Ambiguity attitudes of individuals and groups in gain and loss domains

Minnich, Aljoscha; Lange, And eas
A icle — Published Ve sion
Ambigui y a i udes o indi iduals and g oups in gain and
loss domains
Theo y and Decision
P o ided in Coope a ion wi h:
Sp inge Na u e
Sugges ed Ci a ion: Minnich, Aljoscha; Lange, And eas (2024) : Ambigui y a i udes o indi iduals and
g oups in gain and loss domains, Theo y and Decision, ISSN 1573-7187, Sp inge US, New Yo k, NY,
Vol. 98, Iss. 3, pp. 373-403,
h ps://doi.o g/10.1007/s11238-024-10009-9
This Ve sion is a ailable a :
h ps://hdl.handle.ne /10419/323472
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/
Vol.:(0123456789)
Theo y and Decision (2025) 98:373–403
h ps://doi.o g/10.1007/s11238-024-10009-9
Ambigui y a i udes o indi iduals andg oups ingain
andloss domains
AljoschaMinnich1· And easLange2
Accep ed: 23 Sep embe 2024 / Published online: 29 No embe 2024
© The Au ho (s) 2024
Abs ac
This s udy measu es he di e ences in ambigui y a i udes o g oups and indi idu-
als in he gain and loss domains. We elici ambigui y a e sion and ambigui y-gen-
e a ed insensi i i y o na u al empe a u e e en s. We do no ind signi ican di -
e ences be ween indi iduals and g oups in ou main sample, ye highe ambigui y
a e sion and ambigui y-gene a ed insensi i i y esul o g oups in he gain domain
when cons aining he sample o g oups and indi iduals wi h a be e unde s and-
ing o he expe imen . The g oup e ec on he ambigui y-gene a ed insensi i i y is
sign-dependen .
Keywo ds Ambigui y a i udes· G oup decision making· Gain and loss domain
1 In oduc ion
Decisions a e o en made unde subs an ial unce ain ies: indi iduals decide on job
oppo uni ies, make in es men s, selec hei pa ne s, o jus make e e yday deci-
sions like selec ing clo hes depending on wea he o ecas s. Many impo an deci-
sions, howe e , a e aken in g oups. Spouses need o decide on educa ional p ospec s
o hei child en, sea ch commi ees collec i ely choose p ospec i e job ma ke can-
dida es, socie y needs o decide on policies, e.g., on clima e policy.
This pape compa es indi idual o g oup decisions on ambiguous na u al e en s
when acing po en ial losses s. gains. Wi h his, we in es iga e o wha ex en
* And eas Lange
and eas.lange@uni-hambu g.de
Aljoscha Minnich
aljoscha.minnic[email p o ec ed]
1 Uni e si y o Hambu g, Depa men o Economics andCen e o Ea h Sys em Resea ch
andSus ainabili y (CEN), G indelbe g 5, 20144Hambu g, Ge many
2 Uni e si y o Hambu g, Depa men o Economics, Von Melle Pa k 5, 20146Hambu g,
Ge many
374
A.Minnich, A.Lange
indi iduals’ and g oups’ ambigui y a i udes a e sign-dependen . Speci ically, we
elici p e e ences on wea he - ela ed e en s, i.e. on empe a u e anges. Conside -
ing ambigui y-a i udes owa ds empe a u e e en s has he ad an age ha mos indi-
iduals a e amilia wi h wea he - ela ed decisions (c . on Gaudecke e al., 2022,
who compa e ambigui y a i udes in ela ion o s ock ma ke s and empe a u e ises;
Minnich e al., 2024).
A la ge and i al li e a u e has deal wi h decision-making unde unce ain y,
o igina ing p ominen ly om Knigh (1921). Besides a la ge heo e ical li e a u e
on ambigui y p e e ences and decisions (see, e.g. E ne e al., 2012; Büh en e al.,
2021), a subs an ial empi ical li e a u e has e ol ed and sugges s po en ial di e -
ences in ambigui y p e e ences in he gain s. he loss domain (e.g., Baillon and
Bleich od , 2015; Koche e al., 2018; T au mann and Van De Kuilen, 2015; Büh en
e al., 2021). Ye , he li e a u e compa ing ambigui y a i udes be ween indi iduals
and g oups in hese wo domains is unde de eloped. I la gely ocuses on hypo-
he ical o e y abs ac decision con ex s (e.g., Ma quis and Rei z, 1969; Agga wal
e al., 2022).
We employ he me hod sugges ed by Baillon e al. (2018) and conside ambigu-
i y a i udes owa ds wea he ( empe a u e) e en s. We place ou wo k in he la ge
li e a u e in Sec .2. In ou expe imen , g oups o h ee can cha wi h each o he
and mus each an unanimous solu ion. Ou esul s sugges s no signi ican di e -
ences be ween how g oups’ and indi iduals’ a i udes a e o mula ed on a e age.
Con olling o comp ehension o ou expe imen , we ind la ge ambigui y a e sion
among g oups han among indi iduals in he gain domain. The second measu e,
he so-called ambigui y-gene a ed insensi i i y index (Baillon e al., 2018) is also
sign-dependen .
Beyond his, ou wo k con i ms p e ious s udies ha indi iduals a e mo e
ambigui y a e se in he gain han in he loss domain. Explo ing he mechanisms
h ough which indi idual a i udes a e agg ega ed in o g oup decisions, we ind
ha he median playe s in he g oup appea o be decisi e. Tha is, ambigui y a i-
udes be ween indi iduals and g oups may depend on he speci ic dis ibu ion o
p e e ences, i.e. i he expec ed median p e e ence in a g oup is below o abo e he
expec ed mean.
The emaining pape is s uc u ed as ollows. Sec ion2 in oduces a b ie li e a-
u e e iew o he di e ences in ambigui y p e e ences be ween g oups and indi id-
uals. Sec ion3 is abou he expe imen al design and p ocedu es. Sec ion4 p esen s
ou esul s. We discuss ou esul s and conclude in Sec .5.
2 Li e a u e e iew
Ou pape on g oup a i udes owa ds ambigui y complemen s he ex ensi e li e -
a u e agg ega ion o isk a i udes in g oups (e.g., Zhang and Casa i, 2012; Bail-
lon e  al., 2016, Fuku omi e  al., 2022). This li e a u e uses ela ed expe imen al
designs. Baillon e al. (2016), o example, also use g oups o h ee and allow o
communica ion. They ind s onge agg ega ion and communica ion (g oup) e ec s
o an unanimi y ule compa ed o he majo i y ule which guided ou design choice.
