ndings have not been adequately integrated with knowledge about the background features of bullies. For example, few studies have examined whether low SES, family dysfunction and impulsivity are characteristics of female as well as male bullies or whether female bullies possess different social, family or individual features from their male counterparts. Research tends to collectively term both males and females bullies and subsequently analyze them as a unitary group. Exceptions to this include Roland (1989) who found that female bullies, unlike their male counterparts, were not below average on intelligence, but were in fact higher than average, and Rigby (1994) who found family dysfunction in the background of female as well as male bullies. How individual and background features inuence the frequency of involvement in bullying has also not been well addressed. This is likely because of the different denitions of bullying used in research. While some have argued that only those who take part in bullying once a week or more often should be considered bullies (e.g., Olweus, 1993), others have considered bullies to be anyone who reports bullying three or more times in the past year (e.g., Haynie et al., 2001). Others still have considered that bullying behaviour exists along a continuum (e.g., Bosworth et al., 1999). Just as it is important to identify the social, family and individual background features of those who bully (the relationship with prevalence), it is also important to identify those features that are associated with the amount of bullying (the relationship with frequency). For example, family dysfunction may increase the likelihood of becoming involved in bullying, but high impulsivity may increase both the likelihood of involvement in bullying and the frequency. It is important to investigate these distinct issues. Obviously there are many ways in which bullies can harass, humiliate and torment their classmates. While research in the past tended to operationalise bullying as physical (e.g., hitting or kicking; Olweus,1978), contemporary research usually includes this as well as forms of verbal bullying (e.g. name calling) and indirect bullying (e.g. purposefully avoiding someone). Although it is useful to draw distinctions between the different types of bullying, when examining the background features of bullies these types are often treated as a single behaviour. It seems unlikely that the factors that increase the likelihood of bullying another child in a violent manner would be exactly the same as those that increase the likelihood of calling another child names. This paper will investigate the relationship between individual and social background variables and the different facets of bullying (prevalence, frequency and type) for males and females separately. To the degree that these variables differentiate bullies, their independent relationships with bullying will be tested. The key issue is whether low cognitive or affective empathy is related to bullying independently of other variables that have a relationship with bullying. Research questions 1. What relationships are evident between the individual and social background factors (including empathy) and the selfreported prevalence and frequency of bullying for males and females? 2. Is cognitive and/or affective empathy independently associated with either the prevalence or frequency of bullying among males and females when other individual and social background factors are controlled? 3. What relationships are evident between the individual and social background factors (including empathy) and the different types of bullying for males and females? 4. Is cognitive and/or affective empathy independently associated with any of the different types of bullying among males and females when other individual and social background factors are controlled? Method Participants Data for this study were obtained from 720 adolescents (376 males, 344 females) in Year 10 (mean age 14.8, SD .41, range 1317) from 3 comprehensive schools in the UK. Originally seven schools were invited to take part, but four declined. The schools that took part would be considered suburban as they were located in a Borough adjacent to London. Over 90% of the sample was Caucasian, while 4.3% were Asian and 2.6% were Afro-Caribbean. There were no signicant differences in age, racial composition or socioeconomic status between students at the three schools. Measures Empathy Empathy was measured using the Basic Empathy Scale (BES; Jolliffe & Farrington, 2006b). This 20-item scale assesses both cognitive (9 items) and affective empathy (11 items) and was designed to measure the degree to which a person understands

D. Jolliffe, D.P. Farrington / Journal of Adolescence 34 (2011) 5971

63

and shares the emotions of another (Cohen & Strayer, 1996). An example of a cognitive item would be It is hard for me to understand when my friends are sad, and an example of an affective item would be I usually feel calm when other people are scared. The affective and cognitive items were separately scored on a ve-point Likert scale. A total empathy score was also produced. Tests of the validity of the BES have been carried out in French (DAmbrosio, Olivier, Didon, & Besche, 2009) and Italian (Albeiro, Matricardi, Speltri, & Toso, 2009) samples of youths. The reliability of the overall BES for this sample was alpha .87 (.85 males, .83 females), with the reliability of the cognitive scale alpha .79 (.79 males, .78 females) and the reliability of the affective scale alpha .85 (.79 males, .74 females). Socioeconomic status (SES) SES was assessed by asking the children if their parents were employed and also their type and place of employment. A value from 1 (e.g. teachers and other professionals) to 8 (long-term unemployed) was assigned according to the system developed by the Ofce of National Statistics for the 2001 Census (Walker et al., 2002). In order to generate a single measure of SES for each child, the higher value of either maternal or paternal SES was used. This also compensated for the difculty presented by those children with only a mother or a father at home. The mean score on this measure was 3.6 (SD 1.7). Non-intact family Respondents were asked to report whether they lived with both of their biological parents. Over 42% of the sample reported living in non-intact families. Parental supervision Four questions were used to assess the typical parenting that the child received. These were derived from a large scale longitudinal study of UK adolescents (Wikstrom & Butterworth, 2006). For example, When you are out in the evenings by yourself or with friends do your parents (or step-parents) tell you what time to return home?. These 4 items were scored on a 5-point Likert scale and summed to produce a measure of poor parental supervision. Cronbachs alpha for this sample was .72 (males .71, females .74). In line with expectation (e.g., Moftt et al., 2001), females reported signicantly better supervision than males [mean for females 8.1 (SD 2.3) compared to mean for males 8.8 (SD 2.3), t 3.7, p < .0001]. Verbal uency The Thurstone Verbal Fluency Task was included as a measure of executive function. This task requires subjects to name as many words in a given category as possible in 1 min. This measure was one of two verbal components included in Thurstones (1938) conception of intelligence, and verbal uency continues to be an important component of present day intelligence theories and tests (e.g. Carroll, 1993). This test has been shown to differentiate patients with frontal lobe damage (e.g. Milner, 1964). Furthermore, both positron emission tomography (PET) (Frith, Friston, Liddle, & Fracknowiak, 1991) and functional magnetic resonance imaging (fMRI) scans (Phelps, Hyder, Blamire, & Shulman, 1997) indicate that the left frontal cortex was activated during verbal uency tests. The mean score on this test was 14.4 (SD 4.1) for the entire sample with scores ranging from 1 to 28. In line with previous research regarding the greater verbal ability of females (e.g., Moftt et al., 2001), females scored signicantly higher than males on this measure (15.3 compared to 13.5, t 6.0, p < .0001). Impulsivity The 12-item Urgency component of the UPPS (Urgency, [lack of] Premeditation, [lack of] Perseverance, Sensation seeking; Whiteside & Lynam, 2001) was used to assess impulsivity in this sample. High scorers on the urgency scale are likely to engage in impulsive behaviours in order to alleviate negative emotions despite the long-term harmful consequences of these actions (Whiteside & Lynam, 2001: p. 685). A typical item is: It is hard for me to resist acting on my feelings (agree). This scale had an alpha reliability of .88 for the entire sample (males .88, females .87). Bullying A self-report bullying questionnaire was used to assess the amount of bullying. The items of this measure were derived from those used by Whitney and Smith (1993) in their large scale assessment of bullying in the UK. One item assessed the prevalence of bullying (i.e., number of individuals involved), while another assessed the frequency of bullying (i.e., approximate incidence e.g. once or twice, sometimes, about once a week, several times a week) this school year, which was an interval of approximately nine months. Furthermore, the respondents were also asked to indicate the types of bullying that they had undertaken on a dichotomous scale. The possible types were name calling, racial name calling, physically hurting someone, threatening someone, spreading rumours or writing grafti about someone, avoiding speaking to someone or taking someones belongings away. The responses to these questions were not mutually exclusive. For the purposes of analysis the eight types of bullying were separated into three categories on the basis of theoretical and statistical association. The rst category included name calling and racial name calling. The second category included indirect bullying which included purposefully avoiding others and spreading rumours or writing grafti about others, all of which were signicantly related. The nal category involved physical assault, threatening and taking belongings away. Physical assault was related to threatening and signicantly related to taking belongings away, while threatening and taking belongings away were also highly related.

