Journal of Adolescence 35 (2012) 133142

Contents lists available at ScienceDirect

Journal of Adolescence
journal homepage: www.elsevier.com/locate/jado

Identity in university students: The role of parental and romantic attachmentq

Marisa vila*, Joana Cabral, Paula Mena Matos*
University of Porto, Portugal

a b s t r a c t
Keywords: Parental attachment Identity development Romantic relationships Mediation University students

In this study with 236 Portuguese university students, we examined two competing models for understanding identity. In the rst model, the direct independent effects of parental and romantic attachment on identity were tested. The second model examined the mediating role of romantic attachment representations in the link between parental attachment and identity. The participants completed measures of parental and romantic attachment at the beginning of the freshman year (Wave 1), and a measure of identity development 18 months later (Wave 2). Structural equation modeling for the independent model indicated that identity is exclusively predicted by romantic attachment at this life stage, and not by parental attachment. Regarding the second model, our mediational hypothesis was supported, since the association between parental attachment and identity was totally mediated by romantic attachment representations. Both models presented adequate indices of adjustment, and provided a complementary understanding of the differential role that romantic relations play in contributing to explaining identity development. 2011 The Foundation for Professionals in Services for Adolescents. Published by Elsevier Ltd. All rights reserved.

The inuence of attachment on psychosocial development has been both theoretically and empirically supported. Research has pointed out that secure parental attachment promotes adaptive psychological functioning (Leondari & Kiosseoglou, 2000; Lopez & Brennan, 2000) and encourages the development of identity (rseth, Kroger, Martinussen, & Marcia, 2009; Kennedy, 1999; Marcia, 1983; Quintana & Lapsley, 1987). Parental attachment also affects competence in romantic relationships. Adolescents who are more securely attached have closer and more durable relationships with their dating partners (Collins & Sroufe, 1999; Mayseless & Scharf, 2007). Using an alternative, yet complementary approach, researchers began examining peer relationships in late adolescence, more recently focusing on romantic relationships, exploring the inuence of this relational context on identity development (Vogensen, 2003). Although the value of these studies and reviews is unquestionable, the assessment of clear links between the nature of the parent-adolescent relationship and the development of identity remains difcult to examine (Meeus & Minet, 2007). This is likely due to the complexity of this developmental task and its consequent assessment and measurement. Moreover, the emergence of new relational contexts during this life phase might inuence both parental attachment relationships and identity development, factors that must be considered. The role of romantic relationships in the development of adolescents has recently appeared and merits further examination (Collins, 2003; Collins, Welsh, & Furman, 2009). There is a lack of

q This study was supported by a research grant (PTDC/PSI/65416/2006) from the Portuguese Foundation for Science and Technology (FCT). This study has also been supported by a PhD. scholarship grant (SFRH/BD/18357/2004) attributed to the second author. * Corresponding authors. Marisa vila or Paula Mena Matos, Faculty of Psychology and Science Education, University of Porto, Rua Dr. Manuel Pereira da Silva, 4200-392 Porto, Portugal. Tel.: 351 226 079 700. E-mail addresses: marisavila@fpce.up.pt (M. vila), pmmatos@fpce.up.pt (P.M. Matos).
0140-1971/$ see front matter 2011 The Foundation for Professionals in Services for Adolescents. Published by Elsevier Ltd. All rights reserved. doi:10.1016/j.adolescence.2011.05.002

