j. Cosmet. Sci.,59, 127-137 (March/April2008) Whywomenusemakeup: Implication of psychological traits in makeupfunctions RODOLPHE KORICHI, DELPHINE PELLE-DE-QUERAL, GERMAINE GAZANO, and ARNAUD AUBERT, LVMH Recherche, Parrums etCosmgtiques, Rued'Enj•r,45800 SaintJeande Braye,(R. K., D. P.-d.-Q.,G. G.), andDESCO,Facultg des Sciences, 37200 Tours(A. A.), France. Accepted for publication September 27, 2007. Synopsis Makeup actsandstimulates threeof oursenses: touch(which encompasses all sensations fromthebody surface), smell(fragrance), andsight(theprocess ofbecoming andlooking beautiful). Thepositive stimulation ofthese senses bymakeup caninduce sensory aswellaspsychological pleasure. In order tounderstand therelationship ofwomen totheirmakeup, weinterviewed different groups ofwomen ontheirquality of lifeandmakeup habits. Then,through fourstandard well-validated psychometric self-questionnaires, we examined thepossible relation between theneedtomakeuponeself andspecific psychological features. Our firstresults clearly showed thatmakeup couldsupport twoopposite "up"functions, i.e.,"camouflage" vs "seduction." Concerning theirpsychological profiles, results showed thatwomen of thefunctional class "camouflage" aremore anxious, defensive, andemotionally unstable compared tothose ofthefunctional class "seduction," whoappear tobemoresociable, assertive, andextroverted. Further analyses revealed adivision ofthetwoclasses intosubclasses ofvolunteers withopposed personality andpsychological profiles. Thisnew classification allowed ustodefine moreprecisely therelations existing withinthesubjective experience of women during themakeup process. In conclusion, ourstudy revealed thatbeyond thesimple application of colorful products ontheface,makeup hastwomajorfunctional implications depending onspecific psychologicalprofilesof women. INTRODUCTION Variation in thephysical appearance of humans is emphasized byanthropologists asan important factorin thedevelopment ofpersonality andsocial relations. In particular physical attractiveness should playa crucial rolesince it provides easily accessible nonverbalinformation abouta person to others. Physical attractiveness isoneof themost important determinants ofinterpersonal attraction in theearlystages ofmanyrelationships (1-3).Mostof thestudies on"physical attractiveness" focus basically on"facial attractiveness" since manystudies reveal thatfacialfeatures arethemainfactors within Addressall correspondence to RodolpheKorichi. 127 128 JOURNAL OF COSMETIC SCIENCE thewholephysical attractiveness. Forexample, NielsenandKernaleguen (4) showed that facialattractiveness, but not bodyattractiveness, influences subjectiveevaluationof overallphysicalattractiveness, as well as socialand professional satisfaction and social desirability.One of the mostimportantwaystoday'swomenincrease their perceived facialattractiveness is throughthe useof commercialcosmetics. Suchproductsallow womento conformto actualfemininebeautystandards by artificiallymodifyingthe appearance of a setof facialfeatures,e.g., enhancingthe visualimpactof eyesand lips, narrowingeyebrows, reddeningcheeks, dyeinggreyhairs,or maskingwrinklesand"age spots."Moreover,a numberof studiessuggestthat displayingyouthfulor slightly immaturefacialfeatures(e.g.,largeeyes,smallnose,full lips, smallchin,delicatejaw) enhancesfemaleattractiveness (5,6). If cosmetics(standardand corrective)are widely portrayedas improvingtoolsfor facialattractiveness, little research hasbeenpublished to objectivize theirefficiency in thisregard.A reviewof theliteraturereveals onlya few studiesthat assumed that cosmetics canefficientlyandobjectivelyenhance attractiveness and can be usedto manipulatephysicalattractiveness and the implicit messages cos- meticsconveyaboutourselves (7-9). The "psychology of cosmetics" appears therefore as a newfield concerning the characterization of the beneficialeffectsof cosmetic practices (10). It