University of Groningen
The Breast Size Satisfaction Survey (BSSS)
Tran, Ulrich S; Barron, David; Afhami, Reza; Aimé, Annie; Almenara, Carlos A; Alp Dal,
Nursel; Amaral, Ana Carolina Soares; Andrianto, Sonny; Anjum, Gulnaz; Argyrides, Marios
Published in:
Body Image
DOI:
10.1016/j.bodyim.2020.01.006
IMPORTANT NOTE: You are advised to consult the publisher's version (publisher's PDF) if you wish to cite from
it. Please check the document version below.
Document Version
Publisher's PDF, also known as Version of record
Publication date:
2020
Link to publication in University of Groningen/UMCG research database
Citation for published version (APA):
Tran, U. S., Barron, D., Afhami, R., Aimé, A., Almenara, C. A., Alp Dal, N., Amaral, A. C. S., Andrianto, S.,
Anjum, G., Argyrides, M., Atari, M., Aziz, M., Banai, B., Borowiec, J., Brewis, A., Cakir Kocak, Y., Campos,
J. A. D. B., Carmona, C., Chaleeraktrakoon, T., ... Vidal-Mollón, J. (2020). The Breast Size Satisfaction
Survey (BSSS): Breast size dissatisfaction and its antecedents and outcomes in women from 40 nations.
Body Image, 32, 199-217. https://doi.org/10.1016/j.bodyim.2020.01.006
Copyright
Other than for strictly personal use, it is not permitted to download or to forward/distribute the text or part of it without the consent of the author(s) and/or copyright holder(s), unless the work is under an open content license (like Creative Commons).
Take-down policy
If you believe that this document breaches copyright please contact us providing details, and we will remove access to the work immediately and investigate your claim.
Downloaded from the University of Groningen/UMCG research database (Pure): http://www.rug.nl/research/portal. For technical reasons the number of authors shown on this cover page is limited to 10 maximum.
ContentslistsavailableatScienceDirect
Body
Image
j o u r n al ho me p ag e :w w w . e l s e v i e r . c o m / l o c a te / b o d y i m a g e
The
Breast
Size
Satisfaction
Survey
(BSSS):
Breast
size
dissatisfaction
and
its
antecedents
and
outcomes
in
women
from
40
nations
Viren
Swami
a,b,∗,
Ulrich
S.
Tran
c,
David
Barron
b,
Reza
Afhami
d,
Annie
Aimé
e,
Carlos
A.
Almenara
f,
Nursel
Alp
Dal
g,
Ana
Carolina
Soares
Amaral
h,
Sonny
Andrianto
i,
Gulnaz
Anjum
j,
Marios
Argyrides
k,
Mohammad
Atari
l,
Mudassar
Aziz
m,
Benjamin
Banai
n,
Joanna
Borowiec
o,
Alexandra
Brewis
p,
Yeliz
Cakir
Kocak
q,
Juliana
Alvares
Duarte
Bonini
Campos
r,
Carmen
Carmona
s,
Trawin
Chaleeraktrakoon
t,
Hong
Chen
u,
Phatthanakit
Chobthamkit
t,v,
Bovornpot
Choompunuch
w,
Togas
Constantinos
x,
Aine
Crumlish
y,
Julio
Eduardo
Cruz
z,
Simon
E.
Dalley
A,
Devi
Damayanti
B,
Joanna
Dare
C,
Stacey
M.
Donofrio
A,
Anja
Draksler
D,
Michelle
Escasa-Dorne
E,
Elaine
Frances
Fernandez
F,
Maria
Elisa
Caputo
Ferreira
G,
David
A.
Frederick
C,
Antonio
Alías
García
H,
Shulamit
Geller
I,
Alexias
George
x,
Louai
Ghazieh
J,
Cosmin
Goian
K,
Colin
Gorman
y,
Caterina
Grano
L,
Jonathan
Eliahu
Handelzalts
I,
Heather
Horsburgh
M,1,
Todd
Jackson
N,
Lady
Grey
Javela
Javela
Delgado
O,
Marija
Jovi ´c
P,
Marko
Jovi ´c
Q,
Adam
Kantanista
o,
Sevag
K.
Kertechian
J,
Loes
Kessels
R,
Magdalena
Król-Zieli ´nska
o,
Garry
Kuan
S,
Yee
Cheng
Kueh
T,
Sanjay
Kumar
U,
Ingela
Lundin
Kvalem
V,
Caterina
Lombardo
L,
Ernesto
Luis
López
Almada
W,
Christophe
Maïano
e,
Mandar
Manjary
U,
Karlijn
Massar
R,
Camilla
Matera
X,
Juliana
F.
Figueiras
Mereiles
G,
Norbert
Meskó
Y,
Hikari
Namatame
Z,
Amanda
Nerini
X,
Felix
Neto
aa,
Joana
Neto
ab,
Angela
Nogueira
Neves
ac,
Siu-Kuen
Ng
ad,
Devi
R.
Nithiya
ae,
Salma
Samir
Omar
af,
Mika
Omori
ag,
Maria
Serena
Panasiti
ah,
Irena
Pavela
Banai
n,
Eva
Pila
ai,
Alessandra
Pokrajac-Bulian
aj,
Vita
Postuvan
D,
Ivanka
Prichard
ak,
Magdalena
Razmus
al,
Catherine
M.
Sabiston
am,
Reza
N.
Sahlan
an,
Jacob
Owusu
Sarfo
ao,
Yoko
Sawamiya
Z,
Stefan
Stieger
ap,aq,
Cindi
SturtzSreetharan
p,
Eugene
Tee
F,
Gill
A.
ten
Hoor
R,
Kulvadee
Thongpibul
ar,
Arun
Tipandjan
as,
Otilia
Tudorel
at,
Tracy
Tylka
au,
Zahir
Vally
av,
Juan
Camilo
Vargas-Nieto
aw,
Luis
Diego
Vega
ax,
Jose
Vidal-Mollón
s,
Mona
Vintila
at,
Deborah
Williams
ay,
Amber
Wutich
p,
Yuko
Yamamiya
az,
Danilo
Zambrano
aw,
Marcelo
Callegari
Zanetti
Aa,
Ivanka ˇZivˇci ´c-Be ´cirevi ´c
aj,
Martin
Voracek
caSchoolofPsychologyandSportScience,AngliaRuskinUniversity,Cambridge,UnitedKingdom bCentreforPsychologicalMedicine,PerdanaUniversity,Serdang,Malaysia
cDepartmentofBasicPsychologicalResearchandResearchMethods,SchoolofPsychology,UniversityofVienna,Vienna,Austria dDepartmentofArtStudies,TarbiatModaresUniversity,Tehran,Iran
eDepartmentofPsychoeducationandPsychology,UniversitéduQuébecenOutaouais,Saint-Jérôme,Canada
fInstituteforResearchonChildren,Youth,andFamily,FacultyofSocialStudies,MasarykUniversity,Brno,CzechRepublic gMidwiferyDepartment,MunzurUniversity,Tunceli,Turkey
hFederalInstituteofEducation,Science,andTechnologyofSoutheastMinasGerais,SãoBarbacena,Brazil iDepartmentofPsychology,IslamicUniversityofIndonesia,Yogyakarta,Indonesia
jDepartmentofSocialSciencesandLiberalArts,InstituteofBusinessAdministration,Karachi,Pakistan
∗ Correspondingauthorat:SchoolofPsychologyandSportScience,AngliaRuskinUniversity,Cambridge,CambridgeshireCB11PT,UnitedKingdom. E-mailaddress:viren.swami@anglia.ac.uk(V.Swami).
