Leaving home for college - Taylor & Francis Online

44 downloads 0 Views 281KB Size Report
Leaving home for college: A potentially stressful event for .... The transition to college is generally viewed as a positive step up in life involving many new.
Attachment & Human Development, June 2005; 7(2): 171 – 185

Leaving home for college: A potentially stressful event for adolescents with preoccupied attachment patterns

ANNIE BERNIER1, SIMON LAROSE2, & NATASHA WHIPPLE1 1

University of Montreal, Canada, and 2Laval University, Quebec City, Canada

Abstract Sixty-two high school students, 28 of which were planning on leaving home to attend college, completed the Adult Attachment Interview and self-report questionnaires pertaining to their relationship with their parents. Compared to their autonomous counterparts, preoccupied students who had left home reported having a more negative relationship with each parent and experiencing more family-related stress. However, they reported having more contact with each parent. In contrast, no attachment differences with regards to perceptions of the parent-adolescent relationship were found among students who lived at home while in college. This suggests that individual differences related to attachment state of mind in adolescence may be magnified by a stressful life experience.

Keywords: parent-adolescent relationships; college transition; attachment measures; personal distress

Introduction Despite strong theoretical claims that the core function of attachment state of mind is to guide perception and interpretation of attachment-relevant information, and although it is assessed based on individuals’ discourse regarding their relationship with their parents, researchers generally agree that attachment state of mind is unrelated, or weakly related, to people’s perceptions of their relationships with their parents (Crowell, Treboux & Waters, 1999). Such findings could wrongfully be interpreted as contradictory to the conceptualization of attachment state of mind. Two main hypotheses have been proposed to explain this lack of association: the type of measurement utilized (Crowell, Fraley & Shaver, 1999) and the level of consciousness of the constructs tapped (Maier, Bernier, Pekrun, Zimmermann & Grossmann, 2004). Both of these hypotheses, which are supported by empirical research, are based on the view that defensive cognitive processes may interfere with attachment measures that are based on self-report only, and thus decrease associations with an interview-based measure based on discourse analysis. Empirical research on attachment state of mind and parent-adolescent interactions suggests a third explanation for the lack of association between attachment state of mind and perceptions of attachment relationships. Research has consistently shown that adolescent attachment state of mind is associated with parent-adolescent interaction Correspondence: Annie Bernier, Department of Psychology, University of Montreal, P.O. Box 6128, Downtown Station, Montreal, H3C 3J7, Canada. E-mail: [email protected] ISSN 1461-6734 print/ISSN 1469-2988 online ª 2005 Taylor & Francis Group Ltd DOI: 10.1080/14616730500147565

172

A. Bernier et al.

patterns (Allen & Hauser, 1996; Becker-Stoll & Fremmer-Bombik, 1997; Kobak, Cole, Ferenz-Gillies, Fleming & Gamble, 1993). For instance, adolescents with an autonomous state of mind engage with their parents in problem-solving discussions characterized by less dysfunctional anger and less avoidance (Kobak et al., 1993). This would appear to stand in sharp contrast with studies showing weak relations between state of mind and perceptions of the relationship with the parents. One explanation for these apparently discrepant findings is that studies addressing parent–adolescent interactions typically make observations in contexts that are likely to induce negative emotions or some mild form of personal distress, such as situations of conflict resolution. In contrast, studies that assess individuals’ perceptions of their relationships with their parents generally do so in neutral situations, such as during the completion of questionnaires at home or in a university lab. This suggests that the presence of some form of stress, perhaps leading to the activation of the attachment system, may be a necessary condition in order to observe state of mind differences in perceptions and behaviors related to the parent-adolescent relationship. One of attachment theory’s main assumptions is that attachment representations, and their related behaviors, are activated in times of personal distress (Bowlby, 1982). In line with Bowlby’s ideas, the importance of stress to reveal attachment differences is at the core of the procedures used to assess attachment patterns in infancy (Ainsworth, Blehar, Waters & Wall, 1978) and preschool (Cassidy & Marvin, 1992). In contrast, the stress-induced activation of the attachment system remains largely speculative in adolescence and adulthood. However, social psychology researchers have begun to find evidence for such activation. They found that subliminal presentation of threat-related words (e.g., failure, separation) increases the cognitive availability of representations of attachment figures, and that overall, anxious-ambivalent individuals have higher accessibility to attachment representations than secure or avoidant individuals (Mikulincer, Birnbaum, Woddis & Nachmias, 2000; Mikulincer, Gillath & Shaver, 2002). Related evidence has begun to emerge in developmental research on attachment state of mind and parenting. Studies by Adam, Gunnar and Tanaka (2004) and by Phelps, Belsky and Crnic (1998) have shown that non-optimal parenting (e.g., intrusiveness, lack of warmth) emerges among parents with insecure parental states of mind only when they also experience at least moderate levels of personal distress (stress, depression). According to Adam et al., such findings suggest that the correlates and consequences of attachment state of mind are more likely to appear under stressful conditions. The aim of the present study is to investigate this assumption in the context of the parent–adolescent relationship. This study uses a naturalistic separation occurring in late adolescence, that is, leaving home to attend college, as a context that may potentially activate adolescents’ attachment system, and thus magnify state-of-mind differences in their perceptions of their relationships with their parents. The experience of leaving home The transition to college is generally viewed as a positive step up in life involving many new opportunities for students. However, this transition is not without challenges, especially for those students who must leave home to attend college. This represents a marked break from previous routines and lifestyle, as well as adaptation to a completely new environment involving academic, social and residential challenges (Fisher & Hood, 1987). In a short period of time these adolescents leave home, move into an apartment or dormitory without adult supervision, learn to manage their own affairs and assume adult responsibilities, in

