Attachment as a moderator of the effect of security in ... - SAGE Journals

9 downloads 0 Views 119KB Size Report
quality with college teachers. Simon Larose. Laval University. Annie Bernier & Nathalie Soucy. University of Montreal. ABSTRACT. The purpose of this study was ...
06 larose (ds)

18/5/05

3:23 pm

Page 399

Attachment as a moderator of the effect of security in mentoring on subsequent perceptions of mentoring and relationship quality with college teachers Simon Larose Laval University

Annie Bernier & Nathalie Soucy University of Montreal

ABSTRACT

The purpose of this study was to examine if college students’ attachment insecurity, as evaluated by the Adult Attachment Interview, moderates the effect of affective security in mentoring on subsequent perceptions of the mentoring program and relationship quality with other teachers. Academically at-risk students were involved in a 10-hour mentoring program and completed measures at three points in time. Security in mentoring was associated with a subsequent positive perception of mentoring and with low conflict with teachers, although not with supportive relationships with teachers. As expected, these associations were moderated by attachment insecurity. Security in mentoring was positively related to subsequent perceptions of mentoring only for students showing low preoccupation with attachment, and inversely related to conflict with teachers only for students showing high dismissing attachment tendencies. KEY WORDS:

attachment • mentor • mentoring • moderator variables • protégé • student–teacher relationships

The research described in this article was supported by grants from Le Fonds FCAR (99NC-1705) and from the Social Sciences and Humanities Research Council of Canada (410–98–0892). All correspondence concerning this article should be addressed to Simon Larose, Département d’études sur l’enseignement et l’apprentissage, Faculté des Sciences de l’Éducation, Université Laval, Québec, G1K 7P4, Canada [e-mail: simon.larose @fse.ulaval.ca]. Stanley O. Gaines was the Action Editor on this article. Journal of Social and Personal Relationships Copyright © 2005 SAGE Publications (www.sagepublications.com), Vol. 22(3): 399–415. DOI: 10.1177/0265407505052443

06 larose (ds)

400

18/5/05

3:23 pm

Page 400

Journal of Social and Personal Relationships 22(3)

Empirical research has shown that academic mentoring with a teacher has a positive impact on college students’ adjustment (see Jacobi, 1991 for a review). Very few studies, however, have examined the personal factors that may explain or moderate the positive impact of mentoring. Identification of such factors would allow for a better understanding of individual differences in students’ responses to mentoring, thus providing empirical guidelines that could lead to further improvement of mentoring programs. One possible moderator of the effect of mentoring on students’ adjustment is attachment dispositions. However, attachment researchers have not paid much attention to academic mentoring relationships, despite some functional similarities between mentoring and attachment relationships. Academic mentoring involves interactions with an older adult who serves as a facilitator of the student’s exploration of his or her social and academic environment and as a source of support to cope with the stress of adjusting to a new school. These two functions parallel those of secure base and haven of safety, which characterize an attachment relationship (Feeney & Noller, 1996). Thus, although it does not constitute an attachment relationship, the student–mentor relationship might activate students’ attachment dispositions and, indirectly, influence the benefits that they derive from this relationship (Slade, 1999). The main goal of this study was to investigate, among academically at-risk students, the role of attachment insecurity as a moderator of the associations between affective security in academic mentoring, subsequent perceptions of mentoring, and relationship quality with college teachers. Mentoring and students’ adjustment Research has suggested that informal contacts (outside the classroom) between college students and faculty have a positive impact on students’ academic performance, satisfaction with college life, retention, and educational and career goals (Lamport, 1993). Drawing from these findings, several colleges have developed academic mentoring programs (e.g., the Freshman Mentoring Program at Brewton-Parker College; Wilson, 1994) in an effort to prevent the adjustment problems typically associated with the transition from high school to college. These programs generally match a college teacher with an at-risk freshman and are aimed at providing the student with skills and strategies for dealing with the typical stressful events of the transition (e.g., academic failure or social network disruption). Studies have shown that this type of program improves study skills, motivation, academic adjustment, and personal adjustment (Jacobi, 1991). Factors proposed to explain this positive impact include the feedback provided to students on their coping strategies and the reinforcement of their personal value at a time when it may be severely threatened (Rhodes, Grossman, & Resche, 2000). Affective security in mentoring relationships Another likely explanation for the capacity of mentoring to enhance students’ adjustment is the degree of mutuality, trust, and empathy that

06 larose (ds)

