Content uploaded by Christina Salmivalli
Author content
All content in this area was uploaded by Christina Salmivalli on Oct 09, 2023
Content may be subject to copyright.
,.\GGRESSIvE
I}EH.{VIOR
\blume
22.
pages
l-15 (1996)
Bullying
as
a Group Process:
Participant
Roles
and
Their
Relations
to
Social Status
Within
the
Group
Christina
Salmivalli,
Kirsti Lagerspetz,
Kaj Bjorkqvist,
Karin
Osterman, and
Ari
Kaukiainen
Depaftment
of Psychology,
University of Turku, Turku
(C.5.,
K.L., A.K.)
and Department
of
Psychotogy,
Abo Akademi
lJniversity, Vasa (K.8.,
K.O.), Fintand
Bullring rras
investigated
as a group process,
a social
phenomenon
taking
place
in a
school setting
among 573 Finnish
sixth.grade
children
(286
girls.287
bols) aged l2-13
J
ears. Different Participant Roles
taken bv individual children in the bullr ing process
rlere
examined and related
to a)
self.estimated behavior in
bullling situations,
b) social
acceptance
and
social
rejection,
and
c) belongingness to one of the fi\e
sociometric
sta-
tus Broups
(popular. rejected.
neglected, controrersial, and a|erage). The Participant
Roles assigned
to the subjects
s ere Victim, Bullv, Reinforcer of the bull-ti
Assistant
of
the bull-r', Defender
of the r ictim, and Outsiden There rlere
significant sex differences in
the distribution
of Participant
Roles. Bo1s
uere more
frequintl.v in
the
roles
of Bullr,
Reinforcer
and Assistant, $hile the most frequent roles of the girls u.ere
those of De-
fender and Outsider. The
subjects
\rere
moderatel.\' $'ell
au are
of their Participant Roles,
although
thel underestimated
their
participation
in actile bull-ring
behalior and
em-
phasized
that thel" acted
as Defendcrs
and Outsiders.
The sociometric slatus of the chil-
dren$asfoundtobeconnectedtotheirParticiDantRoles.
o1996lt,ir.r-Liss.tnc.
Kev
sords;
aggressire
behalioq peer
relations, roles,
social acceptance.
social
groups, \ictimization
INTRODUCTION
Bulll,ing in
schools can
be seen as a
category ofaggressive
behavior in
u,hich
rhere
is
an imbalance
of porver.
and the
aggressive act is repeated
over time
IOlu'eus,
l99l;
Smith and Thompson,
l99ll. Bulll,in_l
is social in its nature.
and takes
place
in rela-
tively
permanent
social
-eroups,
in
u
hich the
victim
has little possibility
of
avoiding
his/
Rcceived
for publication
February
I,
Address
reprint requests
to Christrna
20500,
Finland.
O 1996 \\'ile1'.Liss,
Inc.
I 995 accepted April 26, | 99-s.
Saimivalli. Depannrent
of Pslchologv.
Unireruily'of
Turku,
Fh*-
2 Sulmitalli tt al.
her tornrentors.
and
rhc bully ofren gers
supporr
fror'l
orher
group
mernbers
IBjrjrkqvist
et al.. I
9821.
The attacks
are
rnostlv
unprovoked
anci
may
be
phl,sical
or verbal,
direct
or indirect.
Researchers
in rhis
ficld
often crlphasize
thc socjal
chrracrcr
ol'bullyrng.
For instarrce.
Pikas
il9T5ltlcscribes
bullf ine
ls violencc
in e group
contc\t
in \\,hich
the pupils
reinlbrce
cach otltcls'behavior
in
their intcracrion.
This vierv
o1'trull1,i1e
as
e prinlarill'social
activity
is. horvever.
scldonr convcrtccJ
into
entpirical
researcfi.
Stud-
ies
on
bullf ing have
tvpicaill'concenrrared
only
on ihe
bulll'-r.ictim
relarion-ship.
lt is
as if bullf ine
bchavior *'as
regarded
only as
a
l'unction
of cerrain
characterisrics
of the
bulli.and/or
rhe
vicrint,
ri'hile
rhe
_group
conre.\t.
is ser aside,
or fbrsotren.
La::crs;.retz
et al. [1982]
ha'e poinred
out rr'o inrponant
fearures
ofbullying anrorrg
school
children: l) its collecrive
cllrracrer, and 2) rhe fact rhat
ir is bascd
on social
rc'lationships
in the group.
They surgest
that aggression
in a
eroup
can
be studied
as
a
relationship
berrrecn people
teking
dilferent
roles.
or having
roles assigned
ro
thenr.
The present
studv
focuses
on thcse
roies.
resarding
thenr lrs pan ofa process
involr,-
ing a
ri
lrolc group,
the school
class.
The
rnain research problenr
u
as
to investigate what
other
pupils.
or group
nreurbers,
do,
r'hen
a
bully is lrarassin_s
a
victim.
il rvas
assumed
that all the children in the class
are sonrehow involved in, or at least
aware
of, the
bulll in,e process,
even
if thcy
do not actively
artack
rhe vicrirrr.
Bullying often rakes
place
in a situation in s'hich
se'eral rnenrbers
ofthe group
are
present;
even the
ones
not
present
are
usually
arvarc
of*,hat is,eoing
on, due
to the fact
that
bu llyina
by definirion
happens
repeatedlv,
over a
period
of rinre.
Even
i f the
malor-
ity of chiidren in rhe class
cio
not parricipare
in acti'e bully,ing
bcha'ior, rhey rnay
behave
in says rvhich nrake
the beeinning
anc.l
continuarion
of the bullyin,e process
possrble.
what nrarters
more
rhan
their real
attitude
to
bulll,in_r-in fact.
mosi &il.jr.n
disapprove
of bullving beliavior
or say
that rhey
cannor undersrlnd
u,hy
sonre
childrel
bully others
[e.g.,
Boulton and
Underrvood,
1992:
whirnev and
Srnirh.
1993]-is horv
they acrually
behave
in buJIi,in-e
siruations.
Even rhe
ignorine
of uhat is
going on be-
t\\'een the
bulll'and his/her victirl rnay
be inrerprc.tcd
by rhe
bully as
approval
of his/
her
behavior.
Not only do the roles
of the
bu llies and
the
vicri
rns
constitute
an intportant
elenrent
in the bullying process,
but so,
too, do the
roles
of "the others." \!'e refer
ro
thcse roles
^slhe Participatlt Roles
of the
children.
Ri-lby
and
Slee
il9931 suggesr
rhar.
anrons
school
children,
rhere
are rhree
dinren-
sions
of interpcrsonal
relations,
rellecting
(endencies.
first,
to bully others,
second,
ro
be victinrized
by orhcrs, and rhird, to rehle to orhers
in a pro-social
and cooperarrve
manr)er.
Their results
support
rhe lactorial
independence
of these
dinrensioni.
Hou,-
ever.
in a bulli'ing siruation.
prosocial/cooperative
beha'ior could irnpll,
many things,
e.e., takins
sides *'ith
the
victinr,
stayine
unin'ol'ed
and nor
ensa-sinq
in acrive
bully-
ing
behavior,
or perhaps
even
cooperatine rvith
rhe
bully.
whitney
and
Smith
[1993]
studied
school
children's
arrirudes
ro
bullying.
