Ge ech , S enja; Eckmann, Lisa; Wen zel, Daniel; Landweh , Jan R.
A icle — Published Ve sion
An h opomo phic Sad Exp essions Reduce Was e o
“Single” Impe ec Food
Psychology & Ma ke ing
P o ided in Coope a ion wi h:
John Wiley & Sons
Sugges ed Ci a ion: Ge ech , S enja; Eckmann, Lisa; Wen zel, Daniel; Landweh , Jan R. (2024) :
An h opomo phic Sad Exp essions Reduce Was e o “Single” Impe ec Food, Psychology &
Ma ke ing, ISSN 1520-6793, Wiley, Hoboken, NJ, Vol. 42, Iss. 3, pp. 669-683,
h ps://doi.o g/10.1002/ma .22145
This Ve sion is a ailable a :
h ps://hdl.handle.ne /10419/319316
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/
Psychology & Ma ke ing
RESEARCH ARTICLE
An h opomo phic Sad Exp essions Reduce Was e o
“Single”Impe ec Food
S enja Ge ech
1
| Lisa Eckmann
2
| Daniel Wen zel
1
| Jan R. Landweh
2
1
School o Business and Economics, RWTH Aachen Uni e si y, Aachen, Ge many |
2
Facul y o Economics and Business, Goe he Uni e si y F ank u ,
F ank u am Main, Ge many
Co espondence: Lisa Eckmann ([email p o ec ed])
Recei ed: 6 Feb ua y 2024 | Re ised: 11 Oc obe 2024 | Accep ed: 12 Oc obe 2024
Keywo ds: an h opomo phism | emo ional exp essions | ood was e | impe ec ood | poin ‐o ‐sale communica ion
ABSTRACT
Poin ‐o ‐sale an h opomo phism—a ibu ing human cha ac e is ics o nonhuman en i ies—has been shown o inc ease e a-
lua ions o ood ha de ia es in shape o o m om consume expec a ions (“impe ec ood”). Howe e , while p e ious
esea ch has included he use o p ima y emo ions o manipula e an h opomo phism, i is unclea whe he depic ing impe ec
p oduce wi h happy o sad emo ional exp essions is mo e e ec i e in boos ing choice o such p oduce. This esea ch add esses
his gap and addi ionally examines a no el ype o impe ec ion: when p oduce is o n o he bunch i no mally comes in. A
ield s udy conduc ed a a la ge ood e aile in Ge many e ealed ha sales o “single”bananas signi ican ly inc eased when
hey we e an h opomo phized wi h sad emo ions compa ed o when hey we e an h opomo phized wi h happy emo ions o no
an h opomo phized. Ac oss h ee p e egis e ed online expe imen s, his esea ch co obo a es he e ec i eness o sad emo-
ional exp essions in boos ing pu chase in en ions, es ablishes compassion as he unde lying p ocess, gene alizes his e ec o
o he p oduce (i.e., oma oes), and shows ha a p ice discoun elimina es his e ec . The indings o his esea ch a e pa -
icula ly ele an o ood e aile s aiming o inc ease pu chases, and hus educe was e, o esh p oduce ha has become a
loose “single.”
1 | In oduc ion
Acco ding o he UNEP Food Was e Index 2024, mo e han 1.05
billion ons o ood a e was ed globally (UN En i onmen
P og amme 2024)—by 2030, his numbe is p ojec ed o double
o 2.1 billion ons (BCG 2024). Cu en ly, 131 million ons occu
in he e ail sec o alone (UN En i onmen P og amme 2024).
In indus ialized na ions, his was e mainly esul s om ade
no ms and picky consume p e e ences conce ning he lawless
appea ance o esh p oduce (FAO 2019).
In ligh o hese numbe s, a subs an ial body o esea ch has
eme ged ha examines how o inc ease consume accep ance o
impe ec p oduce. The cu en li e a u e concep ualizes impe ec
ood p edominan ly as ui s o ege ables ha de ia e in shape o
colo om p o o ypical p oduce ye a e equally as y, nu i ious, and
sa e o consume (G ewal e al. 2019; Mooke jee, Co nil, and
Hoegg 2021). Acco dingly, p e ious s udies ha e ocused on
objec i e es he ic laws such as shape de o ma ions o educed
eshness (e.g., misshaped po a oes, wea he ed cucumbe s).
This esea ch examines a no el ype o impe ec ion ha has so
a been o e looked by he cu en li e a u e: when he p oduce
has been o n o he bunch i no mally comes in (e.g., bunched
bananas, oma oes on he ine) and hus de ia es om con-
sume expec a ions ega ding he p esen a ion o p oduce a he
poin o sale. Fo ins ance, single bananas ha e been shown o
accoun o he highes amoun bo h o clima e impac and o
This is an open access a icle unde he e ms o he C ea i e Commons A ibu ion License, which pe mi s use, dis ibu ion and ep oduc ion in any medium, p o ided he o iginal wo k is p ope ly
ci ed.
© 2024 The Au ho (s). Psychology & Ma ke ing published by Wiley Pe iodicals LLC.
The i s wo au ho s con ibu ed equally o his s udy.