134

M. vila et al. / Journal of Adolescence 35 (2012) 133142

research on whether identity is related to romantic attachment (Berman, Weems, Rodriguez, & Zamora, 2006), despite the recognition that the attachment system functions on the formation of new bonds in young adults (Bowlby, 1988; Hazan & Shaver, 1987). Therefore, studies that consider the role played by both parental and romantic attachment in the prediction of identity in late adolescence might contribute to a broader and more comprehensive understanding of the articulation of both relational contexts in the lives of late adolescents. Driven by the potential of both parental and romantic attachment relationships in promoting psychosocial growth, namely identity, we designed and tested two competing models. We rst inspected the direct inuence of both attachment contexts (parental and romantic) on identity. We anticipated that romantic attachment representations might operate as stronger predictors of this dimension, considering the signicance of romantic relationships during this period of life. In a second model, we present a mediational hypothesis in which we anticipated that romantic attachment representations might function as a mediating mechanism whereby parental bonds are involved in identity development. In the following sections, we discuss the links between parental attachment and identity in late adolescence and summarize ndings from available empirical studies. We then provide a conceptual and empirical review of romantic attachment in late adolescence and its inuence on the developmental outcome under study. Parental attachment, romantic relations and identity in late adolescence Identity formation is a dynamic and life-long process, characterized by continuous changes in individuals amount of exploration and the strength of their commitments (Erikson, 1968; Marcia, 1983). This process takes place while the parentadolescent relationship is also undergoing major changes. As adolescents become more capable of self-sufciently dealing with lifes challenges, parents may learn to recognize and adjust to their childrens near-adult maturity and encourage their development either implicitly or explicitly. The changes that occur on both sides allow parents and adolescents to move toward a relationship that is less hierarchical and more egalitarian (Matos, 2006). Nevertheless, these relationships continue to be crucial contexts for psychosocial development. In general, affective support and connectedness with parents seems to provide an optimal environment for the encouragement of identity formation (Noom, Dekovic & Meeus, 1999). Previous research supports the developmental advantage of a secure attachment representation. Secure attachment to parents encourages identity development (Meeus, Oosterwegel & Vollebergh, 2002; Samuolis, Layburn, & Schiafno, 2001; Zimmermann, 2002), whereas less secure parental attachment appears to be related with identity diffusion (Matos, Barbosa, Almeida, & Costa, 1999). Additionally, it was found that parental attachment is predictive of social skills, which affect competence in romantic relationships (Engels, Finkenauer, Meeus, & Dekovic, 2001). In a similar study, acceptance and encouragement of independence in the parental relationship contributed to better social competence, which in turn promoted higher capacity for intimacy in romantic relationships (Scharf & Mayseless, 2001). Young adults who retrospectively reported close and autonomous relationships with their parents also reported closer, more intimate and more secure relationships with their romantic partner (Hazan & Shaver, 1987; Furman & Simon, 1999; Kobak & Sceery, 1988). More recent studies reveal that attachment organization in infancy was found to be predictive of the emotional quality of romantic relationships in early adulthood (Simpson, Collins, Tran, & Haydon, 2007). There is also evidence that the quality of observed interaction patterns in the family (e.g., warmth and sensitivity) seems to be positively related to attachment security and similar behaviors between romantic partners (Dinero, Conger, Shaver, Wideman & Larsen-Rife, 2008). Taken together, these studies bring to light the implication of parental attachment on identity development and on the creation and development of romantic relationships. However, in a meta-analysis review, Meeus and Minet (2007) questioned the nature of the association between parent-adolescent relationships and identity development. More specically, in terms of parental attachment, the authors found that few studies revealed direct links. This may be due to the reorganization of the parent-adolescent relationship that occurs at this life stage and to the diverse environmental characteristics involved, namely the importance of peers and the romantic partner. It is plausible to consider that parental attachment may inuence identity development through these other relational contexts. Romantic attachment and identity development in late adolescence During their adolescent years, youth gradually abandon idealized representations of their parents and become increasingly involved in intimate relationships with their peers. These relationships ultimately become primary sources of social support by the time these adolescents enter young adulthood (Overbeek, Engels, Meeus, & Volleberg, 2003; Seiffge-Krenke, 2003). Furman and Shaffer (2003) argue that romantic relationships play an important role in the development of identity, in the development of close relationships with peers and on the level of transformations in family relationships. Romantic partners inuence identity formation in a variety of ways, including by providing encouragement and direction and by serving as role models (Vogensen, 2003). The partners actions, non-verbal signs and verbal expressions provide a rich source from which to infer and adjust ones self-view (Hinde, Finkenauer, & Auhagenc, 2001). More recently, it was discovered that the developmental demand for establishing a separate identity is an important contributor to young adults romantic relationships across most stages of adolescence (Collins et al., 2009). Therefore, there is a common assumption that optimal close relationships are characterized by a balance between self and others (Kegan, 1982; Seiffge-Krenke, Shulman, & Klessinger, 2001). Taking an opposite approach, Berman et al. (2006) found that identity status was signicantly related to romantic attachment style. However, these researchers pointed out the complex nature of the associations between identity and romantic attachment, mentioning that identity statuses do not dictate romantic attachment styles.