64

D. Jolliffe, D.P. Farrington / Journal of Adolescence 34 (2011) 5971

Procedure Anonymous self-report questionnaires were administered in classrooms by an experienced researcher. Respondents were read a set of standardized instructions about how to ll in the questionnaire. Also, in order to ensure that the students were operating with a similar understanding of bullying students were provided with a denition of bullying. This denition was exactly the same as that provided by Whitney and Smith (1993: p. 7) and introduced students to the range of behaviours considered bullying while stressing the difference between bullying and similar aggressive acts. Children were asked not to put their name or any other identifying mark on the questionnaire to ensure condentiality. It was thought that anonymity would increase the rate of reporting behaviours that were not socially acceptable by ensuring that the subjects were aware that their responses could not be identied (Kulik, Stein, & Sarbin, 1968; Wikstrom & Butterworth, 2006). A passive consent procedure was used to obtain tacit approval from the parents of those involved, whereby parents were provided with an opportunity to withhold consent for their childrens participation. In addition to the passive parental consent, active informed consent was sought from all eligible students. Children were informed that their participation was voluntary and that they could withdraw from the study at any time without penalty. Because staff at each school and within each classroom dealt differently with those whose parents did not want them to take part and those who withdrew themselves, it is difcult to establish the exact response rate. However, a total of 903 children were enrolled in Year 10 at the three schools, which would be equivalent to a response rate of 79.7%. The true rate may be higher than this as enrolment numbers tend to overestimate those actually attending each school. If items were missing from questionnaires (e.g. an item on the empathy scale), the entire questionnaire was excluded from analysis. Of the 720 questionnaires that were administered 701 were fully completed. Results Table 1 shows the prevalence and frequency of self-reported bullying by males and females in the nine month time period. More than one-quarter of males (26.9%) and one in seven females (14.8%) reported that they had bullied others in the last school year. This difference was equivalent to an odds ratio (OR) of 2.1 (C.I. 1.53.1). The OR indicates the increased odds of males being involved in bullying (Farrington & Loeber, 2000). As a rule of thumb, an odds ratio of 2.0 or greater (equivalent to a doubling of the risk or greater) indicates a strong relationship (Cohen, 1996). Table 1 shows that males were more likely than females to be involved in bullying once or twice, sometimes, once a week and several times a week. Even though the differences between males and females for once a week and several times a week were not statistically signicant this was likely the result of the small numbers in these categories. In order to accommodate for the limitation of the small numbers involved in bullying sometimes or more often, they were combined into a category of frequent bullying for all subsequent analyses. The bottom of Table 1 shows the prevalence of the different categories of bullying for males and females. Males were signicantly more likely to be involved in name calling (OR 2.1, p < .001) and violent bullying (OR 3.2, p < .0001) than females. However, there was no signicant difference between males and females in the use of indirect bullying. In light of the substantial gender differences found in the prevalence and frequency of most types of bullying, the results of males and females were treated separately for the remainder of this paper. This decision was also supported by the signicant differences in empathy between males and females which have typically been identied (Davis, 1983; Jolliffe & Farrington, 2006b). Inter-relationships between the explanatory variables for males and females Before the independent inuence of empathy on offending was assessed using regression analysis, the inter-relationships of the explanatory variables which were to be used as predictors were assessed using a correlation table. Table 2 shows the inter-correlations of the measures of empathy with the individual and social background variables.
Table 1 Prevalence and frequency of bullying for males and females. N 376 % Males Bullied in last school year Once or twice Sometimes Once a week Several times a week Name calling Indirect bullying Violent bullying Note: ***p < .0001, **p < .001, *p < .05. 26.9 16.0 6.9 2.1 1.9 20.3 9.2 19.8 N 344 % Females 14.8 11.0 2.9 0.3 0.5 10.4 6.4 7.0 2.1 (1.53.1)*** 1.5 2.5 7.5 3.2 (1.02.4)* (1.25.2)** (.959.0) (0.715.7) OR (C.I.)

