Constructing the Self in Early, Middle, and Late Adolescent Boys: Narrative Identity, Individuation, and Well-Being

Abstract

The present study examined aspects of identity development in a sample of adolescent boys from two approaches: individuation and narrative. To extend the more recent research on narrative identity development, we also examined relations between narrative identity, well-being, and age. Narrative meaning making was predicted by themes of individuation in the narratives, specifically the interaction of autonomy and connectedness. Well-being was predicted by different aspects of meaning depending on the kind of meaning and the stage of adolescence. Finally, results showed an age-related increase in meaning-making processes, particularly meaning related to perceiving the self as changing. Results are discussed in terms of the processes of narrative identity development for adolescent boys.

A major task of adolescence is to construct a personal identity (Erikson, 1968). One of the main emphases of research on identity development in Western cultures is on the process of establishing oneself as unique from others, but in the context of close relationships. That is, identity development involves establishing an autonomous sense of self while at the same time developing and sustaining close relationships with others, a process subsumed under the label of individuation (e.g., Grotevant & Cooper, 1985). While many classic studies of identity development have focused on this process of individuation, a more recent approach to the study of identity development has been a narrative approach. Narrative is both a tool to examine identity development as well as the vehicle by which identity is constructed (e.g., McAdams, 1993). From a narrative approach, identity development occurs as one reflects on past experiences in relation to the present and future self (e.g., Habermas & Bluck, 2000; McAdams, 1993); that is, one makes meaning of the past to understand one's sense of self through time.

In the present study we aimed to marry these approaches—individuation and narrative; indeed, we suggest that narrative is a mechanism by which individuation develops. In this study we further aimed to extend the study of narrative identity development to adolescents, as this approach has most often been examined with midlife adult populations. Thus, this is the first large-scale study (of which we are aware) to examine the development of narrative identity in adolescence. Specifically, in a sample of adolescent boys, we examined (1) meaning making in relation to themes of individuation, (2) meaning making and individuation in relation to well-being, and (3) age-related differences in the personal meaning that adolescents made of their pasts.

CONSTRUCTING IDENTITY: TWO APPROACHES

The central tenet of both approaches to identity development, individuation and narrative, is to develop a healthy identity. While individuation and narrative have largely been investigated by two distinct theoretical and empirical traditions, we suggest that there are important relationships between these constructs (Dunbar & Grotevant, 2004).

Individuation

Building on a broad emphasis within developmental psychology to study development within important social and relational contexts, researchers have shown that healthy identity exploration and achievement involves the simultaneous processes of developing an autonomous self within the context of close relationships (e.g., Grotevant & Cooper, 1985; see also Blatt & Blass, 1996; Cleveland & Reese, 2005; Deci & Ryan, 2000). Autonomy involves feelings of assertion and distinction from others, whereas connectedness involves permeability and mutuality with others (e.g., Grotevant & Cooper, 1998).¹ Most research examining individuation processes has examined real-time interactions and negotiations within families, and behavioral coding has focused on assertive (e.g., expressing one's opinion) and mutual (e.g., agreeing with others) behaviors. Research on individuation often operationalizes identity in terms of identity statuses, in which identity is viewed as progressing through stages of personal exploration and commitment to ideals and values (e.g., Marcia, Waterman, Matteson, Archer, & Orlofsky, 1993). Indeed, in these family negotiation tasks the statistical interaction of autonomy and connectedness predicts the identity achievement status, in which one has both explored and committed to an identity (Grotevant & Cooper, 1985).

Narrative Meaning Making

While individuation is a behavior-based approach to identity development, narrative approaches focus on the importance of developing a sense of personal coherence and continuity through time via the construction of a life story (e.g., Chandler, Lalonde, Sokol, & Hallett, 2003; Fivush, 2001; Habermas & Bluck, 2000; McAdams, 1988, 1993). Theoretical work (e.g., Bruner, 1990; McAdams, 1993) suggests that the way that we develop an understanding of ourselves, others, and the world around us is through the construction of narratives. Research on narrative development in childhood, however, shows that narrative skills and processes develop via relational processes (see Fivush, Haden, & Reese, 2006, for a review). Thus, while research on fully formed life stories has not explicitly examined conversational or interpersonal processes in the construction of the life story, developmentally, both individuation and narrative emerge in interpersonal contexts.