Attachment and parent-adolescent relationships

173

addition to having to adjust to changes in the composition and availability of their social network (Kenny, 1987; Larose & Boivin, 1998). Many students present marked reactions to this transition, namely homesickness. Homesickness is defined as ‘a complex cognitive-motivational-emotional state concerned with grieving for, yearning for and being preoccupied with thoughts of home’ (Fisher & Hood, 1987, p. 426). Fisher and Hood (1988) found homesickness to be associated with elevated levels of psychological distress (anxiety, phobias, somatic symptoms, depression) and cognitive failure (i.e., absent-mindedness). In line with this, Bloom (1987) suggests that although adolescent/parent separation is a normal part of life, it may be experienced by adolescents as a major loss. He notes that separation often triggers intense emotional reactions accompanied by attempts to re-attach to the lost person. Emotional reactions vary, but may include nostalgia, depression, anger, and guilt. Kenny and Donaldson (1991) further argue that the process of separation may activate the attachment system. In fact, Kenny (1987) conceptualizes leaving home as analogous to Ainsworth’s Strange Situation procedure. Larose and Boivin (1998) have provided some empirical support for the idea that the transition to college is especially stressful for those adolescents who leave home, and may thus activate their attachment system. They found a decline in perceived social support along with increases in anxiety and loneliness during the transition to college, but only for those students who had left home to attend college. This general increase in personal distress may have led to the activation of these students’ attachment system, given that they also reported more positive relationships with their parents after the transition to college, while no change occurred for those students who stayed home. Consistent with these findings, Sullivan and Sullivan (1980) found that adolescents who separate from their parents when entering college report increases in their parents’ affection toward them, in their own affection toward their parents, as well as in the quality of communication between their parents and themselves. No such changes were observed among commuter students. The present study One may thus argue that leaving home to attend college can activate adolescents’ attachment system, making this transition a useful context to test whether state of mind differences in one’s perception of the parent–adolescent relationship are more manifest when the attachment system is presumed to be activated. The purpose of this study is to test the hypothesis that attachment state of mind is associated with perceptions of the parent– adolescent relationship among those students who leave home to attend college, but not among those who live at home. We focus on four dimensions of adolescents’ perceptions of their relationships with their parents, which parallel some of the major individual differences that distinguish attachment groups in the AAI: memories of childhood experiences with each parent, the reported emotional quality of the current relationship with each parent, the frequency of current contact with each parent, and the overall stress that the parent– adolescent relationship and family life are causing to the adolescent. It is expected that autonomous students who are leaving home to attend college will have more positive memories of their childhood attachment experiences than their insecure (dismissing or preoccupied) peers, and will report better current relationships with their parents, more contacts with them, and less family-related stress. No attachment difference is expected among adolescents living at home. Given that a preoccupied state of mind may be characterized by anger toward one’s parents, while a dismissing state of mind is often accompanied by an idealization of one’s relationship with his or her parents, it is expected

174

A. Bernier et al.

that the differences between autonomous adolescents and preoccupied adolescents will be clearer than those between autonomous adolescents and their dismissing peers.

Method Participants and procedure Sixty-one Caucasian adolescents (30 males and 31 females), aged 16 to 17 years old, were randomly selected from a sample of 298 participants involved in a longitudinal study of adjustment to college (Larose & Boivin, 1998). Twenty-seven (44.3%) of these adolescents left home to attend college, while 34 lived with their parents throughout. Average parental income was between $50,000 and $60,000. Level of education averaged 12.6 years for the mothers and 13.1 years for the fathers. Most participants (82.5 %) came from intact twoparent households. Three meetings were scheduled with each participant; one at the end of high school (T1), one at the beginning of the first (fall) semester in college (T2), and the last one shortly before the Christmas break (T3). At Times 1 and 2, students completed self-report measures pertaining to their past and current relationship with their parents. The AAI, conducted at T3, was used to assess attachment state of mind. We judged it appropriate to conduct the AAI at T3 because of its high stability over periods ranging from one month to one year in adult populations (Benoit & Parker, 1994; Steele & Steele, 1994), and over a two-year period among late adolescents (Zimmermann & Becker-Stoll, 2002). Given this demonstrated stability, it is current practice to conduct the AAI following the assessment of the variables it is meant to be related to (e.g., Dozier, Stovall, Albus & Bates, 2001; Pederson, Gleason, Moran & Bento, 1998). Measures Attachment state of mind: The Adult Attachment Interview (AAI; George, Kaplan & Main, 1996). The AAI is a semi-structured interview focusing on childhood attachment experiences with one’s parents and on the integration of these experiences into a coherent view of self, parents and attachment relationships. Participants are asked to describe their relationship with their parents when they were young, to substantiate these descriptions with specific memories, to recall incidences of distress, and to conceptualize relationship influences. The interview is recorded and transcribed, and its content is rated according to Main and Goldwyn’s (1998) classification system. Each transcript is classified into one of the four following states of mind: autonomous, preoccupied, dismissing, or unresolved. The AAI has been shown to have excellent reliability, discriminant validity, and predictive validity (Bakermans-Kranenburg & van IJzendoorn, 1993; Crowell et al., 1996; Sagi et al., 1994). Two graduate students trained by Mary Main and Erik Hesse conducted the interviews, which were between 45 and 90 minutes long. The interviews took place in a quiet room provided by the college, and were conducted according to the protocol developed by George et al. (1996). The second author of this report, trained by Main and Hesse, rated each interview. Fifteen transcripts were independently rated by a second judge (J. Gagnon), also trained by Main and Hesse. Both raters were blind to other measures, as well as to the identity and gender of the participants. Inter-rater agreement with regards to the 3-group classification was 80 % with a kappa of .60. The agreement rate on the 4-way classification was the same given that none of the 15 interviews were rated Unresolved by either judge.