18/5/05

3:23 pm

Page 401

Larose et al.: Attachment and college mentoring

401

develops between the mentor and the student (Rhodes, 2002). Mentoring research has shown that interpersonal factors closely related to these constructs, such as trust and perceived support from mentor, are associated with student adjustment (Jacobi, 1991; Talmi, 1997). Furthermore, perceived security with the teacher mentor predicts college adjustment above and beyond perceived security with mother and father (Soucy & Larose, 2000). Perceptions of security developed during a mentoring program may, therefore, be an important determinant of the student’s subsequent adjustment. Quality of relationships with college teachers A unique aspect of the current study is its focus on the quality of relationships with college teachers as an outcome variable. Several studies have shown that academically at-risk students have negative perceptions of relationships with teachers. They more frequently perceive these relationships as conflictual (Long & Morse, 1996), more often report that they are treated unfairly (Nelson, Epstein, Bursuck, Jayanthi, & Sawer, 1998), and have more difficulty seeking help from teachers when they are in a situation of failure, compared with other students (Richman, Rosenfeld, & Bowen, 1998). Academically at-risk students therefore appear to enter mentoring relationships with rather negative histories of interactions with teachers. From this perspective, improving students’ perceptions of teacher–student relationships becomes an important objective of mentoring, one that might favor students’ subsequent adjustment and academic success. The first objective of this study was to examine the associations between students’ perceived security in mentoring, on the one hand, and their subsequent perceptions of their relationship with their mentor as well as with other teachers, on the other hand. Five characteristics of students’ interactions with teachers were examined: their perceptions of conflict and unfairness, their comfort in disclosing personal information and in seeking help, and their level of informal contacts with teachers outside the classroom. These characteristics were chosen because they capture the two dimensions of support and conflict that are proposed to be central in understanding student–teacher interpersonal relationships (Wubbels & Levy, 1993). The moderating role of students’ attachment insecurity One of attachment theory’s basic tenets is that an individual’s attachment history is progressively internalized through the development of working models of attachment (WMA; Bowlby, 1988). In adolescence and adulthood, WMA appear as attachment styles (Bartholomew, 1990; Hazan & Shaver, 1987) or states of mind regarding childhood experiences (Main, Kaplan, & Cassidy, 1985), which reveal the behavioral and cognitive strategies used by the individual to process attachment-relevant information. One secure (secure or autonomous) and three insecure (dismissing/avoidant, preoccupied/anxious–ambivalent, and unresolved/fearful) patterns of attachment have been identified by attachment researchers.

06 larose (ds)

402

18/5/05

3:23 pm

Page 402

Journal of Social and Personal Relationships 22(3)

In this study, WMA were assessed using the Adult Attachment Interview (AAI), which focuses specifically on the current integration of early experiences with parents. Based on this interview, four WMA are identified: autonomous attachment (secure) is characterized by an open and coherent discourse about childhood experiences with the parents. In contrast, dismissing attachment (insecure) is recognized by the minimization of the importance of attachment relationships, often combined with idealization of the parents. Preoccupied attachment (insecure) is evidenced by current enmeshment with the parents, indicated by an angry and confused, or vague and rambling, style of discourse. Finally, unresolved attachment (insecure) is characterized by lapses in reasoning or discourse while recounting a loss or a trauma (Main & Goldwyn, 1998). Unresolved attachment is recognized in certain parts of the interview only, which can otherwise be identified as dismissing, preoccupied, or autonomous. Attachment theory suggests that insecure attachments may have a negative impact on the development of mentoring relationships (Bowlby, 1988). Bowlby (1988) proposes that insecure attachment leads to negative relational expectations, characterized by mistrust or uncertainty as well as by a negative evaluation bias for the appraisal of others’ attempts to offer help. Such negative predispositions would interfere with the tendency to turn to others for assistance, and the resulting distortions in the helpseeking process would take different forms depending on the nature of the individual’s insecurity (Slade, 1999). Specifically, individuals presenting a preoccupied attachment pattern are believed to dwell on their own distress, increasing their need for help to such an extent that they would perceive support as unavailable or insufficient to meet their needs (Kobak & Sceery, 1988). Such perceptions might lead students in mentoring to seek the mentor’s help in an overwhelming and inappropriate manner. In contrast, individuals presenting a dismissing attachment pattern would be guided by mechanisms that restrict acknowledgment of distress and limit the associated attempts to seek comfort and support. To resolve their problems, they would revert to ‘self-reliance patterns’ (Bowlby, 1988), which may lead students in mentoring to disclose few personal issues and to hold back from emotional involvement. Consistent with these theoretical arguments, attachment insecurity, assessed with the AAI, has been found to predict difficulties in establishing new relationships with teachers and counselors. Specifically, individuals presenting high dismissing attachment tendencies reported difficulties in seeking help from teachers and low levels of trust in potential supporters (Larose & Bernier, 2001). They were evaluated by treatment providers as disclosing few personal issues, as rejecting of the treatment (Dozier, 1990), and as less committed than others (Korfmacher, Adam, Ogawa, & Egeland, 1997). Individuals presenting high preoccupied attachment tendencies reported mistrust and dissatisfaction with potential supporters, and were perceived by their peers as having difficulties seeking help from teachers (Larose & Bernier, 2001). Other studies using self-reports of romantic attachment styles have shown that among students who acknowledged high

06 larose (ds)

18/5/05

3:23 pm

Page 403

Larose et al.: Attachment and college mentoring

403

problem levels in college, those harboring negative models of others (i.e., more avoidant) were less willing to seek counseling than those with positive models (i.e., more secure) (Lopez, Melendez, Sauer, Berger, & Wyssmann, 1998). Similarly, people who expressed attachment-related distrust in the availability and dependability of others (i.e., more anxious–ambivalent) were less likely to evaluate the early phase of a counseling relationship positively (Satterfield & Lyddon, 1995). Finally, women with less secure representations of their parents and who reported being more avoidantly attached to romantic partners provided less support to their partners in times of stress (Simpson, Rholes, Orina, & Grich, 2002). These empirical findings suggest that attachment insecurity may attenuate the benefits that individuals gain from social support interactions. The second objective of this study was to investigate whether the association between security in mentoring and the expected positive outcomes differs according to the student’s level and type of attachment insecurity. Hypotheses In sum, this study examined whether college freshmen’s perceived security with a teacher mentor, as assessed at the end of an academic mentoring program, predicts subsequent perceptions of mentoring and relationship quality with other teachers. We hypothesized that security in mentoring would be related to subsequent positive perceptions of mentoring, high supportive relationships with teachers and low conflictual relationships with teachers. The moderating role of attachment in these associations was also examined, with the expectation that insecure attachments would decrease the strength of the predicted associations. Method Participants Colleges and academic mentoring programs. Participants were recruited from three colleges that offered an academic mentoring program to their academically at-risk freshmen. These three colleges were selected over other colleges offering academic mentoring because of core similarities in their mentoring programs: (1) voluntary participation; (2) mentoring provided on an individual basis by volunteer teachers; (3) a focus on the student’s integration into college and on the prevention of academic, social and emotional adjustment problems; and (4) the exclusion of support on specific academic matters. The programs matched students with teachers whom they otherwise would not have known, and were designed to help students deal more competently and responsibly with the specific challenges of the college transition. The number of meetings was the same for all participants (ten). The meetings ranged in length from 20 to 70 minutes. At-risk students. The sample was composed of 102 students (31 men, 71 women; mean age 18.1 years), the majority of whom (74%) were from intact two-parent households; 42% had to leave home to attend college. These

