Empirical Article review

Personality Development From Adolescence to Emerging Adulthood: Linking Trajectories of Ego Development to the Family Context and

Identity Formation

Moin Syed University of Minnesota

Inge Seiffge-Krenke University of Mainz

This longitudinal study analyzed personality development using an individual approach by examining changes in ego development across the transition from adolescence to emerging adulthood. Specifically, the study mapped the heterogeniety in ego development growth trajectories and linked the different trajectories to the family context in adolescence and identity development in emerging adulthood. Participants were 98 families with a child who were followed from age 14 to age 24. Latent class growth analysis identified 4 distinct trajectories of growth in ego development of the children over the 10-year period. The results indicated that growth was more rapid during adolescence and tended to taper off in emerging adulthood. In addition, promotion of personal growth within the family and parents’ ego development were particulary instrumental in children’s ego developmental gains in adolescence. Finally, youth who demonstrated continued ego development into emerging adulthood also demonstrated heightened levels of identity exploration.

Keywords: ego development, identity status, family climate, emerging adulthood, individual devlopment

Social scientists have recently highlighted the changing nature of adulthood in industrialized societies (Arnett, 2000; Côté, 2006). Where previously the transition from adolescence to adulthood was clearly marked by entry into stable adult roles and deriving a sense of independence in the early 20s, these developmental mile- stones have been substantially prolonged until age 25–30. Arnett (2000, 2004) has argued that these societal changes have resulted in profound changes in the psychological experience of these young people, or emerging adults. For example, meta-analytic research suggests a delay in identity development and great vari- ation with which mature states of identity are accomplished (Kroger, Martinussen, & Marcia, 2010). Thus, it is critical to understand whether personality development likewise has under- gone similar changes, and how personality develops within impor- tant developmental contexts. We take up this task in the present study by charting individual trajectories of ego development, a

fundamental component of personality (Blasi, 1998), across the transition from adolescence to emerging adulthood, and examine how an important context in adolescence, the family, and key task of emerging adulthood, identity, are related to these trajectories.

Personality Change: Integrating a Developmental Perspective

There is substantial empirical evidence that personality is char- acterized by both stability and change across the life span (Caspi, Roberts, & Shiner, 2005; McAdams & Olson, 2010). Personality change is most frequently assessed in one of two ways: differential continuity, or the degree to which individual differences in trait levels maintain their rank order over time (Roberts & DelVecchio, 2000), and mean levels of traits over time, indicating the average developmental course of traits across the life span (Roberts, Wal- ton, & Viechtbauer, 2006). There are, however, at least two addi- tional methods for assessing stability and change that have re- ceived much less attention in the literature: ipsative change and individual change (Caspi & Roberts, 1999; Donnellan, Conger, & Burzette, 2007; Ozer & Gjerde, 1989; Roberts, Caspi, & Moffitt, 2001).

Ipsative change is similar to differential continuity, but exam- ines stability and change in personality profiles within an individ- ual (e.g., Lönnqvist, Mäkinen, Paunonen, Henriksson, & Verkasalo, 2008). In contrast, individual change is similar to the mean-level change approach, but examines individual variations in change in personality over time that may deviate from the average trajectory (e.g., Donnellan et al., 2007). These less frequently examined forms of change are in need of greater attention given the substantial heterogeneity that has been observed in studies of differential continuity and mean-level change (Roberts & DelVec-

This article was published Online First September 17, 2012. Moin Syed, Department of Psychology, University of Minnesota; Inge

Seiffge-Krenke, Institute of Psychology, University of Mainz, Mainz, Germany.

The research reported here was supported by Bundesministerium für Forschung und Technologie Grant BMFT 0706567 awarded to Inge Seiffge-Krenke and by Deutsche Forschungsgemeinschaft Grant DFG SE408/10-1, 10-3.

Correspondence concerning this article should be addressed to either Moin Syed, Department of Psychology, University of Minnesota, Twin Cities, 75 East River Road, Minneapolis, MN 55455-0344, or Inge Seiffge- Krenke, Johannes Gutenberg-Universität Mainz, Psychologisches Institut, Abteilung Entwicklungs- und Pädagogische, Psychologie Wallstraße 3, 6. OG, D-55122 Mainz, Germany. E-mail: [email protected] or seiffge- [email protected]

Journal of Personality and Social Psychology © 2012 American Psychological Association 2013, Vol. 104, No. 2, 371–384 0022-3514/13/$12.00 DOI: 10.1037/a0030070

371

chio, 2000; Roberts et al., 2006). Indeed, studies of personality development adopting an individual approach have documented substantial individual variability for most personality traits. Im- portantly, these individual patterns can sometimes indicate a very different developmental picture compared with the mean-level change for the sample. For example, Donnellan et al. (2007) highlighted how for traditionalism, a quarter of the sample in- creased and a quarter of the sample decreased, resulting in no mean-level change over time. Thus, although understanding aver- age trajectories of personality change is important for gaining a sense of the normative developmental course of personality, indi- vidual level analyses may allow for a more nuanced and precise view of the causes, consequences, and correlates of personality development.