375
Ambigui y a i udes o indi iduals andg oups ingain andloss…
P e ious indings on di e ences in ambigui y p e e ences be ween indi iduals
and g oups a e a he inconclusi e.1 A i s (and o en o e looked) s udy on isk and
ambigui y p e e ences be ween indi iduals and g oups was conduc ed by Ma quis
and Rei z (1969). They ind ha g oups show mo e isk/ambigui y-lo ing choices in
he gain domain and mo e ambigui y/ isk-a e se beha io occu in he loss domain.
Keck e al. (2014) ind highe ambigui y neu ali y in g oup decisions. In con as ,
Kelle e al. (2007) ind s onge ambigui y a e sion in g oups o wo compa ed o
indi idual decisions bu ely on hypo he ical s a emen s. Lloyd and Dö ing (2019)
examine he isk and ambigui y a i udes o male adolescen s and ind mo e ambigu-
i y seeking in g oups. O he s udies do no ind signi ican g oup e ec s on ambigu-
i y a i udes (B une e e al., 2015; Le a i e  al., 2017). Simon (2017) deals wi h
decisions ega ding ambiguous gambles and s ock in es men s and inds ha g oups
wi h communica ion become mo e ambigui y neu al, while g oup o ma ion wi h-
ou communica ion ends o lead o mo e ambigui y a e sion. Ca bone e al. (2019)
in es iga e in e - empo al indi idual and g oup decisions on consump ion and sa -
ing and ind ha g oups pe o m wo se in isky ou comes and be e in ambigu-
ous ou comes. Agga wal e al. (2022) ely on hypo he ical s a emen s and ind mo e
ambigui y seeking decisions a he indi idual le el in he loss domain, while mo e
ambigui y seeking esul s in g oup decisions in he gain domain. Thus, he e ec s o
g oups on ambigui y p e e ences could be sign-dependen (Ma quis and Rei z, 1969;
Agga wal e al., 2022; c . Lahno, 2014).
By compa ing indi idual and g oup decisions ollowing a cha ing oppo uni y
among g oup membe s, ou s udy also ela es o li e a u e ha examines e ec s o
social in e ac ions on indi idual ambigui y p e e ences: indi idual ambigui y a i-
udes may change (i) i decisions a e subsequen ly disclosed o o he pe sons (e.g.,
Cu ley e al., 1986; Mu huk ishnan e al., 2009; T au mann e al., 2008) o obse ed
by pee s (Tymula & Whi ehai , 2018), (ii) i pa icipan s obse e o he decisions
(e.g., Coope and Rege, 2011; Del ino e  al., 2016; Lahno, 2014), o (iii) due o
di ec social in e ac ions (e.g., Cha ness e al., 2013, Engle e al., 2011, Engle-Wa -
nick e al., 2020, Ahsanuzzaman e al., 2022).
In ou s udy, we explici ly compa e ambigui y a i udes in a gain and a loss
domain. Ex an li e a u e sugges s ha ambigui y a i udes a he indi idual le el can
be sign-dependen (e.g., Baillon and Bleich od , 2015; Koche e al., 2018; Abdel-
laoui e al., 2016). The e iew by T au mann and Van De Kuilen (2015) sugges s a
ou old pa e n o ambigui y a i udes, ha is, ambigui y a e sion o la ge p ob-
abili ies in he gain domain and o small p obabili ies in he loss domain, ambigui y
seeking o low p obabili ies in he gain domain and o la ge p obabili ies in he
loss domain. Baillon and Bleich od (2015) addi ionally ind highe ambigui y-gen-
e a ed insensi i i y o losses.
Ou s udy con ibu es o he li e a u e by explici ly compa ing g oup and indi-
idual decisions unde ambigui y in bo h loss and gain domains. Ou s udy is he
i s o compa e ambigui y p e e ences o indi iduals and g oups using he me hod
o Baillon e al. (2018). This me hod has he ad an age ha wo dis inc componen s
1 A de ailed o e iew o p e ious li e a u e indings can be ound inTable4 in he Appendix.
376
A.Minnich, A.Lange
o ambigui y a i udes can be measu ed: a i s index measu es ambigui y a e sion
( anging om ambigui y seeking beha io o ambigui y indi e ence o ambigu-
i y a e sion), a second index e e s o ambigui y-gene a ed insensi i i y. The la e
index measu es he o e - o unde es ima ion o small and la ge p obabili ies and is
also in e p e ed as a pe cei ed le el o ambigui y (Dimmock e al., 2015). We a e
unawa e o any s udy ha measu es he di e ence in pe cei ed le els o ambigui y
in g oups and indi iduals.
3 Expe imen al design andp ocedu es
This sec ion p esen s ou me hod o elici ing ambigui y a i udes i s , be o e
desc ibing he expe imen al ea men s and p ocedu es and discussing hypo heses.
3.1 Ambigui y measu emen
We use he me hod o Baillon e al. (2018) o he elici a ion o ambigui y p e e -
ences o na u al e en s. I elies on assessing he ma ching p obabili ies on single
success e en s (
E1
,
E2
,
E3
) which pa i ion he ull s a e space and he co esponding
composi e success e en s (
E12
,
E13
,
E23
). He e,
Eij
deno es
Ei∪Ej
(
j≠i
).
The ma ching p obabili ies a e deno ed by
mi
o he single and
mij
o he com-
posi e e en s (
i,j∈{1, 2, 3}
). They co espond o he winning p obabili y o a lo -
e y a which he decision make is indi e en be ween he lo e y o be ing on he
na u al e en E (e.g. Dimmock e al., 2016).
As single e en s a e mu ually exclusi e, hei ma ching p obabili ies add up o 1
o ambigui y-indi e en decision-make s.
Baillon e al. (2018) de ines wo indices ha measu e ambigui y a i udes:
whe e
ms=(m1+m2+m3)∕3
and
mc=(m23 +m13 +m12)∕3
e e o he a e ages
o ma ching p obabili ies o single and composi e e en s, espec i ely.
Index b measu es ambigui y a e sion. I anges om – 1 (minimum a e sion
o maximum ambigui y seeking) o 1 (maximum ambigui y a e sion). Ambigui y
indi e ence is gi en a a alue o 0. Index a measu es he ela ionship be ween he
ma ching p obabili ies o he single and composi e e en s and is labeled as an index
o ambigui y-gene a ed insensi i i y. I heo e ically can ange om – 2 o 4, ye
a maximum o 1 is possible i p e e ences sa is y weak mono onici y (
mc≥ms
).
Ambigui y neu ali y gi es
b=0
and
a=0
(
mc=2∕3
and
ms=1∕3
). I pa icipan s
o e weigh low p obabili ies and unde weigh high p obabili ies, a will be posi i e.
