A dvances in Metho dology and Statistics , 2023, 20 (1), 77–93. https://doi.org/10.51936/alnb3451
Resear ching moral foundations using
the Bay esian appr oach
Kristina Rakinić
a
Univ ersity of Ljubljana, Faculty of So cial Sciences, Ljubljana, Slo v enia
Abstract
Moral foundations the or y describ es a comple x space of human morality along fiv e moral
foundations: car e/harm, fairness/cheating, lo yalty/b etrayal, and sanctity/degradation. The
first aim of this study is to e xplor e the r ole of so cio-demographic variables in endorsing the
fiv e moral foundations and to pr o vide insight into the r ole of moral foundations in attitudes
to war ds homeless p e ople . The se cond obje ctiv e of the study is to pr o vide an e xample of
emplo ying the Bay esian statistical appr oach. Ther e w er e 162 participants in our study ( 27 %
male , mean age 36 y ears). Conv enience sampling metho d was use d. A ge and e ducation
found to b e the most imp ortant pr e dictors of moral foundations, with Bay esian factors
sho wing the str ongest e vidence for age as a p ositiv e pr e dictor of moral foundations. Gender ,
p olitical affiliation, car e and authority w er e found to b e imp ortant pr e dictors of attitudes
to war ds homeless p e ople , with Bay esian factors sho wing much str onger e vidence for moral
foundations than so cio demographic variables. This study lays the gr oundw ork for r esear ch
on moral foundations in Slo v enian samples and pr o vides a practical e xample of Bay esian
appr oach application in so cial science .
K e y w or ds: moral foundations, so cio demographic variables, attitudes, Bay esian statistics
1. Intr o duction
In its essence , morality delineate the distinctions b etw e en right and wr ong. Ho w e v er , e v en
within a shar e d cultural frame w ork, individuals e xhibit variations in discerning sp e cific
actions as morally right or wr ong. Mor e o v er , these judgements can b e fluid, subje ct to
shifts and changes o v er time and conte xt dep endent. These study wants to addr ess the
endorsement of moral foundation in e valuating immoral b ehaviours. The the or y of moral
foundations has so far b e en applie d to numer ous conte xts and so cieties, ho w e v er its appli-
cation in the Slo v enian samples is absent. The aim of these study is to test the applicability
in the Slo v enian conte xt using a r elativ ely underutilise d, although pr omising, Bay esian
appr oach.
Email addr ess: kristina.rakinic@fdv .uni-lj.si (Kristina Rakinić)
ORCID iD: (Kristina Rakinić)

78 Rakinić
1.1. Moral foundations the or y
One of the most widely cite d authors in moral psy chology is K ohlb erg ( 1969 ), who de v elop e d
the six-stage the or y of moral de v elopment. K ohlb erg understo o d morality as justice , that
which is just is also moral. Gilligan ( 1977 ) criticise d K ohlb erg for sample bias ( his the or y
was de v elop e d base d on studies including only b o ys) and argue d that girls, b e cause of their
so cialisation, ar e mor e oriente d to war ds the moral asp e ct of car e rather than justice . Wher eas
K ohlb erg emphasise d the rational part of morality , Gilligan emphasise d the emotional part.
Later , base d on e v olutionar y and anthr op ological the ories, Haidt and Joseph ( 2004 ) br oaden
these notions and de v elop e d Moral Foundations The or y , a descriptiv e the or y of moral
b ehaviour that includes fiv e foundations that ar e univ ersal and generalisable acr oss cultur es.
Haidt and Joseph ( 2004 ) hav e identifie d fiv e univ ersal moral foundations: car e/harm,
fairness/cheating, lo yalty/b etrayal, authority/subv ersi on and sanctity/degradation. The
moral foundations the or y r ests on four assumptions. First, nativism asserts that the initial
structur e of the moral mind is innate . Se cond, cultural learning p osits that the univ ersal
initial form of the moral mind is complete d and shap e d by cultural influences. Thir dly ,
intuitionism suggests that moral judgments ar e base d on moral intuitions that arise quickly ,
sp ontane ously and asso ciativ ely . Fourth, pluralism asserts that ther e is mor e than one moral
foundation. Drawing on the mo del of so cial intuition, the authors argue that our moral
judgements ar e always intuitiv e first; rationalisation and r easoning come later and ser v e to
justify the original, intuitiv e moral judgements.
The car e/harm foundation is base d on our innate ability to car e for our offspring. In the
course of e v olution, w e hav e passe d this car e on to other memb ers of our so ciety . This foun-
dation is activate d when w e p er ceiv e that some one else is suffering and in ne e d; it r epr esents
our compassion for victims of psy chological or physical violence . The fairness/cheating
foundation is ab out r esp onding to p er ceiv e d unfair actions that may arise dir e ctly fr om
interaction with another p erson or thr ough obje cts or thir d parties. This foundation stands
for the maintenance of justice , e quality and trust. The lo yalty/b etrayal foundation stems
fr om e v olutionar y assumptions of co op eration and gr oup formation and comp etition with
other gr oups for sur vival. It r efers to our sacrifice , affiliation and supp ort for the gr oup
to which w e b elong. T o day , this foundation is found in lo yalty to athletes or teams and
lo yalty to brands (Graham et al., 2013 ). The authority/subv ersion foundation stems fr om
the e v olutionar y fact that w e mammals hav e always liv e d in some kind of so cial hierar chy .
In the past it was the alpha males, to day it is the laws, mo dern institutions ( e .g., the p olice)
authority figur es ( e .g., leaders, teachers, par ents). This foundation is linke d to r esp e ct and
submission to tradition ( e .g., Graham et al., 2013 ). Sanctity/degradation is base d on the
moral emotion of disgust, which had a sur vival function in e v olution but no w accompanies,
for e xample , our r eactions to immigrants and se xual de viance . Attempts hav e b e en made to
add other foundations (fr e e dom/r epr ession Iy er et al., 2012 and dividing fairness foundation
into e quality and pr op ortionality Atari et al., 2023 , but only the original fiv e foundations
can b e consider e d univ ersal. Car e/harm and fairness/cheating r epr esent moral b ehaviours
that giv e mor e w eight to individual rights, wher eas lo yalty/b etrayal, authority/subv ersion
and sanctity/degradation giv e mor e w eight to gr oup rights Graham et al., 2013 , r epr esenting
b oth individual-oriente d morality , i.e ., our moral actions in dir e ct interactions with other
individuals, and gr oup-oriente d morality , i.e ., our moral actions within the so cial gr oup in
which w e liv e . Haidt, 2008 has e xtende d morality to the gr oup le v el, arguing that morality is
not only ab out ho w w e b ehav e to war ds each other , but also ab out gr oup cohesion, supp ort
for basic institutions and living a holy and noble life .