06 larose (ds)

18/5/05

404

3:23 pm

Page 404

Journal of Social and Personal Relationships 22(3)

students’ high school grades were relatively poor, but sufficient to be admitted to college. Forty-two percent of students attended a public college in a semirural area, 33% attended a private urban college and 25% attended a public urban college. All participants were either native French speakers or sufficiently fluent in French to attend a college where all classes are provided in French. Mentors. Five female and five male college teachers (M age = 39 years; SD = 8.75) were paired with the 102 students. They had a range of teaching experience from 3 to 37 years (M = 13 years; SD = 9.61), and a range of academic mentoring experience from 30 to 450 hours (M = 150 hours). They all received a 3-hour training session that consisted of an informal discussion about the counselor functions (e.g., acceptance, encouragement, guidance, and information) likely to improve students’ adjustment. Measures Perceived security in mentoring. The 25-item Inventory of Parent and Peer Attachment (IPPA; Armsden & Greenberg, 1989) questionnaire evaluates a constellation of emotional experiences with the partner such as trust, communication, and feelings of alienation. To assess the adolescents’ perceived security in mentoring, we modified the instructions in the parental version slightly by asking students to refer to their relationship with their mentor. Otherwise, the items were exactly the same as those included in the parental version. This adapted questionnaire was proved to be reliable and valid in an earlier study (Soucy & Larose, 2000). The reliability and construct validity of the IPPA are well established (Armsden & Greenberg, 1987). A global relational quality score was obtained by summing all items (␣ = .92). Subsequent perceptions of mentoring. A 16-item questionnaire was developed, aimed at assessing two dimensions of students’ perceptions of the mentoring experience 5 months after the end of the program: Maintenance of the relationship (eight items, ␣ = .86) and Satisfaction with mentoring (eight items, ␣ = .84). The Maintenance scale taps the extent and quality of current contacts with the former mentor (e.g., ‘When I experience difficulties, I tend to go and talk about them with the mentor that I had last semester’). The Satisfaction scale pertains to the perceived usefulness of the previous semester’s mentoring program (e.g., ‘Mentoring was useful for me and I would recommend it to other students’). Because these two scales were strongly related (r = .74), these items were summed as the first dependent variable (Subsequent Perceptions of Mentoring). Quality of relationship with teachers. The Measure of Affective Relationships with College Teacher (MARCT; Larose, Bernier, Soucy, & Duchesne, 1999) is a 20-item questionnaire that taps students’ behaviors, feelings and beliefs about their social and affective interactions with college teachers. Specifically, the MARCT was used to assess the levels of perceived Self-Disclosure in the student–teacher relationship (e.g., ‘I spontaneously share information about myself with my teachers’: ␣ = .66 at Time 1 and .83 at Time 2), Conflict (e.g., ‘My teachers and I always seem to be at odds’: ␣ = .78 at both times), Feelings of Unfairness (e.g., ‘I feel that my teachers treat me unfairly’: ␣ = .78 at both times) and Informal Alliances (e.g., ‘I often think about my teachers when not

06 larose (ds)

18/5/05

3:23 pm

Page 405

Larose et al.: Attachment and college mentoring

405

at school’: ␣ = .82 at Time 1 and .70 at Time 2)). Each subscale contains five items that are answered on a 5-point Likert-type scale. Psychometric properties of the MARCT have been documented previously (Bernier, Larose, Soucy, & Duchesne, 1997; Larose et al., 1999). The Seeking Help from Teacher subscale of the Test of Reactions and Adaptation in College (SHT/TRAC; Larose & Roy, 1995) was used to assess adolescents’ help-seeking behaviors in their relationships with college teachers (e.g., When I experience difficulty in understanding the classroom material, I waste a lot of time before asking my teacher for help). The SHT/TRAC is part of a 50-item questionnaire that measures students’ beliefs, emotional reactions and social behaviors in learning situations. The SHT subscale contains five items that are answered on a 7-point Likert-type scale. It has been shown to have good psychometric properties, including high internal consistency and good construct, concurrent and predictive validity (Larose & Roy, 1995). In this study, Cronbach’s alpha for the SHT/TRAC was .84 at Time 1 and .82 at Time 2. Because strong correlations were found between some of the above scales (e.g., the correlations between Conflict and Feelings of Unfairness were .57 at Time 1 and .68 at Time 2), and to avoid redundant analyses, a factor analysis was performed on the five teacher–student relationship scores at Time 1 (i.e., Self-Disclosure, Conflict, Feelings of Unfairness, Informal Alliances, and Seeking Help from Teachers). Based on the model proposed by Wubbels and Levy (1993), we expected to find the two dimensions of support and conflict that are proposed to be central in understanding student–teacher interpersonal relationships. Indeed, two orthogonal factors were derived. The first factor explains 44.3% of the variance (Eigenvalue = 2.22), with positive loadings for the Conflict (.87) and Feelings of Unfairness scores (.85), and a negative loading for the Seeking Help from Teachers score (–.51). The second factor explains 27.2% of the variance (Eigenvalue = 1.36), with positive loadings for the Self-Disclosure (.85), Informal Alliances (.90), and Seeking Help from Teacher scores (.47). These findings were replicated with the Time 2 teacher–student variables. Consequently, two teacher–student factor scores were derived at Time 1 and Time 2 (Conflictual and Supportive Relationships with teachers) and used as two distinct dependent variables. Students’ attachment. The Adult Attachment Interview (AAI; George, Kaplan, & Main, 1996) is a face-to-face semi-structured interview focusing on childhood attachment experiences with parents. Participants are asked to describe their relationships with their parents when they were young, to substantiate descriptions with specific memories, to recall incidences of distress, and to conceptualize relationship influences. The AAI has been shown to have excellent reliability, discriminant validity and predictive validity (see Hesse, 1999 for a review). Written transcripts are scored on a series of state of mind scales. The scoring procedure yields classification of each transcript into one of three primary states of mind: autonomous (F), dismissing (Ds) and preoccupied (E). The main indicators of an autonomous state of mind are high scores on the Coherence (of Transcript/of Mind) scales, whereas a dismissing state of mind is indicated by high scores on the Lack of Recall, Idealization (mother/father) or Derogation scales, and a preoccupied state of mind, by high scores for Anger (mother/father) or Passivity of Discourse (Main & Goldwyn, 1998). Written

