Alizade, Jeyhun; Ellge , Fabio; G ünewald, Ma ius; Tichelbaecke , Thomas
A icle — Published Ve sion
Does poli ical iolence unde mine desc ip i e
ep esen a ion? The case o women in poli ics
Eu opean Jou nal o Poli ical Resea ch
P o ided in Coope a ion wi h:
WZB Be lin Social Science Cen e
Sugges ed Ci a ion: Alizade, Jeyhun; Ellge , Fabio; G ünewald, Ma ius; Tichelbaecke , Thomas (2025) :
Does poli ical iolence unde mine desc ip i e ep esen a ion? The case o women in poli ics,
Eu opean Jou nal o Poli ical Resea ch, ISSN 1475-6765, Wiley, Hoboken, NJ, Vol. 64, Iss. 4, pp.
2106-2121,
h ps://doi.o g/10.1111/1475-6765.70017
This Ve sion is a ailable a :
h ps://hdl.handle.ne /10419/318276.2
S anda d-Nu zungsbedingungen:
Die Dokumen e au EconS o dü en zu eigenen wissenscha lichen
Zwecken und zum P i a geb auch gespeiche und kopie we den.
Sie dü en die Dokumen e nich ü ö en liche ode komme zielle
Zwecke e iel äl igen, ö en lich auss ellen, ö en lich zugänglich
machen, e eiben ode ande wei ig nu zen.
So e n die Ve asse die Dokumen e un e Open-Con en -Lizenzen
(insbesonde e CC-Lizenzen) zu Ve ügung ges ell haben soll en,
gel en abweichend on diesen Nu zungsbedingungen die in de do
genann en Lizenz gewäh en Nu zungs ech e.
Te ms o use:
Documen s in EconS o may be sa ed and copied o you pe sonal
and schola ly pu poses.
You a e no o copy documen s o public o comme cial pu poses, o
exhibi he documen s publicly, o make hem publicly a ailable on he
in e ne , o o dis ibu e o o he wise use he documen s in public.
I he documen s ha e been made a ailable unde an Open Con en
Licence (especially C ea i e Commons Licences), you may exe cise
u he usage igh s as speci ied in he indica ed licence.
h p://c ea i ecommons.o g/licenses/by/4.0/
2106 Eu opean Jou nal o Poli ical Resea ch 64: 2106–2121, 2025
doi: 10.1111/1475-6765.70017
Does poli ical iolence unde mine desc ip i e ep esen a ion? The case o
women in poli ics
JEYHUN ALIZADE,1FABIO ELLGER,2MARIUS GRÜNEWALD3,4
& THOMAS TICHELBAECKER5,6
1Mig a ion, In eg a ion, T ansna ionaliza ion Uni , WZB Be lin, Ge many; 2T ans o ma ions o Democ acy Uni ,
WZB Be lin, Ge many; 3Eu opean Uni e si y Ins i u e, I aly; 4Eu opean Cen al Bank, Ge many; 5Chemni z
Uni e si y o Technology, Ge many; 6Depa men o Poli ics, P ince on Uni e si y, USA
Abs ac . Do a acks agains poli icians exace ba e he poli ical unde ep esen a ion o ma ginalized g oups?
Exis ing esea ch sugges s ha candida es and o iceholde s om unde ep esen ed g oups a e mo e likely o
become a ge s o poli ical iolence, bu li le is known abou he consequences o indi ec exposu e o poli ical
iolence o desc ip i e ep esen a ion and poli ical ambi ion. Focusing on he case o women in poli ics, we
s udy how he p e alence o poli ical iolence a ec s bo h he desc ip i e ep esen a ion and he poli ical ambi ion
o women in Ge many – Eu ope’s la ges democ acy. Combining an analysis o obse a ional da a measu ing
c imes agains poli icians wi h e idence om o iginal and p e- egis e ed su ey expe imen s, we i s demons a e
ha a acks on poli ical eli es a e no associa ed wi h ewe emale candida es on pa y lis s o local elec ions.
Examining poli ical ambi ions and unde lying mic o ounda ions wi h di e en samples o esponden s a ying in
hei likelihood o conside ing poli ical candidacy, we p o ide su ey-expe imen al e idence ha in o ma ion abou
he p e alence o poli ical c ime does no educe willingness o un o o ice o engage in poli ics among emale
esponden s wi h high poli ical in e es bu may do so among hose wi h low poli ical in e es . Taken oge he , his
s udy highligh s he esilience o unde ep esen ed g oups in he ace o inc easing poli ical iolence. Howe e , we
also show ha poli ical iolence may c ea e a pipeline p oblem i i de e s he wide popula ion o women om e en
conside ing o un o o ice.