1 HeatherHorsbughisnowattheSchoolofEducationandSocialSciences,UniversityoftheWestofScotland,Paisley,UnitedKingdom.
https://doi.org/10.1016/j.bodyim.2020.01.006
SchoolofHealthSciences,NeapolisUniversityPaphos,Paphos,Cyprus
lDepartmentofPsychology,UniversityofSouthernCalifornia,LosAngeles,UnitedStates
mDepartmentofSocialSciences,ShaheedZulfikarAliBhuttoInstituteofScienceandTechnology,Karachi,Pakistan nDepartmentofPsychology,UniversityofZadar,Zadar,Croatia
oDepartmentofPhysicalEducationandLifelongSports,Pozna´nUniversityofPhysicalEducation,Pozna´n,Poland pSchoolofHumanEvolutionandSocialChange,ArizonaStateUniversity,Tempe,UnitedStates
qMidwiferyDepartment,EgeUniversity, ˙Izmir,Turkey
rDepartmentofFoodandNutrition,SchoolofPharmaceuticalScienceofSãoPauloStateUniversity,SãoPaulo,Brazil sDepartmentofResearchMethodsandDiagnosisinEducation,UniversityofValencia,Valencia,Spain
tDivisionofPsychology,FacultyofLiberalArts,ThammasatUniversity,Pathumthani,Thailand uSchoolofPsychology,SouthwestUniversity,Chongqing,China
vSchoolofPsychology,UniversityofKent,Canterbury,UnitedKingdom
wDepartmentofEducationalPsychologyandCounseling,FacultyofEducation,KhonKaenUniversity,KhonKaen,Thailand xDepartmentofPsychology,PanteionUniversityofAthens,Athens,Greece
ySchoolofPsychology,UlsterUniversity,Coleraine,NorthernIreland,UnitedKingdom zDepartmentofPsychology,UniversitydelosAndes,Bogotá,Colombia
AFacultyofBehaviouralandSocialSciences,UniversityofGroningen,Groningen,Netherlands BDepartmentofPsychology,AhmadDahlanUniversity,Yogyakarta,Indonesia
CDepartmentofPsychology,ChapmanUniversity,Orange,UnitedStates
DSloveneCentreforSuicideResearch,AndrejMarusicInstituteandDepartmentofPsychologyFAMNIT,UniversityofPrimorska,Koper,Slovenia EDepartmentofAnthropology,UniversityofColoradoColoradoSprings,UnitedStates
FDepartmentofPsychology,HELPUniversity,ShahAlam,Malaysia GFederalUniversityofJuizdeFora,JuizdeFora,Brazil
HDepartmentofEducation,UniversityofAlmeria,Almeria,Spain
ISchoolofBehavioralSciences,AcademicCollegeofTel-AvivYaffo,Yaffo,Israel JDepartmentofHumanResources,SaintJosephUniversity,Beirut,Lebanon KDepartmentofSocialWork,WestUniversityofTimis¸oara,Timis¸oara,Romania LDepartmentofPsychology,SapienzaUniversityofRome,Rome,Italy
MSchoolofSocialSciences,UniversityofSouthampton,Southampton,UnitedKingdom NDepartmentofPsychology,UniversityofMacau,Taipa,China
ODepartmentofPsychology,UniversidaddelRosario,Bogotá,Colombia
PDepartmentofMarketingManagementandPublicRelations,UniversityofBelgrade,Belgrade,Serbia QDepartmentofPlasticandReconstructiveSurgery,SchoolofMedicine,UniversityofBelgrade,Belgrade,Serbia RDepartmentofWorkandSocialPsychology,MaastrichtUniversity,Maastricht,Netherlands
SSchoolofHealthSciences,UniversitiSainsMalaysia,KubangKerian,Kelantan,Malaysia
TUnitofBiostatisticsandResearchMethodology,SchoolofMedicalSciences,UniversitiSainsMalaysia,KubangKerian,Kelantan,Malaysia UDepartmentofPsychology,D.A.V.College,Muzaffarnagar,India
VDepartmentofPsychology,UniversityofOslo,Oslo,Norway
WFacultyofHumanities,UniversidadCatólicaCampusItapúa,Encarnación,Paraguay
XDepartmentofEducation,Languages,Intercultures,Literatures,andPsychology,UniversityofFlorence,Florence,Italy YDepartmentofGeneralandEvolutionaryPsychology,InstituteofPsychology,UniversityofPécs,Pécs,Hungary ZDepartmentofPsychology,UniversityofTsukuba,Tsukuba,Japan
aaDepartmentofPsychology,UniversityofPorto,Porto,Portugal
abDepartmentofPsychology,CatholicUniversityofBrasília,Taguatinga,Brazil
acDivisionofResearch,PhysicalEducationCollegeoftheBrazilianArmy,RiodeJaneiro,Brazil adPhysicalEducationUnit,ChineseUniversityofHongKong,HongKong,China
aeDepartmentofPhysiology,MahatmaGandhiMedicalCollegeandResearchInstitute,SriBalajiVidyapeth,India afDepartmentofDermatology,Venereology,andAndrology,AlexandriaUniversity,Alexandria,Egypt
agDepartmentofPsychology,OchanomizuUniversity,Tokyo,Japan
ahSantaLuciaFoundation,ScientificInstituteforResearchandHealthcare,Rome,Italy aiSchoolofKinesiology,UniversityofWesternOntario,Ontario,Canada
ajDepartmentofPsychology,FacultyofHumanitiesandSocialSciences,UniversityofRijeka,Rijeka,Croatia akCollegeofNursingandHealthSciences,FlindersUniversity,Adelaide,Australia
alInstituteofPsychology,MarieCurie-SkłodowskaUniversity,Lublin,Poland
amFacultyofKinesiologyandPhysicalEducation,UniversityofToronto,Toronto,Canada anDepartmentofClinicalPsychology,IranUniversityofMedicalSciences,Tehran,Iran aoDepartmentofNursing,AllNationsUniversityCollege,Koforidua,Ghana
apDepartmentofPsychologyandPsychodynamics,KarlLandsteinerUniversityofHealthSciences,KremsanderDonau,Austria aqDepartmentofPsychology,ResearchMethods,Assessment,andiScience,UniversityofKonstanz,Konstanz,Germany arDepartmentofPsychology,FacultyofHumanities,ChiangMaiUniversity,ChiangMai,Thailand
asIndiraGandhiGovernmentGeneralHospitalandPostgraduateInstitute,Puducherry,India atDepartmentofPsychology,WestUniversityofTimis¸oara,Timis¸oara,Romania
auDepartmentofPsychology,OhioStateUniversity,Columbus,UnitedStates
avDepartmentofPsychologyandCounseling,UnitedArabEmiratesUniversity,AlAin,UnitedArabEmirates awDepartmentofPsychology,FundaciónUniversitariaKonradLonrez,Bogotá,Colombia
axDepartmentofPsychology,UniversidadLatinadeCostaRica,SanJosé,CostaRica ayCollegeofHealthSolutions,ArizonaStateUniversity,Phoenix,Arizona,UnitedStates azDepartmentofUndergraduateStudies,TempleUniversityJapan,Tokyo,Japan AaDepartmentofPhysicalEducation,SãoJudasTadeuUniversity,SãoPaulo,Brazil
a
r
t
i
c
l
e
i
n
f
o
Articlehistory:
Received13September2019
Receivedinrevisedform24January2020 Accepted25January2020
Availableonline4February2020
a
b
s
t
r
a
c
t
TheBreastSizeSatisfactionSurvey(BSSS)wasestablishedtoassesswomen’sbreastsizedissatisfaction andbreastedexperiencesfromacross-nationalperspective.Atotalof18,541womenwererecruitedfrom 61researchsitesacross40nationsandcompletedmeasuresofcurrent-idealbreastsizediscrepancy,as wellasmeasuresoftheorisedantecedents(personality,Westernandlocalmediaexposure,andproxies ofsocioeconomicstatus)andoutcomes(weightandappearancedissatisfaction,breastawareness,and
Keywords: Breastsize Breastideals Breastawareness Personality Psychologicalwell-being Cross-culturalpsychology
psychologicalwell-being).Inthetotaldataset,47.5%ofwomenwantedlargerbreaststhanthey cur-rentlyhad,23.2%wantedsmallerbreasts,and29.3%weresatisfiedwiththeircurrentbreastsize.There weresignificantcross-nationaldifferencesinmeanidealbreastsizeandabsolutebreastsize dissatisfac-tion,buteffectsizesweresmall(2=.02–.03).Theresultsofmultilevelmodellingshowedthatgreater Neuroticism,lowerConscientiousness,lowerWesternmediaexposure,greaterlocalmediaexposure, lowerfinancialsecurity,andyoungeragewereassociatedwithgreaterbreastsizedissatisfactionacross nations.Inaddition,greaterabsolutebreastsizedissatisfactionwasassociatedwithgreaterweightand appearancedissatisfaction,poorerbreastawareness,andpoorerpsychologicalwell-beingacrossnations. Theseresultsindicatethatbreastsizedissatisfactionisaglobalpublichealthconcernlinkedtowomen’s psychologicalandphysicalwell-being.
©2020ElsevierLtd.Allrightsreserved.
1. Introduction
Whenawoman placesherhandover herheart,it liesonand betweenherbreasts.Ifherchestisthehouseofherbeing,from whichradiatesherenergytomeettheworld,herbreastsarealso entwinedwithhersenseofself.[...]Formanywomen,ifnotall, breastsareanimportantcomponentofbodyself-image;awoman maylovethemordislikethem,butsheisrarelyneutral(Young, 1992,p.215).
Thebodyisahighlysymbolicandimportantsiteforthe cre-ation,negotiation,andperpetuationofsocialandpowerrelations
(Backett-Milburn&McKie,2001;Jeffreys,2005;Scheper-Hughes&
Lock,1987;Thompson,Heinberg,Altabe,&Tantleff-Dunn,1999).