06 larose (ds)

406

18/5/05

3:23 pm

Page 406

Journal of Social and Personal Relationships 22(3)

transcripts were rated according to Main and Goldwyn’s (1998) system by the second author, who has met all reliability criteria set by Mary Main. A second judge, also certified by Mary Main, independently rated 30 interviews. Agreement on the classifications was 86.7% (␬ = .71). In order to maximize statistical power, and in line with Furman (2001) and Larose and Bernier (2001), we used a factor approach to AAI data analysis. The state of mind scales were reduced in three steps. First, the Derogation, Metacognition, Unresolved Loss and Unresolved Trauma scales were removed because they had little variance in this sample (mean on a 9-point scale = 1.41 Derogation, 1.16 Metacognition, 1.74 Unresolved Loss, and 1.44 Unresolved Trauma). This is in line with Main and Goldwyn’s (1998) argumentation that these indicators of state of mind can be rare in normative samples. Eight scales were left: Idealization/mother, Idealization/father, Lack of Recall, Anger/mother, Anger/father, Passivity, Coherence of Transcript and Coherence of Mind. Main and Goldwyn argue for the existence of a general state of mind with respect to attachment that integrates experiences with both parents. Thus, the subscales Anger/mother and Anger/father (r = .61), Idealization/mother and Idealization/father (r = .71), and Coherence of Transcript and Coherence of Mind (r = .98) were averaged. Third, the five remaining scales (Anger, Idealization, Coherence, Passivity and Lack of Recall) were submitted to a principal component analysis. Two factors emerged: the factor loadings for Lack of Recall (.88), Idealization (.89) and Coherence (–.92) suggest that the first factor taps the dismissing pattern, whereas the second factor measures preoccupied strategies (i.e., Passivity = .79; Anger = .82). No cross-loadings were observed. This factor structure is consistent with that found in previous independent samples (Furman, 2001; Larose & Bernier, 2001). Because of their theoretical soundness and empirical clarity, the resulting factor scores were used in the remainder of the analyses. Scores (z-scores) on the dismissing dimension varied from –1.28 to 2.64, and those on the preoccupied dimension, from –1.23 to 4.02. Intraclass correlations coefficients (ICC) for the subscale scores ranged from .68 (Passivity) to .88 (Lack of Recall) with a mean of .81. The ICC coefficients were .88 for the dismissing factor and .79 for the preoccupied factor. The factor scores are systematically related to attachment classifications. Participants classified as Autonomous (n = 67) obtained low scores on both dimensions (M = –.59 dismissing dimension and .01 preoccupied dimension), participants classified as Dismissing (n = 28), elevated scores on the dismissing dimension (M = 1.33) and low scores on the preoccupied dimension (M = –.35), and participants classified as Preoccupied (n = 7), low scores on the dismissing dimension (M = -0.09) and elevated scores on the preoccupied dimension (M = 2.47). All differences are significant (p < .001). Procedure Students completed the MARCT and the SHT/TRAC at the very beginning of their first semester in college (Time 1). At this point, they were instructed to refer to their experiences with high school teachers. Throughout the first semester, they participated in a 10-hour academic mentoring program with a volunteer teacher who addressed a range of affective, social and academic issues. At the very end of the semester (Time 2), the IPPA-mentor version was administered and the MARCT and SHT/TRAC were re-administered. Students were instructed to complete these two latter questionnaires by referring to their

06 larose (ds)

18/5/05

3:23 pm

Page 407

Larose et al.: Attachment and college mentoring

407

college teachers, and to exclude their mentor from their appreciation. Finally, all students were met again 5 months following the end of the mentoring program (Time 3), and were administered the questionnaire pertaining to subsequent perceptions of mentoring as well as the AAI.