The two processes of meaning making that we examined were (a) the degree to which people connect past events to the self, and (b) the sophistication or complexity of what one has learned about the self through reflecting on the past. Pasupathi, Mansour, and Brubaker (2007) have defined self-event connections as explicit connections that are about self-stability or self-change (e.g., I am a good person, I became a better person).² Making these connections is a critical part of the process of integrating one's disparate autobiographical experiences into a coherent life story. McLean and colleagues (e.g., McLean, 2005; McLean & Pratt, 2006; McLean & Thorne, 2003; Thorne, McLean, & Lawrence, 2004) have defined the degree to which one has learned about the self in reflecting on past events as a second kind of meaning making, here termed sophistication. For example, one might report not gaining knowledge about the self in recalling the past, learning something basic about the self, such as a behavioral lesson, or at the most complex level, reporting an insight about the self, which often involves some kind of transformation in self-understanding.

In thinking about the distinctions between these two types of meaning making, we suggest that self-event connections underlie the process of creating an autobiographical account of a coherent self, while sophistication reflects the degree of complexity and engagement in that process of creating a coherent self. Thus, to construct a life story, one needs, at minimum, to find connections between experience and the self. Beyond that, one may draw on more complex narrative strategies to not only state connections, but also to explain one's reasoning about how one came to be the person that he or she is (e.g., Linde, 1993; Pals, 2006), reflecting a more sophisticated narrative account. Therefore, as we were interested in meaning-making processes more generally, as well as age-related differences in meaning making, we wanted to examine the most basic aspects of meaning making to capture age-related shifts as well as more complex processes.

Linking Meaning Making and Individuation

The two approaches to identity development (i.e., individuation and narrative) have different theoretical histories, but we suggest that there may be important relations between them. These themes may have a motivational quality for story-making, as an emerging sense of one's self as autonomous and/or transitions related to connectedness may motivate the construction of a life story. Of course, the converse may also be true: the process of weaving together one's experiences into a coherent narrative may provide new insights that lead to greater autonomy from and/or connectedness to others.

Nevertheless, we see a mild tension between these two approaches in terms of a focus on behavior (or action) versus story (see also Dunbar & Grotevant, 2004). That is while agreement and disagreement—behaviors and interactions that serve as markers of autonomy and connectedness—have clearly been shown to predict important aspects of identity, we suggest that it is through narrative that individuals make meaning of these interactions and incorporate them into one's identity. Indeed, this idea has also been proposed by Dunbar and Grotevant (2004), who suggested that “identity work” is “accomplished and presented through the process of narrative construction” (p. 138). Therefore, the first hypothesis of this study was that the interaction of autonomy and connectedness would predict greater meaning making. In contrast to previous work that has focused on behavioral interactions, in the present study we coded narratives for the independent presence of autonomy and connectedness. Given that narrative approaches to identity development are newer than the individuation approach, we wanted to ask some more specific questions about narrative development, particularly in relation to well-being and age.

MEANING MAKING AND WELL-BEING

We know that the process of individuation is often related to higher well-being in adolescence (e.g., Allen, Hauser, Bell, & O'Connor, 1994), which raises the question of whether narrative patterns of identity development are also related to well-being. In fact, there is a strong body of work on adult narratives showing that those adults who make meaning of past events are better adjusted in terms of self-esteem, depression, psychological well-being, physical health, psychological maturity, and life satisfaction (Bauer & McAdams, 2004; Bauer, McAdams, & Sakeda, 2005; King, Scollon, Ramsey, & Williams, 2000; McAdams, Reynolds, Lewis, Patten, & Bowman, 2001; Pals, 2006; Pals & McAdams, 2004; Pratt, Hunsberger, Pancer, Allard, & Alisat, unpublished data). Much of this research has focused on difficult life events, and researchers have shown that the ability to engage narrative processes to manage and resolve negative experiences is critical to positive functioning (Breen & McLean, 2009; King, 2001; Pals & McAdams, 2004).

Research with children has also found links between narrative processes and well-being. For example, Reese, Bird, and Tripp (2007) found that parents who referred to a greater number of positive emotions and evaluations in discussing past events had children with higher self esteem. Further, parents who used more explanations during conversations about past negative events and more explanations and confirmations of past positive events had children with higher self-esteem. However, in these studies parents were active scaffolders in talking about the past, and Fivush and colleagues have suggested that narrating past difficult events on one's own is a developmental accomplishment. For example, in a sample of children writing about stressful events, Fivush, Marin, Crawford, Reynolds, and Brewin (2007) found that those children who used more explanations in their narratives had lower levels of well-being. These authors suggest that in this writing task these children may not have had the necessary contextual support to create a coherent narrative. While this research suggests that scaffolding is necessary in order for children to establish well-being through narrating difficult experiences, we do not yet know how far this need for contextual support extends into adolescence.