In this view,macro-level culturalfactors – such as patriarchal structures (Bartky, 1990; Jeffreys, 2005), masculine aesthetics
(Dworkin,1974),andbeautysystems(Bordo,1993;Turner,1984)
– tie women’s moral, economic, and socio-political worth to theirphysicalappearanceandwillingnesstoenactpersonalbody projects.Concomitantly,thosestructuresorsystemsalso propa-gateprescriptiveandunrealisticbeautyideals(Swami,2015,2020), rewardingwomenwhostriveandattainthoseidealsandpunishing ordenigratingthosewhodonot(Forbes,Collinsworth,Jobe,Braun,
&Wise,2007;Kimmel,2011;Ramati-Ziber,Shnabel,&Glick,2019;
Swami,Colesetal.,2010).Totheextentthatthosebeautyideals
areinternalisedbywomen,itoftenresultsinahostof detrimen-taloutcomes(e.g.,negativebodyimage,symptomsofdisordered eating,poorer psychologicalwell-being)when women perceive themselves as being unable to embody those ideals (Rodgers,
Campagna,&Attawala,2019;Thompsonetal.,1999;Tomiyama
etal.,2018).
Muchof the academic literature hasfocused on attainment ofathinideal,butsomeworksuggeststhistheorisingmayalso applytowomen’sbreasts(Gripsrud,Ramvi,Froggett,Hellstrand,
&Manley,2018;Lee, 1997)and “breastedexperiences”(Young,
1992,p.215).Thisisunsurprisingconsideringevidencethatthe breastsplayanimportantroleinjudgementsandobjectificationof women(e.g.,Bareket,Shnabel,Abeles,Gervais,&Yuval-Greenberg,
2019;Dixson,Grimshaw,Linklater,&Dixson,2011;Swami&Tovée,
2013b),andarereportedtobeafocusofwomen’sbodyanxieties
(e.g.,Beck,Ward-Hull,&McLear,1976;Lee,1997;Millsted&Frith,
2003;Young,1992).Yet,comparativelylittleiscurrentlyknown
aboutwomen’sbreastdissatisfaction,particularlyintermsof cross-nationalvariations,aswellasantecedentsandoutcomes.Torectify thisgapintheliterature,wereporthereontheresultsoftheBreast SizeSatisfactionSurvey,across-nationalsurveyofbreastsize ide-alsandbreastsizedissatisfactioninwomenacross40nations.Not onlydoesthissurveyrepresentthelargestmulti-sitestudyever conductedontheissueofbreastsizedissatisfaction,italsoprovides importantinsightsthatwillbeofusetohealthcarepractitioners, policy-makers,andactivists.
1.1. Breastsizeanddissatisfaction
Women’sbreastsvaryalongmanydimensions,suchasshape, symmetry,andfirmness(Atiye&Chahine,2018),butmany con-temporarybeautysystemsreportedlyobjectifywomeninterms of theirbreastsize (Calogero &Thompson,2010; Mazur,1986;
Seifert,2005;Tantleff-Dunn,2001).Inparticular,largebreastsare
commodifiedandfetishisedinmanyformsofmainstreamWestern media(Einon,2012;Gerald&Potvin,2009;Ward,Merriwether,&
Caruthers,2006),particularlymediathat(re)produce
stereotypi-calgenderrolediscourse(Graff,Murnen,&Krause,2013;Yockey,
King,Vidourek,Burbage,&Merianos,2019)andheteronormative
expectations(Murnen,Poinsatte,Huntsman,Goldfarb,&Glaser,
2015;Rodriguez&Hernandez,2019;Rogers&Liebler,2017;for
a counter-point,seeWebb, Vinoski,Warren-Findlow,Burrell, &
Putz, 2017).Thatis,in contemporary culturalimagery,“female
breasts arenot celebrated orscrutinized forwhat theydo, but for how theyaresupposed to look”(Naugler,2009, p.101). In addition to de-coupling breasts fromtheir functional purposes (e.g.,child-feeding),contemporarypopularculturealsotieslarge breaststoconceptionsofwomanhood,somuchsothatlargebreasts havebecomeadefiningcharacteristicoffemininity(Dozier,2005;
Gripsrud,2008;Young,1992).
Thiscommodificationandscrutinisationofbreastscaninfluence howwomenfeelabouttheirownbodies(Bonillas,2009;Goldsmith
&Byers, 2016; Millsted &Frith, 2003; Webb, Jacox,&
Temple-Oberle,2019).StudiesofNorthAmericanandWesternEuropean
womenhavedocumentedhighlevelsofbreastsizedissatisfaction –typicallyoperationalisedasadiscrepancybetweencurrentand idealbreastsizesand,lessfrequently,asattitudesaboutbreastsize –sinceatleastthe1950s(Forbes&Frederick,2008;Forbes,Jobe,&
Revak,2006;Jacobi&Cash,1994;Jourard&Secord,1955;
Tantleff-Dunn,2002;Tantleff-Dunn&Thompson,2000).Forexample,one
largesurveyofNorthAmericanheterosexualwomen(N=26,703) reportedthatthemajority(70%)weredissatisfiedwithsomeaspect oftheirbreastsandthat,oftheseparticipants,28%wantedlarger breasts(Frederick,Peplau,&Lever,2008).InWesternEurope, stud-ieshavereportedthatamajorityofsampledwomenintheUnited Kingdom(Swami,Cavelti,Taylor,&Tovée,2015,Study3=54.3%, Study4=53.3%) or closeto amajority in theUnitedKingdom
(Swami,Caveltietal.,2015,Study2=49.7%;Swami&Furnham,
2018:44.2%)andItaly(Lombardo,Panasiti,Vacca,Grano,&Swami, 2019:44.4%)wantedlargerbreaststhantheycurrentlyhad.
Muchlessresearchhasexaminedratesofbreastsize dissatis-factionoutsideNorthAmericaandWesternEurope,aneglectthat isimportantbecausetheethnographicrecordsuggestssome vari-ationinidealbreastsize,atleasthistorically(Ford&Beach,1951). Inparticular,someculturesmayessentialisetheperformativityof womanhood andgender identitythroughsymbolicassociations with largebreasts. For example, somescholars have discussed theheightenedculturalpressureexperiencedbyLatinAmerican
womentodemonstratefemininityand sexualitythroughbreast fullness(Correa&Shohamy,2018;Guimarãesetal.,2015). Con-sistentwiththisview,Junqueiraetal.(2019)reportedthat65.5% ofBrazilianwomenintheirstudydesiredlargerbreaststhanthey currentlyhad,althoughitshouldbenotedthesampleconsisted primarilyofyoungwomen.Incontrast,EastandSoutheastAsian womentypicallyhavesmallerbreaststhanwomeninotherparts oftheworld(Limetal.,2018),andsomeliteraturehasproposed thatbreastshavehistoricallyplayedasubsidiaryroleintermsof theculturalobjectificationofwomeninthisregion(Miller,2003;
seealsoMenon,2019)andareemphasisedlessoftenin
percep-tionsoffemininity(B ˛ak-Sosnowska,Pawlicka,&Warchał,2016;
seealsoChing&Xu,2019).Likewise,somenon-Western
commu-nitiesmayconstruelargebreastsasproblematic,eitherbecause largebreastsareperceivedasacauseofdisease(e.g.,breastcancer;
Naanyuetal.,2015)orbecausebeinglarge-breastedis
stereotypi-callyassociatedwithhypersexuality(Liebelt,2019)andnarcissism
(seeSmith,2017).
Nevertheless, it is also possible that breast size ideals are becomingincreasinglyhomogenisedacrosstheglobe.For exam-ple,in relation to thethin ideal, scholars have discussed how theinter-relatedforces ofglobalisation(particularlyinterms of theproliferationofWesternmassmediaandWesternisedbeauty ideals)andmodernisation(e.g.,changingnutritionalandlifestyle patterns, disruptions to gender inequality and traditional gen-derroles)haveservedtoenforceahomogenisedpreference for thinnessacrossmuchoftheglobe(Anderson-Fye,2018;Brewis,
Wutich,Falletta-Cowden,&Rodriguez-Soto,2011;Swami,2015).
Thisissupportedbyempiricaldatashowingarelatively standard-isedpreferenceforfemalethinnessandsimilarratesofactual-ideal weightdiscrepancyacrossnations (e.g.,Swami,Fredericketal., 2010).Similarprocesseshavenotbeenexaminedvis-à-visbreast size,althoughsomescholarshavesuggestedtheincreasing com-modificationandobjectificationoflargebreastsinsomepartsof theworld(e.g.,inEastandSoutheastAsia;Rongmuangetal.,2011;
Wong,2019),aswellasamarkedshiftawayfromafocusonthe
functionsofthebreaststowardtheirtreatmentinpurelyaesthetic terms(Miller,2006).Intheabsenceofempiricaldata,however,it isdifficulttoknowtowhatextentahomogenisingpatterncould beobservedglobally,andsoafirstaimofthepresentstudywasto examinebreastsizeidealsandbreastdissatisfactionacrossdiverse nationalsites.