Results The results are presented in three steps. First, we examine the correlations between students’ attachment dispositions (i.e., dismissing and preoccupied dimensions) and their perceptions of security in mentoring. Second, the hypothesis linking perceived security in mentoring with subsequent perceptions of mentoring and relationship quality with teachers (i.e., conflictual and supportive relationships) is examined through multiple linear regressions. Finally, the hypothesis of the moderating effect of attachment is tested by adding two interaction terms to the same regressions: ‘Dismissing Tendencies by Security in Mentoring’ and ‘Preoccupied Tendencies by Security in Mentoring’. Any significant interaction is then broken down using the procedure proposed by Aiken and West (1991). Because student participants were nested within mentor, the question of interdependence of scores was examined following the guidelines proposed by Kashy and Kenny (2000). Intraclass correlations for group, with students nested within each mentor treated as a group, were estimated on the three dependent variables (Conflictual and Supportive Relationships with Teachers and Subsequent Perception of Mentoring) and the independent variable (Security in Mentoring). The four intraclass correlations were respectively .038, –.048, .021, and .016 (p values > .05), thus suggesting the absence of interdependence in our data. Students’ attachment and their perceptions of security in mentoring Attachment was weakly associated with perceived security in mentoring. The correlations between the levels of dismissing and preoccupied tendencies in the AAI and the levels of security in mentoring were respectively –.21 (p < .05) and –.22 (p < .05). Security in mentoring, subsequent perception of mentoring, and relationship quality with teachers Tables 1 and 2 present the results of the regression analyses. The pattern of analysis is essentially the same for each dependent variable. We first entered the initial score at Time 1 (when available) in order to predict the outcome at Time 2 after controlling for initial levels of the same construct. The two attachment scores were entered in a second step, followed by the security in mentoring score. This third step allowed us to test the hypothesis that the perception of security in mentoring predicts subsequent perceptions of mentoring and of teacher–student relationships. The last step included the ‘Attachment ⫻ Security in Mentoring’ interactions, and allowed us to examine the moderating effect of attachment. For clarity purposes, these interaction effects are discussed in a separate section. Table 1 shows that security in mentoring predicts 6% of the variance of subsequent perceptions of mentoring (see Step 2). Students who perceived higher levels of security in mentoring were more satisfied with the mentoring

06 larose (ds)

18/5/05

3:23 pm

Page 408

408

Journal of Social and Personal Relationships 22(3)

TABLE 1 Standardized regression coefficients for subsequent perception of mentoring: moderating influences of students’ attachment insecurity Predictor 1. AAI 3. Dismissiveness score (Ds) 3. Preoccupied score (Pr) 2. Security in mentoring (SM) 3. Interaction 3. Ds ⫻ SM 3. Pr ⫻ SM Adjusted R2 F change

Step 1

Step 2

Step 3

–.19 –.03

–.13 .04 .29**

–.12 –.08 .26*

.08 7.54***

.07 –.38*** .26 7.99**

.02 1.76

*p < .05; **p < .01; ***p < .001.

TABLE 2 Standardized regression coefficients for relationships with teachers: moderating influences of students’ attachment insecurity Predictor 1. Conflictual relationships at Time 1 2. AAI 2. Dismissiveness score (Ds) 2. Preoccupied score (Pr) 3. Security in mentoring (SM) 4. Interaction 2. Ds ⫻ SM 2. Pr ⫻ SM Adjusted R2 F change 1. Supportive relationships at Time 1 2. AAI 2. Dismissiveness score (Ds) 2. Preoccupied score (Pr) 3. Security in mentoring (SM) 4. Interaction 2. Ds ⫻ SM 2. Pr ⫻ SM Adjusted R2 F change

Step 1 .73***

Step 2 .72*** .02 .03

.52 101.51*** .61***

.52 .11 .61*** .03 –.02

.36 54.47***

.36 .09

Step 3 .64***

Step 4 .63***

–.05 –.03 –.29***

–.07 .01 –.23**

.58 15.03***

–.17** .07 .60 3.00*

.61***

.61***

.06 .01 .10

.06 –.02 .11

.36 1.25

–.01 –.09 .35 .46

*p < .05; **p < .01; ***p < .001.

program and maintained more contacts with their mentor 5 months after the end of the program. Table 2 indicates that high security in mentoring predicts low levels of conflictual relationships with teachers, after controlling for conflictual relationships at Time 1 (see Step 3). The variance explained was 6%.

06 larose (ds)

18/5/05

3:23 pm

Page 409

Larose et al.: Attachment and college mentoring

409

The perception of security in mentoring does not predict supportive relationships with teachers (see bottom of Table 2, Step 3). The moderating effect of attachment on mentoring The moderating role of attachment in the links between security in mentoring, subsequent perceptions of mentoring and relationship quality with teachers was examined in Step 3 of Table 1 and Step 4 of Table 2. Table 1 shows that the interaction effects are significant, explaining 18% of the variance in subsequent perception of mentoring scores. The beta scores show that the preoccupied dimension explains this variance. Three simple regression analyses linking security in mentoring to the outcome were conducted in order to decipher the interaction pattern (Aiken & West, 1991). These analyses were performed on the basis of students’ standardized preoccupation score using the values suggested by Cohen and Cohen (1983): one standard deviation below the mean, at the mean, and one standard deviation above the mean. The relation between security in mentoring and subsequent perceptions of the relationship varied according to students’ level of preoccupation with attachment. Among students showing low preoccupied tendencies, security in mentoring was positively related to subsequent perceptions of mentoring (␤ = .82, p < .001), whereas among students showing high preoccupied attachment tendencies, the linear relation between security in mentoring and subsequent perceptions was nonsignificant (␤ = –.02, ns). Table 2 shows that the attachment scores interact with security in mentoring to predict Conflictual Relationships with teachers. These interactions explain only a small amount of variance (2%), which is nevertheless significant. The beta scores show that it is the dismissing dimension that contributes to this variance. The relation between security in mentoring and conflictual relationships with teachers varied according to students’ level of dismissing tendencies. Among students showing high dismissing tendencies, high security in mentoring predicted low levels of conflictual relationships with teachers at Time 2 (␤ = –.63, p < .000), after controlling for conflictual relationships at Time 1, whereas among students showing low dismissing tendencies, the linear relation between security in mentoring and conflictual relationships with teachers was nonsignificant (␤ = –.17, ns). This interaction effect is in sharp contrast with our hypotheses. No Security by Attachment interaction effect was found significant in predicting supportive relationships with teachers (see bottom of Table 2, Step 4). To summarize, the tendency to show a preoccupied attachment pattern in the AAI decreases the strength of the positive link between perceived security in mentoring and subsequent perceptions of mentoring, whereas the tendency toward dismissing attachment increases the positive link between perceived security in mentoring and quality of interactions with teachers (i.e., low conflictual relationships). The findings involving the dismissing dimension are counterintuitive and in fact, contradictory to our hypotheses.