669 o 950Psychology & Ma ke ing, 2025; 42:669–683
h ps://doi.o g/10.1002/ma .22145
ood was ed a e aile s (Ma sson, Williams, and Be ghel 2018).
Speci ically, single bananas a e much mo e likely o be dis-
ca ded han hei bunched coun e pa s i hey a e no sold by
he end o he day such ha cu en ood was e a oidance
p ac ices explici ly lis single bananas as a sou ce o a oidable
ood was e (BCG 2024; EcoWo lde 2022; Rees 2023).
Independen o he p oduce in ques ion, one s eam o li e a u e
has ocused on an h opomo phism—asc ibing human cha -
ac e is ics o nonhuman agen s—as a po en ial solu ion o
mi iga ing consume ejec ion o impe ec ood. These s udies
ound posi i e e ec s o a ious o ms o an h opomo phism on
he e alua ion and choice o impe ec ood, including adding
acial elemen s (Shao e al. 2020,2021; Walsh and Da ie 2020),
e aming de o ma ions as body pa s (Coo emans and
Geuens 2019), o mimicking human beha io (Koo, Oh, and
Pa ick 2019).
Mos ele an o he p esen esea ch is p io wo k ha has
examined emo ional exp essions as a o m o an h opomo ph-
ism. Exis ing s udies ha e ocused on he e ec o depic ing
happy (T. Chen e al. 2021; Coo emans and Geuens 2019) o sad
(Mukhe jee, Mukhe jee, and Iye 2021) emo ional exp essions
on impe ec p oduce. While happiness and sadness induce
di e en a ec i e and beha io al esponses and ha e been
shown o di e in e ec i eness when used in ma ke ing mes-
saging in o he domains (Ke on and Nale elich 2019; Small and
Ve ochi 2009), he cu en impe ec ood li e a u e lacks a
di ec compa ison o he e ec i eness o hese emo ions. Ye ,
he a ibu ed capaci y o expe ience emo ions is a powe ul
indica o o pe cei ed humanness, which wa an s u he
in es iga ion in o he di e en ial e ec s o happiness and sad-
ness as p ima y emo ions in an h opomo phism esea ch. This
esea ch seeks o ill his esea ch gap by explici ly compa ing
he e ec i eness o happy and sad emo ional exp essions on
an h opomo phized single p oduce in boos ing choice o such
p oduce.
Ou wo k has di ec p ac ical implica ions. Among di e en
an h opomo phism o ms, endowing impe ec p oduce wi h
emo ions ia acial exp essions is easily iable o many p oduce
ca ego ies, making i an impac ul ye cos ‐e ec i e in e en-
ion. Hence, gene a ing insigh s in o which emo ion is mos
e ec i e in boos ing choice o single p oduce is highly ele an
om a manage ial pe spec i e. Impo an ly, ou wo k also
shows how an h opomo phizing such p oduce in e ac s wi h
ano he s a egy ypically used o p omo e such p oduce—p ice
discoun s— he eby p o iding e aile s wi h clea guidelines on
how o e ec i ely communica e a he poin o sale.
2 | Theo e ical Backg ound
2.1 | An h opomo phism in Impe ec Food
An h opomo phism means imbuing “nonhuman agen s wi h
humanlike cha ac e is ics, mo i a ions, in en ions, o emo-
ions”(Epley, Way z, and Cacioppo 2007, p. 864), allowing in-
di iduals o d aw on hei knowledge ega ding human agen s
o in e p e and make sense o a nonhuman a ge (Epley,
Way z, and Cacioppo 2007). Fo ins ance, seeing a “ ace”in a
nonhuman objec allows indi iduals o e alua e i as i hey
would e alua e a human's ace. Ma ke e s in en ionally inco -
po a e an h opomo phism in o hei p oduc s o help consum-
e s es ablish a humanlike connec ion wi h he p oduc . While
any ype o an h opomo phism ha in ol es some human
componen is concei able, s udied ypes include le ing a
p oduc speak, le ing a p oduc engage in human beha io ,
o gi ingap oduc a acialexp ession,whichha ebeen
ound o posi i ely in luence p oduc e alua ions (Agga wal
and McGill 2007), b and liking (Delbae e, McQua ie, and
Phillips 2011), and p oduc sales (Landweh , McGill, and
He mann 2011), espec i ely.
In he con ex o sus ainabili y, an h opomo phism has no only
been shown o posi i ely in luence en i onmen al conse a ion
beha io s such as esou ce use, p oduc eplacemen , and e-
cycling (Ahn, Kim, and Agga wal 2014; Chandle and
Schwa z 2010; Ke on and Nale elich 2019; Wu, Mal e , and
Joha 2023) bu also o enhance e alua ions and choice o
impe ec p oduce—p oduce ha de ia es in appea ance om
wha consume s conside ypical. Simila o esea ch on
impe ec ood in gene al (e.g., G ewal e al. 2019; Mooke jee,
Co nil, and Hoegg 2021), exis ing s udies in he domain o
an h opomo phism ha e ocused p ima ily on shape abno -
mali ies o educed eshness (see Table 1 o an o e iew) as
mani es a ions o impe ec ions. In pa icula , hese s udies
ound posi i e e ec s on pu chase in en ion, willingness‐ o‐pay,
and p oduc e alua ions o such p oduce om a ange o di -
e en an h opomo phism o ms, including p esen ing oddly
shaped ui and ege ables wi h a humanlike smile (T. Chen
e al. 2021) o sad acial ea u es (Mukhe jee, Mukhe jee, and
Iye 2021), aming shape abno mali ies as body pa s
(Coo emans and Geuens 2019), a anging old p oduce o
esemble a human ace o imi a ing human beha io (Koo, Oh,
and Pa ick 2019), and adding eyes as human acial ea u es
(Shao e al. 2020,2021; Walsh and Da ie 2020).
