Annotated Bibliography:
Attraction, Selection, and Satisfaction --
Understanding what makes happy couples
Duffy, S.M., & C.E. Rusbult. (1986). Satisfaction and Commitment in Homosexual and Heterosexual Relationships. Journal of Homosexuality 12 (2).
In this study, the authors examined differences in relationship investment and satisfaction in heterosexual and homosexual relationships. Their sample consisted of four groups of 25 people each: gay men, straight men, lesbian women, and straight women. All participants were relatively young, between 20 and 40 years old, and demographically equivalent in terms of income, educational level, and race (all Caucasian). (The equivalence is important to the methodology, but the lack of diversity in age and race makes the findings a bit hard to generalize.) Each participant was asked to recall one relationship, either ongoing or in the past, and respond to both general and specific prompts about the costs, rewards, and investments in the relationship and the availability and attractiveness of alternatives to the partner in question. In general, most participants were satisfied with their relationships and committed to them; high rewards and low costs predicted satisfaction. Perhaps the most interesting broad finding was that sex differences had much more powerful effects than sexual orientation differences. The authors speculate that this is true because people are socialized first to be men or women and later to like men or women -- that is, sexual orientation becomes salient later in development than gender. Noteworthy sex differences: Men's satisfaction was less sensitive than women's to the costs of the relationship. Men of any sexual orientation were thought their partners were more likely to stray outside the relationship than did women, and men also felt they had given up more freedom to their relationships. More women than men reported that their relationships were monogamous. Noteworthy sexual orientation differences: The reward value of the relationship for gay men did not correlate with commitment; they also reported liking casual dating more than any other group. (Lesbian women liked it least of all.) Similarly, gay men reported the least monogamy; lesbians reported the most. (Comment: the distribution of number of lifetime partners that I have seen in other studies shows monotonic increase from lesbian women to straight women to straight men to gay men. It appears to be a sex difference, with heterosexual men and women meeting in the middle of women's preference for fewer partners and men's preference for more.)
Doherty, R.W., Hatfield, E., Thompson, K., & P. Choo. (1994). "Cultural and ethnic influences on love and attachment." Personal Relationships 1: 391-398.
This study examined the relationship between attachment and ethnicity. The authors recruited 124 men and 184 women from the University of Hawaii (mean age 23.7 years). All but 6 percent were in a relationship of some sort; 20 percent were married. Each subject belonged to one of four ethnic groups: European-American, Japanese-American, Pacific Islander, or Chinese-American. All spoke English. As expected, these groups differed in individualism-collectivism, with European-Americans most independent, Chinese-Americans most collectivist, and Japanese-Americans and Pacific Islanders in the middle. The authors found that ethnicity did not predict the degree of passionate or companionate love reported by any of the groups. Individualism had a small negative correlation with passionate love, but individualism-collectivism had no other relationship with type of love. Rather, adult attachment style related to the kinds of love reported. Anxious-avoidant people reported greater passionate love, while secure people reported greater companionate love; these findings held across all four ethnic groups.
Goodwin, R., & C. Findlay. (1997). "We were just fated together" ... Chinese love and the concept of yuan in England and Hong Kong. Personal Relationships 4: 85-92.
This study explores the Chinese notion of yuan, or destined love ("relational fatalism"). In Chinese culture, it serves a useful function by justifying relationships socially, and by smoothing over conflict, which was also destined to happen and thus is not the couple's fault. In general, though, it resembles Western notions of fated love, so the authors sought to compare Chinese and English students' beliefs about the concept. They recruited 100 students in Hong Kong and 100 in Bristol, half female in both cases, and gave them an assessment of relational fatalism. Chinese scored significantly above British students; in particular, they more often stated that relationships form outside of our control and that personality and attractiveness don't affect them. Furthermore, the Chinese students valued morality and conformity to family and societal norms more than the British. Gender did not predict scores on the yuan scale. Yuan was positively associated with pragmatic and agapic love styles and negatively associated with erotic love; as we would expect, Chinese scored higher than the British on pragmatic and agapic love but lower on eros.