M. vila et al. / Journal of Adolescence 35 (2012) 133142

135

This study seeks to expand on these previous studies by investigating the differential function of romantic attachment on identity development in a sample of university students, given that we consider their direct application (e.g., with parental attachment) as well as their intervening role (e.g., between the link of parental attachment and the psychosocial processes under study). The present study In this study, we intend to analyze the inuence of parental and romantic attachment on identity development by means of two competing models. We expected that in the independent model (Model A), differences in parental attachment relationships as well as in romantic attachment representations would predict differences in identity. We predicted that relationships characterized by secure attachment dimensions would promote identity development, while relationships marked by insecure attachment dimensions would have the opposite effect. Moreover, given the salience of romantic relationships at this age, as well as the process of reorganizing emotional family bonds, we expected that romantic attachment could function as a better predictor of identity development (Fig. 1). With regard to the second model, we predicted that the inuence of parental attachment on identity could be mediated by romantic attachment (Model B). More specically, emotionally close and supportive parental relationships that allow for movement toward psychological independence and autonomy, without exaggerated separation anxiety, should promote secure romantic relationships. As a result, these late adolescents should be able to develop a positive sense of identity. The opposite effect was expected among adolescents whose parental relationships were characterized by insecure attachment dimensions. Finally, we hypothesized that romantic attachment would mediate the expected association between parental attachment and identity. The conceptual model for the mediational hypothesis is represented in Fig. 2. Method Participants We used longitudinal data gathered from 236 rst-year Caucasian students attending various courses at the University of Porto, the largest university in Portugal, located in Porto. One hundred and sixty-eight females (71.2%) and 68 males (28.8%), ranging in age from 17 to 20 years (M 17.96; SD .53) participated in the two waves of the study. Participants came from medium socioeconomic strata, based on their fathers level of education and current profession. All participants were engaged in a romantic relationship, with durations varying from one month to approximately 4.5 years (M 13.92 months; SD 11.71). Fifty-seven participants (33.5%) were in relationships that ranged from one month to one year, 41 participants (24.1%) were in relationships that ranged from over one year to two years and 42 participants (24.7%) were in relationships ranging more than two years. The maximum relationship duration for this last group was 4.5 years. Procedures Participants were recruited from seven different faculties within the University of Porto: Architecture, Science, Law, Economy, Pharmacy, Medicine and Psychology. Students were asked to voluntarily participate in a larger study concerning the psychosocial development of university students. The main goals of the study were briey presented both orally and in an introductory letter. Participants were assured that their responses to survey questions would be kept condential. These

Parental Attachment Identity

Romantic Attachment

Fig. 1. Conceptual model A.

136

M. vila et al. / Journal of Adolescence 35 (2012) 133142

Parental Attachment

Romantic Attachment

Identity

Fig. 2. Conceptual model B.

students then completed the package of questionnaires during the initial weeks of their rst year, in classrooms and under standardized group administration conditions. In the rst wave of the study, each participant was assigned an identication code that enabled their identication in the subsequent wave of the study. Participants were not offered any incentives for participation, such as extra course credit or monetary compensation, since the use of such incentives in survey research is not common practice in Portugal. In the second wave, midway through the students third year of studies (18 months after the rst administration), data concerning psychosocial development was gathered. We checked for random pattern of sample attrition and found that participants who remained in the second wave (N 236) did not differ from those who dropped out of the study after the rst wave (N 573) in any of the demographic measures or in parental attachment (Pillais trace .01; F (9, 2772) 1.63, ns) or romantic attachment (Pillais trace .01; F (12, 2739) 1.05, ns).

Measures Parental attachment Parental attachment was measured using the Father/Mother Attachment Questionnaire (FMAQ; Matos & Costa, 2001a), a self-report measure that was developed in Portugal and is being widely used in empirical research with Portuguese participants. It assesses attachment representations separately for each parent along three main subscales. The rst, Quality of Emotional Bond, assesses the experience of an emotional, close and supportive relationship with parents (10 items; e.g., I know that I can count on my parents whenever I need them). The second, Inhibition of Exploration and Individuality, assesses the absence of encouragement from parents when exploring the environment as well as the perception that parents are intrusive and inhibiting of psychological differentiation (10 items; e.g., My parents discourage me when I want to try new things.). The third and nal subscale, Separation Anxiety, taps the disruptive experience of levels of separation anxiety (10 items; e.g., I am afraid of being alone if I lose my parents.). Results from psychometric and validity studies with several independent samples indicate that FMAQ is a valid and reliable measure for studying attachment representations in adolescence and young adulthood (Matos & Costa, 2004; Matos et al., 1999). Father/Mother Attachment Questionnaire scores have correlated in expected directions with scores on independent measures of parental/peer attachment, separation-individuation, self-perceptions and family environment. In the present study, however, we used the continuous scores on the three subscales. Our results revealed that secure attached individuals presented high levels of Quality of Emotional Bond, low levels of Inhibition of Exploration and Individuality, and low to moderate levels of Separation Anxiety. On the other hand, insecure attached individuals tended to present low levels of Quality of Emotional Bond, high levels of Inhibition of Exploration and Individuality and either very low or very high levels of Separation Anxiety. In the present sample, internal reliability was high for scores on all FMAQ subscales (Cronbachs alpha ranged from .89 to .89 and from .85 to .86 for father and mother, respectively). Romantic attachment The brief version of the Romantic Attachment Questionnaire (RAQ, Matos & Costa, 2001b) was used to assess attachment representations regarding the current romantic relationship. This instrument was developed and validated with Portuguese participants. We considered the following subscales for this study: Trust assesses the extent to which the romantic partner is seen as a source of emotional support (5 items; e.g., I know that I can count on my boyfriend/girlfriend whenever I need him/ her); Dependence assesses the need for physical and emotional proximity, separation anxiety and fear of loss (5 items; e.g., When I cant be with my boyfriend/girlfriend, I feel abandoned); and Avoidance reects the secondary role of the romantic partner in the fulllment of attachment needs (5 items; e.g., When I have a problem, I prefer being alone instead of being with my boyfriend/girlfriend). The questionnaire presented adequate internal consistency (Cronbachs alpha ranged from .70 to .91). Predictable relationships between RAQ and other constructs (rejection sensitivity, social support and network orientation) were found to evidence construct validity (Matos, 2002; Santos & Matos, 2007). In addition, based on the subscales of the Experiences in Close Relationships Scale (Brennan, Clark, & Shaver, 1998), it was observed that anxiety correlated signicantly with dependence (r .61, p < .001) and avoidance correlated with avoidance (r .31, p < .001). Avoidance was negatively correlated with trust (r .71, p < .001).

