Read 5 articles and answer questions
wiluiana
olweus_d._2013.pdf
CP09CH27-Olweus ARI 25 February 2013 18:29
School Bullying: Development and Some Important Challenges Dan Olweus Research Centre for Child and Youth Mental Health, Uni Health, Uni Research, and University of Bergen, PB 7810, Bergen NO-5020 Norway; email: [email protected]
Annu. Rev. Clin. Psychol. 2013. 9:751–80
First published online as a Review in Advance on January 3, 2013
The Annual Review of Clinical Psychology is online at http://clinpsy.annualreviews.org
This article’s doi: 10.1146/annurev-clinpsy-050212-185516
Copyright c© 2013 by Annual Reviews. All rights reserved
Keywords
definition, bully, victim, cyber bullying, peer relations research
Abstract
After sketching how my own interest and research into bullying problems began, I address a number of potentially controversial issues related to the definition and measurement of such problems. The importance of maintain- ing the distinctions between bullying victimization and general victimization and between bullying perpetration and general aggression is strongly em- phasized. There are particular problems with the common method of peer nominations for purposes of prevalence estimation, comparisons of such es- timates and mean levels across groups and time, and measurement of change. Two large-scale projects with time series data show that several recent claims about cyber bullying made in the media and by some researchers are greatly exaggerated and lack scientific support. Recent meta-analyses of the long- term outcomes for former bullies and victims provide convincing evidence that being involved in such problems is not just a harmless and passing school problem but something that has serious adjustment and public health con- sequences that also entail great costs to society. Another section presents my view of why the theme of bullying took quite some time to reach the peer relations research community in the United States and the role of a dominant research tradition focusing on “likeability” in this account. In a final section, I summarize some reasons why it may be considered impor- tant and interesting to focus both research and intervention on bully/victim problems.
751
A nn
u. R
ev . C
lin . P
sy ch
ol . 2
01 3.
9: 75
1- 78
0. D
ow nl
oa de
d fr
om w
w w
.a nn
ua lr
ev ie
w s.
or g
by S
ta te
U ni
ve rs
ity o
f N
ew Y
or k
- B
uf fa
lo o
n 07
/2 7/
13 . F
or p
er so
na l u
se o
nl y.
CP09CH27-Olweus ARI 25 February 2013 18:29
Contents
INTRODUCTION . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 752 THE BEGINNINGS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 753 MEASURING BULLY/VICTIM PROBLEMS WITH THE OLWEUS
BULLYING QUESTIONNAIRE . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 755 DEFINITION OF BULLYING . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 756
Intentionality . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 757 Some Repetitiveness . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 757 Power Imbalance . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 757 What About the Three Criteria as Regards Cyber Bullying? . . . . . . . . . . . . . . . . . . . . . . 758
SOME MEASUREMENT ISSUES. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 759 Distinctions Between Bullying Victimization and General Victimization and
Between Bullying Perpetration and General Aggression . . . . . . . . . . . . . . . . . . . . . . . 759 Does the Power Imbalance Make a Difference?. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 760
SOME MEASUREMENT ISSUES RELATED TO THE OLWEUS BULLYING QUESTIONNAIRE . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 761 Reliability . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 761 Convergent Validity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 762
FACTORS CONTRIBUTING TO LACK OF CORRESPONDENCE AND PROBLEMS WITH PEER NOMINATIONS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 763
ISSUES RELATED TO CYBER BULLYING . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 765 Claims About Prevalence and Time Trends . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 766 Degree of Overlap of Traditional Bullying with Cyber Bullying . . . . . . . . . . . . . . . . . . . 767 Are There Negative Effects of Cyber Bullying, and How Do We Find Out? . . . . . . . 768 Unfortunate Consequences of the Distorted Media Picture . . . . . . . . . . . . . . . . . . . . . . . 768 How Can We Counteract and Prevent Cyber Bullying? . . . . . . . . . . . . . . . . . . . . . . . . . . 768 A System-Level Proposal . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 769
THE LONG-TERM OUTCOMES FOR BULLIES. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 769 THE LONG-TERM OUTCOME FOR VICTIMS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 770 EFFECTIVENESS OF ANTIBULLYING PROGRAMS . . . . . . . . . . . . . . . . . . . . . . . . . . 771 WHY BULLY/VICTIM PROBLEMS CAME LATE IN THE
UNITED STATES . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 771 SOME REASONS FOR FOCUSING ON BULLY/VICTIM PROBLEMS . . . . . . . . . 775 CONCLUDING WORDS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 775
INTRODUCTION
Research on bully/victim problems over the past 15 years or so has exploded, and there has been a parallel increase in attention from the media, schools, and relevant authorities. In my view, this has been a successful and fortunate development by and large. Like any relatively new field, however, the field of bullying research is to some extent plagued by problems, disagreements, and unresolved issues. To achieve a solid and coherent body of knowledge with well-replicated findings, it is important to reach some consensus on important conceptual and methodological issues. In this review, I take a largely critical-analytical approach and focus more on problems than on successes. I begin by describing how my own interest and research on these problems started
752 Olweus
A nn
u. R
ev . C
lin . P
sy ch
ol . 2
01 3.
9: 75
1- 78
0. D
ow nl
oa de
d fr
om w
w w
.a nn
ua lr
ev ie
w s.
or g
by S
ta te
U ni
ve rs
ity o
f N
ew Y
or k
- B
uf fa
lo o
n 07
/2 7/
13 . F
or p
er so
na l u
se o
nl y.
CP09CH27-Olweus ARI 25 February 2013 18:29
and conclude by summarizing some reasons why it may be considered important and interesting to focus both research and intervention on bully/victim problems.
A terminological note: I use the terms bullying, bully/victim problems, bullying others/being bullied, and bullying perpetration/bullying victimization as approximate synonyms in this review. Although we try as much as possible to avoid the labels bully and victim in practical intervention work and otherwise, they are the terms used regularly in the research literature and are also used in the present text.
THE BEGINNINGS
A strong societal interest in the phenomenon of peer bullying first started in Sweden in the late 1960s and early 1970s under the designation mobbning or “mobbing” (Heinemann, 1969, 1972; Olweus 1973a). The term was introduced into the public Swedish debate by a school physician, P.-P. Heinemann, in the context of racial discrimination (Heinemann 1969). Heinemann had “borrowed” the term mobbing from the Swedish version of a book on aggression written by the well-known Austrian ethologist, Konrad Lorenz (1963, 1968). In ethology, the word mobbing is used to describe a collective attack by a group of animals on an animal of another species, which is usually larger and a natural enemy of the group. In Lorenz’s book (1963), mobbing was also used to characterize the action of a school class or a group of soldiers ganging up against a deviating individual. Parenthetically, it was later was found out that it was not Lorenz who used the term mobbing in his German original but the Swedish translator, himself an ethologist by profession (Lagerspetz et al. 1982).
The term mob has been used for quite some time in social psychology (Lindzey 1954) and to some extent by the general public in English-speaking countries, to describe a relatively large group of individuals—a crowd or a mass of people—joined in a common activity or goal. As a rule, the mob has been formed by accident, is loosely organized, and exists only for a short time. In the social psychological literature, distinctions have been made between several types of mobs, including the aggressive mob (the lynch mob) and the panic-stricken mob (the flight mob). Members of the mob usually experience strong emotions, and the behavior and reactions of the mob are considered to be fairly irrational (see Lindzey 1954).
At an early point in this debate, I expressed doubts about the suitability of the term mobbing, as used in ethology/social psychology, to describe the kind of peer harassment that occurred in school settings (Olweus 1973a, 1978). Generally, with my background in aggression research (e.g., Olweus 1969, 1973b), I felt that the connotations implied in the concept of mobbing could easily lead to inappropriate expectations about the phenomenon of bullying and to certain aspects of the problem being overlooked.
One particular point of concern with the term mobbing related to the relative importance of the group versus its individual members. The notion that school mobbing is a matter of collective aggression by a relatively homogeneous group did in my view obscure the relative contributions made by individual members or small subgroups. More specifically, the role of particularly active perpetrators or bullies could easily be lost sight of within this group framework. In this context, I also questioned how often the kind of all-against-one situations implied in mobbing actually occur in school. If harassment by a small group or by a single individual were the more frequent type in schools, the concept of mobbing might, for example, result in teachers having difficulty identifying the phenomenon of bullying in their classrooms. In addition, the concept of mobbing often seems to place responsibility for potential problems with the recipient of the collective aggression, the victim, who is seen as irritating or provoking the majority of ordinary students in one way or another. This was clearly the situation in the early Swedish debate.
www.annualreviews.org • School Bullying 753
A nn
u. R
ev . C
lin . P
sy ch
ol . 2
01 3.
9: 75
1- 78
0. D
ow nl
oa de
d fr
om w
w w
.a nn
ua lr
ev ie
w s.
or g
by S
ta te
U ni
ve rs
ity o
f N
ew Y
or k
- B
uf fa
lo o
n 07
/2 7/
13 . F
or p
er so
na l u
se o
nl y.
CP09CH27-Olweus ARI 25 February 2013 18:29
Use of the concept of mobbing might also lead to an overemphasis on temporary and situa- tionally determined circumstances: “The mob, suddenly and unpredictably, seized by the mood of the moment, turns on a single individual, who for some reason or other has attracted the group’s irritation and hostility” (Olweus 1978, p. 5). Although I believed that such temporary emotional outbreaks from a group of school children could occur, I considered it more important to direct attention to another kind of situation, in which an individual student is exposed to aggression systematically and over longer periods of time—whether from another individual, a small group, or a whole class (Olweus 1973a, 1978).
An additional problem was that, at that time, there existed basically no empirical research data to shed light on the many issues and concerns involved in the general debate about the bullying phenomenon. Against this background, I initiated in Sweden (I am a native Swede who has lived in Norway for more than 40 years), in the early 1970s, what now appears to be the first systematic research project on bullying by peers. Results from this project were first published as a book in Swedish in 1973 (Olweus 1973a). In 1978, a somewhat expanded version of this book appeared in the United States under the title Aggression in the Schools: Bullies and Whipping Boys (Olweus 1978). A key aim of this research was to sketch a first outline of the anatomy of peer harassment in schools and to seek empirical answers to at least some of the key questions that had been in focus in the public Swedish debate.
In retrospect, I think it is fair to say that this project and later research (e.g., Farrington 1993; Olweus 1978, 1993, 1994) have shown that several of my early concerns were justified. For example, there is no doubt that students in a class vary markedly in their degree of aggressiveness and that these individual differences tend to be quite stable over time, often over several years, if no systematic intervention is introduced (Olweus 1977, 1979). This and similar statements in my writings have sometimes been interpreted as if I don’t apply or appreciate a group perspective on bullying problems. That is a misunderstanding: A group (and larger-context) perspective on bullying problems is certainly both necessary and valuable, but so is an individual difference perspective (see, e.g., Olweus 2001, pp. 14–17, about the bullying circle, among other things).
Similarly, the research clearly shows that a relatively small number of students in a classroom are usually much more actively engaged in the harassment or bullying than others, who are not directly involved in bullying at all or only in more or less marginal roles (Olweus 1993, 2001). Reports from bullied students indicate that they are most often mainly bullied by a small group of two or three students (Olweus & Solberg 1998), often with a negative leader. In addition, a considerable proportion of the victims, some 25% to 35%, report that they are mainly bullied by a single student (Olweus & Solberg 1998). Data from researchers in England, Holland, and Japan, participating in the same cross-national project on bully/victim problems, indicate that this is largely true also in other ethnic contexts with (partly) different cultural backgrounds and traditions ( Junger-Tas & Kesteren 1998, Morita & Soeda 1998, Smith et al. 1999). Further, these and other data (e.g., Rigby & Slee 1991) also show that a considerable proportion of the students in a class have a relatively negative attitude toward bullying and would like to do, or actually try to do (according to self-reports), something to help the victim.
This research-based picture of peer harassment in schools is very different from what is gen- erally implied in the social psychological or ethological concepts of mobbing. Also, the use of the term mobbing (and derivatives of it) by Scandinavians has certainly come to deviate from both the scientific and the ordinary English root meaning of the term. This is particularly evident when we hear a (Scandinavian) student saying “he/she mobbed me today.” Obviously, the word mobbing has gradually, and in part on the basis of highly publicized research findings, acquired a new meaning in Scandinavian everyday language, loosely implying relatively systematic, repetitive harassment of an individual (or possibly a group) by one or more other individuals (usually but
754 Olweus
A nn
u. R
ev . C
lin . P
sy ch
ol . 2
01 3.
9: 75
1- 78
0. D
ow nl
oa de
d fr
om w
w w
.a nn
ua lr
ev ie
w s.
or g
by S
ta te
U ni
ve rs
ity o
f N
ew Y
or k
- B
uf fa
lo o
n 07
/2 7/
13 . F
or p
er so
na l u
se o
nl y.
CP09CH27-Olweus ARI 25 February 2013 18:29
not necessarily by a peer/peers). This new meaning of the word is well established in Norway, Sweden, and Denmark.
At the same time, it was clear at an early stage that, for an English-speaking audience, the terms mob and mobbing are not very useful in describing the phenomenon of bullying; they typically elicit associations in the direction of the social psychological/ethological concepts and the original meaning of the word mob. On the basis of experiences along these lines, I decided to use the term bully/victim (or whipping boy) problems (instead of, or in addition to, mobbing) in my early writings in English (e.g., Olweus 1978). And most of the international attention to these problems over the past 20 years or so concerns some variant of the term bullying and not mobbing.
At the time of initiation of my first research project on bullying, it was not possible, or even desirable, to set forth a very stringent definition of peer harassment or bullying. However, the need for a relatively clear and circumscribed definition became urgent in connection with the government-initiated campaign against bullying in Norway in 1983 (Olweus 1986, 1993). Specif- ically, an important part of this campaign was a nationwide registration of bully/victim problems by means of a student questionnaire that I developed. The basic definition of bullying or peer victimization underlying the construction of the questionnaire was the following: A student is being bullied or victimized when he or she is exposed, repeatedly and over time, to negative ac- tions on the part of one or more other students. This definition emphasized intentionally negative or aggressive acts that are carried out repeatedly and over time. It was further specified that in bullying there is a certain imbalance of power or strength. The student who is exposed to negative actions has difficulty defending himself or herself (for further details, see Olweus 1993, 1999a).
