Ki el, Rebecca C.
A icle — Published Ve sion
Le 's alk populis ? A su ey expe imen on e ec s o
(non‐) populis discou se on o e choice
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: Ki el, Rebecca C. (2025) : Le 's alk populis ? A su ey expe imen on e ec s o
(non‐) populis discou se on o e choice, Eu opean Jou nal o Poli ical Resea ch, ISSN 1475-6765,
Eu opean Conso ium o Poli ical Resea ch, Colches e , Vol. 64, pp. 719-743,
h ps://doi.o g/10.1111/1475-6765.12710
This Ve sion is a ailable a :
h ps://hdl.handle.ne /10419/311287.2
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/
Eu opean Jou nal o Poli ical Resea ch 64: 719–743, 2025 719
doi: 10.1111/1475-6765.12710
Le ’s alk populis ? A su ey expe imen on e ec s o (non-) populis discou se
on o e choice
REBECCA C. KITTEL1,2
1Ins i u e o Eas Eu opean S udies, F ee Uni e si y o Be lin, Be lin, Ge many; 2WZB Be lin Social Science
Cen e , Cen e o Ci il Socie y Resea ch, Ge many
Abs ac . Populism esea ch has ound much schola ly and public a en ion alike in ecen yea s. Mos esea ch has
ocused on how populism can be de ined, assessed o e en measu ed. E en hough he e a e eme ging s udies on
populis messages, ew o hem ha e paid a en ion on causally iden i ying ways in which discou se can a ec
suppo o populis ac o s. This a icle posi ions i sel wi hin his gap and aims o answe which discu si e
elemen s make (non-)populis messages appealing o a ying g oups o people. To answe his esea ch ques ion,
I conduc ed a no el su ey expe imen on o e choice in Ge many om Decembe 2020 o Janua y 2021 wi h N
=3325. Responden s we e asked o choose be ween wo candida e s a emen s ha displayed a ying discu si e
elemen s. Thus, he expe imen causally es ed whe he people-cen ic he o ic, blame a ibu i e languages o
populis s yle ocusing on language complexi y d i e he populis o e. Resul s show ha a neu al o m o
blame a ibu ion, namely owa ds poli icians, had he highes p obabili y o d i ing o e choice, i espec i e o
esponden s’ unde lying ideological p e e ences o populis a i udes. Simple language nea ly always has a nega i e
e ec on o e choice, whe eas people-cen ism adds a posi i e ouch. These esul s show ha he e may be an
inc easing dissa is ac ion wi h democ acy ha is oiced by blaming poli ical eli es o he mal unc ioning o socie y.
Keywo ds: populism; language complexi y; blame a ibu ion; o ing beha iou ; su ey expe imen
In oduc ion
Populism is on he ise. In many coun ies, populis ac o s ha e become e y success ul and
may e en gain go e nmen ep esen a ion (Ellge , 2024; Robe s, 2022; Spoon & Klü e , 2019;
Vachudo a, 2021). As such, he igh -wing populis pa y, Pa y o F eedom, became he la ges
pa y du ing he 2023 Du ch pa liamen a y elec ions wi h a o e sha e o 23.7 pe cen . In Sweden,
he igh -wing populis pa y, Sweden Democ a s, became he second la ges o ce du ing he 2022
pa liamen a y elec ions and e en en e ed a go e nmen coali ion o he i s ime. And, in I aly,
Gio gia Meloni, he igh -wing populis pa y leade o B o he s o I aly, became I aly’s i s emale
P ime Minis e in 2022.
O en, he success o populis pa ies is associa ed wi h democ a ic backsliding o challenges
owa ds libe al democ acy (Albe azzi & Muelle , 2013; Rummens, 2017; U bina i, 2019).
Howe e , despi e populism being s udied om a ious angles (Hawkins e al., 2018; Mudde, 2004;
Tagga , 2000), i is s ill no clea why and when di e en o e s suppo populis candida es.
One majo elemen o en associa ed wi h he success o populis ac o s is hei discu si e
s a egy (Aalbe g e al., 2016; Aslanidis, 2018) as populis communica ion can ha e eno mous
e ec s on ci izens’ opinions and hei poli ical beha iou , including o ing (Bos e al., 2013;
Hamelee s e al., 2017; Muelle , 2016; Reinemann e al., 2019;Wi z,2018). Al hough some
ecen s udies ocus on he e ec o blame a ibu ion in populis discou se (Bos e al., 2018,
© 2024 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.
720 REBECCA C. KITTEL
2020; Busby e al., 2019) o on which a i udes o p e-disposi ions a ec he populis o e
(Akke man e al., 2014; Cas anho Sil a e al., 2018; Hieda e al., 2021), li le is known abou
how elemen s o (non-)populis discou se causally a ec o ing beha iou .
Theo e ical and empi ical conside a ions sugges ha blame a ibu ion is no he only decisi e
elemen o populis discou se. Resea ch also conside s people-cen ic he o ic (Meije s & Van
De Velden, 2022; Rooduijn & Akke man, 2017) and language simplici y (Bischo & Senninge ,
2018) as commonly used ools o populis he o ic. Conside ing hese s ill unde s udied elemen s
when s udying he e ec o populis discou se on ci izens, his a icle asks wha makes he populis
appeal e ec i e.
Focusing on a ying discu si e elemen s, I a emp o answe his esea ch ques ion by looking
a di e en elemen s o populis and non-populis messages, es ing hei e ec on o e choice. To
his end, I conduc ed an ex ensi e o e expe imen in Ge many be ween Decembe 2020 and
Janua y 2021 (N=3325).1The o ced choice igne e expe imen asked esponden s o cas
hei o e choice be ween wo candida e s a emen s. These s a emen s a ied ac oss he h ee
discu si e elemen s o blame a ibu ion, people-cen ism and language complexi y o cap u e
di e en populis and non-populis messages.
Con a y o p io esea ch wi h obse a ional da a, he esul s show ha populis a i udes
a e no he main d i e o populis o ing bu ha blame a ibu ion is he mos decisi e d i e
o o e choice. Blame a ibu ion was measu ed wi h h ee ideological a ian s: le (blaming
capi alis s), igh (blaming e ugees) and neu al (blaming poli icians). Ideology exp essed h ough
blame a ibu ion co esponds highly wi h o e s’ o e in en ions. This shows ha ideology seems
mo e decisi e han populism in explaining o e choice. Simple language, on he o he hand, in
con adic ion o i s heo e ical conside a ions, has an o e all nega i e e ec on o e choice. People-
cen ism has a posi i e e ec on o e choice in mos s a emen con igu a ions. The esul s show ha
people end o like poli icians who c ea e an in-g oup eeling and conside hemsel es pa o he
o e g oup. Su p isingly, he neu al e sion o blame owa ds poli icians d i es o e choice ac oss
all o e g oups no only o e s wi h populis a i udes o o e in en ions. These esul s sugges an
o e all dissa is ac ion wi h he unc ioning o poli ical sys ems and he wo k o poli icians. The
esul s also hold i espec i e o o e s’ ideological p e e ences. Fu he , hey a e in line wi h o he
expe imen al s udies showing ha indi iduals’ ideology p e e ences de e mine o e choice ins ead
o populis a i udes (Cas anho Sil a & W a il, 2023; Dai & Kus o , 2022; Neune & W a il,
2022). Howe e , his s udy goes beyond hese conjoin expe imen s as i no only es s people-
cen ism and an i-eli ism as elemen s o populism bu adds e en mo e a ibu es o en associa ed
wi h populism, such as blame a ibu ion and simple language. S ill, his analysis shows ha a ious
populis elemen s ha e a minimal e ec on an indi idual’s o ing decision, showing ha ideology
is s ill he main d i e o o e choice.
E en hough he s udy was only conduc ed in Ge many, esul s may also be applicable o o he
Eu opean coun ies wi h simila poli ical sys ems. As Hauwae and Van Kessel (2018)aswellas
Reinemann e al. (2019) ha e shown, populis o e s ac oss se e al coun ies in Eu ope display
simila cha ac e is ics and espond o simila pa e ns. The e o e, we can assume ha he main
esul s om his su ey expe imen also a el beyond Ge many.
