Back to Journal Cover Page
Volume 4, Number 3
Submitted: January 9, 1999
Resubmitted: March 5, 1999
Accepted: March 18, 1999
Publication date: March 23, 1999
Susan M. Moore
Swinburne University of Technology, Australia
Victoria University of Technology, Australia
University of Melbourne, Australia
Victoria University of Technology, Australia
The relative contributions of sex and sex role to romantic attitudes were explored through a four-way analysis of co-variance design with sex (male, female), masculinity (high, low), femininity (high, low), and relationship status (in and not in a relationship) as independent variables and age as a covariate. Five romantic attitude scales were the dependent variables. Among 252 men and women aged 17 to over 50 years, there were no main-effect sex differences in romantic attitudes. However, interactions between sex, masculinity, and femininity indicated that stronger romantic attitudes were associated with high femininity for women (regardless of masculinity), and a balance between masculinity and femininity for men (rather than the preponderance of either).
The aim of the current study was to explore sex differences in romantic attitudes/beliefs through investigation of the contribution of sex roles as well as biological sex. The issue of whether there are male/female differences in romantic beliefs, style, or attitudes has already been heavily researched but no unequivocal pattern emerges. For example, Rubin (1970), in his early studies of romantic love, recorded females as tending to be more expressive, focusing on intimacy and concern for their partner, whereas males tended to think about the playful aspects of their relationship. He found that males scored higher than females on romance scales, and suggested that these gender differences reflected the assumption that women acquire the social and economic status of their partners and therefore must be more practical and rational in their mate selection than males. Hong and Bartley (1986) nearly two decades later, also found males to be more romantic than females but interestingly they attributed this to changing sex roles which gave females more flexibility and allowed them to adopt more practical attitudes towards love while men retained traditional attitudes. Conversely, both Stone (1992) and Philbrick (1987) reported females as preferring romantic styles of love relationship. Cimbalo and Novell (1993) answered the question "Who is more romantic, men or women?" by stating that it depends on the dimension under consideration. In their study of college students, women were more likely than men to rate as romantic verbal and behavioral (but non-sexual) expressions of affection, such as receiving flowers. Men on the other hand were more likely to rate as romantic sexual expressions within a relationship. Pederson and Shoemaker (1993), also using a multi-dimensional 'attitudes to romantic love' approach, found no significant gender or marital status differences within their sample of undergraduates. In a study that reflected the state of the literature in this area, Singelis, Choo and Hatfield (1995) examined whether males and females possess different love schemas or styles but, in their words, "We did not really get a clear answer to this question" (Singelis et al, 1995, p. 29).
Singelis et al. contend that theorists have grossly exaggerated existing sex differences and that in the current (liberal) climate differences are getting smaller. Hendrick, Hendrick, Foote and Slapion-Foote (1984) on the other hand argue that studies which show small or non-significant sex differences in romantic love attitudes may be limited by their methodology, in that sex differences become more pronounced when love is considered as a multi-dimensional instead of a uni-dimensional construct. Indeed, differences between studies could relate to several factors, including the age of participants, the type of measuring instrument used, and the conceptualization of romantic attitudes (for example Lee (1973) and Hatfield and Rapson (1995) each use popular, but very different conceptions). Particularly important may be the historical time at which the study was conducted. In relation to this latter variable, changes in male and female norms for behavior are likely to influence variables that relate to interactions between the sexes, such as romantic attitudes. That such norms have changed greatly in the last 20 to 30 years with the relative success of the women's movement points to the possibility that data about sex differences in romantic attitudes collected in the 1970s may no longer be generalizable to today's social climate. Sex role stereotypic views of appropriate behavior may be less common, which may in turn influence the way individuals behave and describe themselves in terms of so-called sex role characteristics.
In the current study, we assessed beliefs about love within a multidimensional, phenomenological framework, using the scale developed by Pedersen and Shoemaker (1993). These researchers explored beliefs about romance through analysis of subject-generated descriptions of romantic behavior and attitudes, a methodology suited for encapsulating current popular beliefs about romance, rather than pre-set researcher conceptions. In addition, in the current study generalizability was enhanced through participation of a wide age range of adults, rather than confining the study to college students, as has often been the case in past research. Age was used as a control variable.
