Different Types of Well-Being? A Cross-Cultural Examination of Hedonicand Eudaimonic Well-Being

David J. Disabato, Fallon R. Goodman,Todd B. Kashdan, and Jerome L. Short

George Mason University

Aaron JardenAuckland University of Technology

A large international sample was used to test whether hedonia (the experience of positive emotional statesand satisfaction of desires) and eudaimonia (the presence of meaning and development of one’spotentials) represent 1 overarching well-being construct or 2 related dimensions. A latent correlation of.96 presents negligible evidence for the discriminant validity between Diener’s (1984) subjectivewell-being model of hedonia and Ryff’s (1989) psychological well-being model of eudaimonia. Whencompared with known correlates of well-being (e.g., curiosity, gratitude), eudaimonia and hedoniashowed very similar relationships, save goal-directed will and ways (i.e., hope), a meaning orientation tohappiness, and grit. Identical analyses in subsamples of 7 geographical world regions revealed similarresults around the globe. A single overarching construct more accurately reflects hedonia and eudaimoniawhen measured as self-reported subjective and psychological well-being. Nevertheless, measures ofeudaimonia may contain aspects of meaningful goal-directedness unique from hedonia.

Keywords: well-being, happiness, positive psychology, eudaimonia, meaning in life

The ability to meaningfully distinguish between types of well-being is debatable. A common framework of well-being distin-guishes between hedonic well-being (hedonia) and eudaimonicwell-being (eudaimonia; Ryan & Deci, 2001). According to Aris-tippus (3rd century BCE), the Greek philosopher and founder ofthe Cyrenaic school of philosophy, hedonia represents the maxi-mization of pleasure and minimization of pain (Tatarkiewicz,1976), whereas eudaimonia represents human flourishing and liv-ing up to one’s full potential (i.e., self-actualization). Eudaimoniastems from Aristotle’s 4th-century BCE (1985) conceptualizationof well-being that extends beyond pleasure-driven happiness toencapsulate being true to oneself and working toward personalgrowth. Although this distinction between hedonia and eudaimoniahas been helpful in describing and directing research on well-being, there is questionable evidence for its theoretical utility andempirical support (Coyne, 2013; Kashdan, Biswas-Diener, &King, 2008, Kashdan et al., 2009; Sheldon, 2013). A systematicunderstanding of well-being is imperative as research and large-scale initiatives develop in line with the positive psychologymovement (e.g., comprehensive soldier and family fitness; Eidel-

son, Pilisuk, & Soldz, 2011). Before executives and policymakersallocate resources toward improving human well-being, a morecomprehensive, scientifically based understanding of what exactlyhumans are striving for is needed. The present study assessed theextent to which hedonia, operationalized as subjective well-being,and eudaimonia, operationalized as psychological well-being, rep-resent distinct types of well-being.

Theoretical Conceptualizations of Hedoniaand Eudaimonia

Psychologists traditionally hold to the philosophical conceptu-alization of hedonia stated above, which departs from a lay defi-nition of hedonia. The pleasure that is maximized refers to a broadarray of human experiences that range from having sex and eatingfood, to feeling excited for a new experience. Psychologists havebuilt on these conceptualizations to define hedonia as how satis-fying one evaluates his or her life to be. The most widely usedconceptualization of hedonia is Diener’s (1984) tripartite model ofsubjective well-being, which contains three components: life sat-isfaction, and the balance between positive and negative affect (fora recent review, see: Busseri & Sadava, 2011). Many studies haveused components or variations of this model to measure well-being(e.g., Vittersø, 2003).

Hedonic definitions of well-being are incomplete for somescholars who have argued that well-being cannot be reduced toonly immediately gratifying experiences (Ryff, 1989; Waterman,1993). Research on eudaimonia draws upon Aristotle’s distinctionbetween pleasure and the good life, with Aristotle defining a goodlife as living to one’s fullest potential in accordance with virtue orexcellence. Unlike research on hedonia, there is less agreementamong modern researchers on what constitutes eudaimonia (Huta& Waterman, 2014). To date, there is no single agreed-upon theoryor methodological approach to studying eudaimonia.

This article was published Online First September 7, 2015.David J. Disabato, Fallon R. Goodman, Todd B. Kashdan, and Jerome L.

Short, Department of Psychology, George Mason University; AaronJarden, Department of Psychology, Auckland University of Technology.

Todd B. Kashdan was financially supported as a Senior Scientist of theCenter for the Advancement of Well-Being, George Mason University.Fallon R. Goodman was supported as Doctoral Research Fellow by thesame center.

For additional developments, see www.toddkashdan.com.Correspondence concerning this article should be addressed to Todd B.

Kashdan, Department of Psychology, MS 3F5, George Mason University,4400 University Drive, Fairfax, VA 22030. E-mail: tkashdan@gmu.edu

Thi

sdo

cum

ent

isco

pyri

ghte

dby

the

Am

eric

anPs

ycho

logi

cal

Ass

ocia

tion

oron

eof

itsal

lied

publ

ishe

rs.

Thi

sar

ticle

isin

tend

edso

lely

for

the

pers

onal

use

ofth

ein

divi

dual

user

and

isno

tto

bedi

ssem

inat

edbr

oadl

y.

Psychological Assessment © 2015 American Psychological Association2016, Vol. 28, No. 5, 471–482 1040-3590/16/$12.00 http://dx.doi.org/10.1037/pas0000209

471

Extant models of eudaimonia vary widely, but they typically have twopoints of convergence: some component of personal meaning andgrowth, and the explicit exclusion of an affect component (Ryan &Deci, 2001). One of the most common theories of eudaimonia ispsychological well-being, which equates to positive functioning (Ryff& Singer, 1998). This theory conceptualizes eudaimonia as a broadtype of well-being rather than a specific adaptive trait (e.g., authen-ticity; Waterman, 1993). The facets of psychological well-being stemfrom the historically rich humanistic–existential psychology litera-tures, including Jahoda’s (1958) focus on the positive side of mentalhealth. The psychological well-being model has been used to studyrelationships between well-being and personality traits, mental andphysical health, healthy aging, family and occupational experiences,and neurological processes (Ryff, 2014).

Different Types of Well-Being?

Despite the popularity of the hedonia–eudaimonia distinctionwithin the well-being literature, there is skepticism about its em-pirical support and theoretical utility (Coyne, 2013; Kashdan et al.,2008, Kashdan et al., 2009; Sheldon, 2013). The primary criticismis a lack of discriminant validity between hedonia and eudaimonia.Discriminant validity is crucial for the measurement of global andabstract constructs such as well-being (Fiske, 1982). Several stud-ies of hedonia and eudaimonia have revealed very large correla-tions ranging from .76 to .92, which shed doubt on their discrim-inant validity. Four prior factor-analytic studies have measuredhedonia and eudaimonia with Diener’s (1984) tripartite model ofsubjective well-being and Ryff’s (1989) model of psychologicalwell-being, respectively. Two of these studies used a large, na-tionally representative sample of middle-aged U.S. adults andfound that hedonia and eudaimonia were highly correlated, r � .84(Keyes, Shmotkin, & Ryff, 2002) and r � .78 (Gallagher, Lopez,& Preacher, 2009, Study 2). In a study among U.S. undergradu-ates, hedonia and eudaimonia were correlated, r � .92 (Gallagheret al., 2009, Study 1). A fourth study of U.K. adults found thathedonia and eudaimonia were correlated, r � .76 (Linley, Maltby,Wood, Osborne, & Hurling, 2009). In a more recent study, re-searchers found an observed correlation, r � .79 between happi-ness (i.e., hedonia) and flourishing (i.e., eudaimonia) in a conve-nience sample of healthy adults (Fredrickson et al., 2013). Theseprior psychometric studies of the subjective well-being model ofhedonia and the psychological well-being model of eudaimoniasuggest substantial overlap between the two constructs.

Hedonia, Eudaimonia, and Correlates of Well-Being

If hedonia and eudaimonia are indeed separate constructs, certaincorrelates of well-being should differentially correlate with each typeof well-being. For example, hedonia might be more strongly related toa pleasure orientation to happiness because of the emphasis on greaterpositive emotional states (Kahneman, Diener, & Schwarz, 1999) andto rumination because it often covaries with negative emotional ex-periences (Keyes, 2005). Similarly, eudaimonia might be morestrongly related to a meaning orientation to happiness and the searchof meaning in life because of eudaimonia’s emphasis on meaning andpurpose (Ryan & Deci, 2001; Ryan, Huta, & Deci, 2008), to gratitudeand loneliness because they involve the quality of social relationships(Ryan & Deci, 2000), to curiosity and engagement because they

involve challenging oneself and searching for personal growth(Compton, Smith, Cornish, & Qualls, 1996), and to hope and gritbecause they involve goal-directedness and striving to reach one’spotential or intrinsic motivation (Ryan & Deci, 2000, 2001). A lack ofdifferential relationships between hedonia and eudaimonia in the theoret-ically expected direction would cast doubt on the distinction between thetwo types of well-being.