Resear ching moral foundations using the Bay esian appr oach 79
1.2. Unv eiling the r ole of so cio demographic variables in shaping moral foundations
Numer ous studies hav e sho wn that ther e is an imp ortant interplay b etw e en so cio demo-
graphic variables and the endorsement of moral foundations. Graham et al. ( 2011 ) r ep orte d
that in a sample of 49 228 w omen and 68 812 men, w omen rate d the foundations of car e ,
fairness and sanctity higher , while men rate d the foundations of lo yalty and authority higher ,
while Miles ( 2014 ) r ep orte d that w omen value d car e and sanctity mor e than men. The r esults
of a meta-analysis (Jaffe e & Hy de , 2000 ) sho w e d that w omen value d only the foundation of
car e mor e . Resear ch has also sho wn a significant r elationship b etw e en moral foundations
and age . In a study inv olving memb ers of thr e e differ ent generations, oldest generation rate
all fiv e moral foundations higher than the tw o y ounger generations (Friesen, 2019 ). All fiv e
moral foundations w er e found to b e p ositiv ely corr elate d with age (Sağel, 2015 ). Mor e o v er ,
older individuals tend to assign higher rating to the foundations of justice , lo yalty , authority
and sanctity higher in comparison to y ounger (Miles, 2014 ). These findings colle ctiv ely
indicate a p ositiv e asso ciation b etw e en age and moral foundations.
Compar e d to gender and age , e ducation is much less r esear che d, and the r esults ar e
much mor e mixe d. This is mainly b e cause most samples consist of students, which makes it
imp ossible to study the effe cts of differ ent le v els of e ducation on the acceptance of moral
foundations. Ne v ertheless, some authors hav e addr esse d this issue . Foundation of car e
demonstrates a p ositiv e r elationship with e ducation, as do es fairness (Efferson et al., 2017 ;
van Le euw en et al., 2014 ). Findings r egar ding the foundation of authority e xhibit mixe d
pattern: individual with lo w er e ducation tend to endorse the foundation of authority mor e
str ongly (Er ceg et al., 2018 ), while concurr ently , individuals with higher e ducation endorse
authority mor e str ongly (Efferson et al., 2017 ). Similarly , lo yalty demonstrates contradictor y
findings; some studies r ep ort a p ositiv e r elationship (Efferson et al., 2017 ), wher eas others
indicate a negativ e r elationship (van Le euw en et al., 2014 ).
The moral foundations w er e analyse d in differ ent cultur es. Participants fr om Eastern
cultur es (East A sia, South A sia) rate d lo yalty and sanctity higher than participants fr om
W estern cultur es (Graham et al., 2011 ), who rate d car e and fairness higher . This can
b e e xplaine d thr ough individualism—colle ctivism distinctions and cultur e orientations
(T riandis, 2001 ). Moral Foundations The or y was founde d in cultural psy chology to e xplor e
cultural differ ences in moral orientations, but quickly pr o v e d useful for e xploring the
differ ences b etw e en lib erals and conser vativ es in the Unite d States. Numer ous studies with
differ ent samples and using differ ent metho ds hav e sho wn that lib erals ar e mor e likely to
accept the car e and fairness foundations compar e d to the other thr e e foundations than
conser vativ es, who value all fiv e foundations similarly ( e .g., Do dd et al., 2012 ; Haidt et al.,
2009 ). McA dams et al. ( 2008 ) also conducte d a qualitativ e study in which the y found that
when defining their moral principles, conser vativ es talke d ab out r esp e cting authority and
or der , sho wing lo yalty to family and countr y , and acting in ways that ke ep oneself pur e
and go o d, while lib erals talke d ab out their moral obligations primarily in terms of not
harming others and caring ab out fairness, justice , and e quality . A similar pattern was found
among r eligious p e ople , who w er e mor e likely than non-r eligious p e ople to r efer to the
moral foundations of lo yalty , authority and sanctity (Haidt et al., 2009 ). Moral foundations
the or y states that p olitical attitudes ar e the r esult of the acceptance of moral foundations,
but empirical e vidence is insufficient. The question r emains whether moral b eliefs lead to a
particular p olitical orientation or whether the r elationship is inv erse or r e cipr o cal. Hatemi
et al. ( 2019 ) pr o vide d one answ er to this question. Using a cr oss-lagge d panel analysis for
thr e e differ ent samples, the y conclude d that p olitical ide ology pr e dicts moral foundations,
which could mean that moral foundations r epr esents some situational judgements that

80 Rakinić
justify pr e-e xisting ide ological b eliefs, rather than the other way r ound.
1.3. Moral foundations as basis for attitudes formulation
The acceptance of moral foundations corr elates with various psy chological traits and b e-
haviours, including emotions (Horb erg et al., 2009 ), moral identity (Dawson et al., 2021 ) and
e v en r e cr eational drug use (Kurzban et al., 2010 ). In our study , w e fo cus on attitudes, which
denote a p ositiv e or negativ e e valuation of a particular obje ct and ar e often base d on moral
intuitions. A ccor ding to moral foundation the or y , moral judgements ar e rapid, effortless
and intuitiv e and as such hav e an impact on our further e valuations. K ole va et al. ( 2012 )
demonstrate d the conne ction b etw e en moral foundations and attitudes to war d differ ent
pr essing so cial issues, such as same-se x marriage , ab ortion etc. T o b e mor e pr e cise , higher
ratings of car e and fairness moral foundations pr e dicte d mor e fav ourable attitudes to war ds
homose xuals, while the foundation of sanctity inv ersely influence d attitudes (Rosik et al.,
2013 ). Similarly , p ositiv e attitudes to war ds the p o or w er e asso ciate d with higher ratings
of car e and fairness, while negativ e attitudes w er e asso ciate d with the r emaining thr e e
foundations (Lo w & W ui, 2015 ). Lo yalty , authority , and sanctity w er e p ositiv ely asso ciate d
with gr oup hostility and discriminator y tendencies (Kugler et al., 2014 ). Our study , ho w e v er ,
delv es into attitudes to war ds homeless p e ople . W e sele cte d this gr oup as the fo cal p oint
of our r esear ch due to its widespr ead familiarity and the appar ent absence of r esear ch
concerning attitudes to war d homelessness fr om the p ersp e ctiv e of moral foundations.
1.4. Bay esian appr oach to analysing data
In the quantitativ e so cial sciences, the statistical infer ence appr oach dominates, and despite
many advantages and incr easing calls for its use , the Bay esian appr oach is still r elativ ely
uncommon in the so cial sciences (van de Scho ot et al., 2017 ). Bay esian statistics has
many advantages o v er “ classical” statistics. It allo ws the analysis of smaller samples, prior
kno wle dge can b e include d in the analyses, and it allo ws the estimation of a far mor e comple x
mo dels (W agenmakers, 2007 ). Hyp othesis testing with Bay es factors (BFs) is incr easingly
pr op ose d as a concr ete and practical alternativ e to hyp othesis testing with p values (Jeffr e ys,
1961 ; K ass & Rafter y , 1995 ). Hyp othesis testing with Bay es factors compar es the pr e dictiv e
p o w er of tw o comp eting statistical mo dels, ranking the e vidence pr o vide d by the data on
a continuous scale and quantifying the change in b elief that the data pr o duce for the tw o
mo dels under consideration. Bay es factors hav e se v eral practical advantages: the y allo w
quantification of e vidence , and this e vidence can b e continuously monitor e d as the data
accumulate (Rouder , 2014 ; W agenmakers, 2007 ). Fr e quentist hyp othesis testing allo ws for
only tw o outcomes. The null hyp othesis may or may not b e r eje cte d, which do es not mean
it is accepte d, wher eas Bay esian statistics can b e use d to assess the r elativ e plausibility
of one or the other hyp othesis. Ho w e v er , the use of Bay esian metho ds comes with some
costs. The first r elates to the question of prior distributions. Sometimes de ciding which
prior distributions to use can b e difficult (Faulkenb err y et al., 2020 ). Bay esian metho ds
ar e available in sp e cialise d statistical pr ogrammes such as R or JASP and may r e quir e
additional (pr ogramming) kno wle dge (Baldwin & Larson, 2017 ). Bay esian analysis do es not
automatically lead to corr e ct or b etter estimates. W e can just as easily mislead ourselv es with
Bay esian metho ds as with other te chniques ( e .g., playing ar ound with prior distributions
until w e get the r esult w e want or that can b e publishe d; Simmons et al., 2011 ). Ne v ertheless,
w e hav e chosen to pr esent an e xample of a study that uses the Bay esian appr oach to pr omote
kno wle dge in the so cial sciences that differ ent appr oaches to analysing data ar e p ossible
and valid.

