Introduction
Adult attachment theory appears to be a valuable framework to understand romantic relationship functioning (for a review, see Mikulincer & Shaver, 2016). A model of four adult attachment styles has been conceptualized in terms of positive versus negative internal working models, that is, mental expectations about self-worth –model of self– and the supportive availability of others in attachment relationships –model of others– (Bartholomew & Horowitz, 1991). These representations are early developed during childhood with respect to primary caregivers, but affect how individuals interpret and behave in later close relationships. Likewise, adult attachment styles are conceived as emerging from two underlying dimensions: anxiety and avoidance (Fraley et al., 2000). The former represents the degree to which individuals are concerned about being abandoned or rejected by their partner, while the latter reflects the degree to which they feel comfortable with emotional intimacy and closeness. Individuals labeled as secure hold a positive view of the self (low anxiety about abandonment) and a positive view of others (low avoidance of intimacy); preoccupied individuals possess a negative perception of self (high anxiety) and a positive perception of others (low avoidance); fearful-avoidant individuals adopt a negative view both of self and of others (high anxiety and avoidance); and dismissing-avoidant individuals have a positive sense of self (low anxiety) and a negative perception of others (high avoidance).
Seemingly, two different mechanisms may explain the connections between adult attachment style and romantic relationship functioning (Collins et al., 2002): relationship skills (e.g., partners’ ability to regulate their emotions) and mate selection (e.g., the tendency to partner with secure versus insecure individuals). This article focuses its attention on the latter.
There are mainly three hypotheses of attachment-related partner preference and selection, as reviewed by Holmes and Johnson (2009): the similarity hypothesis (preference for partners with the same attachment style as the self), the complementary hypothesis (preference for partners with the opposite style), and the security hypothesis (preference for secure over insecure partners). In line with these three predictions, studies suggest that secure individuals prefer similarly secure partners. Empirical evidence, however, is inconsistent with regard to insecure individuals’ preferences, maybe partly due to the differences in attachment classifications used (i.e. dimensional vs. categorical measures; three vs. four styles) (Holmes & Johnson, 2009).
Theoretically, internal working models of adult attachment may have indirect effects on romantic relationship functioning by leading individuals to select partners who are more or less able to provide emotional warmth and supportiveness, and more or less vulnerable to relationship difficulties (Collins et al., 2002). For instance, adverse outcomes –such as low satisfaction with the relationship– may be partly explained by the selection of partners whose personality predisposes them to dysfunctional styles of relating or behavioral patterns that are detrimental to relationship functioning.
Indeed, as Pietromonaco et al. (2004) highlight, partners’ internal working models seem to shape their perceptions of threat and their goals during couple conflict. Preoccupied individuals would tend to view conflict as a threat because they are afraid of being abandoned or not having their partner available if needed, which fosters intense emotional reactions that hinder constructive communication. Dismissing individuals would tend to perceive conflict as a threat because they feel pressured to self-disclose and engage in behaviors that foster emotional intimacy, which leads them to withdraw or minimize the relevance of conflicts. Fearful-avoidant individuals would display a mixture of behaviors linked with dismissing and preoccupied styles. Finally, securely attached individuals would be more likely to use constructive tactics because they view their partners as available, responsive, and loving and they believe that conflicts are an opportunity to adjust to the partner´s needs (Pietromonaco et al., 2004).
Studies examining the joint effects of both partners’ adult attachment styles on relationship functioning have generally considered three configurations or combinations (Ben-Ari & Lavee, 2005): secure couples —two securely attached partners—, insecure couples —two insecurely attached partners, either preoccupied, dismissing, or fearful—, and mixed couples —one partner securely attached and other insecurely attached. In general, secure couples report higher levels of relationship satisfaction (Guzmán & Contreras, 2012; Kuncewicz & Kuncewicz, 2019; Rivera et al., 2011; Senchak & Leonard, 1992) and healthier conflict communication (Domingue & Mollen, 2009) than the other two couple types.