The
pupils
u'ere
asked
u'hat
they usually
did rvhen
they
sa\\/ someone
being
bullied.
and
ruheiher
they
might
join in rhe
bulli,ing.
About half
of rhe.junior/nriddle
school pupils
reporre<i
lhat the-)'u'ould
rry to help rhe
r,icrim,
rvhereas
only a third of the seiondary
ichool
pupils
felr this *ould be
likely.
of rhose
pupils uho reponed
rheir likelihood
of doing
nothin-s.
sorne
feh they oueht
to help rhe vicrinr.
*hile sorne
rhought
it ri,as
none
of
their
busincss. The
rnajority
of rhe pupils
clid
not rhink
rhey
u'ould
join
in
rhe
bullying;
only'abour
one-fifth reponed
that
rhey
rnishr
do so. Attitudes.
horlever.
tlo not neces-
sariiv corelale ri'ith
rhe
actual
beha'ior
ofrhe
children
in
bullvine
situations.
Bulllirtg
as o Group Process 3
The Panicipant Role an individual
child
takes in a bullying situation is undoubtedly
deternrined by many factors,
personal
as
rvell
as contextual. One
ofthese factors may
be the social
status the child has established
in the
gloup.
The behavioral
pattems
pos-
sible for a
child
dcpcnd on his/hcr social
status.
For exanrple. a low-status
child
Inight
be allaid of being
victirnized
himself,
if he/she
takes
sicles
rvith
the victim. Recipro-
cally, the behavior
of rhe child obviously
has an effect on his/her status
in different
rvays. We wanted
to examine the
connections
bet\vcen
the Panicipant Roles
in the bul-
lying situation
and the sociometric status
of the
childrcn.
Bullying apar1. literature describes
plenty
ofstudies
on children's
group
behavior
in
gerreral
and its correlates
with their sociometric
status.
It has been repeatedly
noted
that
a child's social
approval
(positive
status)
is connected
with obedience
to rules,
with
friendliness,
end
with prosocial interaction.
Bcing socillly rejected
(negative
status)
is
connected
u,ith
ag-r;ession and destructive
behavior
[Coie
ct al.,
1990].
Lancelotta
and Vaughn
[989] studied
the
relations bet*'een
different
subtypes of
aggressive behavior
and the
sociometric
status ofthird- and fourth-grade
children. Sig-
nificant negative correlations
u,ere
found between social
status and all categories of
aggressive behavior
for both sexes,
with the exception
of provoked
physical aggression
in boys.
indirect aggression
(detlned
by them as tattling, stealing. damaging
others'
property) was
the
t)'pe
of aggressive behavior
that
correlated
most highly
with low peer
ratlngs.
Perry et al. [988] studied 9- to l2-1'ear-old
victinrs of peer
aggression.
Children's
victimization scores
were negatively
conelated
with peer acceptance
and positivell'
correlzrted
rvith peer
rejcction.
In the study by Lagerspetz et
al.
I
I
982]
the bullies
were found to be less
pooular
than
both controls and
rvell-adjr.rsted
children.
although not so
unpopular
as the
victims.
In the study by Lindman
and
Sinclair
[988], both
bullies
and
victims
u'ere
found
to
be
less
popular
than olher children.
However,
the fernale bullies
constituted
an
excep-
tional group, rvhose popularity rvas
surprisingly
high (rneasured
by questions Iike
"u'hom
rvould you Iike to sit next lo you in the
classroom?" or "who would be
the leader
in the
class,
if the teacher
rvere
absent?").
Even if some characteristics
typical
of leaders
u'ere
linked to male bullies as
well.
thcy
were not accepted
leaders.
Coie
et al. [982] have suggested
that
instead of two t1'pes of social
stalus,
positive
and negative.
there are five status
types: rejected.
neglected,
averaqe,
popular. and con-
troversial. In their study.
popular
children',vere
vieu'ed
by peers in obviously
prosocial
terms.
uhile rejected children
had an opposite
profile: they
rvere vieu'ed as
disruptrve
and as liable to start
fights. Cotttroversial
children.
u'ho rvere
both hi-ehly
liked and
highly disliked.
u,ere perceived by their peers
as disruptive
and aggressive
(like the
rejected
group)
but also
as social
leaders
(like
popular
children).
The authors
describe
them as visible.
active. and assertive
children.
In a l'unher study by Dodge [19831'
controversial bol's
rvere found frequently
to engage
in both prosocial
and antisocial
behavior.
In a study by Boulton
and Smith
[994]. both bullies
and
victinrs
were
found to
be
overrepresented
in the rejected
status
group.
and to be underrepresented
in the
popular
-eroup.
relative to not-invcrlvcd
children.
In addition. socionretricrlll'rejected
boys
re-
ceived
significantlv nrore bulll irrg notrtinalions
thart
popular. aleraqe.
neglected,
and
other status bol's. Both rejected bo1's
and
girls received significantly
nrore
Victinliza-
tion
nonrinations than
popular. averilge.
and olher status
bo)'s
and
girls.
1 Sultttttulli
tt ul.
The
purposc
of lhe
prescnt
study
$'as
to investigate
the fbllorving
questions: I
)
What
kinti of Participrnt
Roles-in addition
to the traditionalll'studied
roles of
the bully
arttl
rhe
victinr--<io the children
take.
orhave
assigned to
thcrn, in the bulli'ing
situation?
2)
llorv
urllure childre
'r arvarc of
Lhe irorvn Participant
Roles?
3)
Horv
is the social
status
of
childrcn rclrrtcd
to rheir Particitrtnt
Rolcsl
METHODS
Puoils
r\ltogerhcr13classesl'rorn
llFinnishschoolsrreresurvelecl.Thepupilsparlicipat-
ing in the studl,rvere 573
sixth-gradc children
(2,!6
girls,287
boys),
aged
12-13
years.
Class
sizes
varicd frortr l9 to 32 pupiis, nreart cluss size being
24.9.
Questionnaire
The pupils filled in a
questionnairc consisting of l) evaluations
of each
child's be-
havior
in builfing situntions;
both
self- atrd
peer-estintates.
from rvhich
the Participant
Roles
u
ere dcrivcd:
2) identitication
ofthe
victinrs
bullied
by others;
and
3)
a sociometnc
part.
Bcfore starting
lo fill in the
questionnaire, the
puprls
$'ere presented with a definr-
tion of bulll,ing
as
"one
child
being
exposed repeatedly
to harassment
and attacks
flotn
one or sevcral
other children:
I)arassrnent and attacks
rtray be, for example,
shoving
or
hitting rhe other
Lrne, calling nanres or
rnaking
jokes
of hinr,/irer, leaving
him,/trer outside
the
group.
taking
his,4rer things, or any other behavior
rreant to hun the other one."
Participant
Roles in the Bullying Process
In the first
part
ofthe questionnrire. rhe
pupils
evlluated
(on
a three-point scale:
0 =
ncver,
I = sonretimes.
2 = often) horv
u'ell each child in their
class,
including
thern-
selvcs.
fit -50 bull)ing-situation
bchavioral descriptions
(presented
in the Appendix).
Fronr thc 50 iterns. fii'e subscales describing
tendencics
to act as Bulll', Reirtfott'ar
of
tlrc bullt,
,\ssistant oJ'
rltc bullt'. Defendet' tf thc rictim, and
Outsider
were
fonned (two
iterns
uere dropped out
in order to raise the reliability ofthe scales).
