Hendrick and Hendrick Love Attitudes

8/2/2019 Hendrick and Hendrick Love Attitudes

1/11

Journal of Per'~onality and Social Psychology Copyright 1986 by the American Psychological Association, Inc.1986, Vol. 50, No. 2, 392-402 0022-3514/8 6/$00.75

A Theory and Method o f LoveClyde Hendrick and Susan Hendrick

Texas Tech UniversityThis research was part of a larger research program on love and sex attitudes. Earlier work on lovewas reported in Hendrick, Hendrick, Foote, and Slapion-Foote (1984). The work on love extendsLee's (1973/1976) theory of six basic lov e styles: Eros (passionate love), Ludus (game-playing love),Storge (friendship love), Pragma (logical, "shopping list " love), Mania (possessive, dependen t love),and Agape (all-giving, selfless ove). Theory development has proceeded concurrently with the devel-opment of measurement scales. Study I (N = 807) used a 42-item rating questionnaire, with 7 itemsmeasuring each of the love styles. Six love style scales emerged clearly from factor analysis. Internalreliability was shown for each scale, and the scales had low intercorrelations with each other. Significantrelationships we re found between love attitudes a nd several background variables, including gender,ethnicity, previous love experiences, current love status, and self-esteem. Confirmatory Study II (N =567) replicated facto r structure, factor loadings, and reliability analyses of the first study. In addition,the significant relationships between love attitudes a nd gender, previous love experiences, current lovestatus, and self-esteemwere also consistent with the results of Study I. The love scale shows considerablepromise as an instrument for future research on love.

Dur ing the pas t decade, love has become respectable as anarea for study by psy chologists (e.g., Rubin, 1984). Several the-ories of love have been propo sed (e.g. , readings by Coo k & W ilson,1979; Kelley, 1983). Early theorie s that used glo bal conce pts oflove are being replaced by theor ies that use mult idimensionalconstructs that promise greater yields in knowledge.

T h e o r i e s o f L o v eBlau (1964) propose d an exchange theory o f love that char -acter ized the develo pment of love as requir ing a nicely balanced

degree of mutuali ty and the consis tent exchange of rewards be-tween par tners . M ore recently, Clark and Mills (1979) at tempte dto dif ferentia te "exchange" f rom "communal" (e .g. , a l t ruis t ic)relat ionships by showing that a " t i t for ta t" ap proach may beaccepted in an exchange relationship, but such an app roach m ayactually damage a communal relat ionship.

Berscheid and Wals ter (1974) proposed an appro ach to lovewhich descr ibed romantic , pass ionate love as physiologicala r ousa l accompanied by appr opr ia te cogni tive cues such tha t"pass ionate love" is the appro pr iate label for the arousal . Fromthis essentia l ly labeling approach , Wals ter and Wals ter (1978)proposed two general kinds of love: pass ionate love and com-panionate love, with the former near ly always evolving to thelat ter in an endur in g close relat ionship.Moving coun ter to the increas ing emphasis on love as mult i-dimensional , S ternberg and Gra jek ( I 984) proposed tha t thereis a "general factor" of love which is quite consis tent across ro-mantic , famil ia l , and f r iendship relat ionships . However , the au-thors note that a l though the var ious love exper iences may bes imilar, the "c onco mitan ts" o f the exper iences may be quite dif -

Correspondence concerning this article should be addressed to ClydeHendrick, Department of Psychology,Texas Tech University, Lubbock,Texas 79409.

ferent . In a recent scholar ly at tempt to develop a conceptu al andmethodological f ramework within which to ex amine love, Kelley(1983) introdu ced a mod el for what he cal led "pragm atic love"(p. 283) while still recognizing passionate love (e.g. , Berscheid& Walster, 1978) a nd altruist ic love (e.g., Clark & Mills, 1979).Pragmatic love emphasizes t rus t and to lerance and develops withgreater del iberat ion and self -control than do other types of love.As Kelley concluded, " love is typical ly a blend of the dif ferentforms descr ibed by the preceding models" (p. 186). Unfor tu -nately, a comprehensive mo del that re lates and integrates thesedif ferent forms of love has not been available .

S c a l e D e v e l o p m e n tThe pr ima ry thrus t of psychological work on love has beentoward theory building, with a l imited secondary thrus t toward

scale construct ion. Considerable work in the area has been do neby Rub in (1970, 1973, 1974), wh o was one of the first to stud ythe s imilar i t ies and dif ferences between loving and l iking. Rubinviewed the two as conceptually dis t inct though l inked phen om-ena, and he developed two scales to meas ure the two constructs .Although the scales have been widely used, and the dis t inct io nbetween the two concepts was clear, the cor relat ion between thetwo scales was higher than des irable for measures of two inde-penden t constructs . However , some p osi t ive evidence for the va-l idi ty of the dis t inct ion has been ob tained (Derm er & Pyszczyn-ski, 1978; Hill, Rubin, & Peplau, 1976). In contrast, in a con-ceptual analys is , Kelley (1983) identif ied four components ofRu bin's love scale (needing, caring, trust, tolerance) and suggestedthat the l iking scale might bet ter have been nam ed as a measureof respect. Research by Steck, Levitan, McLane, a nd Kelley( 1982) a t tempted to manipula te these pr esumed components o fthe love scale and found quite dif ferent responses to dif ferents t imulus prof iles , even though the total love score at tr ibuted tos t imulus s trangers was constant across s t im ulus prof i les. For ex-ample, a "car ing " prof i le conno ted a greater degree of love than

392


2/11

A THEO RY AND METHOD OF LOVE 393a "needing" profile. Thus i t appears that the init ial simplicityof Rub in's scales masks a m ore complex m ultidimensional reali ty.

An a t tempt to examine the re la t ionship be tween romant iclove and locus of control (Dion & D ion, 1973) produc ed inter-esting results with explicit application to scaling. In the courseof their study, the authors asked 255 subjects to rate their sub-jective experience o f roman tic love on 23 bipolar adjective i tems.The results of a factor analysis of the i tems suggested at leastfive different approac hes (o r styles) o f experie ncing love. Thesewere labeled volatile, circumspect, rational, passionate, and im-petuous.

There have been several other approaches to the scaling oflove (e.g., Hinkle & Spora kowsk i, 1975; Swenso n, 1972), thoughscale development and theory development were not t ightlyl inked. An exception was Mun ro and A dams (1978) , who di f-ferentiated ro man tic from con jugal love and related both to de-velopmental changes in individuals' role structures.

C o l o r s o f L o v eOne of the mo re interesting theories of love was proposed by

Lee (1973/1976) who forged a classification of several differentapproach es to love. After an extensive interview procedure andcomplex data red uction techniques, Lee proposed a typology oflove styles that formed a closed circle. Lee identified three pri-ma ry types of love styles: Eros (roma ntic, passionate love), Lud us(game-p laying love), Storge (friendship love), and three ma insecondary styles: Mania (possessive, dependent love), Pragma(logical, "sh opp ing list" love), and Ag ape (all-giving, selfless love).These secon dary styles were conceived as comp oun ds o f pairs ofprimar y styles. Analogous to chemical comp ounds, the secondarystyles are qualitative transfo rmations of the "base pri mar y ele-ments ." Thus, Mania i s a compou nd of Eros and Ludus, butMania is qualitatively very different from either primary. In thesame fashion, Pragm a is a compou nd o f Storge and Ludus, buthas very different properties. T he sam e holds true for Agape, acomp ound of Eros and Storge. One impl icat ion of the analogyto chemical compounds is that although the six love styles arelogically interrelated, each style has qualitative properties inde-pendent of all of the other styles. Empirically, measures of thesesix love styles should be orthogonal to each other. In sum, thelove styles are all equally valid ways of loving. There is no onetype o f love, but rather ma ny different types.Lee's ty pology is exceedingly rich theoretically, both becauseof i ts multidimensionality and grounding in research, and becauseit encompasses less extensive love theories that have been pro -posed. Fo r instance, exchange theory is probabl y a basis for Lee's( 1973/19 76) P rag ma (logical), wh ereas Clark an d Mills' (1979)co mm un al love is exemplified by Aga pe (selfless). Berscheid andWalster (1978) would recognize Eros as their passionate love,whereas com pani onat e love is probably best represented b y Storge(friendship). Kelley's ( 1983 ) pragm atic love would seem to equalPragma. Even Dion and D ion's (1973) factors appear very similarto Lee's (1973) constructs: Volati le = Mania, C ircumsp ect =Storge, Rational = Pragma, and Passionate = Eros. Thus, Leeoffers multidimensionality within a cohere nt theory.