The Present Study

The primary aim was to test the discriminant validity betweentwo proposed types of well-being—hedonia and eudaimonia—measured by subjective and psychological well-being, respec-tively. We have built on prior studies that have tested the distinc-tion between hedonia and eudaimonia in three ways. First, in thisstudy we replicated previous findings using similar measurementmodels that found low discriminant-validity coefficients (i.e., highlatent correlations) between hedonia and eudaimonia. Second, wessessed the degree to which known correlates of well-being dif-ferentially related to separate hedonia and eudaimonia factors.Third, we used an international sample to assess well-being on aglobal level, as all prior studies have been restricted to the U.S. andUK. The international sample allowed us to test whether measure-ment models of well-being differed across world regions.

Method

Participants and Procedure

Data from the International Wellbeing Study (www.wellbeingstudy.com) were collected between March, 2009 and March, 2013.The sample consisted of 7,6171 late adolescent and adult partici-pants (aged 15 years and older2). Age showed a positively skeweddistribution, with M � 33.5 and SD � 14.2. Over half of theparticipants were married (54%) and most were women (79%).Participants were compensated with a chance to win one of 15$100 vouchers, an email summary of their score reports comparedwith others, and the opportunity to take one of three differentwell-being courses. Participants were recruited through emailedand printed advertisement posters distributed to study investiga-tors’ work colleagues, university departments, businesses, charita-ble organizations, and online forums including listservs.

Organization Into World Regions

Participants were from 109 different countries from six of theseven world continents (Antarctica excluded) and completed theonline questionnaires in 16 different languages. The 109 countrieswere grouped into nine world regions to make meaningful com-

1 The final sample size reflects participants who completed all measures.Of the participants who started the online survey, 82% completed allmeasures. Unfortunately, demographic information comparing completersand noncompleters was not available because demographic questions wereasked at the end of the survey.

2 Our total sample contained people (8.9%, n � 679) who were youngerthan 18 years old. The majority of adolescents were from Finland, Slove-nia, the Philippines, Greece, and New Zealand. Principle investigatorAaron Jarden received ethical permission to include late adolescents fromthe Open Polytechnic of New Zealand Ethics Committee, September, 2009.

Thi

sdo

cum

ent

isco

pyri

ghte

dby

the

Am

eric

anPs

ycho

logi

cal

Ass

ocia

tion

oron

eof

itsal

lied

publ

ishe

rs.

Thi

sar

ticle

isin

tend

edso

lely

for

the

pers

onal

use

ofth

ein

divi

dual

user

and

isno

tto

bedi

ssem

inat

edbr

oadl

y.

472 DISABATO, GOODMAN, KASHDAN, SHORT, AND JARDEN

parisons. Because very few country groupings led to measurementinvariance and the desired degree of homogeneity, a data-drivenapproach of organizing countries into world regions was inappro-priate. A theoretical approach was chosen based on participants’country of residence. Initial world region groupings were selectedaccording to the CIA Factbook, a reputable source of world infor-mation (see www.cia.gov). Past international well-being researchhas used the CIA Factbook’s world regions to help group interna-tional samples (e.g., Tay & Diener, 2011). The CIA Factbookorganizes countries primarily by geographical proximity, whichexcludes important contextual information. Categorizations ofcountries for the current study were made based on additionalinformation about common historical, cultural, linguistic roots, andcurrent economics. For example, due to Mexico’s colonial history,Latin culture, Spanish language, and less developed economy, itwas separated from North American countries (e.g., United States)and grouped with South American countries. The United Statesand Canada were grouped with Northern Europe because of theirsimilar ancestral roots and economic development. Southeast Asiawas separated from East and South Asia because of its greatercolonial and Judeo-Christian influence. South Asia was combinedwith East Asia because of the limited South Asian subsample size(n � 48). Participants from Africa and the Middle East wereexcluded from subsample analyses due to insufficient sample sizes(ns � 64, 49, respectively). The final seven world regions aredisplayed in Table 1.

Measures

The total assessment battery from the International WellbeingStudy (www.wellbeingstudy.com) contained 20 scales (235items), which took participants an average of 29 min. Of these, 12scales (159 items) were used for the present analyses.3 Mostparticipants completed the assessment battery in English (59%).Where scales were already available in one of the desired 16languages, that language translation was used. Where no transla-tion was available from the English version to the required lan-guage, scales were translated by a native speaker of that languagewho had a degree in psychology or higher (most translators weremaster’s- or doctoral-level students in psychology familiar withpsychometrics). Scales were then independently cross-checkedafter translation by a second translator and areas of disagreementidentified and resolved between the two translators.

Hedonia

Satisfaction with life. The Satisfaction With Life Scale (Die-ner, Emmons, Larsen, & Griffin, 1985) is a five-item measure ofpresent, global life satisfaction, which comprise a cognitive judg-ment of a person’s quality of life. Participants responded on aLikert scale from 1 � strongly disagree, to 7 � strongly agree. Anexample item is, “I am satisfied with my current life.” Cronbach’s� ranged from .89 to .91 across world regions.

Happiness. The Subjective Happiness Scale (Lyubomirsky &Lepper, 1999) is a four-item measure of global happiness. The firstitem assesses absolute happiness, the second item assesses relativehappiness compared with peers, and the last two items assess howwell characteristics of happiness describe each participant’s self.Participants responded to items on a Likert scale from 1 to 7. The first

two item anchors are from 1 � less happy to 7 � more happy, and thelast two item anchors are from 1 � not at all to 7 � a great deal.Cronbach’s � ranged from .74 to .86 across the world regions.

Depression. The Center for the Epidemiological Studies De-pression Scale (CES-D; Radloff, 1977) is a 20-item measure ofdepressive symptoms that emphasizes affective components ofdepression (see Ensel, 1986, for an overview of the measure).Participants responded to items on a 4-point Likert scale from 0 �rarely or none of the time to 3 � most or all of the time. Anexample item is, “I felt that everything I did was an effort.”Cronbach’s � ranged from .89 to .92 across the world regions.

Eudaimonia

Psychological well-being. The Scales of Psychological Well-Being (Ryff, 1989) contains six subscales of positive psychologicalfunctioning: Self-Acceptance, Positive Relations With Others, Auton-omy, Environmental Mastery, Purpose in Life, and Personal Growth.A shortened version with three items per subscale was used (Ryff &Keyes, 1995). Participants responded to items on a 7-point Likertscale from 1 � strongly disagree to 7 � strongly agree. The sixsubscales were used individually based on evidence showing that asix-factor model (one factor for each subscale) represents the databetter than a one- or three-factor model of psychological well-being(Chen, Jing, Hayes, & Lee, 2013). Consistent with past studies, thethree-item subscale estimates of internal consistency were low due tothe theoretical, top-down item-selection process. To be sure to capturethe functional diversity within each aspect of well-being, the abbre-viated scale constructors chose to maximize conceptual breadth ratherthan statistical reliability.

The self-acceptance subscale measures positive evaluations ofone’s self and past life (� � .60–.71; e.g., “I like most parts of mypersonality”). The Positive Relations With Others subscale mea-sures the quality of interpersonal relationships (� � .57–.61; e.g.,“I have not experienced many warm and trusting relationships withothers”–reverse scored). The Autonomy subscale measures self-determination and independence, including the ability to evaluateoneself by personal standards (� � .51–.65; e.g., “I have confi-dence in my own opinions, even if they are different from the waymost other people think”). The Environmental Mastery subscalemeasures the capacity to effectively manage one’s life and sur-rounding world. (� � .43–.61; e.g., “I am good at managing theresponsibilities of daily life”). The Personal Growth subscale mea-sures a sense of continued growth and development as a person(� � .41–.66; e.g., “For me, life has been a continuous process oflearning, changing, and growth”). The Purpose in Life subscalemeasures belief that one’s life has meaning, purpose, and a clearsense of direction (� � .02–.344; e.g., “Some people wanderaimlessly thought life, but I am not one of them”). However, dueto the low internal consistency across world regions, this subscale(a) did not adequately measure purpose in life, (b) likely tappedinto other related constructs (e.g., mindfulness, personal growth),

3 Copies of all scales and items used in the International WellbeingStudy are available at the study website (http://www.wellbeingstudy.com).

4 The total sample correlation between two items (i.e., “Some peoplewander aimlessly through life, but I am not one of them” and “I sometimesfeel as if I’ve done all there is to do in life”) in the Purpose in Life subscalewas .01, which explains the low �s.

Thi

sdo

cum

ent

isco

pyri

ghte

dby

the

Am

eric

anPs

ycho

logi

cal

Ass

ocia

tion

oron

eof

itsal

lied

publ

ishe

rs.

Thi

sar

ticle

isin

tend

edso

lely

for

the

pers

onal

use

ofth

ein

divi

dual

user

and

isno

tto

bedi

ssem

inat

edbr

oadl

y.

473DIFFERENT TYPES OF WELL-BEING?

and (c) included large amounts of measurement error. Thus, thePurpose in Life subscale was replaced by a conceptually similarmeasure of meaning in life.5

Meaning in life. The Presence subscale of the Meaning inLife Questionnaire (Steger, Frazier, Oishi, & Kaler, 2006) is afive-item measure of the extent to which a person feels their life ismeaningful (e.g., “I understand my life’s meaning”). Participantsresponded to items on a 7-point Likert scale from 1 � absolutelyuntrue to 7 � absolutely true. Cronbach’s �s ranged from .85 to.92 across the world regions.