Comment: As the authors note, the relatively high score on yuan for British students couple be a methodological artifact (tending toward the center of any scale might produce an inflated score), but its congruence with the popular romantic notion of one true or perfect love is striking. In the West, though, instead of this notion comforting those in unhappy times in their relationship, it seems to serve more to induce despair when the true love can't be found, or the perfect love proves flawed.
Gottman, J.M., & L.J. Krokoff. (1989). Marital Interaction and Satisfaction: A Longitudinal View. Journal of Consulting and Clinical Psychology 57 (1): 47-52.
The authors coded marital conflict behavior in the laboratory and at home and used it to predict both present and future marital satisfaction in two samples of couples. The first sample comprised couples from Champaign-Urbana, Illinois, with mean age in the mid-40s and mean duration of marriage 23.9 years. The second sample consisted of 30 couples from Bloomington, Indiana, with mean age around 30 and mean years married 4.2 years; this sample scored slightly higher in marital satisfaction. Sample 1 couples discussed a contentious issue in their relationship for 15 minutes in the lab; Sample 2 couples picked two issues and were videotaped discussing one in the lab and audiotaped discussing the other at home to alleviate the effect, if any, of a co-present observer.
Different variables predicted current and future satisfaction. Negativity by the husband was associated with lower present satisfaction but greater long-term satisfaction. Similarly, negativity by the wife was associated with lower current satisfaction, but it did not affect long-term satisfaction. Also, compliance and positive verbal behavior by the wife predicted lower long-term satisfaction, whereas compliance had no relationship with present satisfaction and positive verbal behavior was actually associated with greater present satisfaction. Engagement in conflict or acts of disagreement clearly predicted lower short-term satisfaction, but it also was associated almost as strongly with better long-term satisfaction. The wife being stubborn or withdrawn predicted lower satisfaction in the present but did not relate to satisfaction in the future. The same pattern held for the wife showing anger or the husband or wife showing contempt. For the wife's sadness, the pattern was reversed: it had no relation with present satisfaction but was associated with reduced satisfaction in the future. The husband's whining predicted his future dissatisfaction, but not hers (not at p < .05, though the trend was in that direction).
The overarching lesson of this study is that some conflict behaviors might help in the short term but hurt in the long term -- it's convincing, and it's also interesting because the results dramatically contradict cultural notions of what makes a good relationship (e.g., low conflict).
Gottman, J.M., & R.W. Levenson. (1992). Marital Processes Predictive of Later Dissolution: Behavior, Physiology, and Health. Journal of Personality and Social Psychology 63 (2): 221-233.
This study examined a "cascade" process model of dissolution, in which separation and divorce followed from a sequence of events, common at first, but becoming less and less common as the cascade continues -- having some dissatisfaction is more common than considering separation, which is more common than actually separating, which is more common than divorce. The authors had 79 couples (mean age around 30, mean married time 5.2 years) talk in the lab about a problem area in their relationship while connected to physiological measuring devices; they also took an inventory of standard martial issues. Later, each spouse viewed the video of their interaction and used a dial to provide continuous feedback about their feelings during the discussion. Finally, four years later, the authors followed up with the couples about their marital satisfaction, health, and consideration of divorce.
In a neat geometric progression, half the couples considered divorce over the four-year period, half that number separated, and half that number divorced. Structural equation modeling suggests that the hypothesized cascade effect was plausible, but its results can't strictly prove causation. The authors also distinguish between "regulated" and "non-regulated" couples -- the non-regulated pairs fought more, were less joyful, angrier, more defensive and stubborn, whinier, more withdrawn, and paid less attention to their partners. The regulated couples felt more positive as the discussion of their problem area progressed, while the non-regulated couples felt increasingly negative. Non-regulated couples reported lower marital satisfaction scores at both the outset of the study and the four-year checkup; they tended to move farther along the hypothesized path to divorce. The most successful model for differentiating regulated and non-regulated couples was a blend of the huband's and the wife's ratios of positive to negative speech acts. Somewhat surprisingly, physiological signals gathered during the laboratory problem discussion did not provide much distinction between regulated and non-regulated couples.
Carstensen, L.L., Gottman, J.M., & R.W. Levenson. (1995). Emotional Behavior in Long-Term Marriage. Psychology and Aging 10 (1): 140-149.