Lee's research inspired the development o f a 50-item tru e-false scale to mea sure the six love styles (H atk off & Lasswell,1979; Lasswell & Lasswell, 1976). Ea ch sub ject received a score

on each of the six subscales by counting t he number o f trueresponses fo r the items in a given subscale. Thu s, each su bjectcould be profiled on all six subscales. It was reporte d th at aGutman-Lingoes Smallest Space analysis produced the six lovetypes as conceptually distinct (although none of the details o fthe data analyses were reported). Gender differences in love styleswere also reported. Rosenman (1978) correlated the Lasswells'love scale with Rubin's liking and loving scales. Rubin's lovescale correlated positively with the subscales representing Storge,Agape, and M ania, but not w ith Pragma, Ludus, and Eros. Theseresults fit Kelley 's (1983) analysis of Ru bin 's scale, with K elley'sneeding component equaling Mania, whereas caring, trust , andtolerance equaled Agape and Storge. What was missing fromRub in's scale was the passion o f Eros.

Although Lee's typology offers an intriguing combination ofconceptual richness and clarity, n o sustained work ha d been d onewith either Lee's theo ry or the Lasswells' scale until the cu rren tresearch prog ram was initiated.

A previous study (Hendrick, Hendrick, Foote, & Slapion-Foote, 1984) built on the Lasswells' work, using their items andnew ones in a Likert format. Approximately 800 students com-pleted the revised love scale. Fac tor analysis of the item s pro videdpartial suppo rt for Lee's theory. The secon dary styles of Mania,Pragma, and Agape emerg ed clearly as separate factors. However,each of the primar y styles (Eros, Ludus, and Storge) tended tocomb ine with anothe r style instead of emerging as independentfactors, and it was unclear whether Eros existed at all.

Theory bui lding and const ruc t measurement are joint boot-strap operations. Fro m the previou s results it was unclear whetherthe theory was partially wrong, or whether the mixed resultswere due to po or meas ureme nt scales. After much considerationof i tems, we concluded that there were enough q uestionableitems to warrant scale revision. Consequently, the love scale wasrevised substantially, and the cu rren t research was based on thisrevision.

In addition, ou r previous study fou nd fascinating gender dif-ferences on several o f the love subscales. Males w ere clearly mor eludic than females, but females were more pragmatic, storgic,and m anic in love att i tudes than males.

The intent o f Study I was to devise a measurem ent inst rumentthat would measure the six love styles/attitudes clearly, therebyproviding evidence that the six different conceptions o f love trulyexist. Assuming that the love styles exist in fact, we were alsointerested in the general process of examin ing the dom ain o f thetheory o f love styles. Toward that end, several b ackg roun d me a-sures were taken that might be related conceptually to the lovestyles, inclu ding gender, ethn ic differences, age, effects of pre viou sand cu rren t love experiences, an d level of self-esteem. This s tudywas part o f a larger study that also meas ured sexual att itudes.Only the love att i tude data will be considered in the currentreport . S tudy II was a replication o f Study I in a different geo-graphical area w ith slightly revised scales.

S t u d y IMethod

A revised questionnaire entitled Attitudes About L ove and Sex wasdeveloped based on the instrument used in ou r previous study (Hendricket al., 1984). The questionnaire included a brief explanation about thestudy of attitudes, an I l-item Backgroun d Inventory, a section entitled


3/11

39 4 CLYDE HENDRICK AND SUSAN HENDRICKLove Attitudes Scale that contained 42 love items, and a section entitledSexual Attitudes Scale hat contained 58 sexual attitude items. The itemsin the attitude sections were rated on a 5-category basis that was trans-formed into a 5-point numerical basis for data analyses: 1 = stronglyagree, 2 = moderately agree, 3 = neutral, 4 = moderately disagree, and5 = strongly disagree. The responses o all items were made on machine-scorable answer sheets.It should be noted that following he scaling radition in this research,subjects were asked to complete the scale with the current love partnerin mind, in so far as possible. To accommodate subjects not currently inlove, the following nstructions were included for completion of the lovescale:

Some of the items refer to a specific love relationship, while othersrefer to general attitudes and beliefs about love. Whenever possible,answer the questions with your current partner in mind. If you arenot currently dating anyone,answer he questionswith your most recentpartner in mind. If you have never been in love, answer in terms ofwhat you think your responses would most likely be.The questionnaire was administered during the fall semester of 1983and early springsemester of 1984 o groups of students aking ntroductory

psychologyat the Universityof Miami. A total of 807 students completedthe entire questionnaire and were included for data analysis.In addition,during the spring semester, 112 of these students completed the ques-tionnaire a second time at 4 to 6 week intervals in order to gather datafor a test-retest reliability analysis. The test-retest subjects were toldduring the second session that we were interested in whether love andsex attitudes change over ime and to complete he questionnaire n termsof their current feelings.

R e s u l t s

(41%) were age 18 or less, and some 235 stud ents (29%) wereage 19. The remaining 30% of the st udents were 20 or older.Most students were single and had never been married (96%).However, 16% stated they were now, or had in the past, livedwith someone of the opposite sex. Males and females did notdiffer on this item. The ethnic heritage of the sample was asfollows: Black (5.3%), Whi te-n on-H ispa nic 50.4%), White-Hi s-panic (29.0%), Oriental (7.7%), Other (7.6%). A substantial 161students (20%) indicated that they were international students.The religious heritage of the sample was as follows: Protestant(13.3%), Catholic (47.6%), Jewish (16.4%), None (4.7%), Other(18.1%).

One item asked students "How m any times have you been inlove?" Results were none (15.1%), one (36.8%), two (26.1%),three to five (17.7%), more than five (4.2%). Males and femalesdiffered on this item, X2(4, N = 807) = 26.6, p < .01, with thedifference showing up as greater extremes for males. By morethan a 2:1 margin, males had either never been in love or hadbeen in love three or more times.

Males and females also differed o n the questi on "Are y ou inlove now?" For males, 54.5% said no and 45.5% said yes. Forfemales, 36.1% said no and 63.9% said yes, Xz(1, N = 807) =26.2, p < .01. The same results occurred in a previous study(Hendrick et al., 1984).

A final background question attempted to measure self-esteem:"Th e way I feel about myself generally is." The great majo rity(84.6%) rated their esteem very positive or positive , and 12.4%rated themselves n e u t ra l Only 24 stud ents (3.0% of the sample)rated their esteem as negative or very negative.

Each of the six love styles was measured by 7 items. The items,grand means, and standard deviations are shown in Table 1.Because of the way the items were scored (i.e., 1 = strong ly agree,5 = strongly disagree), the lower the score the more a subjectsubscribed to the love style mea sured b y a given item. A perusalof the item means indicates that subjects tended to endorse theitems positively. A large major ity of the item m eans (except forLudus) fell on the ag reement side of the scale neut ral point (i.e.,3). The fact that the grand means tended toward the center ofthe scale suggested th at there were n o pro blems of end effects orscale restriction. Also, the size of the standard deviations indicatedthat degree o f agreement with a given item varied widely acrosssubjects.