Keywo ds: poli ical iolence; desc ip i e ep esen a ion; poli ical beha io ; gende and poli ics; local poli ics
In oduc ion
Violence agains poli ical eli es is on he ise ac oss ad anced indus ialized socie ies. B u al
mani es a ions o his end include he shoo ing o U.S. House Majo i y Whip S e e Scalise o he
mu de o Labou MP Jo Cox in 2016, bu less se e e o ms o poli ical iolence a e widesp ead,
pa icula ly in local poli ics. In a su ey among B i ish candida es, one hi d o all esponden s
indica ed ha hey had expe ienced in imida ing beha iou (Collignon & Rüdig, 2020). In a 2021
su ey, 72 pe cen o o e 1600 Ge man local poli icians epo ed ha hey had been he a ge o
e bal o physical abuse (Kommunalmagazin, 2021).
Beyond he magni ude o iolence, i is conce ning ha membe s o g oups, al eady
unde ep esen ed in poli ics, such as emale poli icians, end o disp opo iona ely become
he a ge s o a acks (Collignon & Rüdig, 2020; Håkansson, 2021). This aises ou cen al
esea ch ques ion: Does indi ec exposu e o poli ical iolence u he exace ba e gende gaps
in poli ical ambi ion and desc ip i e ep esen a ion? This is an impo an ques ion because
desc ip i e ep esen a ion can ansla e in o imp o ed subs an i e ep esen a ion (Mansb idge,
1999; Wängne ud, 2009). In addi ion, he legi imacy o ep esen a i e democ acy can be
© 2025 The Au ho (s). Eu opean Jou nal o Poli ical Resea ch published by John Wiley & Sons L d on behal o Eu opean Conso ium o Poli ical Resea ch.
This is an open access a icle unde he e ms o he C ea i e Commons A ibu ion License, which pe mi s use, dis ibu ion and ep oduc ion in any medium,
p o ided he o iginal wo k is p ope ly ci ed.
DOES POLITICAL VIOLENCE UNDERMINE DESCRIPTIVE REPRESENTATION? 2107
unde mined i he socio-demog aphic composi ion o he poli ical eli e pool de ia es oo s ongly
om he unde lying popula ion (A nesen & Pe e s, 2018).
We a gue ha iolence may de e women mo e s ongly han men om poli ical ac i i y
since poli ical iolence is disp opo iona ely di ec ed agains emale candida es and o iceholde s
(Buckley e al., 2024; Collignon e al., 2022; Daniele e al., 2023; Håkansson, 2021; K ook, 2020;
K ook & Res epo Sanín, 2020). Mo eo e , he p e alence o poli ical iolence may se e as a
signal o ‘nas y’ poli ics, which women end o dislike (F idkin & Kenney, 2011, p. 315; Kan hak
& Woon, 2015). Finally, gende di e ences in isk a e sion could esul in di e en ial eac ions o
poli ical iolence (C oson & Gneezy, 2009).
Al hough closely ela ed o ecen wo k on he consequences o iolence and hos ili y agains
women candida es and elec ed poli icians o women’s poli ical ep esen a ion (Bja negå d e al.,
2022; Buckley e al., 2024; Daniele e al., 2023; Eady & Rasmussen, 2024; E ikson e al.,
2023; Håkansson, 2024a, 2024b; He ick e al., 2021), ou esea ch goes beyond hese s udies by
examining indi ec exposu e o poli ical iolence and i s e ec on willingness o un and desc ip i e
ep esen a ion.1
We ocus on Ge many, whe e women a e s ikingly unde ep esen ed in local poli ics2and
whe e a acks agains poli icians inc eased d ama ically. In 2023, he numbe o a acks g ew
by 30 pe cen compa ed o 2022 (Bundesminis e ium des Inne n, ü Bau und Heima &
Bundesk iminalam , 2024). Su eys o local poli icians sugges ha women a e disp opo iona ely
a ge ed (Kommunalmagazin, 2020).3
Ou empi ical analysis elies on wo sou ces o e idence. Fi s , we examine he ela ionship
be ween a acks on poli icians and he p opo ion o emale candida es in Ge man local elec ions
using no el da a om o e 2000 municipali ies. Second, we employ p e- egis e ed su ey
expe imen s (N=3590) o expose esponden s o in o ma ion abou poli ical iolence and measu e
i s e ec s on poli ical ambi ion.4We use di e en samples based on esponden s’ poli ical in e es s
o assess hei likelihood o conside ing candidacy.
We ind no e idence ha a acks on poli icians educe he p opo ion o emale candida es
in a ec ed municipali ies. Ou expe imen shows ha in o ma ion abou poli ical iolence does
no disp opo iona ely discou age women wi h high poli ical in e es om unning o o ice;
i any hing, i may educe men’s willingness o engage in poli ics. While isk a e sion and a
p e e ence o coope a ion do no mode a e ea men e ec s, ou ea men inc eases he pe cei ed
isk o being a ge ed. Among women wi h low poli ical in e es , he e is some e idence ha
poli ical iolence may diminish poli ical ambi ion in his g oup.