Discussion The first goal of this study was to examine the associations between college students’ perceived security in mentoring, their subsequent perceptions of

06 larose (ds)

410

18/5/05

3:23 pm

Page 410

Journal of Social and Personal Relationships 22(3)

the mentoring experience, and the quality of their relationships with other teachers. The results showed that academically at-risk students who perceived the relationship with their mentor as more secure were more likely to maintain links with the mentor 5 months after the end of the program, and to report being still satisfied with their experience. Moreover, the degree of felt security was inversely related to perceptions of conflictual relationships with other teachers at the end of the mentoring program, controlling for initial perceptions. These results suggest that perceived security in mentoring can be transferred to a better appreciation and utilization of relationships with teachers in general, which may help explain the beneficial effects of mentoring on students’ academic and social adjustment (Darling, Hamilton, & Niego, 1994). These results are all the more important because it has been shown that academically at-risk students perceive more conflicts and unfairness in their relationships with teachers and seek their help less frequently when they have problems (Richman et al., 1998). Many studies on academic mentoring have been guided by the premise that simply pairing an adolescent with a mentor from the college environment will help prevent academic adjustment problems. The results of this study, however, suggest that in the absence of relational security with the mentor, mentoring may have little effect on students’ adjustment. The second goal of this study was to examine the moderating role of attachment in the links between the aforementioned variables. Given the parameters of the mentoring relationship (duration, frequency, physical setting, etc.), it appears very unlikely that the dyads developed attachment relationships over the course of the mentoring program. Nonetheless, both insecure attachment dimensions were found to play a role in the benefits drawn from the program. First, more dismissing and preoccupied attachments were associated with lower perceptions of security in mentoring, which supports the hypothesis that insecure WMA influence how students develop mentoring relationships with teachers. Second, preoccupation with attachment was found to moderate the relation between felt security in mentoring and perceptions of the program one semester later: the expected positive association was observed only among students showing low preoccupation with attachment. It could be that the development of a bond with the mentor (indicated by a high score on perceived security) in the context of a student’s tendencies toward preoccupation with attachment can activate his or her attachment system and typical relational patterns. One might speculate that the activation of preoccupied tendencies in the context of academic mentoring triggers dependency on the mentor and an elevated need for support, thereby leading the student to give an exaggerated importance to the mentoring relationship. This, in turn, could result in the student feeling disappointed when the program ends, leading him or her to report low satisfaction with mentoring, and preventing him or her from trying to keep in touch with the former mentor. Finally, the moderating effect of dismissing attachment on the association

06 larose (ds)

18/5/05

3:23 pm

Page 411

Larose et al.: Attachment and college mentoring

411

between perceived security in mentoring and conflictual relationships with teachers was found to be significant, but in an unexpected direction. Felt security in mentoring was negatively related to conflictual relationships with teachers, but only for students showing high dismissing tendencies in the AAI, suggesting that mentoring was more beneficial in this regard to dismissing students than to others. The presumed origins of dismissing attachment should be considered in explaining the counterintuitive results obtained. Adolescents with a dismissing attachment pattern tend to downplay the importance of their relationship with their parents because their experiences with them have presumably been marked by rejection and/or indifference (Main & Goldwyn, 1998). Dismissing adolescents claim that they do not need their parents to meet their emotional needs and often show a compulsive self-reliance pattern in solving emotional problems (Allen, Moore, Kuperminc, & Bell, 1998). The relationship of trust established with a mentor might thus partly compensate for the difficulties that the adolescent’s parents have in providing emotional support, thereby resulting in positive changes in the adolescent’s perceptions of adults. Hence, a positive experience with an adult mentor might be especially beneficial to dismissing adolescents because they need it to a greater degree than those adolescents who already feel like they can freely access support from their parents. Strengths and limitations This study has a number of theoretical and methodological strengths. First, despite the empirical evidence that (i) academically at-risk students have difficult relationships with their teachers, and (ii) mentoring has a beneficial impact on at-risk students’ adjustment, this study is the first to test if mentoring relationships improve perceptions of student–teacher relationships. The findings indicate that security in mentoring helps at-risk students develop less conflictual relationships with teachers, which could be one mechanism through which mentoring favors students’ adjustment. Second, previous studies linking attachment to interpersonal functioning have mainly focused on well-established, ongoing dyads (e.g., adolescent–mother, best friends, romantic partners), which already present a rich interpersonal history at the onset of the study. Examining a relationship characterized by a clear beginning between individuals who did not know each other before has allowed for isolation of the unique role of attachment insecurity, not confounded with that of the dyad’s interpersonal history. Finally, this study has emphasized the importance of not solely examining the direct effect of attachment, but also its moderating effect on interpersonal relationships. Because of these two latter particularities, this study is one of the very few to have tested widely claimed assumptions regarding the role of attachment in the formation of new relationships (Bowlby, 1988; Dozier & Tyrrell, 1998; Slade, 1999). One limitation of the present study is the absence of a control group. Only a randomized design with control groups would have enabled to draw causal inference regarding the impact of mentoring on students’ relationships with teachers. However, the possibility to predict changes in such