The authors studied affect and reciprocity in middle-aged couples who had been married at least 15 years and older couples who had been married at least 35 years; they selected a balance of happy and unhappy couples. Recruitment gave them four groups: satisfied middle-aged (35 couples), dissatisfied middle-aged (47), satisfied older (43), and dissatisfied older (31). Of note, because the researchers generally sought first marriages, age is confounded with length of marriage. Couples were videotaped in the lab discussing a neutral topic, a point of contention in their relationship, and a pleasant topic; each spouse subsequently and separately provided ratings of his or her feelings during the discussion. Older couples showed markedly less negative affect, even during the discussion of a relationship problem, than the middle-aged ones, and when they did display it, they interwove affectional expressions to mitigate it. Older couples were also less likely to introduce negative affect in the first place ("avoiding escalation"); the authors suggest that this fits with other findings that people learn to better regulate their emotions as they age. Those in the dissatisfied couples, predictably enough, showed more frequent instances of negative emotions (anger, contempt, sadness, being domineering), while satisfied couples exhibited more positive expressions (happiness, affection, validation). The satisfied couples also proved better at avoiding the escalation of negative affect. In terms of gender differences, women were generally more emotional, and in unhappy couples tended to be higher than their husbands on overall emotion, negative emotion, and contempt. Men tended to be more defensive and, in unhappy couples, to try to "de-escalate" rising affect.
Carrère, S., Buehlman, K.T., Gottman, J.M., Coan, J.A., & L. Ruckstuhl. (2000). Predicting Marital Stability and Divorce in Newlywed Couples. Journal of Family Psychology 14 (1): 42-58.
The authors performed a longitudinal study of 95 newly married couples (less than 6 months of marriage) in Seattle; the sample was racially and ethnically representative of the metropolitan area. They began with an "Oral History Interview," from which they coded how the couple "tells the story of their relationship" -- some idealize it, some talk about their struggles -- and how the spouses describe each other. From the coding of the interview, a principal component analysis revealed a construct that the authors call "perceived marital bond," which consists of variables including both the husband and the wife's fondness, expansiveness, "we-ness," negativity, and disappointment, as well as couple-level chaos, volatility, glorification, and (much more weakly) gender stereotypy. This marital bond construct was related to but distinct from marital satisfaction.
The initial interviews were conducted over a three-year period. Every year for the next five years, each member of the couples was interviewed by telephone. Using the 11 components of the perceived marital bond factor that loaded at or about .70 (all but volatility, glorification, and gender stereotypy), the authors constructed a discriminant function that predicted which couples would divorce with quite high accuracy: 87 percent after four to six years and 81 percent after seven to nine years (recall that the couples started the study at various times over a three-year period). This rather good predictive accuracy, the authors suggest, emerges because the metric combines the relationship-relevant traits of both partners with coded information about their interactions at the interview and the narrative (as opposed to descriptive) nature of the relational history. It is unclear how well this construct would predict marital satisfaction over time, which could be represented as a continuous rather than binary variable, like divorce, but this would be a good future test of the perceived marital bond factor.
Klohnen, E.C., & G.A. Mendelsohn. (1998). Partner Selection for Personality Characteristics: A Couple-Centered Approach. Personality and Social Psychology Bulletin 24 (3): 268-278.
Using an interesting methodological approach to account for non-zero baseline personality similarity, the authors examined whether 36 college-age heterosexual couples exhibited personality similarity or complementarity. All couples had been together less than five years. The subjects completed Block's California Q-Set for their actual selves, their ideal selves, and for their partners; because of baseline personality similarity, the average Q-correlation for randomly paired individuals from the sample was r = .27. Of course actual couples, like every other pair, were correlated well above zero as well, but to establish whether couples truly exhibited more similarity than random pairs, the authors ranked the correlations between a given subject and all other opposite sex subjects, and considered where the subject's actual partner fell in the rankings. If the actual partner ranked significantly above the midpoint (18.5 for a pool of 36), they could conclude that the couple exhibited better-than-chance similarity.