O e all, ou indings sugges ha indi ec exposu e o poli ical iolence pe se may no
u he unde mine he desc ip i e ep esen a ion o women. They a e also consis en wi h esea ch
showing ha poli ical ambi ion and engagemen can inc ease when women a e exposed o
policy h ea s (Clay on e al., 2023) o h ea s o sexual iolence (K e , 2019). Fu he mo e,
ou s udy adds o ou unde s anding o he condi ions unde which salien iolence may
o may no educe women’s poli ical engagemen (Hadzic & Ta i s, 2019). Ou esul s also
highligh he need o sample ele an popula ions, as mo e likely candida es and highly
engaged cons i uen s may di e in impo an ways om he gene al popula ion (see Magalhães
& Pe ei a, 2024). Finally, ou indings sugges ha poli ical iolence can c ea e a pipeline
p oblem i i discou ages he la ge popula ion o women om e en conside ing unning o
o ice.
© 2025 The Au ho (s). Eu opean Jou nal o Poli ical Resea ch published by John Wiley & Sons L d on behal o Eu opean Conso ium o Poli ical Resea ch.
2108 JEYHUN ALIZADE ET AL.
Theo e ical backg ound and hypo heses
Why should he p e alence o poli ical iolence a ec he gende gap in poli ical ambi ion and
desc ip i e ep esen a ion? We p opose h ee non-mu ually exclusi e ac o s: The gende ed na u e
o poli ical iolence, po en ial gende di e ences in isk a e sion and a po en ially s onge
p e e ence o coope a ion and ci il poli ics among women.
Fi s , indi ec exposu e o poli ical iolence may cu b women’s poli ical ambi ion and c ea e a
gende gap in desc ip i e ep esen a ion i women an icipa e o be a g ea e isk in he poli ical
a ena. Depa ing om he inding ha poli ical iolence is disp opo iona ely di ec ed agains
women (Collignon e al., 2022; Daniele e al., 2023; Håkansson, 2021; K ook & Res epo Sanín,
2020; K ook, 2020; Yan & Be nha d, 2024), schola s ha e ound ha a acked emale poli icians
a e less likely o e- un (Daniele e al., 2023) and a e mo e likely o conside lea ing poli ics
(Håkansson, 2024b). These e ec s o di ec a acks could spill o e o women mo e gene ally gi en
mo e widesp ead expe iences o gende ed physical iolence and abuse among women (see, e.g.,
Mülle e al., 2007). This socie al backg ound ma e s because hos ili y owa ds women in poli ics
is no only pe cei ed by ci izens and poli icians as mo e se e e bu also pe cei ed as a emp s o
emo e women om poli ics (Eady & Rasmussen, 2024). In consequence, e en women who a e
no di ec ly a ge ed may an icipa e he gende ed na u e o poli ical iolence esul ing in a lowe
willingness o un o o ice among women gene ally and highe d opou among emale candida es
compa ed o male coun e pa s.
Resea ch on o he ypes o iolence lends some empi ical suppo o his no ion. Fo ins ance,
Hadzic and Ta i s (2019) show in a su ey expe imen ha making pas wa ime iolence salien
educes women’s willingness o pa icipa e in poli ics mo e han men’s. Simila ly, p o iding c oss-
na ional e idence, Wood (2024) inds ha elec o al iolence is linked o ewe women in elec ed
o ice.
Second, i women a e mo e isk-a e se han men, hey may be mo e likely han men o ac o
in he cos s o unning in a con en ious poli ical en i onmen . Al hough e idence has sugges ed
an exis ing gende gap in isk a e sion in he gene al public (C oson & Gneezy, 2009; Nicholson
e al., 2005; Pa e & Fox, 2018), i is unclea o wha ex en his gende gap gene alizes o po en ial
poli ical candida es and poli icians. Fo ins ance, Magalhães and Pe ei a (2024) show ha emale
candida es a e ac ually less isk-a e se han hei male coun e pa s. Simila ly, Linde and Vis
(2017) ind no gende di e ences in isk p e e ences among Du ch MPs. To ake his he e ogenei y
in o accoun , we use a dis inc , p e- ea men measu e o isk a e sion in ou su eys o es his
mechanism.
Thi d, he p e alence o poli ical iolence ma ks poli ics as con en ious and ‘nas y’. E idence
sugges s ha his may de e women in pa icula because hey dislike unci il campaigning (F idkin
& Kenney, 2011, p. 315) and end o adop a mo e coope a i e s yle o poli ics (Childs, 2004).