Further muddying the matter of how meaning making might be related to adolescent well-being are cognitive developmental stages through which adolescents progress. While cognitive abilities become far more sophisticated and abstract in adolescence, Harter has shown that specific reasoning abilities relating to one's self-concept emerge at different stages of adolescence. For example, the ability to compare abstractions of self emerge in middle adolescence, but this new ability comes with some conflict as adolescents become aware of opposition in self-attributes (Harter & Monsour, 1992). By late adolescence, felt conflicts have begun to diminish, perhaps due to the cognitive abilities that facilitate integrating these characteristics into an abstract system, or unified self. While Harter's research has focused on adolescents' conceptions of their selves in the present, we suggest that thinking about one's self across time invites even greater apparent contradiction in one's selves than exists in thinking about one's self at the same age. In constructing a life story narrative, adolescents face what may be an especially daunting task: to integrate the inevitable physical, emotional, and cognitive changes that occur with development into one coherent self.

Given the complexities of past research in which (1) carefully scaffolded stories in childhood predict higher well-being (Reese et al., 2007), with the reverse being true for unscaffolded stories (Fivush et al., 2007); (2) the clear positive relation between meaning-making processes and well-being in adulthood (e.g., McAdams et al., 2001); and (3) the developmental stages of complex thinking about the self in adolescence (e.g., Harter & Monsour, 1992), we left open the question of how meaning making would be related to well-being in adolescent boys.

We also examined the relation between individuation and well-being, expecting to replicate past work. Indeed, prior research has found that the statistical interaction of autonomy and connectedness predicts positive changes in ego development and self-esteem over the course of adolescence (Allen et al., 1994; Grotevant & Cooper, 1985). However, we note that our method was different from past studies, as we were trying to understand the process of individuation within narratives, rather than in interpersonal interactions, which have been well-studied. We turn now to our final question, which concerned meaning making and age.

MEANING MAKING AND AGE

In terms of age, we expected that there would be age-related increases in meaning making, in terms of making any kind of self-event connections, but particularly for those about change, as well as for sophistication. We based this hypothesis on research findings concerning cognitive development in adolescence. As suggested above, in adolescence new cognitive structures emerge that are critical to the developing self-system (Harter, 2003, 2006; Harter, Bresnick, Bouchey, & Whitesell, 1997; Harter & Monsour, 1992), such as those relating to considering and testing possibilities (Inhelder & Piaget, 1958) and those that allow for the integration of single abstractions into higher order abstractions (Case, 1985; Fischer, 1980). While we were unable to predict how these cognitive changes might influence the relationship between narrative meaning making and well-being in adolescence, our expectations concerning the more general relationship between meaning making and age were clear: we hypothesized that these cognitive developments may underlie one's ability to think about the present self in relation to one's “selves” of both the past and the future (McAdams, 1988) and thus expected to see an age-related increase in meaning making. Indeed, in one of the only studies on life story development in adolescence, Habermas and Paha (2001) showed an age-related increase in the likelihood of making connections between the past and the present self. While this latter study is in line with the expectations of the current study, it was done on quite a small scale (N=12).

THE PRESENT STUDY

We had three hypotheses and one research question in the present study. First, we expected that the interaction of autonomy and connectedness would predict meaning making. Second, we asked the question of how meaning making would relate to well-being in early, middle, and late adolescence. We also examined the relation between individuation and well-being, expecting it to be positive. Finally, we expected that meaning making would show an age-related increase across adolescence.

METHOD

Participants

The sample was drawn from a private school for boys located in a rural area of Ontario, outside of the Greater Toronto Area. Given the difficulty of collecting narrative data with large sample sizes and the relative lack of data on adolescent narrative identity, we note that this all-male sample, though it is to some extent one of convenience, nevertheless presents an important opportunity to begin to explore the constructs of interest.

The majority of the students come from the local area, followed by other parts of Canada, and about 20% are international students representing 21 countries. Approximately one-half of the total students at this school are boarders, and approximately 16% receive some form of financial aid. Our sample consisted of 146 adolescents from this school, ranging in age from 11 to 18 years. Fifty-seven participants from our sample were boarders. The ethnicity of our sample was as follows: Caucasian (n=88), Asian (Chinese, South Asian, Japanese, Korean; n=30), and Other (Aboriginal, Black, Arab, Latino, Other, and Mixed; n=18).