hasbeenarguedthat personality traitssuchaspublicself-consciousness, public body-consciousness, socialanxiety,andfacetsof bodyimagearesystematically relatedto variationsin facial cosmeticuse(11). Moreover,Cashand Cash(12) showedthat women who felt relativelydissatisfied with variousaspectsof their own bodyor with their generalphysicalappearance reportedhigheror recentlyexpanded patternsof cosmetic use.It couldbe arguedthat suchpatternscouldreflecta compensatory effortto correct or balancea flawedself-image. Moreover,in a stimulatingarticle,LevSque (13) revealed a relationshipbetweenappearance andhealth,andemphasized the supportingeffectsof makeupproductson the mentalhealthof womenaffectedby severeillnesses. Fromthe cleardemonstration thatfeelingconfident aboutone'sappearance hasa beneficial impact on one'smood(14), somepositiveinfluencesof cosmetics on the well-beingand selfesteemthat result in positiveemotionalstateshave beenreported.For example,emo- tionswith a positivevalence(e.g., inducedby a pleasantodoror colorfulcosmetics) inducea decrease in heartrate,whereas negativevalencestimuliinduceopposite effects (15). Therefore,sincecosmetics areusedto modifyphysicalappearance andattractiveness,onecouldpredictthat sucha usecouldbe relatedto stablepsychological factors. Hence, the purposeof the presentstudywas to investigatein womenthe underlying correspondence betweenpersonality andself-experience with makeup.The first stageof ourstudyaimedto evaluate thesubjective affective experience of subjects with makeup, usinga self-assessment questionnaire built according to the informationobtainedfrom interviewsof womenon the quality of life and makeup.Thereafter,the secondstage consisted in establishing psychological profilesof our subjectsby usingwell-validated psychometricself-questionnaires. MATERIALS AND METHODS SUBJECTS Seventyfemalesubjectsbelongingto four differentagegroupswererecruitedfor this study:group1:25-34 yr (n = 21); group2:35-44 yr (n = 14); group3:45-54 yr (n = 25); andgroup4:55-65 yr (n = 10). All subjects werecustomary cosmetics users. WHY WOMEN USE MAKEUP 129 EVALUATION OF PERSONALSUBJECTIVEEXPERIENCEWITH MAKEUP A self-assessment questionnaire wasbuilt duringa qualitativesurveyfromtheinterviews with womenabouttheir relationship with makeup.In our study,makeupis the applicationof facialcosmetics, includingfoundations, but alsoeye shadow,lipsticketc... The choiceof productsandmodeof application werefreefor eachsubject.The instruction givento the subjects wasto applymakeupastheyusuallydid, in orderto analyze themoststablemakeuppatternsforeachsubject,andtherefore themorerepresentative. Threedifferent agegroups ofwomen,all makeup products users, wererecruited forthis studyandinterviewed (thefirstgroup:18-30 yr, n = 10; the second group:30-40 yr, n = 10; and the third group >40 yr, n = 10). More than 560 sentences were first generated andclassified into sixdifferentcategories: (i) assertiveness level,(ii) protection, idealizednaturalness, (iii) womenmulti-facets,(iv) aestheticvaluation,(v) seduction,(vi) makeupfor oneself, and(vii) withoutmakeupandhabits.The numberof sentences was thenreducedto 325 according to theirindependence, exhaustiveness, andpertinence. In a secondstep,a questionnaire wasusedduringa consumer studydonewith 57 women 24- to 58-years-old, in orderto selectmorecomprehensive, independent, andrelevant sentences. The definitiveformof thequestionnaire wassetup, with the remaining140 sentences proposed with a five-pointLikert response scalerangingfrom 0 to 4 (0 meaningstronglydisagreeing and4 stronglyagreeing). PSYCHOMETRIC ANALYSES Psychological characterization of the subjects wasdoneusingfourwell-established and validatedpsychometric self-questionnaires: The State-TraitA.xiety I.ve. tory(STAI-T) (16) consists of 20 itemsdesigned to assess trait anxietyandinstructs subjects to report how they "generally"feel by ratingthemselves on a four-pointfrequencyscaleranging