2.2 (1.43.4)*** 1.5 (0.82.7) 3.3 (2.05.5)***

D. Jolliffe, D.P. Farrington / Journal of Adolescence 34 (2011) 5971 Table 2 Correlations of empathy to individual and social background variables. BES cog BES cognitive BES affective BES total U. imp. Thur. Par. sup. SES Non-int families X .43* .78* .08 .14* .09 .14* .05 BES aff .41* X .90* .11* .16* .12* .10 .10 BES total .80* .88* X .04 .18* .12* .13* .10 U. imp. .08 .01 .05 X .1 .35* .02 .12* Thur. .08 .01 .05 .02 X .02 .23* .03 Par. sup. .12* .21* .20* .18* .08 X .05 .15* SES .01 .06 .06 .10 .10 .05 X .12*

65

Non-int families .03 .03 .03 .12* .05 .08 .12* X

Note: The correlations for males are above the diagonal and the correlations for females are below the diagonal. *p < .05. U Imp urgency impulsivity, Thur. Thurstones verbal uency, Par. Sup. parental supervision, SES socioeconomic status.

For males, cognitive empathy was found to be signicantly and positively correlated with affective empathy (r .41), and signicantly negatively correlated to parental supervision (r .12). Affective empathy was also negatively related to parental supervision (r .21). For females, the correlation between the two measures of empathy (r .43), was similar in magnitude to that of males. Interestingly, females showed low but signicant correlations between all measures of empathy and verbal uency, and also between cognitive empathy and SES while for males these correlations were very small.2 Independent relationship of empathy to the prevalence and frequency of bullying in males Table 3 shows the results of the comparison of those males who did not report bullying others with those who reported bullying, and a sub-group of the bullying category, those who reported frequent bullying (operationalised as bullying sometimes or more often). For example, those who did not bully had a mean score on cognitive empathy of 32.4 compared to 31.5 for those who reported bullying. This difference was not statistically signicant and was equivalent to a standardized mean difference effect size (d) of .17. Similarly, there was no statistical difference between those who bullied compared to those who did not on affective empathy (d .14). A number of methods of interpreting the magnitude of effect sizes have been proposed (Lipsey & Wilson, 2001). A widely used convention is that proposed by Cohen (1988). An effect size of about .20 is considered small, while an effect size of around .50 is considered medium and an effect size greater than .80 is considered large. However, this convention seems too conservative. A more meaningful way of interpreting an effect size can be provided by converting the results to the differences in proportions who have low empathy between those who do or do not bully. First, the standardized mean effect size (d) is converted to a phi correlation r.3 This results in an r value of approximately half that of d, and this value of r or phi is, in turn, equal to the difference in proportions between the two groups (Farrington & Loeber, 1989). Therefore, the d value of .17 for cognitive empathy corresponds approximately to 59% of bullies having low cognitive empathy compared to 50% of nonbullies (difference in proportions .09). In contrast to the results with empathy, there were clear differences between males who reported bullying and those who did not on the other individual and social background variables. Table 3 shows that those who bullied had signicantly greater impulsivity (d .57, p < .0001), and signicantly lower levels of parental supervision (d .26, p < .05) compared to those who did not. These would be equivalent to differences in proportions of 27% (77% of bullies with high impulsivity) and 13% (63% of bullies with poor parental supervision) respectively. Male bullies did not have signicantly lower verbal uency or SES and were not more likely to come from non-intact families. In order to determine whether either affective or cognitive empathy was related to bullying by males independently of other individual and social background variables a stepwise logistic regression was undertaken. This method of regression is appropriate in exploratory research of this nature (Cramer, 2003). In this regression bullying by males was the dependent variable and the independent variables were affective empathy, cognitive empathy, impulsivity, verbal uency, parental supervision, SES and family status (intact v not intact). These variables were those which were signicantly related to bullying in a zero-order relationship, or with a standardized mean effect size of d .10 or greater.4 The results suggested that two variables were independently related to bullying by males; these were impulsivity (LRCS 22.8, p < .0001) and affective empathy (LRCS 4.3, p < .04). Much previous research has suggested that only those who bully frequently (i.e. sometimes or more often) should be considered bullies (Olweus, 1980). Therefore, those who bullied sometimes or greater were compared to those who did not bully on the measures of empathy and the other individual and social background variables. Table 3 shows that frequent bullies scored signicantly lower than those who did not bully on cognitive, affective and total empathy (d .31, p < .05,

2 Aside from the artifactual correlations of the affective and cognitive empathy scales to their total, no correlations were greater than r .50. This suggests that multi-collinearity would not be an issue in regression and all variables may be entered into the regression equation simultaneously. 3 d 2r/O1 r2 when sample sizes are about equal (Lipsey & Wilson, 2001: p. 199). 4 This is approximately equivalent to a 5% difference between those who bullied and those who did not.

66

D. Jolliffe, D.P. Farrington / Journal of Adolescence 34 (2011) 5971

Table 3 Comparison of empathy and individual and social background variables to bullying for males. Not bully (NB) (N 265) M BES cognitive BES affective BES total Urgency impulsivity Thurstones verbal uency Parental supervision SES Non-intact family 32.4 32.4 64.8 38.8 13.6 8.6 3.4 (%) 34.0 Bully (B) (N 97) M 31.5 31.5 63.0 43.6 13.2 9.2 3.6 (%) 41.0 Freq. bully (FB) (N 40) M 30.5 30.1 60.6 44.5 12.9 9.6 3.8 (%) 49.0 d .17 .14 .18 .57*** .10 .26* .11 .12 d .31* .35* .44** .67*** .17 .41** .22 .20 NB v B NB v FB

Note: M mean score, % percent non-intact family. *p < .05, **p < .01, ***p < .001.

d .35, p < .05 and d .44, p < .05 respectively). There were also signicant differences between frequent bullies and those who did not bully on impulsivity (d .67, p < .0001) and poor parental supervision (d .41, p < .001). In the regression predicting frequent bullying by males, again all the individual and social background variables were included (e.g., affective empathy, cognitive empathy, impulsivity, verbal uency, parental supervision, SES and family status). The results again suggested that impulsivity was the most important independent predictor of frequent bullying (LRCS 18.8, p < .0001), followed by affective empathy (LRCS 7.0, p < .008). However, coming from a non-intact family was also found to independently predict frequent bullying by males (LRCS 4.7, p < .03). Independent relationship of empathy to prevalence and frequency of bullying in females Table 4 shows the results of the comparison of those females who did not report bullying others compared with those who reported bullying, and those who reported frequent bullying. It can be seen that females who bullied scored signicantly lower on affective empathy (d .32, p < .05) than those who did not, but females who bullied did not have signicantly lower cognitive empathy. Females who bullied differed signicantly from those who did not on a number of the other individual and social background variables. Those who reported bullying others scored signicantly higher than those who did not on impulsivity (d .75, p < .0001), but scored lower on verbal uency (d .44, p < .01), SES (d .47 p > .0001) and parental supervision (d .39, p < .01). A stepwise logistic regression was carried out with bullying by females as the dependent variable and with affective empathy, cognitive empathy, impulsivity, verbal uency, parental supervision, SES and family status as the explanatory variables. The results showed that impulsivity was a signicant independent predictor of bullying amongst females (LRCS 22.9, p < .0001), as was SES (LRCS 8.4, p < .004), and verbal uency (LRCS 5.4, p < .02). Table 4 also shows the results of the comparisons between females who did not bully and those who took part sometimes or more often (frequent bullies). Because of the small number of frequent bullies (N 12) these results should be considered