Procedure

Letters were sent to all parents explaining the purpose of the study and requesting consent for their sons' voluntary participation. Hard copies of the letter were mailed to parents in the summer, and follow-up copies were emailed to parents in the fall. Of the 531 boys enrolled at the school, parental consent was obtained for 181 (34%). In the weeks before data collection, letters were sent to students who had parental consent explaining the purpose of the study and requesting their consent to participate; 82% (N=146) agreed to participate. Students were tested in two 90-minute sessions, with one session focused on survey measures and one session focused on the narrative assessment. Students were compensated for their participation with pizza following each testing session. Only measures related to the present study are described here.

Tasks and Measures

Depression

Level of depressive symptomatology (e.g., feelings of sadness, guilt, and sleep disturbance) was assessed through the Beck Youth Depression Inventory (Beck, Beck, & Jolly, 2001). The inventory presents respondents with 20 statements and asks them to indicate on a scale from 0=never to 3=always how often the statement has been true for them during the past 2 weeks (α=.90).

Self-esteem

Level of self-esteem was assessed through the 10-item Rosenberg Self-Esteem scale (Rosenberg, 1965). The Rosenberg scale asks participants to indicate on a 4-point scale from “strongly agree” to “strongly disagree” the extent to which they agree with statements such as “I feel that I have a number of good qualities” (α=.85).

Narratives

Participants were asked to write about four separate autobiographical memories: a high point, a low point, a turning point, and a continuity experience (in that order). A high point is a “peak experience” in a person's life that is seen as a highly positive emotional experience. A low point is defined as an experience during which the individual experiences extremely negative emotions, such as sadness, disappointment, fear, guilt, or shame. A turning point is an episode in the life story during which the individual experiences a significant change in understanding of the self. A continuity experience shows how one has stayed the same person over time. In all four narratives, participants were asked to describe the event, including what happened, when it happened, who was involved, what they were thinking and feeling, why the experience was significant, and what the experience says about the participant. The instructions for these narrative elicitations were modified from McAdams' guided autobiography instructions (McAdams, 2006: http://www.sesp.northwestern.edu/foley/instruments/guided/) to be made suitable for a younger group. For all analyses below, narrative codes were aggregated across memories for each participant.

Narrative Coding

All written responses were transcribed and participants' responses were scored by the first two authors. Reliability was conducted on 20% of the cases with an independent reliability coder who was blind to the age of the participants and the hypotheses of the study. Kappas and intraclass correlations for interrater reliability are reported for each code below.

Autonomy and connectedness

Narratives were coded for autonomy and connectedness with a 3-point scale for each dimension. Autonomy was defined as a self-focus in which the self is viewed as unique or different from others, often reflected by self-assertion, independence, mastery, or being better or worse than others (r=.73). Connectedness was defined as feelings of being in relation with others, or describing how one is connected to others (r=.71). This can be reflected in interactions with others, the importance of others, or how others connect to the self.

Self-event connections

Based on Pasupathi's system (e.g., Pasupathi & Mansour, 2006), each narrative was coded for self-event connections with three mutually exclusive categories: none, explanation, or change (overall κ=.84). No connection occurs in a narrative in which no explicit statement is made about how the event connects to the self (κ=.85). Explanation refers to a connection that explains a stable aspect of the self (e.g., “This event shows what a nice person I am”; κ=.82). Change refers to a connection about some kind of change in the self (e.g., “I became more independent”; κ=.87).

Sophistication of meaning

Each narrative was coded for sophistication of meaning, based on McLean and Pratt's (2006) recent adaptation of the system developed by McLean and Thorne (McLean, 2005; McLean & Thorne, 2003; Thorne et al., 2004). Narratives were coded according to a 4-point scale that assesses the level of complexity in autobiographical reasoning with acceptable reliability (r=.75). A score of 0 was assigned to narratives that contained no explanation of the meaning of the event to the self. Narratives were scored as 1 if there was mention of a specific lesson that the reporter learned from the event. A score of 2 was assigned to narratives that contained “vague meaning;” narratives of this sort describe some growth or change in the self, but the specifics of the change are not clear. Narratives were scored as 3 if there was evidence that the reporter gleaned specific emotional, psychological, or relational insight from the event that applies to broader areas of the reporter's life.

RESULTS

Preliminary Analyses

We first tested for ethnic and boarding differences on our main variables of interest. A one-way ANOVA with ethnicity as the predictor variable and narrative variables as outcome variables (sophistication, self-event connections, autonomy, and connectedness) showed that there were some differences by ethnicity, F(2, 133)=3.65, p<.05. Post hoc Tukey's tests showed that this difference was in the report of no connections, with those in the Other category (m=.69) scoring higher on this dimension than those in the Caucasian category (m=.41). Given that there were no other differences between ethnic groups and because the “other” category is difficult to interpret, ethnicity is not considered further. In terms of whether or not participants boarded, an independent t-test was run (board or not) for all of the narrative variables, and there were no significant differences between groups.