M. vila et al. / Journal of Adolescence 35 (2012) 133142

137

Identity The Erikson Psychosocial Inventory Scale (EPIS; Rosenthal et al., 1981) assesses the six stages of psychosocial development proposed by Erikson. The questionnaire was translated into Portuguese according to International Test Commission (Hambleton, 2005) guidelines. In this study, we used the subscales of Autonomy (6 items, a .83; e.g., I like to make decisions on my own.) and Identity (6 items, a .86; e.g., I have a sense of who I am and what I want from my life.), both of which presented adequate internal consistency. Overview of analyses We sought to test two comprehensive yet parsimonious models involving our key variables. Given the high correlation between separate father and mother scores on the IEI, QEB and SA dimensions (all rs > .80, p < .01), each FMAQ dimension was represented in the model as a latent variable combining both the father and mother scores as indicators. The same procedure was applied to the psychosocial dimensions being studied, considering the moderately high intercorrelation between Autonomy and Identity scores (r .79, p < .001). In addition, taking into account their conceptual overlap and the proximity between both processes, we combined scores on these measures to create a composite variable that we labeled as Identity. Latent variables using item parcels as observed variables were used for all the variables in this study. Applying a domain sampling rationale, we assigned each item to each one of the parcel groupings, randomly and without replacement. Random assignment of item groupings resulted in parcels that contain roughly equal common factor variance (Little, Cunningham, Shahar, & Widaman, 2002). Both hypothetical causal models contained manifest (measured) and latent variables. These models were examined using Structural Equation Models (SEM; with EQS, version 6.1), using maximum likelihood estimation. In the rst model, we tested the direct independent affects of parental attachment and romantic attachment on identity, imposing covariances between the predictors (latent variables). In the second model, the analyses were conducted using the same method and following the procedures outlined by Baron and Kenny (Kenny, 2006) for testing mediational effects. Results Preliminary analyses Table 1 presents the intercorrelations as well as the means, standard deviations and ranges of the scores for the studys key dimensions. Regarding parental attachment, Quality of Emotional Bond (QEB) and Inhibition of Exploration and Individuality (IEI) correlated negatively (r .47, p < .001), while QEB correlated positively with Separation Anxiety (SA) (r .52, p < .001). Quality of Emotional Bond correlated positively with Trust (r .21, p < .01), while IEI correlated positively with Avoidance (r .22, p < .01) and SA correlated positively with Dependence (r .29, p < .001). Results were consistent according to the conceptualization of the attachment dimensions. Also as expected, QEB positively correlated with Identity (r .20, p < .01) and IEI negatively correlated with Autonomy (r .16, p < .05). No signicant correlations were found between SA and any of the psychosocial dimensions. Intercorrelations of scores on romantic attachment and psychosocial measures were also as expected. Dependence was negatively correlated with Autonomy (r .21, p < .01), while Trust positively correlated with Identity (r .16, p < .05). No signicant correlations were found between Avoidance and any of the psychosocial dimensions. Measurement model A This measurement model indicated an adequate t: c2 (84, N 236) 131.97, p < .0001; CFI .98; RMSEA .05 (90% CI: lower bound .033, upper bound .066); and SRMR .06. Table 2 shows that all loadings of the measured variables on the latent variables were statistically signicant, therefore all of the latent variables appear to have been adequately measured by their respective indicators. Measurement model A was used to test the structural model, which will be described ahead.