The itenrs of rhe Bully Scale
(a = .93) described
active,
initiative-taking,
leaderlike
bulli'ing behavior.
Thc itenrs oftheAssistant Scale
(o = .81)
also described aclive.
but
rnore
lbllou'er than leaderlike builying behavior.
On the Reinforcer Scale
(o = .91),
therc ucre items describins
tendencies to act in ta1's uhich reitrforce
the builf ing
behar
ior. like lauchine. conrins lo see
u'h:.tt
is happening,
and being
present. Ihus
pro-
vidinq rn "uudicnce"
for the
bullv.
inciting
the bulli,,
etc. On Ihe Defender Scale
(cr
=
.93).
rhe itcnrs describcd supporlive. consoling
side-taking
rvith
the
victint as
well as
active
c1'forls to nrrke
others
stop
bullf ing. The itenrs of the
Outsider
Scele
(a = .39)
describcd
"doing
nothing."
staving
outside
the builf ing situations.
Tlrc pccr-c:tinrrtcrl
scores on the Bully, Reinforccr. Assistunt,
Defender.
and Out-
sidcrSceles\\ereslar)dlrrdizedbl'class(i.e..themeanscoreof
eachclass=0,SD=
l).
The scores
rrere
used to idenrit,r,children
u,ith
corresponding
bulii,ing-situation
roles.
In thc follou ing
u
e
re llr to thcse ls Participant
Roles.
For instanct-. a child
ri'as
considcred to
have
the Participant
Role of bein,e
a Bully.
il
l) hc/she scLrred
above the nrcun
(0-00)
on tlre standurdized
Bully Scale,
and 2) he/she
scored
highcr on lhrt scale than
on any of the other scales.
Since sonre
chiidren
received alnrost
equal scores on tso or more scales. an addi-
Bulltittt
Lls a
Grlup Proce
ss J
tional criterion had to be
es(ablished:
ifthe difference
benveen
a pupil's hi,qhest
score
and his/her second
highest score
\\,as less than 0.
l. he/she
was
regarded
as not havins a
clearll,definable
Participant
Role. Those
not scoring
abore the ntearl on any ofthe
scales
s'ere
equally consitlered
not to have
a Participant
Role.
Tlle
percentage
of pupils
rvirhout
eny l)articipant
Role
u'as
12.77r.
ldentif ication of Victims
The
victirns were picked out separately by rhe
pupils; on
the second
part
ofthe ques-
tionnaire,
rhe
pupils $'rote the nanres ofthe peers they thought
ri'ere
the
victinls
bu-llied
by the others.
If 307c
or
nrore
of the classmates
named
somebodl'
as a
victir1l. that
rvas
considered
to be her/his
role, regardless ofhorv she,he scored
on
the Bully,
Reinforcer,
Assistant,
Defender. and Outsider
scales.
The idcntification of the
victims
in this
way rvas
based
on
the assunlption
that being
a
victinr
rvas.
after
all, the
"prirnary
role''
of a child so assessed
by 30Va or
nrore of her/
his
peers, no rnatter horv she,/tre
behaved in a situation
rvhere sotneone else
was the
target of attacks.
Sociometry
In the sociometric part,
the pupils rvere
asked
to nominate
the three
female and
the three
rnale classmates
u'hotn they liked the most and the three
fenrale
and the
three
rnale classnlates
rvhom they liked the lerst. These norninations
\\'ere
stan-
dardized
by class
(see
above), and
the
variables social acceptance
(like
rnost nontt-
nations) and social rejectioft
(like least nonrinations) we
re formed. Furthermore,
the
pupils
rvere
divided into the status
groups
ofpopular.
rejected.
neglected,
con-
troversial, and
avera-9e suggested by Coie
et
al. The procedure
was
the sanle
used
by these authors
[Coie
et al., 1982,
p. 561]. The popular
status
group (n = 116)
consisted of children
with rnany
positive and only ferv negative
nominations.
On
the contrary, rejected
children
(n
= 137) had few
positive,
but
nrany
negative nomr-
nations.
Both
positive
and
negative
nominatiot.ts of the average
status
group
(n =
156)
were
near the
nrean. Neglected
children
(n = l5) had ferv
positive and
few
negative
nominatiolts,
while
controversial
children
(n
= l9) had many
positive
and
many
negative nominations.
RESULTS
Participant
Roles in the Bullying Process
In accordance
rvith
the
procedure described.
it ri'as possible to assi-cn
a Participant
Role to 87% of the
pupils.
The rnost cornmon
Participant
Roles
u'ere
Outsider,
Rein-
forcer.
and Defender.
There
nas
a slatislically
higher significant
sex difference
[i(t
(5)
=239.5'P <.001
] tn
the distribution
of the Participant
Roles.
There rvere r.nore Defenders
anong the
girls
(30.
l9'r of rhe
girls.4.5% of the bo1's) as
uell irs Outsiders
('10.27c
of the
-qirls.7.37c
of
the bo1's).
',r'hile
anronq the boys,
the Participant
Roles
of Reinforcer
(37.17r
of the
boys.
1.7%
of the
girls)
artdAssistant
(12.2?i
of the bo1's,
l.'1% of the
girls)
were
rnore
frequent.
N{ore children
uere also designated
as Bullies
rnrong
lhe boys
(10.5%) than
among lhe
girls
(5.9%).
The frequency
of the
Victirns
was
about
the same
for both
boys
(l LE9t)
and
girls
(l 1.57c, Fig.
l).
t
Stlntivalli
et ol.
n
160
1
40
40
20
120
100
BO
60
Victim3 Bulli.s Reinlorcers
E cirls UZ eoys
Fie. l. Percentual distribution of thc subjccts
(girls
and
bo1's) between
the different Participant Roles.
Self-Estimated
Behavior in a Bullying Situation
The self-estimated scores
on the Buil),, Reinforcer, Assistant,
Defender,
and Outsider
scales
rvere positivel),corelated rvith the corresponding
peer-estimated
scores.
Self-estinrated
scores on the Bully Scale
rvere associated not only with the peer-
estinrated Bully Scale, but
also
ivith
the
peer-estimated
Reinforcer andAssistant Scales.
Respectively, both the self-estimated Reinforcer
and Assistant
Scales
correlated
posi-
tively with all the
peer-estimated
scales Bully, Reinforcer, and Assistant. The self-esti-
nratcd scores on the
Defender Scale
conelated
positively not only rvith
the
peer-estimated
scores on tlre Defender Scale.
but also
rvith
the Outsider Scale.
The self-estimated
scores
on the Outsider scale
correlated
positively u,ith
the
peer-estimated
scores on the
De-
fendcr Scale. Intercorrelations of all the self- and
peer-estimated
Panicipant Role scales
are presented
in Table I.
TABI-E I. Correlations
Bet\reen
Self-
and Pcer-Estinrations on the Participant Role
Scales
Sclf-estimatcs
Bullv Reinforccr Asststanl Defendcr Out si
de
r
Peer-cstirrrllcs
B ulll'
Reinforcc'r
Assi
stant
De
fe ndcr
Outsidcr
.10** .1J"*
5l.* ..11**
-19-" .18*"
11r* tl*
l!t** ,.20**
.16'-
l3**
l9-
*
_.19,-
..lo*'
.ll+'
-.