from 1 (almostnever)to 4 (almostalways).The Coopersmith Self-Esteem I.ve.tory(SEE adult version)(17) is designedto measureand evaluateone'sself-opinionregarding social,academic, family,andpersonalareasof experience. The inventorycontainsa total of 58 itemsanswered "likeme"or "unlikeme."Scoring yieldsseparate scores forthefour subscales: generalself(26 items),socialself-peers (eightitems),academic (eightitems), andhome(8 items).The scaleallowsthe calculationof a globalself-esteem scorethat can be used as an indicator of self-esteem and valuation of the self. The Rath•s Assertive.ess Schedule (RAS) (18) is used to measureassertiveness level. The scaleincludes 30 items that arescoredon a six-pointLikert-typescale,rangingfrom -3 (leastlike me) to +3 (most like me). The Eyse. c• Perso.a/ityI. ve.tory(EPI, B form) (19) was designedto measuretwo pervasive,independentdimensions of personality,neuroticism-stability (24-items)andextroversion-introversion (24 items).The "extroversion" dimensionrepresentsthe tendencyto be positivelyand activelyengagedwith one'senvironment (interpersonal interactionandsociability), whilethe neuroticism dimensioncharacterizes high levelsof negativeaffect suchas depression and anxiety.The scaleallowsthe calculationof an "extroversionscore"(E) and a "neuroticismscore"(N). STATISTICS Sentences from the makeupquestionnaire werecomparedby the chi-squarestatistical method.The research intothepsychological profilesof thevolunteers wascarriedout by 130 JOURNAL OF COSMETIC SCIENCE usingprincipalcomponent analysis (PCA)andanalysis ofvariance (ANOVA).Statistical significancewas set to a maximum risk, o• = 0.05. RESULTS CLASSIFICATIONOF SUBJECTSFROM SELF-ASSESSMENT QUESTIONNAIRES In ourstudy,datafromLikertscales havebeenreduced to a nominallevelbycombining all responses intotwocategories of "agree" and"disagree." Thesignificant percentage of agreement forspecific sentences is described in TableI. Theresults showtwolargeand distinctive groupsof volunteers. The first(class A) regroups 21 femalesubjects (mean age47 years+ 12), whilethe other(classB) is composed of 49 femalesubjects (mean age43 years+ 10). The responses of the femalesubjects in classA refergloballyto negative self-assessment orwithdrawal fromothers(TableI). Forexample, thesesubjects typically fullyagreewithpropositions suchas:"Withoutmakeup, I don'tlike myself" (62%) or "I alwaysmakemyselfup whenI am in contactwith people"(95%). On the otherhand,the responses of classB subjectsreferto a positiveself-assessment or approach to others(TableI). Forexample,thesesubjects typicallyfully agreewith such propositions as:"I makemyselfup in a moreintensivewayin the evening"(92%) or "WhenI'm made-up,I feelsensual" (76%). Overall,thepatternof responses of classA subjects regarding makeupusecanbeconsidered asa "camouflage" profile(i.e.,aiming Table I Analysisof the MakeupQuestionnaire (n = 70) Percentage of agreement for specificsentences Class A Class B Sentences (n = 21) (n = 49) Significance WhenI'm made-up, I feelglamourous With makeup, I wantto smile I always makemyself upwhenI amin contact withpeople WhenI'm made-up, I feelnatural WhenI'm made-up naturally, I feelmyself I'm notafraid to change mymakeup process to remain 25^ 14^ 95• 43^ 81^ 71• 63B 47B 61^ 33^ 86^ 20^ S(p< 0.01) S(p= 0.02) S(p< 0.01) NS NS S(p< 0.01) 57^ 76u 48^ 76^ 24^ 92• 47^ 67^ 55^ 10^ S(p< 0.01) S(p= 0.05) NS NS NS 48^ 38^ 76• 62u 43^ 59^ 69B 47^ 33^ 76• faithful to myself I makemyself upin a moreintensive wayin theevening A woman whodoes notmakeherself up,it isa pity In theevening, I makemyself upto allure Withoutmakeup, I'm insipid A woman whodoes notmakeherself upisa woman who overrates intellectual dimension to the detriment of appearance I usemakeup tocontrol theimage ofmyself WhenI'm made-up, I wantto allure I don'tlikemyfacewithoutmakeup Withoutmakeup, I don'tlikemyself WhenI'm made-up, I feelsensual NS S (p = 0.03) s (p = 0.05) S (p = 