As such, his analysis con ibu es o he c ucial deba e on populism, i s ising popula i y
in ecen yea s, and he challenges i poses owa ds ep esen a i e democ acy. Mo e b oadly,
his analysis con ibu es o he deba e on o e choice and how ce ain discu si e elemen s can
in luence o ing beha iou . I p o ides a causal link be ween (non-)populis discou se and how
© 2024 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.
LET’S TALK POPULIST? 721
i a ec s o e s, showing ha a ious o e g oups espond su p isingly posi i ely o simila
language pa e ns.
Populis discou se: Theo e ical conside a ions and hypo heses
Resea che s ha e discussed many concep s when de ining populism. Main esea ch s ands include
populism as a s a egy (Weyland, 2001), populism as an o ganisa ional o m (Robe s, 2006)
o a cul u al s yle (Os iguy, 2017). Mos p ominen ly, schola s ha e discussed he idea ional
app oach es ablished by Mudde (2004). Especially when analysing discu si e elemen s o
populism o populis communica ion s a egies, schola s ag ee on co e ea u es included wi hin
he idea ional app oach (Hunge & Pax on, 2022). As such, his a icle is also based on he main
assump ions o he idea ional app oach and combines hem wi h ideas o how hey can be enac ed
h ough language.
As such, discou se is no he only elemen ha ma e s when de ining populism. Howe e ,
schola s ag ee ha i is impossible o unde s and populism wi hou analysing i s discu si e pa e ns
(Aslanidis, 2018). Gi en ha speci ic language pa e ns a e c ucial o populism, hey can be
conside ed ele an o he success o populis s. The e o e, we mus unde s and how hese language
pa e ns a ec o e s and, mo e p ecisely, o e choice, i we aim o unde s and why populis ac o s
ha e become mo e success ul a ound he globe.
In his pape , I econcep ualise populism in e ms o h ee elemen s displayed in populis
discou se: people-cen ism, ideologically oned blame a ibu ion, and language complexi y. The
i s wo elemen s conside he con en side o he discou se, whe eas he las elemen ocuses on
a s ylis ic elemen . I chose o analyse a combina ion o hese h ee elemen s as hey a e e e enced
as key ea u es in he li e a u e on populism based on populism as discou se and he idea ional
app oach. Howe e , he h ee elemen s a e o en discussed sepa a ely ins ead o connec ed o each
o he . The e o e, I explain people-cen ism, ideologically oned blame a ibu ion, and language
complexi y and discuss how he in e play o hese h ee elemen s c ea es he heo e ical ounda ion
o populism.
People-cen ic he o ic
E en hough he e is no unique and ully encompassing de ini ion o populism, cu en schola ship
om Wes e n democ acies la gely ag ees on he idea ional app oach owa ds populism (Rooduijn,
2019). This app oach classi ies he wo ld o poli ics in o a dualis ic iew o good e sus e il. As
such, populis s claim o be he ue democ a s who de end he will o he common and ‘good’
people agains he co up and ‘e il’ eli es (Cano an, 1999; Mudde, 2004). Following he populis
logic, he people and hei so e eign y a e he co e o democ acy, which is h ea ened by he eli es;
he people a e conside ed he basis o a good socie y (Albe azzi & McDonnell, 2008;Mény&
Su el, 2002).
This de ini ion cap u es he ecognisable ea u es schola s ag ee on, ha is, he ‘us’ e sus
‘ hem’ men ali y. This unde s anding is ansmi ed h ough hei he o ic and plays a majo ole
in how populis s communica e wo ldwide. I will e e o his he o ical s a egy in his a icle as
people-cen ic he o ic. This he o ic s a egy unde lines ha populis ac o s use ‘us’ o alk abou
he common people o whom hey unde s and hemsel es as ad oca es. Following he populis
wo ld iew, he common people a e usually classi ied as a so e eign, homogeneous, pu e and
© 2024 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.
722 REBECCA C. KITTEL
i uous g oup (Akke man e al., 2014). Populis s aim o c ea e an in-g oup eeling ha common
people can eel pa o . Consequen ly, hey use he e e ence o he ‘o he ’ o c ea e an opposing
en i y. This e e ence helps he ‘us’-g oup o in-g oup o dema ca e i sel be e , bu i also c ea es
a ‘ hem’-g oup o ou -g oup ha can be opposed and blamed o socie al o poli ical p oblems.
Populis ac o s align hese g oups wi h hei populis iew o a dualis ic wo ld. This wo ld iew
is ansmi ed oci e ously h ough hei he o ic, unde lining ha people-cen ic he o ic is a
signi ican elemen o populis discou se.
E en hough using people-cen ic he o ic o e e ences o he ‘us’-g oup is a s ong
cha ac e is ic o populis discou se, populis ac o s do no uniquely apply i . Mains eam ac o s
also use people-cen ic he o ic in hei messages o c ea e a eeling o p oximi y o hei
elec o a e. Thus, people-cen ic he o ic can be unde s ood as a discu si e elemen ha populis
and mains eam ac o s employ. Ne e heless, populis ac o s end o o e -emphasise his elemen
and speak o people who a e mo e esponsi e o i . Thus, I expec a posi i e e ec o people-cen ic
he o ic in poli ical messages ac oss all o e g oups. Howe e , populis ac o s bene i e en mo e
om he use o people-cen ic he o ic. This leads o he ollowing hypo hesis.2
People-Cen ism Hypo hesis (H1). People-cen ic he o ic posi i ely a ec s o e choice
in gene al, bu i has an e en s onge e ec among indi iduals wi h populis o e in en ions
o a i udes.
Blame a ibu i e language
Besides he undamen al di ision o ‘us’ e sus ‘ hem’, schola s a gue ha populis ac o s exploi
social iden i y ames o emphasise in- and ou -g oups e en mo e (Bos e al., 2020). Commonly,
populis ac o s aim o unde line in-g oup a ou abili y and ou -g oup hos ili y. Using discu si e
elemen s, hey c ea e a subjec i e sense o un ai ea men among people o he in-g oup enac ed
h ough people o he ou -g oup o socie y (Elcha dus & Sp uy , 2016). Iden i ying a g oup o
people by c ea ing an opposing g oup o people makes hem eel mo e a ached o hei own g oup,
as i c ea es a clea iden i ica ion wi hin g oup bounda ies.
Acco ding o social psychological heo y, populis ac o s apply psychological mic o-
mechanisms ha c ea e iden i y h ough discou se. Thus, I ely mo e p ecisely on social iden i y
heo y in his analysis. Commonly, social iden i y heo y a gues ha an in-g oup c ea es a
posi i e assessmen o i s own g oup in dema ca ion om he o he (ou -)g oups. Addi ionally,
he heo y poin s ou ha in-g oup membe s usually sha e emo ional in ol emen and achie e
social consensus wi hin he g oup (Taj el & Tu ne , 2004), which uni es hem as g oup membe s.
In ela ing hese assump ions o people who a e p one o populis ideas, esea ch shows ha
such indi iduals o en ha e p oblems wi h inding a posi i e social iden i y (Sp uy e al., 2016).
As he wo ld has become mo e complex, people ha e s uggled o ind hei ole and iden i y in i –
o as Ros bøll (2023) pu s i : ‘many people ha e los hei o me ly secu e eelings o iden i y
and s a us, and hey a e inc easingly u ning hei us a ion a hei loss owa d he poli ical
sys em by suppo ing populis pa ies and leade s’ (Ros bøll, 2023, p. 2). Huddy (2001) emphasises
ha i is, he e o e, impo an o conside iden i y choices. She emphasises ha choosing o be
pa o a g oup olun a ily inc eases sel -es eem and g oup cohesion leading o highe in-g oup
iden i ica ion bu also mo e disc imina ion agains ou -g oups (Huddy, 2001, p. 139) (see also
Pe eaul & Bou his, 1999 and Tu ne e al., 1984). Thus, when people choose o suppo a populis
© 2024 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.