I 8**
.n*
.l:* *
*P=0.0-5.
**P=0.01
(two,rrilcd
rests).
l:::..,:::)
i .'. .. .
ri
Btrllyirtt
us rt
Crottp Prott'ss 7
The scale rreans
of the self- and peer
ratings on the Participant Role scales
are
presented in Table II. As thc table reveals, self-estin)ared
scores
were siqnificantly
lou,cr than
peer-cstinrated
ones on the Bully Scale. This sugqcsts
that
pupils
underes-
timated their tendency to act as
bullies. On
the other hand, self-estinrated scores turned
out significentiy,
Irigher than
lhe
pcer-estinrated
ones on
the Reinforcer, Defender, and
Out-
sider Scales. Obviouslt,,
pupils
overeslinrate(l
their tendency to act in these
u'avs.
On the
Assistant Sclrle. there
rr,as
no signillcant dillerence Lret* een self'- and
pecr-estimated scores.
As Table II shorvs, self'-estinratcd scores differed on rnost sclles
significantly
frorn the
corresponding
peer-estimatiorls.
Horvever.
rve
had a closer look at how children
tt'ith
dif-
ferent Panicipant Roles actually evaluated their orvn behavior in bulll ing situation.
Bullies, Assistants. and Reinforcers scored
higher than thc other
status
groups
on
the self-estjnrated Bulli' Scale
[F(5,.12
l) = I1.34, P < .001].r On the self-estinrated
Reinforcer
Scale.
Assi
stants.
B ull ies. and Reinforcers had the highest scores
tF(5.12
I
)
= 13.37,
P <.00ll.: On the
self-estinraled
Assistant
Scale,
the
ones
scoring highest
were
again Bullies, Assistants, and Rcinforcers
[F(5.a21)
= 9.-56,
P < .001].r
Dct'end-
ers, Outsiders.
and Victirrs
scored highest on thl:
self-eslimated
Defcnder
scale
IF(5,,12.1)
= 10. ll,P<0.0011.'Ontheself-c-stimatedOutsiderScale,OutsidersandDefenders
had the highest scores
[F(5,422)
= 5.87,
P <.001].5
Of all the
children identified as Victirns.49.3%
assigned themselves
as
victims,
23.9Vantmed someoncelseasbeingavictim,
and26.9Vo didnotfillinthispartof the
questionnaire
at all. That is,67.4Va of those
victinrs who responded to this
particular
question,
merrtioned
thenlselVes
as
victirrs.
Secondary Roles of the Victims
The Participant Roles
(Victinr,
Bullr,, Reinforcer, Assistant, De fender, and Outsider)
were. by definition, considered as mutually exclusive. It rvas.
horvevcr,
possible
to
exanrine the secondary roles
of the
victirns-i.e., horv the
victims
bchaved in a situa-
tion in rrhich someone else
rvas
attacked. It should be remenrbered that the
victinrs
were
picked
out by a separate
peer-rrorlination
procedure.
The scores
of the victirns
on rhe Bully, Reinforcer, Assistant, Defender, and Out-
sider
scales
u'ere
conrpared
rvith
the scores
of nonvictirnized children
(those
nol as-
si-ened as
victinrs
by anl,one). The results su_egest that, compared
rvith
thern.
the victinrs
scored significantly higher on the Defender Scale'
[t(77.0.1)
=2.59,
P < .05] and on the
Outsider
Scale:
It(-14-5)
= 2.01, P <.05]. but sli*ehtly-although
not significantly-
lower on the
Bully,s Reinforcer,o and Assistants Scales.
Among all
Victims,
rhcre
rvas
only one
girl (3.07c
of the fenrale
Victinrs)
and three boy s
(8.87o
of the
nrlle Victirns), uhose actual
"secondary
role" *'as that of a Bully, and
riho
thus can be tlrought to concspond
to the
bull1,/victinrs
described in the literature.
Tukev
s
honestll'-.isnificrnt
difference
(HSD)
tesl.
(,05
Icvel):
'Bullies
>
Outsiders.
Defenders.
Victirns.
Reinforcers.
Assistanrs
>
Outsiders. Defenders
rAssistrnls.
Bullies
>
Outsiders. Defenders.
Victinrs
Reinforcers
>
Outsiders. Deftnders
rBullies
>
Outsiders. Defendcrs.
Victims.
Rcinforccrs.
Assistants
>
Outsidcrs.
Dcfentlers
Reinforcers
>
Outsiders
'Defendcrs
> Reinlorccrs.
Bullics. Assistants
Outsiders.
Victinrs >
Rcinforcers
uulrrder\ > bullrt\- \ rLlrnr\
Dcfenders
> Bullics.
\'Iem scores of r,ictirns
and
ncrnlictinrs
on the
fir'c
scales:
rrjctirns
.1J6.
n(rn\ictirns .108:
:rictirns
.191,
nonviclinrs l-19:'rictirns.07{.nonrictinrs.0S-l:\ictinrs.195.nonrictinrs.l96:1r'ictinrs.0S8.nonlictims.i09.
8 Sulntivulll et al.
T,.\BI-F]
II- Scalc
)Ieans of Self-
and Pcer-Estinrations.
and the Significancies
of Thcir Differences
Scale Self-esl
i nrate l)eer-estimate tdf P<
0l
001
n. s.
001
001
Bu tl1'
Re i nforcc r
A ss it ant
De
fendcr
Outsider
0.06
0.26
0.r0
0.27
0.33
0.09
0.21
0.l l
0.
l3
0.26
t.63
1:6
0.7 3
1t.66
1.51
.19
I
i9r
.19
I
.19.1
192
Participant
Role and Sociometric Status
When the connectiolr bet\\'een
Participant
Roles
and
sociometric
status
in the class
was
examined,
the fr;llorving
aspccts
were
taketl into
corlsideration.
First,
the children's
social
acceptance
(like rnost nominations)
and social
rejection
(like
least
nominations)
\\'ere
cornpared
rvith
their Participant
Roles. Second,
the children's
Participant
Roles
were rclated to their belonging to one of the five status
groups (popular, rejected,
ne-
glected,
controversial,
or average).
In the
final set oi analyses, behirvioral
profiles were
fomred for children belonging to the different status
-qroups.
Social Acceptance and Social Rejection
A significant
rnultivrriate etfect of Participant
Role
[F(
10,976)
= 9.99,
P < .001]
rvas
found. In the
univariate analysis, a significant
effect
of Panicipant
Role appeared in the
case of both social
acceptance
[F(5,.19)
= 6.1'1,
P < .001
] and social rejection
[F(5,19)
=
19.03, P <.0011 of the child.
---
Victims Bullies ReinforcersAssistantsDefenders
Oulsiders
IE Girts i-'l eovs
Fi-q. 2. fr'lcans for social acceptance ofgirls and boys
in the
diffcrcnt Panicipant Ro)es
s,
.:
:::
I
I
I
)
lI
I
I
-1
l
I
I
l
0.6
o.4
o.2
0
-o.2
-
o.4
-
0.6
-
0.8
-.i:!!9-rt"l99l=
I
l_.
wf'v
core
i
I
I
I
I
I
I
l
i
I
I
II
I
li
II
t.c
1
u.c
0
-
o.5
Bullying as a OrouSt Process
social rejection
(z-scores)
Victims Bullies. ReinforcersAssistants
Defenders Outsiders
E ctrts Z eoys
Fig. 3. \,leans for
social rejccrion of
eir)s
and boys in
rhe different
Panicipmt Rolcs.