1.2. Antecedentsofbreastsizedissatisfaction
Here,wealsoconsideredantecedentsofbreastsizeidealsand breastsizedissatisfaction,whichhavenotbeeninvestigatedinany depthinpreviousresearch(Lombardoetal.,2019).Onesuch poten-tialantecedentissocioeconomicstatus:intermsofthethinideal, forexample,cross-sectional (e.g.,Mintem, Horta,Domingues,&
Gigante,2015;Swami,Fredericketal.,2010;Swami,Jones,Einon,
&Furnham,2009;Swami&Tovée,2005b,2005b),experimental
(Swami&Tovée,2006),andethnographic(Anderson-Fye&Brewis,
2017)researchhasdocumentedaninverserelationshipbetween socioeconomicstatusandidealisationofathinideal.Inexplanation, ithasbeensuggestedthatindividualsfromsitesofrelativelylow socioeconomicstatusmayidealiselargerbodysizesbecausebody fatissymbolicofresourcesecurity(e.g.,accesstofoodorwealth), whereasthinnesscomestobeassociatedwithwealthandstatusin sitesofrelativelyhighersocioeconomicstatus(forreviews,seeFox,
Feng,&Asal,2019;Swami,2015,2020).Asimilarpatternhasbeen
proposedforbreast sizeideals, withcross-sectionaland experi-mentalresearchsuggestingapreferenceforlargerbreastsizeswith decreasingsocioeconomicstatus(Dixson,Vaseyetal.,2011;Swami
&Tovée,2013a;butseeGray&Frederick,2012).Assuch,itmight
beexpectedthattherewouldbevariationinwhatisperceivedto
betheidealbreastsizeasafunctionofinter-individualdifferences insocioeconomicstatus.
Ontheotherhand,ratesofbreastsizedissatisfaction(i.e.,a dis-crepancybetweencurrentandidealbreastsizes)mightbeexpected tobelower among participantsof lower socioeconomicstatus. For example,based ondata fromresearchsites in two nations (MalaysiaandSouthAfrica)thatdifferedinsocioeconomicstatus,
Swami,Fredericketal.(2010)reportedsignificantlylower
actual-idealweightdiscrepancy inparticipantsfromsites ofrelatively lowcomparedtohighsocioeconomicstatus.Similarfindingshave alsobeenreportedinBrazil(Laus,Miranda,Almedia,BragaCosta,
&Ferreira,2012),MalaysianBorneo(Swami,Kannan,&Furnham,
2012),andMexico(Penelo,Negrete,Portell,&Raich,2013),with participantsfromsitesofhighersocioeconomicstatusreporting morenegativebodyimagecomparedtothosefromsitesoflower socioeconomicstatus.Thissuggeststhatparticipantsofrelatively lowsocioeconomicstatuswouldreportlowerratesofbreastsize dissatisfaction.Thereareanumberofreasonsforexpectingthis patternofresults(forreviews,seeAnderson-Fye&Brewis,2017;
Swami,2015), includinglowerpressurefrombeautysystemsto
attainbeautyidealsinsitesofrelativelylowsocioeconomicstatus andlowerexposuretoWesternmassmediathatconstructlarge breastsasideal.
In terms of the latter, previous cross-national research has shown that exposureto Western mediais significantly associ-atedwithboththeidealisationofthinfemalebodiesandwomen’s bodydissatisfaction(Swami,Fredericketal.,2010).Someresearch hasexaminedtheimpactof exposuretomassmediaonbreast sizeideals and breastsize dissatisfaction.For example, qualita-tiveresearchwithwomenfromtheUnitedStateshasimplicated exposuretolargebreastsdepictedinmassmediaasanimportant influenceonbreastsizedissatisfaction,despiteparticipants’stated dislikeof those mediated images (Goodman & Walsh-Childers, 2004).Similarly,exposuretoidealisedbodiesinmassmediahas beenassociatedwithapreferenceforlargerbreastsincollege-aged NorthAmericanwomen(Harrison,2003)andhasalsobeen associ-atedwithnegativefeelingsaboutwomen’sownbreasts(Frederick,
Daniels,Bates,&Tylka,2017).OnestudywithruralNicaraguans
foundthatexposuretoWesterntelevisionwasassociatedwitha preferenceforlargerbreasts(Thornborrow,Jucker,Boothroyd,&
Tovée,2018),althoughthestudyonlyconsideredmen’sratingsof
women.
Alternatively,somescholarshavesuggestedthatmediaeffects havebeenexaggerated(e.g.,Ferguson,2018;Swami,2020).For example,Ferguson’s(2013)meta-analysis,which included stud-iesthatweremainlyconductedinadvancedeconomies,concluded therewasnooveralleffectofmediaexposureonwomen’sbody dissatisfaction.Instead,theyconcludedthatmediaeffectsonbody dissatisfactionwerelimitedtothosewithpre-existingbody dissat-isfactionorpossiblythosewithvulnerablepersonalitytraits,such ashighNeuroticism.Infact,asidefromthispotentialmoderating role,Neuroticismhasalsobeenfoundtobedirectlyassociatedwith morenegativebodyimage(Frederick, Sandhu,Morse,&Swami,
2016;Sutin&Terracciano,2016;Swami,Taylor,&Carvalho,2011,
2013).Indeed,onerecentsystematic reviewreportedthat Neu-roticismwaspositivelyassociated withnegativebody imagein women(Allen&Walter,2016),possiblybecauseindividualshighin Neuroticismaremoreself-consciousandvulnerabletobodyimage threats.Relationshipsbetweenotherpersonalitytraitsandbody imagearemoreequivocal(Allen&Celestino,2018;Allen&Walter,
2016;Swami&Furnham,2016),thoughitshouldbenotedthatno
studyhasexaminedassociationswithbreastsizedissatisfaction specifically.
1.3. Outcomesofbreastsizedissatisfaction
Intermsofthebodyimageliterature,breastsizedissatisfaction istypicallyconceptualisedasafacetofnegativebodyimagemore generally(Ålgarsetal.,2011;Swami,Caveltietal.,2015).Thus, wewouldexpectsignificantassociationsbetweenbreastsize dis-satisfactionandotherindicesofnegativebodyimage.Indeed,the availableevidencesupportssuchassociations:greaterbreastsize dissatisfactionhasbeenfoundtobesignificantlyassociatedwith greaterbodydissatisfaction(Forbes&Frederick,2008;Frederick
etal.,2008;Swami,Caveltietal.,2015;Swami&Furnham,2018),
greater actual-idealweight discrepancy (Junqueira et al., 2019;
Swami,Caveltietal.,2015), lowerbodyappreciation(Junqueira
etal.,2019;Swami,Caveltietal.,2015),greaterappearance
pre-occupation(Koff&Benavage,1998),andgreaterdriveforthinness
(Swami,Caveltietal.,2015).Thestrengthofassociationshas
gen-erally been weak-to-moderate (e.g., Swami & Furnham, 2018), althoughitshouldbenotedthatatleastonestudyfoundno sig-nificantassociationwithbodyimagedisturbance(Tantleff-Dunn
&Thompson,2000).Beyondindicesof bodyimage,twostudies
have reported that greater breast size dissatisfaction is signifi-cantly,albeitweakly, associatedwithlowerself-esteem(Koff &
Benavage,1998;Swami,Caveltietal.,2015;butforanulleffect,
seeTantleff-Dunn&Thompson,2000).
Anotherimportantoutcomeofbreastsizedissatisfactionmaybe decreasedbreastawareness,whichinvolvesfamiliaritywithone’s breasts(i.e.,howtheynormallylookandfeel,andhowtheychange throughouta woman’slife),as wellas theconfidence tonotice anddetectanychange(e.g.,throughbreastself-examinationfor breastcancersymptoms)andtoseekimmediatemedicalhelpwhen achangeis detected(McCready,Littlewood,&Jenkinson,2005;
Thornton&Pillarisetti,2008).Althoughhealthcarepoliciesinmany
countriescontinuetoadvocate breast self-examinationalone, a growingbody ofevidencesuggeststhat,whendoneonitsown andwithoutpropertraining,breastself-examinationmaynot nec-essarilyimprovebreastcancerdetectionandmortalityreduction (forareview,seeKösters&Gøtzsche,2003).Instead,amore holis-ticapproachbasedonbreastawarenessappearstoofferimproved efficacyintermsofbreastcancerdetection(Harmer,2011;Mant, 1991).Importantly,inacohortofBritishwomen,greaterbreastsize dissatisfaction–butnotbodydissatisfaction–wassignificantly associatedwithlowerbreastawareness(i.e.,lessfrequentbreast self-examination, lower confidence in detecting breast change, andgreaterdelayinseekingprofessionalhelpfollowingabreast change) (Swami &Furnham,2018).In explanation,theauthors suggestedthat theseassociationsmaystem fromwomen seek-ingtoavoidfocusingonthesiteoftheirdissatisfactionornegative emotions(e.g.,shameandembarrassment)thatresultfromvisual inspectionofone’sbreasts.