0.04) s (p = 0.02) When thereis a significant difference betweenclasses, we indicateby letters(A, B) the classto whichit belongs. If the classes havethe sameletter,theyarecomparable; if not,theyaresignificantly different. WHY WOMEN USE MAKEUP 131 at decreasing the loadof theirnegatively perceived attractiveness and/orglobalselfimage),whilethoseof class B subjects referratherto a "seduction" functional profile (i.e.,aimingat supporting andpromoting a positiveself-image). If we analyze more precisely thesetwoclasses, weobserve a statistically significant division of the"seduction" class(i.e., classB) into three differentsubclasses (Table II), dependingon the relative valuation ofthenaturalandmade-up faceofeachsubject. Thesameanalysis for class A did not reveal sustainable subclasses. In the firstsubclass (subclass B1) of classB, the made-upfaceis considered to bemore valuablethanthenaturalfaceandlessvariabilityis introduced in themakeupstructure. Forexample, thesesubjects claim"Withoutmakeup, I'm insipid"(81%),"I don'tlike my facewithoutmakeup" (69%),or "WhenI'm made-up naturally,I feelmyself" (100%). For the third subclass (subclass B3) and in comparison to subclass B1, the "natural"faceis ratherperceived asmorevaluablethanthemade-upface.Forinstance, thesesubjects typically agreewith"I usemakeup to control theimageofmyself" (30%) or "In the evening, I makemyselfup to allure"(100%).Concerning the intermediate subclass (subclass B2), "natural" and"made-up" faces areconsidered asvaluable aseach otherrelative to thesubject's personal andprofessional life.Forexample, womenofthis subclass claim"I alwaysmakemyselfup whenI am in contactwith people"(70%), Table II Analysis of theMakeupQuestionnaire for ClassB (n = 49) Percentage of agreement for specificsentences SubclassB1 SubclassB2 SubclassB3 Sentences (n = 16) (n = 23) (n = 10) WhenI'mmade-up, I feelglamourous Withmakeup, I wanttosmile I always make myself upwhen I amin contact 56B 44^• 56^ 61B 44^• 70^B 80• 60• 50^ s (•: 0.04) 48B 87^• 26^ 40^• 60^ 20^ S (•: 0.04) 88^• 91B 100B S (p < 0.01) 56^• 35^ 60^• NS 44^ 81B 0^ 70^• 39^ 4^ 100• 50^• 40^• s (p = o.oD s (p = 0.03) s (p = o.oD 75• 75B 69• 50B 69^• 61^• 61^• 35^ 17^ 74^• Significance NS S (• = 0.03) with people WhenI'mmade-up, I feelnatural WhenI'mmade-up naturally, I feelmyself I'm notafraid tochange mymakeup process to 6^ 100• 13^ S (• = 0.05) S • < 0.01) remainfaithful to myself I make myself upin a moreintensive way in the evening A woman whodoes notmakeherself up,it is a pity In theevening, I make myself uptoallure Withoutmakeup, I'm insipid A woman whodoes notmakeherself upisa woman who overrates intellectual dimension to the detrimentof appearance I usemakeup tocontrol theimage ofmyself WhenI'mmade-up, I wanttoallure I don'tlikemyface without makeup Withoutmakeup, I don'tlikemyself WhenI'm made-up, I feelsensual 30^ 80• 40AB 40A• 90• NS NS s (• = 0.04) s (• = 0.04) s (p = o.oD Whenthereis a significant difference amongclasses, we indicate by letters(A, B) theclass to whichit belongs. If theclasses havethesameletter,theyarecomparable; if not,theyaresignificantly different. 132 JOURNAL OF COSMETIC SCIENCE but theyalsoenjoytheirfacewithoutmakeup("Withoutmakeup,I don'tlike myself" 17%). Finally, Table III summarizesthe classifications detailed above. ORGANIZATION OF PSYCHOMETRICVARIABLESIN SUBJECTS As a first result,the ANOVA revealed,interestingly,that psychological variablesobtained from psychometric self-questionnaires were not statisticallyrelatedto aging (TableIV). Thereafter, a principalcomponents analysis (PCA)wascarriedout in order to furtherinvestigate the relationsbetweenmeasured psychometric parameters, andto reducedatadimensionality. The resultsperformedon the nine psychometric variables retained (i.e., the STAI score,the extroversionscore,the neuroticismscore,the RAS score,the SEI global score,and the four SEI sub-scores) are shownin the correlation circle represented