But for an insecurely attached individual, it is more beneficial to be in a relationship with a similarly-attached partner or with a securely-attached one? According to some data, mixed and insecure couple groups do not significantly differ in the use of mutual constructive communication (Domingue & Mollen, 2009), or in conflict management or relationship satisfaction (Rivera, 2006). Conversely, however, Myers et al. (2016) observed significantly lower levels of conflict engagement, withdrawal, and compliance and higher levels of constructive problem solving and relationship satisfaction in mixed couples than in insecure ones. Moreover, consistently with the above-mentioned evidence, these authors found that secure couples reported lower levels of unhealthy conflict-resolution styles than insecure and mixed couples, and higher levels of problem solving and satisfaction than insecure couples, thus concluding that “advantages of attachment matches among couples are instead advantages only of the presence of securely attached partners” (p. 46). Similarly to Myers et al., the study by Kuncewicz and Kuncewicz (2019) regarding relationship quality indicated that: a) a combination of two secure patterns was the most, while that of two insecure the least beneficial; b) a combination of secure-insecure patterns (mixed couples) showed to be more favourable than a combination of two insecure ones; c) a combination of secure-insecure patterns coexisted with satisfaction at a level similar to that observed in a combination of secure-secure patterns.
Research in this area presents limitations that should be considered. First, most studies have been conducted in English-speaking countries, so it cannot be assumed that their results are generalizable to the Spanish population. Moreover, this evidence is generally obtained from small-size samples −e.g., 53 couples (Domingue & Mollen, 2009), 35 couples (Myers et al., 2016)−. Furthermore, respondents are generally of a wide age range, from young people in their twenties to elderly in their sixties or even seventies (Domingue & Mollen, 2009; Guzmán & Contreras, 2012; Kuncewicz & Kuncewicz, 2019), which hinders the extrapolation of results to emerging adults. It should be noted that, during emerging adulthood, individuals have to deal with more age-specific tasks and life decisions than at any other stage, not only about couple relationships but about their studies, careers and work as well. Young adults, therefore, face the unique challenge of coordinating romantic commitment and life plans, that is, addressing their own live tasks and integrating them with those of their partner. The strategies used to manage these individual deliberations and joint discussions and resolutions with the romantic partner are relevant in the transition to a long lasting and satisfactory romantic partnership (Shulman & Connolly, 2013).
In brief, to date, a relatively small number of studies have addressed the influence of adult attachment style matching and combination on conflict resolution style and/or perceived relationship quality among romantic couples —to our knowledge, none of them focusing on youngsters—, so more research into this field is warranted, particularly in Spain.
Therefore, by using a fairly large sample of Spanish young-adult heterosexual couples, the main objective is to examine whether partners’ conflict resolution style and perceived relationship quality, considering males and females separately, are influenced by the dyadic configuration of adult attachment. Specifically, the aim is to test the impact of: a) the matching of attachment styles, considering matched (secure-secure, preoccupied-preoccupied, dismissing-dismissing, fearful-fearful) versus unmatched couples; and b) the number of couple members with a secure versus insecure style.
Based on prior findings, it is hypothesized that couples with matched vs. unmatched adult attachment styles significantly differ in perceived relationship quality and use of diverse conflict resolution styles. Likewise, a secure attachment effect −rather than a match effect− is expected, i.e., the higher the number of couple members with a secure style, the higher the relationship quality and the lower the use of dysfunctional conflict resolution styles. Thus, it is assumed that secure couples (with two secure members) show higher levels of relationship satisfaction and constructive problem solving, and lower levels of conflict engagement, withdrawal, and compliance, than mixed and insecure couples (with one or none secure partners, respectively).
Method
Participants
The criteria for inclusion in the sample were: a) being involved in a heterosexual, dating relationship of at least 3 months, and b) at least one member of the couple being 18-25 years. A total of 405 heterosexual couples were recruited. Mean age of women and men was 21 years (SD = 2.11) and 22 years old (SD = 2.28), respectively, with a maximum age difference between partners of 7 years (the oldest person being the man) and 6 years (the oldest being the woman). The majority of female (87%) and male (66%) participants were students, generally at the university (86% and 65%, respectively). On average, the duration of the relationship was 31 months, and only a minority (10%) were living together.