Whilethesestudiesareimportantintheirownright,itis impor-tanttonotethat theyaregenerallylimited topopulationsin a smallhandfulofworldregions(i.e.,NorthAmericaandWestern Europe)anditisuncleartowhatextentsimilarassociations(i.e., withnegativebodyimage, psychologicalwell-being, and breast awareness)mightbeexpectedinsamplesfromagreaterdiversity of populations.The associationbetweenbreast size dissatisfac-tionandbreastawarenessinparticulardeservesgreaterattention acrossnations. For example, ifscholars areto design interven-tionstopromoteimprovedbreastawarenessbasedonreducing breastsizedissatisfaction,itwillfirstbeimportanttoshowthat theassociationsbetweenthesevariablesarerobustacrossnational groups.However,cross-nationaldifferencesinculturalbeliefsand attitudes(e.g.,bodyorbreast embarrassment,lackof a preven-tivehealthorientation,andfatalism;forareview,seeLee,2015) mayattenuateanyassociationbetweenbreastsizedissatisfaction andbreastawarenessinsomenationalgroups.Forthesereasons,
weexaminedassociationsbetweenbreastsizedissatisfactionand keyoutcomevariables inthepresentstudy,namelyweightand appearancedissatisfaction, psychologicalwell-being, and breast awareness.
1.4. TheBreastSizeSatisfactionSurvey
Whilethenumberofstudiesfocusedonbreastsize dissatisfac-tionhasgrown,animportantlimitationofthisresearchisthatithas primarilyconsideredtheexperiencesofwomeninNorthAmerica andWesternEurope,totheexclusionofwomeninotherpartsof theworld.Thisisnotablebecauseitshouldnotbeassumedthatall populationswillsharesimilarexperiencesoftheirbreastsorthat findingsfromWestern,educated,industrialised,rich,and demo-cratic(WEIRD)societieswillgeneralisetoothersettings(Henrich,
Heine,&Norenzayan,2010).Moreover,as discussedby Swami,
Cavelti et al.(2015), there areadditional limitations that
ham-per our understanding of women’s breasted experiences, even in WEIRD populations, including wide variation in the manner inwhichbreastsizedissatisfactionisoperationalised,theuseof measuresofbreastsizedissatisfactionwithunknownoruntested psychometricproperties,andarelianceonsamplesofcollege-aged women.Inshort,thereisasyetnosystematicinvestigationofbreast sizeidealsandbreastsizedissatisfactionacrossalargeanddiverse setofnations,andscholarshavecalledforfurtherworkthatfocuses onwomen’sbreastedexperiences(Gillen&Markey,2019;Prieler
&Choi,2014).
Toaddresstheseissues,wereportonthefindingsoftheBreast SizeSatisfactionSurvey(BSSS),whichinvolved18,541womenin 61researchsitesacross40nations completingavalidated mea-sure of perceptual breast size dissatisfaction (i.e.,current-ideal breastsizediscrepancy;Swami,Caveltietal.,2015).Thisallowed us to examine the extent to which there exist cross-national differencesinratingsofidealbreastsizeand breastsize dissat-isfaction.Basedonpreviouscross-nationalresearchonbodyimage (e.g.,Swami,Fredericketal.,2010),weexpectedtheretobe sig-nificant cross-national differences across both scores, but also expectedthose differencestobenegligible-to-small instrength (Hypothesis1).Thisalsomirrorstheargumentspresentedabove, which suggestthatbreast sizeidealsare becomingincreasingly homogenisedoruniformacrosstheglobeasaresultofthe pro-cessesofglobalisationandmodernisation.Inadditiontotestingfor thesecross-nationaldifferences,theBSSSdatasetalsoallowedusto examineantecedentsandoutcomesofbreastsizedissatisfaction.
In terms of antecedents, we includedvalidated measuresof theBigFivepersonalityfacetsandnovelmeasuresofexposureto Westernandlocalmedia,whichweexaminedforcross-national measurement invariance here. In addition, we also included measures of financial security and urbanicity as proxies for socioeconomicstatus and, for exploratorypurposes,participant age. In terms of outcomes,we includedvalidated measures of appearanceandweightdissatisfaction,breastawareness(breast self-examinationfrequency,confidenceindetectingbreastchange, anddelayinseekingprofessionalhelpfollowingbreastchange), and psychological well-being (subjective happiness and self-esteem).Multilevelmodellingwasusedtoexaminetherobustness of these antecedents and outcomes of breast size dissatisfac-tion across nations. In terms of antecedents, we hypothesised thatpersonality(specifically,higherNeuroticism),greater expo-suretoWesternand localmedia,greaterfinancialsecurity,and higherurbanicitywouldbesignificantlyassociatedwithgreater breast sizedissatisfaction (Hypothesis2).Interms ofoutcomes, wehypothesisedthatgreaterbreastsizedissatisfactionwouldbe significantly associated with greater appearancedissatisfaction, greaterweightdissatisfaction,lowerbreastawareness(i.e.,lower breastself-examinationfrequency,lowerconfidenceindetecting
Fig.1. SchematicRepresentationoftheHypothesisedAssociationsbetweenAntecedentsandOutcomesofBreastSizeDissatisfaction.Note:Personality=Opennessto Experience,Conscientiousness,Agreeableness,Extraversion,andNeuroticism;BSE=BreastSelf-Examination;Confidence=ConfidenceinNoticingBreastChange.
breastchange,andgreaterdelayinseekingprofessionalhelp fol-lowing breast change), lowerhappiness, and lower self-esteem (Hypothesis3).Fig.1presentsaschematicrepresentationofthese hypothesisedassociations,whichweexpectedtobestableacross allnationssurveyedintheBSSS.
2. Method
2.1. Breastsizesatisfactionsurveyoverview
ThedatareportedherewerecollectedaspartoftheBSSS,a col-laborativeresearchprojectinvolving104scientistsworkingacross 61researchsitesin40nations.Acallforcollaboratorswasplaced onasocialnetworkingsiteforscientistsinSeptember2017,sent outtopotentialcollaboratorsinternationallybythefirstauthor, anddisseminatedopportunisticallyanddirectlytointerested par-tiesuntilDecember2018.TheBSSSdataset,therefore,represents sitesthatwereselectedonaconveniencebasis.Thisisreflected inthefactthat,ofthe61researchsites,mostwereinAsia(23) andEurope(22).Incontrast,therewerefewerresearchsitesin SouthAmerica(8),NorthAmerica(5),Africa(2),andAustralia(1). OnceacollaboratoragreedtobeinvolvedintheBSSS,theywere expectedtoapplyforlocalethicsapprovalwhererequired, trans-late(wherenecessary)andadapt theBSSSsurveyfor localuse, recruitaminimumof200women(relaxedinfourcases:Koforidua, Osaka,Rijeka,andTelAviv)fromthecommunity(relaxedinfour cases–Germany,Hungary,theNetherlands,andPeru–where par-ticipantswererecruitedonline)tocompletetheBSSSsurvey,and returnthedatatothefirstauthorbyApril2019.
2.2. Participants
The BSSS dataset included a total of 18,541 women from 40 nations, ranging in age from 19 to 94 years (M=34.19, SD=13.71)andinself-reportedbodymassindex(BMI)from12.04 to49.78kg/m2(M=23.57,SD=4.54).Intermsofethnic/racial
affil-iation,78.9%self-reportedasbeingmembersoftheethnic/racial majorityoftheirrespectivenationsand9.6%asmembersofan ethnic/racialminority,whereas 11.5%wereunsure.Interms of educationalattainment,0.7%hadnoformaleducational qualifi-cation,4.0%hadcompletedprimaryeducation,21.8%secondary education,21.6%werestillinfull-timeeducation,30.8%hadan undergraduatedegree,16.6%apostgraduatedegree,and4.5%had someotherqualification.Interms oftheircurrentplaceof resi-dence,31.0%ofparticipantslivedinacapitalcity,14.3%incapital citysuburbs,26.1%inaprovincialcity(morethan100,000
resi-dents),16.7%inaprovincialtown(morethan10,000residents), and11.8%inruralareas.Withregardstotheirfinancialsecurity, 22.5%ofparticipantsreportedthattheyfeltlesssecurerelativeto othersoftheirownageintheircountryofresidence,57.8%equally secure,and19.8%moresecure.
