Resea ch on Child and Adolescen Psychopa hology
h ps://doi.o g/10.1007/s10802-025-01339-5
unde an imbalance o powe d i en by collec i e no ms;
(iii) which can be pe cei ed epea edly, ei he by he e-
quency o cybe -agg essions o e ime o by he se e i y
o hei nega i e e ec s; (i ) and which mani es s i sel as
unwan ed beha iou ha o en causes emo ional and social
ha m, including in he long- e m (UNESCO Chai , 2024).
Among adolescen s and young people, pos ing hu ul com-
men s o umou s, whe he once o se e al imes, a e he
mos common ypes o cybe -agg essions (And ade e al.,
2021), o example o make un o he ic im’s appea ance.
These cybe -agg essions can be pe pe a ed by one o mo e
indi iduals om anonymous p o iles and can be wi nessed
by a wide audience by being epos ed o sha ed wi h o he s
(Menin e al., 2021).
The ex en o pee - o-pee cybe bullying is a public heal h
conce n wi h se ious implica ions o a conside able p o-
po ion o adolescen s and young people a ound he globe.
A ecen me a-analysis by Li and colleagues (2024) ha
included 42 s udies om Eu ope, Aus alia, No h Ame ica,
Asia, and he Middle Eas , wi h 266,888 pa icipan s, ound
In oduc ion
The in e na ional e ision o he e m ‘cybe bullying’, pub-
lished by UNESCO Chai in 2024, highligh ed ha his
ype o ela ional iolence h ough he elec onic media is
cha ac e ised by: (i) being a damaging social p ocess, i.e.
i occu s in a ne wo k o in e pe sonal ela ionships; (ii)
Rosa io Del Rey
[email p o ec ed]
1 Depa men o Psychology, Uni e sidad Loyola Andalucía,
Dos He manas (Se ille, Andalusia), Spain
2 Facul y o Psychology, Uni e sidade de Lisboa, Lisboa,
Po ugal
3 Depa men o De elopmen al and Educa ional Psychology,
Uni e sidad de Se illa, Se ille (Andalusia), Spain
4 Facul y o Educa ional Sciences, Uni e si y o Se ille
[Uni e sidad de Se illa], Pi o ecnia S ee , Se ille
41013, Spain
Abs ac
Episodes o cybe bullying ic imisa ion ha e se ious consequences among adolescen s, which wo sen when hei in ol e-
men is pe pe ua ed o e ime. I is he e o e impo an o unde s and wha ac o s lead o long- e m cybe ic imisa ion
o p e en i . This one-yea longi udinal s udy examines signi ican socioemo ional ac o s in he o igin and dynamics o
cybe ic imisa ion, no ye join ly explo ed in i s pe pe ua ion. Pa icipan s we e 427 Spanish 7 h -, 8 h -, 9 h - and 10 h
-g ade s uden s (52.9% boys, 46.8% gi ls, 0.2% o he ), aged 12–17 (Mage = 13.08, SD = 1.01). Adolescen s comple ed a
se ies o sel - epo ed ques ionnai es assessing pee p essu e, ange dys egula ion, and cybe ic imisa ion. The esul s
e ealed ha : (a) o he o al sample, 5.6% we e cybe ic ims only a T1 and 8.0% only a T2, and 3.6% we e long- e m
cybe ic ims; (b) all a iables we e signi ican ly and posi i ely co ela ed a T1 and T2, excep cybe ic imisa ion a T1
and ange dys egula ion a T2; (c) sco es a ied acco ding o gende and age; and (d) pee p essu e can inc ease he isk
o long- e m cybe ic imisa ion, speci ically in cases whe e ange dys egula ion le els a e mode a e o low. The esul s
highligh he impo ance o conside ing mode a ing mechanisms in ol ed in inc easing he isk o long- e m cybe ic i-
misa ion as well as he need o p omo e posi i e pee g oup dynamics and emo ion managemen o a oid pe pe ua ing
he p oblem.
Keywo ds Cybe bullying · Vic imisa ion · Pe sis ence · Pee s · Ange
Accep ed: 29 May 2025
© The Au ho (s) 2025
Adolescen s Su e ing om Long- e m Cybe bullying Vic imisa ion:
Pee P essu e and Ange Dys egula ion as Risk Fac o s
Espe anzaEspino1· Ana Ma ga idaVeiga-Simão2· Paula Cos aFe ei a2· Vi giniaSánchez-Jiménez3·
Rosa io DelRey3,4
1 3
Resea ch on Child and Adolescen Psychopa hology
ha he mean p e alence o cybe ic ims was 11.1%. Simi-
la ly, he esul s o a s udy wi h 12,285 Spanish adolescen s
aged 11–18 yea s showed ha he p e alence o cybe ic ims
o e he pas wo mon hs was 8.1% (González-Cab e a e
al., 2020). Findings ega ding gende and age di e ences
in cybe ic imisa ion a e mixed. While some s udies epo
highe a es among gi ls and om middle adolescence, e.g.
ages 13–15 (Tsi sika e al., 2015; Wo ld Heal h O ganisa ion
[WHO], 2020), o he s ha e ound no signi ican di e ences
(Smi h e al., 2019).
A Longi udinal Pe spec i e on In ol emen in
Cybe bullying Vic imisa ion
Gi en he signi ican p e alence a es o cybe bullying,
conside able esea ch has ocused on unde s anding i s e o-
lu ion and impac , pa icula ly among cybe ic ims (Chu e
al., 2018; Gámez-Guadix e al., 2015).