in Figure l a. The PCA identifiedthree main factorsthat explain 71.8% of the total variance. The first two factors account for 62.6% of the initial variability of the data, while factor 3 accountsfor 9.1% of the total variance.Their projectionin the factors'spaceshowsthat the axiscorresponding to factor1 (accounting for 51.6% of the totalvariance) is essentially relatedto the SEI generalscore,the RAS score,and the STAI T-anxietyscore,which is inverselycorrelatedto the SEI general score.The axiscorresponding to factor2 represents 11.1%of the remainingvariance and is relatedto the EPI extroversion andneuroticism scores. The projectionof the psychometricvariablesin the otherfactors'space,represented in Figure1b, showsthat the axis corresponding to factor3 is only relatedto the SEI home-parents scoreandindependent of the otherspsychometric variables. Theseresultsshowclearlythat five psychometric variablesout of the nine are ableto describesatisfactorily the differences betweenthe volunteers. Note that the projectionof the psychometric variables in the factors'spaces (F1 vs F2 and F1 vs F3) confirmsthat the variable"age"is independentof all these variables.Concerningthe psychometric classification of our subjects,significantdifferencesbetweenclasses(classA and classB) were observedfor the five psychometric variables(seeTableV for a statisticalsummary). For generalself-esteem,the mean(+ SD) scorewas 12.4 (-+3.9)for the volunteersof class A (n = 21) and 20.6 (+2.4) for the volunteersof classB (n = 49). Theseresultsreveal that the femalesubjectsof classB express a betterself-esteem than thoseof classA. Concerningthe meanof the STAI T-anxietyscores,data obtainedin classA were significantly higherthanthoseof classB, with respective scores of 46.6 (+8.1) and35.6 (-+5.5).The volunteersof classA canthereforebe labeledas"anxious," oppositethoseof classB. Concerning the RASscore,participants of classA (-5.2 + 16.2)werelabeledas Table III Classification of Volunteers Accordingto Their Useof the MakeupProcess (n = 70) ClassB (n = 49) Seduction ClassA (n = 21) Main background for makeup Camouflage Subclass B1 Subclass B2 Subclass B3 (n = 16) (n = 23) (n = 10) Self-perceived valueof Naturalimage< Natural image < Natural image = Natural image > naturalfaceand made-upimage made-upimage made-upimage made-upimage make-upface WHY WOMEN Table USE MAKEUP 133 IV RelationBetweenPsychometric VariablesandAgeGroups Group1 Group2 Group3 (n = 21) (n = 14) (n = 25) (n = 10) (50yr + 3) (59yr + 3) (30.2yr + 4) (40.2yr + 3) Variables Group4 Mean +SD Mean +SD Mean +SD Mean +SD Significance STAI(T-anxiety state) 37.3 6.2 36.8 8.5 40.8 8.7 40.0 9.5 NS SEI(General selfscore) 18.6 4.7 18.4 4.3 17.7 4.8 17.7 5.7 NS SEI(Social self-peers score) 6.6 SEI(Home-parents score) 6.7 1.6 1.5 6.9 6.5 0.8 1.5 6.2 5.9 1.7 1.8 5.8 5.3 1.6 1.9 NS NS SEI(Academic score) 6.4 SEI(Totalself-esteem score) 40.3 1.2 7.4 6.1 40.9 1.4 6.3 6.3 39.3 1.7 5.5 1.9 9.4 38.2 11.4 NS NS RAS score 19.0 20.5 22.9 24.4 14.4 EPI (formB) 14.1 3.7 16.4 3.6 13.8 8.9 5.4 10.6 3.8 9.6 20.8 27.8 NS 3.2 12.8 8.9 4.3 NS 3.9 4.7 NS Extroversion-introversion EPI (formB) 8.0 Neuroticism-stability "non-assertive," whilethoseofclass B (26.1 + 17.8)werelabeledas"strongly assertive." Theanalysis of "extroversion" and"neuroticism" variables, whicharetwocoredimensions ofhuman personality, revealed some interesting results. Indeed, weobserved aclear opposition in thepersonality traitsbetween ourtwoclasses of volunteers. Themean neuroticismscoreof classA wasfoundto be lowerthan in classB (respectively, 11.9 + 2.8 and15.3+ 3.6),whilethemeanextraversion score ofclass A washigherthanin class B (respectively, 13.5+ 3.4and7.6 + 3.6).Thisinteraction between extroversion and neuroticism, knownto bea strong predictor of satisfaction withlife,mood,andsubjective well-being (20),tendsto showthatthevolunteers ofclass A (higherin neuroticismandlowerin extroversion thanthoseof