Resear ching moral foundations using the Bay esian appr oach 81
1.5. Aims of the study
This study has tw o main obje ctiv es. Firstly , w e aim to assess the applicability of the
moral foundations the or y in the Slo v enian conte xt. Se condly , w e se ek to demonstrate the
application of Bay esian statistical appr oach within the r ealm of so cial sciences. Our aim is
to determine whether the structur e of moral foundations identifie d in pr e vious r esear ch can
also b e r e cognise d in the Slo v enian cultural envir onment. Furthermor e , majority r esear ch
has sho wn that so cio demographic variables play an imp ortant r ole in the endorsement of
moral foundations. W e aim to inv estigate this within the Slo v enian sample , sp e cifically
e xamining which so cio demographic variables influence moral foundations endorsement and
in what way . In addition, w e want to e xplor e the r elationship b etw e en moral foundations
and attitudes, as w e ar e inter este d in whether and ho w endorsement of differ ent moral
foundations pr e dicts attitudes to war ds homeless p e ople . A s a se condar y obje ctiv e , w e
aim to demonstrate the practical application of Bay easian appr oach in so cial science using
a r elativ ely simple e xample . In doing so , w e aim to illustrate ho w differ ent statistical
appr oaches can b e emplo y e d to deriv e meaningful and informativ e conclusions.
2. Metho ds
2.1. Study design
W e conducte d a quantitativ e , non-e xp erimental cr oss-se ctional study base d on a question-
nair e metho d. The questionnair e containe d questions on so cio demographic data, vignettes
on moral foundations and a scale on attitudes to war ds homeless p e ople .
2.1.1. Moral Foundations Vignettes. The purp ose of the Moral Foundations Vignettes (Cliffor d
et al., 2015 ) is to measur e endorsements of moral foundations. Each vignette r epr esents a
b ehaviour that violates a particular moral foundation. Vignettes b egin with the description
“Y ou se e […]” and continues with “[…] an emplo y e e lying ab out ho w many hours she
w orke d during the w e ek” ( violation of the fairness/cheating foundation). Participants
answ er on a fiv e-p oint scale ranging fr om 1 ( not at all wr ong ) to 5 ( e xtr emely wr ong ).
The original scale was translate d fr om English using a back translation pr o cess by tw o
indep endent translators. A final v ersion was formulate d base d on mutual agr e ement. The
questionnair e also measur es the adde d sixth moral foundation of lib erty/r epr ession. The
six-factor structur e of the questionnair e was confirme d in our sample , χ
2
(153) = 739.94 ,
𝑝 < 0.001 , TLI = 0.91 , CFI = 0.88 , RMSEA = 0.05 , SRMR = 0.06 . The ω r eliability
co efficient for the lib erty/oppr ession subscale was 0.40 , so w e de cide d to include only the
other fiv e moral foundations in the analysis. The ω co efficients for the other fiv e scales
range d fr om 0.59 to 0.74 . W e confirme d the ne w fiv e-factor structur e of the questionnair e ,
χ
2
(105) = 618.26 , 𝑝 < 0.001 , TLI = 0.91 , CFI = 0.89 , RMSEA = 0.07 , SRMR = 0.06 .
2.1.2. Sur v e y of Attitudes T o war d Homeless Pe ople . Sur v e y of Attitudes T o war d Homeless
Pe ople (Sno w-Hill, 2019 ) measur es negativ e attitudes to war ds homeless p e ople on a six-p oint
scale ranging fr om 1 ( str ongly disagr e e ) to 6 ( str ongly agr e e ), with higher o v erall mean scor es
indicating mor e negativ e attitudes to war ds homeless p e ople . An e xample of an item is
“Homeless p e ople ar e lazy . ” Sno w-Hill ( 2019 ) r ep orte d high internal r eliability and construct
validity of the questionnair e . The original scale was translate d fr om English using a back
translation pr o cess by tw o indep endent translators. A final v ersion was formulate d base d on
mutual agr e ement. The one-factor structur e of the scale was also confirme d in our sample ,
χ
2
(27) = 52.61 , 𝑝 < 0.001 , TLI = 0.95 , CFI = 0.93 , RMSEA = 0.08 , SRMR = 0.05 . The ω
r eliability co efficient was 0.89 in our sample .

82 Rakinić
2.1.3. So cio demographic variables. The participants indicate d their le v el of e ducation, which
range d fr om primar y scho ol to a do ctorate . Political orientation was measur e d on an 11-
p oint scale , with a scor e of 0 indicating a left-wing p olitical orientation and a scor e of
11 indicating a right-wing p olitical orientation. Participants w er e also aske d ab out their
r eligious affiliation, with the options of cho osing a sp e cific r eligion ( e .g., Catholic, Je wish,
Islamic, Buddhist, etc.), stating no r eligious affiliation or sele cting “ other” as an alternativ e .
The question was r e co de d for the purp ose of analysis, wher e a value of 0 indicates not b eing
r eligious and a value of 1 indicates b eing r eligious.
2.2. Data colle ction
The data colle ction to ok place in April and May 2023. The questionnair e was made available
thr ough the 1KA online application (Centr e for So cial Informatics, 2023 ). Participants
r e ceiv e d a link to the questionnair e in the invitation to participate in the sur v e y . At the end
of the questionnair e , each participant was able to leav e their contact details ( email addr ess)
if the y wishe d to b e enter e d into a prize draw for a change to win a cinema ticket to se e the
film of their choice . Each participant was guarante e d anonymity of their data. The data is
pr esente d in anonymise d form and only at gr oup le v el.
2.3. Sampling and sample description
The conv enience sampling metho d was use d. Individuals who w er e of legal age w er e invite d
to participate . The invitation to participate in the study was sent thr ough the r esear cher’s
p ersonal netw ork and p oste d on various online forums.
The study inv olv e d 162 participants, most of whom w er e female ( 69 % ). Fiv e participants
did not want to indicate their gender and chose the option “ other” . The participants’ av erage
age was 36 y ears ( SD = 15 , min = 18 , max = 84 ). The sample was dominate d by students
( 30.2 % ), follo w e d by emplo y e es in the public se ctor ( 25.9 % ) and in private companies ( 25.3 % ),
with 7.4 % r etir e d and less than 4 % of participants in the other categor y (unemplo y e d, other ,
self-emplo y e d). Most participants in the sample had a four-y ear finishe d se condar y scho ol
( 37 % ) and a univ ersity degr e e ( 33.3 % ). Fiv e of the sample had a P h.D ., while four participants
had complete d primar y scho ol or less. Most participants gr e w up in a rural place , a village
( 41.4 % ), follo w e d by participants in a large ( 24.1 % ) or small to wn ( 23.5 % ). Slightly mor e
than half of the participants in the sample r ep orte d not b elonging to any r eligion ( 54.6 % ),
while 35.8 % identifie d themselv es as Catholic.
2.4. Data analysis
Data w er e analyse d using SPSS (IBM Corp oration, 2020 ), R (R Cor e T eam, 2022 ) and JASP
(JASP T eam, 2023 ). First, descriptiv e statistics w er e calculate d for the variables include d
in the analysis. Further w e p erform Bay esian pair e d samples t-test, corr elation analysis
and linear r egr ession. W e use d noninformativ e priors due to lack of solid kno wle dge of
inv estigate d phenomena in Slo v enian conte xt. In our interpr etation of the r esults, w e
adher e d to Jeffr e ys’s ( 1939 ) guidelines, wher e a Bay es factor b etw e en 1 and 3 signifies
w eak e vidence , b etw e en 3 and 10 indicates mo derate e vidence , ab o v e 10 r epr esents str ong
e vidence , and b e y ond 30 suggests v er y str ong e vidence . These classifications ar e use d only
as general rules to facilitate communication and interpr etation of the str ength of e vidence .
Inde e d, one of the advantages of the Bay es factor is that it allo ws the assessment of e vidence
on a continuous scale (van Do orn et al., 2021 ).

Resear ching moral foundations using the Bay esian appr oach 83
3. Results
3.1. Descriptiv e statistics
On av erage , participants rate d as the most morally wr ong the statements r elating to the
foundation of car e/harm (the b o y thr o wing stones at the co ws) and fairness/cheating (the
r efer e e making unfair de cisions in fav our of his fav our e d team). On av erage , participants
found the statement that r eferr e d to the foundation of lo yalty/b etrayal, namely that the
wife helps her husband’s opp osing team, to b e the least morally wr ong (T able 1 ).
T able 1. Descriptiv e statistics of Moral Foundations Vignettes
Factor Y ou se e […] Min Max M (SD )
Car e […] a b o y telling a w oman that she lo oks just like her
o v er w eight bulldog.
1 5 4.34 ( 0.84 )
[…] a b o y thr o wing r o cks at co ws that ar e grazing in
the lo cal pastur e .
2 5 4.72 ( 0.57 )
[…] w oman spanking her child with a spatula for getting
bad grades in scho ol.
1 5 4.29 ( 1.07 )
Fairness […] a runner taking a shortcut on the course during the
marathon in or der to win.
2 5 4.45 ( 0.80 )
[…] a r efer e e intentionally making bad calls that help
his fav our e d team win.
1 5 4.72 ( 0.62 )
[…] an emplo y e e lying ab out ho w many hours she
w orke d during the w e ek.
1 5 3.93 ( 1.09 )
Lo yalty […] a former Army General fr om y our countr y saying
publicly he w ould ne v er buy any of y our countr y’s pr o d-
ucts.
1 5 3.05 ( 1.41 )
[…] a head che erleader b o oing her high scho ol’s team
during a home coming game .
1 5 3.28 ( 1.28 )
[…] the coach’s wife sp onsoring a bake sale for her
husband’s rival team.
1 5 2.93 ( 1.37 )
A uthority […] a te enage girl coming home late and ignoring her
par ents’ strict curfe w .
1 5 3.14 ( 1.16 )
[…] a staff memb er talking loudly and interrupting the
may or’s sp e e ch to the public.
1 5 3.94 ( 0.98 )
[…] a man turns his back and walk away while his b oss
questions his w ork.
1 5 3.43 ( 1.16 )
Sanctity […] a drunk elderly man offering to hav e oral se x with
any one in the bar .
1 5 4.14 ( 1.23 )
[…] a man in a bar using his phone to watch p e ople
having se x with animals.
1 5 4.43 ( 1.01 )
[…] a stor y ab out a r emote trib e eating the flesh of their
de cease d memb ers.
1 5 3.10 ( 1.40 )
On av erage , participants rate d the most morally wr ong actions e xpr essing the moral
foundation of car e/harm ( M = 4.45 ; SD = 0.59 ) and fairness/cheating ( M = 4.36 ; SD = 0.66 ).