Attachment and parent-adolescent relationships

175

The three cases (20%) on which judges initially disagreed were conferenced in order to reach final classification. Attachment classifications were distributed as follows: 34 participants were classified as autonomous (55.7%), 17 as dismissing (27.9%), and 10 as preoccupied (16.4%). Only one participant was classified as unresolved. In line with the procedure proposed by Hesse (1999), this participant was assigned his secondary classification (preoccupied in this case). This distribution is consistent with those reported in normal adolescent samples (e.g., van IJzendoorn & Bakermans-Kranenburg, 1996). Memories of childhood experiences: Mother–father–peer scale (MFP; Epstein, 1983). Two subscales of the MFP were used to explore adolescents’ memories of their childhood experiences with each parent. The first subscale (Independence vs. Overprotection; 13 items) evaluates the extent to which individuals believe their parents encouraged them to be autonomous, to have confidence in themselves, and to develop their personal abilities. A low score on this subscale suggests that participants believe their parents overprotected them during childhood; worried a great deal about their health and security; and failed to help them become independent. The second subscale (Acceptance vs. Rejection; 10 items) evaluates the extent to which individuals believe that their parents communicated their love to them, accepted them, and appreciated them during their childhood. A low score suggests that subjects believe that they bothered their parents, were not wanted, and were a source of misfortune and disappointment. The construct and convergent validity, as well as the internal consistency of the MFP are excellent (Epstein, 1983). The internal consistency scores for the Independence–Mother, Independence–Father, Acceptance–Mother and Acceptance–Father subscales were .79, .76, .83 and .88 respectively at T1, and .79, .78, .84 and .88 at T2. Emotional quality of the current relationships: Inventory of parent and peer attachment (IPPA; Armsden & Greenberg, 1987, 1989). This 25-item questionnaire assesses individuals’ perceptions of their current relationship with their father and mother by exploring a constellation of emotional experiences such as trust, communication, and feelings of alienation and anger. The reliability and construct validity of the IPPA are well-established (Armsden & Greenberg, 1987). High test-retest reliability (r = .93) has been reported over a three-week period (Armsden & Greenberg, 1989). As recommended by Armsden and Greenberg (1989), a global score was obtained for each target figure by summing up all items. In this study, the internal consistency values of the scores pertaining to adolescents’ perceived relationship with their mothers and fathers were .78 and .75 at T1, and .77 and .73 at T2. Frequency of interactions. Adolescents’ perceptions of the frequency of their contacts with their mothers and fathers were assessed by two questions, one pertaining to each parent: ‘I am in contact with my father/mother (i.e., speak with him/her on the phone, in person or otherwise)’. . . 1- less than once a month; 2- between once a month and every other week; 3more than every other week but not weekly; 4- at least weekly, but not daily; 5- one or more times a day. The correlation between frequency of contacts with mother and frequency of contacts with father was .31 at T1 and .48 at T2. Family-related stress: Measure of perceptions of social support specific to worrisome events (MPSSE; Larose & Boivin, 1997). The MPSS-E assesses expectations of support with regards to specific worrisome events. In the first section of the questionnaire, participants evaluate the extent to which they worry about 36 life events related to late adolescence and college

176

A. Bernier et al.

transition. Only this first section was used in the current study. The content and ecological validity of the MPSS-E have been previously documented (Larose & Boivin, 1997). Five items that focused on stress related to parents and family members were selected (e.g., ‘I am concerned with the lack of support my parents provide with regards to my studies’; ‘I am concerned with conflicts in my family’). This subscale of family-related stress showed satisfying internal consistency (T1 a = .77 and T2 a = .71). Leaving-home experience The experience of leaving home was defined as a new and prolonged separation from the family that limited regular face-to-face contact. Participants must not have left their families before, indicating that leaving for college was their first real and permanent physical separation. In addition, they had to be living at least fifty miles away from home, either by themselves or with non-family members. Based on these criteria, forty-four percent of participants were rated as having left home during the transition.

Results Preliminary analyses In light of the high correlations between the IPPA and the MFP subscales (ranging from .73 to .81 for mother-T1; from .59 to .81 for mother-T2; from .54 to .77 for father-T1; and from .50 to .76 for father-T2), we submitted the scores pertaining to the mothers and those pertaining to the fathers to two principal component analyses each (T1 and T2), followed by VARIMAX rotations. Each rotation yielded a one-factor solution. The four resulting factor scores, representing 63% of the variance for the mother scores and 67% for the father scores at T1, and 61% of the variance for the mother scores and 69% for the father scores at T2, were used in the remainder of the analyses as global scores of the perceived mother– adolescent and father–adolescent relationships.1 We examined whether demographic variables (student gender and age; parents’ education, income, and marital status) were associated with any of the dependent variables. Results showed that parents with more years of education and a higher income had adolescents reporting more frequent contacts with them (r = .42 for the mother and .27 for the father) and (marginally) less family-related stress (r = 7 .21), at Time 2 only. Given the correlations (ranging from .27 to .58) between maternal and paternal income and education, we created an aggregate score representing the family’s socio-economic status (SES). For simplicity purposes, this SES index was used as a covariate in all subsequent analyses. Classification on the AAI was unrelated to the demographic and family variables mentioned above. Furthermore, state-of-mind classifications did not differ according to residential status, X2 = 0.38, ns, which supports the status of the AAI as independent from the immediate impact of stressful life events. Finally, we examined whether there were differences between students who were to leave home and those who were to stay at home in terms of their perceptions of the parent– adolescent relationship prior to the transition to college (T1). These analyses were carried out in order to rule out the possibility that students with less favorable perceptions of their relationship with their parents would be more likely to choose to leave home when entering college, which would obscure the nature of the association between the variables. There were no differences between the two groups of students in terms of their pre-transition