Although we present narrative data in an aggregated form to preserve statistical power and to be consistent with prior work (e.g., Blagov & Singer, 2004; McAdams et al., 2001; McLean, 2005; McLean & Thorne, 2003), for descriptive purposes we report the means of each of the narrative variables in Table 1. These data show that sophistication is most likely to appear in turning point memories. When self-event connections are present they are more likely to be about change in turning point and low-point memories, with change and explain connections relatively equivalent in high-point memories and, not surprisingly, explain connections being more common in continuity memories. Finally, autonomy is least common in low-point memories, and connectedness does not show great variation across memory types.

Table 1.  Means and Percents of Narrative Variables Across Memory Type

	High Point	Low Point	Turning Point	Continuity
Sophistication (0–3)	0.56 (1.00)	0.66 (1.16)	1.33 (1.27)	0.66 (1.17)
Change connections (%)	23	29	60	16
Explain connections (%)	18	13	7	31
No connections (%)	46	49	21	34
Autonomy (1–3)	2.26 (.77)	1.96 (.71)	2.44 (.74)	2.38 (.73)
Connectedness (1–3)	1.83 (.80)	1.70 (.81)	1.47 (.69)	1.49 (.69)

Degrees of freedom vary in the analyses that follow because some students either did not attend both sessions or did not complete all parts of the survey.

MEANING MAKING AND INDIVIDUATION

Table 2 shows the intercorrelations of the narrative variables. Notably, autonomy was positively related to sophistication and to change connections, and negatively to making no connection; connectedness was only related to sophistication, not self-event connections. Regression analyses were then conducted to test the unique and combined effects (i.e., main effects, interaction effects) of autonomy and connectedness on change connections and sophistication of meaning. Effects were tested sequentially: first, the main effects for autonomy and connectedness were entered into the model; then their interaction effect was tested. With respect to sophistication of meaning, there was both a main effect of autonomy, β=.30, p<.001, and of connectedness, β=.18, p<.05, and as expected, a significant interaction effect between autonomy and connectedness, β=.21, p<.05. As can be seen in Figure 1, connectedness was significantly and positively correlated to sophistication when levels of autonomy were high; when levels of autonomy were low, connectedness and sophistication were less correlated. With respect to change connections, there was a main effect of autonomy, β=.41, p<.001, but no main effect for connectedness and no interaction effect. Thus, meaning that involves learning something about the self is associated with perceiving the self as distinct from others, while simultaneously in relation to others, and constructing self-event connections is solely related to autonomy.

Table 2.  Intercorrelations of Narrative Variables

	1.	2.	3.	4.	5.	6.
^* p<.05; ^** p<.01.
1. Sophistication	—
2. No connections	−.40^**	—
3. Explain connections	−.13	−.31^**	—
4. Change connections	.60^**	−.40^**	−.14	—
5. Autonomy	.30^**	−.48^**	.13	.43^**	—
6. Connectedness	.18^*	−.14	.14	.01	−.01	—

Figure 1.  Autonomy and connectedness predicting the sophistication of meaning.

MEANING MAKING AND WELL-BEING

Regression analyses were conducted to test the unique and combined effects (i.e., main effects, interaction effects) of meaning making and age on well-being (averaged self-esteem and reverse-scored depression).³ We conducted two regressions, one with change connections and one with sophistication. Effects were tested sequentially: first, the main effects for change/sophistication and age were entered into the model; then their interaction effect was tested. In the first regression there was a main effect for change connections, β=.26, p<.01, but no main effect for age and no interaction effect. When sophistication was entered in place of change connections, there was a main effect of sophistication in predicting well-being, β=.28, p<.01, and no main effect of age. This main effect was qualified, however, by an interaction between age and sophistication in predicting well-being, β=−.26, p=.01. As can be seen in Figure 2, younger boys with more sophisticated meaning scored lower on well-being, but this inverse relation between sophistication and well-being disappeared by late adolescence. Thus, positive well-being appears to be a function of the presence of change connections, and the interaction of age and the sophistication of meaning, with sophistication relating to poor well-being in early adolescence.

Figure 2.  Meaning making (sophistication) and well-being.

We also examined individuation as a predictor of well-being using linear regression to predict well-being from autonomy, connectedness, and their interaction. Effects were tested sequentially: first, the main effects for autonomy and connectedness were entered into the model; then their interaction effect was tested. Unexpectedly, there were neither main effects nor a significant interaction effect.