84 Rakinić
In contrast, actions e xpr essing the foundation of lo yalty/b etrayal ( M = 3.09 ; SD = 1.10 ) w er e
rate d as least morally wr ong, follo w e d by the foundation of authority/subv ersion ( M = 3.51 ;
SD = 0.86 ), and the foundation of sanctity/degradation ( M = 3.89 ; SD = 0.92 ) b eing in the
middle . W e teste d for differ ences b etw e en all ten pairs of moral foundations. Bay esian
pair e d samples 𝑡 tests ( with Cauchy prior 𝑟 = 0.71 ), sho w str ong supp ort ( BF
10
> 100 ) for
differ ences b etw e en nine pairs of moral foundations (Figur e 1 ), for e xample , participants
endorse the car e foundation mor e than the authority foundation. Ther e was no supp ort for
the differ ence b etw e en car e and fairness foundations, indicating that participants in our
sample value these tw o foundations e qually .
1.0
1.5
2.0
2.5
3.0
3.5
4.0
4.5
5.0
5.5
Care Fairness Loyalty Authority Sanctity
Score M (SD)
Moral foundations
Figur e 1. Endorsement of moral foundations pr esente d as M (SD )
3.2. So cio demographic variables as pr e dictors of moral foundations
T able 2 sho ws the corr elations (using a noninformativ e prior; any value b etw e en −1 and 1 is
e qually likely ) b etw e en so cio demographic variables and moral foundations. The str ongest
e vidence for a corr elation is found b etw e en female gender and car e and age and fairness,
authority and lo yalty .
T able 2. Corr elations b etw e en so cio demographic variables and moral foundations
Car e Fairness A uthority Lo yalty Sanctity
V ariable 𝑟 BF
10
𝑟 BF
10
𝑟 BF
10
𝑟 BF
10
𝑟 BF
10
Gen – 0.30 168.35 – 0.02 0.10 – 0.17 0.93 – 0.10 0.22 – 0.02 0.10
A ge 0.23 5.99 0.40 89 360.80 0.38 25 675.15 0.31 307.89 0.10 0.20
Edu – 0.13 0.37 – 0.11 0.24 0.01 0.10 – 0.11 0.25 – 0.19 1.85
Rel 0.01 0.10 0.14 0.51 0.21 3.20 0.24 10.12 0.20 2.57
PolOr – 0.14 0.47 0.21 3.66 0.15 0.29 0.06 0.17 0.09 0.17
Legend : Edu = Education Le v el, Gen = Gender , PolOr = Political Orientation, Rel = Religiosity .
Notes : Baseline le v els for categorical variables ar e Female ( Gen), Mor e right-wing orientation (PolOr ), and Not
b eing r eligious (Rel). Se e te xt for further details.
Further , w e p erforme d Bay esian linear r egr ession, to understand the r ole of imp ortant
so cio demographic pr e dictors in moral foundations. W e use d a uniform prior mo del, which
assumes that each mo del has the same prior pr obability , r egar dless of the numb er of variables

Resear ching moral foundations using the Bay esian appr oach 85
include d in the mo del. W e also use d the Jeffr e ys-Zellner-Sio w ( JZS) prior distribution to
estimate the r egr ession co efficients (Liang et al., 2008 ; Zellner & Sio w , 1980 ). The sele cte d
mo dels w er e also teste d for conv ergence . W e assess the
̂ 𝑅 values and the effe ctiv e sample
size . The
̂ 𝑅 values w er e 1 sho wing that the chains conv erge d to the same p osterior (Gelman
& Hill, 2007 ). Furthermor e , the effe ctiv e sample sizes (ESSs) w er e ab o v e 5000 , sho wing that
our estimates ar e stable (Baldwin & Larson, 2017 ).
3.2.1. Car e . The mo del with the variables gender , age , e ducation and p olitical orientation
pr o v e d to b e the most plausible . The r esults sho w e d that this mo del could e xplain the data
12 times b etter than the empty mo del, these pr e dictors e xplaine d 22 % of the variance in the
foundation. The BF
10
scor e was ab o v e 100, indicating str ong supp ort for this mo del. W e
can b e 95 % confident that the true value of car e foundation de cr eases b etw e en −0.57 and
−0.21 ( M = −0.41 ) for men, incr eases b etw e en 0.01 and 0.02 ( M = 0.01 ) for e v er y one unit
incr ease in age , de cr eases b etw e en −0.10 and 0.00 ( M = −0.05 ) for e v er y one unit incr ease
in e ducation, and de cr eases b etw e en −0.07 and 0.00 ( M = −0.02 ) for e v er y one unit incr ease
in mor e right-wing p olitical orientation. Gender and age sho w v er y str ong supp ort for
inclusion in the mo del ( BF
Inc
> 100 ), while e ducation ( BF
Inc
= 2.96 ) and p olitical orientation
(BF
Inc
= 1.22 ) sho w rather w eak supp ort for inclusion in the mo del.
3.2.2. Fairness. The mo del that pr o v e d to b e the most plausible in pr e dicting fairness foun-
dation containe d the pr e dictors age , e ducation and p olitical orientation. This mo del can
e xplain the data 13 b etter than the null mo del, e xplaining 22 % of the variance in the fairness
foundation. BF
10
> 100 , indicating str ong supp ort for this mo del. W e can b e 95 % confident
that the true value of the fairness foundation incr eases b etw e en 0.01 and 0.02 ( M = 0.02 )
with an incr ease of one unit in age , de cr eases by b etw e en −0.12 and 0.00 ( M = −0.06 ) for
an incr ease of one unit in e ducation, and incr eases by b etw e en 0.00 and 0.09 ( M = 0.04 ) for
each incr ease in a mor e right-wing p olitical r egime . The age variable sho ws v er y str ong
supp ort for inclusion in the mo del ( BF
Inc
> 100 ), while the e ducation ( BF
Inc
= 4.30 ) sho ws
mo derate and p olitical orientation (BF
Inc
= 2.48 ) w e ek supp ort for inclusion in the mo del.
3.2.3. A uthority . The mo del that pr o v e d the most plausible in pr e dicting the authority
foundation containe d pr e dictors gender , age and r eligiosity . This mo del can e xplain the data
27 times b etter than the null mo del, e xplaining 23 % of the variance in the fairness foundation.
BF
10
> 100 , indicating str ong supp ort for this mo del. W e can b e 95 % confident that the true
value of the authority foundation de cr eases b etw e en −0.59 and 0.00 ( M = −0.35 ) for males,
de cr eases by b etw e en −0.10 and 0.01 ( M = −0.01 ) for a one unit incr ease in e ducation and
incr ease b etw e en 0.00 and 0.52 ( M = 0.28 ) for r eligious individuals. The age sho ws v er y
str ong supp ort for inclusion in the mo del ( BF
Inc
> 100 ), gender ( BF
Inc
= 12.08 ) str ong and
r eligiosity mo derate (BF
Inc
= 7.00 ) supp ort for inclusion in the mo del.
3.2.4. Lo yalty . The mo del that pr o v e d the most plausible in pr e dicting the lo yalty include d
pr e dictors for gender , age , e ducation and r eligiosity . This mo del can e xplain the data 15
times b etter than the null mo del, e xplaining 19 % of the variance in the fairness foundation,
BF
10
> 100 , indicating str ong supp ort for this mo del. W e can b e 95 % confident that the true
value of the fairness foundation de cr eases b etw e en −0.57 and 0.01 ( M = −0.19 ) for males,
incr eases b etw e en 0.01 and 0.03 ( M = 0.02 ) for each y ear mor e of age , de cr eases by b etw e en
−0.20 and 0.00 ( M = −0.09 ) for a one-unit incr ease in e ducation, and incr eases by b etw e en
0.00 and 0.69 (M = 0.41 ) for the r eligious. The age sho ws v er y str ong supp ort for inclusion
in the mo del ( BF
Inc
> 100 ), r eligiosity ( BF
Inc
= 12.38 ) str ong e vidence , e ducation sho ws
mo derate supp ort ( BF
Inc
= 3.49 ), and the gender w e ek ( BF
Inc
= 1.49 ) supp ort for inclusion