Attachment and parent-adolescent relationships

177

perceptions of their relationships with their mothers, F (1,60) = 0.05, ns, and fathers, F (1,60) = 0.02, ns. Main analyses The five dependent variables (perceptions of the relationship with the mother, perceptions of the relationship with the father, frequency of contact with mother, frequency of contact with father, and family-related stress) were each submitted to a 2 (residential status : living at home or leaving home) 6 3 (attachment state of mind: autonomous, dismissing and preoccupied) 6 2 (Time 1 and Time 2) repeated measure multivariate analysis of covariance (controlling for family SES). Two a priori contrasts were specified on attachment state of mind: autonomous vs. dismissing (contrast 1), and autonomous vs. preoccupied (contrast 2). In order to estimate effect sizes, eta squared statistics are provided for all significant contrasts. Means and standard deviations of the five dependent variables are presented in Table 1. Significant interactions are displayed in Figure 1. With regards to perceptions of the mother–adolescent relationship, the main AAI effect was significant, F(2,53) = 10.10, p 5 .001, and was qualified by a significant AAI 6 Residence interaction, F(2,53) = 3.97, p 5 .05. With respect to this AAI x Residence interaction, only Contrast 2 was found to be significant, t = 7 2.72, p 5 .01, Z2 = .12. Preoccupied students had more negative perceptions of their relationship with their mother than autonomous students, but this was true only among students who had left, or were going to leave, home (t = 6.39, p 5 .001 vs t = 1.27, ns). The main Residence effect was not significant, but there was a significant Time effect, F(1,54) = 4.10, p 5 .05. Overall, students had more positive perceptions of their relationship with their mother after the transition to college than before. With regards to perceptions of the father–adolescent relationship, the main AAI effect was significant, F (2,53) = 3.12 , p 5 .05, and was qualified by a significant AAI 6 Residence 6 Table I Means (and standard deviations) of the five dependent variables according to time of measurement, residential status, and attachment state of mind High school (T1) Leaving home

Living at home

College (T2) Left home

Living at home

F Ds E F Ds E F Ds E F Ds E (n = 16) (n = 7) (n = 4) (n = 18) (n = 10) (n = 6) (n = 16) (n = 7) (n = 4) (n = 18) (n = 10) (n = 6) Relationship .39 7.62 73.30 with mother (1.23) (1.27) (1.04) Relationship .17 .24 71.41 with father (.97) (.80) (.90) Contact with 5.00 5.00 4.00 mother (.00) (.00) (2.00) Contact with 5.00 4.43 3.00 father (.00) (1.51) (2.31) Family-related 2.35 2.51 5.35 stress (1.35) (1.09) (1.59)

7.37 7.04 71.05 .51 7.30 72.73 (1.44) (1.39) (2.09) (.86) (.71) (.57) 7.21 .15 7.50 .18 .46 71.06 (1.19) (1.13) (.92) (.87) (.83) (.58) 5.00 5.00 4.50 1.75 2.00 2.50 (.00) (.00) (1.23) (1.24) (1.53) (1.73) 4.56 4.30 5.00 1.36 2.00 3.25 (1.29) (1.49) (.00) (.63) (1.53) (2.06) 1.78 1.84 2.13 1.56 1.60 4.41 (.75) (.77) (1.30) (.55) (.48) (1.68)

.09 7.36 (1.42) (1.12) 7.06 7.31 (1.31) (1.03) 5.00 4.20 (.00) (1.69) 4.39 3.90 (1.42) (1.79) 1.49 1.78 (.45) (.67)

7.85 (1.61) 7.81 (1.11) 3.67 (2.07) 4.33 (1.63) 2.63 (1.49)

Note: F = Autonomous; Ds = Dismissing; E = Preoccupied. Relationship with mother = Perceived quality of the relationship with the mother, factor score; Relationship with father = Perceived quality of the relationship with the father, factor score; Contact with mother = Frequency of contact with the mother; Contact with father = Frequency of contact with the father.

178

A. Bernier et al.

Figure 1. Attachment state of mind differences according to residential status and time of measurement.

Time interaction, F(2,53) = 3.65, p 5 .05. Contrast 1 was significant on this triple interaction, t = 2.41, p 5 .05, Z2 = .10, whereas a marginal trend was found for Contrast 2, t = 1.80, p 5 .08, Z2 = .06. The results of Contrast 1 were attributable to opposite nonsignificant trends, suggesting the following pattern: dismissing students tended to have similar or more positive perceptions of their relationship with their father than autonomous students in all cases, except after the transition for the group of students who stayed at home. Among them, dismissing students tended to have more negative perceptions of their relationship with their father than their autonomous counterparts. With respect to Contrast 2, preoccupied students who had left, or were going to leave home, had more negative perceptions of their relationship with their father than their autonomous counterparts, and this difference was more pronounced before they left home (T1: t = 3.58, p 5 .002; o2 = 0.37) than after they left (T2: t = 2.35, p 5 .03; o2 = 0.18). In contrast, no significant difference was observed between autonomous and preoccupied students who lived at home, either at Time 1 (t = .83, ns) or at Time 2 (t = 1.23, ns). No other main effects or interaction effects were found to be significant. With respect to the frequency of contacts with the mother, the main Residence effect was found significant, F(1,54) = 15.66, p 5 .001, and was qualified by a significant AAI 6 Residence 6 Time interaction, F(2,55) = 4.29, p 5 .05. This triple interaction was