classB) reportmorenegativeemotional experiences. Overall, ourfindings reveal a cleartwo-class division, withsubjects ofclass B appearing asmoresociable, optimistic, calm,emotionally stable, non-anxious, and assertive thanthoseof class A, whoaremorereserved, anxious, of a lowerself-esteem, and non-assertive. Concerning thesubclasses ofvolunteers, weobserved opposed personality andpsychologicalprofiles (cf.TableVI). Firstof all, no significant differences wereobserved between subclasses (B1,B2, andB3) fortheSTAIT-anxietyscale. FortheSEIgeneral self-esteem score, volunteers of subclass B1 andsubclass B2 expressed a significantly lowermeanself-esteem score(19.4 + 1.9 and20.0 + 2.1) thanthoseof subclass B3 (23.6+ 1).Asfortheassertiveness level,theRASmean score ofsubclass B3(42+ 15.4) wassignificantly higherthanthose ofsubclass B2 (26.7+ 16.5)andsubclass B1 (15.3 + 13.3).Concerning theindependent dimensions ofpersonality, themeanneuroticismstability score ofsubclass B3wasfoundtobesignificantly lowerthanthatofsubclasses B2 andB1 (respectively, 3.1 + 1.7, 9.1 + 2.9, and8.1 + 3.2).Moreover, themean extroversion scores of subclasses B3 andB2 werehigherthatthoseof subclass B1, with scoresof 17 + 2.3 and 11.7 + 3, respectively. Overall,theseresults suggest thatfemalesubjects of subclass B3 aremoresociable, optimistic, calm,emotionally stable, andassertive, withahigher self-esteem, thanthose of subclassB1. In the caseof the intermediatesubclassB2, the volunteershad a 134 JOURNAL OF COSMETICSCIENCE Variables (F1 and F2: 62.6%) • B)-extraversion-introversion PI (form B)-neuroticism-st bility self-peersscore) S! self score) ' SEI (acad Fl Variables (F1 and F3: 60.6%) RAS o• !mic) STAI(T-anxiety state) (form O• EPI(form ' traversion-introversion •lf-peersscore) (generalselfscore) (home-parentsscore) F1 (51.5%) Figure1. PCAanalysis ofpsychometric variables cakingintoaccount (a)twofactors, F1 andF2 (62.6%of thetotalvariance), and(b) twofactors, F1 andF5 (60.6%of thetotalvariance), based uponninevariables. A correlation circleidentifies theparameters responsible œor thesevariances. WHY WOMEN USE MAKEUP Table 13 5 V Identificationof Two Significance Classes of Volunteers Variables Class A (n = 21) ClassB (n = 49) Mean Mean + SD + SD Significance STAI(T-anxiety state) 46.6B 8.1 35.6^ 5.5 SEI (Generalselfscore) 12.4^ 3.9 20.6B 2.4 RAS score 17.8 - 5.2^ 16.2 26.1• EPI FormB (Extroversion-introversion) 11.9^ 2.8 15.3• 3.6 EPIFormB (Neuroticism-stability) 3.4 7.6^ 3.6 13.5• S (p < 0.01) S (p < 0.01) S (p < 0.01) S (p < 0.01) S (p < 0.01) When thereis a significant difference betweenclasses for eachpsychometric variable,we indicateby letters (A, B) the classto whichit belongs.If the classes havethe sameletter,theyarecomparable; if not, theyare significantlydifferent. psychological profile relativelycloseto that of subclass B3, but with a higher level of assertiveness and emotional stability. The same analysisfor classA did not reveal sustainablesubclasses. Indeed, the statisticalsubdivisionsrevealeda main subclass that regroups90% of the subjectsof classA, with a response profilesimilarto the general classA profile. DISCUSSION The first resultof our studyhasclearlyshownthat the womenof ourpanelexpressed througha self-assessment questionnaire two oppositefunctionsin facialmakeup.The firstfunctionis ratherintendedto decrease a negativeself-perception, whichwe labeled as "camouflage." The secondis moreorientedto the desireto please,which we labeled as "seduction," with some variations. For one subclassin which the "natural" face is perceived asmorevaluablethanthe made-upface,makeupis ratherusedin "playful"or "enjoyable" contextsand, hence,variabilityin its featuresis allowed.For the other subclass, the made-upfaceis considered asmorevaluablethanthe naturalfaceandtends to supportsthe self-image,thusrestrictingvariabilityin the makeupstructure.In order to gofurtherin theunderstanding of thesefunctionaldifferences, we triedto