The major innovation of the current study however was a reframing of the question about sex differences in romantic attitudes as a question about the influence of learned patterns of behavior traditionally associated with, but not equivalent to, biological sex -- in other words, the influence of sex roles. Sex or gender roles are characteristics, behaviors and interests defined by a society or culture as appropriate for members of each sex. In Western society, traditionally appropriate sex roles for men have been as worker, breadwinner, head of the household and leader in the community, activities assumed to require so-called masculine personality traits such as assertiveness, confidence, bravery and independence. In classic sex role theory, these instrumental traits were seen as the opposite to stereotypically feminine, expressive traits such as warmth, nurturance, dependency and co-operation, traits more suited to feminine sex-typed behaviors of child rearing, responsibility for family relationships and household duties (Bem, 1974). Bem (1974) argued that masculinity and femininity were independent dimensions, with the possibility of both sets of characteristics co-existing in the one individual. In more sex-typed societies and social groups, such co-existence, as epitomized in the androgynous individual, is more likely than in more traditionally sex typed societies.
Sex role theory implies that positive attitudes toward non-sexual expressions of love (as are typically reflected in 'romance' scales) are more likely to be associated with traditionally feminine (expressive) traits than with traditionally masculine (instrumental) traits. Following this, the major hypothesis of the study was that there would be sex role differences in romantic attitudes. Individuals (not just females) high on stereotypically feminine traits were hypothesized to have stronger romantic attitudes than individuals (not just males) high on stereotypically masculine traits. If sex differences occurred, they would be a function of differential distribution of masculine and feminine sex roles between the sexes. Thus in a sample which was not particularly sex-role stereotyped, sex differences would be unlikely, while in a highly sex-stereotyped sample, sex differences would be more evident. Indeed, the 'de-stereotyping' or 'androgynizing' of male and female roles which is postulated to have occurred in the last 20 years, particularly in the courtship domain (e.g., McCabe & Collins, 1990) suggests that sex differences would not be marked.
The participants comprised a convenience sample of 252 adults, including 84 third year psychology students from a university in Melbourne, Australia and 168 members of the general community recruited by the students. There were 98 males (38.9%) and 154 females (61.1%) in the sample. Age data was collected in categories, with ages ranging from 17 to over 50. The largest age category was the 21 to 30 years group (43%). Participants were spread approximately evenly across the 17 to 20 group, the 31 to 40 group, and the 41 to 50 group (about 18% each). Only 3% were over 50 years. One hundred and fifty-nine (63.1%) individuals in the sample were married, living with a partner, or currently in a romantic relationship; 93 (36.9%) were not currently in a relationship. Most of the sample were born in Australia or another English-speaking western country (N =190; 75.4%), with 32 (12.7%) born in non-English speaking European countries, 16 (6.3%) born in Asian countries and 14 (5.6%) born elsewhere.
A three-part survey was constructed to measure sex roles, romantic attitudes and demographic characteristics of the subjects.
Part A of the survey comprised the Bem Sex Role Inventory (Bem, 1974), which consists of 60 self-descriptive, personality-characteristic adjectives designed to measure psychological masculinity and femininity as two independent variables. All adjectives are designed to be positively toned, or socially desirable. Twenty items assess masculinity, 20 assess femininity; the others are neutral. Each adjective is ranked on a Likert scale where 1 = 'never or almost never true' and 7 = 'always or almost always true'. Scores on each dimension can range from 20 to 140. Individuals were classified on the basis of median splits as 'masculine' (high masculinity, low femininity), 'feminine' (high femininity, low masculinity), 'androgynous' (high masculinity and femininity), or 'undifferentiated' (low masculinity and femininity).