A significant interaction benveen sex and Parricipant Role appeared iF(10.976) =
2.22), P < 0.051 in rhe
mulrivariate tesr.
Furrher univariare
analysis
reverled a signifi-
cant inferaction
benveen sex and Participant
Role in regard
to social rejection
tF(5,49)
= 3.48, P < 0.0051.
There
was
no signitlcanr inreracrion
of Panicipant
Role and sex
in
regard to the
social acceptance
of the child, ho*,ever.
The standardized
score
means for social acceptance
(like most nominations)
and
social rejec!ion
(like
least nominations)
ofthe children with different
Participant
Roles
are
presented
in Figures 2 and
3.
Both male
and fernale Victims scored low on social acceptance
and
high on social
rejection-they rvere clearly rejected
children. Orher children rvirh a sirlilar partem
(low acceptance,
high rejection) were
nrale Bullies, female
Reinforcers,
and female
Assistants. Female
Bullies,
however, scored
above average
both
on social rejecrion and
social acceptance-seenring
to have a
profile
like rhe
children describcd
as controver-
sial by Coie et al. [982]. \4ale
Reinforcers had a profile like popular
children
(high
acceptance,
lorv
rejection), while nrale Assistants
scored
near average
on both social
rejection and social acceptatrce. The children who scored
highest on socirl acceptance
(scoring
also low on social
rejection),
hou'cver,
\\/ere
the Defenders
ofthe victinr.This
was
true
in the
case
of borh
sexes.The Outsiders
(borh
female
and male)
scored below
average
on both social acceptance
and social rejection.
Status
Groups
Follo*'ing the procedurc
suegested
bl,Coie et
al.
f19821,473
pupils u,ere
selecred,
frorn
the original pool
of 573, as firring one of rhe five
social srarus
groups
(popular.
rejecled. neglccted,
controversial,
and average).
!0 .Sulntiyalli cr ul.
TAIII-F, Itl. Distribution of Children in thr DilTerent
Participant
Roles in Four
Status Groups
(Populrr,
Rcjccled.
Contror ersial. and ..\r
cragc)
Prnicip3nr role
Stulus Vic
t inr Rcin forcer/Assi stant Dcfendcr Ou t s ider
Populrr
Rejcclcd
Controvcrsirl
Averarc
Toral
3.61
11 .4(,t
-1.6tt
2l.lcio
r00(a
I
fl.::
{1
-i
I .3C;
l().t%
]t.lt;
I 00tt
19.0(;
r9.5(,i
1.lt/c
31 tq.
I 00ci
.13.0ta 29.ic,
2l.sct :0.8%
3.89c 1.01c
10.+9e 45.5%
I 007c I 00'Z
Sincc the nurnber
of childrcn
selected in the
neglected status
_qroup
was ver),srnall
(n
= I 5; only I 2 of thern had
a dcfinable Participant
Role), rhat
group was
dropped
fiom
the
chi square anal),sis
(thrce
Victims,
lbur Re
inl'orcers, one Defender, and four
Outsid-
ers bclonged to the neglccted
status
croup). in addition,
rhe Reinforcers and theAssis-
tanls
\\'ere
truncaled
to one catcgory
due to their small cell frequencies.
The analysis
reveeled
slatistically si_gnificarrt
differences bet*een rhe
status
groups
of the children
with
different Paniciparrt
Roles
[1:
(12)
= 31.9, P <.001].
The distribution
of children having
difl'erent Parricipant Roles in the
starus
groups
of
popular,
rejected, average, and
controversial are
presented
in
Table
III.
The controversial
status
group
\\,as
not thc most
tvpical one for any,ofthe role
groups,
*hichispartl),duetotherelativell'smallnunrberof
rhepupils(n= l9)belongingtorhe
contror,ersial
eroup.
To shed more ligitt
on the behavior
ofcontroversial children, among rhe
olher status
groups.
the
behavioral
profiles
(peer-estinrated
scores
on tlre
scales Bully, Reinforcer,
Assistant. Defender, end
Outsider) of the five
status
_groups
rvere
exarnined. The results
are
presented
in Table lV.
r\s can be seen in Tablc lV there l'ere sisnilicent differences betueen the scores
of
childrcn
ri'ith
different
status
groups
in all
scales e)icept the
Outsider Scale. Controversial
children
scored above the meirn
on
all
scales except
one
(the
Outsider Scale). On the Bully,
Reinforcer.
xnd Assistant
Scales they scored
significantly higher thut several orher status
groups (Tukey's
HSD lest).
It sc-erns that
thhou-ch engagrng in acrive
bullying behavior,
the
controversial
children also tend
to act as Defenders
of the victim or as Reitrforcers and
TABI-l- I\l Peer-Estinraled
Scores on lhe Ilultr', Reinforcer.
Assistrnt. Defender. and
Outsider
Scales as
a Function
of Pecr Group
Social Slatus
Sociome tric
slarus
group
Scalc
(z-scores)
Assislrnl Dc fendcr Outsider
Bully' Reinforcer
l. Popular
(,\=1.16)
2. Rcjectcd
(N=137)
3.
Neglccted
(N=
I
5)
.1.
Contror e rsirl
1\= I9)
5. Arcrrrc
(N=l-56)
F
df
P<
Tuker''s IISD test
(.05
Ier
cl t
0.05
0.1.1
0.c].1
0. l7
0.0tr
t0.
l-r
{
.001
:>1.5
{>1.5
0.10
0.t-l
0.rs
0.16
O.
IE
-1
.(15
0.09
0.
r5
0.06
0. l7
0.09
5._r I
l
.()0
I
2>t.5
0.r3
0.t2
0.07
0. l.l
0.tr
?.93
.l
.05
0.26
0.26
0.15
0.1.1
0. l6
0.07
4
n.
s.
Bullting us
a
Grotrp
Proct'ss l1
Assistants-but
not as
passive
Outsiders.
The results suppon
lhe
vierv
of thent
as
visible,
lctive.
and
asserlive children,
as they
\\,ere
described by Coie
et al.
[
1982].
The behavioral
profiles
of the other
status
groups
conoborale the results
already
presented in Figures
2 and 3 and
in Table III.
The ones
scoring highest on
the Defender
Scirle
rverc popular chrldren. Surprisingll'.
rlso rejectcd children
scored
relalively
high
on the
Delcnder Scale;
this ntay be
due to the
Victinls'tendcncl'lo act its
Defenders,
s'hen
sontebody
else is harassed.'l'he
ones
scoring
high on the
Bully' Reinforcer,
and
Assistant
Sclles
w'ere
rejected children
(in
addition to controversial
children,
u'ho
scored
higher
than all others
on these scales).
DISCUSSION
The results support
the notion
that bullying
nray be
regarded as
a
group phenonrenon
in ri,hich
most children of a school
class have
a dellnable Participant
Role.
Boys seenred,
irt
general,
to be nrore
actively
involved in the
bullfing process:
therr
most frequent
Participant Roles
were
those
of Reinforcer
and Assistant,
rvhile anrong
the
girls
the Participant
Roles of Outsider
and Defender
were the most common
ones.