MEASURING BULLY/VICTIM PROBLEMS WITH THE OLWEUS BULLYING QUESTIONNAIRE
In my first research project on bullying comprising some 900 boys who were 13 to 15 years old at the first time of measurement, I used a combination of teacher nominations and peer ratings to classify students as victims (whipping boys), bullies, and “control boys” (Olweus 1973a, 1978). The project also used a number of other data sources including self-reports, mother reports, stress hor- mone data, projective techniques, and psycho-physiological measurements (e.g., Olweus 1980a,b; Olweus et al. 1988). Although a number of self-report items related to bullying and victimization were included in the project, they were not used in the classification of the students into the vari- ous bully/victim categories. However, extensive experience with the questionnaire I developed in the context of Norway’s nationwide 1983 campaign against bullying and the associated interven- tion project from 1983 to 1985 (Olweus 1991, 2005) convinced me that a carefully constructed questionnaire can be an excellent tool for the measurement of bully/victim problems.
Although the basic definition of bullying (involving the three criteria listed above) has been retained unchanged, the “definition” presented to the students in a revised version of the question- naire, the Revised Olweus Bullying Questionnaire (OBQ) (in earlier writings, often referred to as the Olweus Bully/Victim Questionnaire; Olweus 1996, 2007), has been somewhat expanded. In the latest version of the questionnaire (Olweus 2007) this (copyrighted) definition reads as follows:
We say a student is being bullied when another student, or several other students,
� say mean and hurtful things or make fun of him or her or call him or her mean and hurtful names � completely ignore or exclude him or her from their group of friends or leave him or her out of
things on purpose � hit, kick, push, shove around, or lock him or her inside a room
www.annualreviews.org • School Bullying 755
A nn
u. R
ev . C
lin . P
sy ch
ol . 2
01 3.
9: 75
1- 78
0. D
ow nl
oa de
d fr
om w
w w
.a nn
ua lr
ev ie
w s.
or g
by S
ta te
U ni
ve rs
ity o
f N
ew Y
or k
- B
uf fa
lo o
n 07
/2 7/
13 . F
or p
er so
na l u
se o
nl y.
CP09CH27-Olweus ARI 25 February 2013 18:29
� tell lies or spread false rumors about him or her or send mean notes and try to make other students dislike him or her
� and other hurtful things like that.
When we talk about bullying, these things happen repeatedly, and it is difficult for the student being bullied to defend himself or herself. We also call it bullying when a student is teased repeatedly in a mean and hurtful way. But we do not call it bullying when the teasing is done in a friendly and playful way. Also, it is not bullying when two students of about the same strength or power argue or fight.
After a general or global question about being bullied in the past couple of months (or bullying other students, in a different section of the questionnaire), taking all possible forms of bullying into account, the students are asked to respond to questions about nine specific forms of bullying they may have been exposed to. These various forms of bullying comprise direct physical and verbal (including racial and sexual) harassment and threatening and coercive behaviors, as well as more indirect or relational ways of harassment in the form of intentional social isolation, having rumors spread, and manipulation of friendship relationships. There are also some questions about digital or cyber bullying (below).
The questionnaire can be administered in both anonymous (e.g., Olweus 2005) and confidential (e.g., Olweus 1991, Solberg & Olweus 2003) modes. To make the measuring instrument more sensitive to change, most of the questions refer to a specific, relatively short reference period, “the past couple of months.” For most of the questions, the response alternatives are frequency alternatives, and they are made as concrete as possible: “I have not been bullied at school in the past couple of months,” “it has only happened once or twice,” “2 or 3 times a month,” “about once a week,” and “several times a week,” usually coded from 0 to 4. The questionnaire also contains several questions about the reactions of “others” to bullying, as perceived by the respondents; that is, the behavior and attitudes of teachers, peers, and parents. These questions provide important information about the school’s efforts to counteract bullying and in which areas additional efforts may be particularly needed.
DEFINITION OF BULLYING
As defined above, bullying is a subset of aggressive behavior, which in turn is generally defined as “behavior intended to inflict injury or discomfort upon another individual” or in similar terms (Berkowitz 1993, Olweus 1973b, Tedeschi & Felson 1994). Since most bullying occurs without apparent provocation on the part of the targeted child or youth, it is usually considered to be a form of proactive aggression (in contrast to reactive aggression; Coie & Dodge 1998). Bullying is thus aggressive behavior with certain special characteristics such as an asymmetric power relationship and some repetitiveness.
Use of the three criteria of intentionality, some repetitiveness, and imbalance of power for classification of a behavior as what can be called traditional or conventional bullying seems to have been well accepted among both researchers and practitioners for a substantial number of years (e.g., Smith & Brain 2000, Smith et al. 2012). With the advent of cyber bullying, that is, bullying via electronic forms of contact (Smith et al. 2008) or communication, concerns have been raised about whether and possibly how both the repetitiveness and the power imbalance criteria in the general definition of traditional bullying can be applied to cyber bullying. It is obvious, for example, that some ways of being cyber bullied such as having been exposed to a personally embarrassing picture or video on a website are usually single—and not repeated—acts for both target and perpetrator but can spread quickly to a large group of people. And how should one conceive of the power
756 Olweus
A nn
u. R
ev . C
lin . P
sy ch
ol . 2
01 3.
9: 75
1- 78
0. D
ow nl
oa de
d fr
om w
w w
.a nn
ua lr
ev ie
w s.
or g
by S
ta te
U ni
ve rs
ity o
f N
ew Y
or k
- B
uf fa
lo o
n 07
/2 7/
13 . F
or p
er so
na l u
se o
nl y.
CP09CH27-Olweus ARI 25 February 2013 18:29
imbalance in episodes of cyber bullying? Before addressing such questions, I want to make a few comments on the three basic criteria used in the general definition of (traditional) bullying.
Intentionality
As mentioned above, the concept of aggressive behavior(s) implies an intent or desire/aim to inflict harm (injury or discomfort) upon another. Some people have raised concerns about the use of intentionality of bullying behavior as a form of aggressive behavior: How do we know that the perpetrator really intends to harm the other person? And following up: If this is part of the definition, should not the investigator/observer/researcher make sure/document that this is actually the case? The issue of intentionality has a long history of discussion among aggression researchers, and with regard to most behaviors that are considered aggressive the assessment of intentionality typically does not come from asking the perpetrator about his or her intentions, for example, but rather from an analysis and understanding of the context. If it can be implied or assumed that the perpetrator(s) knows or understands that the exerted behavior is or will be perceived as unpleasant and maybe distressing or harmful by the targeted person, such presumed awareness on the part of the perpetrator is usually enough to classify the behavior as aggressive. Such assessments are also likely to be an important part of the targeted person’s evaluation of the situation and the motive(s) behind the behavior. Some aggression researchers also want to add that the behavior the target is exposed to should be “unwanted” by the target, in order to clearly differentiate it from potentially hurtful behaviors in a sadomasochistic context. With regard to intentionality, an exception is also made for potentially hurtful behaviors when used in the context of a “social role” such as that of dentists or surgeons who intentionally “inflict injury and discomfort” upon their patients but where there is typically no reason to assume an underlying intent/desire to hurt or harm (Buss 1961).
Some Repetitiveness
A major reason why I introduced the criterion of repetitiveness was that one can then be more certain that the negative behavior is intended. However, I never thought of this as an absolutely necessary criterion (Olweus 1993, 1999a). In the definition of bullying presented in the OBQ, the original formulation is “When we talk about bullying, these things happen repeatedly,” but the response alternatives also include “once or twice” (in the past couple of months). In some recent versions of the questionnaire, including the one used in Norway, we have made a slight modification by stating that “these things may happen repeatedly” or “are usually repeated,” depending on what is most natural with the relevant language.
Power Imbalance
The third criterion of a power imbalance favoring the perpetrator(s) is quite important and also closely associated with the general issue of who is to define when bullying has occurred or occurs. In my view, the ultimate “power of definition” must reside with the targeted student—at least as a basic point of departure. Given the sometimes complex and subtle behaviors involved in bullying and maybe a particular historical personal background for the behavior, it is often very difficult for a “reasonable outside” person (cf. Smith et al. 2012) to assess the situation correctly. At the same time, it is clear that a closer investigation of the situation may sometimes reveal that things are more complex than they appear on the surface and that perspectives from the outside may be very helpful in finding solutions to the problematic situation. In this context, however, one must
www.annualreviews.org • School Bullying 757
A nn
u. R
ev . C
lin . P
sy ch
ol . 2
01 3.
9: 75
1- 78
0. D
ow nl
oa de
d fr
om w
w w
.a nn
ua lr
ev ie
w s.
or g
by S
ta te
U ni
ve rs
ity o
f N
ew Y
or k
- B
uf fa
lo o
n 07
/2 7/
13 . F
or p
er so
na l u
se o
nl y.
CP09CH27-Olweus ARI 25 February 2013 18:29
also be aware of the tendency of some schools (as potential judges) to portray, when confronted, even obviously abusive situations as “regular peer conflicts” between equals as a way of escaping responsibility.
What this implies is that we are here dealing with a power imbalance as perceived by the targeted youth. The perceived imbalance is likely to be associated with such objective factors as physical strength or difference in numbers but may also relate to differences in self-confidence, popularity/status in the peer group, and the like. One important aspect here is very likely the targeted student’s perception of how difficult it would be for him/her to defend him/herself in a reasonable way and maybe with some success. Another aspect is of course how the student perceives the motives behind the behavior s/he is exposed to, whether or not it is perceived as aiming to hurt and create problems for him/her.
What About the Three Criteria as Regards Cyber Bullying?
The issue of whether and how the three key criteria of traditional bullying can be applied to cyber bullying is discussed in considerable detail in a recent chapter by Smith et al. (2012). After thought- ful consideration of a number of different aspects, the authors reach the tentative conclusion that the three key criteria defining traditional bullying are largely applicable to cyber bullying as well. They suggest that the imbalance of power can be assessed “in terms of differences in technological know-how between perpetrator and victim, relative anonymity, social status, number of friends, or marginalized group position” (p. 36). In partial contrast to my position stated above, however, they do not think the imbalance of power should be based on the victim’s perception but rather from an “outside” perspective as much as possible.
It is of vital importance for the building of a solid and cohesive body of knowledge about bullying that researchers can reach and use in their research a reasonably common conceptual definition of the phenomenon. But empirical analyses can also help inform such a development. In the following analyses, the main focus is on whether the ordinary response alternatives of the OBQ can be meaningfully used also with cyber bullying items.
The aim of these analyses was to examine if cyber bullying items (being bullied) would relate differently than the global traditional bullying items to a variable that these items can be expected to correlate with (Olweus 2012a). Poor self-esteem is one variable that has been found to be linearly related to the five frequency response alternatives of the global being bullied question in the OBQ (varying from “not bullied in the past couple of months” to “several times a week”; Solberg & Olweus 2003). The background for this comparison is that some forms of cyber bullying can refer to single—and not repeated—events but still be shared with a large group of people and be read again and again. Being exposed to such cyber behavior can certainly be very distressing for the target, and it might be considered a “misclassification” if the targeted student were categorized as “not being cyber bullied” because the event happened only once (with the common “2 or 3 times a month” as a lower-bound criterion for classifying a respondent as being bullied). If such misclassifications occurred for several episodes of cyber bullying, one would not expect a regular linear increase in poor self-esteem with increasing frequency of cyber bullying but rather some different pattern with an elevated level of poor self-esteem for the “once or twice” category.
This was clearly not the result of our empirical analyses, where we found a linear-increasing re- lation between the five response alternatives for the general being cyber bullied question and poor self-esteem. Students who were exposed to cyber bullying more often tended to have systematically poorer self-esteem. This result is very similar to what has been found in previous analyses of tradi- tional global being bullied questions. Similar results were also obtained when the relation between
758 Olweus
A nn
u. R
ev . C
lin . P
sy ch
ol . 2
01 3.
9: 75
1- 78
0. D
ow nl
oa de
d fr
om w
w w
.a nn
ua lr
ev ie
w s.
or g
by S
ta te
U ni
ve rs
ity o
f N
ew Y
or k
- B
uf fa
lo o
n 07
/2 7/
13 . F
or p
er so
na l u
se o
nl y.
CP09CH27-Olweus ARI 25 February 2013 18:29
five different means or channels of being bullied—by email, instant messaging, chat room, website and mobile/cell phones—and poor self-esteem was examined (for details, see Olweus 2012a,b).
In this context, it is also worth pointing out that some forms of traditional bullying, for example, that of having false rumors spread, are somewhat similar to the special forms of cyber bullying with regard to the implicated spreading mechanisms (Smith et al. 2012, Ybarra et al. 2012). For false rumors (and other traditional forms) there was also a regular linear relation between the five frequency categories and poor self-esteem. Applying the standard criterion of some repetition for defining somebody as being bullied thus appears to have functioned quite well also for this form of (indirect) bullying. What these results indicate is that cyber bullying items seem to function in roughly the same way as items on traditional bullying—which could be a desirable result from the perspective of simplicity or parsimony. It must be emphasized, however, that only one “external” psychosocial adjustment variable has been used in these analyses, poor self-esteem. This variable is no doubt a meaningful and important variable for this kind of analyses, but to find out if the obtained results can be generalized more broadly, it is obviously necessary to use a larger set of external variables and more channels or means of cyber bullying. There is clearly a need for more research on such issues.
SOME MEASUREMENT ISSUES
Distinctions Between Bullying Victimization and General Victimization and Between Bullying Perpetration and General Aggression
As mentioned, bullying is conceived of as a special form of aggressive behavior in the context of a power-imbalanced relationship. From a conceptual point of view, the most important charac- teristic distinguishing bullying from generally aggressive behavior—often without an identified or consistent target—is no doubt the power imbalance between the perpetrator(s) and the target. We have repeatedly argued on purely conceptual grounds that not including the power imbalance in the measurement procedures in one way or another is likely to have some unfortunate con- sequences (Olweus 2010, Solberg et al. 2007). One consequence relates to the fact that students who themselves initiate many aggressive interactions are likely to be exposed to aggressive acts from their opponents also when they are clearly the winners of the aggressive interaction. These students will then correctly report that they have been exposed to aggressive acts and as a con- sequence also be included in the total group of victims (including bully-victims). However, such students are likely to be partly different from students who have been exposed to the aggressive acts in the context of a bullying relationship with a clear power imbalance. To classify such aggressive students as victims would also increase the overlap between bullies and victims as well as the cor- relation between victimization and aggression/bullying variables in dimensional analyses, possibly leading to largely incorrect conclusions that bullies and victims are the same students, as has been reported in some research (see Solberg et al. 2007). Generally, the distinctions mentioned above also lead naturally to a separation of three groups of key actors involved in bully/victim problems, representing very different reaction patterns and personality profiles: pure bullies or bullies only, pure victims or victims only, and bully-victims (Olweus 1978, 1993; Solberg et al. 2007, 2010).