Table 4 Comparison of empathy and individual and social background variables to bullying for females. Not bully (NB) (N 290) M BES cognitive BES affective BES total Urgency impulsivity Thurstones verbal uency Parental supervision SES Non-intact family 35.1 40.6 75.7 40.7 15.6 8.0 3.6 (%) 36.0 Bully (B) (N 49) M 34.6 38.8 73.4 46.4 13.9 8.9 4.4 (%) 49.0 Freq. bully (FB) (N 12) M 33.5 35.7 69.2 44.0 12.5 8.6 4.8 (%) 46.0 d .13 .32* .28 .75*** .44** .39** .47*** .20 d .41 .85** .78** .42 .82** .22 .70** .09 NB v B NB v FB

Note: M mean score, d standardized mean difference, % percent non-intact family. *p < .05, **p < .01, ***p < .001.

D. Jolliffe, D.P. Farrington / Journal of Adolescence 34 (2011) 5971

67

exploratory only. Despite the small numbers signicant differences were found between those who did not bully and those who bullied frequently. Females who bullied frequently scored signicantly lower on affective and total empathy than those who did not bully (d .85, p < .01 and d .78, p < .01 respectively). Frequent bullies also scored signicantly lower on verbal uency (d .82, p < .01) and SES (d .70, p < .01). The small number of female frequent bullies prevented the investigation of the independent inuence of empathy on frequent bullying among females. The relationship between empathy and type of bullying for males Table 5 shows the comparison of the different types of bullying (name calling, indirect bullying and violent bullying) compared to empathy as well as the other individual and social background features for males. The results suggest that males involved in name calling did not differ from those who did not take part in this type of bullying in either cognitive or affective empathy. The only factor that signicantly differentiated males involved in name calling from those who did not take part in this was impulsivity (p < .0001). In a logistic regression predicting name calling by males, affective empathy, cognitive empathy, impulsivity and parental supervision were included as explanatory variables. Impulsivity was found to be the only independent predictor of male bullying by name calling (LRCS 14.0, p < .001). Similar to males involved in name calling, males involved in indirect bullying did not have signicantly lower cognitive or affective empathy than those who did not in a zero-order relationship. Also, similar to males involved in name calling, males involved in indirect bullying had signicantly greater impulsivity (p < .0001), while boys involved in indirect bullying also had signicantly lower levels of SES (p < .05). A logistic regression was undertaken to predict indirect bullying by males including all explanatory variables (except non-intact family and cognitive empathy which had d values 0.0). The results suggested that impulsivity (LRCS 11.6, p < .001) and affective empathy (LRCS 5.3, p < .02) were the only variables that were independently related to indirect bullying among males. There was some indication that males who were involved in violent bullying might have lower total empathy than those who did not, but this difference was only signicant in a one-tailed analysis (p < .10). However, male violent bullies had signicantly greater impulsivity and signicantly lower levels of verbal uency and parental supervision. Logistic regression was used to determine those variables that were independently related to violent bullying. All explanatory variables were included (e.g., affective empathy, cognitive empathy, impulsivity, verbal uency, parental supervision, SES and family status). The results suggested that impulsivity accounted for the most variance of in violent bullying (LRCS 35.2, p < .0001), followed by affective empathy (LRCS 5.7, p < .02) and verbal uency (LRCS 5.8, p < .02). The relationship between empathy and type of bullying for females Table 6 shows the comparison of the different types of bullying (name calling, indirect bullying and violent bullying) compared to empathy and the other individual and social background features for females. The results suggested that females involved in name calling did not signicantly differ from those not involved on cognitive or affective empathy. However, females involved in name calling had signicantly greater impulsivity (p < .001), and signicantly lower verbal uency (p < .05), parental supervision (p < .01) and SES (p < .004). A stepwise logistic regression predicting involvement in name calling included affective empathy, cognitive empathy, impulsivity, verbal uency, parental supervision, SES and family status as independent predictors. Only impulsivity (LRCS 14.6, p < .0001) and SES (LRCS 7.8, p < .005) were independently related to name calling among females.

Table 5 Comparison of empathy and individual and social background variables to different types of bullying for males. Not bully (NB) (N 265) M BES cognitive BES affective BES total Urgency impulsivity Thurstones verbal uency Parental supervision SES Non-intact family 32.4 32.4 64.8 38.8 13.6 8.6 3.4 % 34.0 Name calling bully (NCB) (N 54) M 32.1 32.2 64.3 43.1 13.6 8.9 3.5 % 32.0 Indirect bully (IB) (N 26) M 32.4 31.1 63.5 44.9 13.1 8.9 4.1 % 34.0 Violent bully (VB) (N 53) M 31.2 31.2 62.4 45.6 12.8 9.5 3.8 % 45.0 d .06 .03 .05 .51*** 0.0 .13 .06 .03 d .0 .20 .14 .72*** .12 .13 .39* 0.0 d .23 .18 .24 .82*** .20 .38** .22 .19 NB v NCB NB v IB NB v VB

Note: M mean score, % percent non-intact family. *p < .05, **p < .01, ***p < .001.