The boys
\\,ere also ntore frequently
designated
as Bullies
than
rvere the
girls'
which
is
a result obtained
in several
previous studies
as
u'ell [Lagerspetz
et al.,
1982; Boulton
and Undenvood,
1992;
Whitney and Smith,
19931.
Several
interpretalions of these dilTerences
are conceivable.
For boys,
at least
physi-
cally aggressive
*'ays of being together
are nrore
comnlon and also more approved
of---€\'cn
expecte(l: bo1's use
aggression
to cre3te soci:rl
ordcr. To be
accepted
in their
peer
group.
bol s are expccted to
join, at least
to sonre extent.
in rough-and-tunrble
play'
rnutual
"testing."
and bully,ing behavior.
The idealization of aggression
antong bo1's,
but not among
girls,
has been found in several studies
[e.g.,
Bjcirkqvist
et al., 19821
Rauste-von-Wright,
19891. Cirls. on the
other hand,
are expected
to behave
in more
prosocial,
caretaking,
and helping
ri'ays-it is a
part
of the fenrale
social
role. Sex dif-
l'erences
in aggression,
as
well as in helping
behavior,
can
be
explained
referring
to the
different
social
roles
of males and
lemales
[Eagl1'.
I 987]. A further explanation
for the
different Participant Roles of the bo1's
and the
girls
nright be the
girls'better developed
ability
for empathy
[e.g.,
Hoffnran.
I
977
]. Accord in-e
to
Envin [
1993.
p. 168], the school-
aged
girls'intinrate. dl,adic relationships
lead to a style ofmoral reasoning
ivhich
enr-
phasizes
en)pathy
and sensjtivity.
The results on self-estimated
behavior
in bullf
ing situations suggest
that the
children
\.\,ere
moderately
u'ell
au,are
of their own Panicipant
Roles. Bullies,
as
u'ell as Rein-
forcers
and
Assistants.
seemed
to realize
their central role in the
process, even ifthey
underestinrated
their
participation
in active bullying behavior
and
rnostly claimed
that
they acted as
Reinforcers. This
explains the fact
that
peer-estimated score
means on the
Reinforcer Scale
u'ere
higher than self-estintated
ones;
not only Reinforcers,
but also
those
assigned as Bullies
andAssistants
qave
lherrtselves
high
scores
on
the Reinforcer
scale.
Defenders and Outsiders
u,ere also auare oftheir prosocial
and/or
passire role.
In all Participant Role
groups
the
pupils tended, ifconrpared
to peer-estimations.
to
underestinrate
their ag-eressive behavior
and entphasize
prosocial and
u ithdratving
be-
havior. both of which are nrore socialll'desirable.
Ostennan
et
al.
[199'1]
hai'e
refened
to
a
"sell'-serving
attribution
bias" explainine
the
differences in self'-
and
peer-estimated
l2 Sulntitttlli
er
ul.
aggressive
bchavior:
individuals
nrake
attributions
rvhich favor their self-perceptiorl
airJ
support
their sclf-esreern.
Thc senre
rnechanisttt
seems
to appear in the results
of
thc
prcsent
studv
as
rvell.
Alntost olte fourrh
(2_3.9ft)
of rhe
Vicrinrs
(i.e., those
$'ho rvere
assessed so by at
least l09t of thcir
pcels)
did noI ntentiorl
thctnselvcs.
but sonleorle
else. ls a victinl.
One.
pcrhxp5 unlikclv.
reasorr tirr
this r|a),be
tlrat thel,re3lly
do not have the
subjective
expericnce
of being
harasserl.
r\nothcr
possibilitl,
is that the
\/ictinls denl'the
fact
that
they
ere bLrllieti bl,ihcir pcers. It,shguld
bc notccl thet
almost
one third of the
Victinls
did nor ans\\,cr rhis
perticular
question at all.
s hich can
be seen
as
a funher sign
of this
denial/rcpression.
Tlis brings up
rhe n)crhodological
question about
* hether
the
I'ictints
ofbullf ing shoultl
be
ic6ntilled bl,self-rating
or pecr-raring
procedures.
[n sonte
earlier studies,
tlre
victints
hat,e
bcen idcnriilt'd
bl,asking the
subjr'cts
thettlselves,
uhether
c)r
not they
have Lrecn
bullitd [e.g..
Bou]torr and Undenvood.
19921.
rvhile
some
hilve used
peer-assessmenl
pro-
cedurcs
alone
[e.g.,
[ren1,et
al.. 1938]
or t()gethcr
$,ith
terclrer's
nominations
[olueus.
199-tl. Our result
speaks for the
pecr ritlings
as a superior
procedure to selfratings:
using
peer
ratings.
the
Victints'tendenc)'to den)'their
situation
does not
have any
effect'
When sonreone
else
rvls attacked,
tlre Mctirns lended
either
to defend
hinr,4rer or keep
thenrselves
outside. So rhey did not, \\'hen
thcy got an opponunity, eagerly
join in the
bulll ing.
The Panicipant
Roles ofthe Bully
and the
Victim
\!'ere
clearly separate:
only
ferv
children
could be assigned
as bullr,/victinrs.This
is in
accordance
ri'ith
Olu'eus'view
ofthe
Mctirls as not
aggressive
or teasing. but
ntostly
pxssive and submissive
in
their behavior
[Ol$eus.
1991,I991].Alsoinastudy,bl'BoultonandSrnith[199-1].nonrorethur'1.-l7cof
the
chiltlren
could bc tlesignated
as bulli,/r'ictirns.
Funher
studies
of the behavior
of the
Victims arc in
progress
in ourresearclt
group
[Sulnrivalli
et al.' in press].
Among both boi's
and
girls. \''ictinrs hld a louer status thatr
lhe other
groups. They
scorcd
high in social
rejection.
lorv in social
acceptcnce,
and their
nlost fiequent
status
groul)
\\'as
rhat
of lieing
rejccted. This
corroborates
the resulls
of n]any
earlier studies
[e.g..
Lrgcrspetz
ct al.. 1982:
Lindman
and Sinclair,
1988:
Perry
et al., 1988]'
The
unpopullrity ofVictirns
can Lre seen both
as a cause and
a result ofcontinuous
bullf ing.
One reaso6 for rheir being
picked
on
and harassed
in the
first
place nray be
their origi-
nal
unpopularity'*'it[irt
the
group.
On the other
hend, as Oi*eus [1991]
pointed out
when tiescribing the
group
ntechanisttts
inVolved
in bullfing. there are
gradual
cogni-
tive changes
in the
perceptions ofthe victinr
b1'the
pecrs. As the bullying
continues.
the),
slurt to see the
Victirr as deViant.
\\'orlhless.
and ahnost desen'ing
of being
ha-
rassed:
along
u ith these cognitive
changes. the
victinl
becolres
even
lllore unpopular-
It
beconrcs
a social nonn of the
group not to like hinr,fter.
Dclcnders of the
victim had rhc highest
status.
This suppons
previous results
about
the
high-status
prosocially behai ing childrcn
hrve ICoie
et al., 1990].
Tu'o
interpreta-
tions are conceivable:
I
) Defenders
lrave a hi-qh slatus
just
because
they
react to
bullv-
ine in
that
panicular \\'a)':
dcfending
thc
victinr
is
appreciatt'd
b1'the
peers: 2) a
high-status
chilcl does not have to be
afraid ol-being
r
ictirniz,ed
himself. ei'en
if he
takes sides
rvith
the
victinr.