There may also be unfortunate consequences when it comes to intervention. General aggressive behavior is sometimes interpreted as a natural and maybe healthy response to an oppressive society or a disadvantaged situation, whereas being a victim of bullying by powerful peers constitutes a violation of fundamental human rights (Olweus 1993). A school’s moral obligation to counteract and prevent the latter kind of malignant behavior will certainly be much more compelling than to reduce aggressive behavior in general.
www.annualreviews.org • School Bullying 759
A nn
u. R
ev . C
lin . P
sy ch
ol . 2
01 3.
9: 75
1- 78
0. D
ow nl
oa de
d fr
om w
w w
.a nn
ua lr
ev ie
w s.
or g
by S
ta te
U ni
ve rs
ity o
f N
ew Y
or k
- B
uf fa
lo o
n 07
/2 7/
13 . F
or p
er so
na l u
se o
nl y.
CP09CH27-Olweus ARI 25 February 2013 18:29
Does the Power Imbalance Make a Difference?
But does it make a difference if the power imbalance aspect is included in the empirical mea- surement procedures? In the OBQ, the power imbalance implied in bullying is introduced as a qualifying part of the general definition presented to the students (see definition above). Based on the picture of the pure victims/victims only generated from the students’ responses (Solberg et al. 2010), it is our clear impression that the students have an understanding of the difference between being exposed to aggressive behavior in a power-imbalanced relationship, that is, being bullied, and being exposed to (recurrent) aggressive acts in a fight or conflict between students of roughly equal power or strength. However, we have not ourselves studied the problem systematically from an empirical point of view.
In a relatively recent paper, Hunter et al. (2007) took a more direct approach to the power imbalance issue. In their study of approximately 1,400 students in the 8 to 13 years age range, the researchers asked the participants to indicate via self-report how often they had been exposed to a number of aggressive behaviors in the past two weeks. In addition, the participants were asked to indicate if the aggressor(s) in question was more “powerful” as reflected in greater physical strength, higher popularity, or a situation in which the perpetrator was part of a group. Victims of bullying, that is, students exposed to aggression in a power-imbalanced relationship, perceived significantly more threat and less control over their situation in addition to being more depressed, engaging in more wishful thinking, and seeking more social support than the other group. In conclusion, the authors emphasized the importance of making a distinction between peer-victimized students and bullied students with the presence of a power imbalance as the differentiating criterion. The results of this study clearly suggest that, from the perspective of the targeted students, bullying is a more serious and hurtful form of peer aggression.
A couple of conclusions from a paper by Ybarra and colleagues (2012) are also relevant in this context. In two different studies reported in the same article, members from an online panel were randomized to different conditions varying the form of administration of a questionnaire on bullying and cyber bullying. An important result was that using “a behavioral list of bullying experiences without either a definition or the word ‘bully’ results in higher prevalence rates and likely measures experiences beyond the definition of ‘bullying’ ” (Ybarra et al. 2012, p. 5). The prevalence data for the various forms were also checked for possible misclassification, and a follow- up question about the power imbalance between the perpetrator(s) and the target was clearly most effective in reducing misclassification.
It is also interesting to note that in a meta-analysis of predictors of bullying and victimization in childhood (Cook et al. 2010), there was a clearly significant moderator effect of “measurement approach” (table 8, p. 76): The “bully group” was (also) characterized by internalizing problems when measurement was performed via behavioral descriptors of general aggression (average effect size r = 0.17) rather than by specific reference to bullying (average effect size r = 0.02). (Sur- prisingly, however, and in conflict with these results, the discussion section in the paper concludes that “effect sizes were not influenced by measurement approach,” Cook et al. 2010, p. 78.) These moderator results are consistent with the arguments presented above about the increased overlap between bullies and victims when bullying is measured with general aggression items and without its defining criteria.
Generally, although there is a need for more empirical research on these issues, conceptual arguments and the reported empirical findings so far clearly support the usefulness of a distinction between a bullying perpetration/bullying victimization research line on the one hand and a general aggression/general victimization line on the other (also see an early identification of this distinction in Schäfer et al. 2002). Victims of bullying are likely to overlap in part with and can be seen as
760 Olweus
A nn
u. R
ev . C
lin . P
sy ch
ol . 2
01 3.
9: 75
1- 78
0. D
ow nl
oa de
d fr
om w
w w
.a nn
ua lr
ev ie
w s.
or g
by S
ta te
U ni
ve rs
ity o
f N
ew Y
or k
- B
uf fa
lo o
n 07
/2 7/
13 . F
or p
er so
na l u
se o
nl y.
CP09CH27-Olweus ARI 25 February 2013 18:29
a subgroup of peer-victimized students with special characteristics. Similarly, students who bully others are likely to overlap in part with and can be seen as a subgroup of aggressive students with special characteristics. Although there is a good deal of overlap between the subgroups, they are definitely not identical. At this point in time, it is essential to be clear about these likely differences, to be more correct and precise in describing and interpreting the results of empirical studies, and not to use the relevant terms as synonymous without empirical support. With regard to the latter point, authors should be careful not to present results as applying to bullying when they actually have measured general victimization and general aggression (also see Hunter et al. 2007). A special case in point is when the results from studies based on questionnaires designed to measure general aggression are reported as being research on bullying or bully/victim problems. Such practices will only create confusion and prevent building of a cohesive and consistent knowledge base.
SOME MEASUREMENT ISSUES RELATED TO THE OLWEUS BULLYING QUESTIONNAIRE
Reliability
According to Cornell & Bandyopadhyay (2010), the OBQ “is the most important and widely used bullying survey worldwide, yet there is little published information about its reliability and valid- ity.” Although issues of the reliability and validity of the OBQ have certainly not been neglected or understudied by me and my colleagues, the comment about the lack of published information has some merit. Accordingly, it is appropriate to briefly report about relatively recent studies with such information.
With regard to the reliability of the bullying perpetration and bullying victimization scales of the eight or nine various forms of bullying included in the OBQ, there are now available seven empirical studies with such information from independent researchers from several different countries (Fekkes et al. 2005, Felix & McMahon 2006, Ferguson et al. 2009, Hartung et al. 2011, Jenson et al. 2010, Kyriakides et al. 2006, Stavrinides et al. 2010, Strohmeier et al. 2011, Theriot et al. 2005). Results have been reported in the form of internal consistency coefficients like Cronbach’s alpha or item response theory (IRT) technology and have been uniformly quite good with reliabilities in the 0.80–0.90 range. Similar findings are obtained in our own IRT analyses for the bullying others scale based on a very large sample of Norwegian students (n = 48,926; Breivik & Olweus 2012).
As reported in Solberg & Olweus (2003), the correlation between school-level prevalence estimates based on the dichotomized global being bullied and similar estimates based on the (in this study, seven) specific forms of being bullied was quite high, r = 0.79. The corresponding correlation for school-level prevalence estimates based on the dichotomized global and specific bullying others items, respectively, was 0.77. After application of the Spearman-Brown formula (Guilford 1956) to estimate the school-level reliability of the combined prevalence estimates, these correlations approached 0.90, indicating that such estimates did an excellent job in differentiating the schools from one another. These results have now been largely replicated in the recent Finnish KiVa project (Kärnä et al. 2011a). The results also attest to the usefulness of reporting prevalence data based on the dichotomized (in contrast to continuous) global being bullied and bullying others items to the schools. As is well known, the use of dichotomized variables may imply some loss of information, but the choice of dichotomized versions for prevalence estimation (percentages) has been largely dictated by considerations of the need to give the schools feedback that is easy to understand and helpful in the planning of their intervention work (Solberg & Olweus 2003).
www.annualreviews.org • School Bullying 761
A nn
u. R
ev . C
lin . P
sy ch
ol . 2
01 3.
9: 75
1- 78
0. D
ow nl
oa de
d fr
om w
w w
.a nn
ua lr
ev ie
w s.
or g
by S
ta te
U ni
ve rs
ity o
f N
ew Y
or k
- B
uf fa
lo o
n 07
/2 7/
13 . F
or p
er so
na l u
se o
nl y.
CP09CH27-Olweus ARI 25 February 2013 18:29
Convergent Validity
There has been no systematic review of studies that have used both the OBQ and some form of peer nominations. However, in an unpublished meta-analysis presented at the Society for Re- search on Child Development symposium in Tampa, Florida in 2003, Card (2003) reported on 21 studies of the correlation between self- and peer reports of victimization. Several of the studies included had used the OBQ. The peer reports were both nominations and ratings. The aver- age correlation across the 21 studies was 0.37. Considering the fact that the studies included in the meta-analysis were of varying quality and several of them were not particularly designed to maximize correspondence between data from the two sources, this is a quite respectable result.
In the Finnish KiVa project (Kärnä et al. 2011a), data were obtained both with the OBQ and peer ratings (n.b. not nominations) about physical, verbal, and indirect bullying victimization and similar variables about participation in bullying perpetration (Salmivalli et al. 1996). For the OBQ, both the global being bullied/bullying others items and a scale of seven different items reflecting the various forms of bullying were used in the analyses. The three peer ratings were averaged into two summary scores, with a reliability of 0.84 for being bullied and 0.90 for participation in bullying. The correlations between the global item and the scale of being bullied versus the peer ratings of bullying victimization were 0.39 and 0.40, respectively. Corresponding correlations for the global bullying others item and the scale of bullying with the relevant peer ratings were 0.36 and 0.34. It is interesting to note that the results for the single global items of the OBQ had equally strong correlations with the peer ratings as the scales representing the various forms of bullying (cf. Fuchs & Diamantopoulus 2009). These results were obtained on very large samples of students in grades 3–12 (n = 19,780) before implementation of the KiVa intervention program (C. Salmivalli, personal communication). With correction for unreliability (attenuation correction), the convergent validity correlations would be even higher (Guilford 1956).
These results are in stark contrast to the results obtained in some studies conducted by Cornell and his associates (Branson & Cornell 2009, Cornell 2010, Cornell & Brockenbrough 2004, Lee & Cornell 2010), where the correspondence between self-reports and standard peer nominations of victim and bully status have been much lower, in the 0.10–0.20 range in some cases. There may be several reasons for this discrepancy, but in the most directly relevant of these studies, with the title “Concurrent Validity of the Olweus Bully/Victim Questionnaire” (Lee & Cornell 2010), the sample used was exceptionally small considering its purpose, only 202 students. The very small sample size resulted in the self-identification of only two bullying students when the standard criterion of “2 or 3 times a month” (Solberg & Olweus 2003) was applied. There were slightly more self-reported victims (n = 13), and for being bullied, the correlation with peer nominations was actually, and somewhat surprisingly, as high as 0.42, much in line with the results just reported. In addition, as also acknowledged by the authors of all cited papers, the schools included in the research studies were all in the process of using the Olweus Bullying Prevention Program (OBPP), which may have compounded the assessment situation further. Under such conditions, strong claims about the inadequacy or lack of validity of the OBQ or self-reports about bullying and victimization in general cannot be taken quite seriously.
Summing up, the correlations in the Card meta-analysis and the KiVa project were in the 0.35– 0.40 range, maybe with somewhat lower values for bullying perpetration. This result is clearly higher than the average association between self-reports and peer data described in a well-known meta-analysis of cross-informant reports on child behavior problems (r = 0.26; Achenbach et al. 1987), for example.
762 Olweus
A nn
u. R
ev . C
lin . P
sy ch
ol . 2
01 3.
9: 75
1- 78
0. D
ow nl
oa de
d fr
om w
w w
.a nn
ua lr
ev ie
w s.
or g
by S
ta te
U ni
ve rs
ity o
f N
ew Y
or k
- B
uf fa
lo o
n 07
/2 7/
13 . F
or p
er so
na l u
se o
nl y.
CP09CH27-Olweus ARI 25 February 2013 18:29
FACTORS CONTRIBUTING TO LACK OF CORRESPONDENCE AND PROBLEMS WITH PEER NOMINATIONS
Although the reported results indicate that the correspondence between self-reports about bully/victim status with the OBQ and peer ratings/nominations concerning similar dimensions can be considered rather satisfactory, the overlap is of course far from perfect. And there are good grounds for not expecting such an overlap. Some of these reasons are considered in the next sections, in which I have deliberately chosen to focus mainly on problems with peer nominations. The reason for this particular focus is that peer nominations is a very popular method, in particular in the United States, for collecting information about peer relations including bullying and that the methodological problems associated with its use for several purposes have received little research attention.
In a typical peer nomination procedure, students are presented with a roster of the names (and possibly pictures) of their classmates and asked to nominate a fixed number, often three, or an unlimited number of usually same-gender peers in their classroom or grade who fit one or several descriptions of victimization/being bullied and aggression/bullying other students (Who is the kid who. . . ; He/she is picked on. . . ; Find the names of three classmates who. . .). The number or usually some percentage of nominations received is used as the student’s score on the relevant dimension. The scores are often standardized within gender, grade, and/or classroom.
One likely reason for discrepancies between self-reports and peer nominations is the fact that a good deal of bullying is of a subtle and somewhat secretive nature that may be difficult for peers to observe but is clearly perceived by the targeted student and, accordingly, likely to be reported in the OBQ.