These descriptive data suggested that the items were quitesuitable for intercor relation and subsequent factor analysis.

S a m p l e C h a r a c t e r i s t i c sThe Un iversity of Mia mi is a large, ur ban, privat e university.

It enrolls students from many states and has a large contingentof internationalstudents. Many Hispanic students, mostly Cuban,attend the University. Because large numbers of students takeintroductory psychology, the sample was reasonably represen-tative of the student population. The great diversity of the studentpopulation was especially valuable in developing an instr umentto measure love attitudes.

Selected sample characteristics were as follows. There were466 males (58%) and 341 females (42%). Some 330 students

Factor Structure of the Love Ite msThe 42 items were intercorrela ted and factored. Several POS-sible principal componen t solutions were examined. The best

solution extracted six factors, with uni ties in the diagonal, usingvarimax rot ation of the factors. The factor loading for each item(on its factor) is shown in the third col umn o f Table 1. Beforerotation, the percent of total variance accounted for by each ofthe factors was as follows: Eros (6.2), Ludus (6.8), Storge (4.3),Pragma (9.3), Mania (4.8), Agape (12.9). Agape was the firstfactor extracted and (by definition) it accounted for the mostvariance. However, it is worth noting that this first factor wasnot very general and did no t overshadow the other factors. In-stead, the factors were roughly comparable i n size, with on lymodest decreases in var iance per factor. Further, all six factorsaccounted for a healthy 44.2% of the total variance.

The factor loadings shown in Table 1 were substantial. Forthe critical seven items that defined a love scale, in two cases(Ludus, Pragma), no loading was less than .50. For two others(Agape, Mania), only one of the seven items had a loading lessthan .50. Eros had two items loading at less than .50, and Storgehad three such items. However, inspection of Storge suggests t hatit was a substantial factor.The full factor matrix showed quite remark able results. The7 critical variables for a scale showed strong loadings on a factor,and the loadings for the remaining 35 variables were low on th atfactor, often approaching zero. O nly two items were questionable.Item 5 (Eros) showed loadings of.25 to .35 on three other factorsin addit ion to Eros. Item 36 (Agape) also showed a loadin g of


4/11

A T H E O R Y A N D M E T H O D O F L O V ET a b l e 1Means, Standard Deviations, and Factor Loadings for the Love Attitude I tems

3 9 5

Study IFac tor

I t em M SD loadings MStudy II

FactorSD loadingsEros1. M y lover and I were at t racted to each o therimmediately after we fi rst met . 2.4 1.22. My lover and I have the right physical"che mist r y" between us. 2.0 1.03. Ou r lovemaking is very intense andsatisfy ing. 2.1 1.14. I feel that my lover and I were meant foreach other. 2.2 1.15. My lover and I becam e physical ly involvedvery quickly. (My lover and I b ecam eemot ional ly involved rather quickly.) 3.0 1.46. My lover and I real ly unders tand eachother. 2.1 1.17. My lover fit s my ideal standards o f physicalbeauty /handso mene ss. 2.1 1.0L u d u s8. I t ry to keep my lover a li t tle uncertainabout my comm i tment t o h im/her . 3 .2 1 .39. I bel ieve that wh at my lover doe sn't knowabout me won 't hur t him/her. 3.2 1.410. I have somet im es had to keep two of mylovers from finding out about each other. 3.6 1.5I 1. I can get over love affairs pretty easily andquickly . 3.5 1.312. My lover would get upset i f he/she kn ew ofsome of t he t h ings I 've done w i th o therpeople. 3.0 1.413. Wh en m y lover gets too depe nden t on me,I wan t to back off a l i t t le. 2.9 1.214. I enjoy playing the "ga me of love" wi th anum ber of di fferent partners. 3.8 1.3Storge15. I did not re alize that I was in love until Iactual ly had been for som e t ime. (It i shard to say exact ly where friendshipends and love begins.) 2.7 1.216. I cann ot love unless I first had caring fo rawhi le. (Ge nuine love first requirescaringfor awhile.) 2.2 1.117. I st il l have good fr iendships wi th almosteveryone wi th whom I have ever beeninvolved in a love relat ionship. (I expectto always be friends wi th the one I love.) 2.5 1.218. The b est kind o f love grows out of a longfriend ship. 2.3 1.219. It i s hard to say exact ly when m y lover andI fel l in love. (Our friendship merged

gradu ally into love over time. ) 2.6 1.120. Love is really a deep friendship, not amysterious, myst ical emot io n. 2.5 1.321. My m ost sat i sfying love relat ionships havedeveloped from good friendships. 2.5 1.2Pragma22. I consider what a person is going tobecome in l i fe before I commi t myse l fto him/he r. 3.1 1.323. I t ry to plan my l i fe careful ly beforechoo sing a lover. 2.9 1.224. It i s best to love some one wi th a simi larbackground. 2.7 1.225. A ma in considera t ion in choosing a lover i show he/she reflects on my family. 3.2 1.226 . An im por t an t fac to r in choos ing a par t ner

is wheth er or not he /she wi ll be a goodpare nt. 2.5 1.2

.48 2.3 1.2 .53

.76 2 .0 .9 .78

.68 2.1 1.1 .61

.65 2.1 1.1 .64

.36 2.4 1.2 .37

.57 2.0 1.0 .56

.59 1.9 1.0 .57

.70 3.4 1.4 .61

.67 3.6 1.4 .66

.69 3.8 1.4 .68

.55 3.9 1.2 .52

.57 3. l 1.4 .47

.50 3.0 1.3 .52

.72 4.2 1.2 .66

.36 2.5 1.2 .33

.50 1.6 .9 .33

.38 1.5 .9 .32

.69 2.3 1.1 .76

.49 2.6 1.3 .77

.57 2.5 1.3 .56

.69 2.7 1.3 .80

.69 2.9 1.3 .58

.68 2.8 1.2 .52

.54 2.6 1.1 .57

.69 2.8 1.2 .73

.69 2. 0 1. 0 .6 6(Table 1 cont inues on next page.)


5/11

3 9 6T ab le 1 (cont inued)

C L Y D E H E N D R IC K A N D S U S A N H E N D R IC K

Item

Study I Study IIFactor FactorM SD loadings M SD ioadings

Pragma (continued)27. One consideration in choosing a partner ishow he/she will reflect on my career. 2.8 1.228. Before getting very involved with anyone, Itry to figure out how compatible his/herhereditary background is with mine incase we ever have children. 3.3 1.2Mania29. When things aren 't right with my lover andme, my stomach gets upset. 3.0 1.330. When my love affairs break up, I get sodepressed t hat I have even thought ofsuicide. 4.3 1.131. Sometime s I get so excited about be ing inlove tha t I can 't sleep. 2.7 1.232. When m y lover does n't pay attention tome, I feel sick all over. 3.0 1.233. When I am in love, I have troubleconcen trating on anythin g else. 2.9 1.234. I canno t relax if I suspect tha t m y lover iswith someon e else. 2.3 1.235. If my lover ignores me for a while, Isometimes do stupid things to get his/her attention back. 3.0 1.3Agape36. I try to use my own strength to h elp mylover through difficult times. (I try toalways help my lover through difficulttimes). 1.7 .837. I would rather suffer myself than let mylover suffer. 2.2 1.038. I cannot be happy unless I place my lover'shappiness befo re my own. 2.6 1.139. I am usually willing to sacrifice my ow nwishes to let my lover achie ve his/hers. 2.7 1.140. Whate ver I own is my lover's to use as he/she chooses. 2.4 1.241. Wh en m y lover gets angry with me, I stilllove hi m/ he r fully and unconditionally. 2.1 1.042. I would endure all things for the sake ofmy lover. 2.8 1.2

.72 2.5 1.1 .67

.71 3.2 1.2 .56

.54 2.5 1.3 .54

.45 4. I 1.2 .46

.63 2.2 1.2 .64

.76 2.8 1.2 .74

.67 2.7 1.2 .72

.58 2.2 1.2 .57

.59 2.8 1.2 .50

.30 1.3 .6 .30

.74 1.9 .9 .68

.79 2.5 1.1 .83

.77 2.4 1.0 .77

.67 2.4 1.2. .64

.56 1.9 1.0 .52

.77 2.4 1.1 .69Note. Item s 5, 15, 16, 17, 19, 36 were revised from Stu dy I to Stud y II. The revision is shown in parentheses. Und er Study I, the data show n are fo rthe original version o f the item; under Study II, the data shown are for the revised item.