Table 1 Means, standard deviations, and correlations among dimensions. M 1. 2. 3. 4. 5. 6. 7. 8. Inhibition of Exploration and Individuality Quality of Emotional Bond Separation Anxiety Trust Dependence Avoidance Autonomy Identity 2.84 5.17 3.53 4.99 3.19 2.52 4.67 4.61 SD .91 .65 .85 .75 .78 .86 .73 .82 1 2 -.47*** 3 -.16* .52*** 4 -.27*** .21** .02 5 .01 .00 .29*** .37*** 6 .22** -.16* -.17** -.69*** -.59*** 7 -.16* .10 -.11 .10 -.21** .01 8 -.16* .20** .04 .16* -.06 -.11 .78***

Note. N 218. *p < .05 **p < .01 ***p < .001.

138 Table 2 Factor loadings for measurement model A. Measure and variable Trust Trust Parcel 1 Trust Parcel 2 Trust Parcel 3 Dependence Dependence Parcel 1 Dependence Parcel 2 Dependence Parcel 3 Identity Identity Parcel 1 Identity Parcel 2 Identity Parcel 3

M. vila et al. / Journal of Adolescence 35 (2012) 133142

B 1.00 1.00 .96 1.00 1.27 1.36 1.00 .87 .88

SE .07 .07 .13 .14 .08 .08

Z 14.93 13.86 10.16 10.08 11.77 11.77

.86*** .86*** .81*** .69*** .84*** .83*** .88*** .88*** .88***

Note: N 236. According to Byrne (1994), in order to guarantee the specication of the measurement model, some of the factor loadings should be constrained to a known value, typically 1.00. ***p < .001.

Independent affects of romantic attachment on identity Results from the structural model supported some of our expectations, as illustrated in Fig. 3. The observed affects from the nal model demonstrated that only romantic attachment dimensions, which were assessed at the beginning of participants rst year of university, had a direct inuence on predicting identity, which was assessed midway through the third year of university (18 months after the rst administration). No dimension of parental attachment was found to be directly associated to identity when simultaneously considering the direct inuence of romantic attachment dimensions. With the exception of Avoidance, all romantic attachment dimensions were related to Identity through different paths and magnitudes. Dependence was a negative predictor of Identity (r 34, p < .001), while Trust was found to forecast (r .55, p < .001). Measurement model B This measurement model indicated an adequate t: c2 (125, N 236) 208.82, p < .001; CFI .97; RMSEA .06 (90% CI: lower bound .041, upper bound .067); and SRMR .06. The model did not include any correlated error terms. Table 3 illustrates that all loadings of the measured variables on the latent variables were statistically signicant; therefore, all of the latent variables appear to have been adequately measured by their respective indicators. Measurement model B was used to test the structural models described next. Mediational hypothesis The results from the structural model partially supported the mediational hypothesis, as illustrated in Fig. 4. In our nal structural model, the Avoidance dimension was removed, since it was found to have no relation to the nal variable in the

Fig. 3. Model A: Independent effects of romantic attachment on identity.

M. vila et al. / Journal of Adolescence 35 (2012) 133142 Table 3 Factor loadings for the mediational model. Measure and variable Inhibition of Exploration and Individuality Parcel 1 Parcel 2 Parcel 3 Quality of Emotional Bond Parcel 1 Parcel 2 Parcel 3 Separation Anxiety Parcel 1 Parcel 2 Parcel 3 Trust Parcel 1 Parcel 2 Parcel 3 Dependence Parcel 1 Parcel 2 Parcel 3 Identity Parcel 1 Parcel 2 Parcel 3 B 1.00 .96 .96 1.00 1.49 1.49 1.00 1.02 1.02 1.00 1.02 1.02 1.00 .85 .85 1.00 1.13 1.13 SE .06 .06 .12 .12 .07 .07 .05 .05 .06 .06 .06 .06 Z 15.71 15.71 12.72 12.72 14.39 14.39 21.06 21.06 13.50 13.50 17.80 17.80

139

.85*** .77*** .85*** .69*** .90*** .93 .83*** .79*** .80*** .88*** .89*** .90*** .78*** .83*** .76*** .83*** .88*** .89***

Note: N 236. According to Byrne (1994), in order to guarantee the specication of the measurement model, some of the factor loadings should be constrained to a known value, typically 1.00. ***p < .001.

study. First, we observed an indirect effect of Inhibition of Exploration and Individuality (IEI) on Identity, through a connection with Trust. More specically, we found that IEI constitutes a negative predictor of Trust (b .24, z 3.54, p < .001). These lower levels of Trust are predictive of lower levels of Identity (b .22, z 2.59, p < .01). Considering the role of Quality of Emotional Bond (QEB), we found that the relationship between QEB and Identity is partially mediated by Dependence, since QEB continues to directly predict the nal outcome to a lesser degree. Quality of Emotional Bond is a negative predictor of Dependence (b .33, z 4.45, p < .001), which in turn discourages Identity development (b .23, z 2.72, p < .01). As for the link between Separation Anxiety (SA) and Identity, we found this to be totally mediated by Dependence, meaning that the association between SA and Identity is completely explained by Dependence. We found that SA is a positive predictor of Dependence (b .53, z 5.37, p < .001) since higher levels of SA are associated with higher levels of Dependence, which in turn forecasts lower levels of Identity development (b .23, z 2.72, p < .01).