We aim o add o his body o li e a u e by examining “single-
ness”as ano he , cu en ly unde s udied mani es a ion o
impe ec p oduce. Speci ically, we call esh p oduce “single”
when i is o n o he bunch i o iginally comes in and is
usually p esen ed in (e.g., bunched bananas, oma oes on he
ine) and concep ualize i as impe ec as i de ia es om
consume expec a ions ega ding how i should be p esen ed a
he poin o sale (i.e., in i s o iginal bunched s a e). S udying
such single p oduce is pa icula ly ele an in he con ex o
an h opomo phism as sepa a ion om pee s e lec s a unda-
men al human conce n (Baumeis e and Lea y 1995).
2.2 | Emo ional Exp essions
While he p e iously discussed body o esea ch con ibu es o
ou unde s anding o he e ec i eness o di e en o ms o
an h opomo phism, he impac o di e en emo ions con eyed
h ough an h opomo phism has ecei ed less a en ion. Sig-
naling emo ions and in e p e ing emo ional exp essions shape
e e yday social in e ac ions. The ace is conside ed he main
non e bal medium o communica ing emo ions and he
670 o 950 Psychology & Ma ke ing, 2025
TABLE 1 | O e iew o s udies on an h opomo phism in selling impe ec p oduce (IP).
Resea ch S imuli
Independen
a iable
An h opomo phism
h ough …
Dependen
a iable(s) P ocess Mode a o
S udy
ype Key esul s
T. Chen
e al. (2021)
Typical/
misshapen
po a oes/
ca o s/
apples
An h opomo phism,
ype o
an h opomo phism
Happy ace, ex Pu chase
in en ion
Empa hy P oduc
popula i y
Online An h opomo phizing IP
e okes consume
empa hy due o
ejec ion. Empa hy
inc eases accep ance
and pu chases o IP.
P oduc popula i y
weakens he e ec .
Coo emans
and
Geuens
(2019)
Typical/
misshapen
ca o /
oma o/
cucumbe
P oduc ype ( ypical/
impe ec ),
an h opomo phism
Happy ace, adding
“body pa s”
Pu chase
in en ion,
choice,
willingness‐
o‐pay
Posi i e
a ec i e
eac ions, as e
pe cep ions
En i onmen al
conce n
Lab,
online
Due o posi i e a ec i e
eac ions ha enhance
as e pe cep ions,
an h opomo phizing IP
leads o highe pu chase
in en ions and choice.
This e ec is weake o
consume s high on
en i onmen al
conce ns.
Koo, Oh, and
Pa ick
(2019)
F esh/
wea he ed/
spo ed
cucumbe /
banana/zuc‐
chini
(Im)pe ec ion,
an h opomo phism
A anging elemen s o
esemble neu al human
ace, IP engaging in
ypical human beha io
(sunba hing)
P oduce
e alua ion
Feelings o
wa m h
Lay belie s
abou aging
Lab,
online
An h opomo phizing
old p oduce inc eases
e alua ion due o
enhanced eelings o
wa m h, con ingen on
“old is gold”belie s.
Mukhe jee,
Mukhe jee,
and Iye
(2021)
(S udy 2)
“An
asso men
o bo h
ui s and
ege ables”
(Im)pe ec ion,
an h opomo phism
“sad aces ha mimic
human acial ea u es”
Pu chase
in en ion,
e aile
pa onage
in en ion
Feelings o
emba assmen
—Online An h opomo phism
inc eases pu chase
in en ion and e aile
pa onage in en ion due
o dec eased eelings o
emba assmen when
pu chasing IP.
Shao
e al. (2020)
Misshaped
po a o
An h opomo phism Eyes, neu al beak,
ha , ex
Pu chase
in en ion
—Tempo al
dis ance,
message
aming
Online An h opomo phism
inc eases pu chase
in en ion owa ds IP.
Gain‐ amed messages
a e mo e e ec i e when
se in he dis an u u e,
(Con inues)
671 o 950
TABLE 1 | (Con inued)
Resea ch S imuli
Independen
a iable
An h opomo phism
h ough …
Dependen
a iable(s) P ocess Mode a o
S udy
ype Key esul s
while loss‐ amed
messages a e mo e
e ec i e when desc ibed
in he nea u u e.