Correlates of Well-Being

Happiness orientations. The Orientations to Happiness Scale(Peterson, Park, & Seligman, 2005) is an 18-item measure con-sisting of three six-item subscales representing three differentpathways by which individuals seek happiness: A Pleasure Orien-tation, (e.g., “Life is too short to postpone the pleasures it canprovide”), developed from hedonic theories of happiness; an En-gagement Orientation (e.g., “I am always very absorbed in what Ido”), developed from psychological flow; and a Meaning Orien-tation (e.g., “I have a responsibility to make the world a betterplace”), developed from eudaimonic theories of happiness. Partic-ipants responded to items on a 5-point Likert scale from 1 � not atall like me to 5 � very much like me. Cronbach’s �s ranged from .75to .81 for the Pleasure Orientation, .54 to .73 for the EngagementOrientation, and .72 to .84 for the Meaning Orientation.

Hope. The (Dispositional) Adult Hope Scale (Snyder et al.,1991) is a 12-item measure of a motivational state oriented towardachieving goals. The measure consists of two 4-item subscales:Agency, or goal-directed determination (e.g., “I energetically pur-sue my goals”), and Pathways, or planning ways to meet goals(e.g., “I can think of many ways to get the things in life that areimportant to me”). The scale also contains four filler items that are

excluded from analyses. Participants responded to items on an8-point Likert scale from 1 � definitely false to 8 � definitely true.Cronbach’s � ranged from .85 to .91 across the world regions.

Gratitude. The Gratitude Survey–Six-Item Form (Mc-Cullough, Emmons, & Tsang, 2002) is a measure of one’s generaltendency to appreciate the positive in the world. Participantsresponded to items on a 7-point Likert scale from 1 � stronglydisagree to 7 � strongly agree. An example item is “I am gratefulto a wide variety of people.” Cronbach’s � ranged from .75 to .84.

Curiosity. The Curiosity and Exploration Inventory–II (Kash-dan et al., 2009) is a 10-item measure of the motivational systemassociated with the pursuit of novelty and challenge. The measureconsists of two 5-item subscales: Stretching, or seeking out newknowledge and experiences (e.g., “I actively seek as much informa-tion as I can in new situations”), and Embracing, or willingness toembrace novelty and uncertainty (e.g., “I am the type of person whoreally enjoys the uncertainty of everyday life”). Participants responded toitems on a 5-point Likert scale ranging from 1 � very slightly or not atall to 5 � extremely. Cronbach’s � ranged from .84 to .90.

Grit. The Grit Scale (Duckworth, Peterson, Matthews, &Kelly, 2007) is a 12-item measure of trait-like perseverance andpassion for long-term goals. The measure consists of two 6-itemsubscales: Perseverance of Effort, (e.g., “Setbacks don’t discour-age me”), and Consistency of Interests (e.g., “I often set a goal but

5 We conducted all analyses with the Purpose in Life subscale included(instead of the Presence of Meaning in Life subscale). We obtained verysimilar results for the one- and two-factor models (e.g., latent correlationsbetween Hedonia and Eudaimonia ranged from .81 to .93 across the worldregions with a total sample correlation of .91). However, as expected, themodel fit was slightly worse for the one-factor models, �2(35) � 2060.35,p � .001; CFI � .93; NNFI � .91; RMSEA � .087; SRMR � .042, andtwo-factor models, �2(31) � 1597.55, p � .001; CFI � .95; NNFI � .93;RMSEA � .078; SRMR � .037.

Table 1Relative Frequency by World Region and Country

RegionRelative

frequency (%) Count RegionRelative

frequency (%) Count

Oceania 25.7 1,964 Southern Europe 11.5 874New Zealand 21.1 1,605 Slovenia 3.9 296Australia 4.6 439 Greece 2.4 181

Anglo Nations & Northern Europe 25.2 1,909 Germany 1.7 128United States 11.4 870 Portugal 1.1 81England 5.2 393 Spain .5 37Finland 4.2 319 France .5 36Canada 1.7 128 Other 1.5 115Norway 1.3 97 Latin America 9.2 719Ireland .7 54 Mexico 4.1 315Other .6 48 Colombia 2.7 204

Former USSR & Eastern Europe 21.6 1,637 Brazil .8 59Hungary 14.1 1,076 Guatemala .5 35Czech Republic 3.4 258 Other 1.4 106Russia 2 155 East and South Asia 3.6 273Slovakia 1.3 96 China 2.8 210Other .7 51 India .6 45

Other .2 18Southeast Asia 1.8 138

Philippines 1.4 108Malaysia .2 15Other .2 15

Thi

sdo

cum

ent

isco

pyri

ghte

dby

the

Am

eric

anPs

ycho

logi

cal

Ass

ocia

tion

oron

eof

itsal

lied

publ

ishe

rs.

Thi

sar

ticle

isin

tend

edso

lely

for

the

pers

onal

use

ofth

ein

divi

dual

user

and

isno

tto

bedi

ssem

inat

edbr

oadl

y.

474 DISABATO, GOODMAN, KASHDAN, SHORT, AND JARDEN

later choose to pursue a different one”–reverse coded). Participantsresponded to items on a 5-point Likert scale from 1 � not at all like meto 5 � very much like me. Cronbach’s � ranged from .73 to .82.

Search for meaning.6 The Search subscale of the Meaning inLife Questionnaire (Steger et al., 2006) is a 5-item measure of theextent to which a person is actively seeking meaning in his or herlife (e.g., “I am seeking a purpose or mission for my life”).Participants responded to items on a 7-point Likert scale from 1 �absolutely untrue to 7 � absolutely true. Cronbach’s �s rangedfrom .83 to .90 across the world regions.

Rumination. The 6-item Rumination Scale, a new short formcreated from the 22-item Ruminative Response Style subscale ofthe Response Styles Questionnaire, assesses responses to depres-sive symptoms that focus on their meanings, causes, and conse-quences (Nolen-Hoeksema, 1991). Participants were promptedwith, “In the past 3 months would you say you . . .?” and respondedto six items using this stem. Two items came from the Broodingfactor, or moody and self-critical pondering (e.g., “Thought: ‘Whycan’t I handle things better?’”), and four from the Depression-Relatedfactor, thought to directly tap into depression symptoms themselves(e.g., “Thought: ‘Why can’t I get going?’” Treynor, Gonzalez, &Nolen-Hoeksema, 2003). Participants responded to items on a 7-pointLikert scale from 1 � strongly disagree to 7 � strongly agree.Cronbach’s � ranged from .83 to .88.

Loneliness. The Three-Item Loneliness Scale (Hughes,Waite, Hawkley, & Cacioppo, 2004) is a short form of the RevisedUCLA Loneliness Scale (Russell, Peplau, & Cutrona, 1980),which measures feelings of being isolated, disconnected, and lack-ing social connectedness. An example item is “How often do youfeel isolated from others?” Participants responded to items on a5-point Likert scale ranging from 1 � never to 5 � all of the time.Cronbach’s � ranged from .79 to .87.

Results

Data Analytic Overview

Whether validated, self-report measures of hedonic and eudai-monic well-being indicated one of two types of well-being wastested in an international sample. Confirmatory factor analysis(CFA) model-fit indices for the one-factor and two-factor modelswere compared to determine whether a one-factor or two-factormodel of well-being best represented the data.

Discriminant validity between Hedonia and Eudaimonia wastested in two ways. First, the magnitude of the latent correlationbetween Hedonia and Eudaimonia was examined. A very strongcorrelation would suggest negligible unique variance between the twofactors. Second, convergent validity coefficients were compared. Weestimated correlations between both types of well-being and relatedvariables. Smaller differences between the Hedonia and Eudaimoniafactor correlations would suggest less discriminant validity, indicatingthat the factors are assessing the same construct.

Because it is possible that one large international sample mightnot adequately represent all regions of the world, analyses wererepeated with subsamples based on the geographic world regions.Results are first presented with the total sample, followed bysubsample results organized by world regions.

Total Sample CFAs

Structural equation modeling using full maximum-likelihoodestimation was conducted with the software Analysis of MomentStructures, Version 22.0 (AMOS; Arbuckle, 2013) and the latentvariable analysis (lavaan) package in R (Rosseel, 2012). The totalsample-observed correlation matrix of well-being indicators usedis presented in Table 2. To evaluate model specification, the �2log-likelihood �2 value, comparative fit index (CFI), non-normed fit index (NNFI), root-mean-square error of approxi-mation (RMSEA), and standardized root-mean-square residual(SRMR) were used to test model fit. An excellent fitting modelhas a small, nonsignificant �2 value; however this is regarded asan overly stringent model-fit criterion for large sample sizes(Kline, 2011). Although conventional model-fit cutoffs exist forthe CFI, NNFI, RMSEA, and SRMR (e.g., Hu & Bentler, 1999;CRI/NNFI � .95; RMSEA/SRMR � .08), these standards areshown to vary across model types (Fan & Sivo, 2007; Kenny &McCoach, 2003) and may be too stringent (Marsh, Hau, & Wen,2004). Rather than focus solely on covariance discrepancies, weinterpreted our models based on building both accurate and parsimo-nious theories for the science of well-being (Mulaik et al., 1989).

Two CFAs were performed to determine whether Hedonia andEudaimonia were better modeled as one factor or two correlatedfactors. A one-factor model was first conducted in which all nineHedonia and Eudaimonia indicators were loaded onto the sameoverarching well-being factor. The second CFA was a correlatedtwo-factor model in which the three Hedonia indicators wereloaded onto a Hedonia factor and the six Eudaimonia indicatorswere loaded onto a Eudaimonia factor.