In he case o unde weigh ed low p obabili ies and o e weigh ed high p obabili-
ies, he index will be nega i e (Anan anasuwong e al., 2019). The wo indices a e
o hogonal (Baillon e al., 2021).
(1)
b
=1−ms−mca=3×(
1
3
−(
mc−ms
))

377
Ambigui y a i udes o indi iduals andg oups ingain andloss…
3.2 Expe imen al ea men s andna u al e en s
The ambiguous be is abou he empe a u e in an undisclosed ci y on he known
da e, Oc obe 18, 2020, a 2 pm (CEST). Pa icipan s we e in o med ha indica -
ing he CEST ime zone does no au oma ically mean ha he ci y is loca ed in he
CEST zone. The e en s co espond o empe a u e anges.
E1
e e s o he empe a-
u e being below 8 deg ees Celsius,
E2
o he empe a u e ange highe han o equal
o 8 deg ees Celsius and lowe han 14, and
E3
o highe han o equal o 14 deg ees
Celsius. Co espondingly,
E12
e e s o a empe a u e below 14 deg ees,
E23
o a
empe a u e abo e o equal o 8 deg ees, and
E13
o empe a u e below 8 deg ees o
weakly abo e 14 deg ees Celsius. The o de o e en s was andomized in he expe i-
men o pa icipan s o a he g oup le el o g oup ea men s.
The ea men s a y he payo domain be ween gains and losses. In he gain
domain, choosing he empe a u e be o e a lo e y pays 10 eu os in case ha e en
E ma e ializes. In he loss domain, he ealiza ion o he gi en e en E leads o a
loss o 10 eu os. Simila o Baillon e al. (2018), we use choice lis s o de e mine
he ma ching p obabili ies o each single and each composi e e en . The ma ching
p obabili ies a e gi en by he p obabili y o a isky lo e y in he gain (loss) domain
ha makes he pa icipan s indi e en be ween he isky lo e y and he ambiguous
empe a u e be .
We conside ou di e en ea men s: Indi idual-gain (IG), g oup-gain (GG),
indi idual-loss (IL), g oup-loss (GL). In IG and IL, he indi idual subjec s decide
alone. In GG and GL, g oups o h ee subjec s ha e o each a uni ied decision on
he p obabili y a which hey jus p e e a isky be o e an ambiguous one.
3.3 Expe imen al p ocedu es
The expe imen was conduc ed as an online labo a o y expe imen wi h a s uden
pool om he WiSo Resea ch Labo a o y a he Uni e si y o Hambu g. E hical
app o al was ob ained h ough he WiSo Resea ch Labo a o y. The s udy was p e-
egis e ed (Lange & Minnich, 2022). The expe imen was p og ammed wi h oT ee
(Chen e al., 2016), and h oo was used o ec ui men (Bock e al., 2014). Fou
sessions we e held in Feb ua y 2022 (9 h a 4:00 pm, 15 h a 9:00 am, 17 h a 9:00
am, 21s a 12:00 pm), and 382 pa icipan s ook pa , o whom 367 comple ed he
expe imen .2 A all ou sessions, all ea men s an a he same ime, and ea -
men a ilia ion was andomized. In he las session, wo g oups o he g oup-loss
2 Two g oups o he g oup-gain and h ee g oups o he g oup-loss ea men did no comple e he expe -
imen . Since ou g oups in he g oup-loss and wo g oups in g oup-gain ea men s did no a i e a
common solu ions in a leas one o he six ambigui y decisions, ou inal da a se includes 349 pa -
icipan s. The sample size was in o med by he numbe o independen obse a ions in p e ious s udies
(e.g., Keck e al., 2014), ins ead o being based on an explici powe es as ini ial e idence on he e ec
size and s anda d e o s using he me hod by Baillon e al. (2018) we e missing.
378
A.Minnich, A.Lange
ea men we e also un sepa a ely o equalize he numbe o obse a ions o each
ea men .3
In he ollowing, we desc ibe he imeline o he expe imen , which can also be
ound in Table 1. The ins uc ions o ou expe imen a e based on Baillon e  al.
(2018), Li (2017), and Anan anasuwong e al. (2019), and an example o he ins uc-
ions o he ea men g oup-loss is gi en in Online Appendix A.
Fi s , he pa icipan s ge a desc ip ion o he ambigui y ask. They hen ha e o
co ec ly answe wo con ol ques ions abou he ambigui y ask. Pa icipan s had o
s ay in he ins uc ions o a leas 15seconds pe page and a leas h ee minu es in
o al.
Second, he indi idual and g oup decisions conce ning he ambigui y ake place
o he h ee single and h ee composi e e en s. The o de o he six e en s was an-
domized a he indi idual le el o a he g oup le el o g oup ea men s. Following
Baillon e al. (2018), pa icipan s ace choice lis s o six empe a u e e en s in an
undisclosed ci y. The g oups can cha oge he , while he indi iduals can only cha
alone. The g oups ha e h ee chances pe choice lis o each a common unanimous
solu ion (Zhang and Casa i, see 2012). The g oups a e in o med ha hey will au o-
ma ically ecei e he wo s payou (0 in he gain domain and – 10 eu os in he loss
domain) i a choice lis o an e en wi hou a common solu ion is andomly selec ed
o hei payou . Figu e5 in heAppendix shows ha he g oups need ewe a emp s
o each unanimi y o e ime. A e wa d, he e was a one-minu e b eak.
Thi d, subjec s answe i e ques ions o a cogni i e e lec ion ask. They ge paid
(addi ional/ less 2 eu os pe igh /w ong answe in loss s. gain domain) o balance
ou he payou s (see Koche e al. (2018) o a simila paymen p ocedu e). The cog-
ni i e e lec ion es has he ad an age o being ela i ely quick. I also has al eady
Table 1 Expe imen al se up in he espec i e ea men s
T ea men s
Sequence G oup-gain Indi idual-gain G oup-loss Indi idual-loss
Phase 1 Explana ion o he expe imen , Con ol ques ions
Phase 2 Ambigui y ask
+10 eu os o 0 -10 eu os o 0
Phase 3 Cogni i e e lec ion es (5 ques ions) o balance he payou s
-2 eu os pe w ong answe +2 eu os pe  igh answe
Phase 4 Ques ionnai e and payou
Pa icipan s 129 47 126 47
Independen
obse a-
ions
43 47 42 47
3 We use he con ol a iables o check he andomiza ion o ou ea men s (see Table5 in heAppen-
dix). We do no ind sys ema ic di e ences be ween ea men s. Only mino di e ences exis in some
demog aphic a iables (gende , pa en ship, income, p e ious expe imen pa icipa ion, acul y) and com-
p ehension asks.