In he s udy o cybe ic imisa ion ajec o ies, g ea e
a en ion has been gi en o ole changes – such as a cybe -
ic im becoming a cybe bully – han o he pe sis ence o
ch oni ica ion o cybe ic imisa ion o e ime (Ak e al.,
2015; Camacho e al., 2021). Mo eo e , when pe sis ence
has been examined, i has mo e o en been in he con ex o
adi ional school bullying a he han cybe bullying (Cal-
e e e al., 2018; Rome a e al., 2020). The limi ed exis -
ing esea ch on pe sis en cybe ic imisa ion sugges s ha
mos cybe ic ims do no emain in his ole consis en ly
bu ins ead end o disengage (i.e. spo adic cybe ic ims) o
shi oles o e ime (Fes l e al., 2017). Howe e , he e is a
small p opo ion o hem who expe ience cybe -agg essions
o e p olonged pe iods o ime, leading o mo e se e e
consequences and making in e en ion mo e challenging
(Johande e al., 2023). Fo ins ance, González-Cab e a
and colleagues (2023) ound in hei p ospec i e s udy o
3 wa es o e 13 mon hs ha among 1,142 s uden s aged
11–18, 6% we e s able cybe ic ims h oughou he en i e
pe iod analysed (mos ly gi ls, wi h 4.3%).
Despi e g owing in e es in cybe bullying, he isk ac-
o s unde lying long- e m cybe ic imisa ion emain la gely
unde explo ed. Few s udies ha e examined his issue; some
ha e iden i ied dep essi e symp oms (Huang e al., 2022)
and poo -quali y iendships (Tian e al., 2023) as po en ial
isk ac o s, while o he s poin o ai s like mind ulness
as p o ec i e (Royuela-Colome e al., 2018). Ad ancing
esea ch in his a ea is c i ical, as cybe ic imisa ion is
known o be associa ed wi h se ious nega i e ou comes –
including emo ional dis ess, dep ession, impai ed social
unc ioning and suicidal idea ion (Coelho & Ma chan e,
2018; an den Eijnden e al., 2014). I s ands o eason ha
he longe he cybe ic imisa ion pe sis s, he mo e se e e
and en enched hese consequences a e likely o become
(González-Cab e a e al., 2020, 2023). Acco dingly, his
s udy examines pee p essu e and ange dys egula ion – wo
ac o s ypically s udied in isola ion in ela ion o spo adic
cybe bullying in ol emen – bu no join ly in he con ex o
long- e m ic imisa ion.
The E ec o Pee P essu e in Cybe bullying
Vic imisa ion
‘Pee p essu e’ is unde s ood as he in luence ha o he s
exe o e one’s own hough s o beha iou o p io i ise
g oup no ms as a sign o social iden i y (San o e al., 2000).
Adolescen s and young people oday may be inc easingly
suscep ible o such in luence in bo h ace- o- ace and online
in e ac ions, whe e social iden i y is o en cons uc ed and
alida ed by pee s (Gámez-Guadix e al., 2013). While pee
p essu e has been linked o he pe pe a ion o cybe bullying
– pa icula ly when agg essi e beha iou is ein o ced o
encou aged by he pee g oup (Piccoli e al., 2020) – i s ole
in cybe ic imisa ion emains less explo ed. The ew exis -
ing s udies on pee p essu e and cybe ic imisa ion poin ed
ou ha his ela ionship is a om i ial (Li azo i’c &
Ham, 2019). Longi udinal esea ch has shown, o ins ance,
ha pee p essu e can media e he ela ionship be ween
cybe ic imisa ion and i s nega i e psychological conse-
quences (Gao e al., 2021). Howe e , u he heo e ical and
empi ical wo k is needed o unde s and whe he and how
pee p essu e may se e as a isk ac o o he pe sis ence
o cybe ic imisa ion o e ime.
F om a heo e ical pe spec i e, pe cei ing high pee
p essu e may con ibu e o long- e m cybe ic imisa ion by
ein o cing maladap i e pee no ms ha no malise o ol-
e a e agg ession, while simul aneously educing he cybe -
ic im’s abili y o willingness o esis o disengage om
ha m ul pee dynamics. In his sense, pee p essu e can indi-
ec ly sus ain con inued cybe ic imisa ion by comp omis-
ing asse i eness and heigh ening sensi i i y o social cues
and app o al. Suppo ing his, ecen e idence by Ondi i
and colleagues (2024) in a c oss-lagged longi udinal design
ound ha lowe esis ance o pee p essu e was associa ed
wi h inc eased u u e cybe ic imisa ion. This sugges s ha
adolescen s who eel unable o esis g oup expec a ions
may be less likely o de end hemsel es, epo abuse, o
disengage om online en i onmen s ha expose hem o
ha m – he eby inc easing hei ulne abili y o e ime.
In his ein, he e is ample e idence ha he desi e o
social connec ion and pee accep ance in digi al con ex s
may inc ease exposu e o hu ul judgmen s online (And ade
e al., 2021; Gámez-Guadix e al., 2013). Adolescen s who
ha e expe ienced cybe ic imisa ion o en exp ess a heigh -
ened need o belonging, companionship, and es o a ion
o damaged social s a us (Fang e al., 2020; Na a o e al.,
1 3
Resea ch on Child and Adolescen Psychopa hology
2015; Rome a e al., 2020). This need may d i e hem o
s ay connec ed in pee spaces whe e hey con inue o be a -
ge ed, especially i hey ea exclusion o social isola ion
(Gao e al., 2021).