06 larose (ds)

412

18/5/05

3:23 pm

Page 412

Journal of Social and Personal Relationships 22(3)

relationships, through the use of initial scores at Time 1, increases the confidence one may have regarding the direction of the association. A second limitation pertains to the assessment of initial perceptions of relationships with teachers in reference to the high school context. This was a necessary procedure in order to assess these initial perceptions very early on in the first semester, prior to the beginning of the mentoring program. Yet, there could be some qualitative differences between students’ perceptions of their relationships with high school teachers and with college professors. However, we argue that having a baseline level, although very imperfect, strengthens the internal validity of the results if one is aware of the potential limitations associated with the assessment of this baseline (Kazdin, 2003). A third limitation pertains to the moment when the AAI was administered (at Time 3). Given that previous studies have shown changes in WMA during the transition to college (Davila, Burge, & Hammen, 1997), one may argue that attachment tendencies could have changed in some students. However, in the latter study, changes were found in romantic attachment, not attachment to parents, using self-report methodology. In the present study, attachment insecurity was assessed with an interview focusing on relationships with parents. The AAI has shown high stability over a 2-year period among late adolescents (Zimmermann & Becker-Stoll, 2002). Despite this empirical support, the possibility that WMA of academically at-risk students may have shifted during the transition to college remains an open question and would need further examination. Finally, although our measure of subsequent perceptions of mentoring presents high internal consistency, its convergent and divergent construct validity is yet unknown. The finding that security in mentoring predicted subsequent perceptions of the mentoring experience should be considered as a first step in the validation of this scale. Directions for future research A number of implications emerge from the results of the present study. First, in evaluating mentoring programs, it is important to consider not only the presence or absence of a mentor, but also the interpersonal mechanisms operating in mentoring (Rhodes, 2002). The present study has focused on one interpersonal factor (i.e., perceived security) that potentially mediates the impact of mentoring. Additional interpersonal factors to be investigated include the mentor–mentee agreement on the content and goals of the intervention, and on the type of support needed. Second, the demonstration that attachment dispositions moderate the relation between perceived security in mentoring and academic interpersonal functioning calls for a more thoughtful examination of the role of attachment in the formation, maintenance and termination of mentoring relationships. For instance, it could be of interest to examine if the moderating effect of attachment is more pronounced at the beginning of the mentoring relationship than at the end. Attachment researchers would posit that attachment dispositions are stable and drive the mentee’s interpersonal style (Dozier & Tyrell, 1998), thus affecting each stage of the mentoring relationship equally. In contrast,

06 larose (ds)

18/5/05

3:23 pm

Page 413

Larose et al.: Attachment and college mentoring

413

mentoring researchers would be more inclined to suggest that mentoring may, under certain circumstances, buffer the negative impact of attachment insecurity (Rhodes, 2002), thus viewing attachment as a malleable characteristic that may change progressively throughout the mentoring intervention. Longitudinal studies involving repeated assessments of interpersonal functioning throughout the mentoring experience are needed to clarify these matters. Finally, when evaluating mentoring outcomes, studies should also consider the possible contribution of the mentor’s attachment dispositions (Bernier, Larose, & Soucy, 2005). Like many relational contexts, the mentoring relationship involves two people who are pursuing a common objective but who each have their own interpersonal history. It is very likely that one’s adjustment to the relationship is partly influenced by the other’s history of interactions, and vice versa. Conclusion This study suggests that greater knowledge on the part of mentors of the different attachment tendencies that characterize their students may help them better understand students’ varied responses to their interventions. By learning more about the implications of these tendencies for relational functioning, the mentor will be able to put the impact of his or her interventions into perspective, and possibly adjust his or her practices. For most mentors, this is not to suggest that they should try to modify the student’s attachment state of mind (as a therapist would do), but rather that they optimize their interventions by taking attachment models into account. Mentors should be aware that help excessively centered on students’ emotional experience could cause some of them to withdraw (e.g., students who are more dismissing) and others to become too dependent (e.g., students who are more preoccupied). Mentors may, therefore, choose to adapt their interventions to the student’s predispositions in order to maximize the benefits of mentoring.

REFERENCES Aiken, L. S., & West, S. G. (1991). Multiple regression: Testing and interpreting interactions. Newbury Park, CA: Sage. Allen, J. P., Moore, C., Kuperminc, G., & Bell, K. (1998). Attachment and adolescent psychosocial functioning. Child Development, 69, 1406–1419. 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. Armsden or M. T. Greenberg, Department of Psychology, Box 351525, University of Washington, Seattle, WA 98195–1525, USA.) Bartholomew, K. (1990). Avoidance of intimacy: An attachment perspective. Journal of Social and Personal Relationships, 7, 147–178. Bernier, A., Larose, S., & Soucy, N. (2005). Academic mentoring in college: The interactive role of student’s and mentor’s interpersonal dispositions. Research in Higher Education, 46, 29–51.