Shao, Jeong,
and
Jang (2021)
Misshaped
po a o
An h opomo phism Eyes, neu al beak,
ha , ex
Pu chase
in en ion
—Gende , age Online In he younge age
g oup, emale
consume s' pu chase
in en ion owa ds IP is
posi i ely in luenced by
an h opomo phism
while men's is no . The
posi i e e ec applies o
bo h gende s in he
olde age g oup.
Walsh and
Da ie
(2020)
Misshaped
lemon
An h opomo phism Googly eyes Willingness‐
o‐pay
——Online Pa icipan s we e
willing o pay mo e o
oddly shaped lemons in
he an h opomo phized/
whimsical condi ion
compa ed o hose in he
con ol condi ion.
This
esea ch
Single/
bunched
bananas/
oma oes
An h opomo phized
acial exp ession/
emo ion
Happy e sus sad
ace, ex
Choice Compassion P ice discoun ,
cogni i e load
Field,
online
An h opomo phized
sadness is mo e e ec i e
in d i ing consume s'
choice o single IP han
happiness o no
an h opomo phism. The
e ec (sad s. happy) is
media ed by compassion
and elimina ed in he
p esence o a p ice
discoun (sad s.
con ol).
672 o 950 Psychology & Ma ke ing, 2025
li e a u e has pa icula ly ocused on s udying p ima y emo-
ions (e.g., happiness and sadness), which a e uni e sally con-
eyed h ough dis inc acial exp essions (Kel ne e al. 2003).
Impo an ly, p e ious wo k has ound ha he capaci y o ex-
pe ience emo ions (e.g., joy) has a s onge e ec on people's
pe cep ion o wha i means o ha e a mind—and hus, be
human— han he capaci y o ac (e.g., planning; G ay, G ay,
and Wegne 2007). Hence, s udying con eyed emo ions (e.g.,
ia “ acial”exp essions) is di ec ly ele an in he con ex o
an h opomo phism.
Howe e , ecen esea ch on an h opomo phizing impe ec
p oduce has ei he no used speci ic emo ions (Koo, Oh, and
Pa ick 2019; Shao e al. 2020,2021; Walsh and Da ie 2020),
used happy exp essions only (T. Chen e al. 2021; Coo emans
and Geuens 2019), o used sad exp essions only (Mukhe jee,
Mukhe jee, and Iye 2021). To ou knowledge, no p io esea ch
has explici ly es ed whe he con eying happy o sad emo ional
exp essions is mo e e ec i e in inc easing consume choice o
impe ec ood. Ye , a di ec compa ison o he e ec o happy
and sad emo ional exp essions seems highly ele an in his
con ex , as he p ima y emo ions o happiness and sadness can
be unde s ood as opposi es (Feldman Ba e and Russell 1998)
and a e hus likely o e oke di e en a ec i e and beha io al
esponses.
Speci ically, we p opose ha he ex en o which an h opo-
mo phism a ec s consume p e e ences o single p oduce
is s onge o sad han o happy emo ions. O e all, p e ious
esea ch has demons a ed ha con eying sad emo ions is
e ec i e in inc easing message compliance ac oss a ious
domains, including p osocial beha io (Bagozzi and
Moo e 1994), en i onmen al conse a ion (Ahn, Kim, and
Agga wal 2014; Ke on and Nale elich 2019), and ood choices
(Choueiki, Geuens, and Ve mei 2021). Mo e impo an ly, he e
is empi ical e idence ha con eying sadness is mo e e ec i e in
di ec compa ison wi h con eying happiness. Fo ins ance,
Small and Ve ochi (2009) showed ha dona ion appeals we e
mo e e ec i e when he ic ims po ayed in hese appeals
exhibi ed sad a he han happy emo ions. Addi ionally, Ke on
and Nale elich (2019) es ed happy and sad an h opomo phic
cues and, while inding no di e ences be ween he emo ions
hemsel es, unco e ed ha consume s a e mo e likely o
comply wi h en i onmen al appeals when being exposed o sad
e sus no an h opomo phic cues in he decision con ex . Sim-
ila ly, a p oduc ea u ing a sad exp ession may ins ill an u ge o
sa e he p oduc om i s “mise y,”leading o a highe pu chase
likelihood. Con e sely, a p oduc wi h a happy exp ession may
be pe cei ed as being con en in i s con ex . As a esul , con-
sume s a e less likely o expe ience a need o ac ion and may
hus be less inclined o buy ha p oduc . Hence,
H1. Depic ing a sad ( s.happy o no)emo ional exp ession on
single p oduce inc eases cus ome choice o ha p oduce.
2.3 | Media ion by Compassion
As heo ized in he p e ious sec ion, happy and sad emo ional
exp essions on impe ec ood should in luence consume
beha io di e en ly as hey elici di e en a ec i e esponses.
Fo happy exp essions, pas esea ch in he domain o impe ec
ood has shown ha imbuing impe ec p oduce wi h a smiling
“ ace”leads o highe eelings o wa m h compa ed o when he
p oduce is no an h opomo phized (T. Chen e al. 2021;
Coo emans and Geuens 2019). Simila ly, Koo, Oh, and Pa ick
(2019) a gue ha he e ec o an h opomo phizing old p oduce
s ems om “sen imen al eelings o wa m h”(p. 348). Wa m h
pe cep ion is a p ima y de e minan o pe son pe cep ion and
e alua ion and e lec s how well‐in en ioned, iendly, and
us wo hy a pe son appea s (Fiske, Cuddy, and Glick 2007;
Wojciszke 2005). Essen ially, hese s udies a gue ha an h o-
pomo phizing an impe ec p oduc ac i a es a posi i e human
schema, which, in u n, elici s a posi i e a ec i e eac ion (see
also Agga wal and McGill 2007).