One-factor model. Our first CFA tested how well one overallwell-being factor represented the data. Well-being was modeled asa single factor indicated by the nine well-being (sub)scales. Figure1 depicts the standardized results. Model fit: �2(27) � 970.85, p �.001; CFI � .963; NNFI � .950; RMSEA � .068; SRMR � .030.

Two-factor model. The second CFA tested how well a cor-related two-factor model of well-being represented the data. TheHedonia factor was based on Diener’s (1984) model of subjectivewell-being and consisted of three scales: the Satisfaction with LifeScale, the Subjective Happiness Scale, and the CES-D scale (Rad-loff, 1977).7 The Eudaimonia factor consisted of five subscales ofthe Scales of Psychological Well-Being (Ryff, 1989): Self-Acceptance, Positive Relationships With Others, Autonomy, En-vironmental Mastery, and Personal Growth, as well as the Pres-ence subscale from the Meaning in Life Questionnaire (Steger et

6 The observed total sample correlation between the Presence and Searchsubscales of the Meaning in Life Questionnaire was –.16.

7 Factor analyses suggest that the CES-D (Radloff, 1977) consists offour factors rather than one global factor (Shafer, 2006). We tested thetwo-factor model with each of the CES-D factors as distinct leading to 6indicators of Hedonia. We obtained very similar results (e.g., latent cor-relations between Hedonia and Eudaimonia ranged from .78 to .94 acrosssix of the seven world regions) with a total sample correlation of .89, exceptfor East and South Asia, which had a latent correlation of .54). The divergentEast and South Asia latent correlation is likely due to the high loading of theCES-D somatization factor and low loading of the CES-D positive affectfactor, which is consistent with the cross-cultural depression literature (Klein-man, 2004).

Thi

sdo

cum

ent

isco

pyri

ghte

dby

the

Am

eric

anPs

ycho

logi

cal

Ass

ocia

tion

oron

eof

itsal

lied

publ

ishe

rs.

Thi

sar

ticle

isin

tend

edso

lely

for

the

pers

onal

use

ofth

ein

divi

dual

user

and

isno

tto

bedi

ssem

inat

edbr

oadl

y.

475DIFFERENT TYPES OF WELL-BEING?

al., 2006).8 Figure 1 depicts the model standardized results. Bothlatent-factors were standardized (i.e., variances were set equal to 1)and correlated. Model fit: �2(26) � 908.83, p � .001; CFI � .965;NNFI � .952; RMSEA � .067; SRMR � .029.

Model comparisons. Relative and absolute model-fit indiceswere assessed across the one-factor and two-factor models todetermine which best represented the data. The two-factor model’sCFI and NNFI indices were .003 and .002 larger than for theone-factor model, respectively. The two-factor model’s RMSEAand SRMR were both .001 smaller than those for the one-factormodel. The small discrepancies between model-fit indices sug-gested that the one-factor and two-factor models equally repre-sented the data. A �2 difference test comparing the one-factor andtwo-factor models was significant, indicating that the two-factormodel had superior model fit, ��2(1) � 62.02, p � .001. However,the �2 difference test has a greater chance of being significant withlarge samples, as in this study (Rigdon, 1999). Therefore it ispossible that the significant �2 difference test was simply a con-sequence of the large sample size and not a difference in modelmisspecification.

Factor correlation.9 The magnitude of the correlation be-tween Hedonia and Eudaimonia in the two-factor model was alsoused to determine whether each construct represented separatetypes of well-being. Figure 1 shows an estimated .96 correlationbetween the two factors. This indicates that Hedonia and Eudai-monia share 92% of their variance.

Two-Factor Model Correlations With RelatedMeasures of Well-Being

As an additional test for the validity of the two-factor model,known correlates of well-being were used to compare convergentvalidity coefficients between hedonia and eudaimonia. A con-struct’s validity includes identifying unique relationships betweena construct and related variables (Cronbach & Meehl, 1955). Tothe extent two factors have different correlations with the samevariable, their separate construct validity is supported. To theextent two factors have the same correlations with the samevariable, their construct validity is not supported. If the latter istrue then the two factors likely represent the same construct. Totest the relationship between the well-being factors and correlate

variables, correlates of well-being were added to the two-factormeasurement model. The factor loadings were fixed to the two-factor solution and correlations from the Hedonia and Eudaimoniafactors to the correlates were estimated. Each correlate of well-being was added to a separate structural equation model, so therewere 10 different models.

Beyond the modeled correlations, the size of the discrepancybetween the Hedonia and Eudaimonia correlations for the samecorrelate variable was examined (see Table 3, column |�r|). Theaverage correlation difference between Hedonia and Eudaimoniawas .07, indicating a high degree of similarity between the Hedo-nia and Eudaimonia constructs. Hypothesis tests (Steiger’s z test)were conducted to determine significant differences between theHedonia and Eudaimonia correlations for each correlate variable(Meng, Rosenthal, & Rubin, 1992). All correlation magnitudedifferences were significant at p � .001. Notably, due to the largesample size (n � 7617), the correlation-magnitude difference hadto be less than .01 to be nonsignificant. Overall, the Hedonia andEudaimonia correlations were in large part similar except for hope,meaning orientation, and grit.

Well-Being by World Region

Whether a one-factor or two-factor model was more valid foreach distinct world region was tested, as well as how these mea-surement models varied between world regions. Identical confir-matory analyses used for the total sample were conducted on each

8 Some researchers have suggested that Scales of Psychological Well-Being consists of three rather than six factors, whereas environmentalmastery, self-acceptance, meaning in life, and personal growth are groupedtogether into one factor (Abbott et al., 2006; Burns & Machin, 2009). Wetested the two-factor model with this Eudaimonia structure and obtainedvery similar results (e.g., latent correlations between Hedonia and Eudai-monia ranged from .85 to .95 across the world regions with a total samplecorrelation of .92).

9 To highlight the importance of controlling for measurement error, thetotal sample-observed correlation between Hedonia and Eudaimonia was.46. The correlation was calculated after standardizing each well-beingindicator and taking the mean of the standardized variables that encom-passed the Hedonia and Eudaimonia factors.

Table 2Zero-Order Correlations, Means, and Standard Deviations for Well-Being Measures

Construct Measure 1 2 3 4 5 6 7 8 9

Hedonia 1. SWLS2. SHS .583. CES-D �.50 �.58

Eudaimonia 4. SA .62 .63 �.545. PR .37 .49 �.43 .506. AU .22 .31 �.26 .32 .207. EM .46 .53 �.53 .55 .40 .318. ML .43 .51 �.41 .48 .39 .30 .389. PG .24 .36 �.30 .37 .34 .27 .33 .37Mean 22.21 4.84 14.69 16.18 16.33 15.67 15.43 25.1 18.60SD 7.52 1.27 11.12 3.73 3.75 3.46 3.42 6.85 2.70

Note. SWLS � Temporal Satisfaction With Life Scale; SHS � Subjective Happiness Scale; CES-D � Center for the Epidemiological Studies DepressionScale; SA � self-acceptance; PR � personal relations with others; AU � autonomy; EM � environmental mastery; ML � meaning in life; PG � personalgrowth. All correlations significant at p � .001.

Thi

sdo

cum

ent

isco

pyri

ghte

dby

the

Am

eric

anPs

ycho

logi

cal

Ass

ocia

tion

oron

eof

itsal

lied

publ

ishe

rs.

Thi

sar

ticle

isin

tend

edso

lely

for

the

pers

onal

use

ofth

ein

divi

dual

user

and

isno

tto

bedi

ssem

inat

edbr

oadl

y.

476 DISABATO, GOODMAN, KASHDAN, SHORT, AND JARDEN

world region subsample. Multigroup CFAs showed significantmeasurement variability across world regions’ factor loadings,��2(54) � 218.1; p � .001. World-region factor loadings were notfixed to the total sample values to allow for natural differences inHedonia and Eudaimonia, as well-being may manifest differentlyaround the world (Joshanloo, 2014). Model-fit indices across mod-els and convergent validity coefficients within the same worldregion were compared.

The two-factor measurement models better represented the datathan the one-factor models across each world region. Table 4outlines the model-fit indices of the one-factor and two-factormodels and their associated ��2 tests. In general, poorer model fitindices were obtained for East and South Asia and Southeast Asia.The estimated factor correlation between Hedonia and Eudaimoniawas very large in the two-factor model for every world region.10

Although correlations between the Hedonia and Eudaimonia fac-tors significantly differed across world region, ��2(6) � 16.1; p �.05, all exceeded .90 (see Table 4). Notably, the two world regionswith the largest correlations were Oceania and the Anglo nationsand Northern Europe; the regions with the smallest correlationswere Latin America and Southeast Asia.

To further test the validity of the two-factor models, Hedoniaand Eudaimonia convergent validity coefficients were comparedacross world regions. The sizes of the discrepancy between theHedonia and Eudaimonia correlations for each correlate variablewere tested (see Table 5). The average correlation-magnitudedifference between Hedonia and Eudaimonia did not exceed .08.However, the correlation differences were consistently larger forhope, meaning orientation, and grit. On average, the very similarcorrelations between hedonia and eudaimonia with a correlate of

well-being suggest there is near complete overlap between the twotypes of well-being.