86 Rakinić
in the mo del.
3.2.5. Sanctity . The mo del that pr o v e d to the most plausible in pr e dicting sanctity containe d
the e ducation and r eligiosity . The mo del pr e dicte d the data nine times b etter compar e d to
the null mo del and e xplaine d 9 % of the variance . BF
1
= 27.58 sho ws str ong supp ort for
this mo del compar e d to the null mo del. W e can b e 95 % confident that the true value of the
sanctity de cr eases b etw e en −0.21 and 0.00 ( M = −0.12 ) as e ducation incr eases by one unit
and incr eases b etw e en −0.01 and 0.51 ( M = 0.18 ) for r eligious individuals. The pr e dictor of
e ducation sho ws str ong supp ort for inclusion in the mo del ( BF
Inc
= 20.27 ), as r eligiosity
sho ws w e ek supp ort (BF
Inc
= 1.83 ).
T able 3 summarises the r esults of the Bay esian linear r egr ession. In the Slo v enian
conte xt, so cio demographic variables such as female gender , older age and lo w er e ducation
sho w r obust e vidence supp orting their inclusion as pr e dictors of endorsement of moral
foundations. In addition, r eligiosity pr o v es to b e a p ositiv e pr e dictor of binding moral
foundations. The so cio demographic variables consider e d together account for a considerable
amount of the variance in moral foundations, e xplaining b etw e en 9 % and 23 % of the
obser v e d variability .
T able 3. Summar y of Bay esian r egr ession r esults sho wing e vidence of so cio demographic
variables as pr e dictors
V ariable Car e Fairness A uthority Lo yalty Sanctity
Gen ++++ −− −
A ge ++++ ++++ ++++ ++++
Edu − −− −− −−−−
PolOr + +
Rel ++ +++ +
Legend : Edu = Education Le v el, Gen = Gender , PolOr = Political Orientation, Rel = Religiosity .
Notes : Baseline le v els for categorical variables ar e Female ( Gen), Mor e right-wing orientation (PolOr ),
and Not b eing r eligious (Rel). The numb er of signs r epr esents the str ength of e vidence for the
variable as a pr e dictor (ranging fr om w eak to v er y str ong). Positiv e signs ( + ) indicate a p ositiv e
pr e dictor , while negativ e signs ( − ) indicate a negativ e pr e dictor .
3.3. Attitudes to war ds homeless p e ople
Having first e xplor e d ho w so cio demographic variables pr e dict moral foundations, w e w er e
also inter este d in ho w moral foundations, as intuitiv e , quick judgments ab out right and
wr ong, might pr e dict mor e sp e cific judgments, namely attitudes to war ds homeless p e ople .
On av erage , participants in our sample e xpr esse d lo w negativ e attitudes to war d homeless
p e ople ( M = 2.53 , SD = 0.90 , min = 1 , max = 5.44 ). W e ran a linear Bay esian r egr ession. W e
use d a uniform prior mo del, which assumes that each mo del has the same prior pr obability ,
r egar dless of the numb er of variables include d in the mo del. W e also use d the JZS prior
distribution to estimate the r egr ession co efficients (Liang et al., 2008 ; Zellner & Sio w , 1980 ).
W e teste d the conv ergence of the mo del, which sho w e d that our r esults ar e stable (
̂ 𝑅 values
of pr e dictors ≅ 1 , ESS > 5000 ). The r esults sho w e d that the b est mo del was the one with
the pr e dictors gender , p olitical orientation, moral foundation car e and authority , which
e xplaine d 23 % of the variance in attitudes to war ds homeless p e ople ( BF
10
> 100 , T able 4 ).
Gender , p olitical orientation and authority sho w mo derate e vidence of supp ort for inclusion

Resear ching moral foundations using the Bay esian appr oach 87
as pr e dictors, while car e sho ws v er y str ong supp ort for inclusion ( BF
Inc
> 100 ). Male gender ,
right-wing p olitical orientation, and moral foundation authority pr e dicts a mor e negativ e
attitude to war ds the homeless, while moral foundations car e pr e dicts mor e p ositiv e attitudes
(T able 5 ).
T able 4. Mo del comparison of Bay esian r egr ession for pr e dicting attitudes to war ds homeless p e ople
Mo del 𝑃 ( M ) 𝑃 ( M | D ) BF
M
BF
10
𝑅 2
Null 9.8 × 10
−4
1.9 × 10
−4
1.9 × 10
−4
1.0 0.00
Gen + PolOr + C + A 9.8 × 10
−4
5.7 × 10
−2
6.2 × 10
1
3.0 × 10
5
0.23
PolOr + Gen + C + A + L 9.8 × 10
−4
4.9 × 10
−2
5.2 × 10
1
2.6 × 10
5
0.25
PolOr + Gen + S + C + A 9.8 × 10
−4
2.8 × 10
−2
3.0 × 10
1
1.5 × 10
5
0.24
PolOr + Gen + C + F + A + L 9.8 × 10
−4
2.8 × 10
−2
2.9 × 10
1
1.5 × 10
5
0.25
PolOr + Gen + C + F + A 9.8 × 10
−4
2.2 × 10
−2
2.3 × 10
1
1.2 × 10
5
0.24
Gen + C + A 9.8 × 10
−4
2.1 × 10
−2
2.2 × 10
1
1.1 × 10
5
0.21
PolOr + Gen + S + C + A + L 9.8 × 10
−4
1.8 × 10
−2
1.9 × 10
1
9.7 × 10
4
0.25
PolOr + A ge + Gen + C + A 9.8 × 10
−4
1.6 × 10
−2
1.7 × 10
1
8.6 × 10
4
0.23
PolOr + Edu + Gen + C + A + L 9.8 × 10
−4
1.6 × 10
−2
1.7 × 10
1
8.6 × 10
4
0.25
Legend : A = A uthority , C = Car e , F = Fairness, L = Lo yalty , S = Sanctity; Edu = Education Le v el, Gen = Gender ,
PolOr = Political Orientation, Rel = Religiosity .
Notes : Baseline le v els for categorical variables in the r egr ession mo dels ar e Female ( Gen), Mor e right-wing
orientation (PolOr ), and Not b eing r eligious (Rel). T able displays only a subset of 10 b est mo dels. Se e te xt for
further details.
T able 5. Posterior summaries of co efficients of Bay esian linear r egr ession
Co efficient 𝑃 ( Inc ) 𝑃 ( Exc ) 𝑃 ( Inc | D ) 𝑃 ( Exc | D ) BF
Inc
M (SD ) 95 % CI
Inter cept 1.000 0.000 1.000 0.000 1.00 2.55 ( 0.07 ) [ 2.42 , 2.68 ]
Gen 0.500 0.500 0.786 0.214 3.67 0.26 ( 0.19 ) [ 0.00 , 0.58 ]
A ge 0.500 0.500 0.236 0.764 0.31 −1.96 ( 0.00 ) [ −0.01 , 0.01 ]
Edu 0.500 0.500 0.254 0.746 0.34 0.01 ( 0.02 ) [ −0.03 , 0.07 ]
Rel 0.500 0.500 0.245 0.755 0.33 −0.01 ( 0.07 ) [ −0.24 , 0.14 ]
PolOr 0.500 0.500 0.786 0.214 3.67 0.06 ( 0.04 ) [ 0.00 , 0.13 ]
C 0.500 0.500 0.995 0.005 214.28 −0.50 ( 0.14 ) [ −0.77 , −0.21 ]
F 0.500 0.500 0.296 0.704 0.42 −0.03 ( 0.08 ) [ −0.30 , 0.04 ]
A 0.500 0.500 0.860 0.140 6.13 0.23 ( 0.14 ) [ 0.00 , 0.44 ]
L 0.500 0.500 0.547 0.453 1.21 0.08 ( 0.10 ) [ −0.01 , 0.28 ]
S 0.500 0.500 0.352 0.648 0.54 0.03 ( 0.07 ) [ −0.04 , 0.21 ]
Legend : A = A uthority , C = Car e , F = Fairness, L = Lo yalty , S = Sanctity; Edu = Education Le v el, Gen = Gender ,
PolOr = Political Orientation, Rel = Religiosity .
Notes : Baseline categories for variables in the r egr ession mo dels ar e female ( Gen), mor e right-wing orientation
(PolOr ), and not b eing r eligious (Rel). T able displays only a subset of 10 b est mo dels. Se e te xt for further details.