68

D. Jolliffe, D.P. Farrington / Journal of Adolescence 34 (2011) 5971

Table 6 Comparison of empathy and individual and social background variables to different types of bullying for females. Not bully (NB) (N 290) M BES cognitive BES affective BES total Urgency impulsivity Thurstones intelligence Parental supervision SES Non-intact family 35.1 40.6 75.7 40.7 15.6 8.0 3.6 % 36.0 Name calling bully (NCB) (N 29) M 34.8 39.3 74.2 46.1 14.2 9.1 4.5 % 47.0 Indirect bully (IB) (N 19) M 34.1 38.4 72.4 44.7 15.1 8.3 4.6 % 50.0 Violent bully (VB) (N 20) M 34.6 38.7 73.3 47.6 13.0 9.0 5.3 % 50.0 d .07 .23 .18 .70*** .36* .48* .53* .15 d .25 .38 .39 .52* .13 .14 .59** .15 d .13 .33 .30 .89*** .67** .43 .99*** .15 NB v NCB NB v IB NB v VB

Note: M mean score, % percent non-intact family. *p < .05, **p < .01, ***p < .001.

Based on an examination of the effect sizes there was some indication that females who bullied other indirectly might have lower empathy, especially affective empathy, than those who did not, but this was only signicant in a one-tailed analysis (p < .09). However, females who bullied others indirectly did have signicantly greater impulsivity (p < .03) and signicantly lower SES (p < .01). When a logistic regression was undertaken predicting indirect bullying by females, SES (LRCS 5.6, p < .02) and impulsivity (LRCS 5.3, p < .02) were found to be independently related to indirect bullying among females. Again, it appeared as though females who bullied violently might have lower affective empathy than those who did not (d .33), but this difference was not signicant in a two-tailed analysis. Females who bullied violently had signicantly greater impulsivity (p < .0001) and signicantly lower levels of verbal uency (p < .003) and SES (p < .0001) compared to those who did not. A stepwise logistic regression predicting violent bullying by females was undertaken with affective empathy, cognitive empathy, impulsivity, parental supervision, SES and family status as the independent variables. The results suggested that only impulsivity (LRCS 15.6, p < .0001) and SES (LRCS 13.0, p < .0001) were independently related to this behaviour. Discussion There was considerable evidence that males and females involved in bullying had low empathy and this was particularly the case with affective empathy. Males who bullied had lower cognitive and affective empathy than those who did not bully, and this difference was signicant when frequent bullying (sometimes or greater) was examined. Furthermore, males who bullied indirectly and violently had somewhat low affective empathy while the cognitive empathy of violent male bullies was also low (but not to a statistically signicant degree). Females who bullied had signicantly lower affective empathy than those who did not bully and this difference was even more pronounced amongst frequent female bullies. Female bullies had somewhat low cognitive empathy but this was not signicant. Similarly, females who bullied indirectly and violently appeared to have low affective empathy (and to a lesser degree cognitive empathy amongst females that bullied indirectly and violently), but again the low numbers of females involved in bullying probably prevented this from reaching statistical signicance. The key issue of this research was whether the cognitive or affective empathy decits identied were independently related to various forms of bullying (separately for males and females) when other individual and social background variables were controlled. Importantly, the results suggested that the relationship between bullying and low empathy was different for males and females and for cognitive and affective empathy. The results of a series of regressions suggested that low affective empathy was independently related to male bullying, frequent bullying by males, indirect bullying and violent bullying. Cognitive empathy was not independently related to male bullying. Importantly however, impulsivity was independently related to all forms of male bullying. This suggests that male bullies may very well possess a reduced ability to understand the emotions of others, but this does not appear to be functionally related to bullying. Instead, it appeared that a failure to consider the consequences before acting and the absence of emotional restraint that might be provided by sufcient affective empathy were key aspects of bullying by males. For females the results of the regressions suggested that neither cognitive nor affective empathy were independently related to bullying. Instead the key variables independently related to female bullying as well as female name calling, indirect bullying and violent bullying were high impulsivity and low socioeconomic status, while low verbal uency was also independently related to female bullying. This nding suggests that cognitive and affective empathy decits may be apparent in female bullies, but it is a failure to consider the consequences of their actions on others and low socioeconomic status that