Discussion

Using a large, comprehensive community-based sample fromaround the globe, we examined the uniqueness of two proposedfactors of well-being: Hedonia and Eudaimonia. Our results sug-gest that self-report measures of well-being based upon Diener’s(1984) subjective well-being model and Ryff’s (1989) psycholog-ical well-being model reflect one overarching well-being con-struct. More specifically, Hedonia and Eudaimonia showedlittle evidence of discriminant validity, but instead were highlycorrelated (r � .96) and similarly related to other well-beingvariables (|�r| � .07). The one-factor model of well-beinggenerated satisfactory model-fit indices, albeit slightly weakerthan those for the two-factor model. Similar evidence was foundfor a one-factor model of well-being throughout different worldregions.

Discriminant Validity: Factor Correlations

The large magnitude of the correlation between Hedonia andEudaimonia (r � .96) suggests that they primarily capture one

10 In a multigroup CFA in which each world region was a separate groupand all factor loadings were constrained to be equal, we obtained verysimilar results (e.g., latent correlations between Hedonia and Eudaimoniaranged from .93 to .99; model fit: �2(236) � 1336.4, CFI � .96, NNFI �.95, RMSEA � .025).

Figure 1. Standardized factor loadings and variance explained for the one and two factor models of well-being.SWLS � Satisfaction With Life Scale; SHS � Subjective Happiness Scale; HM � Happiness Measures;CESD � Center for the Epidemiological Studies Depression Scale; SA� self-acceptance; PR � personalrelations with others; AU � autonomy; EM � environmental mastery; ML � meaning in life; PG � personalgrowth; e � error term.

Thi

sdo

cum

ent

isco

pyri

ghte

dby

the

Am

eric

anPs

ycho

logi

cal

Ass

ocia

tion

oron

eof

itsal

lied

publ

ishe

rs.

Thi

sar

ticle

isin

tend

edso

lely

for

the

pers

onal

use

ofth

ein

divi

dual

user

and

isno

tto

bedi

ssem

inat

edbr

oadl

y.

477DIFFERENT TYPES OF WELL-BEING?

overarching well-being construct. There is some debate about howlarge a correlation magnitude must be to suggest poor discriminantvalidity between two factors or measures. For example, someresearchers have suggested that a latent correlation magnitudegreater than .85 indicates poor discriminant validity (e.g., Kline,2011). The r � .96 Hedonia–Eudaimonia correlation found in thisstudy far exceeds this cutoff and leaves little unique variancebetween the two factors.

Hedonia, Eudaimonia, and RelatedWell-Being Variables

Even if the correlation between two constructs is high, uniquerelationships to other variables support their discriminant validity.The validity of a psychological construct often involves showingdistinct placement in a nomological network of related constructs(Cronbach & Meehl, 1955). Hedonia and Eudaimonia—as mea-sured in this study—showed little evidence to support distinctplacement. Average correlation differences were small (|�r| �.07), and only three of the 11 variables’ correlation differencesexceeded .06.1 Notably, other studies that have argued for a lack ofdiscriminant validity between well-being constructs found corre-lation differences much larger than our own (e.g., Judge, Erez,Bono, & Thoresen, 2002; Smith, Pope, Rhodewalt, & Poulton,1989). Thus, results from nomological net analyses substantiatethe lack of discrimination between Hedonia (operationalized assubjective well-being) and Eudaimonia (operationalized as psy-chological well-being).

Three exceptions to this trend were hope, meaning orientation,and grit: Each differentially related to Hedonia and Eudaimonia(see Table 5). One explanation for these findings is that hope (asmeasured in this study) and grit involve goal-directed behavior.Eudaimonia might more strongly relate to goal-directed behaviorthan hedonia because theoretical conceptualizations of eudaimoniainvolve striving to reach one’s full potential (i.e., self-actualiza-tion; Maslow, 1943) and intrinsic motivation (Ryan & Deci, 2000).These goals may be in the pursuit of meaning, supported by thegreater correlation difference between a meaning orientation tohappiness and a pleasure and engagement orientation. Alterna-tively, the correlation difference for meaning orientation may bedue to use of the same word “meaning” in the meaning orientationand meaning in life items (Arnulf, Larsen, Martinsen, & Bong,2014).

Latent Factors Versus Scale Scores

Prior researchers have obtained support for the distinction be-tween Hedonia and Eudaimonia through the use of scale scoresrather than latent factors. Although Keyes et al. (2002) conducteda CFA with latent factors, the authors used the results of logisticregression and discriminant function analyses with scale scores tosupport the discriminant validity of Hedonia and Eudaimonia. Themain advantage of latent factors over scale scores is the reductionof measurement error12 (Cohen, Cohen, Teresi, Marchi, & Velez,1990). Analyzing data with latent factors ensures that measurescapture only the construct of interest rather than additional con-founding constructs. In the context of this study, measurementerror includes the unique variance of each of the nine well-beingindicators. Although not random error, this unique variance is

distinct from both Hedonia and Eudaimonia. When there is asignificant correlation between the Eudaimonia scale score andanother variable, it is difficult to determine which aspect of Eu-daimonia is driving this relationship. For example, curiosity mightrelate to Eudaimonia’s scale score due to the unique variance ofpersonal growth, and loneliness might relate because of the uniquevariance of positive relations with others. Thus, the analysis ofscale scores may be inappropriate for testing a measurement modelof well-being and lead to misleading statistical inference.

Hedonia and Eudaimonia May Be the SameWell-Being Construct

The large magnitude correlation and small average correlationdifferences between Hedonia and Eudaimonia may suggest thatthey reflect the same overarching well-being construct. Somereaders may find these results surprising given the rich philosoph-ical history behind the distinction between Hedonia and Eudaimo-nia (Telfer, 1990). Five explanations are proposed for this coun-terintuitive conclusion.

“Happiness” is a term that has been personalized such thatdefinitions can deviate drastically from one person to another.Some researchers portray hedonia as happiness (Ryan & Deci,2001; Waterman, 2008). Philosophical hedonism is often definedas the maximization of pleasure coupled with the minimization ofpain (Tatarkiewicz, 1976). However, in a Wittgensteinian sense, itis possible that the term “happiness” has changed in our modernculture to mean more than just maximizing pleasure and minimiz-ing pain. It may now encompass traditional eudaimonic qualities ofmeaning and fulfillment.

Hedonia is not operationalized in psychology as pleasure. Al-though debate exists in the philosophical literature, hedonism istraditionally associated with pleasure as opposed to happiness(e.g., Haybron, 2001; Morris, 2011). Unfortunately, none of theindicators of Hedonia directly taps into the experience of pleasure.Even positive affect, as measured by the Positive and NegativeAffect Schedule (PANAS; Watson, Clark, & Tellegen, 1988), maybe a poor operationalization of maximizing pleasure and minimiz-ing pain. The PANAS contains items reflecting philosophicallyeudaimonic emotions such as determination, interest, and inspira-tion (Watson et al., 1988). The emotions of excitement and enthu-siasm from the PANAS may capture effort toward meaningful goalpursuits rather than pleasure (Deci & Ryan, 2000).

There is no clear distinction between hedonic and eudaimonicexperiences. Some philosophers and psychologists prefer to inter-pret Hedonia as a high ratio of positive to negative subjectiveexperiences (e.g., Huta & Waterman, 2014). However, it becomesdifficult to categorize positive subjective experiences as hedonic oreudaimonic with this broadened interpretation of Hedonia. Forexample, is maximizing one’s meaning in life while attempting tominimize feelings of meaninglessness hedonic or eudaimonic?Meaning is traditionally viewed as a eudaimonic experience, butthe example also fits the broadened hedonic definition of maxi-mizing positive experiences and minimizing negative experiences.

11 Leaving out hope, meaning orientation, and grit, the average correla-tion difference between Hedonia and Eudaimonia was .04.

12 We acknowledge the possibility of covarying measurement error (e.g.social desirability in each item) present in a latent factor.

Thi

sdo

cum

ent

isco

pyri

ghte

dby

the

Am

eric

anPs

ycho

logi

cal

Ass

ocia

tion

oron

eof

itsal

lied

publ

ishe

rs.

Thi

sar

ticle

isin

tend

edso

lely

for

the

pers

onal

use

ofth

ein

divi

dual

user

and

isno

tto

bedi

ssem

inat

edbr

oadl

y.

478 DISABATO, GOODMAN, KASHDAN, SHORT, AND JARDEN

Philosophical theory, rather than empirical investigation, pro-motes the distinction between Hedonia and Eudaimonia. Someresearchers have argued that a purely empirically driven approachto the study of well-being is harmful and that scientists mustrespect theoretical complexity (Ryan & Huta, 2009; Waterman,2008). Empiricism without theory can be harmful to psychologicalscience, but theory without empiricism can be equally harmful.Scientific theories should be evaluated with special attention toempirical evidence. The Hedonia and Eudaimonia distinction hasstrong philosophical roots, but it lacks sufficient empirical supportto be labeled a scientific theory (see Kashdan et al., 2008).

The philosophical underpinnings of the distinction between He-donia and Eudaimonia suggest that it represents an ethical ratherthan a scientific debate. Scientists may argue that specific well-being constructs (e.g., meaning in life, gratitude) are more impor-tant outcomes than global well-being constructs. In light of thisconsideration, the focus of this study was not to test whether one

specific well-being outcome was ethically superior to another, butinstead to investigate the measurement structure of well-being.