Fo sad exp essions, an h opomo phism is likely o lead o
eelings dis inc om wa m h. Wi nessing sadness in o he s is
likely o e oke compassiona e esponses, pa icula ly in he
con ex o social exclusion. The need o belong is one o he
mos basic human mo i a ions (Baumeis e and Lea y 1995)
such ha people pe cei e being in olun a ily sepa a ed om
pee s as a e si e. We ollow he compassion concep ualiza ion
by Goe z, Kel ne , and Simon‐Thomas (2010) who de ine
compassion as “ he eeling ha a ises in wi nessing ano he 's
su e ing and ha mo i a es a subsequen desi e o help”(p. 2)
and who conside simila cons uc s such as sympa hy, pi y,
and empa hic conce n as a amily o compassion‐ ela ed s a es
ha sha e he cen al desi e o alle ia e ano he 's su e ing.
Indeed, p io esea ch has demons a ed ha such conce n o
o he s is associa ed wi h inc eased helping beha io (e.g.,
Do idio, Allen, and Sch oede 1990). Acco dingly, consume s
exposed o a single p oduce con eying sadness a e likely o
expe ience compassion, which will ansla e in o a desi e o
amelio a e he p oduc 's su e ing. As he compassion e oked
by sad exp essions—in con as o he wa m h e oked by happy
exp essions—should hus igge an immedia e beha io al
impulse o “help”(i.e., buy) he p oduce, he e ec o an h o-
pomo phizing single p oduce on consume choice should be
s onge o sad exp essions han o happy exp essions:
H2. The e ec o depic ing sad emo ional exp essions o single
p oduce on choice o such p oduce is media ed by compassion.
2.4 | Cogni i e Load and P ice Discoun s as
Po en ial Mode a o s
Based on ou p oposed mechanism h ough compassion, we
examined bo h a consume ‐ ela ed and a con ex ‐ ela ed mod-
e a o wi h high p ac ical ele ance ha migh a enua e his
p ocess in a shopping con ex .
We ocused on cogni i e load as a consume ‐ ela ed mode a o
as se e al s udies ha e documen ed ha he a ailabili y o
cogni i e esou ces impac s decision‐making (e.g., G eene
e al. 2008; Hinson, Jameson, and Whi ney 2003). Acco ding o
cogni i e load heo y (Swelle 2011), he wo king memo y has
limi ed capaci y o p ocessing in o ma ion. When his capaci y
is exceeded, p ocessing in o ma ion can become impai ed.
673 o 950
Acco dingly, consume s' cogni i e load migh in luence he
e ec i eness o an h opomo phizing impe ec p oduce a he
poin o sale.
On he one hand, expe iencing compassion in esponse o a sad
p oduc may equi e su icien cogni i e esou ces (Goe z,
Kel ne , and Simon‐Thomas 2010), such ha he di e ence
be ween happy and sad exp essions may ei he disappea unde
high load o e e se gi en ha happy an h opomo phism can
be decoded unde high cogni i e load. On he o he hand,
consume s migh expe ience a spon aneous, gu ‐le el esponse
(Rameson, Mo elli, and Liebe man 2012) when exposed o an
impe ec p oduc exp essing sadness, sugges ing ha he e ec
o sad exp essions is obus ac oss di e en le els o cogni i e
load. Hence, while we hypo hesized ha a sad emo ion depic ed
on single p oduce would inc ease choice mo e han a happy
emo ion unde low cogni i e load, we did no hypo hesize a
speci ic di ec ion o he e ec o high cogni i e load bu
speci ied di e en possible pa e ns o esul s as ollows:
H3. Depic ing sad ( s.happy)emo ions on an h opomo phized
single p oduce is mo e e ec i e in inc easing cus ome choice o
ha p oduce unde low cogni i e load.Unde high cogni i e load,
di e en pa e ns o e ec s a e heo e ically plausible (i.e., sad is
mo e han,less han,o equally e ec i e as happy).
In addi ion o examining cogni i e load, we ocused on p ice
discoun s as an ex e nal, con ex ‐ ela ed mode a o . P ice dis-
coun s a e he p omo ional ool mos equen ly used by ma -
ke e s (Da ke and Chung 2005) and he main s a egy o boos
sales o impe ec ood (G ewal e al. 2019). Hence, i is
impo an o unde s and how discoun s in e ac wi h ou p o-
posed an h opomo phism in e en ion. P e ious esea ch has
demons a ed ha in oducing p ice discoun s can unde mine
o he in e en ions a ge ed a p omo ing impe ec ood by
signaling i s lowe quali y (Mooke jee, Co nil, and Hoegg 2021).