For overall similarity, this did not hold; couples were no more likely than a random pair to be similar. But it turned out that subjects with higher congruence of actual self to ideal self were more similar to their partners; furthermore, those high on actual-ideal congruence tended to have partners who were also high, which suggests assortative mating on this self-esteem-like metric. More significantly, it was the perception of similarity that mattered more than actual similarity: subjects' descriptions (perceptions) of their partners were much more similar to the subjects' own ideal-self descriptions than chance would predict. Subjects' perceptions of their partners seemed reasonably good, but it turned out that their descriptions of their partners bore no more than chance similarity to the partners' actual self-descriptions. So in great part it is ideal-driven perceptual processes, not reality, that influence how similar people believe their partners to be.
This paper was methodologically creative for two reasons: the focus on pairwise ("couple-centric") Q-correlations rather than trait-wise correlations over the whole sample, and the use of rank tests to account for the high baseline correlations between even randomly paired individuals. It strongly suggests that we should consider carefully assumptions that the baseline is zero or the curve is normal for studies involving matching or pairing.
Klohnen, E.C., & S. Luo. (2003). Interpersonal Attraction and Personality: What Is Attractive -- Self Similarity, Ideal Similarity, Complementarity, or Attachment Security? Journal of Personality and Social Psychology 85 (4): 709-722.
To investigate how attachment style influences attraction, the authors assessed adult the attachment styles and procured self- and ideal-self-descriptions of more than 700 undergraduates, approximately half female. They also gathered ratings of the attractiveness of hypothetical partners from the subjects. Some received scenarios in which the hypothetical partner behaved in a way consistent with a particular attachment style; other subjects received experimenter-fabricated self-descriptions by purported previous participants that exemplified particular attachment styles.
On the whole, most subjects found secure (hypothetical) individuals most attractive. Beyond that, subjects found the descriptions attractive that matched their own attachment styles. Some mismatched combinations were particularly aversive, including avoidant subjects evaluating secure and preoccupied profiles and dismissing subjects evaluating preoccupied profiles, even though on average preoccupied profiles were the second-most attractive. Perception also played a major role: subjects perceived most of the profiles as more secure, more self-similar, and more ideal-self-similar than they really were. (In many cases, the ideal self was a close approximation of a secure attachment style.) In fact, these perceptions did more to determine attraction than did the actual characteristics of the profile.
Watson, D., Klohnen, E.C., Casillas, A., Simms, E.N., Haig, J., & D.S. Berry. (2004). Match Makers and Deal Breakers: Analyses of Assortative Mating in Newlywed Couples. Journal of Personality 72 (5): 1029-1068.
The authors examined assortative mating across a wide variety of dimensions among 291 newly married couples from the Iowa Marital Assessment Project. All couples had been married less than a year, and both members of each couple were less than 50 years old. Participants provided demographic information and completed a large battery of assessments of personality, affect, temperament, attachment, intelligence, and relationship satisfaction. Analysis showed that the strongest positive assortment occurred on age (r = .77), followed by religiosity (r = .75) and political conservatism (r = .63). Assortment correlations for "ability" variables, including vocabulary, education, and IQ, were around r = .45; correlations were lower for non-verbal abilities. The authors also found assortment on age and educational level, so they employed partial correlation to control for these factors; the results indicated that the findings are not merely a result of social homogamy. Furthermore, they found that couples generally do not converge on these dimensions over time, which supports the hypothesis of initial assortment. (In fact, the couples tended to diverge slightly over time on Positive Emotions and Openness.) Finally, the authors found that similarity to one's spouse predicts little about relationship satisfaction. Much stronger predictors of satisfaction were a participant's own scores on BFI personality scales, disinhibition, ego resiliency, positive and negative affect, and adult attachment style.
The paper highlights the importance of active assortment -- deliberate choice of a spouse with certain characteristics. The effect of active assortment varies depending on how consensual the characteristic is; as the authors note, certain attributes, like intelligence and agreeableness, are almost universally desired, so their assortative effects seem weaker than something like political views, for which there is no universally preferred value. Political views and religiosity exemplify characteristics with little consensus as to what is preferred, and as such, they show the highest variability in rankings of attribute importance.
Robins, R.W., Caspi, A., & T.E. Moffitt. (2002). It's Not Just Who You're With, It's Who You Are: Personality and Relationship Experiences Across Multiple Relationships. Journal of Personality 70 (6): 926-964.