MEANING MAKING AND AGE

Analyses of meaning variables and age were conducted treating age as a continuous variable and controlling for word count averaged across narratives. As can be seen in Figure 3, and as expected, with age there was an increasing likelihood of making connections overall, partial r(122)=.24, p<.01, and change connections, partial r(117)=.30, p<.001. Also as expected, the sophistication of meaning showed an age-related increase, partial r(117)=.19, p<.05. There were no age-related trends for either explanatory connections or no connections.

Figure 3.  Age and meaning making.

DISCUSSION

“Relationships are central, from the very beginning of and throughout life, to the constitution and expression of the self” (Josselson, Lieblich, & McAdams, 2007, p. 3).

Josselson's quote neatly expresses some of our central findings, which show that constructing an identity in adolescence involves the integration of self and relationships via the reflection on one's past. These results also showed that while viewing the self as changing was related to positive well-being across adolescence, early adolescent boys who reported more sophisticated meaning showed poor well-being, though this relation disappeared by late adolescence. Finally, this is one of the first studies to show that adolescent boys possess an increasing ability to engage in the complex processes involved in making meaning of the past.

Meaning Making and Individuation

The interaction of autonomy and connectedness predicted sophisticated meaning, suggesting that a major part of understanding oneself is perceiving self-distinctiveness from and relatedness to others, supporting research done in other areas (Allen et al., 1994; Grotevant & Cooper, 1985) and reflecting the idea that meaning making is both a personal and a relational process (Fivush et al., 2006; McLean, Pasupathi, & Pals, 2007; Pasupathi, 2001; Thorne, 2000). These results also echo research showing that the content of late adolescents' and emerging adults' self-defining memories is most often about relationships, particularly feelings of closeness and separation (McLean & Thorne, 2003).

For example, the following narrative comes from a 16-year-old we call Andy, who wrote about a breakup with his girlfriend:

It was significant because I learnt a lot about myself, had a chance to start fresh, I grew closer to my parents by sharing, I found out a lot of my friends cared … I did some serious thinking and reflecting→I grew SO MUCH and matured … I know what I want in life, I learned to share and express myself, I learned about the characteristics that I wanted in friends and in a partner, I grew to be a well rounded whole person …

In writing about this painful experience Andy describes how he was able to learn to express himself (an example of autonomy), while at the same time learning about and deepening his connection to others, nicely exemplifying the intertwining of these concepts.

Interestingly, this interaction effect was only seen in predicting sophistication. Change connections were predicted solely by autonomy, which suggests that seeing the self as transforming is related to the processes of assertion and distinction. In other words, distinguishing the current self from the past self (change) appears to be combined with distinguishing self from others (autonomy).

This is the first study to explore meaning making and individuation in adolescence and, as such, it is important to consider how these processes relate to one another. We suggest that individuation and meaning making may be complementary and interrelated developmental processes. While behavioral indicators of individuation are important, so are representational indicators, and we suggest that narrative may be a vehicle for the process of individuation. It is also possible that behavioral changes within relationships motivate the construction of a narrative identity to accommodate these new ways of relating into one's self-story. Finally, given the importance of others in helping to construct life stories (e.g., Fivush et al., 2006; McLean et al., 2007; Thorne, 2000), developing and maintaining a sense of connectedness to others may be critical to life story development and narrative meaning-making skills.

In trying to understand how individuation develops in relation to meaning-making processes, there are some plausible links between childhood narrative practices and adolescent perceptions of relationships. For example, elaborative narration encourages the development of the child's own perspective and feelings about the event (autonomy) while simultaneously creating a context of coconstruction in which the child is encouraged to participate in the conversation (connectedness; e.g., Cleveland & Reese, 2005; Fivush et al., 2006). Therefore, it is possible that even in early childhood parental reminiscence practices are implicitly scaffolding the skills needed for important adolescent developmental tasks.

Meaning Making and Well-Being

Interestingly, while research with adults has consistently found that meaning making is related to positive well-being and research with children has shown that experiencing careful scaffolding of narratives is related to positive well-being, our results suggest a somewhat more complex relationship between meaning making and well-being for adolescent boys. While change connections were related to positive well-being across the board, sophistication was related to poor well-being in early adolescence. Why do we see a difference between self-event connections and sophistication in predicting well-being? It is possible that self-event connections reflect the most basic aspects of autobiographical reasoning, which is a normative developmental process. Sophistication, however, involves reflecting on the processing of one's thinking, which is far more abstract (and perhaps also more difficult) for early adolescents who are not able to coordinate more than one abstraction related to the self (Harter & Monsour, 1992; Harter et al., 1997). Thus, engaging in these kinds of cognitive efforts before one has the skills to integrate abstractions may lead to felt distress.