. 3 9 o n th e E ro s f ac to r. A f t e r ca re fu l co n s id e ra t io n , i t w as d ec id edto r e t a in t h ese tw o i t em s o n th e i r s ca l es fo r fu r th e r an a ly ses .

Formal Scale AnalysesE ach o f t h e l o v e s ty l e sca l es w as su b jec t ed to t h e s t an d a rd

rel iab i l i ty analysis o f the S tat is t ical Package fo r the Social Sciences(S P S S ) p ro g ram . In ad d i t i o n , t e s t - r e t e s t co r r e l a t i o n s w ere o b -t a in ed fo r a su b sam p le o f 1 1 2 su b jec t s . T h e r e su l t s a r e sh o w nin th e t o p p an e l o f T ab le 2 .

T h e a lp h a co e f f i c i en t s w ere su b s t an ti a l . A l l w ere . 7 0 + ex cep tS to rge, wh ich was .62 , a resu l t consis ten t w i th the re la t ively lowerfac to r l o ad in g s fo r t h e i t em s o n th i s s ca le ( s ee T ab le 1 ). T h e t e s t -r e t e s t co r r e l a t i o n s r an g ed f ro m a l o w o f .6 0 fo r E ro s t o a h ig ho f . 7 8 fo r P rag m a . A l th o u g h b ased o n a sm a l l e r N , w h ich m ig h thave affected s tab i l i ty , these resu l ts suggested so me rela t ive sh i f t -

i n g o f t h e l o v e s ty le s co res o n a sh o r t - t e rm b as is . O n e v e ry t en -t a t i v e co n c lu s io n i s t h a t t h e l o v e s ty l e s ca l e s a r e m easu res o fr e l a t iv e ly ch an g eab le a t t i t u d es , r a th e r t h an in d i ces o f en d u r in gpersonal i ty t ra i ts .

B ecau se o f th e n a tu re o f p r in c ip a l c o m p o n en t s f ac to r an a ly si s ,t h e f ac to r s w ere o r th o g o n a l t o each o th er . T h u s , i t em s w i th h ig hlo ad in g s o n o n e f ac to r an d lo w lo ad in g s o n o th e r f ac to r s w o u ldn ecessa r i ly b e r e l a t i v e ly i n d ep en d en t o f each o th e r . H o w ev er , i td o es n o t fo l lo w th a t s ca le s co res b ased o n a su m o f i t em s ( fo rspecif ic facto rs) wi l l be ind epen den t o f each o ther. To assess degreeo f lo v e sca l e i n d ep en d en ce , s u m sco res w ere co m p u ted fo r eacho f t h e s ix scal e s an d in t e r co r r e l a t ed . T h e r e su l t s a r e sh o w n inth e t o p p an e l o f T ab le 3 .

Because o f the large N, very smal l co rrela t ions we re s ign if ican t .S ev e ra l o f t h e s ig n i f ican t co r r e l a t i o n s sh o w n in T ab le 3 w eret r i v i a l i n s i ze. T h e o n ly sca l e w i th p o ss ib le p ro b lem s w as A g ap e .


6/11

A THEORY AND METHOD OF LOVE 397Table 2Reliability Analyses o f the Love Scales

Measure N Eros Ludus Storge Pragma Mania AgapeStudy I

Mean interitem correlation 807 .27 .31 .19 .38 .28 .43Alpha 807 .70 .76 .62 .81 .73 .84Standardized item alpha 807 .72 .76 .62 .81 .73 .84Test-Retest correlations 112 .60 .72 .72 .78 .75 .73Study II

Mean interitem correlation 567 .26 .29 .24 .30 .27 .41Alpha 567 .70 .74 .69 .74 .72 .83Standardized item alpha 567 .71 .74 .68 .75 .72 .83Test-Retest correlations 55 .74 .82 .74 .71 .70 .81Note. There were seven items in each subscale. Test-retest correlations were based on the sum of the seven items.

It was significantlycorrelated wi th four of the oth er scales. How-ever, the largest of these correl ations (.30) accoun ted for only 9%of the vari ance in scale scores. It seems likely that the modestlevel of the correlations reflected the c omm on method varianceof the rating instrument.

In summ ary, the analyses suggested a viable set of scales tomeasure the six love styles. The scales emerged nice ly from factoranalysis of the items, demonstrate d suitable inter nal reliability,and reasonable independence from each other when consideredas additive scales.Backg round Var iables and At t i tudes

A num ber of specific questions pertainin g to the subjects'background were included to assess in an orderly way some ofthe theoretical supposi tions about how different love styles func-tion. The approach used treated each background variable as aninde pend ent variable, using subjects' sum scores on a given lovescale as a depe ndent variable. One-way analyses of variance wereperfo rmed on the data for each love style. (Attempts at two-wayanalyses by crossing two independen t variables yielded nonor-thogonal effects because of the wide disparity in cell frequencies.)

Table 3lntercorrelations Am ong L ove Scale Sum ScoresScale Ludus S to r ge P r a g m a Mania Agape

Study IErosLudusStorgePragmaManiaStudy IIErosLudusStorgePragmaMania

.00 -.0 5 -.05 .07 .27*- - . 0 3 . 1 2 " - . 0 5 - . 2 8 *

- - . 2 3 * . 0 6 . 1 3 "- - . 1 1 " . 0 5

- - . 3 0 *

-.22* -.0 4 -.0 4 .13" .32*- - - . 0 5 . 0 9 * - . 0 3 - . 4 2 *

- - . 2 5 * . 0 1 . 1 5 "- - . 1 3 " . 0 4

- - . 2 3 "

Note. N = 567 in Study II; N = 807 in Study I.*p < .01.

The means and F ratios for the six love styles are shown inTable 4 for several background variables. The F ratio for eachone-way analysis is shown at the top of the relevant column ofmeans. The means reported in Table 4 were derived by summ ingthe seven items on a scale, an d takin g the average. Thus, mean scould vary from 1.0 to 5.0, an d the lower the mean the strongerthe p arti cular love style.