LET’S TALK POPULIST? 723
pa y o mo emen , i is because hey ha e los pa o hei o me iden i y o eel disappoin ed
by an in-g oup hey used o be pa o . These nega i e eelings uni e hem, and le hem c ea e
a new iden i y – an iden i y o disappoin men in he cu en poli ical sys em o , mo e b oadly,
socie y in gene al. Populis ac o s a e awa e o hei suppo e s’ disappoin men . They use his
disappoin men s a egically o o m a new iden i y and in-g oup among hese people.
As such, blame a ibu ion is added as an addi ional laye o people-cen ic he o ic. I o ges
iden i y h ough an in-g oup e sus ou -g oup unde s anding o he wo ld. The mos gene ic o m
o blame a ibu ion is blame agains poli icians o , mo e b oadly, agains he poli ical sys em.
Commonly, all populis s blame poli icians as i helps populis ac o s o c ea e a gene ic ou -g oup
ha can be held accoun able o he mal unc ioning o socie y. Usually, blame agains poli icians
is de ached om any ideological placemen . The e o e, blame a ibu ion agains poli icians can
be seen as a neu al e sion o populis blame a ibu ion and can se e o de ec he gene al
unc ioning o populis he o ical s a egies. Howe e , blaming poli icians wi hou using any o m
o people-cen ism is no gene ically populis . Mains eam ac o s can also blame o he poli ical
eli es in hei discou se. The e o e, i is essen ial o highligh ha populis he o ic consis s always
o bo h elemen s, people-cen ism as well as an i-eli ism (Mudde, 2004).
As blame a ibu ion is enac ed h ough language, i can be conside ed o p oduce aming
e ec s (Busby e al., 2019). Thus, i is impo an o no ice ha aming e ec s can be en o ced
h ough p e-disposi ions, such as alues ha accompany blame-a ibu i e language. Indi iduals’
unde lying a i udes o alues can a ec how much ce ain ames in luence hem. Chong and
D uckman (2007) poin ou ha ‘indi iduals who ha e s ong alues a e less amenable o ames
ha con adic hose alues’ (Chong & D uckman, 2007, p. 111). I indi iduals a e jux aposed
wi h he di e en a gumen s o an issue, hey commonly choose he al e na i e close o hei
p e-exis ing alues and unde lying p inciples (Snide man & The iaul , 2004). This can be ela ed
o populis he o ic and, mo e p ecisely, blame a ibu ion. Indi iduals wi h unde lying populis
a i udes will acco dingly espond mo e o ames ha build up on hese p e-disposi ions. As well,
indi iduals who sha e populis a i udes will a ou ames om populis ac o s ha apply hese
he o ical s a egies. Thus, indi iduals who sha e populis a i udes o o e in en ions should be
mo e likely o o e o a candida e who blames poli ical eli es o socie al p oblems. This leads o
he ollowing hypo hesis.3
Poli ician Hypo hesis (H2a). Blame a ibu ion agains poli icians posi i ely a ec s o e
choice o people wi h populis o e in en ions o a i udes.
So a , I ha e discussed populis ac o s as one uni y. Howe e , as ou lined ea lie , populism is
o en a ached o a hos ideology. Populism is commonly combined wi h adical igh and adical
le ideologies (Mudde & Kal wasse , 2012). E en hough, i is essen ial o no e ha cen is
pa ies can also adop populis s a egies, such as he Fi e S a Mo emen in I aly. Howe e , in his
analysis, I will ocus on populis ac o s and hei suppo e s om he adical le and he adical
igh , as hese a e he pa ies ha mos commonly adop populis s ances (Akke man e al., 2014).
As such, blame a ibu ion a ies depending on he hos ideology o a populis ac o . By
e e encing ce ain in- and ou -g oups, populis ac o s can di ec hei ideological posi ioning.
As le -wing (populis ) ac o s usually apply a discou se ha cen es a ound social class, hey
o en c ea e an ou -g oup ha consis s o he economic eli es o simply ‘ he ich’. Righ -wing
(populis ) ac o s, on he o he hand, apply an e hic and na i is -guided discou se. They usually
a ge immig an s and/o e ugees as an ou -g oup and blame hem o he mal unc ioning o
© 2024 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.
724 REBECCA C. KITTEL
a ious socie al p oblems. Thus, ideological di e ences can be emphasised and e en s eng hened
by blaming a chosen o he .
Howe e , he success o blame a ibu ion os e ed by populis pa ies on he le o he igh will
co ela e wi h o e s’ ideological p e-disposi ions, jus as a ibu ing blame o poli ical eli es does
oo. Consequen ly, placing blame on immig an s and/o e ugees will mos ly a ec people who
al eady ha e exis ing igh -wing a i udes o o e in en ions. Co espondingly, a ibu ing blame o
economic eli es will a ac indi iduals wi h le -wing a i udes o o e in en ions. This leads o he
ollowing wo hypo heses:4
Mig a ion Hypo hesis (H2b). Blame a ibu ions agains e ugees posi i ely a ec o e choice
only o indi iduals who sha e igh -wing populis o e in en ions o a i udes.
Economic-Eli e Hypo hesis (H2c). Blame a ibu ion agains economic eli es posi i ely a ec s
o e choice only o indi iduals who sha e le -wing populis o e in en ions o a i udes.
Populis s yle
A long adi ion in esea ch on populism has poin ed ou ha a decisi e cha ac e is ic o populis
ac o s is communica ing simply (Aalbe g e al., 2016; Cano an, 1999; Mudde, 2004). This
communica ion s yle is mo eo e en o ced h ough a populis wo ld iew and poli ical s a egies,
o , ‘populism’s endency no only o communica e in a simple and di ec manne bu also o
o e solu ions ha a e di ec and simple’ (Mo i & To mey, 2014, p. 387). Populism o en
p o ides a simple and a ac i e al e na i e o he mo e complex and con adic o y ideas o libe al
democ acies (Mudde, 2007). As such, populis ac o s apply a simple iew o he wo ld, which is
o en based on in- and ou -g oup hinking, esul ing in ‘us’ e sus ‘ hem’ he o ical s a egies.
Addi ionally, hey posi ion hemsel es as he g ounded al e na i e o he cu en poli ical o
economic es ablishmen . Thus, people should pe cei e populis s as he simple and maybe e en
c ude, bu hones , al e na i e o he uling eli es ha common people pe cei e as being oo
p e en ious. Populis s’ simple communica ion s yle signals hese poli ical s a egies and ideas.
Hamelee s e al. (2016) unde line ha ‘ he communica ion s yle o populis s dis inguishes i sel by
using simple and di ec language, one-line s, and p o oca i e s a emen s di ec ed owa d he eli e
and o he ou -g oups’ (Hamelee s e al., 2016, p. 147).
Nex o quali a i e e idence, schola s ha e also a emp ed o measu e he language complexi y
o populis ac o s quan i a i ely. Bischo and Senninge (2018) show ha mani es os o populis
pa ies use simple language han non-populis pa ies. Decad i and Boussalis (2020) e en p o ide
causal e idence and emphasise ha MPs who change pa ies adap hei language complexi y
depending on whe he hey a e pa o a populis pa y o no . MPs o populis pa ies use
signi ican ly simple language han MPs o non-populis pa ies. In he U.S. con ex , Oli e and
Rahn (2016) ha e shown ha Donald T ump used simple language han his compe i o s in he
2016 p esiden ial elec ion. O he empi ical esul s, howe e , e eal a mo e mixed pic u e. By
analysing he speeches o poli ical leade s, McDonnell and Ondelli (2020) show, o example,
ha populis leade s some imes e en use mo e complex language han non-populis leade s.
Ne e heless, gi en he idea ha simple language leads o poli ics ha a e easie o
comp ehend and ha o e s would a he o e o someone whom hey can unde s and as hey can
mo e easily hold ha poli ician o accoun (Bischo & Senninge , 2018), I will s ill conside simple
language as ha ing a gene ally posi i e e ec on all o e s, and o e s wi h populis a i udes in
© 2024 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.
LET’S TALK POPULIST? 725
pa icula . The heo e ical conside a ions abou populis beha iou en o ce hese assump ions e en
hough he esul s ega ding hei usage a e con adic o y. This leads o he inal hypo hesis.5
Simple-Language Hypo hesis (H3). Simple language posi i ely a ec s o e choice in gene al,
bu i has an e en g ea e e ec on people wi h populis o e in en ions o a i udes.