The E ec o Ange Dys egula ion in Cybe bullying
Vic imisa ion
Resea ch on cybe bullying and ange is ex ensi e. In ela-
ion o cybe ic imisa ion, wo complemen a y pe spec-
i es ha e eme ged. Some s udies sugges ed ha ange is
a de ining ai among cybe ic imised adolescen s and
young people who s uggle o exp ess his emo ion in adap-
i e ways (Lonig o e al., 2014; P a e al., 2014). O he s
a gued ha ange o en a ises as a eac ion o cybe bullying
(den Hame & Konijn, 2016), wi h cybe ic ims display-
ing poo impulse con ol and eac i e beha io s (Camacho
e al., 2021; O ega e al., 2012;). Speci ically, hose who
a e cybe ic imised may eac wi h ange -ou , i.e. agg es-
si ely, o wi h ange -in, i.e. by supp essing hemsel es (Ak
e al., 2015). Fo ins ance, cybe ic ims a e mo e likely o
umina e, which is i sel an exp ession o ange (Liu e al.,
2020; Na a o e al., 2015). Gi en ha emo ional dys egu-
la ion – pa icula ly ange dys egula ion – has been linked
o bo h cybe bullying in ol emen and he de elopmen o
psychia ic symp oms (den Hame & Konijn, 2016; Zsila e
al., 2019), i is impo an o explo e whe he i also con ib-
u es o he pe sis ence o cybe ic imisa ion o e ime. I is
plausible ha cybe bullies a e ein o ced by he dys egu-
la ed emo ional eac ions o hei cybe ic ims, which could
inc ease he likelihood o con inued pe pe a ion.
The P esen S udy
While he combined ole o pee p essu e and ange has
been p e iously examined in ela ion o cybe pe pe a ion
(e.g. Yang e al., 2023), hei in luence on cybe ic imisa ion
emains unde explo ed – pa icula ly in long- e m cases.
Yang and colleagues (2023) ound ha ai ange p edic ed
pe pe a ion, especially among adolescen s expe iencing
high pee p essu e. This aises he ques ion o whe he simi-
la mechanisms may also con ibu e o con inued cybe ic i-
misa ion. Exis ing e idence sugges s ha , independen ly,
bo h con o mi y o pee p essu e and poo ange manage-
men may ac as maladap i e coping s a egies, inc easing
ulne abili y o cybe ic imisa ion and exace ba ing i s con-
sequences (Camacho e al., 2021; Gao e al., 2021). In his
ega d, we p opose ha cybe ic ims exposed o bo h high
pee p essu e and poo ange egula ion a e especially ul-
ne able o ongoing cybe ic imisa ion o e ime (T1–T2).
Pee p essu e may unde mine hei asse i eness, while
ange dys egula ion can lead o eac i e beha iou s ha
p o oke u he cybe -a acks. When hese ac o s co-occu ,
hey may signi ican ly impai cybe ic ims’ abili y o cope,
esis coe ci e dynamics, o seek suppo – ul ima ely ein-
o cing o wo sening hei cybe ic im s a us. Mo eo e , he
ole o pee p essu e on long- e m cybe ic imisa ion may
depend on cybe ic ims’ le els o ange egula ion, which
could ei he ampli y o bu e i s in luence (see Di ec and
Mode a ing E ec s below).
The e o e, he p esen s udy explo ed whe he pee p es-
su e and ange dys egula ion exe di ec and mode a ing
e ec s on cybe ic imisa ion a Time 1 (T1) and one yea
la e a Time 2 (T2), wi h he aim o cla i ying hei ole in
he pe sis ence o cybe ic imisa ion o e ime, while con-
olling o gende and age (see Fig. 1).
Fig. 1 Theo e ical model wi h
wo mode a o s
No e. X = independen a iable;
Y = dependen a iable; W = i s
mode a o ; Z = second mode a o
1 3
Resea ch on Child and Adolescen Psychopa hology
he main analysis. Hence, he pe cen age o eligible s uden s
ha ac ually pa icipa ed was 91% ( o mo e de ails on he
missing da a analysis, see below).
P ocedu e
The p esen s udy was app o ed by he Andalusian E hical
Coo dina ion Commi ee o Biomedical Resea ch, which
is associa ed wi h he Vi gen Maca ena and Vi gen del
Rocío hospi als (code: 1223-N-18). To conduc a longi u-
dinal s udy in which 12 mon hs elapsed be ween wo ime
poin s, a con enience sample was ec ui ed om ou sec-
onda y schools in sou he n Spain. As s ep 1, we explained
and asked o he collabo a ion o he managemen eams by
email and elephone. All seconda y schools ha con i med
hei in e es and willingness o pa icipa e in he longi u-
dinal p ocess we e included. A s ep 2, he signed w i en
consen o he amilies o legal gua dians was eques ed, as
well as he assen o he au ho ised s uden s. They pa ici-
pa ed olun a ily and anonymously in a pape -and-pencil
sel - epo su ey du ing 15–20 min o class ime and we e
supe ised by eache s and ou esea ch eam membe s a
T1 and T2. Da a we e collec ed o e he cou se o 2021.
Measu es
Fi s , pa icipan s we e asked abou hei gende , age
and g ade and, addi ionally, h ee alida ed scales we e
adminis e ed:
Cybe bullying ic imisa ion. In ol emen in cybe ic i-
misa ion a T1 and T2 was measu ed using he ic imisa-
ion subscale o he Eu opean Cybe bullying In e en ion
Di ec E ec s
Rega ding di ec e ec s, we hypo hesised: (H1) cybe ic i-
misa ion a T1, (H2) pee p essu e and (H3) ange dys egu-
la ion a e posi i ely associa ed wi h cybe ic imisa ion a
T2 (see Fig. 2).
Mode a ing E ec s
Rega ding mode a ing e ec s, we hypo hesised: (H4) pee
p essu e and (H5) ange dys egula ion can mode a e he
associa ion be ween cybe ic imisa ion a T1 and T2; (H6)
ange dys egula ion can mode a e he associa ion be ween
pee p essu e and cybe ic imisa ion a T2; and (H7) ange
dys egula ion can mode a e he condi ional in luence o
pee p essu e on he pe pe ua ion o cybe ic imisa ion (see
Fig. 2).