D. Jolliffe, D.P. Farrington / Journal of Adolescence 34 (2011) 5971

69

appears to account for these empathy differences. Interestingly, the nding that low SES and low verbal uency (a measure of executive function) account for the low empathy apparent in the zero-order relationships between empathy and female bullying mirror previous ndings with respect to empathy and offending (Jolliffe & Farrington, 2004). The nding that low SES was independently related to female bullying was in line with research which has found that social factors have a much more signicant impact on female offending than male offending (Farrington & Painter, 2004). Low socioeconomic status probably had an indirect (non-causal) effect on bullying by females likely through family or individual factors that were not assessed in the current research (e.g. Gottfredson, McNeil, & Gottfredson, 1991; Rutter, 1981). Future research should consider the mechanisms whereby low SES might impact on the bullying behaviour of females. Given the link between low SES and intelligence (e.g. Lynam, Moftt, & Stouthamer-Loeber, 1993) it may be that more comprehensive measures of intelligence might account for this nding. Overall, the results suggest that anti-bullying programmes that address low affective empathy might be useful in reducing bullying by males. It appears, however, that most anti-bullying programmes address cognitive as opposed to affective empathy. For example, Olweus (1993) suggested the use of role-playing as a method of having students experience bullying from the perspective of the victim. Similarly, Tattum (1989) emphasized the need for bullies to see things from the point of view of the victim, and Lowenstein (1991) proposed similar methods of making the bully aware of the victims feelings. These are clearly methods of addressing perceived decits in the understanding of anothers feelings. The extent to which roleplaying and various other perspective taking methods can inuence actual emotional experience is questionable (e.g. Hart & Hare, 1997). Therefore, this focus on cognitive empathy might not only prove ineffective at addressing bullying, but if cognitive empathy is used to facilitate effective bullying or recruit others into bullying (e.g. Sutton, Smith, & Swettenham, 1999), these interventions methods might be creating more skilled, and therefore, better bullies. A more suitable treatment target might be the high levels of impulsivity seen in all types of male bullies and female bullies. High impulsivity has been addressed in Australia through a programme called STOP-THINK-DO whereby children are encouraged to clarify problems, consider solutions and act appropriately, but this programme has not been evaluated (Peterson, 1994). Impulsiveness was also a treatment target in the Viennese social competence training program provided to students and evaluated by Atria and Spiel (2007). Furthermore, an experimental intervention which addressed impulsivity through social skills training with a group of 249 disruptive boys aged 6 resulted in signicantly reduced bullying at age nine (Tremblay et al., 1991). The clear relationship between high impulsivity and bullying also provides support for the previously identied link between attention-decit hyperactivity disorder (ADHD) and bullying (Unnever & Cornell, 2003). This suggests that it may be useful to screen those who exhibit bullying behaviour for ADHD, and also that interventions which reduce the impulsivity associated with ADHD may result in reduced bullying. While there are relatively few anti-bullying programmes which specically address the impulsivity of bullies, there are a number of very successful offender treatment programmes designed to decrease the impulsivity of offenders. For example, a recent systematic review and meta-analysis of the Reasoning and Rehabilitation (R&R) offender treatment programme found a 14% decrease in recidivism for programme participants compared to controls (Tong & Farrington, 2008). This programme was designed to address the cognitive decits of offenders by teaching offenders how to think and targeting their high impulsivity. New anti-bullying programmes could borrow methods and ideas from these programmes and incorporate them into the school curriculum. As an offenders behaviour would likely be much more intransigent than that of school bullies, successful translation of the impulsivity components of the R&R programme might be expected to have a signicant impact on reducing bullying. This research is not without its limitations. For example, there are issues with this investigation that may limit its comparability with other studies of bullying. First, the period of time for which bullying was investigated in the current study was approximately 9 months. This is longer than other studies of bullying (e.g., 3 months in Olweus, 1993), and may have led to less accurate recall of bullying incidents. However, the prevalence and frequency of bullying identied in the current study was relatively similar to that of most other studies (e.g., Farrington, 1993) suggesting fair comparability. Similarly, because data collection involved a one-off administration of a questionnaire it may be that those who completed the questionnaire were in some way different with respect to the key variables than those who did not. Perhaps most importantly is the cross-sectional nature of the research. It may be that high impulsivity and low affective empathy are causally related to male bullying, but it is equally possible that impulsivity and affective empathy are decreased after bullying. In addition, it is well recognized that bullying is a group phenomenon (e.g. Salmivalli, 1999) with individuals usually adopting different roles. It was not possible in this research to determine if the individual who reported undertaking bullying initiated the incident or simply followed along with others who had initiated an incident. This might have a signicant impact as we might expect those who initiate incidents to have lower affective empathy, but potentially higher cognitive empathy which they could use to manipulate others to follow along. Many of these issues could be addressed by undertaking a prospective longitudinal study with repeated assessments whereby empathy was measured before bullying had commenced, and where questions were asked about the roles undertaken in various bullying incidents. Also, bullying could be assessed by multiple methods (e.g., self, teacher, peer reports), which would reduce shared method variance. This longitudinal methodology could help to more rmly establish the causal ordering of empathy and bullying and also the relationship of levels of empathy to the role adopted in any bullying. This method might also allow for trajectories of the seriousness of bullying to be assessed. It may be that there are distinct typologies of bullies, some of whom use name calling and subsequently desist from this behaviour as they age, and for some of