Attachment and parent-adolescent relationships

179

explained by Contrast 2, t = 2.85, p 5 .01, Z2 = .13. In all cases, autonomous students reported having more frequent contact with their mothers than their preoccupied peers (all ts 5 1.82, p 5 .05), except after the transition among students who had left home. Among them, preoccupied students tended to report having more contacts with their mothers than their autonomous peers, although this difference was non-significant (t = 7 1.01, ns). There was also a significant main Time effect, F(1,55) = 73.76, p 5 .001, which was qualified by a significant Residence 6 Time interaction, F(1,55) = 24.06, p 5 .001. This interaction, in turn, was further qualified by the significant AAI 6 Residence 6 Time interaction discussed above. Regarding the frequency of contacts with the father, the main Residence effect was found significant, F(1,52) = 12.32, p 5 .001, as well as the Residence 6 Time interaction, F(1,53) = 15.10, p 5 .001, and the AAI 6 Time interaction, F(2,53) = 6.00, p 5 .01. In turn, these two interactions were qualified by a significant AAI 6 Residence 6 Time interaction, F(2,53) = 10.16, p 5 .001. This triple interaction was explained by Contrast 2, t = 4.48, p 5 .001, Z2 = .27. Autonomous adolescents who were going to leave home reported having more frequent contact with their fathers than their preoccupied peers (T1: t = 3.80, p 5 .001), whereas the opposite was true once in college; preoccupied students who had left home reported more contacts with their fathers than their autonomous counterparts (T2: t = 7 2.20, p 5 .05). However, as shown in Figure 1, this latter result is explained by a drop in the frequency of contact with father in autonomous adolescents; preoccupied students maintaining a rather constant rate of frequency of contact. No significant differences were found among students who stayed at home (T1: t = 7 .83, ns and T2: t = .08, ns). No other main or interaction effects were found to be significant. Finally, with respect to family-related stress, all three main effects were found significant: AAI, F(2,54) = 24.88, p 5 .001; Residence, F(1,54) = 11.51, p 5 .001; and Time, F(1,55) = 4.95, p 5 .05. The AAI and Residence effects were qualified by an AAI 6 Residence interaction effect, F(2,54) = 8.93, p 5 .001. This interaction was explained by Contrast 2, t = 4.08, p 5 .001, Z2 = .24. Hence, although preoccupied students reported more stress related to family matters than their autonomous peers overall, this difference was more pronounced among those students who had left, or were going to leave, home (t = 7 5.83, p 5 .001; o2 = 0.62) than among those who lived with their parents (t = 7 2.10, p 5 .05; o2 = 0.15). The main Residence and Time effects were qualified by a Residence 6 Time interaction effect, F(1,55) = 6.20, p 5 .05. At Time 1, those adolescents who were going to leave home reported more family-related stress than those who were going to stay at home (t = 7 3.12, p 5 .01), whereas no such difference was observed at Time 2, (t = 7 .80, ns). The AAI 6 Time, and AAI 6 Residence 6 Time interactions were not significant. To sum up, the findings revealed a significant AAI 6 Residence interaction effect on all five dependent variables. This indicates that the link between attachment state of mind and perceived parent–adolescent relationships was moderated by the experience of leaving home. This link was either reversed, greater, or significant only, among those students who leave home, which suggests that leaving home magnifies attachment differences. Furthermore, three of the five AAI 6 Residence interactions were qualified by a triple AAI 6 Residence 6 Time interaction, whereas two were not. This suggests that anticipating leaving home may have a somewhat similar impact on adolescents’ perceptions of their relationships with their parents as the actual experience of leaving. Lastly, it is noteworthy that nearly all significant contrasts involved the autonomous vs. preoccupied distinction, whereas very few convincing findings were found with regards to the autonomous vs. dismissing distinction.

180

A. Bernier et al.

Because the clearest findings of this study were attributable to a group of only 10 preoccupied students compared to a group of 34 autonomous students, one might speculate that a statistical artefact contributed to the significant contrasts. It is arguable that the scores obtained by the 34 autonomous students were more stable (more reliable) than those obtained by their 10 preoccupied peers, which could be subject to sampling error (especially when broken down according to residential status). We thus randomly selected subsamples of 10 autonomous and 10 dismissing students in order to re-run all analyses with more equal cell sizes. This random selection yielded the following distribution: 5 autonomous students who left home and 5 who stayed at home, 6 dismissing students who left home and 4 who stayed at home, to be compared with the 4 preoccupied students who left home and the 6 who stayed at home. The scores of these 30 students on the five dependent variables were submitted to the same analyses as described above for the whole sample.2 Despite the inevitable loss of statistical power inherent to the smaller sample size, the double or triple interactions involving the AAI and residential status either remained significant (for quality of relationship with father, frequency of contact with mother, and frequency of contact with father), or became non-significant trends that followed the same pattern as before (quality of relationship with mother and family-related stress). Overall, the findings involving smaller but more even cell sizes are similar in shape to the results pertaining to the whole sample, but some suffer from the expected decrease in statistical significance due to the smaller sample size. While suggesting that the results pertaining to quality of the relationship with the mother and family-related stress may be less robust and thus require replication, these analyses do rule out the hypothesis that the overall pattern of findings is attributable to unequal cell sizes and sampling error with the preoccupied students. Discussion The aim of this study was to test the hypothesis that attachment state of mind is associated with perceptions of the parent–adolescent relationship among those students who leave home to attend college, but not among those who remain at home. The results supported this hypothesis for the autonomous vs. preoccupied distinction. Preoccupied students who had left or anticipated leaving home reported different perceptions and behaviors with respect to their relationship with their parents than their autonomous peers. In contrast, few clear attachment differences were observed among students who stayed at home while attending college. In fact, all main AAI effects were qualified either by an AAI 6 Residence interaction, or by an AAI 6 Residence 6 Time interaction. This suggests that individual differences related to attachment state of mind were magnified by the actual or anticipated experience of leaving home. The findings pertaining to the preoccupied state of mind follow a pattern that is strikingly consistent with preoccupied individuals’ discourse in the AAI. Our results showed that preoccupied students who left home reported having a more negative relationship with their father, a more negative relationship with their mother, and experiencing more stress related to their family than their autonomous peers, yet they also reported having more contact with each parent. Such a blend of negative perceptions of the relationship with the parents, together with increased contact, closely resembles some preoccupied individuals’ discourse in the AAI. These individuals show anger and give very negative evaluations of their relationship with their parents, yet are enmeshed in the relationship to a point where they have difficulty stepping back to take a healthy distance from their parents. For instance, some individuals will report an extremely difficult relationship with a parent, and will go on