Procedure
The convenience sample was recruited by disseminating the study among Spanish universities, juvenile organizations, and vocational training centers. Those willing to participate indicated their own and their partner's e-mail address in order to receive the link to the online survey. Participation was voluntary and anonymous, and participants were requested to respond individually. Informed consent was required to access the survey questions. Responses to all questions were mandatory. Own and partner´s date of birth and email address were requested to facilitate the matching of couples’ responses (the latter data were subsequently deleted from the databases to preserve anonymity). This study received the ethical approval of the Bioethics Committee of the University of Salamanca.
Measures
Demographic information. This section contained questions about age, gender, relationship duration, studies, and occupation.
Adult attachment. The short-form of the Experiences in Close Relationships-Revised (ECR-R; Fraley et al., 2000) validated in Spanish samples (Fernández-Fuertes et al., 2011) was selected to assess Attachment Anxiety (9 items; e.g., “I often wish that my partner's feelings for me were as strong as my feelings for him or her”) and Attachment Avoidance (9 items; e.g. “I get uncomfortable when a romantic partner wants to get very close”). Participants had to consider their current relationship when rating statements on a seven-point scale ranging from 1 (strong disagreement) to 7 (strong agreement). The internal consistency of this Spanish adaptation is adequate for both Anxiety (α = .80) and Avoidance subscales (α = .86). Cronbach's alphas obtained were .86 and .76, respectively. Higher scores indicated higher attachment anxiety or avoidance.
Conflict-resolution style. The 16-item respondent version of the Conflict Resolution Styles Inventory (Kurdek, 1994) assessed the frequency with which participants employed four conflict-resolutions styles: Positive Problem Solving (e.g., “finding alternatives that are acceptable to each of us”), Conflict Engagement (e.g., “launching personal attacks”), Compliance (e.g., “not defending my position”) and Withdrawal (e.g., “tuning the other person out”). Each style was composed by four items that were rated on a five-point scale ranging from 1 (never) to 5 (always). Cronbach alphas were: .68, .80, .76, and .66, respectively. Higher scores indicated higher frequency of use of the style to deal with arguments or disagreements with the partner.
Relationship quality. Four questions proposed by Conger et al. (2000) were selected to measure the degree of satisfaction, happiness and commitment to the current relationship. These questions are: “How happy are you, all things considered, with your relationship? (0 = not happy at all, 5 = absolutely happy); “All in all, how satisfied are you with your relationship?” (1 = not at all satisfied, 5 = completely satisfied); “How much do you want your relationship with your partner to continue and be a success?” (1= I do not want it, 5 = I want it desperately); “How hard are you willing to work to make your relationship a success?” (1 = I would do nothing; 5 = I would do anything). Responses, therefore, were provided on a five-point scale from 1 to 5. Similarly to prior studies (Conger et al.), Cronbach's alpha for the present study was .75.
Analyses
Participants were first classified into four attachment styles according to their scores on anxiety and avoidance (Fraley, 2012): secure (scores on both scales are lower than the median), fearful (scores on both scales are equal to or higher than the median), dismissing (avoidance score is equal to or higher than the median, and anxiety score is lower than the median), or preoccupied (anxiety score is equal to or higher than the median, and avoidance score is lower than the median).
Secondly, couples were categorized according to the similarity of their members’ attachment style, considering those couples whose members have the same style (secure-secure, preoccupied-preoccupied, dismissing-dismissing, fearful-fearful) as matched, and the rest of attachment style combinations as unmatched.