It is also possible that the boys who reported that they have gained insights about their experiences in early adolescence are doing this out of necessity. That is, they may be experiencing especially disruptive events (such as parental divorce or marital conflict) that require abstract lessons and insights, but that may also cause lower well-being. It is interesting to note, however, that this is not the case in adult samples. That is, adults who learn from and resolve difficult life events tend to have higher well-being than those who do not, even if all are experiencing difficult life events (e.g., Pals, 2006).

In terms of the failure of individuation to predict well-being, as it has in past developmental studies (e.g., Allen et al., 1994), we suggest that this may be due to the fact that our method was not interpersonal in nature, but was rather representational, and perhaps captures different aspects of individuation compared with behavioral interactions.

Meaning Making and Age

While there was a linear increase in meaning making with age (for self-event connections overall, change connections, and sophistication), inspection of Figure 1 reveals that middle adolescence (around age 15–16) is a crucial time for increasing meaning making, which has now been replicated in a subsequent study (McLean & Breen, 2009). Indeed, middle adolescence is the stage in which adolescents are able to see contradictions in concurrent aspects of self (Harter & Monsour, 1992), and perhaps the ability to see contradiction in the self is critical to meaning making.

For example, Bob (11 years old) is an early adolescent who reported the following turning point narrative without meaning:

When I came to the [name of school] from my old school I was confused and sad. I was confused because I have never been there before and it was really big. I was sad because all my old friends were gone and I didn't know anyone.

Bob reports this as his turning point, but he does not explicate how he changed. Thus, while it is possible that he may see this as a transformative experience, he may not yet have the cognitive skills to integrate the story and to narrate it as such (at least in this context). He may also simply need more time to develop the meaning of this story, as this experience occurred within a year of his writing.

Moving through the adolescent stages Dennis, 15 years old, reported the following low-point:

My grandfather died at the start of school last year. I went down to meet him with my mom, dad, and sister but he died fifteen minutes before. I was very upset and so was my entire family. It felt very strange to know that he was gone. It is very significant to myself as a maturing person. I thought that it was terrible but it felt like it was meant to happen.

Dennis is able to draw some personal meaning from his experience, which is that this is important to his maturation, but his meaning is still vague. Further, consistent with expectations for the middle adolescent stage of cognitive development, he is able to see some contradiction (that this event was both terrible and meant to happen), but has not yet merged these contradictions into a story that explains his views or his self.

In contrast, 17-year-old Henry reported the following turning point:

My turning point happened shortly after grade eight. I had always been a goof in class, and never took class seriously. One time, a girl in my class was playing with my phone, and it rang. Instead of telling the teacher she had it … she tried to hide it, and pretend she found it, and I played with it alone. My teacher, realizing the lie, told us off. I then realized I had to be more mature, & responsible for my actions. It was a subtle turn, but since then I have been getting 80's in school, and have stopped being the clown in class.

Henry is tying this specific event to a change in how he views himself and in how he acts, which counts as a self-event connection about change and as high sophistication. He is able to create continuity in the self by connecting past experiences to reflections on the current self; he sees two aspects of himself (potential contradictions) and chooses a path to take in life, which he substantiates with this narrative. In comparison with Bob and to Dennis, Henry has more developed meaning in which he has integrated a past experience into an ongoing sense of himself as a person. This may be due to the more advanced cognitive abilities that he has as a late adolescent and perhaps also due to greater practice at storying himself.

One interesting question that emerges from these results is what is it that prompts adolescent boys to begin engaging in this process of meaning making from early to late adolescence? We suggest that there are likely both personal and interpersonal reasons motivating this process. As adolescents enter into the stage of identity crisis they are pressed, by both personal and cultural factors, to define the self in terms of continuity through time and hence to develop a life story (Erikson, 1968). Further, as boys begin to develop more intimate relationships with girls (as friends or romantic partners), they may be required by those audiences to begin to disclose meaning-laden stories about themselves for the purposes of intimacy development (e.g., McLean, 2005).