379
Ambigui y a i udes o indi iduals andg oups ingain andloss…
been used as a con ol a iable o ambigui y p e e ences in Li (2017). The i e
ques ions a e based on F ede ick (2005) and Li (2017).
Fou h, subjec s ill ou a ques ionnai e conce ning demog aphic a iables and
beha io al a i udes, be o e he payou s a e ca ied ou . The ques ions include age,
gende , acul y, numbe o semes e s, numbe o p e ious pa icipa ion in expe i-
men s a he Uni e si y o Hambu g, income, accina ion s a us (Co id-19), pa en -
ship, comp ehension o he ambigui y asks, ou su ey ques ions abou ambigu-
i y p e e ences (Ca a o a & Sch öde , 2019),4 a su ey measu e o isk a i udes
(Dohmen e al., 2011), a wea he - ela ed isk a i ude and he en-i em pe sonali y
in en o y (Gosling e al., 2003; Muck e al., 2007). We decided o collec he big
i e in en o y using he en-i em pe sonali y in en o y because Zhang and Casa i
(2012) ound e ec s o he big i e in en o y on how g oups eached a decision
unde isk. In addi ion, pa icipan s o g oups answe ed wo mo e ques ions abou
how hey a i ed a he g oup decision and how hei own p e e ences con ibu ed o
he g oup decision.
We used a andomized incen i e scheme o pay o he ambigui y ask, i.e. he
decision one o he six e en s ma e ed o inal paymen s o which one andom line
o he decision able (lo e y s. ambiguous be ) was selec ed. Fo he na u al e en s,
pa icipan s o g oups we e andomly assigned o one o 30 ci ies which hen de e -
mined he payo .5 The espec i e ci y was only e ealed a e he expe imen on he
paymen sc een. Paymen s om his ask hus we e ei he plus 10 eu os (gain ea -
men s), minus 10 eu os (loss ea men ) in case o ealiza ion o he assessed e en
o 0 eu o o he wise.
The paymen o he pa icipan s consis s o a s a ing amoun o 10 Eu o, he
paymen based on andomly selec ed ambigui y asks (10, –10, o 0), and he pay-
o om he cogni i e e lec ion asks (adding (loss ea men s) o sub ac ing (gain
ea men s) 2 Eu o pe igh /w ong answe ). Pa icipan s can hus ea n be ween 0
and 20 eu os in e e y ea men .
The a e age paymen o pa icipan s who comple ed he expe imen was 12.72
eu os (IG=12.47, GG=11.88, IL=12.43, GL =13.74).6 The a e age comple ion
ime o hese pa icipan s was 38.5min (IG=21, GG=43, IL=21, GL=46).
4 Due o o empo al conciseness, we decided agains he i h measu emen o Ca a o a and Sch öde
(2019), namely he dynamic Ellsbe g wo-colo u n hough expe imen measu emen . Acco dingly, o
he calcula ion o he ambigui y sco e we only use he con e sion o he Like scales and do no use he
cons an o 130.
5 Unknowingly o pa icipan s, ci ies we e selec ed such ha each single success e en being ue o en
ci ies and each composi e e en being ue o 20 ci ies co espondingly. The numbe o ci ies is se so
high ha no use ul in o ma ion can be exchanged be ween he pa icipan s be ween he di e en sessions.
6 Pa icipan s who could no inish he expe imen because a eam membe had d opped ou we e paid a
kind o hou ly wage depending on hei ime commi men .
380
A.Minnich, A.Lange
3.4 Coding heda a
As desc ibed, we use choice lis s o de e mine he ma ching p obabili ies. The
choice lis s con ain 28 ows and a e adjus ed o a oid middle bias.7 Fo consis ency
o answe s, pa icipan s may only indica e a single swi ching poin om when hey
p e e he isky lo e y. The ma ching p obabili y (indi e ence poin ) is coded as
he midpoin be ween he wo alues o he isky lo e y whe e hey swi ched p e -
e ences wi h wo excep ions a he ex emes: in he gain (loss) domain, we se he
ma ching p obabili y o 0 (100) i he lo e y is always p e e ed and 100 (0) i he
ambiguous be is always p e e ed.8
Based on he ma ching p obabili ies, he ambigui y indices b and a a e calcula ed
(Baillon e al., 2018). Ye , he in e p e a ion o b alues di e in he gain s. loss
domains. A posi i e b e lec s ambigui y a e sion in he gain domain ye ambigu-
i y seeking in he loss domain. In o de o acili a e he in e p e a ion o esul s, we
mul iply b as calcula ed wi h he ma ching p obabili ies wi h
−1
in he loss domain.9
The a iables om he su ey we e used o code con ol a iables, see Table6 in
heAppendix which also gi es he summa y s a is ics o hese a iables. Fo exam-
ple, we coded dummy a iables based on demog aphic a iables, isk and unce -
ain y p e e ences, as well as expe imen -speci ic a iables (cogni i e e lec ion es ,
con ol ques ions e o s, comp ehension). We mos ly use a median spli o he con-
ol a iables, c ea ing dummy a iables as independen a iables.10 We code he
s a emen s on g oups’ decision-making and he inclusion o own p e e ences in he
g oup decision as ca ego ical a iables.
In ou main analyses, we emo ed pa icipan s who did no comple e he expe -
imen . This was he case o g oups when one pe son d opped ou o he expe i-
men .11 Nex , we emo ed he g oups ha did no ind a common solu ion in a leas
one o he six decisions o he ambigui y ask.12
7 I pa icipan s choose op ion B in he gain and loss domain om ow 15 onwa ds, hey ha e a ma ching
p obabili y o 32.5 o indi idual success e en s and 67.5 o composi e success e en s.
8 The choice lis s un om 0 o 100% in he gain domain and om 100 o 0% in he loss domain o
ensu e he simila i y o ea men s (deciding when o p e e a isky lo e y).
9 In ui i ely, losing in case o e en E in he loss domain co esponds o winning i he e en does no
occu (payo 0 ins ead o –10). Acco dingly, he ma ching p obabili ies in he gain domain can be con-
e ed in o he loss domain by calcula ing he ma ching p obabili ies o he coun e e en s
ms
=1−mc
and
mc
=1−ms
. Thus
b=1−(1−mc)−(1−ms)=−1∗(1−ms−mc)
. The ambigui y-gene a ed
insensi i i y s ays he same
a
=3×(
1
3
− ((1−ms)−(1−mc))) = 3×(
1
3
−(
mc−ms
))
10 Fo he calcula ion o he medians, we only used he da a whe e g oups eached a common solu ion
o e e y decision and only da a om pa icipan s in g oup ea men s.