Finally, couples were also classified in three categories : “secure” (secure-secure), “mixed” (secure-dismissing, secure-preoccupied, secure-fearful) and “insecure” (dismissing-dismissing, dismissing-preoccupied, dismissing-fearful, preoccupied-preoccupied, preoccupied-fearful, fearful- fearful) to examine whether the observed effects in conflict-resolution style and relationship quality are due to the number of couple members with a secure versus insecure style –two in the secure category, one in the mixed and none in the insecure–, rather than to the matching of attachment styles.
In order to test the “match” effect and the “secure attachment” effect, MANOVA for conflict-resolution style, and ANOVA for relationship quality, respectively, were conducted. Couple's members’ gender was also included in analysis as an independent variable to examine its interactions with “matching of attachment styles” and with “number of members with a secure style”. Effect sizes and post-hoc tests were calculated. The level of significance was set at .01.
Results
The matches and non-matches of adult attachment style for couples are shown in Table 1. In 36 percent of cases, couple members’ styles were matched, that is, both were secure (35%), preoccupied (7%), dismissing (12%) or fearful (46%). In addition, a significant relationship between partners´ adult attachment style was found [χ2(9, 405) = 38.77, p < .001, Cramer´s V = .18]. Concretely, there was a significantly higher portion of couples with two securely (ARss= 5.0) or fearfully (ARff= 3.7) attached partners, and a significantly lower portion of couples with a securely attached male and a fearfully attached female (ARsf= −3.6).
Table 1 Matches and non-matches of attachment styles for couples
Males’ attachment style | ||||||
---|---|---|---|---|---|---|
Females’ attachment style | Secure n (%) | Dismissing n (%) | Preoccupied n (%) | Fearful n (%) | Total n (%) | |
Secure (%) | n | 51 (12,6) | 13 (3,2) | 19 (4,7) | 31 (7,7) | 114 (28.1) |
Dismissing n (%) | n (%) | 18 (4,4) | 18 (4,4) | 22 (5,4) | 27 (6,7) | 85 (21.0) |
Preoccupied n (%) | n | 18 (4,4) | 17 (4.2) | 11 (2,7) | 18 (4,4) | 64 (15.8) |
Fearful n (%) | n | 23 (5,7) | 30 (7,4) | 22 (5,4) | 67 (16,5) | 142 (35.1) |
Total n (%) | n | 110 (27,2) | 78 (19.3) | 74 (18.3) | 143 (35,3) | 405 (100) |
Analyses did not show a significant effect of “matching of attachment styles” on relationship quality (p > .01), that is, both matched (secure-secure, preoccupied-preoccupied, dismissing-dismissing, fearful-fearful) and unmatched couples (other combinations) showed similar scores. Likewise, the interaction between gender and matching of attachment styles was not significant (p > .01).
Similarly, no significant effect of matching of attachment styles nor its interaction with gender was found for any of the four conflict resolution styles (p > .01) (see Table 2).
Table 2 Descriptives of relationship quality and conflict resolution styles according to adult attachment style matching and the number of couple members with a secure style.