70

D. Jolliffe, D.P. Farrington / Journal of Adolescence 34 (2011) 5971

whom name calling is the beginning of a career of bullying of increasing frequency and severity. There has never been such a longitudinal study of empathy and bullying. There is also a need for randomized experiments to test the effectiveness of anti-bullying programmes informed by the results of this research. These programmes should include those elements that have been identied as successful at decreasing bullying (e.g. parent training, improved playground supervision, disciplinary methods, school conferences, etc.) in a recent systematic review and meta-analysis (Tto & Farrington, 2009; Tto, Farrington, & Baldry, 2008). References
Albeiro, P., Matricardi, G., Speltri, D., & Toso, D. (2009). The assessment of empathy in adolescence: a contribution to the Italian validation of The Basic Empathy Scale. Journal of Adolescence, 32, 393408. Atria, M., & Spiel, M. (2007). Viennese social competence (ViSC) training for students: program and evaluation. In J. E. Zins, M. J. Elias, & C. A. Maher (Eds.), Bullying, victimization and peer harassment: A handbook of prevention and intervention (pp. 179197). Binghamton, NY: Haworth Press. Baldry, A., & Farrington, D. P. (2000). Bullies and delinquents: personal characteristics and parental styles. Journal of Community and Applied Social Psychology, 10, 1731. Beinart, S., Anderson, B., Lee, S., & Utting, D. (2002). Youth at risk? London: Communities That Care. Bernstein, J. Y., & Watson, M. W. (1997). Children who are targets of bullying: a victim pattern. Journal of Interpersonal Violence, 12, 483498. Bjorkvist, K., Lagerspetz, K. M. J., & Kaukiainen, A. (1992). Do girls manipulate and boys ght? Developmental trends in regard to direct and indirect aggression. Aggressive Behavior, 18, 117127. Bosworth, K., Espelage, D. L., & Simon, T. R. (1999). Factors associated with bullying behavior in middle school students. Journal of Early Adolescence, 19, 341362. Bowers, L., Smith, P. K., & Binney, V. (1992). Cohesion and power in the families of children involved in bully/victim problems at school. Journal of Family Therapy, 14, 371387. Bryant, B. K. (1982). An index of empathy for children and adolescents. Child Development, 53, 413425. Camodeca, M., Goossens, F. A., Terwogt, M. M., & Schuengel, C. (2002). Bullying and victimization among school-age children: stability and links to proactive and reactive aggression. Social Development, 11, 332345. Carroll, J. B. (1993). Human cognitive abilities. Cambridge: Cambridge University Press. Cohen, J. (1988). Statistical Power Analysis for the Behavioral Sciences (2nd ed). Hillsdale, NJ: Erlbaum. Cohen, D., & Strayer, J. (1996). Empathy in conduct-disordered and comparison youth. Developmental Psychology, 32, 988998. Cohen, P. (1996). Childhood risk for adult symptoms of personality disorder: method and substance. Multivariate Behavioral Research, 31, 121148. Connolly, I., & OMoore, M. (2003). Personality and family relations of children who bully. Personality and Individual Differences, 35, 559567. Coolidge, F. L., DenBoer, J. W., & Segal, D. L. (2004). Personality and neuropsychological correlates of bullying behavior. Personality and Individual Differences, 36, 15591569. Cramer, D. (2003). Advanced quantitative data analysis. Maidenhead: Open University Press. DAmbrosio, F., Olivier, M., Didon, D., & Besche, C. (2009). The basic empathy scale: a French validation of a measure of empathy in youth. Personality and Individual Differences, 46, 160165. Davis, M. H. (1983). Measuring individual differences in empathy: evidence for a multidimensional approach. Journal of Personality and Social Psychology, 44, 113126. Duyme, M. (1990). Antisocial behaviour and postnatal environment: a French adoption study. Journal of Child Psychology and Psychiatry, 31, 699710. Endresen, I. M., & Olweus, D. (2002). Self-reported empathy in Norwegian adolescents: sex-differences, age trends, and relationship to bullying. In D. Stipek, & A. Bohart (Eds.), Constructive and destructive behavior. Implications for family, school, society (pp. 147165). Washington, DC: American Psychological Association. Espelage, D. L., Bosworth, K., & Simon, T. R. (2001). Short-term stability and prospective correlates of bullying in middle-school students: an examination of potential demographic, psychosocial and environmental inuences. Violence and Victims, 16, 411426. Espelage, D. L., & Holt, M. L. (2001). Bullying and victimization during early adolescence: peer inuences and psychosocial correlates. Journal of Emotional Abuse, 2, 123142. Espelage, D. L., Mebane, S. E., & Adams, R. S. (2004). Empathy, caring, and bullying: toward an understanding of complex associations. In D. L. Espelage, & S. M. Swearer (Eds.), Bullying in American Schools: A socialecological perspective on prevention and intervention (pp. 3762). New Jersey: Lawrence Erlbaum. Farrington, D. P. (1993). Understanding and preventing bullying. In M. Tonry (Ed.), Crime and justice, Vol. 12 (pp. 381458). Chicago: University of Chicago Press. Farrington, D. P., & Baldry, A. C. (2006). Factores de riesgo individuales. In A. Serrano (Ed.), Acoso y Violencia en la Escuela (pp. 107133). Barcelona: Ariel. Farrington, D. P., & Loeber, R. (1989). RIOC and phi as measures of predictive efciency and strength of association in 2 2 tables. Journal of Quantitative Criminology, 5, 201213. Farrington, D. P., & Loeber, R. (2000). Some benets of dichotomization in psychiatric and criminological research. Criminal Behaviour and Mental Health, 10, 100122. Farrington, D. P., & Painter, K. A. (2004). Gender differences in offending: Implications for risk-focussed prevention. London: Home Ofce. Online Report 09/04. Feshbach, N. D. (1975). Empathy in children: some theoretical and empirical considerations. The Counseling Psychologist, 5, 2530. Flouri, E., & Buchanan, A. (2003). The role of mother involvement and father involvement in adolescent bullying behavior. Journal of Interpersonal Violence, 18, 634644. Frith, C. D., Friston, K., Liddle, P. F., & Fracknowiak, P. (1991). Willed action and the prefrontal cortex in man: a study with PET. Proceedings of the Royal Society of London, 244, 241248. Gottfredson, D. C., McNeil, R. J., & Gottfredson, G. D. (1991). Social area inuences on delinquency: a multilevel analyses. Journal of Research in Crime and Delinquency, 28, 197226. Hart, S. D., & Hare, R. D. (1997). Psychopathy assessment and associations. In D. M. Stoff, J. Breiling, & J. D. Maser (Eds.), Handbook of antisocial behavior (pp. 2235). New York: Wiley. Haynie, D. L., Nansel, T., Eitel, P., Crump, A. D., Saylor, K., Yu, K., et al. (2001). Bullies, victims and bully/victims: distinct groups of at risk youth. Journal of Early Adolescence, 21, 2949. Hogan, R. (1969). Development of an empathy scale. Journal of Consulting and Clinical Psychology, 33, 307316. Jolliffe, D., & Farrington, D. P. (2004). Empathy and offending: a systematic review and meta-analysis. Aggression and Violent Behavior, 9, 441476. Jolliffe, D., & Farrington, D. P. (2006a). Examining the relationship between low empathy and bullying. Aggressive Behavior, 32, 111. Jolliffe, D., & Farrington, D. P. (2006b). The development and validation of the basic empathy scale. Journal of Adolescence, 29, 589611. Kraemer, H. C., Lowe, K. K., & Kupfer, D. J. (2005). To your health? How to understand what research tells us about risk. Oxford: Oxford University Press. Kulik, J. A., Stein, K. B., & Sarbin, T. R. (1968). Disclosure of delinquent behaviour under conditions of anonymity and nonanonimity. Journal of Consulting and Clinical Psychology, 32, 506509. Kumpulainen, K., Rasanen, E., & Henttonen, I. (1999). Children involved in bullying: psychological disturbance and the persistence of the involvement. Child Abuse and Neglect, 23, 12531262.