This well-conceived study followed 712 people (384 women) from a birth cohort in New Zealand, giving them personality assessments at 18 and 26 and relationship assessments at 21 and 26. The personality instrument was Tellegen's Multidimensional Personality Questionnaire, which comprises three main factors: negative emotionality, positive emotionality, and constraint. The relationship assessments consisted of 50-minute interviews that evaluated relationship quality, conflict, and abuse. On the whole, the cohort showed increases in relationship quality and decreases in conflict and abuse from 21 to 26; they also showed increases in positive emotionality and constraint and decreases in negative emotionality from 18 to 26. Negative emotionality was associated with lower levels of relationship quality and greater conflict and abuse; positive emotionality and constraint were associated with the inverse.
Because of the staggered longitudinal design, the authors can examine antecedent personality and relationship outcomes. Negative emotionality at 18 was associated with a decrease in relationship quality and an increase in conflict and abuse from 21 to 26. Positive emotionality and constraint were associated with an increase in relationship quality during the same period. These changes were noted whether or not the subjects were with the same partner at 21 and 26; it appears that personality might help to determine relationship experiences. In the other direction, low relationship quality and high conflict and abuse at age 21 were associated with an increase in negative emotionality from 18 to 26; the magnitude of this was not large, but it suggests that relationship experiences might help shape personality over time.
Kaplan, D.L., & C.B. Keys. (1997). Sex and relationship variables as predictors of sexual attraction in cross-sex platonic friendships between young heterosexual adults. Journal of Social and Personal Relationships 14 (2): 191-206.
The authors gave a questionnaire to evaluate cross-sex friendship to a racially diverse group of 186 undergraduates (103 female), most between 18 and 20 years old. The questionnaire asked subjects about three relationships: with their best cross-sex friend, their best same-sex friend, and their romantic partner. Men and women were similarly attracted to their romantic partners, but men were generally more attracted to their cross-sex platonic friends. More than half of the women reported no attraction to their cross-sex friend, as compared to only a quarter of men. Fifty-seven percent of men reported moderate or high attraction, versus 32 percent of women. This difference lies in the "moderate" category, however; about 10 percent of both men and women reported high attraction. In a regression model to predict sexual attraction, the factor that explained the most variance was not biological sex but "loving," defined in this study as emotional bonding. It is mediated by sex, though -- women's sexual attraction to their male friend was more strongly associated with loving than was men's sexual attraction to their female friend. Interestingly, single men were less attracted to their cross-sex friends than were men in relationships. The authors suggest that this might happen because feeling some attraction for a friend is "safe" when one is not available to date her; I would further speculate that the man might see his friend as an alternative to his girlfriend (in the sense of the investment model), but without first-hand knowledge of the negatives of being in a relationship with the friend, thus inflating her perceived attractiveness. Men were also more attracted to their cross-sex friends in short-term friendships; although the present study was cross-sectional, not longitudinal, this suggests that men's attraction might diminish as platonic friendship grows. This time finding did not hold for women, whose attraction did not relate to relationship duration.
Bailey, J.M., Gaulin, S., Agyei, Y., & B.A. Gladue. (1994). Effects of Gender and Sexual Orientation on Evolutionarily Recent Aspects of Human Mating Psychology. Journal of Personality and Social Psychology 66 (6): 1081-1093.
This study examined differences in reproductive psychology among heterosexual and homosexual men and women. Subjects (between 65 and 72 in each of the four groups) were assessed on several scales: uncommitted sex, visual stimuli, partner's status, jealousy, youth of partner, partner's physical attractiveness, and sociosexuality. As Duffy and Rusbult found (see above), sex predicted far greater differences than did sexual orientation. Homosexual and heterosexual men were similar in their interest in uncommitted sex, visual sexual stimuli, unconcern about their partner's status, and the importance of their partner's physical appearance. Heterosexual and homosexual women were similar in their interest in uncommitted sex and their sociosexuality. Homosexual women indicated greater interest than heterosexual women in visual sexual stimuli; they were also less concerned than heterosexuals with the relationship status with their partner. Both of these differences place homosexual women in between men and heterosexual women on these scales. The authors report a similar finding for homosexual men, who were less sexually jealous than heterosexual men (scoring in between that group and the groups of women). Gay men were also less likely to prefer younger partners than were heterosexual men. Again, they fell in between straight men and women on this dimension. It seems that some aspects of romantic and sexual behavior depend as much on what sex you like as on what sex you are. One potential difficulty with this study, the authors note, is that subjects were recruited through ads in alternative weekly newspapers, so in addition to a self-selection bias, the demographics of the publications might skew toward people with liberal attitudes, which might extend to sex.