88 Rakinić
4. Discussion
Our study aime d to offer insights into the assessment of moral foundations base d on Moral
foundation the or y (Haidt & Joseph, 2004 ) in Slo v enian sample . A dditionally , w e se ek to
illustrate the application of Bay esian statistical appr oach. Moral foundations the or y emerge d
as a r esp onse to earlier the ories within moral psy chology that define d morality as either
justice (K ohlb erg, 1969 ) or car e (Gilligan, 1977 ). Moral foundations the or y e xpands our
understanding to fiv e foundations: car e , justice , lo yalty , authority and sanctity , with the
last thr e e r epr esenting gr oup-oriente d morality that tends to pr eser v e so ciety .
Participants in our study value d the foundations of car e and fairness higher than the
other thr e e foundations. This is consistent with r esear ch sho wing that memb ers of W estern
so cieties place gr eater imp ortance on the foundations r epr esenting individual-oriente d
morality ( e .g., Graham et al., 2011 ) due to a str onger orientation to war ds individualism
compar e d to Eastern so cieties. W e e xamine the r ole of so cio demographic variables in
pr e dicting the acceptance of a particular moral foundation. W e found v er y str ong supp ort
for the inclusion of gender as a pr e dictor of the car e foundation. Females value car e
foundation higher compar e d to men (Eisenb erg et al., 1991 ; Graham et al., 2011 ; Güner ,
2020 ). Gilligan ( 1977 ) was one of the first authors to argue that w omen r ely mor e heavily
on the idea of car e when making moral judgements and de cisions. Our r esults additionally
sho w that female gender pr e dicte d higher endorsement of authority and lo yalty . One r eason
for this could b e that w omen o v erall hav e a higher moral r esp onse than men (Mainier o et al.,
2008 ), so the y p er ceiv e all b ehaviour as mor e immoral, r esulting in higher endorsement of
these foundations.
Bay esian analysis sho w e d v er y str ong e vidence for the inclusion of age as a pr e dictor
of four foundations ( e xcept sanctity ). This is consistent with se v eral studies sho wing that
older p e ople hav e higher scor es on all moral foundations (Friesen, 2019). The link b etw e en
age and gr eater acceptance of moral foundations can b e e xplaine d thr ough de v elopmental-
psy chological lens. Older p e ople e xpr ess mor e concern for others and place mor e value
on adhering to so cial norms, r esp e cting authority and upholding traditions than y ounger
p e ople (Robinson, 2013 ). Y ounger individuals, esp e cially those in the transition to adultho o d,
ar e mor e pr one to r efle ct and tr y out differ ent options also r egar ding their identity formation
( Arnett, 2004). This op enness and inde cisiv eness may ther efor e also b e r efle cte d in their
lo w er acceptance of moral foundations, namely their higher tendency to question act as
immoral. Another p otential factor could b e attribute d to so cietal changes, as studies hav e
indicate d that y ounger generations (those b orn after 1990) e xhibit a gr eater tendency to war d
r elativistic thinking (Stein & Dawson- T unik, 2004 ). Conse quently , the y ar e less incline d to
categorize b ehaviours as immoral, r efle cting the influence of r elativism, one of the imp ortant
characteristic of contemp orar y p ostmo dern so ciety . W e found e vidence for the inclusion of
e ducation as negativ e pr e dictor of sanctity , lo yalty , fairness, and car e . Which in contrar y to
some studies sho wing the opp osite r elationship . W e turn again to the concept of r elativism
to e xplain this phenomenon. Mor e e ducate d individuals in our sample may rate actions as
less immoral due to r elativism. Higher le v els of r elativism ar e asso ciate d with higher le v els
of e ducation (Perr y , 1970 ), which may lead to higher le v els of moral r elativism. When p e ople
hav e higher le v els of moral r elativism, the y incr easingly b elie v e that certain b ehaviours
cannot b e define d as immoral a priori, but instead situational factors ar e consider e d mor e
when making judgements (Sulsky et al., 2015 ).
Moral foundation the or y in the conte xt of American so ciety was dominantly use d to
e xplain the differ ence b etw e en lib erals and conser vativ es. In the conte xt of Slo v enian
sample , p olitical orientation was not r e cognise d as str ong pr e dictor of moral foundations.

Resear ching moral foundations using the Bay esian appr oach 89
The e xpr ession of a mor e right-wing p olitical orientation in the Slo v enian sample did not
pr e dict a gr eater appr e ciation of authority , lo yalty , and sanctity , as e xp e cte d, but it did
pr e dict car e and fairness, although sho wing only a w e ek supp ort. It has b e en argue d
that the the or y of moral foundations fo cuses to o much on the p oints of disagr e ement
b etw e en lib erals and conser vativ es in US (Frimer et al., 2013 ), which calls into question its
generalisability of the the or y acr oss cultur es. The left-right sp e ctrum in Slo v enia do es not
neatly align with the lib eral-conser vativ e dichotomy in American so ciety . Conse quently ,
the application of the moral foundations the or y to e xplain differ ences in p olitical orientation
in the Slo v enian conte xt is not straightfor war d. Bay esian analysis sho w e d e vidence for the
inclusion of b eing r eligious in the pr e diction of lo yalty , authority , and sanctity foundation.
All thr e e foundations r efle ct gr oup-oriente d morality (Graham et al., 2011 ). Religious
individuals typically affiliate with a sp e cific r eligious community , which can influence their
p er ception of certain actions, esp e cially r egar ding b etrayal of the gr oup , as mor e morally
r epr ehensible compar e d to non-r eligious individuals. Religion, as Durkheim argue d, is a
so cial phenomenon r o ote d in the colle ctiv e actions and b eliefs of a community . Religions
ar e base d on colle ctiv e b eliefs and rituals that unite the gr oup . These colle ctiv e asp e cts of
r eligion not only shap e the definition of morality but also help e xplain the internalisation
of moral norms (Bader & Finke , 2010 ).
Bay esian analysis sho w e d that the b est mo del ( BF
10
> 100 ) for pr e dicting attitudes
to war ds homeless p e ople was the mo del with the pr e dictors gender , p olitical orientation and
moral foundation of car e and authority . Both moral foundations sho w v er y str ong e vidence
for inclusion in the mo del, higher than gender and p olitical orientation. The car e foundation
r efle cts our compassion and empathy to war ds victims and activates a helping state in us
( e .g., Haidt et al., 2009 ), which can influence our p er ception of homeless p e ople who ne e d
help and car e due to their situation and conse quently lead to less negativ e attitudes. The
higher w e rate the authority foundation, the mor e negativ e our attitudes to war ds homeless
p e ople ar e . One e xplanation for this could b e that homeless p e ople ar e p er ceiv e d as a gr oup
that violates public and/or or der , commits criminal acts and do es not abide by the rules.
Which ar e in fact, principles that constitute the moral foundation of authority (Haidt et al.,
2009 ). Our r esults sho w that at least tw o moral foundations ar e imp ortant pr e dictors of
p e ople ’s attitudes. Since moral foundations ar e mor e fundamental than sp e cific attitudes,
the y should b e e xplor e d when studying the acceptance of various attitudes to war d differ ent
so cial gr oups.
The study has a primar y limitation due to the use of conv enience sampling metho d,
which means our sample lacks r epr esentativ eness, hindering generalisability . A further
limitation r efers to the use of Moral Foundation Vignettes, namely our study sho ws supp ort
for the fiv e-factor structur e , as the foundation of Lib erty was not include d as it sho w e d
lo w r eliability co efficient. W e did not test other asp e cts of validity , such as conv ergent,
discriminant validity of the questionnair e , which is a limitation w orth p ointing out. Also ,
metric invariance of the questionnair e should also b e e xamine d in the futur e in at least
r egar ding gender and age gr oups. Our r esults should b e interpr ete d with this in mind.
Regar ding futur e r esear ch dir e ction, b eside e xamining the validity of the questionnair e
mor e pr e cisely , w e b elie v e that it w ould b e inter esting to e xplor e the p ossible r ole of moral
r elativism (Sulsky et al., 2015 ) in the r elationship b etw e en moral foundations and e ducation,
as our r esults sho w e d a negativ e r elationship b etw e en four foundation and e ducation,
indicating that participants with higher e ducation, endorse those four foundations to a
lesser e xtent, which is contrar y to some pr e vious studies ( e .g., Efferson et al., 2017 ). It w ould
also b e inter esting to e xamine the r elationship b etw e en moral foundations and attitudes