Acco dingly, Kan hak and Woon (2015, pp. 603–604) ind no gende gap in willingness o un
o o ice when campaigning is gua an eed o be hones and u h ul and he cos o compe ing is
low. Simila ly, P eece and S odda d (2015) ind ha p iming he compe i i eness o he poli ical
a ena educes women’s in e es in poli ical o ice bu no men’s. Finally, Schneide e al. (2016)
p o ide e idence ha women’s lack o in e es in powe and con lic can pa ly explain he gende
gap in poli ical ambi ion. This hypo hesized dynamic is no speci ic o he poli ical domain:
P e ious esea ch has, o ins ance, linked sexual ha assmen o gende -speci ic labou ma ke
lows (Adams-P assl e al., 2024; Ba u e al., 2022; Folke & Rickne, 2022). These s udies show
© 2025 The Au ho (s). Eu opean Jou nal o Poli ical Resea ch published by John Wiley & Sons L d on behal o Eu opean Conso ium o Poli ical Resea ch.
DOES POLITICAL VIOLENCE UNDERMINE DESCRIPTIVE REPRESENTATION? 2109
ha women a e less likely o apply o jobs in wo kplaces wi h a highe isk o sexual ha assmen
e en i al e na i e jobs o e lowe wages. In conclusion, women’s p e e ence agains ‘nas y
poli ics’ could cause indi ec exposu e o poli ical iolence o widen he gende gap in poli ical
ambi ion.
All h ee a ionales sugges ha he p e alence o poli ical iolence inc eases he gende gap
in poli ics. Combining obse a ional da a on poli ically mo i a ed c imes and he sha e o emale
candida es on local pa y lis s, we es he ollowing obse a ional implica ion o his a gumen :
H1: Fewe women will un o poli ical o ice in municipali ies ha ha e expe ienced mo e
c imes agains poli icians.
Focusing on indi iduals’ poli ical ambi ion, we es he ollowing indi idual-le el implica ion
o ou a gumen in an expe imen al se ing:
H2: In o ming indi iduals abou he p e alence o poli ical iolence will educe hei
willingness o un o o ice and engage in poli ics mo e s ongly among women han
among men.
To shed ligh on he unde lying mechanisms, we ocus on wo po en ial mode a o s discussed
abo e, namely isk a e sion and p e e ence o coope a ion in poli ics:
H3: In o ming indi iduals abou he p e alence o poli ical iolence educes hei willingness
o un o o ice and engage in poli ics. This e ec should be s onge when in e ac ing
wi h he ea men wi h isk a e sion.
H4: In o ming indi iduals abou he p e alence o poli ical iolence educes hei willingness
o un o o ice and engage in poli ics. This e ec should be s onge when in e ac ing
wi h he ea men wi h a p e e ence o coope a ion o e compe i ion in poli ics.
Empi ical s a egy
Obse a ional da a
Sample. We combine adminis a i e da a on c imes agains poli icians and poli ical pa ies wi h
adminis a i e da a on he sha e o emale candida es on pa y lis s in municipal elec ions in i e
Ge man s a es. Ou sample on he sha e o emale candida es co e s 4886 pa y lis s in 2079
municipali ies om he 2014 and 2019 municipal elec ions in Baden-Wü embe g, B andenbu g,
Saxony, Thu ingia as well as he 2016 and 2021 elec ions in Hesse. The i e s a es a e inhabi ed
by 23 million people, o oughly 25 pe cen o he Ge man popula ion. To hese da a on emale
candidacies, we ma ch he numbe o poli ically mo i a ed c imes in he espec i e municipali y,
o igina ing om p e iously classi ied police epo s o he Ge man Fede al C ime O ice (BKA).
We desc ibe he sample and da a o igin in mo e de ail in he Online Appendix Sec ions C.1 and
C.2.
Independen a iables. Ou independen a iables cap u e he p e alence o c ime in a
municipali y in he 12 mon hs ahead o he municipal elec ions. We dis inguish be ween h ee
o ms o a acks agains poli icians. Fi s , we c ea e a bina y and a coun a iable, espec i ely,
© 2025 The Au ho (s). Eu opean Jou nal o Poli ical Resea ch published by John Wiley & Sons L d on behal o Eu opean Conso ium o Poli ical Resea ch.
2110 JEYHUN ALIZADE ET AL.
o Any Poli ical C ime(s) agains poli ical ep esen a i es o pa y in as uc u e in a municipali y.
This includes mino c imes agains pa y campaign ma e ial, such as he demoli ion o pos e s. We
u he c ea e a bina y and a coun a iable cap u ing any A ack(s) agains Poli icians. Finally, we
also assess he in luence o Se e e A acks agains Poli icians on emale candidacies by ocusing on
se ious physical and/o psychological abuse. We ind ha he se e e a acks a e egula ly co e ed
by local media ou le s and should be salien enough o sway a i udes o o he local poli icians
in he a ea (see MDR Sachsen, 2024, and NDR, 2024, o examples o local co e age). The da a
on c ime s em om he o icial da abase o poli ically mo i a ed c imes (PMC) by he Ge man
Fede al C ime O ice, which was published due o pa liamen a y inqui ies. Mo e in o ma ion on
hei o igin and he documen s ha con ain in o ma ion on a acks is p esen ed in Sec ions C.1 and
Table C.1 o he Online Appendix.