Attachment and parent-adolescent relationships

181

to explain in great detail how they spent an hour on the phone arguing with that parent the night before. The fact that a parallel pattern was observed among preoccupied students who left home, but not among those who stayed home, is consistent with the hypothesis that the experience of leaving home, whether actual or anticipated, is sufficiently stressful to activate students’ attachment system (for preoccupied students at least). Using a developmental attachment measure, the present report thus provides a rare observation of the normative activation of the attachment system, which adds to similar demonstrations previously made by social psychologists (e.g., Mikulincer et al., 2000; 2002). It also supports Adam et al.’s (2004) proposition that individual differences related to attachment state of mind are magnified, and thus more easily observed, under conditions of stress. The fact that most attachment differences were already present in high school, before the actual physical separation occurred, adds an interesting dimension to the results. This finding suggests that, even for late adolescents, anticipating an upcoming separation from one’s parents and current environment might generate nearly as much stress as does the actual experience of being away from home and having to cope with a host of new personal, social and academic challenges. These results are also consistent with classic (Kobak & Sceery, 1988) and recent (Creasey, 2002) evidence suggesting that preoccupied adolescents experience a good deal of psychological distress in college. According to Creasey (2002), such findings can be explained by Allen and Land’s (1999) proposition that a major development task of late adolescence is learning to explore and to live as an emotionally autonomous individual. Given that enmeshment in the relationship with the parents is a defining feature of preoccupied attachment, one might expect the period of late adolescence to be a particular challenge for preoccupied adolescents. In fact, Creasey (2002) found that the link between preoccupied attachment and psychological distress was attributable to the anger/father scale: college women expressing more involving anger toward their father during the AAI reported more psychological symptoms. Anecdotally, it may be worthy of note that our data might suggest a parallel pattern: among the four preoccupied students who left home, three were classified as E2 (characterized by involving anger), and one as E1 (characterized by passivity of discourse). In contrast, among the six preoccupied students who remained at home, four were classified as E1, and two as E2. Although this difference is statistically non-significant, it raises the possibility that late adolescents who are classified as preoccupied because they show involving anger during the AAI may be those who have the most negative perceptions of their relationship with their parents, and might leave home more quickly when given the opportunity. In contrast to the numerous clear findings differentiating autonomous and preoccupied students, no convincing difference between the dismissing and autonomous groups was found. Although contrary to our a priori expectations, this is consistent with dismissing individuals’ discourse in the AAI. They often provide fairly positive descriptions of their relationship with their parents, not much unlike the descriptions provided by autonomous individuals. However, they fail to substantiate these global evaluations with specific examples of warm, supportive exchanges with their parents. One may argue that a similar pattern of idealization is active when dismissing individuals are asked to report on their relationship with their parents by means of a paper-and-pencil questionnaire, thus obscuring the potential distinction between dismissing and autonomous adolescents. In order to overcome this limitation, future research should use external reports (from parents, peers, etc.) of dismissing individuals’ behaviors with regards to their parents in stressful situations. Previous research has shown that such external reports may provide information that would

182

A. Bernier et al.

have been lost through the sole use of dismissing individuals’ self-reports (e.g., Dozier & Lee, 1995; Kobak & Sceery, 1988; Larose & Bernier, 2001). The one difference that was found significant between these two groups, however, is worthy of mention. We found that dismissing students had similar or more positive perceptions of their relationship with their fathers than autonomous students, except after the transition where dismissing students who were still living at home tended to have more negative perceptions of their relationship with their fathers than their autonomous peers. One possible explanation for this finding is that the mechanism of idealization, which may partially account for the relatively elevated scores obtained by dismissing students in most cases, collapsed among dismissing students who lived at home during college. One may argue that the experience of still living with one’s parents at a time when personal autonomy is expected, and the transition to adulthood imminent, is particularly threatening to dismissing adolescents’ valuing of autonomy, personal strength, and independence from parents. This may pose a stress sufficient to challenge these students’ strategy of idealization, thereby forcing them to face certain aspects of the relationship with their parents which are unsatisfying to them. However, this pattern was not found to be true of students’ perceived relationship with their mother, and was attributable to opposite non-significant trends. One must therefore be very careful in interpreting this result, which may be a random finding that would not be replicated with a different sample. This study presents a number of limitations that call for caution in interpreting and generalizing the results. In terms of methodology, it should be noted that two of the five dependent variables (frequency of contact with each parent) were assessed through one item each, as opposed to multiple-item subscales. These two scores therefore cannot be submitted to standard tests of reliability, such as internal consistency. The moderate correlations found at both times between frequency of contact with mother and with father are sensible, but cannot substitute standard reliability indices. However, the fact that these two constructs revealed patterns of findings that complemented, in theoretically-consistent ways, those found with the other dependent variables, provides some confidence in the results. In addition, the mean levels of these two frequency variables showed the expected patterns, i.e., Residence 6 Time interactions, suggesting that regardless of attachment state of mind, the lowest frequency of contact with each parent was reported at Time 2 by adolescents who had left home. This measure would also have provided richer information if we had assessed not only the frequency of contacts, but some qualitative aspects thereof: who was the initiator of the contacts, the adolescent’s thoughts and feelings about these contacts (i.e., desired vs. undesired contact, level of resistance, etc.), the nature of the contacts, the reason for the contacts, etc. Such information could be beneficial to further our understanding of attachment state of mind differences in parent-adolescent current relationships. A conceptual issue with the present study is the inferential leap we make in speculating that the greater attachment differences found among students who leave home suggests that the experience of leaving home activates the attachment system. Although the current findings are consistent with such an assumption, and even if a similar conclusion was suggested by the Larose and Boivin (1998) study, it has not been observed per se. The assumption could be further corroborated by external observations. For instance, peers’ or parents’ reports to the effect that students who leave home show increases in attachment behaviors (e.g., proximity-seeking, need for reassurance, etc.) would strengthen the conclusion that leaving home for the first time is sufficiently stressful to activate the attachment system. However, regardless of the potential activation of the attachment system, the findings presented in this report support the hypothesis that leaving home, or