Attachment style matching | Number of couple members with a secure style | |||||
---|---|---|---|---|---|---|
Unmatched Couples N=147 M (SD) | Matched couples N=258 M (SD) | Insecure couples N=232 M (SD) | Mixed couples N=122 M (SD) | Secure couples N=51 M (SD) | ||
Relationship quality | Male | 4.62(0.43) | 4.56 (0.43) | 4.49 (0.44) | 4.68 (0.40) | 4.86 (0.28) |
Female | 4.55 (0.43) | 4.43 (0.56) | 4.37 (0.52) | 4.64 (0.38) | 4.79 (0.28) | |
Conflict engagement | Male | 1.84 (0.78) | 1.89 (0.88) | 1.92 (0.85) | 1.81 (0.75) | 1.69 (0.79) |
Female | 2.27 (0.89) | 2.15 (0.87) | 2.33 (0.88) | 2.20 (0.90) | 1.78 (0.74) | |
Withdrawal | Male | 2.06 (0.78) | 2.09 (0.74) | 2.21 (0.79) | 1.95 (0.71) | 1.76 (0.58) |
Female | 2.33 (0.80) | 2.27 (0.75) | 2.44 (0.80) | 2.23 (0.75) | 1.90 (0.60) | |
Positive problem solving | Male | 3.95 (0.69) | 3.91 (0.67) | 3.88 (0.69) | 4.01 (0.66) | 4.00 (0.67) |
Female | 3.85 (0.70) | 3.87 (0.69) | 3.75 (0.69) | 3.98 (0.68) | 4.08 (0.68) | |
Compliance | Male | 1.99 (0.78) | 2.03 (0.79) | 2.06 (0.80) | 1.91 (0.72) | 1.97 (0.84) |
Female | 1.55 (0.66) | 1.57 (0.57) | 1.67 (0.65) | 1.43 (0.59) | 1.37 (0.50) |
NOTE: Matched: secure-secure, preoccupied-preoccupied, dismissing-dismissing, fearful-fearful Unmatched: other combinations
On another hand, 13% of couples were labeled as secure, 57% as insecure, and 30% as mixed. There was a significant effect of the number of couple members with a secure style on relationship quality [F(2,423) = 40.46, p < .001, η2 =.16]. The interaction with gender was not significant. For both male and female participants, post-hoc tests showed that partners in insecure couples reported lower levels of relationship quality than those in both secure and mixed couples (p <.001) (see Table 2).
With regard to conflict resolution styles, the number of members with a secure style showed a significant effect on conflict engagement [F(2,423) = 7.67, p = .001, η2 =.04], withdrawal [F(2,423) = 20.57, p < .001, η2 =.09], compliance [F(2,423) = 6.77, p = .001, η2 =.03] and positive problem solving [F(2,423) = 8.47, p < .001, η2 =.04].
Regarding conflict engagement style, post-hoc tests only identified significant differences between insecure vs. secure couples (p = .001), with partners in insecure couples using this style more frequently than those in secure ones.
With respect to withdrawal, post hoc tests showed significant differences between insecure vs. secure (p < .001) and mixed couples (p < .001). Again, the members of insecure couples were more likely to use this strategy when involved in conflicts.
Similarly, significant differences in the use of positive problem solving were found between insecure vs. secure (p = .01) and mixed couples (p = .003), with those in insecure couples being less likely to use this style, as compared to the two other types.
As for compliance, partners in insecure couples significantly differed from those in mixed couples (p = .004): the former used this style more often than the latter.
The interaction of gender and number of members with a secure style was not significant for any of the four conflict resolution styles (see Table 2).
Discussion
In this study, the distribution of adult attachment combinations (13% secure, 57% insecure, 30% mixed) shows some consistencies –the similar portion of mixed couples– and inconsistencies –the remarkably higher portion of insecure couples, especially fearful ones–, with data obtained from other samples (Guzmán & Contreras, 2012; Rivera, 2006). However, this appears to be in line with cross-cultural analyses of adult attachment styles reporting higher mean levels of fearful romantic attachment in Spain than in Chile, as well as a weaker significant negative correlation between secure and fearful attachment among the Spaniards as compared to the Chilean, which suggests that in the former country, the three forms of insecure attachment might coalesce slightly and form a cluster that contrast with the secure scale (Schmitt et al., 2004).
Furthermore, this distribution indicates that preoccupied youngsters in Spain more frequently have a dismissing or fearful partner, while dismissing ones tend to date a preoccupied or fearful partner, and those securely attached are more often romantically involved with another secure person. These data are in line with previous evidence and hypothesis that secure individuals show preference for one another, and are consistent with the conclusion by Holmes & Johnson (2009) that studies on actual matching between partners primarily support the complementarity hypothesis for insecure individuals, while research on attraction to hypothetical/ideal partners generally add evidence to the similarity hypothesis (Holmes & Johnson, 2009).