It will be useful to consider the results in terms of the back-ground variables.Age of subjects. There were no sig nificant mai n effects of agefor any of the love styles. Therefore, the mean s of the love stylesas a functi on of age are not reported in Table 4. There was ofcourse not much variation in age for this college sample. Theitem was includ ed to rule out shifts in preferred love styles fromfresh man to juni or years. The lack of results suggests that suchshifts did not occur.Gender differences. Males were significantlymore ludic thanfemales. Females were significantlymore storgic, pragmatic , andman ic than males. Males and females did no t differ on Eros andAgape. The pa ttern of these results was very similar to the out-come of a previous study (Hendr ick et al., 1984).Ethnic background. Two questions tapped ethnic/culturalbackground. One item stated "My ethnic heritage is," with fiveresponse categories. The second item stated "I am an inte rna-tional student" with ye s or no response categories. For ethnicheritage, there were no differences on Lu dus and Mania. ForEros, Black, White-non-Hispanic,and White-Hispanic studentswere at about the same level, whereas Oriental students weresignificantly ess erotic in or ientation. The om nibus category of"Other" yielded a mean similar to the Oriental mean, althoughdue to small sample size for the Other category, its mean did notdiffer from the Black mean for Eros. As might be expected, therewas also a tendency for Oriental students to be more storgic andpragmatic than were the other three ethnic groups. The tendencywas strongest for Pragma. Finally, Black students were leastagapic, although not all mean comparisons were significant dueto the small numb er of Black subjects (N = 43).

International students (a diverse grouping) rated themselvesas less erotic, more ludic, storgic, an d pragmatic than U.S. citi-zens. The two groupings did not differ on Ma nia or Agape.Number of times in love. This question asked directly "How


7/11

398 CLYDE HENDRICK AND SUSAN HENDRICKman y times have you been in love," and provided five responsecategories ranging from n o n e to m o r e t h a n f i ve . This last categorywas endorsed by only 34 of 807 subjects, and mean compari sonsinvolving it were no t very reliable. As expected, s tudents whohad never been in love were least erotic, and there were no dif-ferences between the Eros means for one, two, three to five, orfive or more times in love. The critical cut was n o n e versus allother categories.

Because Eros is supposed to value the intens ity of love, thennever to have been in love is never to have been an Eros. It wasnoted previously that relatively more males than females hadnever been in love before. However, because the difference be-tween males and females for Eros was not significant, the resultsfor number of times in love cannot be attributed to a confoundwith gender.

Ludus is theoretically conceived as "love as a game." Loveaffairs for Ludus types should not have the fire of passion thatis true for Eros. In fact, depending on how a ludic person defineslove, he or she should either have been in love many times (eachcasual affair is "love") or none (each casual affair is defined asa "casual affair"). The means for num ber of times in love showedprecisely this result for Ludus. Subjects were significantly more

ludic if they had never been in love or had been in love five ormore times, than were subjects in the intermediate categories.There were no differences due to number of times in love forPragma. The main effect of Storge, though significant, did n otsuggest any particular interpretation of the means. There weresignificant main effects for both Mani a and Agape. Subjects whohad never been in love or who had been in love five or moretimes were less manic than the other groups. Also, as expected,subjects who had never been in love were less agapic than theother groups.

A r e y o u i n l o v e n o w ? This item was important for theorytesting. It was expected that Eros types would tend to be i n love,Ludus types would not, and because Storge is friendship andPragma is practical, these two types would not differ. Predictionsfor Mania and Agape were uncertain. Turning the situationaround in terms of the current independent-dependentvariableconvention, the following results emerged. Subjects "in love now"were more erotic, more storgic, more manic, more agapic, andless ludic than subjects "not in love now." Clearly, the perceptionof being currently in love (or not) cued off an entire responsepattern that involved five of the six love styles. These results posecertain in teresting theoretical issues to which we will return.


8/11

A THEORY AND METHOD OF LOVE 399Self-esteem. Because of the low f requency of the subjects

rat ing themselves as negative or very negative in self-esteem, thesetwo categor ies were collapsed with neutral to form a category ofneutral or lower. Clear p r ed ic t ions could be m ade f or Er os andMania. Eros gives fully, intensely, and takes r isks in l ov e- - i t re-quires substantial ego strength. Conversely, people high in self-es teem shou ld mo re l ikely be erot ic than people low in self-es-teem. The sam e reasoning should ap ply in reverse for Mania. Infact , one reason mani c lovers are man ic is because of uncer ta intyof self in the relat ionship.

The results in Table 4 sup por t these predict ions . There was aclear posi t ive relat ion between self -es teem and agree ment w iththe Eros scale. There was a c lear negative relat ion between self -es teem and agreem ent with the Man ia scale . I t is of interes t tonote that subjects with very posi t ive self -es teem were mo re ludicthan were ei ther of the other two self -es teem groups . This resultmakes sense, but was not predicted. Apparently, i t takes goodego s trength to play ser iously at love as a game.

There were no dif ferences due to self -es teem on Storge,Pragma, or Agape. None was expected for S torge or Pragma;however, w e rathe r expe cted peop le with very positive self-esteemto be m ore agapic . But such was no t the case.

Althoug h the results revealed success in the technical aspectsof scale developme nt as well as some theoret ical conf irmatio n ofthe love s tyles , i t was fel t that a c onf irm atory s tudy w as neededto substantia te Lee ' s theory and the Love Att i tudes Scale as amajor new development in research on love.

S t u d y I IM e t h o d

The Love Attitudes Scale was subjected to limited revision. One itemeach from Eros and Agape was revised and four items from Storge wereeither partial ly or fully revised.The new revision of the Attitudes Ab out Love and Sex scale includeda bri ef explanation abou t the study o f attitudes, a 17-item BackgroundInventory (6 items added), a section entitled Love Attitudes Scale thatcontained 42 items, and a section entitled Sexual Attitudes Scale thatcontained 46 items. It was administered to students taking introd uctorypsychology courses at Texas Tech University during the fall of 1984. Thefinal sample consisted of 368 fem ales and 199 males, 567 subjects in all.Selected sample characteristics sh owed differences from the tw o previousSouth Florid a samples. Fewer students were Jewish (.7%) and more w ereProtestant (49%). More students were White-non-Hispanic (83%) andfewer were White-H ispanic (11%), Black (2.5%), and Oriental (2%). Fewerof the Texas students had never been in love or had been in love three o rmore times. The samples were relatively similar on other variables suchas age and self-esteem.

R e s u l t sExamina t ion of the i t em means ind ica ted tha t pa t te rns of en-

dorseme nt c losely replicated those of S tudy I , though there wass l ightly more reject ion of Ludus and s l ightly more endorsem entof the other scales ( see Table 1). The grand means , s tandarddeviat ions , and factor loadings of the i tems are shown in the las tthree colu mns of Table 1. Revis ions of i tems are shown in pa-rentheses beneath the or iginal i tems. Thus the f inal vers ion ofthe scale is represented in Table 1.

The love scale was subjected to a pr incipal compo nents analys is

(as in Study I ) , and s ix factors were extracted. Before rotat ion,the percent of total var iance accounted for by each factor wasEros (4. l) , Lu dus (8.5), Storge (4.5), Prag ma (7.1), Ma nia (5.2),and Agape (13.6) , accounting for 43.1% o f the total var iance(s tructure almost identical to that in Study I ) . The factor loadingsnear ly replicated those of S tudy I , with no loading on Ludus ,Pragma, or Mania less than .45, and one E ros i tem, one Agapeitem, and three Storge i tems loading less than .45 (but above.30) . The scale s tructure was very clear , with only four i temsloading on mo re than one factor. In each of the four cases , thei tem loaded mos t h igh ly on i t s appr opr ia te f ac tor and had anegative loading on a second factor . These loading pat terns wereal l conceptually congruent with Lee ' s theory.

Reliability analysis yielded standa rdized ite m alpha coefficientsvarying f rom .68 for S torge to .83 for Ag ape (s imilar to Study I ,see Table 2). Test- retes t re l iabi l i ty values (bottom panel o f Table2) indicated coefficients of .70 or above for all scales. Intercor-relat ions of sum med scale scores revealed sui table independenceand a pa t tern of cor relat ions s imilar to that shown previously(see Table 3) , a l though the cor relat ion of Agape with L udus wassomewhat higher , probably due to the s tronger rejection of Lud usby the sample.