Fig. 4. Model B.

140

M. vila et al. / Journal of Adolescence 35 (2012) 133142

Discussion The aim of this study was to examine the prospective inuence of parental and romantic attachment on identity in Portuguese university students and, more specically, to test two competing models wherein the distinct qualities of romantic attachment representations could inuence identity development both directly (in the independent model) and indirectly (in the mediational model). Our study revealed that romantic attachment representations have a greater impact on identity development of late adolescents, as can be seen in the independent model. Only romantic attachment remains a direct predictor of Identity, through Trust (a positive predictor of Identity) and Dependence (a negative predictor of Identity). Moreover, we discovered that these dimensions underlie the inuence of parental attachment on identity development (through mediating effects and one indirect effect). These ndings were consistent with our expectation that romantic relationships represent crucial contexts for identity development in late adolescence. During this life stage, romantic partners begin to assume the role of primary attachment gures, such that these relationships should inuence attachment representations just as relationships with parents once did (Feeney, 2004). In addition to encouraging a sense of relatedness to others, romantic relationships may help adolescents successfully establish autonomy and identity development as they explore extra-familial relationships and come to rely less on their parents (Berman et al., 2006; Furman & Shaffer, 2003). The mediational model illustrated that romantic attachment representations intercede the expected associations between parental attachment and identity. Parental attachment operates on identity development only in the presence and interference of the Trust and Dependence developed in the romantic relationship. Regarding Inhibition of Exploration and Individuality, we found this dimension to be a negative predictor of Trust, in turn hindering Identity development. These ndings were consistent with our expectation that intrusive parenting and/or lack of encouragement toward autonomy would fail to create a secure base condition. This parental inhibition of the exploration process may reduce opportunities for late adolescents to develop intimacy skills with peers and learn how to trust others, subsequently inhibiting the process of Identity development. In relation to this nding, some studies showed that the experience of nurturing parenting in the family of origin (e.g., warm, supportive and low in hostility) is predictive of competent behaviors in young adults romantic relationships (Conger, Cui, Bryant & Elder, 2000). As for the link between Quality of Emotional Bond and Identity, we found that this dimension remains a direct predictor of identity, albeit with a decrease in the magnitude of its direct effect due to the partial mediational process that occurs through Dependence. We could assume that this parental attachment dimension has a more stable and long-term effect on psychosocial development in comparison to the other two dimensions assessed in our study (namely IEI, which appears to be a more relational dimension of attachment to parents). These results highlight the developmental signicance of having an emotionally close and condent relationship with parents and experiencing this relationship as a source of support and comfort, which seems to function as a developmental context for the exploration and commitment processes that underlie the construction of identity. In relation to this nding, a number of past studies have revealed the inuence that nurturing parenting practices have on fostering open and exible identity exploration (Smits et al., 2008). Finally, regarding Separation Anxiety and Identity, we found these to be totally mediated by Dependence. This means that anxiously attached late adolescents appear to have higher levels of Dependence on romantic partners, constituting a negative predictor of Identity. These youth display entrenched concerns of being abandoned, of being taken advantage of or of failing to meet their needs for a greater sense of security, likely due to their experiences with higher levels of anxiety (Simpson et al., 2007). Relationships characterized by emotional over-involvement may not create the necessary conditions for selfdifferentiation and separation-individuation processes, leading to the persons identity being dened mainly in terms of the relationship itself (Kegan, 1982). In general, we found that both the models discussed in this study provide insight into the differential meaning of romantic relationships on identity development. Studies on the conjoint effects of both attachment contexts and identity suggest, on one hand, that the presence of both attachment dimensions better predicts identity than does the presence of only one or another variable. On the other hand, this mediational model is consistent with attachment theorists arguments that secure attachment relationships provide the foundation for the development of ones self (Sroufe, 2002). This study presents some limitations that impede the generalization of the results. Besides being a convenience sample, the sample was constituted by only Caucasian university students and was unbalanced in terms of gender, since the majority of participants were female. In addition, the length of the romantic relationship was quite varied, again pointing to the fact that romantic relationships were in different phases of attachment development. As in any other study using self-report measures, these results have some limitations given the spontaneous nature and reliability of responses. Future research should rely on multi-informant designs, since the evaluation of parental attachment and of romantic partners representations of attachment might offer a more comprehensive approach to the dimensions in the study. Acknowledgments We thank Bart Soenens and Wim Beyers for hosting the second author at the Ghent University and supporting her in the development of competence in Structural Equation Modeling. We thank Frederick Lopez and Maria Emlia Costa for the inspiring discussions about this study.