Second, in the OBQ the students report on the frequency with which they have been bul- lied/bullied other students, whereas the peer nomination method measures frequency of nomina- tions of “extreme students,” not the frequency or seriousness of the implicated behaviors. It may be reasonable to assume that a high proportion of the nominations that more extreme individuals are likely to receive actually reflects some frequency/seriousness (and maybe degree of visibility) dimension, at least roughly. However, for less extreme individuals, it is by no means evident that an average number of nominations, for example, directly translates into an average level or fre- quency/seriousness of problem behaviors. Also, in most peer nomination studies a considerable proportion of students receive no nominations at all (e.g., Espelage et al. 2003), and individu- als with zero scores cannot without further analyses be assumed to have identical low levels of the characteristic in question. In addition, it can be shown that the variance and distribution of nominations for a classroom (or grade) are substantially influenced by the number of nominators (classroom size) and the degree of internominator agreement. As a consequence, behaviorally fairly nonextreme students also may well be selected into the extreme group (for example, students with a standard score +1 above the mean) from relatively smaller classrooms and/or classrooms with poor internominator agreement (typically with smaller variance). The effects of such mechanisms, which are not well described and understood, are also likely to reduce correspondence with self- report data. In this context, it should be emphasized that peer ratings in which all students in the classroom or other relevant group usually are being assessed escape several of the problems with nominations mentioned in this paragraph.
Third, as has been documented in our studies (Olweus 1993, 1999b), a good deal of bullying is carried out by older students toward younger ones. To be bullied by older students will be captured in the key questions in the OBQ but is less likely to be registered in the peer nominations, which typically only refer to students in the nominators’ own classroom or grade.
www.annualreviews.org • School Bullying 763
A nn
u. R
ev . C
lin . P
sy ch
ol . 2
01 3.
9: 75
1- 78
0. D
ow nl
oa de
d fr
om w
w w
.a nn
ua lr
ev ie
w s.
or g
by S
ta te
U ni
ve rs
ity o
f N
ew Y
or k
- B
uf fa
lo o
n 07
/2 7/
13 . F
or p
er so
na l u
se o
nl y.
CP09CH27-Olweus ARI 25 February 2013 18:29
Fourth, as has also been documented in our studies (Olweus 1993, 1999b), a good deal of bullied girls report that they are mainly bullied by boys. Such cross-gender bullying will be captured in the key questions in the OBQ but probably not in peer nominations that are often restricted to same-gender nominees. In particular, girls bullied by boys and boys bullying girls may not be well identified under peer nomination conditions.
Fifth, the common practice of statistically standardizing peer nominations within class- room, gender, and/or grade will often result in the removal of meaningful between-classroom/ gender/grade variance. To illustrate, if there is a marked difference in the number/proportion of nominations of bullying (extreme) students for grade 4 and grade 5 and these data are stan- dardized within grade, this means that this developmental difference is effectively eliminated. The two distributions of standardized nominations will have the same mean values and standard deviations in spite of the fact that one grade has a much higher level of problems with bul- lying. As a related consequence, the most extreme students in each grade will receive roughly similar standardized values even if the extreme students in the most aggressive grade have re- ceived larger numbers/higher proportions of nominations. Since such differences in the level of problems are likely to be captured in self-reports, the practice of standardizing also may reduce the degree of correspondence between self- and peer reports. Results similar to those achieved through statistical standardization are likely to be obtained by procedurally restricting nomina- tions to students of the same category (classroom/gender/grade) in combination with use of a fixed number of nominations (e.g., Find the three same-gender peers in the class who fit the description. . .”).
In sum, there is actually a good deal of overlap between data from self-reports and peer nom- inations as evidenced by correlations in the 0.35 to 0.40 range. As a more detailed examination reveals, however, there are also clear differences in what the two methods are likely to measure and in how the raw data are used or “transformed” to generate the variables of interest. According to the analysis presented here, most of the reasons for the discrepancies are linked to characteristics of the peer nomination method. Nonetheless, with an acceptable level of internominator agreement and a large enough group of nominators, peer nominations may provide reasonably adequate mea- sures of relatively stable, individual-differentiating bully/victim characteristics, at least as regards more visible or “public” forms of bullying and the more extreme individuals in the peer group. Generally, it is likely that the possible strength of the typical peer nomination method lies more in the identification of distinct extreme groups than in creating distributions of students roughly arranged/ranked according to frequency or seriousness on the problematic behavior dimensions. The latter fact is also the reason why peer nomination data are not well suited for prevalence estimation and for comparisons of prevalence estimates or mean levels across groups and time points (for details, see Olweus 2010, pp. 20–21). Against this background, it is obvious that peer nomination data cannot be considered some kind of gold standard or ultimate criterion of the validity of self-reports on bully/victim problems.
One final point about the limitations of the peer nomination methodology concerns the mea- surement of change. Both peer nominations and self-report instruments such as the OBQ have been used for the measurement of possible effects of antibullying interventions, that is, to measure change. How do these two methods compare in that regard?
Unfortunately, the very nature of the peer nomination procedure will tend to discount possible changes in the level of problems over time. Even if the overall level of problems has decreased substantially in a school after an intervention, many bullying students nominated one year or maybe six months earlier will most likely be nominated again as bullying other students. This is simply because they still are the students who fit the descriptions best even if most of them have considerably reduced their bullying behavior at follow-up. The same will probably apply to
764 Olweus
A nn
u. R
ev . C
lin . P
sy ch
ol . 2
01 3.
9: 75
1- 78
0. D
ow nl
oa de
d fr
om w
w w
.a nn
ua lr
ev ie
w s.
or g
by S
ta te
U ni
ve rs
ity o
f N
ew Y
or k
- B
uf fa
lo o
n 07
/2 7/
13 . F
or p
er so
na l u
se o
nl y.
CP09CH27-Olweus ARI 25 February 2013 18:29
students who have been bullied. The effects of such mechanisms are likely to be accentuated if the students are given instructions to nominate a fixed number of peers, which will tend to serve as a kind of standardization within classroom/grade/time point and will largely reduce or eliminate possibilities for registering change.
Another important difference is that in the questions of the OBQ, for example, it is repeatedly emphasized that the responses concern “the past couple of months.” The absence of such a refer- ence period in most peer nomination methods reinforces the focus on stable or typical behavior patterns or situations rather than change.
In contrast, the self-report questionnaire measures mostly painful subjective experiences of being bullied, and if clear changes occur in the levels of bullying, this is likely to be quickly registered by the targeted student. Similarly, students who bully other students will tend to note if their behavior is questioned, blocked, or confronted by teachers or peers, and maybe reported to parents.
Generally, it seems that the value of peer nominations has been exaggerated and the problems associated with the methodology correspondingly underrated. There is simply a strong need for much more methodological groundwork to find out if at all, and in which cases how, common peer nomination methods can be meaningfully used for prevalence estimation and the study of developmental changes including comparisons of prevalence estimates or mean values across groups and time.
In drawing these conclusions, I have disregarded the possible ethical problems associated with the use of peer nominations for socially undesirable behavior patterns including “like-least nomi- nations.” In several countries, including the Scandinavian countries and Australia, many research projects with such peer nominations very likely would be rejected for ethical reasons by the research evaluation committees.
ISSUES RELATED TO CYBER BULLYING
Over the past five or six years, newspapers, the popular press, and the research community, as well as educators and parents, have paid a good deal of concerned attention to a relatively new form of bullying among school children and youth, usually named cyber bullying or electronic bullying. As suggested in a previous section, cyber bullying is broadly defined as bullying performed via electronic forms of contact or communication such as mobile/cell phones or the Internet. The general picture created in the media—and often also by researchers and authors of books on cyber bullying—is that cyber bullying is very frequent, that it has increased dramatically over time, and that this new form of bullying has created many new victims and bullies in addition to the victims and bullies involved in traditional bullying. In addition, it is often argued or implied that cyber bullying is very difficult for adults to discover and counteract, creating a feeling of powerlessness in adults and maybe students as well.
In the following sections, I argue that claims about cyber bullying made in the media and elsewhere are often greatly exaggerated and that several such claims have very little scientific sup- port. My arguments are based on empirical analyses of several large-scale studies, two of which have cross-sectional or school-level longitudinal data concerning time trends in the United States and Norway, respectively, at four and five time points (for more details, see Olweus 2012a,b). Furthermore, I report some analyses about the possible effects of being exposed to cyber bul- lying. Finally, I also present some thoughts on how to counteract cyber bullying and suggest steps a school or community can take to further reduce and prevent cyber bullying among their students.
www.annualreviews.org • School Bullying 765
A nn
u. R
ev . C
lin . P
sy ch
ol . 2
01 3.
9: 75
1- 78
0. D
ow nl
oa de
d fr
om w
w w
.a nn
ua lr
ev ie
w s.
or g
by S
ta te
U ni
ve rs
ity o
f N
ew Y
or k
- B
uf fa
lo o
n 07
/2 7/
13 . F
or p
er so
na l u
se o
nl y.
CP09CH27-Olweus ARI 25 February 2013 18:29
17.2
4.3 5.0
15.4
18.1 18.4
4.1 4.5
0
2
4
6
8
10
12
14
16
18
20
22
2007 2008 2009 2010 2007 2008 2009 2010
Verbal Cyber
St ud
en ts
b ei
ng b
ul lie
d (p
er ce
nt )
Year
Figure 1 US time series data for 2007–2010 for verbal bullying (being bullied) and cyber bullying (being bullied electronically). Data from all over the United States. Total n = 447,000.
Claims About Prevalence and Time Trends
The first two major media and, in part, researcher claims to be scrutinized assert that cyber bullying is a very frequent phenomenon among today’s children and youth and that the frequency or prevalence of the phenomenon has increased dramatically in recent years.
One of the obvious reasons why some researchers (several of whom seem to come from fields other than psychology) have reported high or very high prevalence figures of cyber bullying is that cyber bullying has been studied “in isolation,” that is, outside a general context of (traditional) bullying and often also without a general student-friendly definition of what is meant by bullying. To put cyber bullying in proper perspective, it is in my view necessary to study it in the context of (traditional) bullying more generally. As implied in previous sections of this review, one cannot talk about a phenomenon as bullying unless a reasonably precise definition has been provided to the respondents or the formulation of the questions or other measures used make it quite clear that the contents conform to what is usually implied in bullying.
The results that are obtained when cyber bullying is studied in the broader context of other forms of bullying (as with the Olweus Bullying Questionnaire) are shown in Figures 1 and 2. Figure 1 illustrates the results for a very large US sample (with approximately 440,000 students) for being exposed to direct verbal bullying and to cyber bullying, respectively, for the four consecutive years between 2007 and 2010. The average across-time prevalence for being verbally bullied is 17.6, and the corresponding figure for being cyber bullied is 4.5%. The average for bullying others verbally is 9.6%, whereas the corresponding figure for cyber bullying others is 2.8% (Figure 2).
The Norwegian data with prevalence figures from 41 schools over a five-year period from 2006 to 2010 show a very similar pattern of results but at somewhat lower prevalence levels (Olweus 2012a).
With regard to the second claim concerning the presumably dramatic increase of cyber bul- lying over time, both figures (as well as the Norwegian data) show that basically no systematic change in prevalence has occurred over the time periods studied. This is true of both being cyber bullied and cyber bullying others as well as of being bullied and bullying others with direct verbal means.
766 Olweus
A nn
u. R
ev . C
lin . P
sy ch
ol . 2
01 3.
9: 75
1- 78
0. D
ow nl
oa de
d fr
om w
w w
.a nn
ua lr
ev ie
w s.
or g
by S
ta te
U ni
ve rs
ity o
f N
ew Y
or k
- B
uf fa
lo o
n 07
/2 7/
13 . F
or p
er so
na l u
se o
nl y.
CP09CH27-Olweus ARI 25 February 2013 18:29
St ud
en ts
b ul
ly in
g ot
he rs
(p er
ce nt
)
9.6
2.5 3.2
9.3 10.0 9.3
2.9 2.7
0
2
4
6
8
10
12
14
16
18
20
22
2007 2008 2009 2010 2007 2008 2009 2010
Verbal Cyber
Year
Figure 2 US time series data for 2007–2010 for verbal bullying (other students) and cyber bullying (other students). Data from all over the United States. Total n = 440,000.
As documented here, cyber bullying is actually a quite low-prevalent phenomenon, representing only some 25% to 35% of the level of traditional bullying with direct verbal means. It is obvious that the “psychological threshold” for endorsing the global items on cyber bullying is much higher than for direct verbal bullying (Breivik & Olweus 2012). The two global questions about cyber bullying are actually among the various bullying items/forms with the lowest prevalence rates. There is no doubt that many more children and youth are involved in traditional verbal bullying than in cyber bullying.
Similarly, and in spite of increasing accessibility and use of mobile phones and computers, there were no indications of increases in prevalence of cyber bullying over time, both as regards being bullied and bullying other students. The same was true of direct verbal bullying.
Degree of Overlap of Traditional Bullying with Cyber Bullying
A third common claim made by the media and some researchers alike is that the new form of cyber bullying has created many new victims and perpetrators of bullying. This claim is based on an assumption and maybe some empirical data to suggest that children and youth who are involved in cyber bullying are to a considerable degree different from those engaged in traditional bullying.
When this claim was checked empirically (Olweus 2012a), the results documented a very high degree of overlap: Of students who had been exposed to cyber bullying in the US sample, 88% had been bullied in at least one traditional way. For cyber bullying others, the overlap also was 88%. The results for the Norwegian schools were similar, with degree of overlap being 93% and 91%, respectively.
In these analyses, there was thus only a very small percentage, about 10% of the participants, who had only been cyber bullied or had only cyber bullied others. These results suggest that the new electronic media have actually created few new victims and bullies. To be cyber bullied or to cyber bully other students seems to a large extent to be part of a general pattern of bullying, where use of the electronic media is only one possible form and, in addition, is a form with a quite low prevalence.
www.annualreviews.org • School Bullying 767
A nn
u. R
ev . C
lin . P
sy ch
ol . 2
01 3.
9: 75
1- 78
0. D
ow nl
oa de
d fr
om w
w w
.a nn
ua lr
ev ie
w s.
or g
by S
ta te
U ni
ve rs
ity o
f N
ew Y
or k
- B
uf fa
lo o
n 07
/2 7/
13 . F
or p
er so
na l u
se o
nl y.
CP09CH27-Olweus ARI 25 February 2013 18:29
These results also suggest that even if most cyber bullying actually occurs outside school hours, as has been documented in several other surveys, many—very likely, most—episodes of cyber bullying originate in the school setting.
Are There Negative Effects of Cyber Bullying, and How Do We Find Out?