Part B comprised the Romantic Attitudes Rating Scale (RARS) designed by Pederson and Shoemaker (1993) to measure attitudes and beliefs about romantic love. It consists of 50 statements which participants rate on a five-point Likert scale ranging from 1= strongly disagree to 5 = strongly agree. There are five subscales of 10 items each. These subscales were derived by Pederson and Shoemaker via factor analysis of their subjects' responses to a larger pool of items about the nature of romance. Pederson and Shoemaker argue that given that factors were derived from subject responses they are more likely to be ecologically valid and hence truly representative of population attitudes and beliefs about romance. The factors were Togetherness, Expression, Concern and Communication, Romancing and Sensitivity and Spontaneity. Togetherness encompasses the degree of approval of activities carried out only with a romantic companion (e.g., Any activity you do with your lover is romantic). Expressions represents the level of approval of verbal and non-verbal behaviors involving the expression of tenderness and affection (e.g., A sweetheart would take a treat to his or her partner on a stressful day). Concern and Communication relates to valuing the importance of awareness of a romantic partner's problems and being able to communicate about problems and issues (e.g., Lovers should be able to tell each other their problems and concerns and try to help each other resolve them). Romancing represents an idealized attitude towards love relationships, emphasizing feelings of excitement in the presence of the loved one (e.g., Asking a spouse for a date after marriage is romantic). Sensitivity and Spontaneity refers to an attitude of being ready to respond to the needs and desires of the loved one (e.g., Lovers should be expected to participate in spontaneous activities with each other). Scores on each sub-scale can range between 10 and 50, with high scores reflecting high value placed on the named attitude.
Part C of the survey measured demographic variables, including age category, gender, country of birth and relationship status.
The survey was administered to students in social psychology laboratory classes. Each student recruited two additional participants outside of class time as part of the requirement for completing the laboratory program. In order to increase the age range studied, students were requested to recruit at least one participant over the age of thirty years. All participants were informed of their right to withdraw from the study and that all individual data would be anonymous and confidential.
Reliability of Scales
All scales used in the study were checked for reliability (internal consistency) using the Cronbach alpha statistic. Adequate to high reliability was shown for most variables, as follows: Masculinity (0.89); Femininity (0.82); Concern (0.65); Expressions (0.74); Sensitivity (0.71); Togetherness (0.68). For the Romancing variable, the reliability was lower (0.58), so that results using this variable should be treated with caution. Item deletion did not significantly improve the reliability of this scale.
Strength of Romantic Attitudes
Mean scores for the five romantic attitude scales were as follows. Concern/ Communication, M= 40.8, SD= 4.5; Expressions, M= 37.5, SD= 5.2; Sensitivity/Spontaneity, M= 37.3, SD= 5.0; Romancing, M= 36.9, SD= 4.7; Togetherness, M= 33.2, SD= 5.3. On average, scores of 10 to 29 indicate lack of approval of the expressed attitudes, scores of 30 to 50 indicate neutrality (30) through to high approval (50). Thus all attitudes were supported by the sample on average, with Concern/ Communication attracting the strongest approval and Togetherness the least.
Sex Role Typing
When individuals were classified into sex role types (masculine, feminine, androgynous and undifferentiated), there was a significant relationship between sex and sex-role (Chi-square = 33.08; df=3; p< .001). Males and females were similarly distributed in the androgynous category (29% males, 24% females) and the undifferentiated category (21% males, 30% females), but disproportionately distributed in the masculine category (42% males, 8% females) and the feminine category (8% males, 38% females).
Effects of Sex and Sex Role on Romantic Attitudes
Five four-way analyses of covariance were conducted on the romantic attitude dimensions. Factors were sex (male/female), masculinity (high/low, based on a median split), femininity (high/low, based on a median split) and relationship status (currently in a romantic relationship vs. not in a relationship). Age category was used as a covariate, and was significant in the case of Romance (F (1,235) = 3.86, p= .05) and Sensitivity (F(1,235) = 5.28, p < 0.05). In both cases, age was negatively correlated with the subscale with older participants less endorsing of Romancing and Sensitivity items.