Ou come. The Change in he Sha e o Female Candida es be ween municipal council elec ions
cons i u es he ou come o in e es . Fo each municipali y, we calcula e he di e ence in he
a e age sha e o emale candida es o he six main pa ies – CDU, SPD, G eens, FDP, Linke and
A D – be ween he wo local elec ions.
Examining he sha e o emale candida es in municipal elec ions is ideal o assessing he
impac o PMCs on women in poli ics. In e iews wi h local ep esen a i es and addi ional
e idence (see Online Appendix Sec ions B and C) sugges ha pa y lis nomina ions ace ewe
con ounding ac o s a he local le el. In ac , 68 pe cen o municipal pa y lis s in ou da a
s uggle o ill posi ions5and hus ely on olun ee s om he communi y, educing he in luence
o pa y quo as and ga ekeepe s. Fo p og essi e pa ies wi h o mal quo as, he a e age emale
candida e sha e is only 17 pe cen and changes in he sha e a e simila o pa ies wi hou quo as
(see Online Appendix Sec ion C.5). Candida e willingness o engage wi h a pa y is he e o e
c ucial in de e mining po en ial candidacies (see he e idence p esen ed in Online Appendix
Sec ion B.1).
Co a ia es and es ima ion. In all models, we con ol o municipal-le el di e ences in popula ion
and popula ion densi y. We also include s a e and elec ion-yea ixed e ec s. S anda d e o s a e
clus e ed by municipali y.
Su ey expe imen s
To measu e he gende ed e ec s o indi ec exposu e o poli ical iolence on poli ical ambi ion,
we conduc ed su ey expe imen s in Ge many.
Sample. We use he online access panel o he su ey i m Bilendi. O e all, we ec ui ed 3590
esponden s ac oss h ee su eys (July 2022; Augus /Sep embe 2022; Decembe 2022), 50 pe
cen o whom a e emale.6Ou main analysis ocuses on esponden s wi h high poli ical in e es
( i e o highe on a se en-poin scale), as hey ha e a highe baseline p obabili y o unning
o public o ice. This decision is based on quali a i e e idence (Alin e al., 2021, pp. 17–18)
sugges ing ha local poli ical candida es in Ge many o en en e poli ics due o in insic mo i a ion
o ec ui men by pee s, a he han s a egic ca ee goals. Gi en he lack o a clea pa hway
in o candidacy, we conside poli ical in e es as a necessa y condi ion ha c ea es a sample ha
© 2025 The Au ho (s). Eu opean Jou nal o Poli ical Resea ch published by John Wiley & Sons L d on behal o Eu opean Conso ium o Poli ical Resea ch.
DOES POLITICAL VIOLENCE UNDERMINE DESCRIPTIVE REPRESENTATION? 2111
a guably has a highe p opensi y o un o local o ice han an al e na i e sample ha would no
be il e ed by poli ical in e es .7
Ou in e iews wi h local pa y o ganize s (see Online Appendix B) e ealed ha ec ui ing
candida es a he local le el is di icul , o en esul ing in un illed spo s on pa y lis s – only 32
pe cen o local lis s ha e a ull sla e. This lack o compe i ion sugges s ha highly poli ically
in e es ed indi iduals may no only be mo e mo i a ed bu also be e posi ioned o un o o ice.
Ne e heless, in a seconda y analysis, we examine whe he poli ical iolence a ec s poli ical
ambi ion di e en ly among esponden s wi h a ying le els o poli ical in e es , using da a om
ou hi d su ey, which includes indi iduals wi h lowe poli ical in e es .
T ea men . We ely on an in o ma ion ea men . In he con ol condi ion, we use a ex o
in o m esponden s abou he impo ance and asks o local poli icians in Ge many. The ea men
condi ion includes an addi ional pa ag aph indica ing ha local poli icians inc easingly become
a ge s o poli ically mo i a ed c imes, accompanied by a g aphical ep esen a ion o he end in
c imes agains poli icians in Ge many be ween 2019 and 2021. The exac ea men wo ding and
he co esponding g aph can be ound in he ques ionnai e in Online Appendix F.5. (See Online
Appendix E.3 o balance es s.)
We e ained om using a ea men ha explici ly men ions he disp opo iona e a ge ing
o women. Fi s , ou heo e ical mechanisms summa ized in H3 and H4 do no ely on p iming
disp opo iona e a ge ing by gende . Ra he , i men and women di e in hei isk a e sion and
hei p e e ence o poli ical s yle, iolence should gene ally esul in gende ed eac ions. Second,
due o a lack o eliable da a on he gende ed na u e o poli ical iolence in Ge many, we e ained
om p o iding esponden s wi h in o ma ion ha we we e no able o ully back up wi h da a.
Finally, we wan ed o ensu e he expe imen al po ion o he s udy o be compa ible wi h ou
obse a ional analysis, whe e we do no di e en ia e poli ical iolence by gende due o da a
a ailabili y easons.