High
stltus enebles
the clcfcndins of
the
victinr.
The tuo interpretations
do not erclutlc cach other.
but
nlai'both have
an influ-
ence. Il'the l'irst
intcrpreialion
were
true
.
del'endin-e
the
victirl rvould
be
an effec-
tive *:iv ol'irnproving
one's
slatus
in the peers'e)'es:
it rvould
be a role \\orlh
taking
on. Onc
nright
thirrk that
in
the case of the
girls this ntakes
sense: many
girls
Bullyirtg
as a Croult
Proccss /-l
(30.1%) llke the role of Defender,
and that
kind of behrvior is
just rvhat
is
ex-
pected from
-qirls.
fleu,'ever,
llle re
\r'ere
no
nlore than
.1.59t
of the
bo1's
rvho
had
the Participant
Role of
a Defender.
For this retson.
lhe second
interpretation
nlisht better
erplain the high
status
of lltale
Dcfenders:
probably
onlv those bot's
rvho
haVe
a good
position
in the
peer
group hare the couri.lge
to take
sides
$'ith the Victit]t----e\en if it is "a-gainst
the
norm" ir.t the
bo)'s'
peer
group.
in the study by Lagerspetz
et al. [19321.
bullies
rvere
found to h1\'e
a low strlus.
It
u,ill be
renrenrbered. hos'ever.
that in
the
study
b1'
Lindrnan
and
Sinclair
[
1
988].
femalc
bullies
u'ere
found
to be high-status
childrcn,
although
male bullies
had
a lorv status.
French found
out in her srudies
I
I 988, I
990] that
the relationship
benveetr
aggressive
behavior
and rejection
\\'as
r.lot as
po*'erful anrong
girls as
it u'as
alnong
bo)s. In the
present
study,
nrale Bullies
u'ere
clearly
lo\\'-stltus
children,
scclritlg
lorv
in social ac-
ceptance
and high in social
rejection.
They riere not, ho\\'ever.
quite as unpopular
as
Victims. Female
Bullies, on the
other
hand. fonned ati exceptiorlal
group: thc)'scorc'd
above
the mean in both
social tcceptance
and
social
rejection.
One explanation
1'or
the
stttus
differences
of male and felnale
Bullics
could be
their
different rvays
of bullfing. Girls use ntore verbal and indirect fonns of a-egression
[Lagerspetz
rnd Bjdrkqvist.
1994].
which is
Inore hidden
and ntore
"socislly
sophisti-
cated"-mavbe for these
rersons also
more approved of-than the
disruptive, olien
ph;'sical
and direct bullying practised
b1'bo1's.
This explanation
does
not, horvever,
corroborate
the resuhs
ofLancelotta and Vaughn [1989], sho noted that
the use
of
indire'ct
aggression
correlated especially
strongly
with lo$' status.
Ma1'be
the female
Bullies are socially
and
verbally
smart
children
*'ho can choose
their
uords and
amuse the others by verbally--directly or indirectll'-attacking their
victims.Accorcling to
the descriptiorrs
given
by their
peers
[Salmivalli,
1992].
the
girls
in the "gang
of bullies"
illso
rate high in tenrrs of the curent )'ourh
culture:
they
are
''tough
girls''
rvho
knorv the ne\\'est
fashion
and the latest
idols. It is
possible tha!
they
are, even
if fri_ehtening.
also admired.
An intportant
practicitl
ilnplicetiorr
can be
drarvn from tlre results
of the
present study:
since
most chilcjren are someho*'involved in rhe bullf ing process.
interventions
in it
should be directed
not only
torvards
the Bullies and
the
Victinrs.
but
tori'erds
the
rvhole
lroup. For exarnple,
the interventiorl
rnethod
of Pikas
[19751
is based
on s]stcmatic
conversations
u ith bullies
and
viclirns, uhile the
"noninvolved" children
are
not lakett
into accounr. Olucus U991, 199.11 lal's
ntore ernphasis
on directing
the inten'entiotl
not only on the individual.
but also on the
group level
(in the lonn ofclass Incetings.
class
rules against bulllin_l
etc.). in our vierv,
the
children
in the
different
Participent
Roles,
lor exanrple. Outsiders
or
Reitrforcers,
should be made
use of u'hen
tt't'in-s
to
put
an end to bully
ing. It may be
that their behavior
is
easier to change
lhan
the
a,egressive
behavior
of the Bullies.
There
is a possibilitl' that through
these changes
even the be-
havior of the Bullies nright
be affected.
ACKNOWLEDGMENTS
This srudi'uas
financialll'supponed
bv
the E. Aaltorlen
Foundalion
grant to
Chris-
tina SalntiValli.
\Ve
thank
Jarkko
Kerhurrcn
lbr his help in the collectin-tr
and
scoring of
the data and
Dr.
Ceorge
l\{aude
for revision of the Enelish.
14 Solmiyulli er ol.
REFERENCES
Bjdrkqvist
K. Ekman K. Lrs!.rsp!'{z
K\1J
(
I98l):
B u i
I
ics
and
v
icr
i nrs: Thcir
cro picrure. idcal
cgo
picture
and
nonnative
ego picturc.
Scandina-
vian
Journal
of P:r'cholouv
ll:-i07-,1 I
l.
Boulton
\1J.
Snirh PK
(
l99J):
Bullr/ricrinr prob-
lcnrs in middle-school
chilcircn:
Srrbilirr,.
self-
perccircd
contpclcncc.
pccr-perceptions
and
peer
acceplince.
British
Journal
of Dcvelop-
mental
Ps1'cholo-11
12:315
,129.
Boulton
\1J. Undcruood
K (1991):
Bulll/vicrinr
problenrs
amon-e ntiddlc
sch<xrl
childrcn.
Brirish
Joumal
of Educalionrl Psvcholocr'
6l:73-87.
Coie J D. Dcrdce
K A.
Copporclli
l.l
(
I 982
):
Dimcn-
sions ord n [rs ol s(xixl sln]s:,A cRrss-age
f{NFrr-
tive.
Developrncnlai Psl
chologv
I 8:
j57-570.
Coie JD. Dodge
KA, Kupcrsnridr
JB (1990):
Pcer
grcupbchariorand
scial srarus.
In,\sherSR.Coie
JD (cds): Peer Rc'jecrion
in Childhood."
Cam-
bridee
Univcrsirv Press:
Crnrbridre.
pp t 7-59.
Dodce
K (198j): Behavioral
rnlecedcnts
of peer
smial sralus.
Child Der
elopmcnr
5t:
l3EG I
j99.
Eagly AH (1987):
"Sc'r
Differcnces
in
Social Be-
havior:
A
Social
Role
Inlerprc{alion."
Hillsdrle,
N.J.: Erlbaunr.
Eru in P
(
|
993):
"Friendship
rnd Per'r Relarions
in
Children."
Chicheslcr:
John
\\'ilev
& Sons.
Frcnch
DC 1l98ti):
Hcrerogencirv
of pccr-rejecred
boy
s: Agrressive
rnd nonrcgressive
subrvpcs.
Child Derelopmcnr
59;976_985.
French DC (
I
990):
Hcrcroreneirv
of pcer-rejecred
girls.
Child Dcr
cloprncnr
6
I
:1013-10_i
I .