Ge many as a case and sample s a egy
The su ey expe imen was implemen ed in Ge many, which can be conside ed an ideal case
because o he ecen success o populis ac o s in his coun y. Cu en ly, bo h a le -wing and a
igh -wing populis pa y a e ep esen ed in he Ge man Bundes ag: Die Linke (The Le )6, which
is conside ed a le -wing populis pa y, and Al e na i e o Ge many (A D), which is a igh -wing
populis pa y (Rooduijn, 2019). Thus, I s udied indi iduals who suppo populis pa ies wi h
opposing ideological posi ions in a compa a i e se ing. Addi ionally, he Ge man pa y sys em
co e s a b oad ange o pa ies’ ideological posi ions, including he ex eme posi ions co e ed by
he Le and he A D, bu also, in gene al, mo e mode a e ideological posi ions like he Ch is ian
Democ a s (CDU/CSU) o Social Democ a s (SPD).
In addi ion, he mixed elec o al sys em in Ge many has allowed me o ame he candida e
choice expe imen in he o m o candida e o ing decisions (see Neune & W a il, 2022). In
Ge man na ional elec ions, ci izens cas wo o es. The i s o e is o a di ec candida e. Th ough
he so-called E s s imme, o e s di ec ly selec a candida e in hei elec o al dis ic ollowing
a i s -pas - he-pos o mula. The second o e is o a pa y and is commonly e e ed o as
Zwei s imme. This su ey expe imen models he o e o he E s s imme and c ea es a somewha
ealis ic scena io.
O he ea u es ha we should conside when looking a Ge many as a case a e possible egional
di e ences as well as o e cha ac e is ics. Ge many’s his o y may d i e egional di e ences as he
coun y was di ided in o Eas Ge many (Ge man Democ a ic Republic – GDR) and Wes Ge many
(Fede al Republic o Ge many – FDR) un il 1990 wi h wo di e en poli ical sys ems. Elec o ally,
populis pa ies a e cu en ly mo e asse i e in Eas Ge many han in Wes Ge many (Loew & Faas,
2019). Signi ican ly, he A D became e y success ul in Eas e n Ge man ede al s a es like Saxony.
When looking a o e cha ac e is ics, we can obse e ha ideological p oximi y, as well as
populis a i udes, cha ac e ise o e s o bo h pa ies. Hansen and Olsen (2022) a gue ha an i-
immig an ideology is a s onge p edic o o A D o e choice han socio-demog aphic a iables.
They also ind ha nega i e a i udes owa ds poli ical eli es a e inc easing he p obabili y o
o ing o he A D. Simila esul s a e ound o o e s o he Le . Goodge (2023) inds ha
also policy p oximi y o le -leaning policies p edic s o e choice o he Le . These esul s a e
in line wi h o he s udies when analysing suppo o le -wing and igh -wing populis pa ies in
Eu ope (Hauwae & Van Kessel, 2018; Reinemann e al., 2019). As such, he esul s p esen ed in
he analysis a e no only ele an o he Ge man case bu can also a el beyond ha and may be
especially applicable o o he coun ies in Eu ope wi h simila poli ical sys ems.
In o al, 3325 esponden s we e su eyed. The sample is d awn om a la ge online panel
p o ided by he su ey company Lucid. Lucid p o ides access o a pool o people ac oss Ge many
wi h a ying ages and educa ional backg ounds, ensu ing a somewha ep esen a i e sample o
he Ge man popula ion. The mean age o he inal sample is 45.3, wi h 51.7 pe cen emale
and 48.3 pe cen male esponden s. The su ey was conduc ed be ween mid-Decembe 2020 and
© 2024 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.
726 REBECCA C. KITTEL
mid-Janua y 2021. I also asked esponden s abou essen ial cha ac e is ics. As such, I was able o
con ol o egional di e ences be ween Eas and Wes Ge many as well as educa ional a ia ion.
Resea ch design: Pai ed igne e expe imen
To shed ligh on he e ec o populis and non-populis he o ical s a egies and which elemen s
a e he d i ing ac o s, I conduc ed a pai ed igne e expe imen in Ge many be ween Decembe
2020 and Janua y 2021 wi h a sample size o 3325 esponden s. This ga e me 9975 igne e
a ings, as each esponden had o answe h ee choice asks. Pai ed igne e designs ollow he
same logic as pai ed conjoin expe imen s and allow esea che s o es ima e he causal e ec s o
mul iple ea men componen s (Hainmuelle e al., 2014). Compa ed o conjoin expe imen s,
igne e designs a e p esen ed in a ex o m ins ead o a able o ma (Hainmuelle e al.,
2015). The e o e, his design seemed mos sui able o es he e ec o poli ical s a emen s wi h
a ying populis and non-populis he o ical s a egies. Responden s ead wo sho pa ag aphs
wi h a ying a ibu e le els in his design speci ica ion. The pai ed igne e design is pa icula ly
bene icial o answe ing he s a ed esea ch ques ion o how elemen s o (non-)populis discou se
causally a ec o ing beha iou . I allows he ea men s o a y ac oss di e en discu si e
elemen s. Thus, combina ions o he o ical and s ylis ic ea u es in poli ical s a emen s can be
modelled and hei causal e ec es ed. A ansla ed English e sion o he ques ionnai e is
p esen ed in online Appendix C.7Below, I will desc ibe he dependen and independen a iables
and possible mode a o s.
Dependen a iable: Vo e choice
In his newly designed su ey expe imen , esponden s we e p esen ed wi h h ee pai s o ic ional
candida es’ s a emen s ha use a ying populis he o ical s a egies and language complexi ies.
These pai s we e andomly d awn om a pool o 16 di e en ly amed s a emen s. Then,
esponden s we e asked o s a e o which candida e’s s a emen om he pai hey would a he
cas hei o e. The e o e, esponden s’ o e p e e ence is he dependen a iable ollowing he
logical s uc u e o conjoin / igne e expe imen s.
Responden s we e also asked o s a e how much hey could imagine o ing o each candida e
in he pai . Thus, esponden s could s a e whe he i is e y unlikely ha hey would o e o a
candida e (1) o e y likely ha hey would o e o a candida e (7).8The ask was epea ed h ee
imes o each esponden . To minimise associa ions o exis ing candida es as well as hei pa ies
and o include s a emen s ha migh no occu in eali y, I s udy o e choice o candida es a he
han pa ies and do no p o ide pa y labels (see Neune & W a il, 2022).
Independen a iable and ea men : S a emen a ia ions
To measu e he e ec o he h ee elemen s o populis discou se (people-cen ic he o ic, blame
a ibu i e language and populis s yle), di e en ic ional s a emen s o poli icians a e amed.
All s a emen s a e ocused on he wo sening o Ge many’s cu en labou ma ke si ua ion. This
issue is well sui ed o analyse a ious discu si e elemen s on he e ec o o e choice. Fi s ,
he labou ma ke is a less poli icised issue by adical le o adical igh ac o s in compa ison
o opics like immig a ion o c ime. In ac , poli ical ac o s om bo h sides o he ideological
© 2024 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.
LET’S TALK POPULIST? 733
Figu e 4. Ma ginal means by pa y g oups.
No e: Ma ginal means a e displayed wi h 95 pe cen con idence in e als.
and no s a is ically signi ican . As such, he esul s hold ac oss egions and a e no d i en by
geog aphical pa e ns.
Populis o e s e sus non-populis o e s
Now, we will u n o he in e ac ion e ec be ween di e en s a emen elemen s and people’s
ideological p e e ences, namely hei o e in en ions. The e o e, depending on people’s o e
in en ion, I will compa e which s a emen combina ions inc ease candida es’ o e sha e. As such,
we can analyse i populis o e s (A D and Linke o e s) a e likelie o o e o candida es ha
employ mo e populis and ideological adical he o ical elemen s.
Figu e 4shows he ull sample di ided by indi iduals’ o e in en ions on he su ey day. Mos
in e es ingly, he esul s show ha blame a ibu ion seems o be he mos decisi e d i e o o e
choice. The esul s show ha blame agains e ugees in a poli ical message posi i ely a ec s igh -
wing populis o e s in Ge many (A D o e s). Fo all o he o e s, i has a nega i e e ec . Simila
esul s can be obse ed o poli ical messages ha blame economic eli es. They d i e he o e
© 2024 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.