Baumeister, R.F., & E. Bratslavsky. (1999). Passion, intimacy, and time: Passionate love as a function of change in intimacy. Personality and Social Psychology Review 3 (1): 49-67.
The authors examine how passion might depend on changing levels of intimacy over time; for this argument, intimacy is defined in terms of three dimensions -- mutual disclosure, feeling favorable toward the person, and communicating affection. Passion in this case refers to "strong feelings of attraction" for the person, arousal, and sometimes sexual attraction. The authors propose that passion changes with the first derivative of intimacy -- the change in intimacy over time. They begin by noting that passion and intimacy operate on very different time-scales, the former quite short (as it is tied to emotions, which we know to be brief) and the latter much longer. In terms of attraction as a proxy for passion, the authors find in the literature that people are more attracted to those who give fluctuating positive and negative evaluations than to those who are consistently positive -- is it the change in reaction that increases attraction-as-passion? Intercourse, one indicator of passion, is frequent at the start of a relationship, when intimacy is rising rapidly. Intercourse declines over the course of a relationship; they argue that this corresponds with the plateau of intimacy at a more stable level. In terms of personality, the authors note that extraverts perform more intimacy-promoting behaviors, including more frequent eye contact, verbal utterances, and disclosures of information. In keeping with their hypothesis, extraverts also have more sex with more partners, and they start having sex earlier. In established intimate relationships, novel experiences seem to temporarily increase passion (though parenthood seems to be a major exception to this). In theory, a fight or argument should promote passion (with "make-up sex"), but perhaps at the expense of intimacy in the long term.
Comment: This concept is intuitively appealing, but of course hard to prove. Not only is a clean functional relationship unlikely, but the problem of operationalizing things like passion and intimacy so that the relationship would appear may be impossible with sufficient rigor. The evidence here is circumstantial, but its purpose is to lay the groundwork for the theory, not to prove it.
Felmlee, D.H. (1998). "Be careful what you wish for...": A quantitative and qualitative investigation of "fatal attractions." Personal Relationships 5: 235-253.
The author investigates "fatal attractions," in which a partner's initially desirable characteristic turns out to be disliked and, perhaps, leads to relationship dissolution. She focuses on various kinds of different characteristics in the partners: different from self, different from average, and different from norms, such as those based on gender. The sample consisted of 301 undergraduates (two-thirds female), almost all heterosexual, who described their terminated relationships (average length 14.5 months). The subjects completed a questionnaire with closed- and open-ended questions about their relationships; their responses were coded according to the qualities that led to initial attraction and those that led to dissolution. About 30 percent of subjects were determined to have experienced a fatal attraction. For example, one man reported being attracted to a woman because she was the "very opposite of myself," but then indicated that they broke up because she was "too different." Logistic regression suggested that characteristics different from average more so than different from self were potential bases for fatal attraction, especially if they were described as "striking, unique, or extreme." Deviations from gender-stereotypic norms were also more likely to lead to fatal attraction. Broadly, dissimilarity seemed problematic when it plays a part in initial attraction.
Comment: It seems that what is noteworthy or noticeable about a partner is what both attracts us and later repels us. But perhaps this is a matter of cognitive process -- we don't really know what it is that attracts and repels us, but when pressed for an explanation, the most readily available (noteworthy, unique, different) aspects of the partner come to mind.
Shibazaki, K., & K.A. Brennan. (1998). When birds of different feathers flock together: A preliminary comparison of intra-ethnic and inter-ethnic dating relationships. Journal of Social and Personal Relationships 15 (2): 248-256.