D. Jolliffe, D.P. Farrington / Journal of Adolescence 34 (2011) 5971

71

Lagerspetz, K. M. J., Bjorkvist, K., & Peltonen, T. (1988). Is indirect aggression typical of females? Gender differences in aggressiveness in 11- to 12-year old children. Aggressive Behavior, 14, 403414. Lipsey, M. W., & Wilson, D. B. (2001). Practical meta-analysis. Thousand Oaks, CA: Sage. Lowenstein, L. F. (1978). Who is the bully? Bulletin of the British Psychological Society, 31, 147149. Lowenstein, L. F. (1991). The study, diagnosis and treatment of bullying in a therapeutic community. In M. Elliot (Ed.), Bullying: A practical guide to coping for schools. Harlow: Longman. Lynam, D., Moftt, T., & Stouthamer-Loeber, M. (1993). Explaining the relation between IQ and delinquency: class, race, test motivation, school failure, or self-control? Journal of Abnormal Psychology, 102, 187196. Marshall, W. L., Hudson, S. M., Jones, R., & Fernandez, Y. L. (1995). Empathy in sex offenders. Clinical Psychology Review, 15, 99113. Mellor, A. (1990). Bullying in Scottish secondary schools. Edinburgh: Scottish Council for Research in Education. Merrell, K. M., Gueldner, B. A., Ross, S. W., & Isava, D. M. (2008). How effective are scholl bullyinh intervention programs? A meta-analysis of intervention research. School Psychology Quarterly, 23, 2642. Milner, B. (1964). Some effects of frontal lobectomy in man. In J. M. Warren, & K. Akert (Eds.), The frontal granular cortex and behavior (pp. 313334). New York: McGraw-Hill. Moftt, T. E., Caspi, A., Rutter, M., & Silva, P. (2001). Sex differences in antisocial behaviour: Conduct disorder, delinquency, and violence in the Dunedin longitudinal study. Cambridge: Cambridge University Press. Moftt, T. E., & Henry, B. (1989). Neuropsychological decit and self-reported delinquency in an unselected birth cohort. Journal of the American Academy of Child and Adolescent Psychiatry, 27, 233240. Monks, C. P., Smith, P. K., & Swettenham, J. (2005). Psychological correlates of peer victimization in preschool: social cognitive skills, executive function and attachment proles. Aggressive Behavior, 31, 571588. Murray, J., Farrington, D. P., & Eisner, M. P. (2009). Drawing conclusions about causes from systematic reviews of risk factors: the Cambridge Quality Checklists. Journal of Experimental Criminology, 5, 123. Olweus, D. (1978). Aggression in the schools. Washington, DC: Hemisphere. Olweus, D. (1980). Familial and temperamental determinants of aggressive behavior in adolescent boys: a causal analysis. Child Development, 16, 644660. Olweus, D. (1993). Bullying at school. Oxford, UK: Blackwell Publishers. OMoore, A. M., & Hillery, B. (1991). What do teachers need to know? In M. Elliot (Ed.), Bullying: A practical guide to coping for schools Harlow: Longman. Peterson, L. (1994). Stop and think learning: motivating learning in social groups and individuals. In M. Tanish, & J. Izard (Eds.), Widening horizons: New challenges, directions and achievements (pp. 7083). Melbourne: Australian Council for Educational Research. Phelps, E. A., Hyder, F., Blamire, A. M., & Shulman, R. (1997). FMRI of the prefrontal cortex during verbal uency. Neuroreport, 8, 561565. Pithers, W. D. (1999). Empathy: denition, enhancement, and relevance to the treatment of sexual abusers. Journal of Interpersonal Violence, 14, 257284. Rigby, K. (1994). Psychosocial functioning in families of Australian adolescent school children involved in bully/victim problems. Journal of Family Therapy, 16, 173187. Rigby, K. (1996). Bullying in schools and what to do about it. London, UK: Jessica Kingsley. Roland, E. (1989). Bullying: the Scandinavian research tradition. In D. P. Tattum, & D. A. Lane (Eds.), Bullying in schools (pp. 2132), Stoke-on-Trent, Trentham, UK. Rutter, M. (1981). The city and the child. American Journal of Orthopsychiatry, 51, 610625. Salmivalli, C. (1999). Participant role approach to group bullying: implications for interventions. Journal of Adolescence 453459. Smith, P. K., & Sharp, S. (1994). The problem of school bullying. In P. K. Smith, & S. Sharp (Eds.), School bullying: Insights and perspectives (pp. 119). London: Routledge. Smith, P. K., & Thompson, D. (1991). Dealing with bully/victim problems in the UK. In P. K. Smith, & D. Thompson (Eds.), Practical approaches to bullying (pp. 112). London: David Fulton. Stephenson, P., & Smith, D. (1989). Bullying in the junior school. In D. Tattum, & D. Lane (Eds.), Bullying in schools (pp. 4558), Stoke-on-Trent, Trentham. Strayer, J. (1993). Childrens concordant emotions and cognitions in response to observed emotions. Child Development, 64, 188201. Sutton, J., Smith, P. K., & Swettenham, J. (1999). Bullying and theory of mind: a critique of social skills decit view of antisocial behaviour. Social Development, 8, 118126. Tattum, D. P. (1989). Violence and aggression in schools. In D. P. Tattum, & D. A. Lane (Eds.), Bullying in schools (pp. 119). Stoke-on-Trent: Trentham Books. Thurstone, L. L. (1938). Primary mental abilities. Chicago: University of Chicago Press. Tong, L. S. J., & Farrington, D. P. (2008). Effectiveness of reasoning and rehabilitation in reducing reoffending. Psicothema, 20, 2028. Tremblay, R. E., McCord, J., Boileau, H., Charlebois, P., Gangnon, C., LeBlanc, M., et al. (1991). Can disruptive boys be helped to become competent? Psychiatry, 54, 148161. Tto, M., & Farrington, D. P. (2009). What works in preventing bullying: effective elements of anti-bullying programmes. Journal of Aggression, Conict and Peace Research, 1, 1324. Tto, M., Farrington, D. P., & Baldry, A. C. (2008). Effectiveness of programmes to reduce school bullying: A systematic review. Stockholm, Sweden: National Council for Crime Prevention. Unnever, J. D., & Cornell, D. G. (2003). Bullying, self-control, and ADHD. Journal of Interpersonal Violence, 18, 129147. Veenstra, R., Lindenberg, S., De Winter, A. F., Oldehinkel, A. J., Verhulst, F. C., & Ormel, J. (2005). Bullying and victimisation in elementary schools: a comparison of bullies, victims, bully/victims and uninvolved preadolescents. Developmental Psychology, 41, 672682. Walker, A., OBrien, M., Traynor, J., Fox, K., Goddard, E., & Foster, K. (2002). Living in Britain. London: Ofce of National Statistics. Warden, D., & Mackinnon, S. (2003). Prosocial children, bullies and victims: an investigation of their sociometric status, empathy and social problem-solving strategies. British Journal of Developmental Psychology, 21, 367385. Whiteside, S. P., & Lynam, D. R. (2001). The ve factor model and impulsivity: using a structural model of personality to understand impulsivity. Personality and Individual Differences, 30, 669689. Whitney, I., Rivers, I., Smith, P. K., & Sharp, S. (1994). The Shefeld project: methodology and ndings. In P. K. Smith, & S. Sharp (Eds.), School bullying: Insights and perspectives (pp. 2056). London: Routledge. Whitney, I., & Smith, P. K. (1993). A survey of the nature and extent of bullying in junior/middle and secondary schools. Educational Research, 35, 325. Wikstrom, P. O., & Butterworth, D. (2006). Adolescent crime: Individual differences and lifestyles. London: Willan.

Вам также может понравиться