06 larose (ds)

414

18/5/05

3:23 pm

Page 414

Journal of Social and Personal Relationships 22(3)

Bernier, A., Larose, S., Soucy, N. & Duchesne, S. (1997, April). Attachment style in adolescence and perceptions of teacher–student affective relationships. Poster presented at the SRCD meeting, Washington, DC. Bowlby, J. (1988). Attachment and loss, Vol. 3. Loss: Sadness and depression. New York: Basic Books. Cohen, J., & Cohen, P. (1983). Applied multiple regression/correlation analyses for the behavioral sciences (2nd ed.). Hillsdale, NJ: Erlbaum. Darling, N., Hamilton, S. F., & Niego, S. (1994). Adolescents’ relations with adults outside the family. In R. Montemayor, G. R. Adams, & T. P. Gullotta (Eds.), Personal relationships during adolescence (pp. 216–235). Thousand Oaks, CA: Sage. Davila, J., Burge, D., & Hammen, C. (1997). Why does attachment style change? Journal of Personality and Social Psychology, 73, 826–838. Dozier, M. (1990). Attachment organization and treatment use for adults with serious psychopathological disorders. Development and Psychopathology, 2, 47–60. Dozier, M., & Tyrrell, C. (1998). The role of attachment in therapeutic relationships. In J. A. Simpson & W. S. Rhodes (Eds.), Attachment theory and close relationships (pp. 25–45). New York: Guilford Press. Feeney, J. A., & Noller, P. (1996). Adult attachment. Thousand Oaks: Sage. Furman, W. (2001). Working models of friendship. Journal of Social and Personal Relationships, 18, 583–602. George, C., Kaplan, N., & Main, M. (1996). Adult Attachment Interview protocol (3rd ed.). Unpublished manuscript, University of California at Berkeley. Hazan, C., & Shaver, P. (1987). Romantic love conceptualized as an attachment process. Journal of Personality and Social Psychology, 52, 511–524. 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 Press. Jacobi, M. (1991). Mentoring and undergraduate academic success: A literature review. Review of Educational Research, 61, 505–532. Kashy, D. A., & Kenny, D. A. (2000). The analysis of data from dyads and groups. In H. T. Reis & C. M. Judd (Eds.), Handbook of research: Methods in social and personality psychology. Cambridge: Cambridge University Press. Kazdin, A. E. (2003). Methodological issues and strategies in clinical research (3rd ed.). Washington, DC: American Psychological Association. Kobak, R. R., & Sceery, A. (1988). Attachment in late adolescent: Working models, affect regulation, and representation of self and others. Child Development, 59, 135–146. Korfmacher, J., Adam, E., Ogawa, J., & Egeland, B. (1997). Adult attachment: Implications for the therapeutic process in a home visitation intervention. Applied Developmental Science, 1, 43–52. Lamport, M. A. (1993). Student–faculty informal interaction and the effect on college student outcomes: A review of the literature. Adolescence, 28, 971–990. 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., Bernier, A., Soucy, N., & Duchesne, S. (1999). Attachment style dimensions, network orientation, and the process of seeking help from college teachers. Journal of Social and Personal Relationships, 16, 225–247. Larose, S., & Roy, R. (1995). Test of Reactions and Adaptation in College (TRAC): A new measure of learning propensity for college students. Journal of Educational Psychology, 87, 293–306. Long, N. J., & Morse, W. C. (1996). Conflict in the classroom: The education of at-risk and troubled student (5th ed.). Austin, TX: Pro-Ed. Lopez, F. G., Melendez, M., Sauer, E.M., Berger, E., & Wyssmann, J. (1998). Internal working models, self-reported problems, and help-seeking attitudes among college students. Journal of Counseling Psychology, 45, 79–83.

06 larose (ds)

18/5/05

3:23 pm

Page 415

Larose et al.: Attachment and college mentoring

415

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 representation. In I. Bretherton & E. Waters (Eds.), Growing points of attachment theory and research. Monographs of the Society for Research in Child Development, 50(1–2). Nelson, J. S., Epstein, M. H., Bursuck, W. D., Jayanthi, M., & Sawyer, V. (1998). The preference of middle school students for homework adaptations made by general education teachers. Learning Disabilities Research and Practice, 13, 109–117. Rhodes, J. E. (2002). Stand by me: The risks and rewards of mentoring today’s youth. Cambridge, MA: Harvard University Press. Rhodes, J. E., Grossman, J. B., & Resche, N. L. (2000). Agents of change: Pathways through with mentoring relationships influence adolescents’ academic adjustment. Child Development, 71, 1662–1671. Richman, J. M., Rosenfeld, L. B., & Bowen, G. L. (1998). Social support for adolescents atrisk of school failure. Social Work, 43, 309–323. Satterfield, W. A., & Lyddon, W. J. (1995). Client attachment and perceptions of the working alliance with counselor trainees. Journal of Counseling Psychology, 42, 187–189. Simpson, J. A., Rholes, W. S., Orina, M. M., & Grich, J. (2002). Working models of attachment, support giving, and support seeking in a stressful situation. Personality and Social Psychology Bulletin, 28, 598–608. Slade, A. (1999). Attachment theory and research: Implications for the theory and practice of individual psychotherapy with adults. In J. Cassidy & P. R. Shaver (Eds.), Handbook of attachment: Theory, research and clinical applications (pp. 575–594). New York: Guilford Press. Soucy, N., & Larose, S. (2000). Attachment and control in family and mentoring contexts as determinants of adolescent adjustment to college. Journal of Family Psychology, 14, 125–143. Talmi, A. (1997, April). Dimensions of support provided by special adults in the lives of young adolescents. Poster presented at the conference of the Society for Research in Child Development, Washington, DC. Wilson, K. B. (1994). Developing a freshman mentoring program: A small college experience. In M. A. Wunsch (Ed.), Mentoring revisited: Making an impact on individuals and institutions. San Francisco: Jossey-Bass. Wubbels, T., & Levy, J. (Eds.). (1993). Do you know what you look like? Interpersonal relationships. London: Falmer Press. Zimmermann, P., & Becker-Stoll F. (2002) Stability of attachment representations during adolescence: The influence of ego-identity status. Journal of Adolescence, 25, 107–124.