Me hod
Pa icipan s
Pa icipan s we e 427 Spanish seconda y school s uden s
(52.9% boys, 46.8% gi ls, 0.2% o he ). Ages anged om
12 o 17 yea s (M = 13.08, SD = 1.01 a T1, and M = 14.00,
SD = 0.97 a T2). Rega ding hei school yea a T1, 47.4%
we e in 7 h g ade, 42.3% in 8 h g ade, 9.9% in 9 h g ade and
0.5% in 10 h g ade. A T2, 3.3% we e in 7 h g ade, 48.5%
in 8 h g ade, 40.3% in 9 h g ade and 8.0% in 10 h g ade. To
add ess he aim o his longi udinal s udy, only s uden s who
pa icipa ed a bo h ime poin s, T1 and T2, we e selec ed in
Fig. 2 S a is ical model wi h wo
mode a o s
No e. X = independen a iable
(T1-CV = cybe ic imisa ion a
T1); Y = dependen a iable (T2-
CV = cybe ic imisa ion a T2);
W = i s mode a o (PP = pee
p essu e); Z = second mode a o
(T2-ADR = ange dys egula ion);
H = hypo hesis
1 3
Resea ch on Child and Adolescen Psychopa hology
(χ² = 2896.928; d = 2539; p <.001). Howe e , he no med
chi-squa e was low (χ²/d = 1.14) and adjus ed o sample
size sensi i i y, which, acco ding o Bollen’s (1989) c i e ia,
suppo s he assump ion ha he da a we e missing a an-
dom (MAR). The o e all esponse a e was 91% (N = 389).
Subsequen ly, eliabili y and alidi y analyses we e ca -
ied ou using EQS .6.4. Reliabili y was es ima ed using
C onbach’s alpha, and cons uc alidi y was assessed ia
he Leas Squa es Robus me hod. The ollowing goodness-
o - i indices we e examined: Sa o a-Ben le scaled chi-
squa e (χ²S-B; Sa o a & Ben le , 2001); Compa a i e Fi
Index (CFI) and Non-No med Fi Index (NNFI), wi h al-
ues ≥ 0.90 conside ed adequa e and ≥ 0.95 conside ed op i-
mal; Roo Mean Squa e E o o App oxima ion (RMSEA)
and S anda dized Roo Mean Squa e Residual (SRMR),
whe e alues ≤ 0.08 indica e adequa e i and ≤ 0.05 indica e
op imal i (Hu & Ben le , 1999).
Desc ip i e s a is ics we e hen pe o med using SPSS
.29. F equencies we e calcula ed o he ou iden i-
ied cybe ic imisa ion ajec o ies: non-cybe ic ims,
T1-cybe ic ims, T2-cybe ic ims, and long- e m cybe -
ic ims (i.e., hose epo ing cybe ic imisa ion a bo h
T1 and T2). Classi ica ion was based on he cu -o c i e-
ia es ablished by he ECIP-Q (O ega-Ruiz e al., 2016),
which de ines in ol emen in cybe ic imisa ion as expe i-
encing any o m o cybe -agg ession ‘a leas once o wice
a mon h’ in he las wo mon hs.
To explo e he ole o pee p essu e and ange dys eg-
ula ion in hese ajec o ies, one-way ANOVAs we e pe -
o med, ollowed by Bon e oni-co ec ed pos hoc es s o
con ol o Type I e o (see Table 1).
Bi a ia e co ela ions we e hen compu ed be ween all
s udy a iables, including age. Addi ionally, independen
samples - es s we e un o compa e means be ween boys
and gi ls (see Tables 2 and 3). The gende ca ego y “o he ”
(n = 1; 0.2%) was excluded om s a is ical analyses due o
insu icien ep esen a ion.
To ensu e he adequacy o ou sample o he main analy-
ses, we i s conduc ed a powe analysis using G*Powe .
Gi en ha ou mode a ed mode a ion model in ol es a
h ee-way in e ac ion (second-o de ) wi hin a mul iple
linea eg ession amewo k, we ollowed Cohen’s (1988)
guidelines. Pa ame e s we e se a a s a is ical powe o 0.95
and a signi icance le el o α = 0.05. Wi h h ee p edic o s
(one independen a iable and wo mode a o s), he analysis
indica ed ha a minimum o 107 pa icipan s was equi ed
o de ec a medium e ec size. Ou sample exceeded his
h eshold, and hus was deemed su icien o he planned
analyses.
Finally, a mode a ed mode a ion model was es ed using
PROCESS mac o .4.0 o SPSS (Model 3; Hayes, 2013),
employing a longi udinal design (see Fig. 1). The model
P ojec Ques ionnai e, ECIP-Q, alida ed o he Spanish
popula ion wi h op imal alues (O ega-Ruiz e al., 2016).
This subscale includes 11 i ems e e ing o agg essions
expe ienced o e he in e ne and on social media si es in
he las wo mon hs (e.g. ‘someone has insul ed me’, ‘ h ea -
ened me’, o ‘igno ed me’). Responses a e gi en on a 5-poin
Like - ype scale, om 0 ‘Ne e ’ o 4 ‘Yes, mo e han once
a week’. Good eliabili y alues we e ob ained (C onbach’s
αT1 = 0.95; C onbach’s αT2 = 0.94), along wi h op imal
esul s in he con i ma o y ac o analysis (CFA) a T1 (χ²
S − B = 113.93; DF = 44; p <.001; NNFI = 0.98, CFI = 0.99;
RMSEA = 0.06, SRMR = 0.07) and T2 (χ² S − B = 157.04;
DF = 44; p <.001; NNFI = 0.96, CFI = 0.97; RMSEA = 0.08,
SRMR = 0.09).
Pee p essu e. The in luence o he g oup no ms on
pa icipan s’ beha iou was measu ed using he alida ed
Spanish e sion o he Ex eme Pee O ien a ion subscale
o he Pee O ien a ion Scale by Fuligni and Eccles (1993)
(see Sánchez-Jiménez e al., 2023). This one-dimensional
scale includes 4 i ems e e ing o he pe cep ion o ypical
ansg essi e beha iou wi hin pee g oup dynamics (e.g.