relatethese Identification Table VI of Three Subclasses of Volunteers ClassB (n = 49) Variables STAI(T-anxiety state) Subclass B1 Subclass B2 Subclass B3 (n = 16) (n = 23) (n = 10) Mean +SD Mean _+SD Mean _+SD Significance 38.3^ 3.9 35.1^ 5.5 32.2^ 6.3 SEI (Generalselfscore) 19.4^ RAS score 15.3^ EPI FormB (Extroversion-introversion) 11.7^ 1.9 13.3 3.0 20.0^ 26.7B 17.0• 2.1 16.5 2.3 23.6• 42.0c 17.0• 1.0 15.4 2.3 EPIFormB (Neuroticism-stability) 8.1• 3.2 9.1• 2.9 3.1^ 1.7 NS S (p < 0.01) S (p < 0.01) S (p < 0.01) S (p < 0.01) When thereis a significantdifferencebetweensubclasses for eachpsychometric variable,we indicateby letters(A, B, C) the subclass to whichit belongs.If the subclasses havethe sameletter,theyarecomparable; if not, theyare significantlydifferent. 136 JOURNAL OF COSMETICSCIENCE differentmakeupfunctionsto possibleunderlyingpsychological features(i.e., selfesteem,socialdesirability,anxiety,andfearof negativeself-evaluation). Our first result clearlyshowedthat agingwas not a discriminating factorin our differentmakeup functions, andhadno influence on the volunteer's psychological characteristics. This interestingresult is consistentwith thoseof McCrae and Costa (21), who tested the personalities of individuals between19 and80 yearsforovertwelveyearsandspecifically measured theirlevelsof neuroticism, extroversion, openness to experience, agreeableness, andconscientiousness. Theauthors concluded thatthesefivepersonality traitsremained relativelystablewith age.However,theyconceded that theirstudies werenotdefinitive andthatvariabilityacross the individualpersonality wasstill possible. Concerning the psychological profilesof our subjects, we observed that womenusing makeupasa camouflage tool(class A) areratherconcerned with anxietyandneuroticism, while thoseusingmakeupas a "seduction" tool (classB) are rathercharacterized by higherself-esteem, extroversion, andassertiveness. According to the literature(22,23), female subjects ofclass A canbeclassified ashavinga negative self-perception, worrying moreoften,anddwellingonfrustrations anddisappointments. Moreover, subjects with higherneuroticism valueswereshownto be moredistressed on average in comparison with individuals with lowervalues, andaremoresusceptible to stressful events (24).On the otherhand,femalesubjects of classB tendto perceivethemselves asbetterthan averagein communaltraits, with more experience of positiveemotions,definedas sociabilityor a tendencyto be activeand social(25). Fromtheseresults,it is clearthat one'sself-image playsa key role in the development of personality. Womenwith a subjective negative feelingabouttheirimagedevelop defensive mechanisms tocopewith low self-esteem andmayneedto "normalize" or manipulate a perceived impairedappearance. CONCLUSIONS The subjective approach from the self-assessment questionnaire revealedtwo clearly distinctive classes of subjects according to theirfunctional useof makeup,i.e., "camouflage"vs "seduction." Thesetwo classes (and furthersubclasses) havebeenassociated with specific emotional andpsychological profiles.It is clearthatournextstepwill be to furtherstudytheimpactof relatedphysical parameters suchasskinradiance, homogeneityof skincolor,andfacialsymmetry (26), aswell asfacialexpression patterns, alongwith the makeupprocess. Finally,we canconclude that beyondthe simpleapplication of colorfulproductsto the face,makeup appears asa holistic technique thatmodifies notonlyone's appearance, but alsohelpsoneto copewith self-image, emotions, andmood.Therefore, makeupapplicationcanbe considered asa dailyroutineto decrease negativeaffectsand/orincrease positiveaffectsrelatedto self-image and one'srelationto the socialenvironment. Our resultsprovideexperimentalsupportto the link betweencosmetics and welfare,and furtherpromote initiatives suchasthe"LookGood. . . FeelBetter"program thatwas developed in 1989by theCosmetic, Toiletry,andFragrance Association (CTFA).Such