734 REBECCA C. KITTEL
Figu e 5. In e ac ions o a ibu e le els wi h s a emen numbe s (only showing people-cen ic s a emen s): ma ginal
means o SPD and CDU/CSU o e s.
No e: Ma ginal means a e displayed wi h 95 pe cen con idence in e als.
choice o le -wing populis o e s (Linke o e s) bu ha e a ze o o nega i e e ec on all o he
o e s. This shows con i ma ion o H2b and H2c. Blame agains poli icians has a su p isingly
posi i e e ec o all o e g oups. Thus, o e s gene ally espond posi i ely o poli ical messages
ha blame poli icians. This con adic s hypo hesis H2a. No blame a ibu ion, howe e , sco es
highes among non-populis pa y o e s. Also, he esul s show ha people-cen ism and a e age
language complexi y posi i ely a ec all pa y o e s on a e age. This suppo s he people-
cen ism hypo hesis (H1) and e u es he simple-language hypo hesis (H3).
Finally, we will look a message combina ions, indi iduals’ unde lying a i udes, o o e
in en ions. This p oduces a close, ealis ic assessmen o how unde lying ideological a i udes
a ec how people espond o poli ical messages. The esul s o message combina ions, including
people-cen ic he o ic, a e p esen ed in Figu es 5–7. Figu es including all s a emen combina ions
(wi h and wi hou people-cen ic he o ic) a e p esen ed in online Appendix F(Figu es 2–4). The
esul s a e spli in o h ee igu es ( i e g aphs) due o be e g aphical ep esen a ion. The mos
in e es ing esul s a e ha blame a ibu ion agains e ugees again has a posi i e e ec only o
people who in end o o e o he A D ( igh -wing populis pa y). Howe e , in combina ion wi h
people-cen ism and simple language, his poli ical s a emen has he highes posi i e e ec on
people who in end o o e o he A D. This shows ha i is no an indi idual componen o a
poli ical s a emen bu a he he combina ion o all h ee populis discu si e elemen s ha make
people wi h co esponding a i udes u n owa ds hese pa ies. This is suppo ed e en u he when
we look a s a emen s ha do no include people-cen ic he o ic. Fo hese s a emen combina ions
© 2024 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.
LET’S TALK POPULIST? 735
Figu e 6. In e ac ions o a ibu e le els wi h s a emen numbe (only showing people-cen ic s a emen ): Ma ginal
means o Linke and B90/GRUENE o e s.
No e: Ma ginal means a e displayed wi h 95 pe cen con idence in e als.
(see online Appendix F, Figu e 4), A D-o ien ed people ha e a lowe chance o cas ing hei o e
o ha candida e and become e en nega i e o he combina ion o simple language, blame agains
e ugees, and lack o people-cen ic he o ic. This shows ha his pa icula combina ion seems
o a ec o e choice. As such, i is impo an o conside a ious aspec s and hei e ec when
discussing populis discou se.
The esul s o he h ee igu es also show ha he combina ion o a e age language complexi y,
blame agains poli icians and people-cen ism p oduces a highly posi i e e ec on candida e o e
choice among all o e g oups. These esul s a e a he su p ising, as hey con adic he poli ician
hypo hesis (H2a) ha assumes blaming poli icians will only posi i ely a ec o e s wi h populis
a i udes o o e in en ions. Bu also o SPD and CDU/CSU o e s, blame a ibu ion sco es
posi i ely in many s a emen combina ions (see Figu e 5). In addi ion, hey show ha indi iduals
who do no cas a populis o e espond posi i ely o elemen s o populis discou se. People, on
a e age, seem o blame poli icians o socie al p oblems. Blaming economic eli es o e ugees,
on he o he hand, depends on he co esponding unde lying a i udes o he indi idual o , mo e
p ecisely, hei o e in en ion.
These esul s a e also con i med when we look a populis a i udes ins ead o o e in en ions.
As a obus ness check, I c ea ed hin and hick populis a i udes scales ollowing Neune and
W a il (2022). Online Appendix Gshows esul s o people displaying le -wing populis a i udes,
igh -wing populis a i udes o cen is no a i udes (online Appendix G, Figu es 9and 10).
The esul s show ha he p e e ed blame a ibu ion is highly connec ed o people’s ideological
© 2024 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.
736 REBECCA C. KITTEL
Figu e 7. In e ac ions o a ibu e le els wi h s a emen numbe s (only people-cen ic s a emen s): Ma ginal means
o A D o e s.
No e: Ma ginal means a e displayed wi h 95 pe cen con idence in e als.
p e e ences. Fu he , he esul s con i m ha simple language is usually less p e e ed han a e age
language complexi y in all s a emen combina ions. When looking a he di ision be ween populis
and non-populis people wi hou con olling o any ideological p e e ences, we can obse e
ha no blame o blame owa ds poli icians has a posi i e e ec on ei he g oup, e en hough
blame owa ds poli icians sco es highe among people wi h populis a i udes (online Appendix G,
Figu es 9and 10).
In addi ion, a sub-sample o N=1748 esponden s was analysed. The sub-sample was c ea ed
h ough an a en ion check ha was asked o esponden s and only included esponden s who
ead all s a emen s ha we e p esen ed o hem. Fo he a en ion check, esponden s we e asked
o s a e o whom each o he six candida es a ibu ed blame. Thus, esponden s we e asked six
imes o indica e whe he blame was a ibu ed o poli ical o economic eli es, e ugees o nobody.
Responden s who answe ed a leas i e o six ques ions co ec ly we e included in he sample.
This measu e conside s ha esponden s could ha e had one slip o he pen bu , in gene al, ead
all o he s a emen s.13 Resul s o he sub-sample a e nea ly equi alen o he ull sample esul s.
Discussion and conclusion
Populis messages consis o a ious discu si e elemen s ha make hem success ul among a ious
o e g oups. Howe e , no e e y elemen is equally success ul o e e y pe son. This depends e y
much on people’s unde lying a i udes and o e in en ions. While p e ious obse a ional wo k
© 2024 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.
LET’S TALK POPULIST? 737
could no de ec causal claims on he e ec o discu si e s a egies on o e choice, his a icle
con ibu es o his gap by showing how di e en o e g oups a e a ac ed o di e en discu si e
s a egies. The analysis gi es new insigh s in o how pa icula discu si e elemen s causally a ec
o e choice and, mo e p ecisely, which populis elemen s also a ec o e s who ul ima ely would
no in end o o e o a populis pa y. Thus, his expe imen showed he ollowing ou esul s.
Fi s , i showed ha people-cen ic he o ic wo ks among all o e g oups; i does no solely
appeal o populis o e s. This may be due o he pleasan eeling o being pa o a g oup.
People-cen ic he o ic inco po a es people who usually like o belong somewhe e. These esul s
also show ha , o e all, people p e e a mo e inclusi e language in compa ison o echnoc a ic
communica ion. As such, people seem o be mo e a ac ed by a populis discou se which includes
hem as ac o s compa ed o echnoc a ic discou se which employs a op-down app oach (Ca amani,
2017). As hese esul s hold o all o e g oups, hey also p o ide impo an implica ions o non-
populis and mains eam ac o s. We can imply ha o e all, people seem o p e e a bo om-up
app oach in which hey a e included in he discou se ins ead o a op-down app oach in which
poli icians claim o be ul ima e expe s who ha e comple e answe s o p oblems. The e o e,
non-populis ac o s can bene i om using a mo e inclusi e and people-cen ic he o ic in hei
communica ion wi h o e s.