There have been many reports in the research literature and the media about the negative ef- fects of being exposed to cyber victimization—but they usually do not take into account possibly concurrent bullying victimization in “traditional” ways. Since a majority or maybe most cyber- bullied children are also exposed to traditional bullying, as discussed in the previous section, it is difficult to know if, and to what extent, the reported negative effects actually are a consequence of cyber bullying, over and above the well-documented negative effects of traditional bullying. In the article referred to (Olweus 2012a), I presented two sets of analyses with poor self-esteem as the outcome variable and concluded tentatively that the issue of possible negative effects of being cyber bullied may be context specific: If a student was bullied only (or mainly) via electronic means (which was the case for a small group of subjects), this seemed to have a negative effect on his or her psychosocial adjustment or well-being. If the student was exposed to both traditional (to some non-negligible degree) and cyber bullying, however, the additional effect of cyber bullying seemed to be trivial. To find out about the possible negative effects of cyber victimization is a complex and challenging research task, and my conclusion above should certainly be considered preliminary and tentative.
Unfortunate Consequences of the Distorted Media Picture
Getting back to the media picture, I want to briefly touch on two likely unfortunate consequences. First, such a distorted portrayal of reality will probably generate a lot of unnecessary anxiety and tension among parents and maybe teachers and students. It may also create feelings of pow- erlessness and helplessness in the face of the presumably “huge” and ubiquitous cyber bullying problem.
Second, such a picture is likely to result in an unfortunate shift in the focus of antibullying work if digital bullying is seen as the key bullying problem in the schools. This would probably also result in funneling a lot of resources in a wrong direction while traditional bullying—which is clearly the most prevalent and most serious problem—would be correspondingly downgraded.
But should the results and conclusions be interpreted to mean that we should just disregard and stop bothering about cyber bullying among children and youth? No, this would not be a helpful line of action. Since both intuition and some empirical research asking about the perceived impact of hypothetical cyber bullying exposures (Smith et al. 2008) clearly suggest that some forms of bully- ing such as posting painful or embarrassing pictures or videos may have markedly negative effects, it is important also to take cyber bullying seriously both in research and intervention/prevention. It has not been my intention to trivialize or downplay cyber bullying, but I definitely think it is necessary and beneficial to place cyber bullying in proper context (with traditional bullying) and to have a more realistic picture of its prevalence and nature.
How Can We Counteract and Prevent Cyber Bullying?
Very briefly, technologically oriented researchers and many book authors tend to emphasize the importance of teaching students, parents, and educators various aspects of “netiquette” concerning such things as Internet safety, how the different technologies function, and how to behave properly
768 Olweus
A nn
u. R
ev . C
lin . P
sy ch
ol . 2
01 3.
9: 75
1- 78
0. D
ow nl
oa de
d fr
om w
w w
.a nn
ua lr
ev ie
w s.
or g
by S
ta te
U ni
ve rs
ity o
f N
ew Y
or k
- B
uf fa
lo o
n 07
/2 7/
13 . F
or p
er so
na l u
se o
nl y.
CP09CH27-Olweus ARI 25 February 2013 18:29
on the Internet. It is probably useful for most students, parents, and educators today to achieve a certain level of such basic knowledge about the new technologies. And many schools could certainly benefit from introducing and strictly enforcing clear rules about the use of cell phones and computers or the Internet in the schools. At the same time, there seem to be basic limitations to what can be achieved with such a technological approach, according to a recent small-scale meta-analysis (Mishna et al. 2011).
Given that traditional bullying is much more prevalent than cyber bullying and that the great majority of cyber-bullied students are also bullied in traditional ways, it is natural to recommend schools to direct most of their efforts to counteracting traditional bullying, preferably using a program with documented effects (cf. Ttofi & Farrington 2009). In some of our own large-scale intervention studies with relatively few measures focusing directly on cyber bullying (Olweus & Limber 2010), we have observed that levels of cyber bullying have gone down substantially in par- allel with reductions in traditional bullying. Such effects have also been convincingly documented in a recent paper from the Finnish KiVa project (Salmivalli & Pöyhönen 2011). The authors conclude that “reducing cyber bullying does not necessarily require programs tailored to target especially these specific forms of bullying” (p. 68).
Such a conclusion is certainly in line with the general thrust of the arguments presented above. At the same time, I think we should also make use of the possibility that attention to cyber bullying cases can lead to a disclosure of what actually goes on in terms of traditional bullying in the school context. In addition, the great focus on cyber bullying in the media and research can probably be used to revitalize societal interest in the phenomenon of school bullying more generally and the need to systematically address this pressing social problem.
A System-Level Proposal
Technological experts tell us that in most cases, it is quite possible, although maybe somewhat time consuming, to disclose the sender(s) of negative bullying messages communicated via electronic means. Accordingly, I have proposed that an important and presumably quite effective measure an individual school or community can take in counteracting and preventing cyber bullying is to invest time and technical competence in disclosing thoroughly a few identified cases of cyber bullying—and then communicate clearly and openly (but anonymously) the results to the students. This system-level strategy will probably substantially increase the perceived risk of disclosure and very likely be able to reduce further the already low prevalence of cyber bullying.
THE LONG-TERM OUTCOMES FOR BULLIES
Although myriads of studies have been conducted on the concurrent characteristics of bullies, victims, and bully-victims (see, e.g., Arseneault et al. 2010, Jimerson et al. 2010 for overviews), much less research has examined the long-term outcomes for these groups based on prospective studies. An early such study, often referred to as the first of its kind, was my study of three consecutive community cohorts of boys from Greater Stockholm (n = 780) who had participated in my first research project on bullying when they were in grades seven to ten (Olweus 1978, 1991, 1993, 2010). When the criminal records of these boys were followed-up in the official Swedish crime registers up to age 24, it was found that the former school bullies, as assessed by a combination of teacher nominations and peer ratings, were heavily overrepresented. Nearly 55% of the school bullies had at least one conviction (any kind of crime resulting in conviction) in the period from age 16 to 24 years, and 36% had three or more convictions, whereas only 9.8% of the students who had not been involved in bullying others had as many convictions. The odds
www.annualreviews.org • School Bullying 769
A nn
u. R
ev . C
lin . P
sy ch
ol . 2
01 3.
9: 75
1- 78
0. D
ow nl
oa de
d fr
om w
w w
.a nn
ua lr
ev ie
w s.
or g
by S
ta te
U ni
ve rs
ity o
f N
ew Y
or k
- B
uf fa
lo o
n 07
/2 7/
13 . F
or p
er so
na l u
se o
nl y.
CP09CH27-Olweus ARI 25 February 2013 18:29
ratio (OR) for the latter variable was 5.12. Thus, as young adults, the former school bullies had an almost fourfold increase in the level of serious, recidivist criminality, as documented in official registers. For violent crimes, the results were even more marked, with 11% of former bullies having two or more convictions in comparison with 1.6% of the nonbullies, the OR being 7.79.
Recently, a group of researchers from the Institute of Criminology at Cambridge University (Ttofi et al. 2011a) has conducted a meta-analysis of all prospective studies they have found on the relation between bullying perpetration and later criminal and antisocial behavior, including the Swedish study. They were able to identify 18 studies with an average follow-up time of 5.4 years (varying between 0.5 and 11 years). The average age at which bullying perpetration was measured was 11 years (varying between 6 and 16 years). The summary effect size was OR 2.60 (CI: 2.15– 3.14; z = 9.90). When controlling for a considerable number of early covariates (seven on the average), the adjusted summary effect size was reduced to OR = 1.82, but this was still a highly significant effect (CI: 1.55–2.13). Most of the studies were based on male participants, but females were included in a few of them (e.g., Renda et al. 2011). Results for females were also significant and in the same direction as for males, but effects were usually weaker. Interestingly, and related to the previous section on the validity of self-reports, several of the studies had assessed bullying perpetration with some version of the OBQ (e.g., Kim et al. 2011, Lösel & Bender 2011). The authors concluded generally that school bullying is “a strong and specific risk factor for later offending” (Ttofi et al. 2011a, p. 81).
In a few studies, bully status had actually been determined on the basis of a single self-report item (Renda et al. 2011) or a single item from the Child Behavior Checklist ( Jiang et al. 2011). This suggests that the findings reported in the meta-analysis very likely represent a conservative estimate of the predictive power of school bullying variables. This conclusion is further reinforced by the fact that several of the covariates (early aggressive behavior, for example) might be better regarded as a natural part of a bullying perpetration pattern than as possible confounds to be controlled. In this way, the covariate adjustment will thus likely represent a certain degree of overcontrol.
THE LONG-TERM OUTCOME FOR VICTIMS
The same group of researchers has conducted a similar meta-analysis of the relation between bul- lying victimization and depression (depressive states or symptoms) later in life (Ttofi et al. 2011b). In this case, 29 different studies were identified, and the summary effect size OR was 2.06 (CI: 1.79–2.38; z = 9.23). The average age at which bullying victimization was measured was 12 years (varying between 8 and 18 years), and the average length of the follow-up period was 6.9 years (varying between 1 and 36 years). With control for a number of early covariates (six on average), the summary effect size was reduced to OR = 1.74, but this was still highly significant (CI: 1.54–1.97; z = 8.77). Effect sizes were of similar magnitude for males and females. As with the meta-analysis for bullying perpetration, the reported results are likely to be conservative estimates due to possible overcontrol and limited reliability of some assessments of bullying victimization status.
These important meta-analyses show convincingly that being involved in such problems as a bully or a victim with some regularity is not just a harmless and passing school problem. Rather, they document that such involvement is likely to result in serious long-term consequences and adjustment problems for both groups, but of a different nature: Depression and other internalizing problems for the victims and criminality/antisocial behavior and other externalizing problems for the bullies. As emphasized in a previous section, being bullied by peers also represents a serious violation of the fundamental rights of the child or youth exposed (Olweus 1993, p. 48).
In addition to resulting in substantial personal suffering and difficulties for partners and family, these problems entail huge costs for society. It has been documented that both former bullies and
770 Olweus
A nn
u. R
ev . C
lin . P
sy ch
ol . 2
01 3.
9: 75
1- 78
0. D
ow nl
oa de
d fr
om w
w w
.a nn
ua lr
ev ie
w s.
or g
by S
ta te
U ni
ve rs
ity o
f N
ew Y
or k
- B
uf fa
lo o
n 07
/2 7/
13 . F
or p
er so
na l u
se o
nl y.
CP09CH27-Olweus ARI 25 February 2013 18:29
victims are “overconsumers” of society’s health and social support systems (Sourander et al. 2007). Due to long sick leaves, unemployment, and early pensioning, revenues from taxation of these groups will also tend to be greatly reduced. All of these likely outcomes underscore the necessity and value of counteracting and preventing these problems, not only for the sake of the victims, which is the main motivation behind interventions, but also for the sake of the bullies.
Two Swedish economists have estimated that the costs to society of a “marginalized life” (up to age 65) amount to approximately two million US dollars (Nilsson & Wadeskog 2008). Although the calculations are not directly based on the lives of bullies and victims, the estimates are likely to be roughly applicable also to a considerable proportion of such individuals. The economists conclude (in translation): “. . .in order for an (intervention) project to be economically profitable for society, the success rate need only be 1.5 percent” (Nilsson & Wadeskog 2008, p. 10).
EFFECTIVENESS OF ANTIBULLYING PROGRAMS
A few years ago, the Cambridge group also conducted a meta-analysis of the effectiveness of all antibullying programs in the world (Ttofi & Farrington 2009, 2011). They were able to locate 59 reports representing 30 programs with scientific evaluations of reasonable quality. As a general conclusion, the authors stated that antibullying programs are effective in reducing bullying and victimization by 20% to 23% in intervention schools as compared with control conditions (Ttofi & Farrington 2009). Although this meta-analysis no doubt is the most comprehensive and thorough of its kind available, the considerable variation in results among programs makes one wonder if a general statement about the expectable effect of “any antibullying program” (from a potential population of programs) is defensible. A closer look at the results suggests that the chances that some programs would result in positive effects seem very small. It should also be remembered that all programs in the meta-analysis except one (the OBPP, the Olweus Bullying Prevention Program with five replications) were efficacy, first-time (nonreplicated) studies, often with extra resources and support from the researchers conducting the study the result is based on (Flay et al. 2005; also cf. the article “The desperate need for replications”, Hunter 2001). Nonetheless, the overall results of this careful meta-analysis are clearly positive and indicate that at least some theory-guided, research-based antibullying programs very likely will lead to considerable reductions in bully/victim problems (such as the OBPP as emphasized by the authors; Ttofi & Farrington 2009, pp. 20, 23). Although not studied directly in the meta-analysis, programs are presumably particularly likely to be successful if they are combined with a good implementation model and if implementation of the program is made with some degree of fidelity.
After publication of the meta-analysis just described, there are at least two additional large-scale intervention projects that merit attention: the Finnish KiVa project, where positive results have already been published (Kärnä et al. 2011a,b), and a three-year intervention project with the OBPP in Pennsylvania, where encouraging results are in the process of being analyzed (Limber & Olweus 2013).
WHY BULLY/VICTIM PROBLEMS CAME LATE IN THE UNITED STATES
In the following section of the present review, I take up again the thread from the beginning concerning the development of the field of bullying research, with a special focus on what happened and did not happen in the United States.
For quite some time after publication of my first study (Olweus 1973a, 1978), I worked relatively alone on research about and intervention work against bully/victim problems at an international level. However, in the early 1980s, the Finnish researcher Kirsti Lagerspetz, who before that time
www.annualreviews.org • School Bullying 771
A nn
u. R
ev . C
lin . P
sy ch
ol . 2
01 3.
9: 75
1- 78
0. D
ow nl
oa de
d fr
om w
w w
.a nn
ua lr
ev ie
w s.
or g
by S
ta te
U ni
ve rs
ity o
f N
ew Y
or k
- B
uf fa
lo o
n 07
/2 7/
13 . F
or p
er so
na l u
se o
nl y.
CP09CH27-Olweus ARI 25 February 2013 18:29
had mainly conducted research on aggression in animals, together with her associates published a couple of studies on school bullying (Björkqvist et al. 1982, Lagerspetz et al. 1982). Later on, in the late 1980s, I was visited in Bergen, maybe a year apart, by two other senior researchers, Peter Smith from the United Kingdom and Kenneth Rigby from Australia, who were interested in doing research and also intervention work on bullying problems. Both of them became quickly involved in such research and have continued to be productive and important contributors to the field, as I believe is well known (e.g., Rigby 1996, Smith & Sharp 1994). Criminologist David Farrington from the United Kingdom was also an important early contributor who saw the value and importance of studying these problems (Farrington 1993).