Hoffman lr1L
(1977):
Sex differc'nces
in crnparhy
and relrted
bchariors.
Psvcholoricri
Bullclin
81:7 l2-722.
Lagcrspetz
K11J, Bjcirkqvisr
K (199J):
Indjrccr
t_e-
srcssion in bors rnd qirls.
In lluesnrann
LR
(ed):
"A-reressire
Bcharior.
Current
Perspcc-
lives."
\erv'tbrk; Plenum
Press. pp l_11-150.
Lase15pctz
K\IJ, Bjcirkqrist
K, Bcrrs
\1. Kine
E
(1982):
Croup arcrcssion
rnron-c
school
chil-
drcn
in thrce
schools.
Scrndinarian
Journal
of
Ps1'chology
2l:J5--sl.
Lanceloila
GX.
Vaushn
S
(
I 9S9):
Relalion
bclwccn
t)pes
ofatsressiorr
rnd
socionretric
st!lus:
peer
and leachcr perccptions.
Jountal
of Educrrional
Pslchologl,l:S6-90.
APPEf..DIX:
Ilcrns
rrf
Ihe Prrticiprnt Role
Scrles
Thc
Bullv
Scrle
l. Stans
bullr inr
?. l\lakcs
rhe
othcrs
join
in rhc bullr
inr
-1.
Alsa\s finds
ncrr
nrrs ofhrrrssing
rhc
rictinr
Lindman R.
Sinclair
S
(
I
988):
Social roles and
as-
pirations
of bullies
and
vicrims.
Paper prescnred
at rhe
Srh
\\/orld
Biennial
ISRA Conference,
Suansca, \\;alcs.
Jull' :-6. 1988.
Olweus D 1l99l): Bulll/r'icrirn
problcnrs
arnon_e
schoolchildren;
Basic facrs
and effecrs
of a
school-based
inrcn,enrion
prograrn.
In Rubin
K,
Pepler D lcds):
"The Der,elopnrent
and Trear-
ment
of
Childhood Argression."
Hillsdale.
N.J.:
Erlbaum, pp 4l l--1-18.
Olrveus D 1
l99J): Bulll
ing ar
schools. Long-rem
outcomes for
thc
vicrims
and an
effective
school-
based inlcncntion
progrrm.
In lluesnrann
LR
(ed):
"Agrressive
Bchavior
Cuncnt Perspec-
rives." Ne\r'York:
Plenunr
Press, pp
97-130.
Peny DG, Kuscl
SJ, Peny LC 11988):
Vicrinrs
of
pecr
aegression. Developntental
Psl,chotogy
6:807-8 l-1.
Pikas
A (1975):
"SC
stoppar
vi mobbning."
srock-
holm:
Prisma.
Rausre-r'on
Utight ivl
(1989):
Physical
and verbal
aggression
in peer groups
among
Finnish
ado-
lescent bovs
and
girls.
Inlernalionai
Joumrl
of
Behavioral
Developmenr
I
?:,173--18.1.
Rigby
K, Slec PT
(1993):
Dimensions
of inrerper-
sonal relalion among
Australian
children
and
implications
for ps; chological *ell-being.
Jour-
nal of Socirl Pslchologv
I33:33-{2.
Salmivalli
C (
I 992):
Bull) ing
as a
group
process.
\laster's thesis.
Deparrnrent
of Pslcholoer',
Universitl of Turku,
Finland.
Salnivalli
C.
Karhunen
K. Larerspetz
K\{J: Ho*'
do the
vicrinrs
rcspond
ro
bullt.ine? Agrressive
Beha!ior
(iD
press).
Smilh PK, Thompson
D (1991):
"Pracrical
Ap-
proaches
to Bullr
ing,"
London:
David Fulton.
whitne)'
I. Snilh PK (
1993): .A
sun,ey
of rhe na-
ture
and ertenl
of bullving
in
junior/middle
and
secondary
schools.
Educarional
Research
Jf:-1-I)-
Osterman
K. Bjtirkqvisr
K, Laeerspetz
K\.1J,
Krukiainen
A, lluesnrann
LR, Frsczek
A
(1994):
Pccr
and self-esrimrtcd
aggrcssion
and victimizarion
in 8-r ear-old
children
from
five crhnic
groups.
A-lgressive
Behavior
l0:41
l--118.
Bull;'in3 os a Group Proccss
'1.
Fetches
more pcople inro rhc
bullr
ins situation
5. Urges
the
others
to harass Ihe
vrctrnr
6. I\4akes
suggesrions rbour
bull\ing someonc
7. Calls
those
uho don t paniciprtc
in the builying
"crl'-brbies"
8.
\1akcs ironic rcmarks rbout Ihe
r
ictim
9.Sr;stolheothers:
he/sheissostupid,ir'sjustrichtlbrhirrrtcrtobeharasscd"
10.
Tclls others not ro bc fricnds
\\irh
thc
viclinr
The Reinforcer Scalc
l. Corrrcs
rruunJ to sce
thc
:iluition
2. Is usually,present. even ifnot doing rnlthing
3. Giggies
4.
Luughs
5.
Incites the bulll bl'shouting
6. Sa)s to
the
bullr':
"Shoq
him,herl"
7. Sa)'s lo the
olhers:
"Conre
to see. someone is br:ins harassed there!"
The.Assistunt Scale
l. Joins in the bulll ing,
u
hen sonreone else has staned it
2. Assisrs the bully
3. Calches
rhe victim
4. Holds the
victinr, rrhen
he/shc is harassed
The Defender
Scale
l. Sa)s
ro
the
victim:
"Don't
care about them"
2. Tells sonre adult about the bulll ing
3.
Threatens to tcll the tercher. if rhc others don't srop bull)'ing
4. Teils the
olhers
rhar it
doesn't
pal,tojoin
in the bullfing
-5. Savs lo
rhe orhcrs
lhat lhe bully is srupid
6. Conrfons the
viclim
in the bully ing situation
7. Atlacks the bulll in ordcr to dcfend lhe
victim
8. Takes revenge on rhe bull) for the
victim
9.
Calls thc bullies narnes in ordcr to defend the
victim
I
0.
Tells the others
lo slop bull)'ing
I l. Fetches
people
to come and help the
viclim
12.
Says
to the
orhers lhat bull)'ine is stupid
I
3.
Tries to rnake the others
stop
bulll,ing
14. Trics lo arbitrate
the differences by talking
1 5. Comfons the
victim aflenvards
I
6. Stal's
rvith
rhe victim durins thc
breaks
17. Coes to tell thc teacher about the
bull)ing
| 8. Encourages the
victim
to lell the lercher about the
bull),ing
19. Is friends
\\irlr
lhe
viclinr
durin-s
Ieisure
tinre
20.
Fetches the
lcach.r in charge
The
Outsider Scale
l. Isn t usualll present
2.
Strl
s ourside rhe
situarion
3.
Prctcnds not to
notice
*har is hrppening
.1.
Doesn'l do anlrhing
5.
Doesn t
evcn
lno\\' rbout
the bull)
ing
6. Dcesn't take sides
rvilh
anr one
7.
Coes
rurl fron thc
spot
Tu'o ilems
u erc dropped
fronr
the scrles:
l. Strvs
ncar and looks
2. Joins in rhe bullr rr
lrcrr the
others
tell him,&er ro
du
so
I5