Mo e impo an ly, eminde s o money ha e been shown o
p ime a sel ‐su icien o ien a ion, making people dislike
dependencies and hus less likely o help o he s (Vohs, Mead,
and Goode 2006). Consequen ly, when he concep o money is
ac i a ed by he p esence o a p ice discoun , consume s should
be less suscep ible o an emo ionalized message asking he
consume o “help”(i.e., buy) an impe ec p oduc . Acco d-
ingly, we a gue ha in oducing p ice discoun s should weaken
he compassion e oked by sad emo ional exp essions on
impe ec ood, making he sad an h opomo phism in e en ion
less e ec i e:
H4. Depic ing sad emo ions on an h opomo phized single
p oduce inc eases cus ome choice o ha p oduce when no
p ice discoun is p esen .In he p esence o a p ice discoun , he
e ec is mi iga ed.
2.5 | Empi ical O e iew
We conduc ed ou empi ical s udies (see Figu e 1 o ou
concep ual model). We i s es he eal‐li e ele ance o ou
key p oposi ion in a ield expe imen , whe e we obse e ac ual
pu chase beha io (S udy 1). We hen explo e he unde lying
p ocess ia compassion (S udy 2) and examine cogni i e load
and p ice discoun as po en ial mode a o s (S udies 3 and 4) in
he con olled en i onmen o online expe imen s. All s udies
we e p e egis e ed and ecei ed e hical ins i u ional pee
app o al. We epo he s udy p o ocols o ou expe imen s
and addi ional analyses in Suppo ing In o ma ion S1:
Appendices Aand B, espec i ely. The da a se s, s a is ical code,
and s imuli o all s udies a e a ailable on he Open Science
F amewo k: h ps://os .io/8k6pb/.
3 | S udy 1: Field Expe imen
S udy 1 (p e egis e ed: h ps://asp edic ed.o g/ 98d-k 74.pd
and h ps://asp edic ed.o g/4ddx- qz9.pd )
1
was conduc ed in
collabo a ion wi h a majo Ge man supe ma ke chain and
es ed he e ec o di e en emo ional exp essions in a ield
se ing by obse ing ac ual pu chase beha io o single bananas.
3.1 | Me hod
The expe imen was conduc ed on 8 days o 12 h each in wo
supe ma ke s o he same chain, yielding 192 h o da a collec-
ion. In he supe ma ke s, single bananas a e placed in a sep-
a a e baske nex o he bunch bananas. To es how consume s'
choices di e ed ac oss di e en ypes o an h opomo phizing,
FIGURE 1 | Concep ual model.
674 o 950 Psychology & Ma ke ing, 2025
we ins alled h ee di e en signs (con ol, happy, sad) nex o
hese baske s. The wo an h opomo phized condi ions we e
based on exis ing esea ch on acial exp essions and emo ions
(e.g., Bassili 1979; Ekman 1993). In he happy (sad) condi ion,
he sign ea u ed a single banana wi h a smile (downwa d‐
sloped mou h) and he in o ma ion “We a e happy (sad) singles
and wan o be bough as well.”In he con ol condi ion, he
sign only displayed he non‐emo ionalized in o ma ion “He e
a e single bananas ha wan o be bough as well.”Following
Mooke jee, Co nil, and Hoegg (2021), he signs we e o a ed
e e y hou o minimize weekday and day ime in e e ences.
When changing he signs, we coun ed he o e all numbe o
single bananas aken (i.e., sold) in he las hou . To keep he
numbe o bananas in he baske cons an , he baske was e-
illed a e each coun ing when equi ed. Addi ionally, we
wa ched he cus ome s om a hidden spo in he s o e and
eco ded whe he hey chose he single o he bunch bananas
and how many single bananas hey chose. These h ee mea-
su es (numbe o single bananas sold in he las hou , bina y
choice o single bananas pe cus ome , numbe o single
bananas aken pe cus ome ) se ed as ou key dependen
a iables. We also eco ded he numbe o cus ome s passing
he banana shel es in each hou and whe he o he bananas
we e on sale as con ol a iables. Impo an ly, he single
bananas we e no sold a a discoun o as he cheapes op ion
a ailable, as his migh ha e implied a lowe p oduc quali y.
Ins ead, hey we e sold a 1.99 €/kg (i.e., he same p ice as
o ganic and b anded bunches, and 0.84 €/kg mo e expensi e
han he chain's cheapes p i a e label).
3.2 | Resul s and Discussion
Because he s imuli and o a ion scheme we e he same ac oss
bo h supe ma ke s, we pooled he da a o analysis. Ac oss he
en i e ime ame o he s udy (i.e., 192 1‐h ime slo s), a o al o
3,810 cus ome s pu chased bananas (N
con ol
= 1226, N
happy
=
1283, N
sad
= 1301). We had o exclude one ime slo (in which
16 cus ome s bough bananas) because he numbe o pu -
chased single bananas du ing his ime slo was a se e e ou lie
(i.e., i e SDs abo e he mean). This spike was due o a sho age
o bunch bananas in one o he b anches, he eby d i ing sales
o single bananas. Hence, we do no conside his ime slo
ep esen a i e and ocus on he emaining 191 ime slo s and
3794 cus ome s o u he analysis. Ne e heless, we also
conduc ed he analysis wi hou excluding his ou lie as a
obus ness check (see Suppo ing In o ma ion S1: Appendix B).