With regard to the development in the United States, David Perry was probably the first researcher to show a serious interest in the problem. But his first article with colleagues (Perry et al. 1988) was titled “Victims of peer aggression,” and Perry made it clear in his writings that he did not directly study bullying as usually defined, but rather the not fully overlapping phenomenon of children and youth being regularly exposed to aggressive behavior from their peers (see previous section on power imbalance).
At the time the article by Perry et al. was written, the dominant North American approach to the study of peer relations was to measure peer acceptance/likeability or status through positive (“like most”) and negative (“like least”) sociometric peer nominations (as briefly described in a previous section). Using a special classification scheme developed by Coie et al. (1982), researchers sorted their subjects into peer status groups of popular, controversial, neglected, rejected, and average children. A special interest focused on low peer status or rejection, and it was generally believed, partly on the basis of research, that aggressive behavior on the part of the child or youth was the key determinant of rejected status. Much of the research from the 1980s and earlier was summarized in an edited volume on peer rejection that was published in 1990 (Asher & Coie 1990) and in a comprehensive meta-analysis conducted by Newcomb et al. (1993).
Research on bullying in the United States did not become frequent or perceived as important until the early part of the first decade of the twenty-first century, more than ten years later than in Europe (outside Scandinavia) and Australia. Given the great volume of US peer relations research, this was somewhat surprising. It is my belief the this slow entry into the US peer relations research community was to a considerable degree related to and actually a partial consequence of the very strong position of the traditional North American research approach focusing on likeability/acceptance as sketched above. My explication of the reasons for the slow development in the United States, in spite of many excellent peer relations researchers, may of course be perceived as provocative by some people, and I will therefore use some space to try and briefly explain the background and basis for my position.
My own disappointment with this general approach increased gradually during the latter half of the 1980s, and in 1989—at a well-attended Society for Research on Child Development symposium in Kansas City—I expressed my major concerns and criticisms of the dominant approach (Olweus 1989, 2001).
My key arguments related to the fact that peer rejection/likeability measures do not focus directly on the behavioral or personality characteristics of the rejected child. They rather reflect the social environment’s (i.e., the peers’) evaluation of the child in the form of a general liking or disliking. Since children may be disliked or liked by their peers for very different reasons, it was reasonable to expect considerable heterogeneity among children or youth who were rejected by their peers. In contrast to the then dominant view that rejection was mainly a consequence of aggression, children and youth in my own research (for example, Olweus 1978), who were bullied by their peers and highly rejected, typically showed a very different pattern of behavior, one of withdrawal and anxiety and very little aggression.
772 Olweus
A nn
u. R
ev . C
lin . P
sy ch
ol . 2
01 3.
9: 75
1- 78
0. D
ow nl
oa de
d fr
om w
w w
.a nn
ua lr
ev ie
w s.
or g
by S
ta te
U ni
ve rs
ity o
f N
ew Y
or k
- B
uf fa
lo o
n 07
/2 7/
13 . F
or p
er so
na l u
se o
nl y.
CP09CH27-Olweus ARI 25 February 2013 18:29
In addition, my data on boys who bullied other students did not support the view that aggressive behavior typically or mostly leads to rejection by peers. Overall, the generally aggressive bullies were relatively popular among both boys and girls, and they were often supported by two or three other boys in the classroom (Olweus 1978).
In summary, in my symposium contribution, I argued that the presumably marked hetero- geneity of rejected children—with several more or less distinct subgroups of such children with very different personality profiles and behavioral characteristics—and the nature of the rejec- tion/likeability variable would lead to problems of prediction, intervention, and understanding (see details in Olweus 2001). Instead, I argued for a stronger and direct focus on the behavioral or personality characteristics of the children or youth in their classroom contexts and, in parallel, on specific reactions on the part of the social environment.
Not long before the symposium mentioned above, and unknown to me, Doran French pub- lished an article in which he had used cluster analysis with groups of rejected boys and found that there were two clear subtypes of rejected boys: aggressive and nonaggressive (French 1988). The demonstrated heterogeneity of the rejected children (also French 1990) was very much in line with my own thinking and empirical results. Gradually, it became necessary for US peer researchers to talk about and do research on rejected-aggressive and rejected-withdrawn children as two sepa- rate groups. The sociometric label “rejected” was thus too broad and included children with very different personalities and behavioral profiles. This heterogeneity could probably also explain why several interventions focusing on rejected children (cf. Asher & Coie 1990) had been relatively unsuccessful, since rejected-aggressive and rejected-withdrawn children would very likely need distinctly different kinds of therapy or interventions (French 1988, Olweus 1989).
An additional problem with the sociometric likeability classification system was evident when Philip Rodkin and colleagues published an article in 2000 titled “Heterogeneity of popular boys: antisocial and prosocial configurations” (Rodkin et al. 2000). The results in this article showed the same kind of heterogeneity with regard to popularity that had previously been demonstrated for rejection. The authors called the two different types of boys “model boys” and “tough boys.” The authors of this article were all students of and/or collaborators with Robert Cairns, who had devel- oped a new technique for the identification of peer cliques or subgroups and the relative centrality of the clique members. This technique was named social-cognitive mapping and was largely based on student reports on “people in the class/school who hang around a lot” (Cairns & Cairns 1994, Cairns et al. 1989). Somewhat earlier than the Rodkin et al. article, Luthar & McMahon (1996) and Parkhurst & Hopmeyer (1998) had shown that sociometric status/acceptance/likeability and per- ceived popularity (centrality, social dominance, visibility) were relatively weakly related and should be best regarded as two distinct dimensions. Obviously, also the sociometric label popular was too broad and included children and youth with very different personalities and behavioral profiles.
A recent paper by Mayeux & Cillessen (2008) has, maybe somewhat indirectly, pointed to yet another potential problem with the dimension or label popular. What they showed was that the correlation between peer and self-perceptions of popular status both in the form of social pref- erence (sociometric status) and perceived popularity was weak, with correlations in the 0.10–0.30 range. Obviously, many (junior high school) students simply did not know—did not have accurate perceptions—of their own social status in their peer group. Such findings raise the issue of the ex- tent to which sociometric labels such as popular or rejected actually correspond to distinct personal experiences and can be conceptualized as meaningful potentially causal variables, for example.
In sum, although a very great number of articles, book chapters, and books have been published within the research paradigm focusing on likeability, it is my conclusion that the yield in the form of meaningful, replicated knowledge of what actually goes on in peer groups has been relatively meager. It appears to me that the availability of a convenient measurement technique
www.annualreviews.org • School Bullying 773
A nn
u. R
ev . C
lin . P
sy ch
ol . 2
01 3.
9: 75
1- 78
0. D
ow nl
oa de
d fr
om w
w w
.a nn
ua lr
ev ie
w s.
or g
by S
ta te
U ni
ve rs
ity o
f N
ew Y
or k
- B
uf fa
lo o
n 07
/2 7/
13 . F
or p
er so
na l u
se o
nl y.
CP09CH27-Olweus ARI 25 February 2013 18:29
N um
be r o
f P sy
cI N
FO d
at ab
as e
ci ta
ti on
s
600
500
400
300
200
100
0
19 90
19 91
19 92
19 93
19 94
19 95
19 96
19 97
19 98
19 99
20 00
20 01
20 02
20 03
20 04
20 05
20 06
20 07
20 08
20 09
20 10
20 11
Year
5 11 15 20 35 42
56 72 61 68
104
162
127
191
268 254
342
394
466 482
566
523
Figure 3 PsycINFO database citations located that use the words bully, bullying, or bullied. Updated on July 24, 2012.
and classification system did take precedence over conceptualization and problem formulation. In my view, this also contributed to preventing new ideas, approaches, and themes such as bullying from coming into focus for quite some time.
In my symposium contribution in 1989, I called for a reorientation, and this call was reiterated in my chapter (Olweus 2001) in the book edited by Juvonen & Graham (2001), this time in a somewhat more optimistic tone. And the developments in the late 1990s and subsequent years with a focus on perceived popularity as largely distinct from likeability/sociometric popularity seem to be a step in the right direction. This research is relatively closely related to the qualitative, ethnographic studies of peer relations conducted by educational sociologists such as Eder (1995) and Adler & Adler (1998). Although it is difficult to know to what extent the often extremely complex and subtle processes described by Eder and the Adlers can be replicated and generalized, the strong emphasis on power mechanisms and struggles in the peer group seems intuitively meaningful and important. The new developments among US peer relations researchers with a stronger focus on perceived popularity characterized by terms such as social dominance, centrality, and visibility also make it clear that the power dimension is a key aspect of peer relations. As documented in the recently published comprehensive Handbook of Peer Interactions, Relationships, and Groups (Rubin et al. 2009), for example, the approach to the study of peer relations taken by present-day researchers is clearly much broader and more varied than the approach taken 10 years ago.
In parallel with this new trend, the past decade has also witnessed a marked general “reorienta- tion” with regard to research on bully/victim problems. Figure 3, which portrays the number of PsycINFO database citations (articles, books, book chapters, etc.) of the words bully, bullying, and bullied, clearly illustrates this development, with an approximately 100-fold increase of citations per year from 1990 to 2012. A similar development from 1978 to 2005 was shown in a paper by Berger (2007), who talked about an “explosion” of research on bully/victim problems and declared that “Within the past 15 years, scholars have shifted from indifference to fascination regarding bullies” (p. 91). This development is also documented in a number of edited books such as those by Juvonen & Graham (2001), Espelage & Swearer (2004), and Jimerson et al. (2010).
774 Olweus
A nn
u. R
ev . C
lin . P
sy ch
ol . 2
01 3.
9: 75
1- 78
0. D
ow nl
oa de
d fr
om w
w w
.a nn
ua lr
ev ie
w s.
or g
by S
ta te
U ni
ve rs
ity o
f N
ew Y
or k
- B
uf fa
lo o
n 07
/2 7/
13 . F
or p
er so
na l u
se o
nl y.
CP09CH27-Olweus ARI 25 February 2013 18:29
SOME REASONS FOR FOCUSING ON BULLY/VICTIM PROBLEMS
Before closing, I would like to briefly point out some aspects of bully/victim problems, which in my view make them attractive and interesting from both a research and an intervention/prevention perspective:
� The implicated behaviors and reactions are anchored in a social context, and there is a primary focus on perpetrator(s)/aggressor(s), targets/victim(s), bystanders, and the peer group(s). Generally, this entails a more or less direct attention to the power dynamics in the peer group.
� A focus on bully/victim problems leads naturally to an examination of larger units or contexts (a multilevel or ecological approach): classes, schools, and communities and related aspects including teacher behavior, classroom norms, social psychological mechanisms, and school policy. With such an approach, multilevel techniques (e.g., Heck et al. 2010, Raudenbush & Bryk 2003) are both necessary for adequate statistical treatment of research data and very helpful in conceptualizing and analyzing relationships and influences at and across different levels.
� A considerable portion of what characterizes bully/victim problems can be conceptualized in a causal-analytic framework. The implicated behaviors and reactions represent real peer experiences (possible causal factors) with often quite serious short- and long-term conse- quences for the individuals involved and their families and partners. Longitudinal studies make it clear that these problems also constitute a major, not fully realized, public health issue with great costs to society (Olweus & Breivik 2013).
� A simple description of the problems usually provides a good rationale for intervention on which it is relatively easy to reach consensus (at least in principle): no acceptance. The suffer- ing of the victim(s) is the main motivation behind intervention, and this can be linked to the UN Convention on the Rights of the Child and to the obligation of schools to provide a safe learning environment for every child (possibly expressed in legislation). Prevention of such problems also represents large economic gains to society (see, e.g., Olweus & Limber 2010).
� Antibullying work, in particular of the whole-school kind, provides a good entry point for positive school change or organizational development more generally, which will benefit all students in the school.
� A focus on bully/victim problems should also be highly rewarding for both researchers and interventionists by providing ample opportunities to make life safer and happier for large numbers of children and youth.
CONCLUDING WORDS
Looking back on the more than 40 years since my research interest in bullying problems was first initiated, there are two developments that I see as particularly striking and rewarding. The first development is the fact that there has occurred a dramatic change in many countries in the world in the perception of bully/victim problems: from being regarded as a natural part of growing up to being viewed as a pressing social and public health issue that schools and the educational authorities must take seriously and counteract. The second is the documentation and (maybe still partial) understanding that it is quite possible to considerably reduce these problems with a research-based intervention program that is implemented with some degree of fidelity.
DISCLOSURE STATEMENT
The author is not aware of any affiliations, memberships, funding, or financial holdings that might be perceived as affecting the objectivity of this review.
www.annualreviews.org • School Bullying 775
A nn
u. R
ev . C
lin . P
sy ch
ol . 2
01 3.
9: 75
1- 78
0. D
ow nl
oa de
d fr
om w
w w
.a nn
ua lr
ev ie
w s.
or g
by S
ta te
U ni
ve rs
ity o
f N
ew Y
or k
- B
uf fa
lo o
n 07
/2 7/
13 . F
or p
er so
na l u
se o
nl y.
CP09CH27-Olweus ARI 25 February 2013 18:29
ACKNOWLEDGMENTS
Some of the research reported in this review was financially supported by the Norwegian Ministry of Children and Family Affairs (BLD) and the Norwegian Ministry of Education and Research, which is gratefully acknowledged. Thanks also to my collaborator Kyrre Breivik for valuable comments on the manuscript.