Ano he conce n wi h he in o ma ion ea men is whe he i p o ides no el in o ma ion o
esponden s, especially hose wi h high poli ical in e es . A sys ema ic analysis using he Nexis Uni
da abase ound only 419 a icles on iolence agains poli icians published in Ge man newspape s
in he las 5 yea s – compa ed o 1708 a icles on iolence agains ano he ype o public o icial
– he police. This sugges s ha public awa eness o iolence agains poli icians is ela i ely low,
making ou ea men likely o in oduce no el in o ma ion o many esponden s.
Ou comes. Ou main ou comes o in e es ap in o esponden s’ willingness o un (WTR)as
a candida e in local council elec ions in hei municipali y and, as a lowe le el measu e o
pa icipa ion, hei willingness o engage (WTE) in he local poli ics o hei communi y. Following
Mendelbe g and Wes (2021), we addi ionally designed a quasi-beha iou al ou come in which
we p o ided esponden s in he i s wo su ey ounds wi h he op ion o lea n mo e abou
adminis a i e s eps equi ed o candidacy by clicking on a sepa a e page (hence o h In o ma ion
ou come). We measu e whe he a esponden op ed o iew he in o ma ion as a bina y ou come.
Finally, as seconda y ou comes, we examine he ea men e ec on esponden s’ pe cei ed
pe sonal dange / isk o candidacy (included in ounds one and wo) and poli ical engagemen
(included in all h ee ounds). The pu pose o hese ou come a iables is o check i he ea men
is s ong enough o ha e any e ec on esponden s’ belie s since exposu e o in o ma ion abou
poli ical iolence should mos di ec ly a ec isk pe cep ions.
© 2025 The Au ho (s). Eu opean Jou nal o Poli ical Resea ch published by John Wiley & Sons L d on behal o Eu opean Conso ium o Poli ical Resea ch.
2112 JEYHUN ALIZADE ET AL.
Table 1. Poli ical c ime and emale candida es in municipal elec ions
Dependen a iable
Change in pe cen sha e o emale candida es on pa y lis s
(1) (2) (3) (4) (5) (6)
Any Poli ical C ime −1.744*
(0.854)
Any Poli ical C ime Coun −0.001
(0.012)
A ack agains Poli ician 0.510
(1.284)
A ack agains Poli ician Coun 0.364*
(0.181)
Se e e A ack agains Poli ician 0.451
(1.753)
Se e e A ack agains Poli ician Coun 0.424
(0.310)
Popula ion log. 0.007 −0.149 −0.173 −0.215 −0.161 −0.184
(0.372) (0.378) (0.383) (0.382) (0.377) (0.378)
Popula ion densi y 0.00004 −0.0001 −0.0001 −0.0002 −0.0001 −0.0002
(0.001) (0.001) (0.001) (0.001) (0.001) (0.001)
S a e+yea FE Yes Yes Yes Yes Yes Yes
Obse a ions 2079 2079 2079 2079 2079 2079
Adjus ed R2 0.006 0.004 0.004 0.005 0.004 0.004
No e: The able p esen s he esul s o mul i a ia e o dina y leas squa es (OLS) eg essions on he municipali y
le el. All models include he log popula ion and popula ion densi y as municipal-le el co a ia es as well as s a e
and elec ion yea ixed e ec s. S anda d e o s a e clus e ed on he municipal le el.*p<0.05; **p<0.01.
A i udinal p e- ea men mode a o s. We measu e isk a e sion using ou i ems om he Gene al
Risk P opensi y Scale (Zhang e al., 2019), c ea ing an index om one o i e, wi h highe
alues indica ing g ea e isk a e sion; his was measu ed in he i s wo su eys. P e e ence o
a coope a i e poli ical s yle is measu ed wi h a bina y indica o : one i he esponden alues
coope a ion and consensus o e compe i ion and con lic in democ a ic poli ics and ze o i hey do
no . This was included in all h ee su eys.
Resul s
Obse a ional analysis
Table 1shows he esul s o six linea eg ession models assessing he e ec o poli ical c ime on
emale ep esen a ion in Ge man local elec ions.
© 2025 The Au ho (s). Eu opean Jou nal o Poli ical Resea ch published by John Wiley & Sons L d on behal o Eu opean Conso ium o Poli ical Resea ch.
DOES POLITICAL VIOLENCE UNDERMINE DESCRIPTIVE REPRESENTATION? 2113
Acco ding o model one, we obse e a decline in he sha e o emale candida es by oughly 1.7
pe cen age poin s when a municipali y was exposed o a leas one poli ical c ime, which includes
less se e e c imes like p ope y damage. Howe e , he coun -based models show ha highe c ime
a es do no ansla e in o ewe emale candidacies. When excluding less se e e c imes in models
3–6, we show ha (se e e) a acks agains poli icians do no educe emale ep esen a ion in pa y
lis s. I any hing, he es ima es skew posi i e and mos ly become s a is ically insigni ican . These
esul s pe sis i we omi any speci ic pa y in ou sample (see Online Appendix Sec ion C.5).