11 The e is one excep ion, as one pa icipan in he g oup-gain ea men did no ha e ime o comple e
he ques ionnai e a he end o he expe imen due o a ollow-up appoin men .
12 This applied o a o al o only 6 g oups ac oss gain (2) and loss (4) domain. Including hese g oups
does no make any di e ence o ou esul s when using he median alue o he las decision ound in
which no common solu ion was ound.
387
Ambigui y a i udes o indi iduals andg oups ingain andloss…
Fo comple eness, Fig.4 shows he co esponding pic u es o ambigui y index
a. He e, he di e en lines do no signi ican ly di e : he likely eason is ha he
di e ence be ween ma ching p obabili ies o composi e and single e en s does no
much di e be ween he indi iduals wi h he minimal, median, o maximal ma ch-
ing p obabili y (and hence ambigui y a e sion) in a g oup o h ee.
Ano he way o in es iga e he mechanisms behind g oup decisions is guided
by he su ey measu es on how pa icipan s in he g oup ea men pe cei ed he
decision-making p ocess. On a e age, mos pa icipan s eached a g oup decision
h ough unanimi y (169), while 55 eached a solu ion wi h majo i y and 30 wi h he
imposi ion o one pe son’s p e e ences in he g oup. In he p ocess, 177 pa icipan s
Fig. 3 Ambigui y index b in gain domain (panel (a)) and loss domain (panel (b)). The lines show cumu-
la i e dis ibu ions based on boo s apping indi idual ma ching p obabili ies in g oups o h ee (1000
andom g oups o h ee). The cd s co espond o he ambigui y indices b i he boo s apped g oups
always chose he minimal, he median, o he maximal ma ching p obabili y in hei g oup o h ee. The
ou h line displays he ac ual cd o index b in he g oup ea men s
Fig. 4 Ambigui y index a in gain domain (panel (a)) and loss domain (panel (b)). The lines show cumu-
la i e dis ibu ions based on boo s apping indi idual ma ching p obabili ies in g oups o h ee (1000
andom g oups o h ee). The cd s co espond o he ambigui y indices a i he boo s apped g oups
always chose he minimal, he median, o he maximal ma ching p obabili y in hei g oup o h ee. The
ou h line displays he ac ual cd o index a in he g oup ea men s

388
A.Minnich, A.Lange
s a ed ha hey adjus ed hei p e e ences, 73 pa icipan s ound ha hey could
en o ce hei own p e e ences, and ou people subjec s assessed ha hey did no
con ibu e hei own p e e ences o he g oup decision. O e all, his is consis en
wi h g oups inding some middle g ound in he assessmen o ma ching p obabili-
ies as indi iduals wi h he la ges o lowes indi idual p e e ences a e mo e likely
needed o adjus hei p e e ences in o de o each a consis en g oup decision.
We now explo e i hese indi idual iews a e d i en by some socio-economic
cha ac e is ics. Table9 in he Appendix shows he esul s om mul inomial eg es-
sions on g oup s a egies, he esul s o he eg essions on he inclusion o own p e -
e ences can be ound in Table10. In bo h ables, wo columns always belong o he
same unde lying eg ession. The esul s can always be in e p e ed in compa ison o
he baseline esponse op ion (unanimi y in Table9 and adjus ed in Table10). O e all,
i is no iceable ha only e y ew con ol a iables signi ican ly a ec he depend-
en a iables. Tha is, he iews on how he decisions we e made in a g oup do no
signi ican ly di e be ween indi iduals wi h di e en socio-economic cha ac e is ics.
4.3 Mechanisms behindg oup decisions
Figu e5 in heAppendix shows ha , on a e age, he g oups ag eed mo e and as e
o e ime, bo h in e ms o gain and loss domains. A signi ican numbe o g oups
eached an ag eemen in he i s ound, only a ew g oups equi ed mo e han wo
a emp s o each an ag eemen .
A o al o 4859 ex messages we e w i en by he 255 pa icipan s in he g oup
ea men s (GL, GG). This co esponds o an a e age o 3.2 ex messages w i en by
a pa icipan in one o he six decision asks. This a e age numbe a ies om 3.1 in
case ha pa icipan s la e s a ed ha he g oup decision was unanimous, 3.6 i hey
s a ed ha a majo i y solu ion was ound, o 3.0 i one pe son’s p e e ences we e
imposed on he g oup. No su p isingly, he numbe o messages sen by indi iduals
who s a ed o no ha e con ibu ed hei own p e e ences o he g oup decision was
much smalle pe decision ask (1.5) han i indi iduals adjus ed hei p e e ences
(3.2) o i hey imposed hei p e e ences (3.1).
A quali a i e assessmen o he cha messages sugges s ha decisions may be
based on comp omises. The messages o en in ol e someone jus asking “Wha do
you hink?”, li le exchanges o a gumen s, and a he someone making a sugges ion,
a coun e sugges ion (“going up” o “going down”), be o e o he s may ag ee. In some
cases, he e a e also a gumen s abou he possible loca ions in he wo ld, he s a e o
day (nigh in some places), he mon h o Oc obe , and he global a e age empe a u e.
In some g oups, he p obabili ies o occu ence a e discussed o some imes coun e -
p obabili ies a e p esen ed. Since he subjec s do no know which loca ion is c ucial
o he empe a u e e en , he e is no easible way o he subjec o signal expe ise.16
16 Gi en he b e i y o mos messages, we do no pe cei e ha a machine lea ning o ex analysis
app oach could p o ide much deepe insigh s as we do no ha e an independen measu e o indi idual
ambigui y a i udes. Elici ing hose appea s bene icial when u he in es iga ing he g oup decision p o-
cesses in u u e esea ch by also compa ing how di e en agg ega ion p ocedu es a ec he necessi y o
( ex ) exchanges be ween g oup membe s.
389
Ambigui y a i udes o indi iduals andg oups ingain andloss…
4.4 Robus ness checks
We p e egis e ed obus ness checks o in es iga e i ea men e ec s a ise o spe-
ci ic subse s o he sample o which one can an icipa e be e -in o med decisions.
A i s obus ness check is based on he comp ehension o he expe imen pa ici-
pan s as sel -assessed in he su ey (see Online Appendix C). A second obus ness
check akes he answe s o he con ol ques ion as a c i e ion o sample quali y (see
Online Appendix D). In doing so, we compa e g oups ha mee he selec ion c i e-
ion a he median wi h indi iduals who mee he c i e ion.