a programconsists in a free,non-medical, brand-neutral, nationalpublicservice program supported by corporate donorsto helpwomenoffsetappearance-related changes from cancertreatment.This pioneerstudy, revealinga psycho-behavioral backgroundfor differences in the useof makeup,urgesfurtherinvestigation in orderto determine underlyingdeterminants. WHY WOMEN USE MAKEUP 137 ACKNOWLEDGMENTS We aregratefulto all volunteers fortheirpatience duringourexperiments. Theauthors thankDr F. Vial ofSpincontrol forproductive discussions andactiveparticipation in this work. REFERENCES (1) (2) (3) A. Marwick,Bea•tyin History(Thames& Hudson,Great Britain, 1988). N. Etcoff,S•rviva/ofthePrettiest: TheSciences ofBea•ty(Little Brown& Co, London,1999). K. A. Nakdimen, Thephysiognomic basis ofsexual stereotyping, Am.J. Psych., 141,499-503(1984). (4) J.P. NielsenandA. Kernaleguen, Influenceof clothingandphysicalattractiveness in personpercep- tion, Percept. RiotorSkil/s,42, 775-780 (1976). (5) M.R. Cunningham, Measuring the physicalin physical attractiveness: Quasi-experiments on the sociobiology of femalefacialbeauty, J. Person. Soc. PsychoL, 50, 925-935 (1986). (6) C.F. Keating, Gender andthephysiognomy ofdominance andattractiveness, SocPsychoL Q•art.,48, (7) R. Mulhern,G. Fieldman, T. Hussey, J. L. L•v&que, andP. Pineau,Do cosmetics enhance Caucasian (8) J.A. GrahamandA.J. Jouhar,Cosmetics considered in the contextof physical attractiveness: A (9) review,Int. J. Cosmet. Sci.,2, 77-101 (1980). C. L. CoxandW. H. Glick, Resumeevaluations andcosmetics use:When moreis notbetter,Sex 312-323 (1985). facialattractiveness, Int. J. Cosmet. Sci.,25, 199-205 (2003). 14, 51-58 (1986). (10) J. A. GrahamandA. J. Jouhar,The effects of cosmetics on person perception, Int.J. Cosmet. Sci.,3, (11) L. C. MillerandC. L. Cox,Forappearances' sake:Publicself-consciousness andmake-up use,Person. 199-210 (1981). SocPsychoL B•//., 8(4), 748-751 (1982). (12) T. F. CashandD. W. Cash,Women'suseof cosmetics: Psychosocial correlates andconsequences, J. Cosmet. Sci.,4, 1-14 (1982). (13) (14) (15) J. L. Lev•que, Apparence et sant&Ler$ledescosm•tiques, Rev.Ried.Liege, 11,721-725(1996). J. A. Graham,Psychology ofCosmetic Treatments (Prager, London,1986). S. Barkat,T. Thomas-Danguin, M. Bensail, M. Rouby,andG. Sicard,Odorandcolorof cosmetic products: Correlations between subjective judgments andautonomous nervous system response, Int.J. Cosmet.Sci., 25,273-283 (2003). (16) C. D. Spielberger, Rian•a/fortheState-Trait Anxiety Inventory (STAI)(Consulting Psychologists Press, Palo Alto, CA, 1983). (17) J. MyhillandM. Lorr,TheCoopersmith self-esteem inventory: Analysis andpartialvalidation of a (18) (19) (20) S.A. Rathus, A 30-itemschedule forassessing assertive behavior. Behar.Ther.,4, 398-406(1973). H.J. Eysenck, "Biological Dimension of Personality," in Handbook ofPersonality: Theory andResearch, 2nded.,L. A. PervinandO. P. John,Eds.(GuilfordPress, NewYork,1999),pp. 244-276. E. Diener,E. M. Suh,R. E. Lucas, andH. L. Smith,Subjective well-being: Threedecades of progress, (21) (22) R. R. McCraeandP. Costa, Jr., Personality in AdMthood (GuilfordPress, NewYork, 1990). W. E. Kelly,Examining therelationship between worryandtraitanxiety, College St•dentJ. (September (23) D. WatsonandL. A. Clark,Negativeaffectivity: The disposition to experience aversive emotional (24) N. BolgerandE. A. Schillings, Personality andtheproblems ofeveryday life:Theroleof neuroticism in exposure andreactivityto dailystressors,J. Person., 59, 355-386 (1991). (25) P. T. Costaand R. R. McCrae,NEO PI-R Professional Rian•a/(Psychological Assessment Resources, modifiedadult form,J. C/in. Psycho/., 34(1), 72-76 (1978). PsychoL BM/., 125,276-302 (1999). 2004). states,PsychoL BMI., 96, 465-490 (1984). Odessa,FL, 1992). (26) D. I. Perrett,D. M. Burt, I. S. Penton-Voak, K.J. Lee,D. A. Rowland,and R. Edwards,Symmetry and human facialattractiveness, EvoLH•man Behar.,20, 295-307 (1999).