Thelistofresearchsitesandnations,includingnation-specific sample characteristics are reported in Table 1 (for individual research sites and associated sample sizes, seeSupplementary Materials). Nations differed with a large effect size in mean age, F(39, 17441)=73.92, p< .001, 2=.14 (lowest: Malaysia,
highest: Norway), with a medium effect size in mean BMI, F(38, 18088)=46.49, p< .001, 2=.09 (lowest: China, highest:
Egypt), and with a small-to-medium effect size in mean self-rated financial security, F(39, 18422)=21.28, p< .001, 2=.04
(lowest:Brazil,highest:UnitedArabEmirates).Therewerealso significantbetween-nationdifferencesinthedistributionof par-ticipants by ethnicity status, 2(78, 18367)=2875.01, p< .001,
W=.40(lowestaffiliationtomajority:CostaRicaandIndia, high-est:ThailandandCroatia),education(secondary/tertiaryvs.other), 2(39,18432)=1891.20,p<.001,W=.32(lowest:China,highest:
Austria),andurbanicity(urbanvs.rural),2(39,18468)=1871.53,
p<.001,W=.32(lowest:Ireland,highest:Colombia,Pakistan,and Peru).
2.3. Measures
2.3.1. Breastsizedissatisfaction
AllparticipantswereaskedtocompletetheBreastSizeRating Scale(BSRS;Swami,Caveltietal.,2015),afiguralratingscale con-sistingof14computer-generatedimagesofwomenwithincreasing breastsize.Imageswerepresentedingreyscaleandwithoutthe appearanceoffacialfeaturessoastominimisetheimpactofthese featuresonratings.Participantswereaskedtoratetheimagethat mostcloselymatchedtheircurrentbreastsizeandtheimagethey wouldmostliketopossess,withresponsesmadeona14-point scale(1representingthefigurewiththesmallestbreastsize,14 rep-resentingthefigurewiththelargestbreastsize).Anindexofbreast sizedissatisfactioncanbecomputedasboththeabsolutedifference betweenidealandcurrentbreastsizeratings(sothathigherscores reflectgreaterbreastsizedissatisfactionregardlessofthedirection ofthisdissatisfaction)orassignedscores(sothatnegativevalues representadesireforsmallerbreastsandpositivevaluesrepresent adesireforlargerbreasts).Bothscoreswereusedinour analy-ses,assignpostedbelow.ScoresderivedfromtheBSRShavebeen showntohavegoodconstructvalidity,acceptabletest-retest
relia-Table1
SampleDescriptionsofDatafromtheBreastSizeSatisfactionSurvey.
Nation n Meanage Meanbodymass
index %Ethnic/racial majority %Secondary/ tertiary education %Urban residence Meanfinancial security
%Preferenceforlarger breasts/smaller breasts/satisfiedwith currentbreastsize
Australia 201 36.30(13.95) 24.54(4.87) 86 90 97 2.17(0.57) 43/28/28 Austria 229 27.20(11.28) 22.56(3.57) 83 91 87 2.13(0.69) 45/24/31 Brazil 822 31.74(10.85) 24.36(4.12) 53 70 95 1.68(0.64) 51/21/28 Canada 463 38.19(15.46) 25.19(5.30) 79 75 90 2.21(0.64) 41/30/29 China 897 24.16(9.19) 20.45(2.73) 91 27 98 1.92(0.56) 73/14/13 Colombia 483 37.72(13.83) 24.62(3.68) 77 64 100 1.91(0.68) 33/11/57 CostaRica 225 34.05(13.16) 24.95(4.65) 44 81 96 2.09(0.65) 41/23/36 Croatia 414 33.62(13.77) 22.81(3.28) 94 69 89 2.16(0.59) 55/17/28 Cyprus 284 34.26(12.21) 23.31(4.67) 90 78 83 1.89(0.64) 41/22/37 Egypt 200 34.27(9.43) 29.07(6.17) 87 77 99 2.13(0.70) 52/36/13 Germany 387 27.21(13.40) 21.97(3.06) 87 83 74 2.19(0.65) 54/20/26 Ghana 174 28.18(9.37) 24.68(6.87) 43 79 94 2.11(0.73) 33/16/52 Greece 1888 34.74(13.15) 23.41(4.05) 81 64 92 1.93(0.62) 47/23/30 Hungary 831 28.50(11.60) 23.52(4.84) 88 54 82 2.02(0.64) 51/19/30 India 441 38.54(9.19) 24.63(4.26) 43 72 75 1.98(0.73) 42/40/18 Indonesia 266 28.77(8.86) 22.66(3.94) 89 72 85 2.04(0.59) 51/19/30 Iran 946 33.92(8.86) 24.71(4.55) 94 89 98 2.02(0.63) 34/33/33 Ireland 219 34.11(16.51) 24.19(4.59) 71 66 36 1.90(0.62) 41/34/26 Israel 187 36.65(13.85) 23.03(3.43) 82 80 83 2.19(0.63) 37/27/36 Italy 747 39.34(13.76) 22.88(4.24) 87 75 89 1.94(0.58) 49/23/29 Japan 423 39.86(17.88) 20.98(2.77) 82 53 94 2.07(0.53) 70/20/10 Lebanon 406 34.87(10.84) 24.45(4.01) 48 90 98 1.79(0.62) 56/20/24 Malaysia 720 22.50(4.81) 22.25(4.67) 71 79 82 1.75(0.66) 55/23/22 Netherlands 512 43.34(15.21) 25.68(5.12) 83 65 69 2.00(0.64) 34/26/40 Norway 254 49.94(15.05) 25.26(4.28) 88 82 71 2.27(0.61) 28/28/44 Pakistan 419 37.68(14.18) 23.55(4.60) 79 85 100 2.20(0.76) 30/49/21 Paraguay 202 37.45(14.32) 26.15(4.91) 80 44 83 2.10(0.70) 31/19/51 Peru 232 31.89(11.65) 24.00(3.56) 47 54 100 2.03(0.69) 51/16/32 Philippines 200 42.29(21.04) N/A 77 85 94 1.93(0.80) 45/22/34 Poland 999 33.33(13.61) 23.21(4.00) 90 71 76 1.92(0.53) 50/22/28 Portugal 203 41.99(15.23) 24.26(3.99) 78 63 96 1.93(0.62) 33/21/45 Romania 428 30.49(13.64) 22.56(5.07) 62 39 79 1.95(0.66) 54/17/29 Serbia 211 34.28(12.90) 22.55(3.86) 83 73 95 2.07(0.70) 50/15/36 Slovenia 343 35.93(13.23) 23.36(3.83) 82 85 72 1.94(0.47) 50/20/31 Spain 712 37.94(14.12) 23.49(4.44) 83 70 72 1.98(0.64) 36/19/45 Thailand 644 29.59(9.24) 22.56(4.89) 95 70 90 1.87(0.59) 60/17/23 Turkey 211 34.13(11.38) 23.65(4.54) 71 64 99 1.69(0.62) 46/26/28 UAE 205 26.71(9.78) 24.23(4.84) 75 64 89 2.30(0.68) 52/28/20 UK 204 36.70(14.62) 24.01(2.75) 90 77 94 1.86(0.73) 56/32/25 USA 699 38.51(14.85) 24.88(5.39) 69 82 91 1.96(0.77) 43/32/25 Total 18541 34.11(13.39) 23.58(13.69) 79 69 88 1.97(0.65) 48/23/30
Note.Numbersaremeansandstandarddeviations(inparentheses),andpercentages(roundedtothenearestinteger)whereindicatedotherwise.Numberswerebasedon allavailabledata.UAE=UnitedArabEmirates,UK=UnitedKingdom,USA=UnitedStatesofAmerica;%Preferenceforlargerbreasts/smallerbreasts/satisfiedwithbreast size=percentagesofparticipantswithsignedbreastsizedissatisfactionscores>0/<0/=0.
bilityupto3months,andadequatepatternsofconvergentvalidity
inwomen(Junqueiraetal.,2019;Swami,Caveltietal.,2015).
2.3.2. Appearanceandweightdissatisfaction
FollowingFredericket al.(2016),participantswereaskedto respondtotwoitemsaskingabouttheirsatisfactionwith physi-calappearanceandweight,respectively(1=Extremelydissatisfied, 7 =Extremelysatisfied). Althoughsingle-item measuresof body imageareunlikelytocapturethecomplexityofbodyimage con-structs, theirscores neverthelesspresent adequate evidence of convergentvalidity(Sandhu&Frederick,2015).Foranalytic pur-poses,scores onboth itemswere reverse-codedsothat higher scoresreflectedgreaterdissatisfactionwithone’sappearanceand weight.