Limitations

The explanations in the previous section are useful for this studyand further consideration by other researchers and practitionersinterested in the structure of well-being. Nonetheless, there arelimitations to this study that offer additional insight into why thereis negligible discriminant validity between Hedonia and Eudaimo-nia (as measured in this study). First, it is possible that theoperational measurements of Hedonia and Eudaimonia used in thisstudy are inadequate and do not capture the underlying constructs.The Hedonia measurement model was derived from Diener’s(1984) model of subjective well-being, but did not include tradi-tional measures of positive and negative affect (i.e., PANAS;Watson et al., 1988). As for Eudaimonia, there is still muchdisagreement among researchers about what constitutes it, andhow to measure it. For example, Waterman (1993, 2008) definedEudaimonia as authenticity to one’s “true self” rather than Ryffand Singer’s (2008) multifaceted definition. The lack of discrim-inant validity between Hedonia and Eudaimonia might be attrib-utable to the operational definition of Eudaimonia used in thisstudy (i.e., the Scales of Psychological Well-Being; Burns &Machin, 2009). A different measure, such as the longer 54- and84-item versions of the Scales of Psychological Well-Being ofEudaimonia, may have led to greater discriminant validity. Sec-ond, the underlying constructs of Hedonia and Eudaimonia mayexist, but participants might have been unable to distinguish be-tween them when subjectively reporting their experiences. Differ-entiation of these constructs might require a high level of aware-ness and insight that participants might have lacked.

Limitations about the sample include the fact that most partic-ipants were likely middle to upper class, given the study require-ment for Internet access. In addition, less than 20% were men.

Table 3Latent Correlations of Hedonia and Eudaimonia WithCorrelates of Well-Being

Variable Hedonia Eudaimonia |�r| Steiger’s z

Gratitude .63 .64 .01 �4.64���

Pleasure orientation .24 .21 .03 10.98���

Rumination �.56 �.52 .04 �17.08���

Meaning in life (search) �.30 �.25 .05 �18.58���

Engagement orientation .33 .39 .06 �23.06���

Loneliness �.69 �.63 .06 �29.33���

Curiosity .36 .42 .06 �23.39���

Hope .67 .77 .10 �57.34���

Meaning orientation .33 .48 .15 �61.07���

Grit .39 .54 .15 �64.18���

Note. |�r| � difference in correlation magnitude.� p � .05. �� p � .01. ��� p � .001.

Table 4Model Fit by World Region

Model Statistic OceaniaAnglo

& N. EuropeUSSR

& E. Europe S. Europe Latin Amer. E. & S. Asia Southeast Asia

n 1954 1909 1637 874 719 273 138�2(27)a 280.23 375.42 228.96 88.28 112.78 77.30 50.29CFI .965 .949 .960 .978 .962 .937 .941

One-factor model NNFI .954 .932 .947 .970 .949 .916 .921RMSEA .069 .082 .068 .051 .067 .083 .079SRMR .032 .039 .034 .026 .030 .048 .050�2(26) 277.00 368.09 209.18 80.74 92.16 75.27 45.76CFI .966 .950 .964 .980 .970 .938 .950

Two-factor model NNFI .953 .930 .950 .972 .959 .914 .930RMSEA .070 .083 .066 .049 .060 .083 .074SRMR .031 .038 .032 .024 .028 .047 .048��2(1) 3.23 7.33�� 19.78��� 7.54�� 20.62��� 2.03 4.53�

Hedonia & eudaimoniacorrelation

.99 .98 .96 .96 .92 .95 .91

Note. Anglo � Anglo nations; N. � northern; USSR � former nations of the Union of Soviet Socialist Republics; E. � East; S. � South; Amer. �America; CFI � comparative fit index; NNFI � nonnormed fit index; RMSEA � root-mean-square error of approximation; SRMR � standardizedroot-mean-square residual; ��2(1) � change in chi-square value.a All �2 values are significant at p � .001.� p � .05. �� p � .01. ��� p � .001.

Thi

sdo

cum

ent

isco

pyri

ghte

dby

the

Am

eric

anPs

ycho

logi

cal

Ass

ocia

tion

oron

eof

itsal

lied

publ

ishe

rs.

Thi

sar

ticle

isin

tend

edso

lely

for

the

pers

onal

use

ofth

ein

divi

dual

user

and

isno

tto

bedi

ssem

inat

edbr

oadl

y.

479DIFFERENT TYPES OF WELL-BEING?

Most participants were also from traditionally Western cultures,and results may not generalize to people from countries of non-Western culture. These sample demographics potentially restrictour findings to middle–upper class women from Western coun-tries.

Last, and perhaps most important to note, this study tested thefactor structure of well-being from a single measurement perspec-tive. Determining whether different types of well-being exist is acomplicated and multifaceted research question that arguably re-quires multimethod assessment extending beyond self-report mea-sures. Rather than concluding that only one type of well-beingexists, results suggest that Hedonia and Eudaimonia as operation-alized and measured in this study are the same type of well-being.

Conclusion and Future Directions

Researchers interested in the distinction between Hedonia andEudaimonia might consider focusing on pathways to reaching astate of well-being. For example, is a person’s conscious orienta-tion in life toward maximizing the pleasure/pain ratio, or is itvirtuous excellence? These types of research questions wouldprevent researchers from conflating the predictors of well-beingwith well-being itself (Kashdan et al., 2008). Preliminary researchin this area has suggested that eudaimonic motivations relate morestrongly to well-being than do hedonic motivations (e.g., Hender-son, Knight, & Richardson, 2014; Huta & Ryan, 2010). The fieldof behavioral genetics also appears promising toward understand-ing the ultimate causes of well-being (Lyubomirsky, Sheldon, &Schkade, 2005). For example, Fredrickson et al. (2013) found thatthe unique variance among Hedonia and Eudaimonia may begenetically determined (although see Brown, MacDonald, Sa-manta, Friedman, & Coyne, 2014 for a critique of the methodologyused).

Although Hedonia and Eudaimonia have become popular withinthe scholarship of well-being, empirical support for their distinc-tion is limited. Conceptualizing Hedonia and Eudaimonia as onehigher order factor with multiple lower order constructs may bemore appropriate than treating them as distinct constructs. Re-searchers are encouraged to measure and study the lower order,specific variables (e.g., character strengths, meaning in life, au-

thenticity), rather than Hedonia and Eudaimonia (Biswas-Diener,Kashdan, & King, 2009). Based on the questions being asked,researchers and applied professionals can determine what level ofmeasurement is most appropriate.

References

Abbott, R. A., Ploubidis, G. B., Huppert, F. A., Kuh, D., Wadsworth,M. E., & Croudace, T. J. (2006). Psychometric evaluation and predictivevalidity of Ryff’s psychological well-being items in a UK birth cohortsample of women. Health and Quality of Life Outcomes, 4, 76. http://dx.doi.org/10.1186/1477-7525-4-76

Arbuckle, J. L. (2013). Amos (Version 22.0) [Computer Program]. Chi-cago, IL: IBM-SPSS.

Aristotle. (1985). Nicomachean ethos (T. Irwin, Trans.). Indianapolis, IN:Hackett. Original work published 4th century BCE.

Arnulf, J. K., Larsen, K. R., Martinsen, O. L., & Bong, C. H. (2014).Predicting survey responses: How and why semantics shape surveystatistics on organizational behaviour. PLoS ONE, 9, e106361. http://dx.doi.org/10.1371/journal.pone.0106361

Biswas-Diener, R., Kashdan, T. B., & King, L. A. (2009). Two traditionsof happiness research, not two distinct types of happiness. The Journalof Positive Psychology, 4, 208 –211. http://dx.doi.org/10.1080/17439760902844400

Brown, N. J., MacDonald, D. A., Samanta, M. P., Friedman, H. L., &Coyne, J. C. (2014). A critical reanalysis of the relationship betweengenomics and well-being. PNAS: Proceedings of the National Academyof Sciences of the United States of America, 111, 12705–12709. http://dx.doi.org/10.1073/pnas.1407057111

Burns, R. A., & Machin, M. A. (2009). Investigating the structural validityof Ryff’s Psychological Well-Being Scales across two samples. SocialIndicators Research, 93, 359–375. http://dx.doi.org/10.1007/s11205-008-9329-1

Busseri, M. A., & Sadava, S. W. (2011). A review of the tripartite structureof subjective well-being: Implications for conceptualization, operation-alization, analysis, and synthesis. Personality and Social PsychologyReview, 15, 290–314. http://dx.doi.org/10.1177/1088868310391271

Chen, F. F., Jing, Y., Hayes, A., & Lee, J. M. (2013). Two concepts of twoapproaches? A bifactor analysis of psychological and subjective well-being. Journal of Happiness Studies, 14, 1033–1068. http://dx.doi.org/10.1007/s10902-012-9367-x

Cohen, P., Cohen, J., Teresi, J., Marchi, M., & Velez, C. N. (1990).Problems in the measurement of latent variables in structural equations

Table 5Validity Correlation Magnitude Differences Between Hedonia and Eudaimonia by World Region

Variable Oceania Anglo & N. Europe USSR & E. Europe S. Europe Latin Amer. E. & S. Asia Southeast Asia

n 1954 1909 1637 974 719 273 138Gratitude .00 .06��� .01 .04��� .06��� .00 .02Pleasure orientation .03��� .07��� .07��� .01 .00 .06�� .02Rumination .04��� .05��� .02�� .04��� .04�� .04� .01Meaning in life (search) .05��� .05��� .00 .08��� .06��� .13��� .08�

Engagement orientation .05��� .02��� .05��� .06��� .03� .10��� .00Loneliness .06��� .09��� .03��� .06��� .06��� .17��� .11���

Curiosity .07��� .04��� .04��� .08��� .04�� .03 .11���

Hope .09��� .08��� .11��� .14��� .10��� .02 .12���

Meaning orientation .11��� .12��� .19��� .15��� .14��� .10��� .19���

Grit .14��� .15��� .12��� .13��� .14��� .15��� .10���

Average |�r| .06 .07 .06 .08 .07 .08 .08

Note. Anglo � Anglo nations; N. � northern; USSR � former nations of the Union of Soviet Socialist Republics; E. � East; S. � South; Amer. �America; |�r| � difference in correlation magnitude.� p � .05. �� p � .01. ��� p � .001.