3.2.1 | Numbe o Single Bananas Sold Pe Time Slo
To es whe he he numbe o single bananas sold pe hou
di e ed depending on he emo ion displayed on he sign, we
conduc ed a one‐way ANOVA using he agg ega ed da a on he
le el o he ime slo s. As expec ed, his analysis e ealed a
signi ican e ec o he expe imen al manipula ion (F(2,
188) = 3.59, p= 0.030, η
2
= 0.037). Subsequen planned con-
as s e ealed a signi ican di e ence in sales be ween he
con ol (M
con ol
= 2.02, SD
con ol
= 2.31) and sad condi ion
(M
sad
= 3.19, SD
sad
= 3.24; p= 0.016), while he e was no sig-
ni ican di e ence be ween he con ol and he happy condi ion
(M
happy
= 2.13, SD
happy
= 2.55; p= 0.818). Mo eo e , he di -
e ence be ween he happy and he sad condi ion was signi i-
can (p= 0.030). This pa e n o esul s emains obus when
including he numbe o cus ome s passing he banana shel es,
b anch, weekday, imeslo , and whe he o he bananas we e on
sale as con ol a iables (see Suppo ing In o ma ion S1:
Appendix B).
3.2.2 | Choice o Single Bananas Pe Cus ome
We used he disagg ega ed da a, ha is, he cus ome s as
he uni o analysis (N= 3,794), and obse ed whe he each
cus ome bough single bananas (= 1) o no (= 0). A χ
2
es
indica ed a signi ican di e ence in he p opo ion o cus ome s
pu chasing single bananas ac oss condi ions (sad: 7%, happy:
5%, con ol: 4%; χ
2
(2) = 10.33, p= 0.006, φ
c
= 0.052) and logis ic
eg essions eplica ed he inding ha he odds o buying single
bananas a e highe in he sad condi ion han in he happy
condi ion (b= 0.38, p= 0.025) and he con ol condi ion
(b= 0.53, p= 0.003).
3.2.3 | Numbe o Single Bananas Taken Pe Cus ome
Again using he cus ome s as he uni o analysis, we conduc ed
a one‐way ANOVA wi h he displayed emo ion as he inde-
penden a iable and he numbe o single bananas aken as he
dependen a iable. Again, a signi ican main e ec o emo ion
is e ealed (F(2, 3791) = 3.28, p< 0.038, η
2
= 0.002) and con as
analyses indica e signi ican di e ences be ween he sad and
he happy condi ions (M
sad
= 0.16, SD
sad
= 0.66; M
happy
= 0.11,
SD
happy
= 0.51; p= 0.027) as well as be ween he sad and he
con ol condi ion (M
con ol
= 0.11, SD
con ol
= 0.57; p= 0.027).
Taken oge he , we ind e idence ac oss all dependen a iables
ha depic ing sad emo ional exp essions on single bananas is
mo e e ec i e in inc easing choice o such p oduce han
depic ing happy o no emo ional exp ession (H1).
4 | S udy 2: Media ion by Compassion
S udy 2 (p e egis e ed: h ps://asp edic ed.o g/5xqn-47x2.pd )
2
was an online expe imen o examine he p ocess unde lying
choice o single p oduce. We hypo hesized ha he e ec o sad
an h opomo phism bu no ha o happy an h opomo phism
would be media ed by compassion (H2). As in he ield ex-
pe imen , we used single ( s. bunched) bananas as ou s imuli.
4.1 | Me hod
In o al, 754 US CloudResea ch pa icipan s (49.1% emale,
49.9% male, 1.0% o he ; M
age
= 40.88) comple ed a one‐ ac o ial
(p oduc an h opomo phism: con ol, happy, sad) be ween‐
subjec s expe imen . Pa icipan s we e shown a pic u e o a
supe ma ke shel wi h single bananas on he le and banana
675 o 950
bunches on he igh side (see Figu e 2). Apa om being
sepa a ed om he bunch, he single bananas we e no abno -
mal in e ms o shape, colo , o size. The displayed signs we e
simila o hose used in S udy 1, excep ha we depic ed a non‐
humanized banana in he con ol condi ion o make i mo e
simila o he o he condi ions and added ea s o he sad
banana o make he emo ion mo e salien (see Figu e 2).
Pa icipan s we e asked o imagine going o he supe ma ke
in ending o buy bananas and o indica e whe he hey would
ake bananas om he le (single) o igh (bunch) side o he
display (1 = de ini ely om he le side, 6 = de ini ely om he
igh side), which se ed as ou dependen a iable ( e e se‐
sco ed such ha highe numbe s indica ed a highe likelihood
o choosing single bananas). Nex , pa icipan s indica ed hei
compassion o he single bananas (“I eel pi y owa ds he
bananas,”“I eel conce n o he bananas,”“I eel sympa hy
owa ds he bananas”; 1 = no a all, 6 = e y much, α= 0.96;
Jackson and LePine 2003) and esponded o ou an h opo-
mo phism manipula ion check (“How s ongly do he bananas
on he le side emind you o a human being?”;“How s ongly
do he bananas on he igh side emind you o a human
being?”; bo h anging om 1 = no a all, 6 = e y much).
