The Same Old Song? -- Continuity and Change in Relationship Schemas from Adolescence to Young Adulthood

Relationship schemas are central to our understanding of interpersonal functioning. The aim of this study is to examine continuity and change in relationship themes across two developmental epochs ? adolescence and young adulthood ? in the stories that people tell about their interactions with others. Using the Core Conflictual Relationship Theme Method, relationship themes were coded from semi-structured interviews conducted in adolescence and again in early adulthood. The sample consisted of 20 male and 20 female participants in a longitudinal study of psychological development. There was considerable stability across time in the frequency with which particular themes were expressed relationship narratives. Significant changes from adolescence to young adulthood included a decrease in the perception of others as rejecting and of the self as opposing others. Young adults saw themselves and others more positively, and used a broader repertoire of themes in their relationship narratives than they had as adolescents. The basic continuity and particular changes in relationship schemas found in this study are consistent with knowledge about the adolescent-to-young-adult transition derived from other clinical and empirical sources. Tracking conflicting wishes and images of self and other over time sheds light on central developmental processes of this period. 3 The Same Old Song? INTRODUCTION That the present is shaped by the past is one of the oldest and most enduring concepts in our understanding of human relationships. Psychoanalytic theory, social psychology, cognitive neuroscience, and folk wisdom are among the lenses through which students of human nature repeatedly observe that people internalize memories of their most important early relationships, and that these internalized memories may profoundly affect their experiences of and behaviors with significant others in the present. Such internalized models of relationships, or relationship schemas, are thought to be core elements of personality that are enduring and slow to change (Baldwin, 1992; Thorne & Klohnen, 1993). Throughout the psychological literature we find the idea that people approach others with pre-formed expectations about what relationships will be like. Freud conceived of each person as having a central relationship pattern -"a stereotype plate (or several such), which is constantly repeated--constantly reprinted afresh--in the course of the person's life" (Freud, 1912, pp.99-100). According to Freud? s concept of transference, we approach new people with personal relationship templates comprised in large measure of the frustrated wishes, longings, and expectations left over from earlier attempts at need gratification. The idea of a central relationship pattern is reflected in Sullivan's (1953) concept of parataxic distortion, in Tomkin's (1963) work on nuclear scripts, and in Bowlby? s (1969) internal working models of attachment. The use of attachment theory to formulate empirically testable hypotheses constitutes a major development in the study of relationship schemas. The careful, systematic observational research spawned by Bowlby? s work provides empirical support for the theoretical links between early interactions with parents and later behavior, particularly in stressful situations (Ainsworth & Eichberg, 1991; Main, Kaplan, & Cassidy, 1985). Relationship schemas generally include: (1) representations of the self, (2) 4 The Same Old Song? representations of the other person, and (3) an interpersonal script that links self and other. This script? like the script of a play? contains declarative elements (e.g., expected dialogue) and procedural elements (e.g., habitual behaviors) for both actors in the relationship, and includes feelings and goals. People who experience specific patterns of relating to others repeatedly in early life may over-learn these patterns and apply them to situations in the present even when those situations are inappropriate (Baldwin, 1992). Indeed, most psychological therapies aim to free people from the habitual misapplication of such overlearned patterns to current relationships (Linehan, 1987; Luborsky, 1984). How relationship schemas develop and change over the life cycle remains poorly understood. Most often, researchers study representations of the self and representations of others separately, and from this work we have evidence of important developmental pathways that need to be examined. For example, it appears that representations of others in significant relationships become more complex and multifaceted over time (Markus, 1977; Sande, Goethals, & Radloff, 1988). Similarly, the concept of self undergoes differentiation into multiple facets of the self, with more categories of self-description and more trait labels added to one? s repertoire over the course of development from childhood through adolescence (Harter, 1990; Montemayor & Eisen, 1977; Rosenberg, 1986). Much empirical evidence points to the fact that our relationship schemas are slow to change. Research on memory demonstrates that people preferentially notice