Res ic ing he sample on indi iduals and g oups who epo a good comp ehen-
sion (see Online Appendix C), we ind di e ence be ween g oups and indi iduals
in he gain domain: he es ima o g oup is posi i e and ma ginally signi ican o
index b (
𝛽1=0.10
, (54)=1.91, p=0.061), and posi i e and signi ican o index
a (
𝛽1=0.36
, (54)=3.21, p=0.002) (see Table OC.4).17 Di e en ly, no signi i-
can di e ences be ween g oups and indi iduals a ise in he loss domain (see Table
OC.5). In ac , he in e ac ion e ec
g oup ×loss
is nega i e and highly signi ican
o index a (
𝛽3=−0.50
, (90)=
−2.80
, p=0.006) (see Table OC.6).18
Simila e ec s esul i we choose he selec ion c i e ion o no mis akes in he com-
p ehension ques ions (see Online Appendix D). In he gain domain, g oup is posi i e
and ma ginally signi ican o ambigui y index b (
𝛽1=0.10
, (55)=1.71, p=0.093)
and posi i e and signi ican o ambigui y index a (
𝛽1=0.24
, (55) = 2.29,
p=0.026) (see Table OD.4).19 In he loss domain, no di e ence be ween g oup and
indi idual ambigui y indices esul s (see Table OD.5). Consis en ly, he es ima o on
he in e ac ion e ec o g oup and loss is nega i e and ma ginally signi ican e ec
o he ambigui y index a (
𝛽3=−0.29
, (115)=
−1.81
, p=0.073) (see Table OD.6).
Ye , hese esul s come wi h a wo d o cau ion as he numbe o obse a ions
sa is ying good comp ehension o no e o in he comp ehension ques ions, espec-
i ely, is no pa icula ly la ge.20 O e all, he obus ness checks ega ding he pa -
icipan s’ unde s anding yield he ollowing esul :
Resul 3 Fo subsamples wi h likely be e -in o med decisions (good comp ehen-
sion o no mis akes in con ol ques ions), signi ican di e ences be ween indi idu-
als and g oups a ise wi h espec o ambigui y a i udes in he gain domain, whe e
17 This e ec is d i en by g oups assigning a lowe ma ching p obabili ies o e en s
E12
(
𝛽1=−14.25
,
(54)=
−3.14
, p=0.003) and
E23
(
𝛽1=−12.92
, (54)=
−3.53
, p=0.001) (see Table OC.4).
18 This e ec is d i en by he posi i e in e ac ion e ec o
g oup ×loss
on
E12
(
𝛽3=21.89
,
(90)=2.83, p= 0.006) and
E23
(
𝛽3=16.20
, (90)=2.02, p=0.047) (see Table OC.6). O e all, he
mean alues o
E12
and
E23
a e in he o de
(IG)>
(GL)>
(GG)>
(IL)
19 This is again d i en by g oups assigning a lowe ma ching p obabili ies o e en s
E12
(
𝛽1=−10.18
,
(55)=
−2.34
, p=0.023) and
E23
(
𝛽1=−12.25
, (55)=
−2.86
, p=0.006) (see Table OD.4).
20 We ha e also elaxed he c i e ia o he subsamples. I we conside only g oups (median) and indi id-
uals wi h good o a he good unde s anding (see Table OC.1, OC.2, and OC.3) o g oups (median) and
indi iduals wi h one o less e o s on he comp ehension ques ions (see Table OD.1, OD.2, and OD.3),
he esul s poin o simila di ec ions.
390
A.Minnich, A.Lange
g oups show la ge ambigui y a e sion and highe ambigui y-gene a ed insensi i i y.
This g oup e ec o ambigui y-gene a ed insensi i i y is domain speci ic signi i-
can ly di e en in he loss domain.
5 Conclusions
We compa e ambigui y a i udes o g oups and indi iduals in he gain and loss domain
in a be ween-subjec design. G oups o h ee can cha wi h each o he in o de o come
o an unanimous decision. We use he me hod o Baillon e al. (2018) which allows
measu ing ambigui y a i udes by dis inguishing ambigui y a e sion (o indi e ence, o
ambigui y seeking beha io ) and ambigui y-gene a ed insensi i i y o na u al e en s.
We apply his me hod o elici ambigui y a i udes ega ding empe a u e e en s.
Compa ing gain and loss domains, we ind mo e ambigui y-seeking beha io and
a highe ambigui y-gene a ed insensi i i y in he loss domain. This esul is con-
sis en wi h li e a u e (e.g., T au mann and Van De Kuilen, 2015; Baillon and Blei-
ch od , 2015) and ex ends hese esul s o g oup decisions and o ambigui y a i udes
owa ds na u ally occu ing empe a u e e en s.
In bo h domains, he ma ching p obabili ies o small(e ) p obabili ies (single
e en s) a e agg ega ed abo e 1, while he co esponding sum o la ge( ) p obabili-
ies (composi e e en s) is agg ega ed below 2. Wi hou agg ega ing hese ma ching
p obabili ies in o he ambigui y indices à la Baillon e al. (2018), ou esul s a e con-
sis en wi h a ou old pa e n o ambigui y a i udes (T au mann and Van De Kuilen,
2015): subjec s a e ambigui y a e se o mo e likely e en s in he gain domain and
o less likely e en s in he loss domain, while hey a e ambigui y-seeking o low
likelihood e en s in he gain domain and high likelihood e en s in he loss domain.
Ye , he di e ences be ween g oups and indi idual a i udes a e ma ginal. In
ou main sample, we do no ind signi ican di e ences be ween indi idual ambi-
gui y a i udes and hose ha esul in g oup decisions. Howe e , la ge ambigui y
a e sion and ambigui y-gene a ed insensi i i y esul s g oups han o indi iduals
in he gain domain when we concen a e on subjec s indica ing a highe comp e-
hension o he expe imen . Fo hose subsamples, he g oup e ec on ambigui y-
gene a ed insensi i i y appea s o be sign-dependen , i.e. indi idual p e e ences
a e agg ega ed di e en ly in o g oup decisions in he gain s. he loss domain.