2.3.3. Personality
ParticipantswereaskedtocompletetheFive-ItemPersonality Inventory(FIPI;Gosling,Rentfrow,&Swann,2003),which mea-surestheBigFivepersonalityfacetsofOpennesstoExperience(“I seemyselfasopentonewexperiences”),Conscientiousness(“...as dependableandself-disciplined”),Extraversion(“...asextroverted andenthusiastic”),Agreeableness(“...aswarmandsympathetic
toothers”),andNeuroticism(“...anxiousandeasilyupset”)with oneitemforeach facet.Allitemswereratedona 5-pointscale rangingfrom1(stronglydisagree)to5(stronglyagree).Although therearelimitationstomeasuringtheBigFive facetsusing sin-gleitems,Gosling etal.(2003)reportedthat scoresontheFIPI haveadequatelevelsofconvergentvalidity(e.g.,significant correla-tionswithothermeasuresoftheBigFive)andadequatetest-retest reliabilityuptotwoweeks.
2.3.4. Mediaexposure
To measureexposureto Westernand localmass media,we adaptedthemediaexposurescaledevelopedandusedbySwami,
Fredericketal.(2010).Initsoriginalversion,thescaleincluded8
itemsthataskedaboutfrequencyofexposuretoWestern(4items) andlocal(4items)televisionshows,movies,magazines,andmusic, butwereplacedthetwomusicitemswithtwoitemsabout expo-suretoInternetsitesin thepresentstudy.Allitemswererated ona5-pointscale(1=Lessthanonceamonth,2=Onceortwicea month,3=Onceaweek,4=Severaltimesaweek,5=Everyday).The dimensionality,internalconsistency,andmeasurementinvariance ofscoresonthismeasurearereportedintheResults.
2.3.5. Breastawareness
Toprovideanindexofbreastawareness,wefollowedSwami
andFurnham(2018)inusingthreeitemsfromtheBreast
Mod-uleoftheCancerAwarenessMeasure(BCAM;Linselletal.,2010). TheBCAMisaself-reportedmeasureofmultipledomainsofbreast cancerawarenesswithadequateconstructvalidityandtest-retest reliabilityupto2weeks.Thefirstitem askedaboutbreast self-examinationfrequency(“Howoftendoyoucheckyourbreast?”) withresponsesmadeona4-pointscale(1=Rarelyornever,2=At leastonceeverysixmonths,3=Atleastonceamonth,4=Atleast onceaweek).Linselletal.(2010)suggestedthatwomenweremore likelytobebreastawareiftheyengagedinbreastself-examination atleastonceaweekoronceamonth.Theseconditemaskedabout participants’confidenceinnoticingachangeintheirbreasts(“Are youconfidentyouwouldnoticeachangeinyourbreasts?),with responsesmadeona 4-pointscale (1=Notatall confident,2= Slightlyconfident,3=Fairlyconfident,4=Veryconfident).Thefinal itemaskedparticipants,usinganopen-endedformat,howsoon theywouldcontactahealthprofessionaliftheynoticedachange intheirbreasts(“Ifyoufoundachangeinyourbreasts,howsoon wouldyoucontactyourdoctor?).Responsestothequestionwere scoredbasedona7-pointscaledevelopedbySwamiandFurnham
(2018):1=Immediatelyorassoonaspossible,2=Withinafewdays,
3=Withinaweek,4=Withinamonth,5=Withinthreemonths,6 =Delayaslongaspossible,7=Wouldnotseedoctor.Researchers ineachsitescoredparticipants’responsesascloselyaspossible totheseresponseoptionsandinsomecasesdiscussed categori-sationswithotherscientistsunaffiliatedwiththeprojectand/or withthefirstauthor. For comparativepurposes, womenin the UnitedKingdomareencouragedtocontacttheirdoctorora health-careprofessionalassoonaspossibleupondiscoveringachangein theirbreasts,asearlydetectionincreasesthelikelihoodofpositive outcomes(NationalHealthService,2015).
2.3.6. Psychologicalwell-being
Twosingle-itemmeasureswereusedtoassessdistinctaspects ofpsychologicalwell-being.First,weaskedparticipantsto com-pletetheGlobalHappinessItem(Bradburn,1969),whichprovides an index of subjectively-assessed happiness(“Taking all things together,howwouldyousaythingsarethesedays?”).Theitem wasratedona3-pointscale(1=Nottoohappy,2=Prettyhappy,3= Veryhappy)andscoreshavebeenshowntohaveadequate conver-gentvalidity(e.g.,Lyubomirsky&Lepper,1999;Swami,2008).The seconditemwastheSingle-ItemSelf-EsteemScale(SISE;Robins,
Hendin,&Trzesniewski,2001),inwhichparticipantsareaskedto
ratethestatement“Ihavehighself-esteem”ona7-pointscale(1 =Notverytrueofme,7=Verytrueofme).SISEscoreshavebeen showntohaveadequateconstructvalidity(Robinsetal.,2001).
2.3.7. Socioeconomicstatus
Becauseunderstandingsofsocioeconomicstatusarelikelyto varybothwithinandacrossnations,andduetodifficultiescreating acommonmetricofsocioeconomicstatusacrossnations,weused twoproxiesforsocioeconomicstatus inthepresentwork.First, followingSwami,Kannanetal.(2012),weaskedparticipantsto self-reporthowfinanciallysecuretheyfeltrelativetoothersoftheir ownageintheircountryofresidence(1=Lesssecure,2=Same,3 =Moresecure).Second,weaskedparticipantsabouttheircurrent placeofresidence(i.e.,urbanicity),withresponseoptionsadapted
fromPedersenandMortensen(2001)asfollows:Capitalcity,Capital
citysuburbs,Provincialcity(morethan100,000residents),Provincial town(morethan10,000residents),andRuralareas.Responseoptions werecollapsedintourbanversusruralfordescriptivepurposesat thenationallevelandwereassignedvalues1–5(intheaboveorder) forfurtherstatisticalanalysis.
2.3.8. Demographics
Participantswereaskedtoprovidetheirdemographicdata con-sisting of age, height,and weight using open-ended questions. Heightandweightwererecodedintokilogramsandmetres,and wereusedtocomputeself-reportedBMIaskg/m2.Self-reported
height and weight data are strongly correlated with measured
data(Spencer,Appleby,Davey,&Key,2002)andwereincludedin
thepresentstudyfordescriptivepurposes.2ImprobableBMI
val-ues(<12or>50kg/m2;Swami,Weis,Barron,&Furnham,2018)
werediscarded(0.1%ofthetotaldataset).Participantsalso indi-catedtheirhighesteducationalqualificationbyselectingoneof several options presented (No formal education, Primary educa-tion,Secondaryeducation,Stillinfull-timeeducation,Undergraduate degree, Postgraduate degree, Other). Response options were col-lapsedintosecondary/tertiary(Secondaryeducation,Undergraduate degree,Postgraduatedegree)versusother(remainingcategories)for descriptivepurposesatthenationallevel.Finally,toensure com-parability of ethnicor racial backgrounds acrossresearchsites, participantswereaskedtoself-describetheirethnic/racial affilia-tionbyselectingoneofthreepresentedoptionsrelativetotheir nationofresidence(Ethnic/racialmajority,Ethnic/racial minority, Notsure).
2.4. Testadaptation
WhereEnglishwasnottheprimarylanguageorlinguafrancain anation,themeasuresincludedintheBSSSsurveyweretranslated intotheappropriatelocallanguage(seeSupplementary Materi-alsforthelistoflanguages).ThiswasdoneusingBrislin’s(1970)
back-translationtechnique,whichusesaniterativeprocessof inde-pendentforward-andback-translationbyindependentbilingual translators.Specifically,abilingual translatorfirstblindly trans-latedthe measuresincludedin theBSSS questionnaire – along withinstructions,responsecategories,aninformationsheet,and aninformedconsentform–fromEnglishtothelocallanguage.A secondbilingualtranslatorthenback-translatedthematerialfrom thelocallanguagetoEnglish.Thetwoversionsofthematerials werecompared for equivalence and, where issues were raised, thesewerediscussedwiththefirstauthorandresolvedthrough consensus.In practice,therewere veryfew translationalissues encountered.Alltranslationsareavailablefromthefirstauthor. 2.5. Procedures
Ethicsapprovalfortheoverallprojectanddatamanagement wasobtainedfromthedepartmentalethicscommitteeatAnglia Ruskin University (approval number: ESH17-006). In addition, mostBSSScollaboratorsobtainedethicsapprovalfromlocalethics committeesorInstitutionalReviewBoards.Threeexceptionsto thisoccurredinAustria,CostaRica,andGermany,wherenational lawsdidnotrequireethicsapproval.Intheseand allcases,the projectwasconducted inaccordance withtheprinciplesofthe Declaration of Helsinki (6th revision, 2008) and following local
institutionalguidelines.OncetheBSSSinstrumentshadbeen trans-latedand/oradaptedforlocaluse,collaboratorsrecruitedsamples ofwomenfromthecommunityusingdirectapproachesinareas ofcongregateactivityineachresearchsite.Attemptsweremade toreduceselectionbiasbysamplingatdifferenttimesoftheday andfromdifferentsitesofcongregateactivity,andinsomecases this method wassupplemented withadvertisements to poten-tial participantsplaced in local media. Four exceptions to this generalrecruitmentmethodoccurredinGermany,Hungary,the
2Thereweredifficultiesinunderstandingandcompletingheightandweightdata
Netherlands,andPeru,wheredatawerecollectedonline follow-ingbest-practiceguidelines(e.g.,Meade&Craig,2012).Inclusion criteriaforallresearchsitesincludedbeingfluentinthelocal lan-guageandbeing18yearsofageorolder.Potentialparticipantswere givenaninformationsheet,whichprovidedbriefdetailsaboutthe studyandanestimatedsurveycompletiontime(10min). Partici-pantswhoagreedtotakepartprovidedwritteninformedconsent or digital informed consent for online studies. Thesurvey was anonymous and participantstook partvoluntarily and without remuneration.Allparticipantsreceivedwrittendebriefing infor-mationupon returnof completedquestionnaires.Alldatawere collectedin2018-2019.