Al hough no cen al o ou heo izing, pa icipan s also indi-
ca ed o wha ex en hey el wa m h owa ds he single
bananas as ini ially p e egis e ed (see Suppo ing In o ma ion
S1: Appendix B). Pa icipan s hen esponded o an a en ion
check (“Wha p oduc was depic ed in he shel ?,”1 = banana,
2 = cucumbe , 3 = bee ) and a se o con ol a iables, including
as e and eshness pe cep ions o he bananas, whe he hey
had seen p oduc s as depic ed on he le side o he shel be o e,
how impo an i is o hem o consume esh ui s and ege-
ables, how conce ned hey a e wi h he en i onmen , whe he
hey a e he e m “single”as a he posi i e o nega i e, and
wha hei ela ionship s a us is (see Suppo ing In o ma ion S1:
Appendix A).
4.2 | Resul s and Discussion
4.2.1 | A en ion Check
All pa icipan s passed he a en ion check so ha we e ained
all pa icipan s o analysis.
4.2.2 | Manipula ion Check
As in ended, he manipula ion check o he pe cei ed
humanness o he single bananas was signi ican (F(2,
751) = 13.82, p< 0.001, η
2
= 0.036). The single bananas we e
pe cei ed as signi ican ly mo e human in he happy (M
happy
=
2.68, SD
happy
= 1.70; p< 0.001) and he sad condi ions (M
sad
=
2.71, SD
sad
= 1.72; p< 0.001) ela i e o he con ol condi ion
(M
con ol
= 2.03, SD
con ol
= 1.51). The e was no di e ence ac oss
he wo an h opomo phized condi ions (p= 0.872). Fu he -
mo e, as expec ed, he pe cei ed humanness o he bunch
bananas did no di e ac oss he condi ions (F(2, 751) = 1.05,
p= 0.351, η
2
= 0.003; M
con ol
= 1.88, SD
con ol
= 1.49; M
happy
=
1.95, SD
happy
= 1.43; M
sad
= 1.77, SD
sad
= 1.23).
4.2.3 | Choice
Figu e 3p o ides means o he dependen a iable choice and
he media o a iable compassion. An ANOVA e ealed a sig-
ni ican main e ec o he expe imen al manipula ion on choice
FIGURE 2 | Expe imen al condi ions and s imuli o S udy 2.
676 o 950 Psychology & Ma ke ing, 2025
Small, D. A., and N. M. Ve ochi. 2009. “The Face o Need: Facial
Emo ion Exp ession on Cha i y Ad e isemen s.”Jou nal o Ma ke ing
Resea ch 46, no. 6: 777–787. h ps://doi.o g/10.1509/jmk .46.6.777.
Swelle , J. 2011. “Cogni i e Load Theo y.”In Psychology o Lea ning
and Mo i a ion, edi ed by J. P. Mes e and B. H. Ross, Vol. 55, 37–76.
San Diego, CA: Academic P ess. h ps://doi.o g/10.1016/B978-
0-12-387691-1.00002-8.
UN En i onmen P og amme. 2024. Food Was e Index Repo 2024.
Think Ea Sa e: T acking P og ess o Hal e Global Food Was e. h ps://
wedocs.unep.o g/20.500.11822/45230.
Vohs, K. D., N. L. Mead, and M. R. Goode. 2006. “The Psychological
Consequences o Money.”Science 314, no. 5802: 1154–1156. h ps://doi.
o g/10.1126/science.1132491.
Walsh, D., and C. Da ie . 2020. “Wha a Cu e Lemon! The E ec o
Whimsical Cu eness on Willingness‐ o‐Pay o Impe ec P oduce.”In
NA ‐Ad ances in Consume Resea ch, edi ed by J. J. A go, T. M. Low ey,
and H. Jensen Schau, Vol. 48, 1232.
Whi e, K., L. Lin, D. W. Dahl, and R. J. B. Ri chie. 2016. “When Do
Consume s A oid Impe ec ions? Supe icial Packaging Damage as a
Con amina ion Cue.”Jou nal o Ma ke ing Resea ch 53, no. 1: 110–123.
h ps://doi.o g/10.1509/jm .12.0388.
Wojciszke, B. 2005. “Mo ali y and Compe ence in Pe son‐and Sel ‐
Pe cep ion.”Eu opean Re iew o Social Psychology 16, no. 1: 155–188.
h ps://doi.o g/10.1080/10463280500229619.
Wu, A. Y., M. S. Mal e , and G. V. Joha . 2023. “Recycle Me!”P oduc
An h opomo phism Can Inc ease Recycling Beha io .”Jou nal o he
Associa ion o Consume Resea ch 8, no. 3: 351–363. h ps://doi.o g/10.
1086/724999.
Suppo ing In o ma ion
Addi ional suppo ing in o ma ion can be ound online in he
Suppo ing In o ma ion sec ion.
683 o 950