Appendix
391
Ambigui y a i udes o indi iduals andg oups ingain andloss…
Table 4 O e iew o s udies wi h g oup decisions conce ning ambigui y a i udes
No e: Indi dual=I, G oup=G; AA=Ambigui y A e sion, AS=Ambigui y Seeking, AN= Ambigui y Neu al
S udy G oup Size & Ambigui y Communica ion Domain Compa ison Main Finding
Se -up Le els Decision ule Me hod Shi o
Ma quis & 3–7 S ake, P obabili y Discussion Gain, Loss S ake AA(Loss( –25%))
Rei z (1969) I–G (Wi hin) Winning P ice Unanimi y Mixed size AS(Gain, Mixed, Loss (–10%)))
Kelle e al 2 P obabili y Discussion Gain WTP AA
(2007) I-G (Wi hin) – (gamble)
Keck e al 3 P obabili y Discussion Gain Ce ain y equi alen o AN
(2014) I-G,G-I (Wi hin) Majo i y Gamble
B une e e al 3 P obabili y No communica ion Gain Risk o No e ec
(2015) I-G, G-I (Wi hin) Unanimi y, Majo i y Ambigui y
Le a i e al 3 P obabili y No communica ion Gain Risk o AA (Delega e 1,2)
(2017) I-G, G-I (Wi hin) Majo i y, Delega e(1,2) Ambigui y
Simon 3 P obabili y Discussion (Cha :yes/no) Gain S ock in es men ; Minimum AN(discussion)
(2017) I-G (Wi hin); I,G (Be ween) Unanimi y, A e age selling p ice (Risk o Ambigu-
i y)
AS (discussion, s ock in es -
men )
Ca bone e al 2 P obabili y Discussion (Cha ) Gain Consump ion/Sa ing Be e planne
(2019) I, G (be ween) – Expe imen (in e - empo al) unde ambigui y
Lloyd & 2 P obabili y Discussion Gain Wheel o Fo une ask AS
Dö ing (2019) I, G (Be ween) – (Ambigui y o Ce ain)
Agga wal e al 5 (pa ly 4 o 6) P obabili y Discussion Gain, Loss Lo e y choices AS (gain)
(2022) I- G (Wi hin) Winning P ice Unanimi y (Risk o Ambigui y)
392
A.Minnich, A.Lange
Table 5 Summa y s a is ics
ea men GG IG GL IL
Va iable N Mean Sd N Mean Sd N Mean Sd N Mean Sd
Gende ( emale)* 128 0.688 0.465 47 0.809 0.398 126 0.587 0.494 47 0.66 0.479
Gende (di e se)* 128 0.00781 0.0884 47 0 0 126 0.0159 0.125 47 0.0213 0.146
Age 128 25.6 4.35 47 24.8 4.17 126 24.9 4.57 47 25.5 4.61
Pa en ship* 128 0.0469 0.212 47 0.0426 0.204 126 0.0476 0.214 47 0.0851 0.282
Income0 (0–600)* 128 0.234 0.425 47 0.213 0.414 126 0.238 0.428 47 0.319 0.471
Income (600–800)* 128 0.234 0.425 47 0.234 0.428 126 0.246 0.432 47 0.191 0.398
Income (800–1200)* 128 0.375 0.486 47 0.383 0.491 126 0.341 0.476 47 0.362 0.486
Income (1200–1600)* 128 0.102 0.303 47 0.149 0.36 126 0.119 0.325 47 0.0426 0.204
Income (1600–2000)* 128 0.0312 0.175 47 0.0213 0.146 126 0.0159 0.125 47 0.0213 0.146
Vaccina ed* 128 0.922 0.269 47 1 0 126 0.937 0.245 47 0.915 0.282
Vaccina ed (no )* 128 0.0156 0.125 47 0 0 126 0.0159 0.125 47 0.0213 0.146
Vaccina ed (no Ye )* 128 0.00781 0.0884 47 0 0 126 0 0 47 0.0213 0.146
Semes e s 128 8.34 6 47 7.85 4.17 126 7.87 4.9 47 8.94 5.89
Expe imen pa icipa ions 128 9.91 13.2 47 10.8 11.6 126 9.94 13.1 47 11 9.23
Facul y (economics, social)* 128 0.445 0.499 47 0.574 0.5 126 0.516 0.502 47 0.468 0.504
Ambigui y index 128 130 17.6 47 131 16.1 126 131 16.6 47 127 18.5
Risk seeking (gene al) 128 5.73 1.87 47 5.81 2.19 126 5.7 1.99 47 5.45 2.2
Risk seeking (wea he ) 128 6.48 2.34 47 6.49 2.62 126 6.35 2.43 47 5.96 2.78
Ex a e sion 128 4.13 1.46 47 4.35 1.29 126 4.15 1.38 47 4.45 1.62
Ag eeableness 128 5.1 1.06 47 5.15 1.16 126 5.17 0.976 47 5.05 1.03
Conscien iousness 128 5.46 1.12 47 5.59 1.16 126 5.28 1.19 47 5.36 1.2
Emo ional_s abili y 128 4.6 1.24 47 4.31 1.29 126 4.67 1.36 47 4.59 1.12
Openness_ o_expe iences 128 5.11 1.14 47 5.06 1.13 126 5.28 1.07 47 5.36 1.12
Cogni i e es (co ec answe s) 129 3.48 1.52 47 3.47 1.33 126 3.96 1.32 47 3.55 1.35

393
Ambigui y a i udes o indi iduals andg oups ingain andloss…
∗
The a iables a e dummy a iables. Fo some o he dummy a iables, he e was ano he esponse op ion, which a e lis ed
as ollows: Gende (male), income (mo e han 2000), accina ed (no speci ied), comp ehension (no )
Table 5 (con inued)
ea men GG IG GL IL
Va iable N Mean Sd N Mean Sd N Mean Sd N Mean Sd
E o s (con ol ques ions) 129 0.667 1.07 47 1.21 2.15 126 1.11 2.2 47 0.702 1.38
Comp ehension (yes)* 128 0.578 0.496 47 0.574 0.5 126 0.429 0.497 47 0.574 0.5
Comp ehension ( a he yes)* 128 0.398 0.492 47 0.362 0.486 126 0.476 0.501 47 0.34 0.479
Comp ehension ( a he no )* 128 0.0234 0.152 47 0.0426 0.204 126 0.0794 0.271 47 0.0851 0.282
394
A.Minnich, A.Lange
Table 6 Composi ion o he dummy a iables o he explo a o y eg essions o g oup decisions
Va iable name 1 0
acul y (economics, social) business, economics and social sciences o he s
age_high median (24.5) and abo e
≤
24
income_high median (800–1200 €) and abo e
≤
800€
semes e s_high median (7) and abo e < 7
pa icipa ion_high median (6) and abo e < 6
isk_1_high (gene al) median (6) and abo e < 6
isk_2_high (wea he ) median (7) and abo e < 7
amb_index_high median (129.5) and abo e < 129.5
ex a e sion_high median (4) and abo e < 4
ag eeableness_high median (5) and abo e < 5
conscien iousness_high median (5.5) and abo e < 5.5
emo ional_s abili y_high median (4.5) and abo e < 4.5
openness_ o_expe ience_high median (5.5) and abo e < 5.5
quiz_high median(4) and abo e <4
e o _less median(0) >0
comp ehension_high median (good comp ehension) o he op ions
395
Ambigui y a i udes o indi iduals andg oups ingain andloss…
Fig. 5 O e iew o numbe o a emp s o achie e unanimi y in he g oup-gain and g oup-loss ea men s
o all six decisions
396
A.Minnich, A.Lange
Table 7 Linea eg essions: ea men e ec s (gain domain)
Dependen a iable
b a E1 E2 E3 E12 E13 E23
(1) (2) (3) (4) (5) (6) (7) (8)
g oup 0.052 0.122 3.537 –1.323 –3.879 –4.966 –3.549 –5.352
(0.050) (0.105) (3.755) (4.189) (3.804) (4.335) (4.562) (3.926)
Cons an 0.043 0.379
∗∗∗
33.649
∗∗∗
40.532
∗∗∗
38.298
∗∗∗
59.617
∗∗∗
55.723
∗∗∗
59.213
∗∗∗
(0.039) (0.084) (2.908) (3.310) (3.141) (3.134) (3.084) (3.267)
Obse a ions 90 90 90 90 90 90 90 90
R
2
0.012 0.015 0.010 0.001 0.012 0.015 0.007 0.020
Adjus ed R
2
0.001 0.004 –0.001 –0.010 0.0003 0.004 –0.004 0.009
Residual S d.