Second, he esul s show ha he ‘neu al’ e sion o blame a ibu ion – namely owa ds
poli icians – inc eases o e choice among all o e g oups. These a e a he su p ising esul s, as
blame-a ibu i e language is conside ed an elemen o populis discou se. E en hough no blame
a ibu ion in candida es’ s a emen s usually gene a es he highes likelihood o a candida e being
elec ed by non-populis o e s, hese esul s show ha populis dynamics and elemen s o hei
he o ical s a egies a e e ec i e o a la ge g oup o o e s. People seem o be ‘discon en , no
only wi h poli ics bu also wi h socie al li e in gene al’ (Sp uy e al., 2016, p. 335). Thus, ce ain
populis discu si e elemen s a ec e en non-populis o e s mo e quickly han expec ed. Howe e ,
mo e esea ch is necessa y o e eal i non-populis ac o s a e awa e o hese mechanisms, and use
hem, o i a gene al inc ease in dissa is ac ion and dis us in poli icians has led o hese esul s.
Thi d, o e s’ ideological p e e ences ha e a highe e ec on o e in en ions han unde lying
populis a i udes. These dynamics a e o en unde s udied in he discussion in esea ch on
populism. Ne e heless, his expe imen has shown ha unde lying ideological a ibu es ha e
an e ec i people espond o a ge ed blame a ibu ion and de ine hei o e choice mo e han
jus hei gene al populis a i udes. This may also open up discussions on whe he populism
and ideology om he le and he igh should be mo e o en discussed as co-exis ing ins ead
o neglec ing one o he o he . Populis he o ic is, a e all, e y much p imed by i s hos ideology.
Fou h, he pape showed ha agains all expec a ions, simple language has a nega i e e ec on
o e choice in nea ly all poli ical s a emen combina ions and o all o e g oups. These esul s
con adic he expec ed esul bu align wi h simila esul s ound by In Bischo and Senninge
(2021). They hin a he ac ha people may p e e poli icians who speak wi h a e age language
complexi y, as hey assign hem mo e compe ence. Howe e , his can also be speci ic o he Ge man
case and Ge man language peculia i ies.
As i is no only ele an o unde s and wha e ec ce ain discu si e elemen s ha e on o e
choice bu also how long-las ing and pe suasi e hey may be, we can ela e he esul s o esea ch
on pe suasion e ec s in poli ical communica ion. Resea ch has shown ha pe suasion s a egies
in campaigns o en ha e s ong bu sho -li ed e ec s on o ing beha iou (Ge be e al., 2011).
In e p e ing hese indings in ela ion o his su ey expe imen leads us o assume ha (populis )
© 2024 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.
738 REBECCA C. KITTEL
discu si e s a egies may a ec o e s in he sho un and can in luence elec ion esul s. Discu si e
s a egies may no change o e s’ unde lying ideological p e e ences in he long un, hough. I
we conside ha blame a ibu ion owa ds e ugees o capi alis s only posi i ely a ec ed o e s
ha al eady displayed unde lying igh -wing o le -wing ideological a i udes, espec i ely, we
can assume ha o he ac o s may a ec ideological change o o e s. In a b oade scope, hese
esul s imply ha unde lying ideological p e e ences p ima ily d i e o e choice. As such, populis
ac o s may no con ince people who do no al eady sha e hei ideology o o e o hem by jus
applying a blame-a ibu i e discou se. This shows ha language does a ec people. Howe e ,
ins ead o changing a i udes quickly, language ins ead ein o ces hem. Ne e heless, hese a e
only in e p e a ions based on ea lie esea ch esul s. Mo e esea ch es ing he long-las ing e ec s
on ce ain discu si e elemen s would be necessa y o ecei e binding esul s.
E en hough he su ey expe imen was only conduc ed as a single case s udy in Ge many and
he Ge man language displays unique linguis ic cha ac e is ics, esul s can a el beyond ha case.
As discussed ea lie , schola s ha e shown ha simila c oss-coun y expe imen s es ing o ing
beha iou and populis a i udes a e o en simila ac oss Wes e n Eu opean coun ies. Gi en ha
pa y sys ems ac oss Eu ope, especially Wes e n Eu ope, a e ela i ely simila , he e is li le eason
o belie e hese esul s do no hold beyond Ge many.
Analysing di e en discu si e elemen s o populis and non-populis discou se and hei e ec s
on a ious o e g oups has shown ha language can be used s a egically o a ec o e choice.
I has also shown ha peoples’ unde lying a i udes g ea ly a ec how a message is pe cei ed. A
gene al dissa is ac ion wi h he unc ioning o he poli ical sys ems – o a leas how poli icians
wo k – seems o lead o a gene al accep ance o blaming poli ical eli es o socie al p oblems.
Eme ging esea ch shows, oo, ha all kinds o pa ies use an i-eli e he o ic s a egically, o en
depending on hei elec o al ci cums ances (Lich e al., 2024). Fu u e wo k should ocus u he
on he in e play o pa ies’ s a egic use o an i-eli e he o ic and how his a ec s o ing beha iou .
This will help o disen angle he challenges ad anced democ acies may ace in he u u e.
Acknowledgemen s
I hank Geo Allen, Ma iana Ca mo Dua e, B uno Cas anho Sil a, Amy Ca alinac, Denis Cohen,
S en-Oli e P oksch, Simon Hix, Ellen Imme gu , Alexand u Moise, Simon Munze , Cos in
Ciobanu, Kons an in Vössing and Ch is ophe W a il as well as he h ee anonymous e iewe s
o p o iding cons uc i e eedback and help ul commen s on ea lie e sions o he manusc ip
as well as he esea ch design. Also, I would like o hank pa icipan s a he MPSA Con e ence
2023, Düsseldo Lunch Talk 2022, Poli ical Beha iou Colloquium EUI 2022, ECPR Gene al
Con e ence 2021 and he CIVICA PhD Panel 2020 o hei commen s and engaging discussions.
The su ey expe imen was unded by he Ea ly S age Resea che G an om he Eu opean
Uni e si y Ins i u e, Villa San elice, 50014 San Domenico di Fiesole, I aly.
Open access unding enabled and o ganized by P ojek DEAL.
E hics s a emen
The su ey expe imen ecei ed e hical app o al om he E hics Commi ee o he Eu opean
Uni e si y Ins i u e on 14 Decembe 2020. I was also p e- egis e ed on EGAP and can be accessed
h ough his link h ps://os .io/7u sx.
© 2024 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.
LET’S TALK POPULIST? 739
Con lic o in e es s a emen
In submi ing my manusc ip , I ha e ca e ully ollowed he ins uc ions o he au ho s. The e o e, I
ha e no con lic s o in e es o decla e, and he manusc ip is cu en ly no unde e iew elsewhe e.
Da a a ailabili y s a emen
Da a om he su ey expe imen and he ep oducible code o he analysis a e a ailable on OSF
Regis ies h ough he ollowing link: h ps://doi.o g/10.17605/OSF.IO/G7V6S.
Online Appendix
Addi ional suppo ing in o ma ion may be ound in he Online Appendix sec ion a he end o he
a icle:
Appendix A De ia ions om he P e-Analysis Plan
Appendix B Real Wo ld Examples - Populis S a emen s
Appendix C Full Ques ionnai e - English Ve sion
Appendix D T ea men Combina ion o Candida e S a emen s - Ge man Ve sion wi h Sc eensho
om he o iginal Expe imen
Figu e 1: Ge man candida e choice ques ion (dependen a iable) as p esen ed o he esponden
in he su ey, S a emen 6 (le -hand side) and S a emen 11 ( igh -hand side) in Appendix C.1.