There were no significant main effects of sex or masculinity on any of the romance variables. Those in a current relationship showed less positive attitudes toward togetherness (M = 32.36) than those not currently in a relationship (M = 34.43) (F (1,235) = 7.99; p < .005). There was a significant femininity effect for all variables except Togetherness, with high feminine males and females showing more positive attitudes in the areas of Concern, Expressiveness, Romancing and Sensitivity than low feminine males and females (see Table 1).
Table 1: Mean Scores on Romantic Attitude Scales for High and Low Feminine Individuals
|Attitude||Low Femininity||High Femininity||F|
* p < .01; ** p < .05; *** p < .10
The only two-way interaction which was significant was sex by femininity for Concern (F (1,235)= 4.27; p < .05). This will be interpreted in the light of the sex by femininity by masculinity interaction, which was significant for three out of the five romance attitude variables, and close to significance for the other two (Expressions and Togetherness). Table 2 shows the means associated with these interactions. The F values associated with these interactions are: Concern (F (1,235) = 4.27, p < .05); Expressions (F(1,235) = 3.56, p = .06); Romancing (F(1,235) = 4.66, p < .05); Sensitivity (F(1,235) = 4.77, p < .05) and Togetherness (F (1,235) = 3.00, p = .08).
Table 2: Mean Scores on Romantic Attitude Scales for Males and Females, High and Low on Masculinity and Femininity
N = 21
N = 46
N = 41
N = 13
N = 8
N = 58
N = 28
N = 37
The pattern for all variables shown in this table is the same (although less strong for Togetherness). For males, androgynous and undifferentiated individuals (those with a relative balance of masculine and feminine traits) showed the strongest romantic attitudes. For females, those with the highest femininity scores (androgynous and feminine) showed the strongest romantic attitudes.
There was also a significant sex by femininity by relationship status interaction for Togetherness (F(1,235) = 6.63, p = .01). Strong approval of togetherness was associated with high femininity both for females and for males not in a relationship. For males currently in a relationship, high femininity was not associated with strong approval of togetherness. The combination of high femininity and being out of a relationship related most strongly to Togetherness.
Within the limits imposed by the study (including the use of a convenience sample and some measures with low to moderate reliability), mean scores on the romance scales indicated that romantic attitudes were strongly endorsed. Issues of concern and communicating with a loved one were particularly affirmed, closely followed in popularity by behaviors interpreted as demonstrating love, measured by the subscales Expressions, Sensitivity/ Spontaneity, and Romancing. Participating in day-to-day activities together, as reflected in the Togetherness subscale, was judged as less romantic than the other dimensions (especially among those already in relationships), but these activities were nevertheless assessed as more romantic than neutral. The reduction of romantic attitudes with increasing age was a weak trend, occurring minimally (but statistically significantly) for two of the dimensions, Romancing and Sensitivity.
Despite the messages evident in romantic fiction, and indeed the findings of some researchers, we did not find clear-cut sex differences in attitudes to romantic love. As in the Pederson and Shoemaker (1993) study, both sexes endorsed romantic attitudes relatively equally. What our study did show however, was that confusion in past research over sex differences in romantic attitudes may be understandable through recourse to sex role as an explanatory concept. As hypothesized, sex roles were associated with romantic attitudes. Specifically, for the most part individuals with strongly endorsed psychological characteristics classified as traditionally feminine (e.g., warm, loyal, sensitive to the needs of others, understanding) were more romantic than low feminine individuals, regardless of sex. Men and women who scored high on masculinity (traits such as self-reliance, independence, assertion and ease in decision-making), so long as they were also high femininity scorers, were just as romantic as traditionally feminine women. Androgyny (defined as high masculinity and femininity) may indeed provide males with a social identity which allows the expression of romantic attitudes and beliefs, or conversely, those masculine males who are 'in touch' with their romantic side are also enabled to admit to traditionally feminine psychological characteristics. Similarly, high masculinity did not mitigate romantic attitudes in women who had also strongly endorsed feminine traits. These results suggest that in times when sex roles become more flexible with more individuals taking on an androgynous self-image, sex differences in romantic attitudes will be weaker, while among groups and during eras when sex roles are more rigid, sex differences in romantic attitudes may be more evident.