Contrary to expectations, and unlike previous findings (Myers et al., 2016), couples with matched styles in our sample do not report being significantly more satisfied, or using less conflict engagement, withdrawal and compliance, and more positive problem solving than unmatched couples. In the study by Myers et al., however, the match effect was found to rather be a secure attachment effect, since almost all of the couples in the matched subsample were securely attached, and subsequent analysis showed significant differences in relationship outcomes according to the number of partners in the couple with a secure style. In the present study, conversely, matched couples are more frequently composed of two fearful partners (16.5%) than of two secure partners (12.6%), and this probably has contributed to reduce the fictitious match effect that was hypothesized.
Indeed, as expected, a significant effect of the number of couple members with a secure style is found, and regardless of the gender of the members of the couple. Specifically, the presence of one secure member vs. none seems to be beneficial since it increases the level of relationship satisfaction, and makes it less likely to use unhealthy styles (withdrawal and compliance) and more likely to use positive problem solving when faced to conflicts. Similarly, the presence of two securely attached members in the couple appears to be advantageous (increased relationship quality and positive problem solving; less conflict engagement and withdrawal) but only as compared to cases of two insecurely attached partners.
Therefore, although these significant differences observed between mixed and insecure couples contradicts some previous results (Domingue & Mollen, 2009; Guzmán & Contreras, 2012), they support other authors’ conclusion that, for insecurely attached individuals, it is generally more beneficial to be in a relationship with a securely-attached partner than with a similarly-attached one (Kuncewicz & Kuncewicz, 2019; Myers et al., 2016). Nonetheless, it appears that the presence of at least one secure partner in the couple is what makes difference, because no evidence is found to support that the higher the number of secure members, the better. In other words, no significant differences are encountered between mixed and secure couples in the use of the diverse conflict resolution styles –unlike in some studies (Domingue & Mollen, 2009; Myers et al., 2016)–, nor in relationship quality –like in others (Kuncewicz & Kuncewicz, 2019; Myers et al. 2016)–. Maybe, the discrepancies observed with prior studies are affected by the differences in the samples considered, as the present investigation examined a high (vs. low) number of couples composed of youngsters (vs. adults of a wide age range) involved in a romantic relationship for a low (vs. high) average number of months/years, that is to say, participants in a different developmental stage and probably, a different relational context.
Certain limitations, however, should be borne in mind when interpreting the results and implications of this study. First, it used a convenience sampling procedure to recruit Spanish heterosexual couples, mainly university students, so the findings may not be representative of all young adult couples. Thus, studies including homosexual couples and non-students samples are encouraged. Second, and given the online format of the survey, it is not possible to ensure that participants replied individually. Finally, the correlational nature of the findings does not allow to determine the direction of influences. Future research could thus benefit from using longitudinal research to assess causal effects.
Overall, our results support prior data from other countries by suggesting that Spanish emerging adults’ attachment insecurities foster the use of dysfunctional styles of conflict resolution and lower the levels of satisfaction in the context of romantic relationships. This seems to influenced by the fact that, while secure individuals usually show independence and comfort with intimacy, those insecurely attached tend to display dysfunctional thoughts and feelings about the self and others (i.e. strong need for intimacy and fear of being rejected; emotional detachment and self-sufficiency). This study also points out the advantages associated with a higher number of couple members with a secure style, since working models of secure attachment (i.e. low attachment anxiety and avoidance) appear to reduce destructive patterns of conflict management and improve communication styles.
Consequently, an attachment-based intervention aimed at promoting a more secure bond between the members of the couple could be beneficial for both partners (Feeney & Fitzgerald, 2019). Thus, the change of the adult attachment style could be considered a main aim of the couples’ counseling and psychotherapy, not only due to the benefits for the dynamics of the relationship —as the results suggest—, but also because of the impact on the therapeutic process and effects of the treatment (Levy et al., 2010). Moreover, during emerging adulthood, romantic partners need to become especially aware of the major role that they may play in the fulfillment or collapse of the other´s aspirations, and develop the ability to negotiate and handle this interdependence as a couple (Shulman & Connolly, 2013).