‘How many imes ha e you done hings which a e ou side
you own capabili ies in o de o please you iends?’ o
‘How o en do you disobey you amily so ha you iends
won’ ejec you?’), wi h esponses being gi en on a 5-poin
Like - ype scale, om 0 ‘Ne e ’ o 4 ‘Always’. An accep -
able eliabili y alue was ob ained (C onbach’s αT2 = 0.79),
along wi h good esul s in he CFA a T2 (χ² S − B = 12.19;
DF = 2; p <.001; NNFI = 0.94, CFI = 0.98; RMSEA = 0.11,
SRMR = 0.04).
Ange dys egula ion. Di icul ies in con olling and
exp essing ange in s ess ul si ua ions we e measu ed
using a subscale o he Emo ional Quo ien In en o y: You h
e sion o Ba -On and Pa ke (2000), al eady alida ed in
Spanish popula ion (see Méndez e al., 2019). This subscale
includes 8 i ems e e ing o ypical exp essions o ange
(e.g. ‘I igh wi h people’ o ‘When I ge ang y, I ac wi h-
ou hinking’). Responses a e gi en on a 5-poin Like - ype
scale, om 0 ‘Ne e ’ o 4 ‘Always’, so high sco es equa e
o poo e ange managemen . Good eliabili y was ob ained
(C onbach’s αT2 = 0.87), along wi h accep able esul s in he
CFA a T2 (χ² S − B = 79.79; DF = 20; p <.001; NNFI = 0.96,
CFI = 0.97; RMSEA = 0.09, SRMR = 0.06).
Da a Analysis
Da a analysis was conduc ed using SPSS .29 and EQS .6.4.
P io o he analyses, da a we e coded and cleaned in SPSS
.29. Only pa icipan s who comple ed bo h T1 and T2 we e
included in he analyses. To assess he na u e o he missing
da a, Li le’s MCAR (Missing Comple ely a Random) es
(Li le, 1988) was pe o med, yielding a signi ican esul
1 3
Resea ch on Child and Adolescen Psychopa hology
Resul s
P elimina y Analysis
O he o al sample, 5.6% we e only T1-cybe ic ims, 8.0%
we e only T2-cybe ic ims, and 3.6% we e long- e m cybe -
ic ims (i.e., T1T2-cybe ic ims). The analysis o a i-
ance showed s a is ically signi ican di e ences be ween
non-cybe ic ims s. T1-cybe ic ims in baseline ange
dys egula ion (M = 2.41 s. M = 2.93; pbon = 0.045), and
be ween non-cybe ic ims s. T2-cybe ic ims (M = 0.65
s. M = 1.03; pbon = 0.002) and long- e m cybe ic ims
(M = 0.65 s. M = 1.33; pbon = 0.016) in la e pee p essu e
examined whe he pee p essu e and ange dys egula-
ion a T2 mode a ed he longi udinal associa ion be ween
cybe ic imisa ion a T1 (independen a iable) and T2
(dependen a iable). Age and gende (0 = boys, 1 = gi ls)
we e included as co a ia es (see Fig. 2). Analyses used
mean-cen ed a iables, 10,000 boo s apped samples (95%
con idence in e al – CI), and HC3 he e oscedas ici y-con-
sis en s anda d e o s (Da idson & MacKinnon, 1993). The
Johnson-Neyman echnique was applied o p obe in e ac-
ion e ec s, wi h signi icance de e mined a p <.05 when he
95% con idence in e al excluded ze o in he lowe o uppe
limi .
Table 1 Means, s anda d de ia ions and one-way analyses o a iance o ange dys egula ion and pee p essu e among non-cybe ic ims,
T1-cybe ic ims, T2-cybe ic ims and long- e m cybe ic ims (T1T2-cybe ic ims)
Measu e Non-cybe ic ims (n = 322) T1 cybe ic ims (n = 22) T2 cybe ic ims (n = 31) Long- e m cybe ic ims (n = 14) F(3,389) η2
M(SD) M(SD) M(SD) M(SD)
T1-ADR 2.41(0.83) 2.93(1.19) 2.54(0.65) 2.93(1.19) 4.11*** 0.03
T1-PP 0.65(0.73) 0.91(0.95) 0.89(0.86) 1.12(0.78) 3.01* 0.02
T2-ADR 2.46(0.83) 2.59(0.83) 2.72(0.80) 3.06(0.92) 2.83* 0.02
T2-PP 0.65(0.66) 0.89(0.73) 1.03(0.69) 1.33(0.81) 7.45*** 0.05
No e. *** p < 001; * p <.05
Abb e ia ions: ADR, ange dys egula ion; PP, pee p essu e
Table 2 Desc ip i e analyses and co ela ions o he s udy a iables
Va iable 1 2 3 4 5 6 7
1. T1-CV -
2. T2-CV 0.39*** -
3. T1-ADR 0.23*** 0.21** -
4. T2-ADR 0.08 0.22*** 0.51*** -
5. T1-PP 0.27*** 0.23*** 0.37*** 0.21*** -
6. T2-PP 0.23*** 0.31*** 0.28*** 0.37*** 0.48*** -
7. Age 0.10* 0.04 −0.03 −0.12* 0.09 0.05 -
No e. * p <.05; ** p <.01; *** p <.001; Cell en ies a e ze o-o de Spea man co ela ion coe icien s. Abb e ia ions: CV, cybe ic imisa ion;
ADR, ange dys egula ion; PP, pee p essu e
Table 3 Desc ip i e analyses and S uden ’s - es s by gende
M SD M SD pd
1. T1-CV 0.14 0.28 Boys 0.15 0.29 –1.10 0.27 0.11
Gi ls 0.12 0.27
2. T2-CV 0.15 0.27 Boys 0.14 0.24 –1.14 0.26 –0.11
Gi ls 0.16 0.29
3. T1-ADR 2.47 0.83 Boys 2.37 0.80 –2.33 0.02* –0.24
Gi ls 2.58 0.89
4. T2-ADR 2.52 0.84 Boys 2.45 0.82 –1.57 0.12 –0.16
Gi ls 2.59 0.86
5. T1-PP 0.71 0.76 Boys 0.75 0.77 1.14 0.26 0.12
Gi ls 0.66 0.73
6. T2-PP 0.72 0.69 Boys 0.77 0.66 1.33 0.18 0.13
Gi ls 0.67 0.72
No e. * S a is ically signi ican di e ences p <.05
Abb e ia ions: CV, cybe ic imisa ion; ADR, ange dys egula ion; PP, pee p essu e
1 3
Resea ch on Child and Adolescen Psychopa hology
p essu e we e s ongly associa ed wi h cybe ic imisa ion
a T2. Only esea ch Hypo heses 1 and 2 we e suppo ed
(see Fig. 2).