Thi

sdo

cum

ent

isco

pyri

ghte

dby

the

Am

eric

anPs

ycho

logi

cal

Ass

ocia

tion

oron

eof

itsal

lied

publ

ishe

rs.

Thi

sar

ticle

isin

tend

edso

lely

for

the

pers

onal

use

ofth

ein

divi

dual

user

and

isno

tto

bedi

ssem

inat

edbr

oadl

y.

480 DISABATO, GOODMAN, KASHDAN, SHORT, AND JARDEN

causal models. Applied Psychological Measurement, 14, 183–196.http://dx.doi.org/10.1177/014662169001400207

Compton, W. C., Smith, M. L., Cornish, K. A., & Qualls, D. L. (1996).Factor structure of mental health measures. Journal of Personality andSocial Psychology, 71, 406–413. http://dx.doi.org/10.1037/0022-3514.71.2.406

Coyne, J. C. (2013). Highly correlated hedonic and eudaimonic well-beingthwart genomic analysis. PNAS: Proceedings of the National Academyof Sciences of the United States of America, 110, E4183. http://dx.doi.org/10.1073/pnas.1315212110

Cronbach, L. J., & Meehl, P. E. (1955). Construct validity in psychologicaltests. Psychological Bulletin, 52, 281–302. http://dx.doi.org/10.1037/h0040957

Deci, E. L., & Ryan, R. M. (2000). The “what” and “why” of goal pursuits:Human needs and the self-determination of behavior. PsychologicalInquiry, 11, 227–268. http://dx.doi.org/10.1207/S15327965PLI1104_01

Diener, E. (1984). Subjective well-being. Psychological Bulletin, 95, 542–575. http://dx.doi.org/10.1037/0033-2909.95.3.542

Diener, E., Emmons, R. A., Larsen, R. J., & Griffin, S. (1985). TheSatisfaction With Life Scale. Journal of Personality Assessment, 49,71–75. http://dx.doi.org/10.1207/s15327752jpa4901_13

Duckworth, A. L., Peterson, C., Matthews, M. D., & Kelly, D. R. (2007).Grit: Perseverance and passion for long-term goals. Journal of Person-ality and Social Psychology, 92, 1087–1101. http://dx.doi.org/10.1037/0022-3514.92.6.1087

Eidelson, R., Pilisuk, M., & Soldz, S. (2011). The dark side of compre-hensive soldier fitness. American Psychologist, 66, 643–644. http://dx.doi.org/10.1037/a0025272

Ensel, W. M. (1986). Measuring depression: The CES-D scale. In N. Lin,A. Dean, & W. Ensel (Eds.), Social support, life events, and depression(pp. 51–70). New York, NY: Academic Press. http://dx.doi.org/10.1016/B978-0-12-450660-2.50011-2

Fan, X., & Sivo, S. (2007). Sensitivity of fit indices to model misspecifi-cation and model types. Multivariate Behavioral Research, 42, 509–529. http://dx.doi.org/10.1080/00273170701382864

Fiske, D. W. (1982). Convergent–discriminant validation in measurementsand research strategies. In D. Brinberg & L. H. Kidder (Eds.), Forms ofvalidity in research: New directions for methodology in social andbehavioral science (pp. 77–92). San Francisco, CA: Jossey-Bass.

Fredrickson, B. L., Grewen, K. M., Coffey, K. A., Algoe, S. B., Firestine,A. M., Arevalo, J. M. G., . . . Cole, S. W. (2013). A functional genomicperspective on human well-being. PNAS: Proceedings of the NationalAcademy of Sciences of Sciences of the United States of America, 110,13684–13689. http://dx.doi.org/10.1073/pnas.1305419110

Gallagher, M. W., Lopez, S. J., & Preacher, K. J. (2009). The hierarchicalstructure of well-being. Journal of Personality, 77, 1025–1050. http://dx.doi.org/10.1111/j.1467-6494.2009.00573.x

Haybron, D. M. (2001). Happiness and pleasure. Philosophy and Phenom-enological Research, 62, 501–528. http://dx.doi.org/10.1111/j.1933-1592.2001.tb00072.x

Henderson, L. W., Knight, T., & Richardson, B. (2014). The hedonic andeudaimonic validity of the orientations to happiness scale. Social Indi-cators Research, 115, 1087–1099. http://dx.doi.org/10.1007/s11205-013-0264-4

Hu, L., & Bentler, P. M. (1999). Cutoff criteria for fit indexes in covariancestructure analysis: Conventional criteria versus new alternatives. Struc-tural Equation Modeling, 3, 424–453.

Hughes, M. E., Waite, L. J., Hawkley, L. C., & Cacioppo, J. T. (2004). Ashort scale for measuring loneliness in large surveys: Results from twopopulation-based studies. Research on Aging, 26, 655–672. http://dx.doi.org/10.1177/0164027504268574

Huta, V., & Ryan, R. M. (2010). Pursuing pleasure or virtue: The differ-ential and overlapping well-being benefits of hedonic and eudaimonic

motives. Journal of Happiness Studies, 11, 735–762. http://dx.doi.org/10.1007/s10902-009-9171-4

Huta, V., & Waterman, A. S. (2014). Eudaimonia and its distinction fromHedonia: Developing a classification and terminology for understandingconceptual and operational definitions. Journal of Happiness Studies,14, 1425–1456.

Jahoda, M. (1958). Current concepts of positive mental health. New York,NY: Basic Books. http://dx.doi.org/10.1037/11258-000

Joshanloo, M. (2014). Eastern conceptualizations of happiness: Fundamen-tal differences with western views. Journal of Happiness Studies, 15,475–493. http://dx.doi.org/10.1007/s10902-013-9431-1

Judge, T. A., Erez, A., Bono, J. E., & Thoresen, C. J. (2002). Are measuresof self-esteem, neuroticism, locus of control, and generalized self-efficacy indicators of a common core construct? Journal of Personalityand Social Psychology, 83, 693–710. http://dx.doi.org/10.1037/0022-3514.83.3.693

Kahneman, D., Diener, E., & Schwarz, N. (Eds.). (1999). Well-being:Foundations of hedonic psychology. New York, NY: Russell SageFoundation.

Kashdan, T. B., Biswas-Diener, R., & King, L. A. (2008). Reconsideringhappiness: The costs of distinguishing between hedonics and Eudaimo-nia. The Journal of Positive Psychology, 3, 219–233. http://dx.doi.org/10.1080/17439760802303044

Kashdan, T. B., Gallagher, M. W., Silvia, P. J., Winterstein, B. P., Breen,W. E., Terhar, D., & Steger, M. F. (2009). The Curiosity and ExplorationInventory–II: Development, factor structure, and psychometrics. Journalof Research in Personality, 43, 987–998. http://dx.doi.org/10.1016/j.jrp.2009.04.011

Kenny, D. A., & McCoach, D. B. (2003). Effect of the number ofvariables on measures of fit in structural equation modeling. Struc-tural Equation Modeling, 10, 333–351. http://dx.doi.org/10.1207/S15328007SEM1003_1

Keyes, C. L. (2005). Mental illness and/or mental health? Investigatingaxioms of the complete state model of health. Journal of Consulting andClinical Psychology, 73, 539–548. http://dx.doi.org/10.1037/0022-006X.73.3.539

Keyes, C. L. M., Shmotkin, D., & Ryff, C. D. (2002). Optimizing well-being: The empirical encounter of two traditions. Journal of Personalityand Social Psychology, 82, 1007–1022. http://dx.doi.org/10.1037/0022-3514.82.6.1007

Kleinman, A. (2004). Culture and depression. The New England Journal ofMedicine, 351, 951–953. http://dx.doi.org/10.1056/NEJMp048078

Kline, R. B. (2011). Principles and practice of structural equation mod-eling (3rd ed.). New York, NY: Guilford Press.