M. vila et al. / Journal of Adolescence 35 (2012) 133142

141

References
rseth, A., Kroger, J., Martinussen, M., & Marcia, E. J. (2009). Meta-analytic studies of identity status and the relational issues of attachment and intimacy. Identity: An International Journal of Theory and Research, 9, 3291. Berman, S., Weems, C., Rodriguez, E., & Zamora, I. (2006). The relation between identity status and romantic attachment style in middle and late adolescence. Journal of Adolescence, 29, 737748. Bowlby, J. (1988). A secure base: Parent-child attachment and healthy human development. London: Basic Books. Brennan, K. A, Clark, C. L., & Shaver, P. R. (1998). Self-report measurement of adult attachment: An integrative overview. In J. A. Simpson, & W. S. Rholes (Eds.), Attachment theory and close relationships (pp. 4676). NY: Guilford Press. Byrne, B. (1994). Structural Equation Modeling With Eqs: Basic Concepts, Applications, And Programming. SAGE Publications. Collins, W. (2003). More than myth: the developmental signicance of romantic relationships during adolescence. Journal of Research on Adolescence, 13, 124. Collins, W., & Sroufe, L. (1999). Capacity for intimate relationships: a developmental construction. In W. Furman, B. B. Brown, & C. Feiring (Eds.), The development of romantic relationships in adolescence (pp. 125147). Cambridge: Cambridge University Press. Collins, W., Welsh, D., & Furman, W. (2009). Adolescent romantic relationships. Annual Review of Psychology, 60, 631652. Conger, D., Cui, M., Bryant, M., & Elder, G. (2000). Competence in early adult romantic relationships: a developmental perspective on family inuences. Journal of Personal and Social Psychology, 2, 224237. Dinero, R., Conger, R., Shaver, P., Widaman, K., & Larsen-Rife, D. (2008). Inuence of family of origin and adult romantic partners on romantic attachment security. Journal of Family Psychology, 22, 622632. Engels, E., Finkenauer, C., Meeus, W., & Dekovic, M. (2001). Effects of parental attachment on adolescents emotional adjustment: the effects of social skills and relational competence. Journal of Counseling Psychology, 48, 428439. Erikson, E. H. (1968). Identity: Youth and crisis. New York: W. W. Norton. Feeney, C. (2004). A secure base: responsive support of goal strivings and exploration in adult intimate relationships. Journal of Personality and Social Psychology, 87, 631648. Furman, W., & Shaffer, L. (2003). The role of romantic relationships in adolescent development. In P. Florsheim (Ed.), Adolescent romantic relations and sexual behaviour: Theory, research and practical implications (pp. 322). Mahwah: Lawrence Erlbaum Associates, Publishers. Furman, W., & Simon, V. (1999). Cognitive representations of romantic relationships. In W. Furman, B. B. Brown, & C. Feiring (Eds.), The development of romantic relationships in adolescence (pp. 7598). New York: Cambridge University Press. Hambleton, R. K. (2005). Issues, designs and technical guidelines for adapting tests into multiple languages and cultures. In R. K. Hambleton, P. F. Merenda, & C. D. Spielberger (Eds.), Adapting educational and psychological tests for cross-cultural assessment (pp. 338). Mahwah, N.J: Erlbaum. Hazan, C., & Shaver, P. (1987). Romantic love conceptualized as an attachment process. Journal of Personality and Social Psychology, 52, 291332. Hinde, R., Finkenauer, C., & Auhagenc, A. (2001). Relationships and the self-concept. Personal Relationships, 8, 187204. Kegan, R. (1982). Evolving self. Harvard University Press. Kennedy, H. (1999). Romantic attachment style and ego identity, attributional style and family of origin in rst-year college students. College Student Journal, 33, 171180. Kenny, D. A. (2006). Mediation. Retrieved February 7, 2006. http://davidakenny.net/cm/mediate.htm. Kobak, R., & Sceery, A. (1988). Attachment in late adolescence: working models, affect regulation and representations of self and others. Child Development, 59, 135146. Leondari, A., & Kiosseoglou, G. (2000). The relationship of parental attachment and psychological separation to the psychological functioning of young adults. The Journal of Social Psychology, 140, 451464. Little, T. D., Cunningham, W. A., Shahar, G., & Widaman, K. F. (2002). To parcel or not to parcel: exploring the question and weighing the merits. Structural Equation Modeling, 9, 151173. Lopez, F., & Brennan, K. (2000). Dynamic processes underlying adult attachment organization: toward an attachment theoretical perspective on the healthy and effective self. Journal of Counseling Psychology, 47, 283300. Marcia, J. E. (1983). Some directions for the investigation of ego-development in early adolescence. Journal of Early Adolescence, 3, 