One-way analyses of var iance were per formed on the back-ground var iable data for each love s tyle, s imilar to the analysesof Study I . The results are shown in Table 5. Consis tent withStudy I , there were no meaningful age differences in love styles.There were gender dif ferences on a num ber of backgrou nd var i-ables ( including times in love, in love now, and self-esteem), aswell as on five of the six love scales. Males were significantlymore ludic ( though both genders were above the mean of 3.0and thus relat ively rejected the ludic s tyle), whereas females weremore erot ic , s torgic , pragmatic , and manic than were males . Thepattern of results replicated almost exactly the results of the pre-vious s tudies . Although ethnic homogeneity precluded thebroader analys is of e thnic groups presented in Study I , exami-na t ion of W hi te - Hispanic and W hi te - non- Hispanic sub jec t sshowed the Hispanics to be s ignif icantly more ludic than thenon-Hispanics .

Subjects who repor ted tha t they had never been in love wereleas t endors ing of Eros and Agape, whereas subjects who hadbeen in love twice were m ost endo rs ing of these two s tyles. Therewere also s ignif icant dif ferences on Ludus , with those who hadbeen in love three o r more t imes m ost ludic ( fol lowed closely bythose who h ad never been in love); those subjects who had beenin love once were leas t ludic . Although s ignif icant differences didnot a ppear on Storge or Mania, the results considered suppor t iveof Lee ' s theory in Study I ( results on Eros , Ludus , Ag ape) werereplicated almost exactly in Stud y I I .

For the quest ion o f whether a subject was in love at the t im eof tes t ing, subjects " in love now" were mo re erot ic and agapicand less ludic and pragmatic than subjects not in love. Theseresults were very s imila r to those of S tudy I . For the self -es teemvar iable , whose results in Study I appeared to suppor t Lee ' stheory (Eros high in self-esteem, Mania low in self-esteem), resultsof Study II were fully consistent. Subje cts whose self-esteem wasvery positi ve most endo rsed Eros (with the neutral category leas tendors ing) , and those wh o were neutral in self -es teem most en-dorsed Mania (whereas very positive subjects endorsed i t leas t) .


9/11

4 0 0 C L YD E H E N D R IC K A N D S U S A N H E N D R IC KT ab le 5Study 1I." Means and F Ratios for Each Love Style as a Function O f Selected Background Variables

Love stylesVariable N Eros Ludus Storge Pragma Mania Agap e

Gen der F = 5.2* F = 22.9* F = 9.5* F = 9.5* F = 7.8* F = 1.8Males 199 2.2a 3.3, 2.4a 2.8a 2.9, 2.2Fema les 368 2. lb 3.7b 2.2b 2.6b 2.7b 2.1Ethn icity F = .1 F = 7.3* F = .1 F = .2 F = 1.1 F = .1Wh ite-n on-H ispa nic 472 2.1 3.6a 2.2 2.7 2.8 2.1Wh ite-H ispa nic 63 2.1 3.3b 2.2 2.6 2.7 2.2No. tim es in love F = 7.0* F = 4.8* F = .2 F = .5 F = .6 F = 8.1"No ne 68 2.4a 3.4~b 2.2 2.6 2.8 2.5,On e 232 2.1b 3.7c 2.3 2.7 2.7 2.1 ~Two 181 2.0b 3.6~ 2.2 2.7 2.8 2.0c3 or mo re 86 2.3a 3.3a 2.3 2.8 2.8 2.2bIn love now F = 99.7* F = 28.7* F = .4 F = 7.2* F = .2 F = 78.5*No 252 2.4, 3.4~ 2.2 2.6~ 2.8 2.4a

Yes 315 1.9b 3.7b 2.3 2.8b 2.8 1.9bSelf-esteem F = 5.4* F = .1 F = .4 F = .5 F = 8.7* F = .6Very positive 103 2.0a 3.6 2.3 2.7 3.0~ 2.2Positive 346 2. la 3.6 2.2 2.7 2.7b 2.1Neu tral or lower 118 2.3b 3.5 2.3 2.7 2.6h 2.2Note. Means could vary from 1.0 to 5.0. The lower the mean, the greater the agreement with the given love style. Within each column, for eachvariable, me ans with n o subscripts in co mm on differed at the .05 level, either by the F test directly for a pair of means o r by the M ultiple Ran ge Testfor three or more means.* p < .05.

T h u s th e r e su l t s fo r S tu d y I I w ere a lm o s t fu l ly co n s i s t en t w i ththose o f S tudy I fo r the the oret ical ly s ign if ican t var iab les o f gender ,n u m b er o f t im es in l o v e , w h e th e r a su b jec t w as i n l o ve a t t h ep resen t t im e , an d se l f - e s t eem . S ev e ra l i s su es m er i t d i scu ss io n ,in c lu d in g t ech n ica l a sp ec t s o f s ca le d ev e lo p m en t , t h e n a tu r e o fth e l o v e s ty l e co n cep t s , t h eo re t i ca l i s su es r ev o lv in g a ro u n d th elo v e s tyl e s, an d so m e n o t io n a b o u t w h ere t h i s r e sea rch w i l l ev en -tu a l ly lead . R esu l t s f ro m th e tw o s tu d ie s a r e so s im i l a r t h a t t h eycan fo r t h e m o s t p a r t b e d i scu ssed to g e th e r .

G e n e r a l D i s c u s s i o nScale Cons truc t ion

O n e s t ro n g m o t iv e fo r t h e p resen t s tu d ie s w as t h e l ack o f f ac to rclar i ty in Hend rick e t a l. (1984) . I t was uncle ar wheth er the theoryo f l ov e s ty le s was i n co r r ec t , o r w h e th e r t h e m e asu r em e n t s ca l en eed ed i m p ro v em en t . W e o p ted fo r t h e seco n d a l t e rn a t iv e , an dth e r e su l t s f ro m th e cu r r en t s tu d i e s i n d i ca t e t h a t L ee ' s (1 9 7 3 /1 9 7 6) t h eo r y i s v i ab l e , an d th a t each o f t h e s ix co n cep t s o f l o v ecan b e m easu red in a c l ea r m an n er .

In t e rm s o f t h e v a r io u s c r i t e r i a fo r s ca l e co n s t ru c t io n an d v a l -id a t io n , t h e r e su l t s w ere ab o u t a s g o o d a s can b e ex p ec t ed f ro mrea l d a t a o b ta in ed f ro m rea l su b jec t s . W i th t h e ch an g es m ad eb e tw een S tu d ies I an d I I , t h e l o v e sca l e m ig h t b e co n s id e red ina n ea r ly f in a l fo rm a t t h i s p o in t . A l th o u g h m o re w o rk sh o u ld b edon e bef o re i t i s u sed cl in ical ly , the scale i s adequate in i t s p resen tfo rm as a r e sea rch in s t ru m en t fo r co r r e l a t i o n w i th o th e r s ca l e s ,p rese l ec t io n o f su b jec ts , an d so fo r th .

W h a t t h e Lo v e S c a l es M e a s u r eC are fu l p e ru sa l o f t h e co n ten t o f t h e i t em s in d ica t ed th a t t h e

ev o lu t io n o f i t em se ts h as r em a i n ed f a i t h fu l t o t h e t h eo re t i ca lco n cep t io n o f each lo v e s ty l e.

1 . Eros: S t rong physical p referen ces , ear ly a t t ract io n , an d in -t en s i ty o f em o t io n a re a t t r i b u t e s o f e ro t i c lo v e , a lo n g w i th s t ro n gco m m i tm en t t o t h e l o v e r . L o v e i s h ig h ly v a lu ed b y E ro s t y p es .T h e i t em s in T ab le 1 r e f l ec t t h ese a t t r i b u t e s .