2.6. Analyticstrategy
Allwithin-nationdatawerepooledsothatanalysesproceeded atthelevel ofthe nation,ratherthan individualresearchsites. AnalysisoftheBSSSdatathenproceededinfourstages.First,we investigatedcross-nationaldifferencesincurrentandidealbreast size,andinabsolutemeanbreastsizedissatisfaction,using analy-sesofvariance,andreportonthepercentagesofparticipantswith apreferenceforlargerbreasts,smallerbreasts,andnodiscrepancy betweenidealandcurrent breastsizes(Hypothesis1).Wethen checkedwhetheridealbreastsizeratingsweresignificantly asso-ciatedwithsocioeconomicstatus(urbanicityandfinancialsecurity) acrossnationsusingamultilevelmodel(Davidovetal.,2018;van
derVijveretal.,2008)withthetwovariablesaslevel-1predictors
andnationasalevel-2predictor.Interceptswereallowedtovary betweennations.Alllevel-1predictorsweregrandmean-centred andMplus8.2wasutilisedforanalyses,usingMonteCarlo integra-tiontoevaluatethelikelihoodfunction(Heck&Thomas,2015).We reportontheoutcomevarianceexplainedbythemodelonlevels-1 and-2(R12andR22)usingapproximateformulaeprovidedbyBryk
andRaudenbush(1992),whichcomparesthemodelofinterestto
abaselinemodelandincludesonlyrandomintercepts.
Second,todeterminewhetherthemediaexposureitemsformed one-dimensionalscales(separatelyforWesternandlocalmedia), multi-groupmodelswerefittedontheitemsassessingmedia expo-sure,testing for measurement invarianceof theseitems (Chen, 2008).Forthisanalysis,datafromAustriaandGermany,Croatia andSlovenia,theUnitedKingdomandIreland,andJapanandChina, respectively,hadtobemergedinordertomakecomputations fea-sible(thatis,wemergeddataacrossnationswithsimilarcultural backgroundsandthatweregeographicneighbours).Wefirst exam-inedconfiguralinvariance(i.e.,whetherscoresonthe4itemsof eachmeasureformedone-dimensionalscoresineverynation)and thenfullmeasurementinvariance(i.e.,whetheritemparameters –seebelow–wereidenticalacrossnations).Iffullmeasurement invariancewasnotobserved,itemparameterswererelaxedin indi-vidualnationstotestforpartialmeasurementinvariance(which meansthattheparametersofsome,butnotall,itemswereequal acrossnations).Items weretreated asordered-categorical vari-ables,usingthematrixofpolychoriccorrelationsandtheWLSMV estimator(weightedleastsquareestimatorusingadiagonalweight matrixwithstandarderrorsandmean-andvariance-adjusted chi-squareteststatisticthatuseafullweightmatrix)inMplus8.2. Missingdata(0.4%)weretreatedusingfullinformationmaximum likelihood(FIML)estimation,whichissuperiortoothermethods fordealingwithmissingdata(Enders&Bandalos,2001).Eachitem wasmodelledwithtwo sets of parameters:one discrimination parameter(itemloading)andm–1(m=numberofitemresponse categories)thresholdparameters,whichdescribetheprobability foreachresponseoption.Setsofitemparameterswerefreedin tandemforpartialmeasurementanalyses(Sass,2011).Wereport totalasmeasureofreliability(Dunn,Baguley,&Brunsden,2014) forthemanifestscoresofthetwomediaexposurescales.However,
standardisedfactorscoresofWesternandlocalmediaexposure wereusedforfurtheranalyses.
Fortheevaluationofmodelfit,thecomparativefitindex(CFI), the Tucker-Lewis index (TLI), and the standardised root mean squareresidual(SRMR)wereused,utilisingguidelinesprovidedby
HuandBentler(1999)andSchermelleh-Engel,Moosbrugger,and
Müller(2003))(CFI,TLI:>.90acceptablefit,>.95goodfit;SRMR:
<.10acceptablefit,<.05goodfit).WLSMVestimatesthemodel degreesoffreedom(df)fromthedata(Muthén,duToit,&Spisic, 1997);thedfarenotderivedbycomparingthenumberofavailable andestimatedparameters,asinmaximumlikelihood(ML) estima-tion.CFIandTLIusethedftopenaliseformodelcomplexity(the TLImorestronglythantheCFI).UnderMLestimation,thisentails TLI<CFI.However,inthemulti-groupcontextofthepresentstudy, estimationofdfbyWLSMVcouldeitherexcessivelydisadvantage TLItoCFIvalues (especiallyinmodelswitha largernumberof estimatedparameters;i.e.,configuralinvariancemodels)or exces-sivelydisadvantageCFItoTLIvalues (especiallyinmodelswith fewestimatedparameters,suchasfullandpartialmeasurements models,leadingtoCFI<TLI,whichisnotpossibleunderML estima-tion).Thus,modelfitwasconsideredacceptableiftheSRMRand eitherCFIorTLIvaluesindicatedanacceptablefit.Forthe vari-ousmeasurementinvarianceanalyses,modelfitwasinterpreted independently,astheWLSMVestimatordoesnotallowfordirect comparisonsbetweenmodels basedonindices likeCFI (Sass, 2011);becauseofthelargesamplesizesinvolved,wealsodidnot relyon2tests,asthesetendtobecometooliberal(rejectingthe
nullhypothesistoooften)withincreasingN(Cheung&Rensvold,
1999).
In the third stage of analysis, using multilevel modelling, we examinedassociations of potential antecedents withbreast sizedissatisfaction,namelypersonality(OpennesstoExperience, Conscientiousness, Extraversion, Agreeableness, Neuroticism), Westernandlocalmediaexposure(factorscoresfromthe afore-mentionedanalyses),socioeconomicstatus(urbanicity,financial security), andage(level-1predictors), usingnationasa level-2 predictorandallowingforrandomnessin theinterceptand the slopes of Western and local media exposure (i.e.,allowing for differencesbetweennationsintheseparameters;slopesand inter-ceptswereallowedtocorrelate)(Hypothesis2).Fig.1 presents a schematic representation of this model. Multilevel modelling wasutilisedbecauseofthenestednatureofthedata(participants withinnations)(Davidovetal.,2018;Hox,1998).Furthermore,it allowedtestingofwhetherassociationsofmediaexposurewith breastsizedissatisfactionvariedacrosscountries.Guidedbythe resultsofthisinitialmodel,level-1predictorswerethenremoved iftheywerenotsignificantlyassociatedwithbreastsize dissat-isfaction,andslopesofWesternandlocalmediaexposurewere estimatedasfixedeffectsiftheseparametersdidnotexhibit sig-nificantamountsofrandomnessaccordingtolikelihoodratio(LR) tests(Hox,1998).Theremainingsetoflevel-1predictorsservedin theanalysesofthefourthstage.
Inthefourthstageofanalysis,thefinalmodelofthethirdstage incorporatedappearanceandweightdissatisfaction,breast aware-ness,andpsychologicalwell-beingpotentialoutcomesofbreast sizedissatisfaction(Hypothesis3)(seeFig.1).Conceptually,we testedinthismultilevelmodelwhetherbreastsizedissatisfaction mediatedtheeffectsoftheantecedentsontheseoutcomes. Differ-encesbetweennationsintheinterceptsofoutcomeswereallowed forinthismodel.Fortheanalysesinthethirdandfourthstages, alllevel-1predictorsweregrandmean-centred(Hofmann&Gavin, 1998)3 .Mplus8.2wasutilised,usingMonteCarlointegrationto
3Grandmean-centeringsubtractsthegrandmeanofapredictorusingthemean