Appendix E Readabili y Sco es
Table 1: Readabili y Sco es Acco ding o T ea men -Vigne e
Table 2: Readabili y Sco es Acco ding o T ea men -Vigne e
Appendix F Full Resul s o In e ac ion Le els
Figu e 2: In e ac ions o A ibu e Le els: Ma ginal Means by Pa ies, SPD and CDU/CSU
in Compa ison
Figu e 3: In e ac ions o A ibu e Le els: Ma ginal Means by Pa ies, Linke and B90/GRUENE
in Compa ison
Figu e 4: In e ac ions o A ibu e Le els: Ma ginal Means by Pa ies, A D
Appendix G Robus ness Checks
Figu e 5: Ma ginal Means o Ci izens Li ing in he Fo me GDR s. Ci izens Li ing in he
Fo me FDR
Figu e 6: In e ac ion o A ibu e Le els: Ma ginal Means o Ci izens Li ing in he Fo me GDR
s. Ci izens Li ing in he Fo me FDR
Figu e 7: In e ac ions o A ibu e Le els (only S a emen s wi h People-Cen ic Rhe o ic):
Ma ginal Means o Populis s and Non-Populis s
Figu e 8: In e ac ions o A ibu e Le els (only S a emen s Wi hou People-Cen ic Rhe o ic):
Ma ginal Means o Populis s and Non-Populis s
Figu e 9: In e ac ions o A ibu e Le els (only S a emen s wi h People-Cen ic Rhe o ic):
Ma ginal Means o Le -Wing Populis s, Righ -Wing Populis s and NonPopulis s
Figu e 10: In e ac ions o A ibu e Le els (only S a emen s wi hou People-Cen ic Rhe o ic):
Ma ginal Means o Le -Wing Populis s, Righ -Wing Populis s and NonPopulis s
Figu e 11: Ma ginal Means by Ideology G ouping (wi hou con olling o populism)
© 2024 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.
740 REBECCA C. KITTEL
Figu e 12: Ma ginal Means by Populis s and Non-Populis s G oups
Figu e 13: Ma ginal Means o Populis s and Non-Populis s
Appendix H Resul s: Explo a o y Hypo heses
Figu e 14: Ma ginal Means o Emo ions s. No Emo ions
Figu e 15: Ma ginal Means by Emo ional Feelings
Figu e 16: Ma ginal Means by Educa ion G oups
No es
1. The s udy was p e- egis e ed wi h EGAP and can be accessed h ough he ollowing link h ps://os .io/
7u sx. Fu he , I ecei ed e hical app o al om he Eu opean Uni e si y Ins i u e E hics Commi ee on 14
Decembe 2020.
2. The hypo hesis is sligh ly e o mula ed han i was ini ially w i en down in he p e-analysis plan, bu he
meaning has emained he same. Fo mo e de ails, see online Appendix A.
3. This hypo hesis was no pa o he p e-analysis plan no we e H2b and H2c. They a e, howe e , closely ela ed
o wo k by Bos e al. (2020) and Hamelee s e al. (2017).
4. As emphasised in H1a, hese wo hypo heses ha e no been pa o he o iginal p e-analysis plan.
5. This hypo hesis combines hypo heses H1 and H2 o he p e-analysis plan.
6. Die Linke has ecen ly expe ienced some le el o splin e ing in he Ge man Bundes ag. Since Oc obe 2023, a
discussion abou a new splin e pa y eme ged, and in Janua y 2024, Sah a Wagenknech Alliance – Reason and
Jus ice o Alliance (Ge man: Bündnis Sah a Wagenknech – Ve nun und Ge ech igkei (BSW)) was ounded
as he splin e pa y o Die Linke. Du ing he pe iod o he su ey expe imen , Die Linke was s ill a e y s able
pa y and uni ied in he Ge man Bundes ag, so I do no expec ha his has any e ec on he esul s.
7. The Ge man e sion o he ea men (s a emen s) is shown in online Appendix D.
8. So a , his measu e is no included in he analysis.
9. I e e o e ugees ins ead o immig an s in he ea men o wo main easons. Fi s , in 2015, Ge many
expe ienced he so-called e ugee c isis, making e ugees a salien ou -g oup. Second, he A D is mos ly
blaming e ugees ins ead o immig an s. The e o e, e ugees a e he ou -g oup ha is blamed o he mos
in he Ge man case by he igh -wing populis ac o . As such, he e e ence o e ugees in he ea men seemed
mos app op ia e.
10. The o iginal Ge man e sion o he s a emen is p esen ed in online Appendix B. Fu he , a quo e om he
same deba e opic o he le -wing populis pa y The Le is also p esen ed in online Appendix B.
11. Resul s based on demog aphic cha ac e is ics a e p esen ed in online Appendix E.
12. I use he c egg package in R (Leepe , 2020). I calcula es MMs as su ey-weigh ed means wi h app op ia e
s anda d e o s.
13. Ano he echnique would be o only include all candida e pai s ha esponden s co ec ly ma ked. This would,
howe e , lead o an une en numbe o esponden obse a ions.
Re e ences
Aalbe g, T., Esse , F., Reinemann, C., S omback, J., & V eese, C. D. (2016). Populis Poli ical Communica ion in
Eu ope. Taylo & F ancis L d.
Akke man, A., Mudde, C., & Zaslo e, A. (2014). How populis a e he people? measu ing populis a i udes in
o e s. Compa a i e Poli ical S udies,47(9), 1324–1353.
Albe azzi, D., & McDonnell, D. (Eds.). (2008). Twen y- i s cen u y populism: The spec e o Wes e n Eu opean
democ acy. Palg a e Macmillan UK.
Albe azzi, D., & Muelle , S. (2013). Populism and libe al democ acy: Populis s in go e nmen in Aus ia, I aly,
Poland and Swi ze land. Go e nmen and Opposi ion,48(3), 343–371.
Ande son, J. (1981). Analysing he eadabili y o English and non-English ex s in he class oom wi h Lix. Pape
p esen ed a he Annual Mee ing o he Aus alian Reading Associa ion, Da win, Aus alia, Augus 1981.
© 2024 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.
LET’S TALK POPULIST? 741
Aslanidis, P. (2018). Measu ing populis discou se wi h seman ic ex analysis: An applica ion on g ass oo s populis
mobiliza ion. Quali y & Quan i y,52(3), 1241–1263.
Bischo , D., & Senninge , R. (2018). Simple poli ics o he people? Complexi y in campaign messages and poli ical
knowledge. Eu opean Jou nal o Poli ical Resea ch,57(2), 473–495.
Bischo , D., & Senninge , R. (2021). Simpli ying poli ics: The impac o linguis ic s yle on ci izens’ poli ical
knowledge and belie s abou poli icians. OSF P ep in s., h ps://doi.o g/10.31219/os .io/zchpx
Bjö nsson, C.-H. (1968). Läsba he .Libe .
Bos, L., Scheme , C., Co bu, N., Hamelee s, M., And eadis, I., Schulz, A., Schmuck, D., Reinemann, C., & Fawzi,
N. (2020). The e ec s o populism as a social iden i y ame on pe suasion and mobilisa ion: E idence om a
15-coun y expe imen . Eu opean Jou nal o Poli ical Resea ch,59(1), 3–24.
Bos, L., Shee s, P., & Boomgaa den, H. G. (2018). The ole o implici a i udes in populis adical- igh suppo .
Poli ical Psychology,39(1), 69–87.
Bos, L., an de B ug, W., & de V eese, C. H. (2013). An expe imen al es o he impac o s yle and he o ic on
he pe cep ion o igh -wing populis and mains eam pa y leade s. Ac a Poli ica,48(2), 192–208.
Busby, E. C., Guble , J. R., & Hawkins, K. A. (2019). F aming and blame a ibu ion in populis he o ic. The
Jou nal o Poli ics,81(2), 616–630.
Cano an, M. (1999). T us he people! Populism and he wo aces o democ acy. Poli ical S udies,47(1), 2–16.
Ca amani, D. (2017). Will s. eason: The populis and echnoc a ic o ms o poli ical ep esen a ion and hei
c i ique o pa y go e nmen . Ame ican Poli ical Science Re iew,111(1), 54–67.
Cas anho Sil a, B., And eadis, I., Anduiza, E., Blanusa, N., Co i, Y. M., Del ino, G., Rico, G., Ru h, S., Sp uy , B.,
S eenbe gen, M. R., & Li ay, L. (2018). Public opinion su eys: A new scale. In K. A. Hawkins, R. E. Ca lin,
L. Li ay, & C. R. Kal wasse (Eds.), The idea ional app oach o populism: Concep , heo y, and analysis (pp.
150–178). Rou ledge.
Cas anho Sil a, B., Jungkunz, S., Helbling, M., & Li ay, L. (2020). An empi ical compa ison o se en populis
a i udes scales. Poli ical Resea ch Qua e ly,73(2), 409–424.
Cas anho Sil a, B., & W a il, C. (2023). Populism and candida e suppo in he US: The e ec s o ‘Thin’ and ‘Hos ’
Ideology. Jou nal o Expe imen al Poli ical Science,10(3), 438–447.