Mode a ing E ec s. The nex se o esea ch hypo heses
add essed he mode a ing ole o pee p essu e and ange
dys egula ion in long- e m cybe bullying ic imisa ion, i.e.
cybe ic imisa ion a T1 and T2 (H4-H6; see Fig. 2). In he
OLS model o hese indi ec e ec s, he wo-way analyses
e ealed ha pee p essu e was no a signi ican mode a o
o he associa ion be ween cybe ic imisa ion a T1 and T2
[β = −0.01, (9,378) = − 0.11, p =.915; see Table 4]. This esul
does no suppo esea ch Hypo hesis 4 (H4). Howe e ,
he wo-way analyses e ealed a signi ican posi i e mod-
e a ion by ange dys egula ion in he associa ion be ween
cybe ic imisa ion a T1 and T2 [β = 0.17, (9,378) = 2.20,
p <.05; see Table 4], he eby con i ming esea ch Hypo h-
esis 5 (H5). In o he wo ds, high sco es o ange dys egula-
ion mode a ed he associa ion be ween cybe ic imisa ion
a T1 and T2. Mo eo e , wo-way analyses e ealed a sig-
ni ican nega i e mode a ion by ange dys egula ion in he
associa ion be ween pee p essu e and cybe ic imisa ion a
T2 [β = −0.07, (9,378) = − 2.83, p <.01; see Table 4], he eby
con i ming esea ch Hypo hesis 6 (H6). In o he wo ds, low
sco es o ange dys egula ion mode a ed he associa ion
be ween pee p essu e and cybe ic imisa ion a T2.
By using he o e all h ee-way in e ac ion model, we
we e able o es ima e he mode a ing e ec o ange dys-
egula ion (second mode a o ) on he condi ional in lu-
ence o pee p essu e ( i s mode a o ) on he associa ion
be ween cybe ic imisa ion a T1 and T2. In he OLS model
o his h ee-way in e ac ion, he esul s indica ed a nega-
i e mode a ion o pee p essu e and ange dys egula ion
in he associa ion be ween cybe ic imisa ion a T1 and T2
[β = −0.22, (9,378) = − 2.09, p <.05; see Table 4]. This sup-
po s esea ch Hypo hesis 7 (H7; see Fig. 3). Fu he mo e,
he simple slope analysis e ealed a signi ican e ec o
he cybe ic imisa ion a T1 x pee p essu e in e ac ion on
cybe ic imisa ion a T2, which was ound o high le els o
ange dys egula ion (β = −0.19, p <.05; see Fig. 3). In o he
wo ds, pee p essu e ac ed as a mode a o be ween cybe -
ic imisa ion a T1 and T2 when ange egula ion was no
su icien ly de eloped. As shown in Fig. 3, he lowes le el
o cybe ic imisa ion a T2 was ob ained o he lowes
sco es o cybe ic imisa ion a T1, pee p essu e and ange
dys egula ion, sugges ing ha low in ol emen in p e ious
cybe ic imisa ion and good socioemo ional abili ies p o ec
agains long- e m cybe ic imisa ion. In con as , he high-
es le el o cybe ic imisa ion a T2 was ob ained o he
highes sco es o cybe ic imisa ion a T1, pee p essu e
and ange dys egula ion, sugges ing ha high in ol emen
in p e ious cybe ic imisa ion and poo socioemo ional
abili ies inc eases he isk o long- e m cybe ic imisa ion.
(see Table 1). The e ec size was small in all cases, wi h e a
squa ed > 0.01 and < 0.06 (Cohen, 1988).
Desc ip i e Resul s O e all and by Gende and Age
The desc ip i e esul s can be ound in Tables 2 and 3. In
he o e all sample, all a iables unde s udy we e signi i-
can ly and posi i ely co ela ed in each wa e, excep o
cybe ic imisa ion a T1 and ange dys egula ion a T2
( =.08, p =.120). By gende , he - es e ealed signi ican
di e ences only in ange dys egula ion a T1 (Mgi ls = 2.58
s. Mboys = 2.37; p =.02). In o he wo ds, gi ls sco ed signi i-
can ly wo se han boys on ange egula ion a baseline. By
age, he bi a ia e co ela ion was only signi ican o cybe -
ic imisa ion a T1 ( =.13, p =.01) and ange dys egula ion
a T2 ( = −.12, p =.02). In o he wo ds, olde adolescen s
showed signi ican ly highe le els o cybe ic imisa ion a
T1, and younge adolescen s showed signi ican ly highe
le els o ange dys egula ion a T2.