The authors studied inter-ethnic and same-ethnic relationships among 100 undergraduates (74 female, median age 19) in Texas. The study involved only one member from each couple; 44 of the participants were in an inter-ethnic relationship, while the remaining 56 were in a same-ethnic relationship (median length 13 months across the whole sample). Most subjects were Caucasian or Hispanic, along with some Asian, African-American, Indian, and Arabic students. The subjects completed questionnaires with open-ended prompts about the initiation of their relationships, which the authors coded for mention of appearance, personality, and similarity of beliefs and interests. They were also assessed on relationship satisfaction, self-esteem, and the degree of approval they had received from parents, friends, and the public.
The inter-ethnic and same-ethnic groups did not differ in relationship satisfaction, but the inter-ethnic group was very slightly lower in self-esteem (significantly so, but needs replication due to small magnitude). Inter-ethnic dating did not predict lower (or higher) perceived approval by parents or friends, but it did predict lower perceived approval by the general public. The two groups were equally concerned about the approval of others. Relationship satisfaction was associated with the degree of approval from parents and friends for both groups, but from the general public only for those in inter-ethnic relationships.
Those dating within their ethnicity identified more strongly with their own ethnic group, and they perceived less availability of partners of other ethnicities than did those who were dating inter-ethnically. Furthermore, those dating across ethnic boundaries were more willing to accept a roommate, romantic partner, spouse, and "parent of my children" than were those dating someone of their own ethnicity.
Caspi, A., & E.S. Herbener. (1990). Continuity and Change: Assortative Marriage and the Consistency of Personality in Adulthood. Journal of Personality and Social Psychology 58 (2): 250-258.
The authors present a study of the interaction of assortative marriage with adult personality consistency; they adopt a methodological approach like Klohnen & Mendelsohn (see above, though this paper came before K&M) in that they use the California Q-set and focus on intra-couple similarity, not sample-wide similarity between all husbands and all wives. They administered the Q-sort to about 125 couples from Berkeley and Oakland at the start of the study and again 11 years later. Subjects' Time 1 and Time 2 results Q-correlated at a mean around .51, which was shown to be greater than the Q-correlation of randomly selected individuals, indicating temporal consistency of personality. Spouse-similarity Q-correlations from Time 1 averaged .32, which also appeared to be greater than chance (i.e., greater than the Q-correlations of random opposite-sex pairs). More interestingly, spousal personality similarity predicted intra-individual personality stability over time. The authors point out that this could be due to other kinds of homogeneity by which people assortatively marry, such as age and educational level, but a follow-up study with a separate sample indicated that this was not the case. It seems likely that the personality stability really was encouraged by spousal similarity. The authors suggest that perhaps spouses similar in personality share common histories of behavioral reinforcement, which are little changed when they live with their similar spouse. Alternatively, marital satisfaction resulting from personality similarity might influence the partners not to try to change each other's behavior, as they are already happy, thus permitting stability. As evidence for this hypothesis, the authors found significant correlations between marital satisfaction and personality stability within the individual spouses.
Kelly, E.L., & J.J. Conley. (1987). Personality and Compatibility: A Prospective Analysis of Marital Stability and Marital Satisfaction. Journal of Personality and Social Psychology 52 (1): 27-40.
In this study, 300 couples engaged to be married were tracked from the mid-1930s until 1980. Most were between 20 and 30 years old at the start of the study. All were white; most lived in or around Connecticut. Data were collected in the mid-1930s, the mid-1950s, and the early 1980s. Additionally, for the first few years of the study, the couples completed annual surveys and interviews, and they submitted names of acquaintances who contributed personality ratings of the couples. Over the course of the study, 22 couples broke up before marriage, and 50 divorced sometime before 1980. Early divorce (before 1955) was associated with high neuroticism, less puritanical attitudes, and more premarital sex. For men, it was additionally associated with low impulse control. For women, unstable family history was also a predictor. Men who divorced later in life (after 1955) were also higher on neuroticism than still-married men, and they were more extraverted and less agreeable. Women who divorced later were very much like still-married women except that they had less puritanical sexual attitudes and histories. Parental divorce was not significantly related to divorce for this sample.