E o
0.236 0.499 17.789 19.891 18.155 20.394 21.334 18.754
F S a is ic 1.085 1.345 0.888 0.099 1.025 1.331 0.621 1.829
Table 8 Linea eg essions: ea men e ec s (loss domain)
No e:
∗
p<0.1;
∗∗
p<0.05;
∗∗∗
p<0.01; he e oscedas ici y consis en s anda d e o s (“HC3”) in pa en heses
and es ima ed wi h he
R package sandwich (Zeileis, 2004; Zeileis e al., 2020)
Dependen a iable
b a E1 E2 E3 E12 E13 E23
(1) (2) (3) (4) (5) (6) (7) (8)
g oup 0.025 –0.084 2.900 0.695 –3.982 9.407
∗∗
–0.427 –0.978
(0.064) (0.099) (4.431) (4.170) (3.789) (4.312) (4.505) (4.514)
Cons an –0.081 0.579
∗∗∗
36.862
∗∗∗
39.936
∗∗∗
40.053
∗∗∗
48.415
∗∗∗
53.213
∗∗∗
57.287
∗∗∗
(0.054) (0.068) (3.635) (3.102) (3.030) (3.496) (3.510) (3.519)
Obse a ions 89 89 89 89 89 89 89 89
R
2
0.002 0.008 0.005 0.0003 0.012 0.051 0.0001 0.001
Adjus ed R
2
–0.010 –0.003 –0.007 –0.011 0.001 0.040 –0.011 –0.011
Residual S d.
E o
0.306 0.461 21.107 19.600 17.983 20.504 21.301 21.345
F S a is ic 0.152 0.733 0.419 0.028 1.087 4.668
∗∗
0.009 0.047
403
Ambigui y a i udes o indi iduals andg oups ingain andloss…
Le a i, M. V., Napel, S., & So ape a, I. (2017). Collec i e choices unde ambigui y. G oup Decision and
Nego ia ion, 26(1), 133–149.
Li, C. (2017). A e he poo wo se a dealing wi h ambigui y? Jou nal o Risk and Unce ain y, 54,
239–268.
Lloyd, A., & Dö ing, A. K. (2019). When do pee s in luence adolescen males’ isk aking? Examining
decision making unde condi ions o isk and ambigui y. Jou nal o Beha io al Decision Making,
32(5), 613–626.
Ma quis, D. G., & Rei z, H. J. (1969). E ec o unce ain y on isk aking in indi idual and g oup deci-
sions. Beha io al Science, 14(4), 281–288.
Minnich, A., Roggenkamp, H. & Lange, A. (2024). Ambigui y a i udes and su p ises: expe imen al e i-
dence on communica ing new in o ma ion wi hin a la ge popula ion sample. Jou nal o Economic
Beha io & O ganiza ion. h ps:// doi. o g/ 10. 1016/j. jebo. 2024. 106778
Muck, P. M., Hell, B., & Gosling, S. D. (2007). Cons uc alida ion o a sho i e- ac o model ins u-
men . Eu opean Jou nal o Psychological Assessmen , 23(3), 166–175.
Mu huk ishnan, A. V., Wa hieu, L., & Xu, A. J. (2009). Ambigui y a e sion and he p e e ence o es ab-
lished b ands. Managemen Science, 55(12), 1933–1941.
R Co e Team (2023). R: A language and en i onmen o s a is ical compu ing. R Founda ion o S a is i-
cal Compu ing, Vienna, Aus ia. h ps:// www.R- p oje c . o g/.
Simon, J. (2017). Auswi kungen on G uppenbildung und Kommunika ion au ambigue En scheidungen.
Sp inge Gable .
T au mann, S. T., & Van De Kuilen, G. (2015). Ambigui y a i udes. In The Wiley Blackwell Handbook o
Judgmen and Decision Making, 2, 89–116.
T au mann, S. T., Vieide , F. M., & Wakke , P. P. (2008). Causes o ambigui y a e sion: Known e sus
unknown p e e ences. Jou nal o Risk and Unce ain y, 36(3), 225–243.
Tymula, A., & Whi ehai , J. (2018). Young adul s gamble less when obse ed by pee s. Jou nal o Eco-
nomic Psychology, 68, 1–15.
on Gaudecke , H.-M., Wog olly, A., & Zimpelmann, C. (2022). The dis ibu ion o ambigui y a i udes.
Cesi o wo king pape , No. 10079.
Wickham, H. (2016). ggplo 2: Elegan G aphics o Da a Analysis. New Yo k: Sp inge .
Zeileis, A. (2004). Econome ic compu ing wi h hc and hac co a iance ma ix es ima o s. Jou nal o
S a is ical So wa e, 11(10), 1–17.
Zeileis, A., Köll, S., & G aham, N. (2020). Va ious e sa ile a iances: An objec -o ien ed implemen a-
ion o clus e ed co a iances in R. Jou nal o S a is ical So wa e, 95, 1–36.
Zhang, J., & Casa i, M. (2012). How g oups each ag eemen in isky choices: An expe imen . Economic
Inqui y, 50(2), 502–515.
Publishe ’s No e Sp inge Na u e emains neu al wi h ega d o ju isdic ional claims in published maps
and ins i u ional a ilia ions.