Chong, D., & D uckman, J. N. (2007). F aming heo y. Annual Re iew o Poli ical Science,10(1), 103–126.
Dai, Y., & Kus o , A. (2022). When do poli icians use populis he o ic? Populism as a campaign gamble. Poli ical
Communica ion,39(3), 383–404.
Decad i, S., & Boussalis, C. (2020). Populism, pa y membe ship, and language complexi y in he I alian Chambe
o Depu ies. Jou nal o Elec ions, Public Opinion and Pa ies,30(4), 484–503.
Deu sche Bundes ag (2017). Plena p o okoll 19/3-22. No embe 2017. h ps://dse e .bundes ag.de/b p/19/19003.
pd .
Elcha dus, M., & Sp uy , B. (2016). Populism, pe sis en epublicanism and declinism: an empi ical analysis o
populism as a hin ideology. Go e nmen and Opposi ion,51(1), 111–133.
Ellge , F. (2024). The mobilizing e ec o pa y sys em pola iza ion. E idence F om Eu ope. Compa a i e Poli ical
S udies,57(8), 1310–1338.
Ge be , A. S., Gimpel, J. G., G een, D. P., & Shaw, D. R. (2011). How la ge and long-las ing a e he pe suasi e
e ec s o ele ised campaign ads? Resul s om a andomized ield expe imen . Ame ican Poli ical Science
Re iew,105(1), 135–150.
Goodge , E. (2023). Policy and populism: Analysing suppo o Die Linke. Poli ical S udies Re iew,0(0), 1–15.
Hainmuelle , J., Hanga ne , D., & Yamamo o, T. (2015). Valida ing igne e and conjoin su ey expe imen s
agains eal-wo ld beha io . P oceedings o he Na ional Academy o Sciences,112(8), 2395–2400.
Hainmuelle , J., Hopkins, D. J., & Yamamo o, T. (2014). Causal in e ence in conjoin analysis: Unde s anding
mul idimensional choices ia s a ed p e e ence expe imen s. Poli ical Analysis,22(1), 1–30.
Hamelee s, M., Bos, L., & de V eese, C. (2018). F aming blame: Towa d a be e unde s anding o he e ec s o
populis communica ion on populis pa y p e e ences. Jou nal o Elec ions, Public Opinion and Pa ies,28(3),
380–398.
Hamelee s, M., Bos, L., & de V eese, C. H. (2016). The Ne he lands: A hea land ull o insigh s in o populis
communica ion. In T. Aalbe g, F. Esse , C. Reinemann, J. S omback, & C. H. de V eese (Eds.), Populis
poli ical communica ion in Eu ope (pp. 138–150). Rou ledge.
© 2024 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.
742 REBECCA C. KITTEL
Hamelee s, M., Bos, L., & de V eese, C. H. (2017). “They did i ”: The e ec s o emo ionalized blame a ibu ion in
populis communica ion. Communica ion Resea ch,44(6), 870–900.
Hansen, M. A., & Olsen, J. (2022). The al e na i e o Ge many (A D) as populis issue en ep eneu : Explaining
he pa y and i s o e s in he 2021 Ge man Fede al Elec ion. Ge man Poli ics. Ad ance online publica ion.
h ps://doi.o g/10.1080/09644008.2022.2087871
Hauwae , S. M. V., & Van Kessel, S. (2018). Beyond p o es and discon en : A c oss-na ional analysis o he e ec
o populis a i udes and issue posi ions on populis pa y suppo . Eu opean Jou nal o Poli ical Resea ch,
57(1), 68–92.
Hawkins, K. A., Ca lin, R. E., Li ay, L., & Kal wasse , C. R. (2018). The idea ional app oach o populism:
Concep , heo y, and analysis. Rou ledge.
Hieda, T., Zenkyo, M., & Nishikawa, M. (2021). Do populis s suppo populism? An examina ion h ough an online
su ey ollowing he 2017 Tokyo Me opoli an Assembly elec ion. Pa y Poli ics,27(2), 317–328.
Huddy, L. (2001). F om social o poli ical iden i y: A c i ical examina ion o social iden i y heo y. Poli ical
Psychology,22(1), 127–156.
Hunge , S., & Pax on, F. (2022). Wha ’s in a buzzwo d? A sys ema ic e iew o he s a e o populism esea ch in
poli ical science. Poli ical Science Resea ch and Me hods,10(3), 617–633.
Ki el, R. C. (2023). The new poli ics o pa y signalling: Language, posi ioning, and elec o al dynamics (Doc o al
disse a ion). Eu opean Uni e si y Ins i u e.
K iesi, H., G ande, E., Lacha , R., Dolezal, M., Bo nschie , S., & F ey, T. (2008). Wes Eu opean poli ics in he age
o globaliza ion. Camb idge Uni e si y P ess.
Leepe , T. J. (2020). C egg: Simple conjoin analyses and isualiza ion. h ps://gi hub.com/leepe /c egg
Leepe , T. J., Hobol , S. B., & Tilley, J. (2020). Measu ing subg oup p e e ences in conjoin expe imen s. Poli ical
Analysis,28(2), 207–221.
Lich , H., Abou-Chadi, T., Ba be a, P., & Hua, W. (2024). Measu ing and unde s anding pa ies’ an i-eli e s a egies.
The Jou nal o Poli ics. Accep ed o Publica ion. h ps://doi.o g/10.1086/730711
Loew, N., & Faas, T. (2019). Be ween hin- and hos -ideologies: How populis a i udes in e ac wi h policy
p e e ences in shaping o ing beha iou . Rep esen a ion,55(4), 493–511.
McDonnell, D., & Ondelli, S. (2020). The language o igh -wing populis leade s: No so simple. Pe spec i es on
Poli ics,20(3), 828–841.
Meije s, M. J., & Van De Velden, M. A. (2022). The people as ‘Volk’ o ‘Bü ge ’? The implica ions o e hnic
and ci ic concep ions o he people o he measu emen o populis a i udes. Eu opean Jou nal o Poli ical
Resea ch,62(4), 1403–1413.
Mény, Y., & Su el, Y. (2002). Democ acies and he populis challenge. Palg a e Macmillan.
Mo i , B., & To mey, S. (2014). Re hinking populism: Poli ics, media isa ion and poli ical s yle. Poli ical S udies,
62(2), 381–397.
Mudde, C. (2004). The populis zei geis . Go e nmen and Opposi ion,39(4), 541–563.
Mudde, C. (2007). Populis adical igh pa ies in Eu ope. Camb idge Uni e si y P ess.
Mudde, C., & Kal wasse , C. R. (2012). Populism in Eu ope and he Ame icas: h ea o co ec i e o democ acy?
Camb idge Uni e si y P ess.
Muelle , J.-W. (2016). Wha is populism? Uni e si y o Pennsyl ania P ess.
Neune , F. G., & W a il, C. (2022). The populis ma ke place: Unpacking he ole o “ hin” and “ hick” ideology.
Poli ical Beha io ,44, 551–574.
Oli e , J. E., & Rahn, W. M. (2016). Rise o he T umpen olk: Populism in he 2016 Elec ion. The Annals o he
Ame ican Academy o Poli ical and Social Science,667(1), 189–206.
Os iguy, P. (2017). A socio-cul u al app oach. In C. R. Kal wasse , P. Tagga , P. O. Espejo, & P. Os iguy (Eds.),
The Ox o d handbook o populism (pp. 73–97). Ox o d Uni e si y P ess.
Pe eaul , S., & Bou his, R. Y. (1999). E hnocen ism, social iden i ica ion, and disc imina ion. Pe sonali y and
Social Psychology Bulle in,25(1), 92–103.
Reinemann, C., S anye , J., Aalbe g, T., Esse , F., & de V eese, C. H. (2019). Communica ing populism: Compa ing
ac o pe cep ions, media co e age, and e ec s on ci izens in Eu ope (1 ed.). Rou ledge.
Robe s, K. M. (2006). Populism, poli ical con lic , and g ass- oo s o ganiza ion in La in Ame ica. Compa a i e
Poli ics,38(2), 127–148.
© 2024 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.