LITERATURE CITED
Achenbach TM, McConaughy SH, Howell CT. 1987. Child/adolescent behavioral and emotional problems: implications of cross-informant correlations for situational specificity. Psychol. Bull. 101:213–32
Adler PA, Adler P. 1998. Peer Power: Preadolescent Culture and Identity. New Brunswick, NJ: Rutgers Univ. Press
Arseneault L, Bowes L, Shakoor S. 2010. Bullying victimization in youths and mental health problems: “much ado about nothing”? Psychol. Med. 40:717–29
Asher SR, Coie JD, eds. 1990. Peer Rejection in Childhood. New York: Cambridge Univ. Press Berger KS. 2007. Update on bullying at school: science forgotten? Dev. Rev. 27(1):90–126 Berkowitz L. 1993. Aggression. Its Causes, Consequences, and Control. New York: McGraw-Hill Björkqvist K, Ekman K, Lagerspetz K. 1982. Bullies and victims: their ego picture, ideal ego picture and
normative ego picture. Scand. J. Psychol. 23(1):307–13 Branson CE, Cornell DG. 2009. A comparison of self and peer reports in the assessment of middle school
bullying. J. Appl. Sch. Psychol. 25:5–27 Breivik K, Olweus D. 2012. An Item Response Theory Analysis of the Olweus Bullying Scale. Bergen, Norway:
Cent. Child Adolesc. Ment. Health Welf. (RKBU Vest). Manuscript submitted Buss AH. 1961. The Psychology of Aggression. New York: Wiley Cairns RB, Cairns BD. 1994. Lifelines and Risks: Pathways of Youth in Our Time. New York: Harvester Wheat-
sheaf Cairns RB, Cairns BD, Neckerman HJ, Ferguson LL, Gariépy JL. 1989. Growth and aggression: I. Childhood
to early adolescence. Dev. Psychol. 25(2):320–30 Card N. 2003, April. Victims of peer aggression: a meta-analytic review. Paper presented at bienn. meet. Soc. Res.
Child Dev., Tampa, FL Coie JD, Dodge KA. 1998. Aggression and antisocial behavior. In Handbook of Child Psychology: Vol. 3. Social,
Emotional, and Personality Development, ser. ed. W Damon, vol. ed. N Eisenberg, pp. 779–862. New York: Wiley. 5th ed.
Coie JD, Dodge KA, Coppotelli H. 1982. Dimensions and types of social status: a cross-age perspective. Dev. Psychol. 18(4):557–70
Cook RC, Williams KR, Guerra NG, Kim TE, Sadek S. 2010. Predictors of bullying and victimization in childhood and adolescence: a meta-analytic investigation. Sch. Psychol. Q. 25:65–83
Cornell DG, Bandyopadhyay S. 2010. The assessment of bullying. See Jimerson et al. 2010, pp. 265–76 Cornell DG, Brockenbrough K. 2004. Identification of bullies and victims: a comparison of methods. J. Sch.
Violence 3:63–87 Eder D. 1995. School Talk: Gender and Adolescent Identity. New Brunswick, NJ: Rutgers Univ. Press Espelage DL, Holt MK, Henkel RR. 2003. Examination of peer group contextual effects on aggression during
early adolescence. Child Dev. 74:205–20 Espelage DL, Swearer SM, eds. 2004. Bullying in American Schools. A Social-Ecological Perspective on Prevention
and Intervention. Mahwah, NJ: Erlbaum Farrington DP. 1993. Understanding and preventing bullying. In Crime and Justice: A Review of Research, ed.
M Tonry, vol. 17, pp. 381–458. Chicago: Univ. Chicago Press Fekkes M, Pijpers FIM, Verloove-Vanhorick S. 2005. Bullying: Who does what, when and where? Involvement
of children, teachers and parents in bullying behavior. Health Educ. Res. 20(1):81–91 Felix ED, McMahon SD. 2006. Gender and multiple forms of peer victimization: How do they influence
adolescent psychosocial adjustment? Violence Vict. 21(6):707–24
776 Olweus
A nn
u. R
ev . C
lin . P
sy ch
ol . 2
01 3.
9: 75
1- 78
0. D
ow nl
oa de
d fr
om w
w w
.a nn
ua lr
ev ie
w s.
or g
by S
ta te
U ni
ve rs
ity o
f N
ew Y
or k
- B
uf fa
lo o
n 07
/2 7/
13 . F
or p
er so
na l u
se o
nl y.
CP09CH27-Olweus ARI 25 February 2013 18:29
Ferguson CJ, San Miguel C, Hartley RD. 2009. A multivariate analysis of youth violence and aggression: the influence of family, peers, depression, and media violence. J. Pediatr. 155(6):904–8.e3
Flay BR, Biglan A, Boruch RF, Castro FG, Gottfredson D, et al. 2005. Standards of evidence: criteria for efficacy, effectiveness and dissemination. Prev. Sci. 6:151–75
French DC. 1988. Heterogeneity of peer-rejected boys: aggressive and nonaggressive subtypes. Child Dev. 976–85
French DC. 1990. Heterogeneity of peer-rejected girls. Child Dev. 65:2028–31 Fuchs F, Diamantopoulus A. 2009. Using single-item measures for construct measurement in management
research. Die Betriebswirtschaft 69:195–210 Guilford JP. 1956. Fundamental Statistics in Psychology and Education. New York: McGraw-Hill Hartung CM, Little CS, Allen EK, Page MC. 2011. A psychometric comparison of two self-report measures
of bullying and victimization: differences by sex and grade. Sch. Mental Health 3(1):44–57 Heck RH, Thomas SL, Tabata LN. 2010. Multilevel and Longitudinal Modeling with IBM SPSS. New York:
Routledge Heinemann P-P. 1969. Apartheid. Liberal Debatt 3:3–14 Heinemann P-P. 1972. Mobbning: gruppvåld bland barn och vuxna (Mobbing: Group Violence Among Children and
Adults). Stockholm: Natur och kultur Hunter JE. 2001. The desperate need for replications. J. Consum. Res. 28:149–58 Hunter SC, Boyle JME, Warden D. 2007. Perceptions and correlates of peer-victimization and bullying. Br.
J. Educ. Psychol. 77:797–810 Jenson JM, Dieterich WA, Brisson D, Bender KA, Powell A. 2010. Preventing childhood bullying: findings
and lessons from the Denver Public Schools Trial. Res. Soc. Work Pract. 20(5):509–517 Jiang D, Walsh M, Augimeri LK. 2011. The linkage between childhood bullying behaviour and future of-
fending. Crim. Behav. Mental Health 21(2):128–35 Jimerson SR, Swearer SM, Espelage DL, eds. 2010. Handbook of School Bullying: An International Perspective.
New York: Routledge Junger-Tas J, Kesteren JV. 1998. Cross-cultural study of bully/victim problems in school: final report for the
Netherlands to the Japanese Ministry of Education. In School Bullying Around the World, ed. Y Morita, H Soeda. Tokyo: Jpn. Ministry Educ.
Juvonen J, Graham S, eds. 2001. Peer Harassment in School. New York: Guilford Kärnä A, Voeten M, Little TD, Poskiparta E, Alanen E, Salmivalli C. 2011a. Going to scale: a nonrandomized
nationwide trial of the KiVa antibullying program for grade 1–9. J. Consult. Clin. Psychol. 79:796–805 Kärnä A, Voeten M, Little TD, Poskiparta E, Kaljonen A, Salmivalli C. 2011b. A large-scale evaluation of the
KiVa antibullying program: grades 4–6. Child Dev. 82(1):311–30 Kim MJ, Catalano RF, Haggerty KP, Abbott RD. 2011. Bullying at elementary school and problem behaviour
in young adulthood: a study of bullying, violence and substance use from age 11 to age 21. Crim. Behav. Mental Health 21(2):136–44
Kyriakides L, Kaloyirou C, Lindsay G. 2006. An analysis of the Revised Olweus Bully/Victim Questionnaire using the Rasch measurement model. Br. J. Educ. Psychol. 76(4):781–801
Lagerspetz KMJ, Björkqvist K, Berts M, King E. 1982. Group aggression among school children in three schools in Finland. Scand. J. Psychol. 23:45–52
Lee T, Cornell D. 2010. Concurrent validity of the Olweus Bully/Victim Questionnaire. J. Sch. Violence 9(1):56–73
Limber SP, Olweus D. 2013. Evaluation of the Olweus Bullying Prevention Program in two large-scale Pennsylvania projects. Manuscript in preparation
Lindzey G, ed. 1954. Handbook of Social Psychology, Vol. 1. Cambridge, MA: Addison-Wesley Lorenz K. 1963. Das sogenannte Böse. Wien: Borotha-Schoeler Lorenz K. 1968. Aggression: Dess bakgrund och natur. Stockholm: Norstedt & Söner Lösel F, Bender D. 2011. Emotional and antisocial outcomes of bullying and victimization at school: a follow-
up from childhood to adolescence. J. Aggress. Confl. Peace Res. 3:89–96 Luthar SS, McMahon TJ. 1996. Peer reputation among inner-city adolescents: structure and correlates. J.
Res. Adolesc. 6:581–603
www.annualreviews.org • School Bullying 777
A nn
u. R
ev . C
lin . P
sy ch
ol . 2
01 3.
9: 75
1- 78
0. D
ow nl
oa de
d fr
om w
w w
.a nn
ua lr
ev ie
w s.
or g
by S
ta te
U ni
ve rs
ity o
f N
ew Y
or k
- B
uf fa
lo o
n 07
/2 7/
13 . F
or p
er so
na l u
se o
nl y.
CP09CH27-Olweus ARI 25 February 2013 18:29
Mayeux L, Cillessen AHN. 2008. It’s not just being popular, it’s knowing it, too: the role of self-perceptions of status in the associations between peer status and aggression. Soc. Dev. 17(4):871–88
Mishna F, Cook C, Saini M, Wu M-J, MacFaddon R. 2011. Interventions to prevent and reduce cyber abuse of youth: a systematic review. Res. Soc. Work Pract. 21:5–14
Morita Y, Soeda H. 1998. Cross-cultural study of bully/victim problems in school: final report for Japan to the Japanese Ministry of Education. See Morita & Soeda 1998
Morita Y, Soeda H. 1998. School Bullying Around the World. Tokyo: Jpn. Ministry Educ. Newcomb AF, Bukowski WM, Pattee L. 1993. Children’s peer relations: a meta-analytic review of popular,
rejected, neglected, controversial, and average sociometric status. Psychol. Bull. 113(1):99–128 Nilsson I, Wadeskog A. 2008. Det är bättre stämma i bäcken än i ån. Stockholm: Skandia försäkringsbolag.
http://www.ideerforlivet.se Olweus D. 1969. Prediction of Aggression. Stockholm: Skandinaviska testförlaget Olweus D. 1973a. Hackkycklingar och översittare: forskning om skolmobning. Stockholm: Almqvist & Wiksell Olweus D. 1973b. Personality and aggression. Neb. Symp. Motiv. 20:261–321 Olweus D. 1977. Aggression and peer acceptance in adolescent boys: two short-term longitudinal studies of
ratings. Child Dev. 48:1301–13 Olweus D. 1978. Aggression in the Schools. Bullies and Whipping Boys. Washington, DC: Hemisphere Olweus D. 1979. Stability of aggressive reaction patterns in males: a review. Psychol. Bull. 86:852–75 Olweus D. 1980a. Familial and temperamental determinants of aggressive behavior in adolescent boys: a causal
analysis. Dev. Psychol. 16:644–60 Olweus D. 1980b. Bullying among school-boys. In Children and Violence, ed. N Cantwell, pp. 97–131. Stock-
holm: Akademilitteratur Olweus D. 1986. The Olweus Bully/Victim Questionnaire. Bergen, Norway: Res. Cent. Health Promot.
(HEMIL), Univ. Bergen Olweus D. 1989 (March). Peer relationship problems: conceptual issues and a successful intervention program
against bully/victim problems. In Aggressors, Victims, and Peer Relationships, K. Rubin (chair). Symp. meet. Soc. Res. Child Dev. (SRCD), Kansas City, MO
Olweus D. 1991. Bully/victim problems among schoolchildren: basic facts and effects of a school based in- tervention program. In The Development and Treatment of Childhood Aggression, ed. D Pepler, K Rubin, pp. 411–48. Hillsdale, NJ: Erlbaum
Olweus D. 1993. Bullying at School: What We Know and What We Can Do. Oxford, UK: Blackwell Olweus D. 1994. Annotation. Bullying at school: basic facts and effects of a school based intervention program.
J. Child Psychol. Psychiatry 35:1171–90 Olweus D. 1996. The Revised Olweus Bully/Victim Questionnaire. Bergen, Norway: Res. Cent. Health Promot.
(HEMIL), Univ. Bergen Olweus D. 1999a. Sweden. See Smith et al. 1999, pp. 7–27 Olweus D. 1999b. Norway. See Smith et al. 1999, pp. 28–48 Olweus D. 2001. Peer harassment. A critical analysis and some important issues. See Juvonen & Graham 2001,
pp. 3–20 Olweus D. 2005. A useful evaluation design, and effects of the Olweus Bullying Prevention Program. Psychol.
Crime Law 11:389–402 Olweus D. 2007. The Olweus Bullying Questionnaire. Center City, MN: Hazelden. http://www.
hazelden.org/olweus Olweus D. 2010. Understanding and researching bullying: some critical issues. In Handbook of Bullying in
Schools: An International Perspective, ed. SS Jimerson, SM Swearer, DL Espelage, pp. 9–33. New York: Routledge
Olweus D. 2012a. Invited expert discussion paper. Cyber bullying: an overrated phenomenon? Eur. J. Dev. Psychol. 9:520–38
Olweus D. 2012b. Commentary. Comments on cyberbullying article: a rejoinder. Eur. J. Dev. Psychol. 9:559–68 Olweus D, Breivik K. 2013. The plight of victims of school bullying: the opposite of well-being. In Handbook
of Child Well-Being, ed. B-A Asher, F Casas, I Frones, JE Korbin. Heidelberg, Ger.: Springer. In press Olweus D, Limber SP. 2010. The Olweus Bullying Prevention Program: implementation and evaluation over
two decades. See Jimerson et al. 2010, pp. 377–401
778 Olweus
A nn
u. R
ev . C
lin . P
sy ch
ol . 2
01 3.
9: 75
1- 78
0. D
ow nl
oa de
d fr
om w
w w
.a nn
ua lr
ev ie
w s.
or g
by S
ta te
U ni
ve rs
ity o
f N
ew Y
or k
- B
uf fa
lo o
n 07
/2 7/
13 . F
or p
er so
na l u
se o
nl y.