Gender

While we were not able to examine gender differences in this study, we feel it is important to raise issues that will be important for future studies with both boys and girls in adolescence, particularly in understanding the relation between meaning making and well-being. One of the contradictions in past research on meaning-making processes is that there are clearer gender differences in childhood than in adulthood, when meaning is defined as elaboration on the personal importance or interpretation of the event, as it was here (cf. meaning defined as internal state language; Fivush & Baker-Ward, 2005). For example, research on childhood narrative practices has shown that parents vary in how elaborative they are in narrating the past with their children (see Fivush et al., 2006, for a review). We view elaboration as a likely precursor to meaning making, as this kind of narration involves greater reflection than nonelaborative narration (McLean et al., 2007). Most relevant to the present study, research has shown that parents in Western cultures are more elaborative with daughters than with sons, particularly when talking about emotional experiences (e.g., Fivush, 1998; Reese, Haden, & Fivush, 1996). Further, girls' narratives are longer, more detailed, and use more language focused on cognition and emotion (see Fivush & Buckner, 2003, for a review), important aspects of meaning making. This suggests that boys may have less opportunity to develop the narrative skills in childhood that should enhance meaning making in adolescence.

In contrast, research with adult populations does not find gender differences in meaning-making processes. For example, studies have found no gender differences on meaning making (McLean, 2005; McLean & Pratt, 2006; McLean & Thorne, 2003), growth themes or integrative memories (Bauer & McAdams, 2004; Bauer et al., 2005), or the report of self-event connections (Pasupathi & Mansour, 2006).

Thus, we propose that our findings may reflect a process in which boys experience less practice at narrative elaboration in childhood, and catch up to girls over the course of adolescence, so that by young adulthood boys are equally savvy at meaning-making processes. Indeed, recent studies have shown that experiences within family conversations may lead to different relations with well-being for boys and girls in early adolescence (Fivush, Bohanek, Marin, & Duke, 2009). Thus, while our results cannot speak to gender differences, emerging data suggest that boys and girls may have different developmental trajectories concerning narrative identity development and well-being, which is an exciting arena for future research.

Limitations and Conclusions

The first limitation is the causal nature of these results, which we cannot determine. We have offered interpretations that demand longitudinal and experimental data, so we offer our interpretations with caution. Further, we have proposed explanations for some of our findings without assessing all related constructs, such as level of cognitive development. Further studies will need to assess all of these constructs to confirm or refute these interpretations.

There are also potential issues of generalizability. Though our sample was somewhat ethnically diverse, these data came from a private boys' boarding school, which might have impacted the results. For example, our findings about the importance of connectedness might be a result of the close relationships that are developed between boys in this particular context. In addition, although we were unable to assess social class through student reports, it may be important to some of our results, which we were unable to assess through student reports. It is also possible that, due to rigorous academic demands of this school and the resources of these families, these boys may have more developed verbal skills than would boys from other schools and backgrounds. If this latter possibility proves to be the case, then the potential vulnerabilities in the wider population of boys may be even greater. That is, if verbal abilities help to construct the self-story, then those with less developed abilities may be at an even greater risk. While we note that these limitations are important to consider, we also emphasize that these are the first data from a large-scale study to examine narrative development in adolescence. Future research in this area will be important to compare with this sample and to further elaborate on important processes at work in the development of narrative identity.

In conclusion, adolescence is a time when identity development takes center stage. Developing a coherent self involves individual as well as interpersonal processes, and these results show a link between these two approaches to identity development. These data also show that while adolescent boys are increasingly capable of engaging in the process of meaning making, parts of this process may be difficult at certain stages of adolescence. Indeed these results underscore the importance of examining different kinds of meaning-making processes to understand developmental change and to predict important outcomes, such as well-being. Overall, this study is part of a growing field devoted to the empirical examination of the importance of stories in our lives, and these data show that adolescence is a critical time in the development of these stories.

Footnotes

1. ¹ We note that while prior research has predominantly examined individuation in the context of parental relationships, we examined individuation in the context of parent and peer relationships, that is, we examined feelings of autonomy and connectedness in all close relationships that were reported.

2. ² Pasupathi and colleagues have also written about connections focused on dismissing aspects of the self or revealing aspects of the self. Because of the low base rate of reveal connections, we included those within change (see Pasupathi & Mansour, 2006), and we did not see any instances of dismiss connections.

3. ³ We ran all regressions with self-esteem and depression as separate outcome variables, and results were identical to the averages results.

ACKNOWLEDGMENTS

We thank the Narrative Lab at the University of Toronto for data collection, data entry, transcription, and discussions concerning these data and Ivan Service for reliability coding. We also thank the school administrators and students for their participation in this study. This study was supported by separate Social Sciences and Humanities Research Council grants to the first and third authors and a Social Sciences and Humanities Research Council Doctoral Fellowship to the second author. Monisha Pasupathi provided helpful comments on an earlier version of this manuscript.