Attachment and parent-adolescent relationships

183

anticipating this event, magnifies attachment-related individual differences, at least with respects to the parent–adolescent relationship. This study has suggested that attachment state of mind is related to perceived parentadolescent relationships to a greater degree among adolescents who leave home to attend college than among those who live at home while in college. Specifically, preoccupied students who had left, or were going to leave home, appraised their relationships with their mothers and fathers more negatively, experienced more family-related stress, and yet reported more frequent contacts with each parent than their autonomous counterparts. In contrast, few attachment differences were found among those adolescents who lived at home while attending college. These findings add to previous results (Adam et al., 2004; Phelps et al., 1998) in supporting Bowlby’s (1982) proposition that stress activates (and thus magnifies) individual differences in attachment organization. These three recent studies thus suggest that the stress-induced activation of the attachment system, which is welldemonstrated in infancy and childhood, also appears to be true among adults and adolescents. Directions for further research include a more detailed examination of the types of conditions that activate the attachment system after childhood, the level of personal distress necessary to observe such an activation, and individual differences in vulnerability to attachment triggers. Acknowledgements We gratefully acknowledge the help of Jocelyne Gagnon and Roland Roy in collecting the data. Also, our appreciation to Genevie`ve A. Mageau, Chantal Mongeau, and Robert Haccoun for their help with different aspects of this project. Notes 1. We also ran all main analyses separately for childhood memories (MFP) and current relationships (IPPA). The pattern of findings was exactly the same as that reported here, both for mother and for father. For the sake of parsimony, we chose to drop the conceptual distinction between past and current relationships, and to report results with aggregated scores instead. 2. Full description of the results is available from the first author upon request.

References Adam, E. K., Gunnar, M. R., & Tanaka, A. (2004). Adult Attachment, Parent Emotion, and Observed Parenting Behavior: Mediator and Moderator Models. Child Development, 75, 110 – 122. Ainsworth, M. D., Blehar, M. C., Waters, E., & Wall, S. (1978). Patterns of attachment: A psychological study of the strange situation. Hillsdale, NJ: Erlbaum. Allen, J. P., & Hauser, S. T. (1996). Autonomy and relatedness in adolescent-family interactions as predictors of young adults’ states of mind regarding attachment. Development and Psychopathology, 8, 793 – 809. Allen, J. P. & Land, D. (1999). Attachment in adolescence. In J. Cassidy & P. R. Shaver (Eds.), Handbook of Attachment: Theory, research and clinical applications (pp. 319 – 335). New York: Guilford. Armsden, G. C., & Greenberg, M. T. (1987). The inventory of parent and peer attachment: Individual differences and their relationship to psychological well-being in adolescence. Journal of Youth and Adolescence, 16, 427 – 454. Armsden, G. C., & Greenberg, M. T. (1989). Inventory of parent and peer attachment: Preliminary test manual. Available from G.C. Department of Community Health Care Systems, University of Washington, Seattle. Bakermans-Kranenburg, M. J., & van IJzendoorn, M. H. (1993). A psychometric study of the adult attachment interview: Reliability and discriminant validity. Developmental Psychology, 29, 870 – 879. Becker-Stoll, F., & Fremmer-Bombik, E. (April 1997). Adolescent-Mother interaction and attachment: A longitudinal study. Poster presented at the SRCD Biennial Meeting, Washington, DC.