Model Resul s: Main and Mode a ing E ec s
The model shown in Fig. 2 was s a is ically signi ican ,
F(9,378) = 10,44, p <.001, R2 = 0.21. This accoun s o 21%
o he a iance, indica ing a medium-sized e ec . Nei he
gende (β = 0.05, p =.074) no age (β = 0.01, p =.625) we e
ound o play a signi ican ole on cybe ic imisa ion a T2.
Di ec E ec s. As shown in Table 4, he o dina y leas
squa es (OLS) eg ession analyses e ealed ha cybe ic i-
misa ion a T1 [β = 0.43, (9,378) = 3.49, p <.001] and pee
p essu e [β = 0.01, (9,378) = 2.91, p <.01] had a signi ican
di ec e ec on cybe ic imisa ion a T2. In con as , ange
dys egula ion [β = 0.02, (9,378) = 1.04, p =.298] had no sig-
ni ican di ec e ec on cybe ic imisa ion a T2. In o he
wo ds, high sco es o cybe ic imisa ion a T1 and o pee
Table 4 Di ec and in e ac ion e ec s in he mode a e mode a ion
model
E ec B SE 95% CI
LL UL
T1-CV 0.43*** 0.12 0.19 0.67
PP 0.01** 0.03 0.03 0.17
ADR 0.02 0.02 −0.02 0.05
T1-CV x PP −0.01 0.05 −0.11 0.10
T1-CV x ADR 0.17* 0.08 0.02 0.33
PP x ADR −0.07** 0.03 −0.12 −0.02
T1-CV x PP x ADR −0.22* 0.11 −0.43 −0.01
Gende 0.05 0.03 −0.01 0.10
Age 0.01 0.01 −0.02 0.03
No e. Analyses pe o med using he PROCESS mac o o SPSS
(Model 3; Hayes, 2013). *p <.05, **p <.01, ***p <.001
Abb e ia ions: CV, cybe ic imisa ion; PP, pee p essu e; ADR,
ange dys egula ion; CI, con idence in e al; LL, lowe limi ; UL,
uppe limi
1 3
Resea ch on Child and Adolescen Psychopa hology
o e a one-yea pe iod. This igu e ep esen s app oxima ely
one- hi d o hose ini ially iden i ied as cybe ic ims a T1
(9.2%) and aligns wi h indings om o he Spanish longi-
udinal s udies using simila cu -o poin s, such as ha o
Gámez-Guadix and colleagues (2015), wi h 5.8% o s able
cybe ic ims, o ha o González-Cab e a and colleagues
(2023), wi h 6%, o hei o al samples. This consis ency
ein o ces he idea ha cybe ic imisa ion is a dynamic
phenomenon, wi h a signi ican po ion o cybe ic ims
emaining in ol ed o e ime. Recognising his con inu-
i y is c ucial: p e alence a es a di e en ime poin s may
e lec he same indi iduals, no new cases.
In addi ion, he o he key con ibu ion o ou s udy lies
in cla i ying he po en ial ole o pee p essu e and ange
dys egula ion as p edic o s o long- e m cybe ic imisa ion,
bo h independen ly and in combina ion:
In e ms o di ec e ec s, we con i med ou hypo heses 1
and 2, showing ha high le els o p io cybe ic imisa ion
and o pee p essu e a e s ong p edic o s o u u e cybe -
ic imisa ion (Came ini e al., 2020; Li azo i’c & Ham,
2019). In his sense, he e is su icien e idence ha being
a a ge o cybe bullying a one poin in ime inc eases he
Discussion
Al hough esea ch on cybe bullying ic imisa ion indi-
ca es ha mos cases end o esol e o e ime, a signi ican
p opo ion pe sis (González-Cab e a e al., 2023) despi e
he implemen a ion o e ec i e in e en ion p og ammes
(Johande e al., 2023), such as ‘ConRed’ (Del Rey e al.,
2016) o ‘Cybe p og am 2.0’ (Ga aigo dobil & Ma ínez-
Valde ey, 2018). Iden i ying he p edic o s o long- e m
cybe ic imisa ion is he e o e essen ial o p e en ing i s
mos se e e and las ing consequences (González-Cab e a e
al., 2020; Huang e al., 2022; Tian e al., 2023). This s udy
sough o examine he combined in luence o wo li le-
explo ed socioemo ional ac o s – pee p essu e and ange
dys egula ion – in his con ex . Gaining insigh in o hei ole
may help explain why ce ain cases pe sis o e ex ended
pe iods (e.g., one yea ), and could in o m he e inemen o
exis ing psychoeduca ional s a egies o mo e e ec i ely
suppo hose who a e su e ing mo e and o longe .
One o he main con ibu ions o ou s udy is he con i -
ma ion ha a no able p opo ion o adolescen s expe ience
long- e m cybe ic imisa ion – 3.6% o he o al sample
Fig. 3 Th ee-way in e ac ion plo o cybe ic imisa ion a T1, pee
p essu e, and ange dys egula ion on cybe ic imisa ion a T2
No e. The mode a ing e ec o ange dys egula ion on he condi ional
in luence o pee p essu e on he associa ion be ween cybe ic imisa-
ion a T1 and T2
1 3
Resea ch on Child and Adolescen Psychopa hology
alone does no p edic long- e m cybe ic imisa ion unless
conside ed alongside ange dys egula ion le els. Fo cybe -
ic ims wi h high ange dys egula ion, pee p essu e did no
signi ican ly change hei isk o u he in ol emen in he
p oblem – possibly because hei emo ional ulne abili y
al eady places hem a high isk. In con as , o hose wi h
mode a e o low ange dys egula ion (i.e. be e emo ional
sel -con ol), pee p essu e did inc ease he likelihood o
long- e m cybe ic imisa ion. This inding sugges s ha
e en emo ionally compe en adolescen s may emain ul-
ne able when social in luences a e s ong.