and recall information for which they have relevant schemas. Failure-based memories may be particularly likely to become part of the personality and to shape expectations about the future (Cantor & Zirkel, 1990; Markus & Nurius, 1986). Moreover, we know that well-developed relationship schemas often lead people to disattend to or explain away information that is inconsistent with their expectations in relationships (Baldwin, 1992; Kulik, Sledge, & Mahler, 5 The Same Old Song? 1986; Swann & Ely, 1984), thereby promoting conservation of these schemas and distorting memory (Fiske, Haslam, & Fiske, 1991; Markus, 1977; Swann & Read, 1981). Numerous studies have demonstrated that social behavior often produces responses from others that confirm our interpersonal expectations and in this way reinforces previously-formed relationship schemas (Miller & Turnbull, 1986; Snyder, 1984). That relationship schemas matter in the real world ? that they guide our behavior toward others ? is a fundamental theoretical premise of most forms of interpersonal psychotherapy. Flexibility in how we respond to human interaction is often considered to be a hallmark of mental health and psychological maturity (Reich, 1987; Vaillant, 1971). Those who apply a limited number of relationship schemas rigidly and inappropriately in adulthood are often diagnosed with Personality Disorders (DSM-IV, 1994). Changes in relationship schemas ? particularly a shift toward more positive perceptions of others -have been found to be pivotal to positive outcomes in some forms of psychotherapy (Grenyer & Luborsky, 1996; Luborsky & Crits-Christoph, 1990). The links between relationship schemas and interpersonal outcomes has received increasing empirical support in recent decades. For example, Simpson and colleagues (Simpson, Rholes, & Nelligan, 1992; Simpson, Rholes, & Phillips, 1996) have shown that knowing individuals? typical attachment patterns allows one to predict with some accuracy how they will behave toward their romantic partner when placed in a stressful situation. In the realm of developmental psychopathology, maltreated children have been found to generalize negative representational models to new situations and relationship figures, thereby increasing the likelihood that they will perpetuate their relationship histories (Lynch & Cicchetti, 1991; Toth, Cicchetti, Macfie, & Emde, 1997; Toth & Cicchetti, 1996). Relationship schemas have most often been studied by attempts at direct assessment, 6 The Same Old Song? asking people to report on what they expect in their dealing with others (e.g., Rempel, Holmes, & Zanna, 1985). For example, Hazan and Shaver (1987) asked participants to read three descriptions of relationships and endorse the one that best characterized their relationships. The direct approach has the particular advantage of being relatively laborefficient. A major disadvantage of this approach, however, is that some people are likely to distort their responses to direct questions about relationship patterns for defensive reasons (Main et al., 1985; Westen, 1991). For example, Main and her colleagues (1985) suggest that people with avoidant attachment styles do, in fact, see others as disappointing them, but will be unlikely to report this because it would be threatening to see themselves as disappointed by ? and therefore dependent on ? others. A second major mode of studying relationship schemas ? the so-called idiographic mode -offers less opportunity for response bias. It involves collecting stories about relationships from individuals? pasts or from their current day-to-day lives, or culling episodes from videotaped psychotherapy sessions (Carlson & Brincka, 1987; Horowitz, 1989; Luborsky, Crits-Christoph, & Mellon, 1986). Judges then examine these episodes for consistencies in themes, inferring that these consistencies constitute ? stereotype plates? or over-learned patterns. Researchers have found that they can reliably group these narratives according to a limited number of interpersonal themes (Luborsky & Crits-Christoph, 1990). The most widely used of these coding systems, and the one used in the present study, is the Core Conflictual Relationship Theme (CCRT) Method (Luborsky & Crits-Christoph, 1990). In the CCRT coding system, judges locate stories about relating to others told by one person (e.g., in a psychotherapy session), and identify in these stories the person? s wishes, perceptions of the other person? s responses, and responses of the self to the interaction. For example, a man may tell several stories about dealings with different people, but in the 7 The Same Old Song? majority of his narratives the theme may be: ? I want to be close to the other person, but the other is distant, and I feel angry and withdrawn.? Using this method, Luborsky and his colleagues have found that people have identifiable core relationship themes that are repeated across multiple stories (Luborsky, 1977). The development of close relationships is of particular importance during the transition from adolescence to young adulthood. According to Erikson (1959), the most important tasks of adolescence include formation of one? s own identity, the development of