184

A. Bernier et al.

Benoit, D., & Parker, K. C. H. (1994). Stability and transmission of attachment across three generations. Child Development, 65, 1444 – 1456. Bloom, M. V. (1987). Leaving home: A family transition. In Bloom-Feshbach, J., Bloom-Feshbach, S., & associates (Eds.), The psychology of separation and loss: Perspectives on development, life transitions, and clinical practice (pp. 232 – 266). San Francisco, California: Jossey-Bass Publishers. Bowlby, J. (1982). Attachment and loss, vol. 1. attachment (2nd ed.). New York: Basic Books. Cassidy, J., & Marvin, R. S. (with the MacArthur Working Group on Attachment) (1992). Attachment organization in 2 1/2 to 4 1/2 years olds: Coding Manual. Unpublished coding manual, University of Virginia. Creasey, G. (2002). Psychological distress in college-aged women: Links with unresolved/preoccupied attachment status and the mediating role of negative mood regulation expectancies. Attachment and Human Development, 4, 261 – 277. Crowell, J. A., Fraley, R. C., & Shaver, P. R. (1999). Measurement of individual differences in adolescent and adult attachment. In J. Cassidy & P.R. Shaver (Eds.), Handbook of attachment: Theory, research and clinical applications (pp.434 – 465). New York: Guilford. Crowell, J. A., Treboux, D., & Waters, E. (1999). The Adult Attachment Interview and the Relationship Questionnaire: Relations to reports of mothers and partners. Personal Relationships, 6, 1 – 18. Crowell, J. A., Waters, E., Treboux, D., O’Connor, E., Colon-Downs, C., Feider, O., Golby, B., & Posada, G. (1996). Discriminant validity of the Adult Attachment Interview. Child Development, 67, 2584 – 2599. Dozier, M., & Lee, S. (1995). Discrepancies between self- and other-report of psychiatric symptomatology: Effects of dismissing attachment strategies. Development and Psychopathology, 7, 217 – 226. Dozier, M., Stovall, K. C., Albus, K. E., & Bates, B. (2001). Attachment for infants in foster care: The role of caregiver state of mind. Child Development, 72, 1467 – 1477. Epstein, S. (1983). The mother-father-peer scale. Unpublished manuscript, University of Massachusetts, Amherst. Fisher, S., & Hood, B. (1987). The stress of the transition to university: A longitudinal study of psychological disturbance, absent-mindedness and vulnerability to homesickness. British Journal of Psychology, 78, 425 – 441. Fisher, S., & Hood, B. (1988). Vulnerability factors in the transition to university: Self-reported mobility history and sex differences as factors in psychological disturbance. British Journal of Psychology, 79, 309 – 320. George, C., Kaplan, N., & Main, M. (1996). Adult attachment interview protocol (3rd Ed.). Unpublished manuscript, University of California at Berkeley. Hesse, E. (1999). The Adult Attachment Interview: Historical and current perspectives. In J. Cassidy & P. R. Shaver (Eds.), Handbook of attachment: Theory, research and clinical applications (pp. 395 – 433). New York: Guilford. Kenny, M. E. (1987). The extent and function of parental attachment among first-year college students. Journal of Youth and Adolescence, 16, 17 – 29. Kenny, M. E., & Donaldson, G. A. (1991). Contributions of parental attachment and family structure to the social and psychological functioning of first-year college students. Journal of Counseling Psychology, 38, 479 – 486. Kobak, R. R., Cole, H. E., Ferenz-Gillies, R., Fleming, W. S., & Gamble, W. (1993). Attachment and emotion regulation during mother-teen problem solving: A control theory analysis. Child Development, 64, 231 – 245. Kobak, R. R., & Sceery, A. (1988). Attachment in late adolescence: Working models, affect regulation, and representation of self and others. Child Development, 59, 135 – 146. Larose, S., & Bernier, A. (2001). Social support processes: Mediators of attachment state of mind and adjustment in late adolescence. Attachment and Human Development, 3, 96 – 120. Larose, S., & Boivin, M. (1997). Structural relations among attachment working models of parents, general and specific support expectations, and personal adjustment in late adolescence. Journal of Social and Personal Relationships, 14, 579 – 601. Larose, S., & Boivin, M. (1998). Attachment to parents, social support expectations, and socioemotional adjustment during the high school-college transition. Journal of Research on Adolescence, 8, 1 – 27. Maier, M. A., Bernier, A., Pekrun, R., Zimmermann, P., & Grossmann, K. E. (2004). Attachment Working Models as Unconscious Structures: An Experimental Test. International Journal of Behavioral Development, 28, 180 – 189. Main, M., & Goldwyn, R. (1998). Adult attachment classification system, draft 6.2. Unpublished manuscript, University of California at Berkeley. Main, M., Kaplan, N., & Cassidy, J. (1985). Security in infancy, childhood, and adulthood: A move to the level of representation. In I. Bretherton & E. Waters (Eds.), Growing points of attachment theory and research. Monographs of the Society for Research in Child Development, 50 (Serial no. 209; pp. 66 – 104).

Attachment and parent-adolescent relationships

185

Mikulincer, M., Birnbaum, G., Woddis, D., & Nachmias, O. (2000). Stress and accessibility of proximity-related thoughts: Exploring the normative and intraindividual components of attachment theory. Journal of Personality and Social Psychology, 78, 509 – 523. Mikulincer, M., Gillath, O., & Shaver, P. R. (2002). Activation of the attachment system in adulthood: Threatrelated primes increase the accessibility of mental representations of attachment figures. Journal of Personality and Social Psychology, 83, 881 – 895. Pederson, D. R., Gleason, K. E., Moran, G., & Bento, S. (1998). Maternal attachment representations, maternal sensitivity, and the infant-mother attachment relationship. Developmental Psychology, 34, 925 – 933. Phelps, J. L., Belsky, J., & Crnic, K. (1998). Earned security, daily stress, and parenting: A comparison of five alternative models. Development and Psychopathology, 10, 21 – 38. Sagi, A., van IJzendoorn, M. H., Scharf, M., Koren-Karie, N., Joels, T., & Mayseless, O. (1994). Stability and discriminant validity of the Adult Attachment Interview: A psychometric study in young Israeli adults. Developmental Psychology, 30, 771 – 777. Steele, H., & Steele, M. (1994). Intergenerational patterns of attachement. In D. Perlman & K. Bartholomew (Eds.), Advances in personal relationships, vol. 5. London: Kinsley. Sullivan, K., & Sullivan, A. (1980). Adolescent-parent separation. Developmental Psychology, 16, 93 – 99. van IJzendoorn, M. H. (1995). Adult attachment representations, parental responsiveness, and infant attachment: A meta-analysis on the predictive validity of the Adult Attachment Interview. Psychological Bulletin, 117, 1 – 17. Van IJzendoorn, M. H., & Bakermans-Kranenburg, M. J. (1996). Attachment representations in mothers, fathers, adolescents, and clinical groups: A meta-analytic search for normative data. Journal of Consulting and Clinical Psychology, 64, 8 – 21. Zimmermann, P., & Becker-Stoll, F. (2002) Stability of attachment representations during adolescence: The influence of ego-identity status. Journal of Adolescence, 25, 107 – 124.