High neuroticism was also related to lower marital satisfaction even among still-married couples. Attitudes did not interact much with satisfaction, except that family role traditionalism was associated with men's satisfaction. Dissatisfied still-married couples were distinguished from divorced couples by the man's low extraversion and agreeableness. The authors suggest that neuroticism is the underlying cause of marital disturbance, but that other factors determine how it affects marital outcomes.
Hendrick, C., & S.S. Hendrick. (1989). Research on love: Does it measure up? Journal of Personality and Social Psychology 56 (5): 784-794.
The authors factor-analyze five instruments for assessing love to discover common dimensions among them: Hazan and Shaver's General Relationship Attitudes, which deals with adult attachment styles; the authors' own Love Attitudes Scale, which has the "love styles" as its components (eros -- erotic, ludus -- playful , storge -- friendship, pragma -- practical, mania -- needful, agape -- selfless); Sternberg's Triangular Theory of Love (intimacy, passion, commitment); Hatfield and Sprecher's Passionate Love Scale (passion, seems to be some combination of eros and mania); and Davis and Todd's Relationship Rating Form (viability, intimacy, passion, care, satisfaction, and conflict). They had 391 undergraduates (202 women), generally white and well-off, complete all these scales. Many of the dimensions were found to be redundant; in particular, the Triangular Theory's components intercorrelated at an average of .75. A combined factor analysis of all the scales yielded five factors. The first is heavily loaded on passion, erotic love, mania, intimacy, and satisfaction. The second appears to be companionate love, loaded on intimacy, care, viability, and lack of conflict and game-playing. The remaining three factors had fewer components and so were more specific. The third loaded on anxious-ambivalent attachment and mania. The fourth was heavily secure attachment, with a negative loading on avoidant attachment. Finally, the fifth factor combined storge and pragma, which appears to be practical, friendship-based love.
Comment: Although I like the clean factor-analysis approach to this work, the authors themselves note that for a topic like this, the use of college-age subjects alone limits the ability to generalize. We can speculate that youth and romantic inexperience might lead them to over-emphasize the passionate or erotic aspects of love and give short shrift to the companionate components, but we can't know for sure without a broader sample.
Ingersoll-Dayton, B., Campbell, R., Kurokawa, Y., & M. Saito. (1996). Separateness and togetherness: Interdependence over the life course in Japanese and American Marriages. Journal of Social and Personal Relationships 13 (3): 385-398.
The authors analyzed narratives about the lives and marriages of 24 older couples, 11 Japanese (mean age 68) and 13 American (mean age 75; three couples African-American, the rest white). Both samples included a mix of blue-collar and white-collar jobs. One major dimension on which the couples differed across cultures was "separateness and togetherness." Most of the Japanese couples' marriages had been arranged, and as such they sometimes reported initial difficulties in living together, in many cases in an extended family home. The early part of these marriages was heavily role-based -- women having their "sphere of influence" in the home and men in the workplace. By contrast, the Americans spoke of convergence, joining two lives, even though gender role traditionalism (and the attendant separation of duties) also existed in their marriages in practice. Americans were more likely to speak in the second-person plural ("we"); they tended to consider interdependence, as opposed to dependence. Additionally, the rules (implicit or explicit) governing the marriages differed. White American couples focused on equity and fairness, "doing your share," though some felt threatened by role encroachment (e.g., retired husbands trying to cook); this focus on equity was less prominent for African-American couples, though the sample was only three in this study. The Japanese, on the other hand, talked about complementary dependence, how their roles were mutually exclusive, though some overlap occurred when one spouse grew ill or weak and the other took over his or her tasks. In terms of intimacy, especially among the Japanese arranged marriages, affection was sometimes lacking early in the partnership. Some considered divorce early on, but many reported growing intimacy as their marriages progressed. (Other studies I have seen showed that arranged marriages peak in intimacy much later than love marriages and sometimes maintain a higher plateau, whereas love marriages peak early and can fall off dramatically with time.) For Japanese couples, major life changes such as the death of a parent could lead to increased intimacy with one's spouse. American love couples were more deliberate about negotiating time and space boundaries. For them, intimacy sometimes increased dramatically after retirement. Across both cultures, spouses seemed quite willing to adapt to relationship changes when a spouse became ill; this seemed to have the greatest effect on Japanese husbands, who were thrust into a nurturer's role if their wives grew frail.