90 Rakinić
to war ds differ ent attitude obje cts ( e .g., homose xuals, drug addicts, immigrants, etc.) as
differ ent obje cts can trigger differ ent foundations ( e .g., K ole va et al., 2012 ).
Finally , one of the main aims of this study was also to apply Bay esian appr oach, which
is a differ ent, mor e intuitiv e appr oach to hyp othesis testing and allo ws differ ent conclusions
to b e drawn. Despite its many advantages o v er the fr e quentist appr oach and the gr o wing
adv o cacy for its utilization, the Bay esian appr oach r emains r elativ ely underutilize d in the
so cial sciences (van de Scho ot et al., 2017 ). One p otential e xplanation for this limite d
adoption could b e that Bay esian statistics is rar ely taught in applie d statistics courses.
Conse quently , many r esear chers may lack kno wle dge and confidence in their ability to
apply these metho ds to their r esear ch inquiries (Faulkenb err y et al., 2020 ). The authors hop e
that in the futur e , so cial scientists will embrace the Bay esian appr oach, ther eby enriching
the array of metho ds emplo y e d in so cial science .
O v erall, our r esults sho w e d that so cio demographic variables pr e dicte d a quarter of the
variance in the four foundations ( e xcept sanctity ), demonstrating that the y ar e an imp ortant
factor in understanding and e xplaining the endorsement of our foundational b eliefs ab out
right and wr ong. Furthermor e , our study she ds light on the fact that the acceptance of
certain moral foundations also influences attitudes. Our study sho ws that the the or y of
moral foundations is applicable to the Slo v enian conte xt and r epr esents an inter esting ar ea
of r esear ch.
Funding
This w ork was supp orte d by the Slo v enian Resear ch and Inno vation A gency under the
Y oung Resear chers Pr ogram ( Grant No . P5-0168).
Refer ences
Atari, M., Haidt, J., Graham, J., K ole va, S., Ste v ens, S. T ., & Dehghani, M. (2023). Morality
b e y ond the WEIRD: Ho w the nomological netw ork of morality varies acr oss cultur es.
Journal of Personality and So cial Psy chology , 125 (5), 1157–1188. https://doi.org/10.10
37/pspp0000470
Bader , C. D ., & Finke , R. (2010). What do es go d r e quir e? Understanding r eligious conte xt and
morality . In S. Hitlin & S. V aise y (Eds.), Handb o oks of so ciology and so cial r esear ch
(pp . 241–254). Springer . https://doi.org/10.1007/978-1-4419-6896-8_13
Baldwin, S. A., & Larson, M. J. (2017). An intr o duction to using Bay esian linear r egr ession
with clinical data. Behaviour Resear ch and Therap y , 98 , 58–75. https://doi.org/10.1016
/j.brat.2016.12.016
Centr e for So cial Informatics. (2023). 1KA (V ersion 23.06.20) [ Computer softwar e]. Univ ersity
of Ljubljana, Faculty of So cial Sciences. https://www.1ka.si/
Cliffor d, S., Iy engar , V ., Cab eza, R., & Sinnott- Armstr ong, W . (2015). Moral foundations
vignettes: A standar dize d stimulus database of scenarios base d on moral foundations
the or y . Behavior Resear ch Metho ds , 47 (4), 1178–1198. https://doi.org/10.3758/s13428-
014-0551-2
Dawson, K. J., Han, H., & Choi, Y . R. (2021). Ho w ar e moral foundations asso ciate d with
empathic traits and moral identity? Curr ent Psy chology , 42 (13), 10836–10848. https:
//doi.org/10.1007/s12144-021-02372-5
Do dd, M. D ., Balzer , A., Jacobs, C. M., Gruszczynski, M. W ., Smith, K. B., & Hibbing, J. R.
(2012). The p olitical left r olls with the go o d and the p olitical right confr onts the bad:
Conne cting physiology and cognition to pr efer ences. P hilosophical T ransactions of

Resear ching moral foundations using the Bay esian appr oach 91
the Ro yal So ciety B: Biological Sciences , 367 (1589), 640–649. https://doi.org/10.1098/rs
tb.2011.0268
Efferson, L., Glenn, A., Remmel, R., & Iy er , R. (2017). The influence of gender on the r ela-
tionship b etw e en psy chopathy and fiv e moral foundations: Influence of gender on
psy chopathy-moral foundations r elationship . Personality and Mental Health , 11(4),
335–343. https://doi.org/10.1002/pmh.1395
Eisenb erg, N., Miller , P . A., Shell, R., McNalle y , S., & Shea, C. (1991). Pr oso cial de v elopment
in adolescence: A longitudinal study . De v elopmental Psy chology , 27 (5), 849–857.
https://doi.org/10.1037/0012-1649.27.5.849
Er ceg, N., Galić, Z., & Bubić, A. (2018). The psy chology of e conomic attitudes: Moral foun-
dations pr e dict e conomic attitudes b e y ond so cio-demographic variables. Cr oatian
Economic Sur v e y , 20 (1), 37–70. https://doi.org/10.15179/ces.20.1.2
Faulkenb err y , T . J., Ly , A., & W agenmakers, E. - J. (2020). Bay esian infer ence in numerical
cognition: A tutorial using JASP . Journal of Numerical Cognition , 6 (2), 231–259.
https://doi.org/10.5964/jnc.v6i2.288
Friesen, A. (2019). Generational change? The effe cts of family , age , and time on moral
foundations. The Forum , 17 (1), 121–140. https://doi.org/10.1515/for-2019-0005
Frimer , J. A., Biesanz, J. C., W alker , L. J., & MacKinlay , C. W . (2013). Lib erals and conser vativ es
r ely on common moral foundations when making moral judgments ab out influential
p e ople . Journal of Personality and So cial Psy chology , 104 (6), 1040–1059. https://doi.or
g/10.1037/a0032277
Gelman, A., & Hill, J. (2007). Data analysis using r egr ession and multile v el/hierar chical mo dels .
Cambridge Univ ersity Pr ess.
Gilligan, C. (1977). In a differ ent v oice: W omen’s conceptions of self and of morality . Har var d
Educational Re vie w , 47 (4), 481–517. https:/ /doi .org/ 10.17763 /haer .47. 4. g6167429416h
g5l0
Graham, J., Haidt, J., K ole va, S., Motyl, M., Iy er , R., W ojcik, S. P ., & Ditto , P . H. (2013).
Moral foundations the or y: The pragmatic validity of moral pluralism. In P . De vine
& A. P lant (Eds.), A dvances in e xp erimental so cial psy chology (pp . 55–130, V ol. 47).
A cademic Pr ess. https://doi.org/10.1016/b978-0-12-407236-7.00002-4
Graham, J., Nosek, B. A., Haidt, J., Iy er , R., K ole va, S., & Ditto , P . H. (2011). Mapping the
moral domain. Journal of Personality and So cial Psy chology , 101 (2), 366–385. https:
//doi.org/10.1037/a0021847
Güner , H. (2020). Examining the moral foundations of high scho ol students. W orld Journal
of Education , 10 (6), 1–13. https://doi.org/10.5430/wje.v10n6p1
Haidt, J. (2008). Morality . Persp e ctiv es on Psy chological Science , 3 (1), 65–72. https://doi.org/10
.1111/j.1745-6916.2008.00063.x
Haidt, J., Graham, J., & Joseph, C. (2009). Ab o v e and b elo w left–right: Ide ological narrativ es
and moral foundations. Psy chological Inquir y , 20 (2–3), 110–119. https://doi.org/10.10
80/10478400903028573
Haidt, J., & Joseph, C. (2004). Intuitiv e ethics: Ho w innately pr epar e d intuitions generate
culturally variable virtues. Dae dalus , 133 (4), 55–66. http://www.jstor.org/stable/2002
7945
Hatemi, P . K., Crabtr e e , C., & Smith, K. B. (2019). Ide ology justifies morality: Political b eliefs
pr e dict moral foundations. A merican Journal of Political Science , 63 (4), 788–806.
https://doi.org/10.1111/ajps.12448