2 . L u d u s : L o v e as an in t e r ac t io n g am e to b e p lay ed o u t w i thd iv e r se p a r tn e r s ap p ea r s t o b e t h e m a in a t t r i b u t e o f L u d u s ty p es .D ecep t io n o f t h e l o v e r i s accep tab l e w i th in p ro p e r ro l e l im i t s .T h e re i s n o t g rea t d ep th o f f ee lin g ; i n d eed , t h e l u d i c l o v e r is w aryo f em o t io n a l i n t en s i ty f ro m o th er s . L u d ic l o v e h as a m an ip u la t i v eq u a l i t y t o i t . T h i s a sp ec t r e su l ts i n a p p a ren t l o w er so c ia l d es ir -ab i li t y, a s r e fl ec t ed in t h e i t em m ean s in T ab le 1 . I t i s im p o r t a n tto n o te , h o w ev er , t h a t t h e re a r e l u d i c a sp ec t s t o m an y , i f n o tm o s t , l o v e r e l a t io n sh ip s . L ee ' s (1 9 7 3 /1 9 7 6 ) p o in t w as t h a t t h i sap p ro a ch to l o v e r e f l ect s an ex i s t i n g r ea l i t y fo r m a n y p eo p le . T h ei t em s an d th e i r l o ad in g s i n T ab le 1 su g g es t t h a t t h i s s ty l e w asm easu red w e l l .

3. S to rg e : T h i s s ty l e r e fl ec t s an i n c l in a t io n to m erg e lo v e an dfr iendsh ip . There is no f i re in s to rg ic love; i t i s so l id , down-to -ea r th , an d p resu m ab ly en d u r in g . T h i s "ev o lu t io n a ry " r a th e r t h an" rev o lu t io n a ry " em p h as i s i s re f l ec ted b y m o s t o f t h e S to rg e it em s ,sev e ra l o f w h ich h av e e x t r em e ly h ig h lo ad in g s .

4 . P rag m a: R a t io n a l ca l cu l a t i o n w i th a fo cu s o n d es i r ed a t -t r i b u t e s o f t h e l o v e r is cen t r a l t o p ra g m at i c l o v e . In f ac t , " l o v ep lan n in g " m ig h t b e an ap t d esc r ip t io n . B ecau se P rag m a ty p es


10/11

A THEORY AND METHOD OF LOVE 401use cr i ter ia matching, i t is easy to view them as "co mpu ter m at-ing" people. The i tems and loadings shown in Table 1 indicatethat this love s tyle was also well measured.

5. Mania: Reading the i tems suggests that M ania is "sympto mlove," based on uncer tai nty o f self and th e lover. I t may be mostcharacter is t ic of adolescents , but examp les of older mani c loversf requently occur . The i tems and their factor loadings indicatesuccess in construct measurement.

6. Agape: Lee did no t f ind this s tyle manifes ted ful ly in actualhum an beings . However , the factor results in Table 1 suggest thati t is a viable s tyle. Clear ly i t is an al l-giving, nond eman ding love.In fact , the i tem with a modest loading ( I tem 36) mentions"s trength " ( in Study I ) , a theme n ot occ urr ing in any of the otherAgape i tems. Even the i tem revis ion with "s tre ngth" deleted wassomewhat low in i ts loading. Perhaps any i tem n ot deno ting self -abnegatio n would load less s trongly on A gape.

The conclus ion of this detai led scrut iny is that the s ix lovescales appear to b e conte nt val id as well as technical ly sound asmeasurement scales .

Personality or Attitudes?Do the love styles measure endur in g personali ty t ra i ts or more

transient a t t i tudes? This interes t ing quest ion canno t be answereddef ini t ively by the present s tudies . Lee (1973/1976) discussedthe love styles as a typology. To psychologis ts this app roach im-plies t ra i ts . But Lee also bel ieved that i t is poss ible to be s imul-taneously in o ne type o f relationsh ip with on e person (e.g., erotic) ,an di n another typ e (e.g., ludic) with a second person. This pos-s ibi l i ty implies that the cause o f the love s tyle lies in the natureof the relat ionship with another person.

Conceptually, i t might be argued that love s tyles par take ofboth tra i t and s tate characteris t ics . One interes t ing aspect of thelove s tyles is that they vary in emotional intensi ty. Eros and Maniaare high in emotion, Agape is average, and Ludus , S torge, andPragm a are al l low. To whatever extent emotio nal express ivi ty isa tempe rame ntal facet of the person, to that extent there maybe a const i tut ional predisposi t io n toward dif ferent love s tyles .

At the same t ime, data f rom the present s tudy also point inan at t i tudinal direct ion. Ethnic and gender dif ferences suggestthe effects of sociali zation differences. The fascinatin g results for"are y ou in love now" also suggest an at t i tu de conception o f thelove styles . S tudents in love now were "m ore " on several of thes tyles than s tudents no t in love now. In contras t , the self -es teemdata may perhaps be interpreted as suppor t ing a t ra i t (personali ty)interpretat ion of the love styles ( those endors ing Eros highest inself -es teem, those endors ing M ania lowest in self -es teem).One theoret ical appr oach is to consider the s ix love s tyles asref lecting a s ix-dimen sional matr ix in each person' s psyche. Ev-eryone has some loca t ion at a given t ime on each o f the dimen -sions. Con stitution al differences may te nd to bolster one (or more)dimensio n. However, specif ic social izat ion pract ices also af fectthe deve lopment of the conceptual love matr ix. Poss ibly somedimensions are more changeable by exper ience than other di-mensions . Relat ive s tanding on the s ix dimensions m ay vary overt ime. S trong exper ience, such as "being in love now," may causea f lare-up on several dimensions , perhaps resulting in tem porar i lycor r e la ted d imens ions tha t a r e uncor r e la ted under condi t ions ofordinary, nonintense emotional exper ience.

This theory needs much elaborat ion. But i t does point theway toward an explanation of why the love s tyles seem to beboth personali ty t ra i ts and malleable at t i tudes . We need n ot beforced to choose in an e i ther /or fashion.

Further Theoretical IssuesThe results in Tables 4 and 5 bear on several issues that mer i t

fur ther considerat ion.The mean dif ferences in love s tyles between males and females

replicated our previous work (Hendr ick et a l . , 1984) . Hatkoffand Lasswell (1979) fou nd roughly s imilar gender dif ferences .To an extent the dif ferences in love at ti tudes paral le l male- femaledifferences in attit udes tow ard s exuality (e.g. , Ferrell, Tolone, &Walsh, 1977; Laner, Laner, & Palmer, 1978; Medora & Wood-ward, 1982; Mercer & Kohn, 1979). In general, males are morepermiss ive and ins trumen tal in their sexual a t t i tudes (Hendr ick,Hend rick, Slapion-Fo ote, & Foote, 1985), a resu lt consistent withmales being more ludic in their love s tyles . Tradit ionally, femaleshave been more conservative in sexual a t t i tudes , a conservatismthat p ar t ia l ly s tems f rom social izat ion to view sex as a preciouscommod i ty tha t mu s t be guar ded . Also , women have h i s tor ica l lybeen social ized to m arry b oth a love par tn er and a potentia lprovider. From such a s ta te of dependence on males , i t wouldbe surpr is ing if females were not m ore pra gmati c than males .The same reasoning may poss ibly account for females being mores torgic than males . The same social ized depend ency m ay alsoaccount for more m anic at t i tudes by females , a l though this ef fectmight be due to an ar t i fact , namely that females repor t moresym ptom s in general than males. In any event, gender differencesin love s tyles is an impor tant topic wor thy of more researcheffort.