While the data from this study indicated a general association between femininity and romantic attitudes, the relative balance between masculine and feminine traits appeared to be a mediating factor in this association for males only. Thus there were sex differences in the sense of sex roles differentially predicting romantic attitudes for men and women. While there were only very few high feminine, low masculine men in the sample, they were quite low scorers on romantic attitudes. Also, men who were low on both masculinity and femininity (undifferentiated) were more romantic than undifferentiated women. Overall, males with a relative balance of masculine and feminine traits (androgynous and undifferentiated) showed stronger romantic attitudes than men who were sex-typed (masculine) or cross sex-typed (feminine). The most romantic females were, however, the most feminine with respect to their psychological traits. Their level of masculinity was irrelevant to whether they supported romantic attitudes. One possible explanation for these differences may be that psychologically feminine males hold a marginal status in today's society and could feel threatened by admitting to romantic feelings. Alternatively these results could suggest that men's assessment of their psychological femininity is tempered by their assessment of their psychological masculinity, while women view these dimensions more independently. A range of different research techniques, including qualitative measures such as interviews, may be useful at this point to tease out in more detail male and female beliefs about the role of romantic attitudes in their perceptions of their own sex-role stereotypy.
Bem, S. L. (1974). "The Measurement of Psychological Androgyny." Journal of Consulting and Clinical Psychology, 42:155-162.
Cimbalo, R. S. & Novell, D. O. (1993). "Sex Differences in Romantic Love Attitudes among College Students." Psychological Reports, 73:15-18.
Hatfield, E. & Rapson, R. L. (1995). A World of Passion: Cross Cultural Perspectives on Love and Sex. New York, NY: Allyn & Bacon.
Hendrick, C., Hendrick, S., Foote, F. H. & Slapion-Foote, M. J. (1984). "Do Men and Women Love Differently?" Journal of Social and Personal Relationships. 1:177-195.
Hong, S. M. & Bartley C. (1986). "Attitudes toward Romantic Love. An Australian Perspective." Australian Journal of Sex, Marriage and Family. 7:166-170.
Lee, J. (1973). Colors of Love. Toronto: New Press.
McCabe, M. & Collins, J. (1990). Dating, Relating and Sex. Sydney: Horowitz Grahame.
Pedersen, D. & Shoemaker, M. (1993). "Dimensions of Romantic Attitudes." Perceptual and Motor Skills, 77:623-633.
Philbrick, J. L. (1987). "Sex Differences in Romantic Attitudes toward Love among Engineering Students." Psychological Reports, 61:482.
Rubin, Z. (1970). "Measurement of Romantic Love." Journal of Personality and Social Psychology, 16:265-273.
Singelis, T., Choo, P., & Hatfield, E. (1995). "Love Schemas and Romantic Love." Journal of Social Behavior and Personality, 10"15-36.
Stone, C. R. (1992). "Love Attitudes of White South Africans and British University Students." Journal of Social Psychology, 132:609-613.
Susan Moore is a social psychologist whose research interests include adolescent sexuality and risk-taking, attitude research, and beliefs about romance and love. She has a Ph.D. from Florida State University and is a research professor in the Department of Psychology at Swinburne University in Melbourne, Australia. E-mail: email@example.com.
Gerard Kennedy is a lecturer in the Department of Psychology at Victoria University in Melbourne, Australia. His research interests include health and social psychology, in particular the psychology of sleep. E-mail: firstname.lastname@example.org.
Brett Furlonger is a Ph.D. student and part-time lecturer at the University of Melbourne. He formerly lectured in Psychology at Victoria University. His major research interest is in psychology of the deaf. E-mail: email@example.com.
Kaylene Evers is a Ph.D. student at Victoria University in Melbourne, Australia and has lectured there in the Department of Psychology. Her major research interest is health psychology. E-mail: firstname.lastname@example.org.
Back to Journal Cover Page