new relationships outside of the family, and increased autonomy from parents. While developing autonomy and maintaining relatedness with parents have often been conceptualized as opposite ends of a relational spectrum, there is growing evidence that a state of ? autonomousrelatedness? is an optimal outcome for the adolescent-parent relationship (Allen, Hauser, Bell, & O'Connor, 1994; Allen, Kuperminc, & Moore, 1997). Adolescents? coexisting concerns about maintaining close relationships and achieving independence would be expected to engender conflicting wishes and perceptions of self and other in relationships with peers as well as parents. The waning of this conflict in young adulthood is thought to pave the way for the achievement of sustained intimate relationships (Erikson, 1959). Given the centrality of relationship schemas to our understanding of personality and to our attempts to effect therapeutic change in relationship patterns, an understanding of how these schemas function through the life cycle is of considerable importance. The present study is the first to use longitudinal data collected during two developmental epochs ? adolescence and young adulthood -to examine relationship narratives across time. Based on prior research (Bachman, O'Malley, & Johnston, 1978; Block, 1971; Kagan & Moss, 1962), we hypothesized that people would demonstrate considerable stability in relationship themes over 8 The Same Old Song? the decade from adolescence to young adulthood; most particularly, that the wish to be close to others would be the most frequently expressed wish in both eras. In terms of specific changes, we hypothesized that: 1) participants would express more wishes for autonomy and more conflicting perceptions of self and other in adolescence than in young adulthood, 2) young adults would demonstrate a broader repertoire of perceptions of self and others than adolescents, and 3) more positive images of others would be expressed in young adulthood than in adolescence. Finally, based on recent theoretical work on gender differences in interpersonal relationships (Chodorow, 1978; Jordan, Kaplan, Baker Miller, Stiver, & Surrey, 1991), we predicted that females would be more overtly concerned with relationships and would tell more complex stories about relationships than males during both developmental epochs. METHODS Design This study is an analysis of data collected from participants in a longitudinal study of psychological development conducted between 1978 and 1991. Transcribed interviews collected from participants in adolescence (ages 14-16) and again at age 24 were analyzed using the Core Conflictual Relationship Theme Method (Luborsky & Crits-Christoph, 1998) described below. Sample and Timing The sample is derived from a cohort of 146 participants and their families first seen in 1978 as part of the Adolescent and Family Development Project (Hauser, with Powers, & Noam, 1991). On entering this longitudinal study at age 14, subjects were members of 9 The Same Old Song? primarily Caucasian middleand upper-middle-class families. Approximately half were recruited from the freshman class of a local high school (n=76), and half were non-psychotic, non-retarded psychiatrically hospitalized adolescents (n=70). There were 73 males and 73 females (mean age = 14.6 years). The rationale for sampling from these two contrasting groups was to obtain a range of levels of social and psychological functioning. The psychiatric sample was originally diagnosed using DSM-II and later reclassified using DSM-III-R through systematic chart review (Rappaport & Ismond, 1990; Spitzer & Williams, 1987), generating following distribution of primary diagnoses: disruptive behavior disorders (e.g., attention deficit/hyperactivity disorder, conduct disorder) = 33; affective disorders = 18; anxiety and stress-related disorders = 11; and other diagnoses = 8. All eligible patients approached during the recruitment period accepted the invitation to participate in the study, as did almost half (n=250) of the freshman class of one high school. From the latter group, 76 students were selected to match demographic characteristics of the psychiatric group. The groups did not differ in terms of age, gender, birth order, number of siblings, or number of parents living at home. The Adolescent and Family Development Project collected data on ego development and other aspects of psychological and interpersonal functioning from 146 adolescents over the period from age 14 to 17. Ninetyeight percent of the original adolescent participants were studied again at age 24 using ageappropriate measures of psychological and interpersonal functioning. This study uses a sub-sample of 20 male and 20 female subjects, selected in equal numbers from the normal and psychiatrically hospitalized groups. Ego development, as measured on the Loevinger Sentence Completion Test, was used as an index of psychological maturity (Hauser, 1976), and this sub-sample was selected so that both the high school and 10 The Same Old Song? hospital groups included participants who ranged from the least mature (pre-conformist) to most mature (post-conformist) in their Loevinger ego development scores.