215223. Matos, M. P. (2006). Relaes romnticas em adolescentes. [Adolescents romantic relationships]. Psychologica, 41, 924. Matos, M. P. (2002). (Des)continuidades na vinculao aos pais e ao par amoroso em adolescentes. [(Dis)continuities in adolescents attachment to parents and to romantic partner]. Porto: Faculdade de Psicologia e de Cincias da Educao. Unpublished Phd Dissertation. Matos, M. P., Barbosa, S., Almeida, H., & Costa, E. M. (1999). Parental attachment and identity in Portuguese late adolescents. Journal of Adolescence, 22, 805818. Matos, M. P., & Costa, E. M. (2004). Assessing attachment representations in adolescence: The Father/Mother Attachment Questionnaire (2004). IX Conference of the European Association for Research on Adolescence Proceedings, Porto, 129. Matos, M. P. & Costa, E. M. (2001a). Father/Mother Attachment Questionnaire. Unpublished manuscript. Faculty of Psychology and Science Education, University of Porto. Matos, M. P., & Costa, E. M. (2001b). Romantic Attachment Questionnaire. Unpublished manuscript. Faculty of Psychology and Science Education, University of Porto. Mayseless, O., & Scharf, M. (2007). Adolescents attachment representations and their capacity for intimacy in close relationships. Journal of Research on Adolescence, 17, 2350. Meeus, W., & Minet, W. (2007). Relationships with parents and identity in adolescence: a review of 25 years of research. In M. Watzlawick, & A. Born (Eds.), Capturing identity: Quantitative and qualitative methods (pp. 131144). University Press of America. Meeus, W., Oosterwegel, A., & Vollebergh, W. (2002). Parental and peer attachment and identity development in adolescence. Journal of Adolescence, 25, 93106. Noom, M., Dekovic, M., & Meeus, W. (1999). Autonomy, attachment and psychosocial adjustment during adolescence: a double-edged sword? Journal of Adolescence, 22, 771783. Overbeek, G., Engels, R., Meeus, W., & Volleberg, W. (2003). Parental attachment and romantic relationships: associations with emotional disturbance during late adolescence. Journal of Counseling Psychology, 50, 2839. Quintana, M., & Lapsley, K. (1987). Adolescent attachment and ego identity: a structural equations approach to the continuity of adaptation. Journal of Adolescent Research, 2, 393409. Rosenthal, A. D., Gurney, M. R., & Moore, M. S. (1981). From trust on intimacy: A new inventory for examining Erikson's stages of psychosocial development. Journal of Youth and Adolescence, 10, 525537. Samuolis, J., Layburn, K., & Schiafno, K. (2001). Identity development and attachment to parents in college students. Journal of Youth and Adolescence, 30, 373384. Santos, R., & Matos, M. P. (2007). Implicaes da sensibilidade rejeio e do conito interparental na vinculao romntica. [Implications of rejection sensitivity and interparental conict in romantic attachment]. Psicologia: teoria, investigao e prtica, 12, 341359. Scharf, M., & Mayseless, O. (2001). The capacity for romantic intimacy: exploring the contribution of best friend and marital and parental relationships. Journal of Adolescence, 24, 379399. Seiffge-Krenke, I. (2003). Testing theories of romantic development from adolescence to young adulthood: evidence of a developmental sequence. International Journal of Behavioral Development, 27, 519531. Seiffge-Krenke, I., Shulman, S., & Klessinger, N. (2001). Adolescent precursors of romantic relationships in young adulthood. Journal of Social and Personal Relationships, 18, 327346.

142

M. vila et al. / Journal of Adolescence 35 (2012) 133142

Simpson, J. A., Collins, W. A., Tran, S., & Haydon, K. C. (2007). Attachment and the experience and expression of emotions in adult romantic relationships: a developmental perspective. Journal of Personality and Social Psychology, 92, 355367. Smits, I., Soenens, B., Luyckx, K., Duriez, B., Berzonsky, & Goossens, L. (2008). Perceived parenting dimensions and identity styles: exploring the socialization of adolescents processing of identity-relevant information. Journal of Adolescence, 31, 151164. Sroufe, L. A. (2002). From infant attachment to promotion of adolescent autonomy: prospective, longitudinal data on the role of parents in development. In J. Borkowski, & S. Landesman (Eds.), Parenting and the childs world: Inuences on academic, intellectual, and social-emotional development (pp. 187202). Mahwah, NJ: Erlbaum. Vogensen, A. (2003). Finding themselves: identity formation and romantic attachment styles in late adolescence. Dissertation Abstracts International, Section B: The Sciences & Engineering, 64, 1529. Zimmermann, P. (2002). Stability of attachment representations during adolescence: the inuence of ego-identity status. Journal of Adolescence, 25, 107124.