In addi ional es s p esen ed in Online Appendix Sec ion C.4, we show ha ou ou come does
no p edic he occu ence o c ime and ha esul s a e simila when looking a le els ins ead o
wi hin-municipali y changes in he sha e o emale candida es. We also do no ind any e idence
ha poli ical c ime is sys ema ically ela ed o he posi ion o women on candida e lis s (see Online
Appendix Sec ion D). Finally, ou esul s a e simila in con ex s wi h high o low le els o emale
ep esen a ion (see Online Appendix Sec ion C.7).
O e all, we conclude ha poli ical iolence – bo h in e ms o quali y and quan i y – does
no sys ema ically co ela e wi h he desc ip i e ep esen a ion o women (in con as o H1). This
sugges s he esilience o emale poli icians in he ace o ad e se poli ical condi ions.
Su ey expe imen s
The esul s o he su ey expe imen e lec he indings om he obse a ional analysis. As
Figu e 1demons a es o ou h ee main ou come a iables, he e is no nega i e in e ac ion e ec
be ween exposu e o in o ma ion abou poli ical iolence and he dummy indica o o emale
esponden s wi h high poli ical in e es (H2).8I any hing, exposu e o in o ma ion abou iolence
agains poli icians may ac ually educe he gende gap, gi en he posi i e in e ac ion e ec o WTE
as he ou come a iable. This esul is d i en p ima ily by male esponden s, who a e de e ed by
ou ea men , while emale esponden s a e una ec ed by i . The e is, howe e , no di e ence in
he ea men e ec be ween males and emales o WTR and he quasi-beha iou al in o ma ion
ou come.9In sum, con a y o ou hypo hesis, in o ming esponden s wi h high poli ical in e es
abou he p e alence o poli ical iolence does no ha e a s onge de e en e ec on women’s
poli ical ambi ions han on men’s.
Tu ning o po en ial mechanisms o he hypo hesized gende di e ence, Figu e 2shows he
ea men ’s in e ac ion wi h ou p e- ea men a i udinal measu es. Fi s , we do no ind di e en ial
e ec s based on an indi idual’s le el o isk a e sion. Thus, in con as o H3, isk a e sion,
which in ou sample ends o be sligh ly highe among women,10 does no in e ac wi h he
in o ma ion ea men . Simila ly, ac oss all h ee ou comes, we ind no e idence ha p e e ence
o a coope a i e s yle o poli ics mode a es ou ea men e ec (in con as o H4).11
In Figu e 3, we explo e he e ec o he ea men on he pe cep ion o he pe sonal dange / isk
associa ed wi h (1) poli ical candidacy and wi h (2) poli ical engagemen on he local le el. The
ea men aises he pe cei ed pe sonal isk/dange o candidacy among bo h gende s. The e is
also a posi i e ea men e ec on he pe cei ed isk/dange o poli ical engagemen in he emale
subsample. These indings a e eassu ing om a esea ch-design pe spec i e since hey sugges
ha ou manipula ion was s ong enough o shape belie s abou he isk/dange o en e ing local
poli ics – an ou come ha should be mos immedia ely a ec ed by he in o ma ion ea men .
This isk pe cep ion, howe e , does no ansla e in o lowe willingness o en e he poli ical a ena
© 2025 The Au ho (s). Eu opean Jou nal o Poli ical Resea ch published by John Wiley & Sons L d on behal o Eu opean Conso ium o Poli ical Resea ch.
2120 JEYHUN ALIZADE ET AL.
He ick, R., S. Thomas, L. F anklin, M. L. Godwin, E. Gnabasik, & J. R. Sch oedel. (2021). Physical iolence and
psychological abuse agains emale and male mayo s in he Uni ed S a es. Poli ics, G oups, and Iden i ies,9(4),
681–698.
Kan hak, K., & J. Woon. (2015). Women don’ un? Elec ion a e sion and candida e en y. Ame ican Jou nal o
Poli ical Science,59(3), 595–612.
Kommunalmagazin. (2020). Meh F auen in Deu schlands Ra häuse . Kommunal. h ps://kommunal.de/meh -
Bue ge meis e innen-EAF
Kommunalmagazin. (2021). A acken au Kommunalpoli ike in de Co ona-Pandemie wei e ges iegen. h ps://
kommunal.de/a acken-kommunalpoli ike -co ona
K e , A.-K. (2019). Responding o sexual iolence: Women’s mobiliza ion in wa . Jou nal o Peace Resea ch,
56(2), 220–233.
K ook, M. L. (2020). Violence agains women in poli ics. Ox o d Uni e si y P ess.