Linley, P. A., Maltby, J., Wood, A. M., Osborne, G., & Hurling, R. (2009).Measuring happiness: The higher order factor structure of subjective andpsychological well-being measures. Personality and Individual Differ-ences, 47, 878–884. http://dx.doi.org/10.1016/j.paid.2009.07.010

Lyubomirsky, S., & Lepper, H. S. (1999). A measure of subjective hap-piness: Preliminary reliability and construct validation. Social IndicatorsResearch, 46, 137–155. http://dx.doi.org/10.1023/A:1006824100041

Lyubomirsky, S., Sheldon, K. M., & Schkade, D. (2005). Pursuing happi-ness: The architecture of sustainable change. Review of General Psy-chology, 9, 111–131. http://dx.doi.org/10.1037/1089-2680.9.2.111

Marsh, H. W., Hau, K. T., & Wen, Z. (2004). In search of golden rules:Comment on hypothesis-testing approaches to setting cutoff values forfit indexes and dangers in overgeneralizing Hu and Bentler’s (1999).findings. Structural Equation Modeling, 11, 320–341. http://dx.doi.org/10.1207/s15328007sem1103_2

Maslow, A. H. (1943). A theory of human motivation. PsychologicalReview, 50, 370–396. http://dx.doi.org/10.1037/h0054346

McCullough, M. E., Emmons, R. A., & Tsang, J. A. (2002). The gratefuldisposition: A conceptual and empirical topography. Journal of Person-

Thi

sdo

cum

ent

isco

pyri

ghte

dby

the

Am

eric

anPs

ycho

logi

cal

Ass

ocia

tion

oron

eof

itsal

lied

publ

ishe

rs.

Thi

sar

ticle

isin

tend

edso

lely

for

the

pers

onal

use

ofth

ein

divi

dual

user

and

isno

tto

bedi

ssem

inat

edbr

oadl

y.

481DIFFERENT TYPES OF WELL-BEING?

ality and Social Psychology, 82, 112–127. http://dx.doi.org/10.1037/0022-3514.82.1.112

Meng, X., Rosenthal, R., & Rubin, D. B. (1992). Comparing correlatedcorrelation coefficients. Psychological Bulletin, 111, 172–175. http://dx.doi.org/10.1037/0033-2909.111.1.172

Morris, S. (2011). In defense of the hedonistic account of happiness.Philosophical Psychology, 24, 261–281. http://dx.doi.org/10.1080/09515089.2011.556604

Mulaik, S. A., James, L. R., Alstine, J. V., Bennett, N., Lind, S., & Stilwell,C. D. (1989). Evaluation of goodness-of-fit indices for structural equa-tion models. Psychological Bulletin, 105, 430–445. http://dx.doi.org/10.1037/0033-2909.105.3.430

Nolen-Hoeksema, S. (1991). Responses to depression and their effects onthe duration of depressive episodes. Journal of Abnormal Psychology,100, 569–582. http://dx.doi.org/10.1037/0021-843X.100.4.569

Peterson, C., Park, N., & Seligman, M. E. (2005). Orientations to happinessand life satisfaction: The full life versus the empty life. Journal ofHappiness Studies, 6, 25– 41. http://dx.doi.org/10.1007/s10902-004-1278-z

Radloff, L. S. (1977). The CES-D scale: A self-report depression scale forresearch in the general population. Applied Psychological Measurement,1, 385–401. http://dx.doi.org/10.1177/014662167700100306

Rigdon, E. E. (1999). Using the Friedman method of ranks for modelcomparison in structural equation modeling. Structural Equation Mod-eling, 6, 219–232. http://dx.doi.org/10.1080/10705519909540131

Rosseel, Y. (2012). lavaan: An R package for structural equation modeling.Journal of Statistical Software, 48, 1–36.

Russell, D., Peplau, L. A., & Cutrona, C. E. (1980). The revised UCLALoneliness Scale: Concurrent and discriminant validity evidence. Jour-nal of Personality and Social Psychology, 39, 472–480. http://dx.doi.org/10.1037/0022-3514.39.3.472

Ryan, R. M., & Deci, E. L. (2000). Self-determination theory and thefacilitation of intrinsic motivation, social development, and well-being.American Psychologist, 55, 68 –78. http://dx.doi.org/10.1037/0003-066X.55.1.68

Ryan, R. M., & Deci, E. L. (2001). On happiness and human potentials: Areview of research on hedonic and eudaimonic well-being. AnnualReview of Psychology, 52, 141–166. http://dx.doi.org/10.1146/annurev.psych.52.1.141

Ryan, R. M., & Huta, V. (2009). Wellness as healthy functioning orwellness as happiness: The importance of eudaimonic thinking (responseto the Kashdan et al. and Waterman discussion). The Journal of PositivePsychology, 4, 202–204. http://dx.doi.org/10.1080/17439760902844285

Ryan, R. M., Huta, V., & Deci, E. L. (2008). Living well: A self-determination theory perspective on Eudaimonia. Journal of HappinessStudies, 9, 139–170. http://dx.doi.org/10.1007/s10902-006-9023-4

Ryff, C. D. (1989). Happiness is everything, or is it? Explorations on themeaning of psychological well-being. Journal of Personality and SocialPsychology, 57, 1069–1081. http://dx.doi.org/10.1037/0022-3514.57.6.1069

Ryff, C. D. (2014). Psychological well-being revisited: Advances in thescience and practice of Eudaimonia. Psychotherapy and Psychosomat-ics, 83, 10–28. http://dx.doi.org/10.1159/000353263

Ryff, C. D., & Keyes, C. L. M. (1995). The structure of psychologicalwell-being revisited. Journal of Personality and Social Psychology, 69,719–727. http://dx.doi.org/10.1037/0022-3514.69.4.719

Ryff, C. D., & Singer, B. (1998). The contour of positive human health.Psychological Inquiry, 9, 1–28. http://dx.doi.org/10.1207/s15327965pli0901_1

Ryff, C. D., & Singer, B. H. (2008). Know thyself and become what youare: A eudaimonic approach to psychological well-being. Journal ofHappiness Studies, 9, 13–39. http://dx.doi.org/10.1007/s10902-006-9019-0

Shafer, A. B. (2006). Meta-analysis of the factor structures of four depres-sion questionnaires: Beck, CES-D, Hamilton, and Zung. Journal ofClinical Psychology, 62, 123–146. http://dx.doi.org/10.1002/jclp.20213

Sheldon, K. M. (2013). Individual daimon, universal needs, and subjectivewell-being: Happiness as the natural consequence of a life well-lived. InA. Waterman (Ed.), The best within us: Positive psychology perspectiveson eudaimonic functioning (pp. 119–137). Washington, DC: AmericanPsychological Association. http://dx.doi.org/10.1037/14092-007

Smith, T. W., Pope, M. K., Rhodewalt, F., & Poulton, J. L. (1989).Optimism, neuroticism, coping, and symptom reports: An alternativeinterpretation of the Life Orientation Test. Journal of Personality andSocial Psychology, 56, 640–648. http://dx.doi.org/10.1037/0022-3514.56.4.640

Snyder, C. R., Harris, C., Anderson, J. R., Holleran, S. A., Irving, L. M.,Sigmon, S. T., . . . Harney, P. (1991). The will and the ways: Develop-ment and validation of an individual-differences measure of hope. Jour-nal of Personality and Social Psychology, 60, 570–585. http://dx.doi.org/10.1037/0022-3514.60.4.570

Steger, M. F., Frazier, P., Oishi, S., & Kaler, M. (2006). The meaning inlife questionnaire: Assessing the presence of and search for meaning inlife. Journal of Counseling Psychology, 53, 80–93. http://dx.doi.org/10.1037/0022-0167.53.1.80

Tatarkiewicz, W. (1976). Analysis of happiness. The Hague, the Nether-lands: Martinus Nijhoff.

Tay, L., & Diener, E. (2011). Needs and subjective well-being around theworld. Journal of Personality and Social Psychology, 101, 354–365.http://dx.doi.org/10.1037/a0023779

Telfer, E. (1990). Happiness. New York, NY: St. Martin’s Press.Treynor, W., Gonzalez, R., & Nolen-Hoeksema, S. (2003). Rumination

reconsidered: A psychometric analysis. Cognitive Therapy and Re-search, 27, 247–259. http://dx.doi.org/10.1023/A:1023910315561

Vittersø, J. (2003). Flow versus life satisfaction: A projective use ofcartoons to illustrate the difference between the evaluation approach andthe intrinsic motivation approach to subjective quality of life. Journal ofHappiness Studies, 4, 141–167. http://dx.doi.org/10.1023/A:1024413112234

Waterman, A. S. (1993). Two conceptions of happiness: Contrasts ofpersonal expressiveness (Eudaimonia) and hedonic enjoyment. Journalof Personality and Social Psychology, 64, 678–691. http://dx.doi.org/10.1037/0022-3514.64.4.678

Waterman, A. S. (2008). Reconsidering happiness: A eudaimonist’s per-spective. The Journal of Positive Psychology, 3, 234–252. http://dx.doi.org/10.1080/17439760802303002

Watson, D., Clark, L. A., & Tellegen, A. (1988). Development and vali-dation of brief measures of positive and negative affect: The PANASscales. Journal of Personality and Social Psychology, 54, 1063–1070.http://dx.doi.org/10.1037/0022-3514.54.6.1063

Received November 6, 2014Revision received April 30, 2015

Accepted June 23, 2015 �

Thi

sdo

cum

ent

isco

pyri

ghte

dby

the

Am

eric

anPs

ycho

logi

cal

Ass

ocia

tion

oron

eof

itsal

lied

publ

ishe

rs.

Thi

sar

ticle

isin

tend

edso

lely

for

the

pers

onal

use

ofth

ein

divi

dual

user

and

isno

tto

bedi

ssem

inat

edbr

oadl

y.

482 DISABATO, GOODMAN, KASHDAN, SHORT, AND JARDEN