The cur rent s tudies merely suggest that e thnic dif ferences inlove s tyles may be a f rui tful direct ion for future research. O neinteres t ing outcome in Study I was that O r iental s tudents seemedrelatively low in affect ( low in Eros, high in Storge and Pragm a).I t may well be that the s ix love s tyles do not captu re prope r lyOriental concep tions of love. In Stud y I I , Hispanic subjects ap-peared more ludic than non-Hispanic subjects . Future cross-cultural research should be sensi t ive to the poss ibi l i ty that morethan s ix s tyles exis t and that dif ferent s tyles are relat ively moredom inan t in dif ferent cultures .

I t would appear , f rom this research, that the love s tyles arenot independent of one ' s cur rent love s i tuat ion, or for that matter ,the num ber of pas t love relat ionships . The issue mer i ts fur thers tudy. I t would be des irable to m onito r poss ible changes in loveatt i tudes as a love relat ionship progresses f rom f irs t encounterto binding commitment. The diverse results for " in love now"versus "not in love now" indicate that something impor tant isgoing on. One theoret ical interpretat io n was noted ear l ier ; otherposs ibi l i t ies should b e ex amined.

Future DirectionsResearch on at t i tudes toward love can lead in m any direct ions:

More work is needed to assess s tabil i ty of love s tyles . I f scoresare relat ively s table, then the scale could become a valuab le toolfor preselect ion of subjects for a wide var iety of interact ion s tud-


11/11

402 CLYDE HENDRICK AND SUSAN HENDRICKies. It is of great intere st to investigate how well males a nd femaleswi th di f ferent dom inan t love s tyles "mesh ," as comp ared wi thcouples wi th s imi lar s tyles .

Descr ipt ive work on the love s tyles held by society would beof great value. I t may b e that ( in Western society a t leas t) peoplego through a kin d of mo dal developmental sequence o f love s tyles.As noted previously, manic love may be most character i s t ic ofadolescents . In ear ly adul thood the prefer red s tyle may evolvetoward Eros , which in turn may evolve toward S torge and Pra gmadur ing the middle and la ter years . And we have al l known atleas t one "o ld" co uple who ap peare d very agapic . Such a devel -opment a l s equence woul d account fo r t he common obse rva t ionnoted by Wals ter and Wals ter (1978) that wh at usual ly s tar t s aspass ionate love (Eros) sooner or la ter set t les down to compa-nionate love (S torge, Pragma) . I f such a " love his tor y" occ ursfor substantia l numb ers of people , then knowledge of the sequencecou l d enab l e peop l e t o i n t e rvene t o change (o r come t o t e rmswith) their histories (e.g., see Gergen, 1973).

Thus, the ramif icat ions for future research app ear indef ini te lybroa d in scope. We bel ieve that a l l of them are wor th pursuing.W hat i s more i m por t an t t han l ove?

R e f e r e n c e sBerscheid, E., & Walster, E. (1974). A lit tle bit ab out love. In T. L. Hus ton

(Ed.), Founda tions of interpersonal attraction. New York: AcademicPress.Berscheid, E., & Walster, E. (1978). Interpersona l attraction (2rid ed.).Reading, MA: Addison-Wesley.Blau, P. M. (1964). Exchange andpow er in social life.New York: Wiley.Clark, M. S., & Mills, J. (1979). Interperson al attraction in exchange andcommun al relationships. Journal ofPersonality and Socia l Psychology,

37, 12-24.Cook, M., & Wilson, G. (Eds.). (1979). Love and attraction: An inter-national conference. Oxford: Pergamo n Press.Dermer, M., & Pyszczynski, T. A. (1978). Effects of erotica upon men'sloving and liking responses for women the y love. Journal ofPersonalityand Social Psychology, 36, 1302-1309.Dion, K . L., & Dion, K. K . (1973). Co rrelates of romantic love.Journalof Consulting and Clinical Psychology, 41, 51-56.Ferrell, M, A., Toione, W. L., & Walsh, R. H. (1977). Ma turation al andsocietal changes in the sexual double-standard: A pan el analysis (1967-1971; 1970-1974). Journal o f Marriage and the Family, 39, 255-271.Gergen, K. J. (1973 ). Social psychology as history. Journal ofPersonalityand Social Psychology, 26, 309-320.Hatko ff, T. S., & Lasswell, T. E. (1979). M ale-fe male simila ritie s anddifferences in con ceptualizin g love. In M. Co ok & G. Wilson (Eds.),Love an d attraction: An international conference. Oxford: PergamonPress.

Hendrick, C., Hen drick, S., Foote, E H ., & Slapion-Foo te, M. J. (1984).Do men and women love differently? Journal of Social and PersonalRelationships, 1, 177-195.Hendrick, S., Hendrick, C,, Slapion-Fo ote, M. J,, & Foote, E H. (1985).Gender differences in sexual attitudes. Journal ofPersonality and S ocialPsychology, 48, 1630-1642.Hill, C. T., Rubin, Z., & Peplau , L. A. (1976). Brea kups before ma rriage:The end o f 103 affairs. Journal o f Sociallssues, 32(1), 147-168.Hinkle, D. E., & Sporakowski, M. J. (1975). Attitudes toward love: Areexamination. Journal o f Marriage and the Family, 37, 764-767.Kelley, H. H. (1983). Love and com mitm ent. I n H. H. Kelley, E. Berscheid,A. Christensen, J. H. Harvey, T. L. Huston, G . Levinger, E. McClin tock,L. A. Peplau, & D. R. Peterson (Eds.), Close relationships. New York:Freeman.Laner, M. R., Laner, R. H., & Palmer, C. E. (1978). Perm issive attit ude stoward sexual behaviors: A clarification of theoretical explanations.Journal of Sex Research, 14, 137-144.LassweU, T. E., & Lasswell, M. E. (1976). I love you but I'm not in lovewith you. Journal of Marriage and Fam ily Counseling, 38, 21 1-224.Lee, J. A. (1973). The colors of ove: An exploration of he ways o f oving.Don Mills, Ontario : New Press. (Popu lar Edition, 1976).Medora, N., & Woodward, J. C. (1982). Premarital sexual opinions ofundergraduate students at a midwestern university. Adolescence, 17,213-224.Mercel; G. W., & Kohn, P. M. (I 979). Gender differences in the inte gratio nof conservatism, sex urge, and sexual behavio rs among college students.Journal of Sex Research, 15, 129-142.Munro , B., & Adams, G. R. (1978). Love American style: A test of rolestructure theory on changes in attitudes toward love.Human Relations,31, 215-228.Rosenman, M. E (1978). Liking, loving, and styles of loving.PsychologicalReports, 42, 1243-1246.Rubin, Z. (1970). Measurem ent of romantic love. Journal o f Personalityand Social Psychology, 16, 265-273.Rubin, Z. (1973). Liking and loving: An invitation to social psychology.New York: Holt, Rinehart, and Winston.Rubin, Z . (1974). Fro m liking to loving: Patterns of attraction in datingrelationships. In T. L. H uston (Ed.), Foundations of interpersonal at-traction. New York: Academic Press.Rubin, Z. (1984). Toward a science of relationships. Contemporary Psy-chology, 29, 856-858.Steek, L., Levitan, D., McLane, D., & Kelley, H. H. (1982). Care, need ,and c onceptions of love. Journal o f Personality and Social Psychology,43, 481-491.Sternberg, R. J., & Grajek, S. (1984). The nature of love. Journal ofPersonality and Social Psychology, 47, 312-329.Swenson, C. H. (1972). The be havior of love. In H. A. O tto (Ed.), Lovetoday. New York: Association Press.Walster, E., & Walster, G. W. (1978). A new look at love. Reading, MA:Addison-Wesley.

Receive d Augus t 13, 1984Revis ion received Apr i l 9 , 1985 9

Documents

Hendrick and Hendrick Love Attitudes