CP09CH27-Olweus ARI 25 February 2013 18:29
Olweus D, Mattsson A, Schalling D, Löw H. 1988. Circulating testosterone levels and aggression in adolescent males: a causal analysis. Psychosom. Med. 50:261–72
Olweus D, Solberg M. 1998. Cross-cultural study of bully/victim problems in school: final report for Norway to Japanese Ministry of Education. See Morita & Soeda 1998
Parkhurst JT, Hopmeyer A. 1998. Sociometric popularity and peer-perceived popularity two distinct dimen- sions of peer status. J. Early Adolesc. 18(2):125–44
Perry DG, Kusel SJ, Perry LC. 1988. Victims of peer aggression. Dev. Psychol. 24(6):807–14 Renda J, Vassallo S, Edwards B. 2011. Bullying in early adolescence and its association with anti-social be-
haviour, criminality and violence 6 and 10 years later. Crim. Behav. Mental Health 21(2):117–27 Rigby K. 1996. Bullying in Schools and What to Do About It. Melbourne: Aust. Counc. Educ. Res. Rigby K, Slee P. 1991. Bullying among Australian school children: reported behaviour and attitudes to victims.
J. Soc. Psychol. 131:615–27 Rodkin PC, Farmer TW, Pearl R, Van Acker R. 2000. Heterogeneity of popular boys: antisocial and prosocial
configurations. Dev. Psychol. 36(1):14–24 Raudenbush SW, Bryk AS. 2002. Hierarchical Linear Models: Applications and Data Analysis Methods. Thousand
Oaks, CA: Sage. 2nd ed. Rubin KH, Bukowsky WM, Laursen B, eds. 2009. Handbook of Peer Interactions, Relationships, and Groups. New
York: Guilford Salmivalli C, Lagerspetz K, Björkqvist K, Österman K, Kaukiainen A. 1996. Bullying as a group process:
participant roles and their relations to social status within the group. Aggress. Behav. 22:1–15 Salmivalli C, Pöyhönen V. 2011. Cyberbullying in Finland. In Cyberbullying in the Global Playground: Research
from International Perspectives, ed. Q Li, D Cross, PK Smith, pp. 57–72. London: Wiley Schäfer M, Werner NE, Crick NR. 2002. A comparison of two approaches to the study of peer treatment:
general victimization and bully/victim problems among German schoolchildren. Br. J. Dev. Psychol. 20:281–306
Smith PK, Brain P. 2000. Bullying in schools: lessons from two decades of research. Aggress. Behav. 26:1–9 Smith PK, del Barrio C, Tokunaga R. 2012. Definitions of bullying and cyberbullying: How useful are the
terms? In Principles of Cyberbullying Research: Definition, Measures, and Methods, ed. S Bauman, D Cross, J Walker, pp. 29–40. Philadelphia, PA: Routledge
Smith PK, Mahdavi J, Carvalho M, Fisher S, Russell S, Tippett N. 2008. Cyberbullying: its nature and impact in secondary schools. J. Child Psychol. Psychiatry 49:376–85
Smith PK, Morita Y, Junger-Tas J, Olweus D, Catalano R, Slee P, eds. 1999. The Nature of School Bullying: A Cross-National Perspective. London: Routledge
Smith PK, Sharp S, eds. 1994. Insights and Perspectives. London: Routledge Solberg ME, Olweus D. 2003. Prevalence estimation of school bullying with the Olweus Bully/Victim Ques-
tionnaire. Aggress. Behav. 29(3):239–68 Solberg ME, Olweus D, Endresen IM. 2007. Bullies and victims at school: Are they the same pupils? Br. J.
Educ. Psychol. 77(2):441–64 Solberg M, Olweus D, Endresen IM. 2010. Bullies, victims, and bully-victims: How deviant are they and
how different? In Self-reported bullying and victimization at school, M Solberg. PhD thesis. Univ. Bergen, Norway
Sourander A, Jensen P, Rönning JA, Niemelä S, Helenius H, et al. 2007. What is the early adulthood outcome of boys who bully or are bullied in childhood? The Finnish “From a Boy to a Man” study. Pediatrics 120(2):397–404
Stavrinides P, Georgiou S, Theofanous V. 2010. Bullying and empathy: a short term longitudinal investigation. Educ. Psychol. 30(7):793–802
Strohmeier D, Kärnä A, Salmivalli C. 2011. Intrapersonal and interpersonal risk factors for peer victimization in immigrant youth in Finland. Dev. Psychol. 47(1):248–58
Tedeschi JT, Felson RB. 1994. Violence, Aggression, and Coercive Actions. Washington, DC: Am. Psychol. Assoc. Theriot MT, Dulmus CN, Sowers KM, Johnson TK. 2005. Factors relating to self-identification among
bullying victims. Child. Youth Serv. Rev. 27(9):979–94 Ttofi MM, Farrington DP. 2009. What works in preventing bullying: effective elements of anti-bullying
programmes. J. Aggress. Confl. Peace Res. 1(1):13–24
www.annualreviews.org • School Bullying 779
A nn
u. R
ev . C
lin . P
sy ch
ol . 2
01 3.
9: 75
1- 78
0. D
ow nl
oa de
d fr
om w
w w
.a nn
ua lr
ev ie
w s.
or g
by S
ta te
U ni
ve rs
ity o
f N
ew Y
or k
- B
uf fa
lo o
n 07
/2 7/
13 . F
or p
er so
na l u
se o
nl y.
CP09CH27-Olweus ARI 25 February 2013 18:29
Ttofi MM, Farrington DP. 2011. Effectiveness of school-based programs to reduce bullying: a systematic and meta-analytic review. J. Exp. Criminol. 7(1):27–56
Ttofi MM, Farrington DP, Lösel F, Loeber R. 2011a. The predictive efficiency of school bullying versus later offending: a systematic/meta-analytic review of longitudinal studies. Crim. Behav. Mental Health 21(2):80–89
Ttofi MM, Farrington DP, Lösel F, Loeber R. 2011b. Do the victims of school bullies tend to become depressed later in life? A systematic review and meta-analysis of longitudinal studies. J. Aggress. Confl. Peace Res. 3(2):63–73
Ybarra ML, Boyd D, Korchmaros JD, Oppenheim J. 2012. Defining and measuring cyberbullying within the larger context of bullying victimization. J. Adolesc. Health 51:53–58
780 Olweus
A nn
u. R
ev . C
lin . P
sy ch
ol . 2
01 3.
9: 75
1- 78
0. D
ow nl
oa de
d fr
om w
w w
.a nn
ua lr
ev ie
w s.
or g
by S
ta te
U ni
ve rs
ity o
f N
ew Y
or k
- B
uf fa
lo o
n 07
/2 7/
13 . F
or p
er so
na l u
se o
nl y.
CP09-FrontMatter ARI 9 March 2013 1:0
Annual Review of Clinical Psychology
Volume 9, 2013 Contents
Evidence-Based Psychological Treatments: An Update and a Way Forward David H. Barlow, Jacqueline R. Bullis, Jonathan S. Comer,
and Amantia A. Ametaj � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � 1
Quitting Drugs: Quantitative and Qualitative Features Gene M. Heyman � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � �29
Integrative Data Analysis in Clinical Psychology Research Andrea M. Hussong, Patrick J. Curran, and Daniel J. Bauer � � � � � � � � � � � � � � � � � � � � � � � � � � � �61
Network Analysis: An Integrative Approach to the Structure of Psychopathology Denny Borsboom and Angélique O.J. Cramer � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � �91
Principles Underlying the Use of Multiple Informants’ Reports Andres De Los Reyes, Sarah A. Thomas, Kimberly L. Goodman,
and Shannon M.A. Kundey � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � 123
Ambulatory Assessment Timothy J. Trull and Ulrich Ebner-Priemer � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � 151
Endophenotypes in Psychopathology Research: Where Do We Stand? Gregory A. Miller and Brigitte Rockstroh � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � 177
Fear Extinction and Relapse: State of the Art Bram Vervliet, Michelle G. Craske, and Dirk Hermans � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � 215
Social Anxiety and Social Anxiety Disorder Amanda S. Morrison and Richard G. Heimberg � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � 249
Worry and Generalized Anxiety Disorder: A Review and Theoretical Synthesis of Evidence on Nature, Etiology, Mechanisms, and Treatment Michelle G. Newman, Sandra J. Llera, Thane M. Erickson, Amy Przeworski,
and Louis G. Castonguay � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � 275
Dissociative Disorders in DSM-5 David Spiegel, Roberto Lewis-Fernández, Ruth Lanius, Eric Vermetten,
Daphne Simeon, and Matthew Friedman � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � 299
viii
A nn
u. R
ev . C
lin . P
sy ch
ol . 2
01 3.
9: 75
1- 78
0. D
ow nl
oa de
d fr
om w
w w
.a nn
ua lr
ev ie
w s.
or g
by S
ta te
U ni
ve rs
ity o
f N
ew Y
or k
- B
uf fa
lo o
n 07
/2 7/
13 . F
or p
er so
na l u
se o
nl y.
CP09-FrontMatter ARI 9 March 2013 1:0
Depression and Cardiovascular Disorders Mary A. Whooley and Jonathan M. Wong � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � 327
Interpersonal Processes in Depression Jennifer L. Hames, Christopher R. Hagan, and Thomas E. Joiner � � � � � � � � � � � � � � � � � � � � � 355
Postpartum Depression: Current Status and Future Directions Michael W. O’Hara and Jennifer E. McCabe � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � 379
Emotion Deficits in People with Schizophrenia Ann M. Kring and Ori Elis � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � 409
Cognitive Interventions Targeting Brain Plasticity in the Prodromal and Early Phases of Schizophrenia Melissa Fisher, Rachel Loewy, Kate Hardy, Danielle Schlosser,
and Sophia Vinogradov � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � 435
Psychosocial Treatments for Schizophrenia Kim T. Mueser, Frances Deavers, David L. Penn, and Jeffrey E. Cassisi � � � � � � � � � � � � � � 465
Stability and Change in Personality Disorders Leslie C. Morey and Christopher J. Hopwood � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � 499
The Relationship Between Personality Disorders and Axis I Psychopathology: Deconstructing Comorbidity Paul S. Links and Rahel Eynan � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � 529
Revisiting the Relationship Between Autism and Schizophrenia: Toward an Integrated Neurobiology Nina de Lacy and Bryan H. King � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � 555
The Genetics of Eating Disorders Sara E. Trace, Jessica H. Baker, Eva Peñas-Lledó, and Cynthia M. Bulik � � � � � � � � � � � � � 589
Neuroimaging and Other Biomarkers for Alzheimer’s Disease: The Changing Landscape of Early Detection Shannon L. Risacher and Andrew J. Saykin � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � 621
How Can We Use Our Knowledge of Alcohol-Tobacco Interactions to Reduce Alcohol Use? Sherry A. McKee and Andrea H. Weinberger � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � 649
Interventions for Tobacco Smoking Tanya R. Schlam and Timothy B. Baker � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � 675
Neurotoxic Effects of Alcohol in Adolescence Joanna Jacobus and Susan F. Tapert � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � 703
Socioeconomic Status and Health: Mediating and Moderating Factors Edith Chen and Gregory E. Miller � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � 723
Contents ix
A nn
u. R
ev . C
lin . P
sy ch
ol . 2
01 3.
9: 75
1- 78
0. D
ow nl
oa de
d fr
om w
w w
.a nn
ua lr
ev ie
w s.
or g
by S
ta te
U ni
ve rs
ity o
f N
ew Y
or k
- B
uf fa
lo o
n 07
/2 7/
13 . F
or p
er so
na l u
se o
nl y.
CP09-FrontMatter ARI 9 March 2013 1:0
School Bullying: Development and Some Important Challenges Dan Olweus � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � 751
The Manufacture of Recovery Joel Tupper Braslow � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � 781
Indexes
Cumulative Index of Contributing Authors, Volumes 1–9 � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � 811
Cumulative Index of Articles Titles, Volumes 1–9 � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � 815
Errata
An online log of corrections to Annual Review of Clinical Psychology articles may be found at http://clinpsy.annualreviews.org
x Contents
A nn
u. R
ev . C
lin . P
sy ch
ol . 2
01 3.
9: 75
1- 78
0. D
ow nl
oa de
d fr
om w
w w
.a nn
ua lr
ev ie
w s.
or g
by S
ta te
U ni
ve rs
ity o
f N
ew Y
or k
- B
uf fa
lo o
n 07
/2 7/
13 . F
or p
er so
na l u
se o
nl y.
- Annual Reviews Online
- Search Annual Reviews
- Annual Review of Clinical Psychology Online
- Most Downloaded Clinical Psychology Reviews
- Most Cited Clinical Psychology Reviews
- Annual Review of Clinical Psychology Errata
- View Current Editorial Committee
- All Articles in the Annual Review of Clinical Psychology, Vol. 9
- Evidence-Based Psychological Treatments: An Updateand a Way Forward
- Quitting Drugs: Quantitative and Qualitative Features
- Integrative Data Analysis in Clinical Psychology Research
- Network Analysis: An Integrative Approach to the Structure of Psychopathology
- Principles Underlying the Use of Multiple Informants’ Reports
- Ambulatory Assessment
- Endophenotypes in Psychopathology Research: Where Do We Stand?
- Fear Extinction and Relapse: State of the Art
- Social Anxiety and Social Anxiety Disorder
- Worry and Generalized Anxiety Disorder: A Review and Theoretical Synthesis of Evidence on Nature, Etiology, Mechanisms, and Treatment
- Dissociative Disorders in DSM-5
- Depression and Cardiovascular Disorders
- Interpersonal Processes in Depression
- Postpartum Depression: Current Status and Future Directions
- Emotion Deficits in People with Schizophrenia
- Cognitive Interventions Targeting Brain Plasticity in the Prodromal and Early Phases of Schizophrenia
- Psychosocial Treatments for Schizophrenia
- Stability and Change in Personality Disorders
- The Relationship Between Personality Disorders and Axis I Psychopathology: Deconstructing Comorbidity
- Revisiting the Relationship Between Autism and Schizophrenia:Toward an Integrated Neurobiology
- The Genetics of Eating Disorders
- Neuroimaging and Other Biomarkers for Alzheimer’s Disease:The Changing Landscape of Early Detection
- How Can We Use Our Knowledge of Alcohol-Tobacco Interactions to Reduce Alcohol Use?
- Interventions for Tobacco Smoking
- Neurotoxic Effects of Alcohol in Adolescence
- Socioeconomic Status and Health: Mediating and Moderating Factors
- School Bullying: Development and Some Important Challenges
- The Manufacture of Recovery