92 Rakinić
Horb erg, E. J., O v eis, C., K eltner , D ., & Cohen, A. B. (2009). Disgust and the moralization of
purity . Journal of Personality and So cial Psy chology , 97 (6), 963–976. https://doi. org/1
0.1037/a0017423
IBM Corp oration. (2020). IBM SPSS Statistics for Windo ws (V ersion 27.0) [ Computer softwar e].
https://www.ibm.com/spss
Iy er , R., K ole va, S., Graham, J., Ditto , P ., & Haidt, J. (2012). Understanding lib ertarian morality:
The psy chological disp ositions of self-identifie d lib ertarians. PLOS ONE , 7 (8), e42366.
https://doi.org/10.1371/journal.pone.0042366
Jaffe e , S., & Hy de , J. S. (2000). Gender differ ences in moral orientation: A meta-analysis.
Psy chological Bulletin , 126 (5), 703–726. https://doi.org/10.1037/0033-2909.126.5.703
JASP T eam. (2023). JASP (V ersion 0.17.2) [ Computer softwar e]. https://jasp-stats.org/
Jeffr e ys, H. J. (1939). The or y of pr obability (1st e d.). O xfor d Univ ersity Pr ess.
Jeffr e ys, H. J. (1961). The or y of pr obability (3r d e d.). O xfor d Univ ersity Pr ess.
K ass, R. E., & Rafter y , A. E. (1995). Bay es factors. Journal of the A merican Statistical A sso ciation ,
90 (430), 773–795. https://doi.org/10.1080/01621459.1995.10476572
K ohlb erg, L. (1969). Stage and se quence: The cognitiv e-de v elopmental appr oach to so cializa-
tion. In D . A. Goslin (Ed.), Handb o ok of so cialization the or y and r esear ch (pp . 347–480).
Rand McNally .
K ole va, S. P ., Graham, J., Iy er , R., Ditto , P . H., & Haidt, J. (2012). T racing the thr eads: Ho w
fiv e moral concerns ( esp e cially Purity ) help e xplain cultur e war attitudes. Journal of
Resear ch in Personality , 46 (2), 184–194. https://doi.org/10.1016/j.jrp.2012.01.006
Kugler , M., Jost, J. T ., & No orbalo o chi, S. (2014). Another lo ok at moral foundations the or y:
Do authoritarianism and so cial dominance orientation e xplain lib eral-conser vativ e
differ ences in ”moral” intuitions? So cial Justice Resear ch , 27 (4), 413–431. https://doi
.org/10.1007/s11211-014-0223-5
Kurzban, R., Dukes, A., & W e e den, J. (2010). Se x, drugs and moral goals: Repr o ductiv e
strategies and vie ws ab out r e cr eational drugs. Pr o ce e dings of the Ro yal So ciety B:
Biological Sciences , 277 (1699), 3501–3508. https://doi.org/10.1098/rspb.2010.0608
Liang, F ., Paulo , R., Molina, G., Cly de , M. A., & Berger , J. O . (2008). Mixtur es of 𝑔 priors for
Bay esian variable sele ction. Journal of the A merican Statistical A sso ciation , 103 (481),
410–423. https://doi.org/10.1198/016214507000001337
Lo w , M., & W ui, M. G. L. (2015). Moral foundations and attitudes to war ds the p o or . Curr ent
Psy chology , 35 (4), 650–656. https://doi.org/10.1007/s12144-015-9333-y
Mainier o , L. A., Gibson, D . E., & Sullivan, S. E. (2008). Retr osp e ctiv e analysis of gender
differ ences in r eaction to me dia co v erage of crisis e v ents: Ne w insights on the justice
and car e orientations. Se x Roles , 58 , 556–566. https://doi.org/10.1007/s11199-007-936
5-3
McA dams, D . P ., Albaugh, M., Farb er , E., Daniels, J., Logan, R. L., & Olson, B. (2008). Family
metaphors and moral intuitions: Ho w conser vativ es and lib erals narrate their liv es.
Journal of Personality and So cial Psy chology , 95 (4), 978–990. https://doi.org/10.1037/a
0012650
Miles, A. (2014). Demographic corr elates of moral differ ences in the contemp orar y Unite d
States. Po etics , 46 , 75–88. https://doi.org/10.1016/j.poetic.2014.09.004
Perr y , W . G. (1970). Forms of intelle ctual and ethical de v elopment in the college y ears . Holt,
Rinehart & Winston.
R Cor e T eam. (2022). R: A language and envir onment for statistical computing (V ersion 4.2.1)
[ Computer softwar e]. R Foundation for Statistical Computing. https : / / www . r - projec
t.org/

Resear ching moral foundations using the Bay esian appr oach 93
Robinson, O . C. (2013). V alues and adult age: Findings fr om tw o cohorts of the Eur op ean
So cial Sur v e y . Eur op ean Journal of A geing , 10 (1), 11–23. https://doi.org/10.1007/s104
33-012-0247-3
Rosik, C. H., Dinges, L. J., & Saav e dra, N. (2013). Moral intuitions and attitudes to war d gay
men: Can moral psy chology add to our understanding of homonegativity? Journal of
Psy chology and The ology , 41 (4), 315–326. https://doi.org/10.1177/009164711304100404
Rouder , J. N. (2014). Optional stopping: No pr oblem for Bay esians. Psy chonomic Bulletin &
Re vie w , 21 (2), 301–308. https://doi.org/10.3758/s13423-014-0595-4
Sağel, E. (2015). A ge differ ences in moral foundations acr oss adolescence and adultho o d [ Un-
publishe d master thesis]. Middle East T e chnical Univ ersity .
Simmons, J. P ., Nelson, L. D ., & Simonsohn, U . (2011). False-p ositiv e psy chology: Undisclose d
fle xibility in data colle ction and analysis allo ws pr esenting anything as significant.
Psy chological Science , 22 (11), 1359–1366. https://doi.org/10.1177/0956797611417632
Sno w-Hill, N. L. (2019). The sur v e y of attitudes to war d homeless p e ople: The validation of
a ne w instrument assessing negativ e attitudes to war d homeless p e ople [ Unpublishe d
do ctoral dissertation]. Univ ersity of South Car olina.
Stein, Z., & Dawson- T unik, T . L. (2004). ”It’s all go o d”: Moral r elativism and the millennial
mind . De v elopmental T esting Ser vice .
Sulsky , L. M., Mar cus, J., & MacDonald, H. A. (2015). Examining ethicality judgements of
theft b ehavior: The r ole of moral r elativism. Journal of Business and Psy chology , 31 (3),
383–398. https://doi.org/10.1007/s10869-015-9418-5
T riandis, H. C. (2001). Individualism‐colle ctivism and p ersonality . Journal of Personality ,
69 (6), 907–924. https://doi.org/10.1111/1467-6494.696169
van de Scho ot, R., Winter , S. D ., Ryan, O ., Zonder van-Zwijnenburg, M., & Depaoli, S. (2017). A
systematic r e vie w of Bay esian articles in psy chology: The last 25 y ears. Psy chological
Metho ds , 22 (2), 217–239. https://doi.org/10.1037/met0000100
van Do orn, J., van den Bergh, D ., Böhm, U ., Dablander , F ., Derks, K., Draws, T ., Etz, A., Evans,
N. J., Gr onau, Q . F ., Haaf, J. M., Hinne , M., Kucharský , Š., Ly , A., Marsman, M., Matzke ,
D ., Gupta, A. R. K. N., Sarafoglou, A., Stefan, A., V o elkel, J. G., & W agenmakers,
E. - J. (2021). The JASP guidelines for conducting and r ep orting a Bay esian analysis.
Psy chonomic Bulletin & Re vie w , 28 (3), 813–826. https:/ /doi. org/10 .3758/ s13423-020 -0
1798-5
van Le euw en, F ., K o enig, B. L., Graham, J., & Park, J. H. (2014). Moral concerns acr oss
the Unite d States: A sso ciations with life-histor y variables, pathogen pr e valence ,
urbanization, cognitiv e ability , and so cial class. Ev olution and Human Behavior , 35 (6),
464–471. https://doi.org/10.1016/j.evolhumbehav.2014.06.005
W agenmakers, E. - J. (2007). A practical solution to the p er vasiv e pr oblems of 𝑝 values.
Psy chonomic Bulletin & Re vie w , 14 (5), 779–804. https://doi.org/10.3758/bf03194105
Zellner , A., & Sio w , A. (1980). Posterior o dds ratios for sele cte d r egr ession hyp otheses.
T rabajos de Estadistica Y de Inv estigacion Op erativa , 31 (1), 585–603. https://doi.org/10
.1007/bf02888369