K ook, M. L., & J. Res epo Sanín. (2020). The cos o doing poli ics? Analyzing iolence and ha assmen agains
emale poli icians. Pe spec i es on Poli ics,18(3), 740–755.
Linde, J., & B. Vis. (2017). Do Poli icians ake isks like he es o us? An expe imen al es o p ospec heo y
unde MPs. Poli ical Psychology,38(1), 101–117.
Lukoscha , H., & R. Köche . (2021). Pa eikul u en Und Die Poli ische Teilhabe on F auen. Eine Empi ische
Un e suchung Mi Handlungsemp ehlungen an Die Pa eien. EAF Be lin. Di e si y in Leade ship. h ps://www.
ea -be lin.de/ ileadmin/ea /Publika ionen/Dokumen e/2022EAFBe linLukoscha K
Magalhães, P. C., & M. M. Pe ei a. (2024). Women unning o o ice a e less isk a e se han men: E idence om
Po ugal. The Jou nal o Poli ics,86(3), 1093–1097.
Mansb idge, J. (1999). Should blacks ep esen blacks and women ep esen women? A con ingen “yes.” The
Jou nal o Poli ics,61(3), 628–657.
MDR Sachsen. (2024). S einwü e und Beleidigung: Wiede Ang i e au Poli ike in Sachsen. h ps://www.md .de/
nach ich en/sachsen/poli ik/ang i e-wahlkamp -poli ike -100.h ml
Mendelbe g, T., & E. A. Wes . (2021). How elec ion cha ac e is ics explain gende gaps in poli ical ambi ion. P e-
analysis plan. h ps://doi.o g/10.17605/OSF.IO/CT3KD
Mülle , U., M. Sch ö le, S. Glammeie , & C. Oppenheime . (2007). Heal h, well-being and pe sonal
sa e y o women in Ge many. A ep esen a i e s udy o iolence agains women in Ge many. Fede al
Minis y o Families, Senio Ci izens, Women and You h. h ps://www.bm s j.de/ esou ce/blob/93906/
9c0076 c66b1be6d0eb28258 e0aa569/ auens udie-englisch-gewal -gegen- auen-da a.pd
NDR. (2024). Ang i e au Poli ike : Neue Zahlen ü Hambu g. h ps://www.nd .de/nach ich en/hambu g/Ang i e-
au -Poli ike -Neue-Zahlen- ue -Hambu g,ang i e116.h ml
Nicholson, N., E. Soane, M. Fen on-O’C ee y, & P. Willman. (2005). Pe sonali y and domain-speci ic isk aking.
Jou nal o Risk Resea ch,8(2), 157–176.
Pa e, J., & R. Fox. (2018). Ge ing pas he gende gap in poli ical ambi ion. Jou nal o Economic Beha io &
O ganiza ion,156, 166–183.
Po mann, L., & N. S ojano i´
c. (2019). Elec o al disc imina ion agains immig an -o igin candida es. Poli ical
Beha io ,41(1), 105–134.
P eece, J., & O. S odda d. (2015). Why women don’ un: Expe imen al e idence on gende di e ences in poli ical
compe i ion a e sion. Jou nal o Economic Beha io & O ganiza ion,117, 296–308.
Schneide , M. C., M. R. Holman, A. B. Diekman, & T. McAnd ew. (2016). Powe , con lic , and communi y: How
gende ed iews o poli ical powe in luence women’s poli ical ambi ion. Poli ical Psychology,37(4), 515–531.
Wängne ud, L. (2009). Women in pa liamen s: Desc ip i e and subs an i e ep esen a ion. Annual Re iew o
Poli ical Science,12, 51–69.
Wood, R. M. (2024). The e ec s o elec o al iolence on women’s legisla i e ep esen a ion. Poli ical Resea ch
Qua e ly,77(3), 915–930.
Yan, A. N., & R. Be nha d. (2024). The silenced ex : Field expe imen s on gende ed expe iences o poli ical
pa icipa ion. Ame ican Poli ical Science Re iew,118(1), 481–487.
Zhang, D. C., S. Highhouse, & C. D. Nye. (2019). De elopmen and alida ion o he Gene al Risk P opensi y Scale
(GRiPS). Jou nal o Beha io al Decision Making,32(2), 152–167.
© 2025 The Au ho (s). Eu opean Jou nal o Poli ical Resea ch published by John Wiley & Sons L d on behal o Eu opean Conso ium o Poli ical Resea ch.
DOES POLITICAL VIOLENCE UNDERMINE DESCRIPTIVE REPRESENTATION? 2121
Add ess o co espondence: Mig a ion, In eg a ion, T ansna ionaliza ion Uni , WZB Be lin, D-10785 Be lin,
Ge many.
Email: [email protected]
© 2025 The Au ho (s). Eu opean Jou nal o Poli ical Resea ch published by John Wiley & Sons L d on behal o Eu opean Conso ium o Poli ical Resea ch.