These indings make sense, as we know ha ange is an
emo ion ha may lead o g ea e loneliness (Fang e al.,
2020), dep ession (Huang e al., 2022) and s ess (Liu e
al., 2020), which in u n heigh en he cybe ic im’s need o
belong (Gao e al., 2021). Cybe ic ims, who o en s ug-
gle wi h social ela ionships (Rome a e al., 2020), may
become inc easingly a aid o ejec ion as he cybe ic imi-
sa ion pe sis s ( an den Eijnden e al., 2014). In an a emp
o egain accep ance, hey may gi e in o pee p essu e –
adop ing ce ain beha iou s o complying wi h pee no ms
o i in (Gao e al., 2021). Howe e , e y suscep ibili y can
make hem less appealing o hei pee s, as being o e ly
imp essionable is o en pe cei ed nega i ely and associa ed
wi h lowe social s a us o popula i y (Gámez-Guadix e al.,
2013; Rome a e al., 2020; an den Eijnden e al., 2014).
This cycle ul ima ely isola es cybe ic ims u he , dep i -
ing hem o pee suppo ha could help p o ec hem om
ongoing cybe ic imisa ion (Na a o e al., 2015).
Hence, ou s udy unde sco es wo key socioemo ional
condi ions – pee p essu e and ange dys egula ion – ha
may allow cybe bullies o p olong hei a acks o e ime.
These indings ha e di ec implica ions o op imising
gene ic p e en ion and in e en ion measu es aimed a com-
ba ing cybe bullying. In pa icula , i seems c ucial o inco -
po a e: (i) emo ional egula ion (e.g., managing impulsi e
eac ions o online p o oca ion), which may be especially
e ec i e o adolescen s wi h high ange dys egula ion (den
Hame & Konijn, 2016); and (ii) aining in pee esis ance
skills, asse i eness, and c i ical hinking abou pee no ms
and beha io s, which may be especially bene icial o hose
wi h lowe emo ional ulne abili y (Ondi i e al., 2024).
Indeed, his is pa icula ly impo an , as p e ious esea ch
sugges s ha enhancing adolescen s’ emo ional egula ion
s a egies can lead o imp o ed social compe ence, and
con e sely, s onge social compe ence can u he enhance
emo ional egula ion (A a ó e al., 2022). The e o e, based
on ou indings, we a gue ha cybe bullying in e en ions
should be in eg a i e, add essing bo h in e nal (emo ional
egula ion) and ex e nal (pee p essu e) isk ac o s simul-
aneously, as some adolescen s emain a isk o long- e m
cybe ic imisa ion.
isk o being a ge ed again la e on (Coelho & Ma chan e,
2018; González-Cab e a e al., 2023). This may be due, in
pa , o he ac ha he imbalance o powe be ween cybe -
ic ims and hei cybe bullies and pee s becomes inc eas-
ingly en enched in social dynamics o e ime, making
i ha de o b eak he cycle o cybe iolence (Chu e al.,
2018). Al hough pee p essu e has ecei ed less a en ion
in he con ex o cybe ic imisa ion in ol emen , ou ind-
ings highligh i s ele ance, as Li azo i’c and Ham (2019)
o Ondi i and colleagues (2024) show c oss-sec ionally and
longi udinally, espec i ely. Speci ically, he inc eased isk
o cybe ic imisa ion among hose who eel s ongly p es-
su ed by hei pee s may be due o hei lowe abili y o
esis pee in luence. Indi iduals who a e mo e imp ession-
able o socially complian may be pe cei ed as ulne able
by pee s, which can lead o social sanc ions (Gámez-Guadix
e al., 2013), such as a decline in popula i y o social p e -
e ence (Rome a e al., 2020). In his espec , suscep ibili y
o pee p essu e may ac as a ma ke o social agili y ha
makes ce ain adolescen s mo e likely a ge s. Rega ding
ange dys egula ion (hypo hesis 3), ou indings did no
suppo i s ole as a di ec p edic o o u u e cybe ic i-
misa ion. This sugges s ha emo ional eac i i y alone may
no inc ease cybe ic imisa ion isk unless i in e ac s wi h
o he ac o s – such as low social compe ence o weak pee
suppo ne wo ks – ha ampli y ulne abili y.
In e ms o indi ec o mode a ing e ec s, ou hypo h-
eses 5, 6, and 7 we e suppo ed, while, unexpec edly, ou
hypo hesis 4 was no . Speci ically, when analysed indi idu-
ally, only ange dys egula ion mode a ed he ela ionship
be ween p io and u u e cybe ic imisa ion–indica ing
ha di icul ies in managing ange inc ease he isk o long-
e m cybe ic imisa ion. The p ominen ole o emo ional
ac o s in long- e m cybe ic imisa ion aligns wi h p e i-
ous esea ch showing ha poo emo ional sel -con ol o en
p edic s non-con ac ic imisa ion (P a e al., 2014). This
is he case wi h cybe ic imisa ion and he use o ine ec-
i e coping s a egies, which ends o in ol e adolescen s
mo e deeply in he p oblem (A a ó e al., 2022; den Hame
& Konijn, 2016). This may e lec he ac ha cybe bullies
seek emo ional eac ions in hei cybe ic ims, p o oking
hem h ough epea ed a acks o e ime. Ou indings hus
highligh ange managemen as a key p o ec i e ac o – no
only o educe he isk o long- e m cybe ic imisa ion bu
also o p e en maladap i e esponses such as umina ion
(Zsila e al., 2019) o agg ession (Camacho e al., 2021).
Howe e , a mo e nuanced pic u e o long- e m cybe -
ic imisa ion eme ged when examining he in e ac ion
be ween pee p essu e and ange dys egula ion. In line wi h
ou hypo hesis 7, he in luence o pee p essu e appea ed o
depend on he le el o ange dys egula ion. Speci ically, ou
h ee-way in e ac ion analysis e ealed ha pee p essu e
1 3