Despite the insights gained from studies of individual change in personality, the existing longitudinal research has consisted pri- marily of only two waves of assessment and characterized change in terms of increases and decreases from Time 1 to Time 2 (Donnellan et al., 2007; Lönnqvist et al., 2008; Ozer & Gjerde, 1989; Roberts et al., 2001). In their review, Caspi et al. (2005) noted this limitation and called for studies that include more than two waves to make use of sophisticated methods of mapping change over time (e.g., latent growth curve modeling, multilevel modeling). Such methods allow for addressing questions of great importance for the study of personality development, such as examining different personality pathways and types of change (e.g., big increases, small increases, curvilinear change), as well as the social-cognitive and contextual factors that are associated with the different pathways (McAdams & Olson, 2010).

One of the few existing studies to adopt such an approach is Johnson, Hicks, McGue, and Iacono’s (2007) study of individual personality trajectories among adolescent girls. The researchers used growth mixture modeling to identify personality trajectory subgroups over a 10-year period and then subsequently examined differences among the groups in terms of antisocial behavior and substance abuse. This approach allows for researchers to explicitly model the heterogeneity of personality development that can lead to new and interesting hypotheses about personality stability, change, and their correlates. Given the fluid nature and the high diversity found in several life domains during emerging adulthood, a developmental perspective can inform changes in personality in this important transition period.

Many authors have discussed the potential benefits for a stron- ger link between personality and developmental psychology (McAdams, 2001; Roberts & Caspi, 2003; Smith, 2005; Thorne, 2004; Westenberg, Blasi, & Cohn, 1998). Wohlwill’s (1973) clas- sic articulation of developmental science specifies that after de- scribing the developmental course of a particular construct, the next step is to understand concurrent psychological processes associated with the observed developmental trajectories. As de- tailed above, research on the developmental course of personality has generally been descriptive, and has placed relatively less emphasis on understanding why the observed changes may be occurring (McAdams & Olson, 2010). A developmental perspec- tive on personality change highlights the need to understand what else is going on at the same time through an examination of the contexts that may be associated with the observed changes (see Neyer & Asendorf, 2001, for the role of relationship status in

personality development). Doing so facilitates a move from de- scription and toward explanation.

Loevinger’s (1976) conception of ego development is an ideal construct for examining personality from a developmental per- spective. Ego development corresponds to the frame of reference through which an individual interprets the self and relationships with others, and has been considered a fundamental component of personality (Blasi, 1998). Its ties to developmental conceptualiza- tions of selfhood is strong, as ego development is a personality model that incorporates notions of individual differences within cohorts with an examination of age-graded qualitative change (Westenberg et al., 1998). As described in more detail below, ego development provides a framework for understanding how and why individuals may develop at different rates. Indeed, unlike most research on change in personality traits, research on ego development has a long history of mapping individual change in development (e.g., Hauser, Borman, Jacobsen, Powers, & Noam, 1991). Thus, in the present study we examined the contexts of individual trajectories of ego development across the transition from adolescence to emerging adulthood as a means for illustrating a fruitful collaboration between personality and developmental science.

Ego Development

Loevinger (1976) conceptualized the ego as the primary synthe- sizing and regulating agent in individuals’ personality develop- ment. Accordingly, ego development fundamentally corresponds to individuals’ mastery of their selves and social contexts (Blasi, 1998). Similarly, McAdams (1998) has referred to the ego as the “master orchestrator” of personality and identity, serving to pro- vide individuals with a subjective sense of their personality, viewed either as traits, motives, or an integrated life story. In this way, ego development is a process of meaning-making within progressive developmental contexts (Blasi, 1998; McAdams, 1998). More concretely, Loevinger’s (1976) model of ego devel- opment suggests an empirically grounded normative developmen- tal sequence that comprises nine levels through which individuals can theoretically pass; Presocial, Impulsive, Self-Protective, Con- formist, Self-Aware, Conscientious, Individualistic, Autonomous, and Integrated. Each level is considered qualitatively different from those adjacent, as the levels represent particular styles of synthesizing the world around. The levels are, however, also quantitative, in that each successive level is considered to be more sophisticated and adaptive than the previous; thus, the levels describe an increasingly complex view of the interaction between self and society. Taken together, ego development corresponds to a developmental pattern of differential meaning, mastery, and self-understanding.

Despite the clearly articulated developmental nature of ego development, the majority of research has been cross-sectional. There have, however, been several longitudinal studies investigat- ing change over time in ego development (Cohn, 1991; Gfellner, 1986; Hauser et al., 1991; Lanning, Colucci, & Edwards, 2007; Loevinger, Cohn, & Bonneville, 1995). Cohn (1998) meta- analyzed the 16 available longitudinal studies by correlating age and ego development level separately for two age groups: adoles- cence and college age. The findings indicated that age is more strongly related to ego development during adolescence compared

372 SYED AND SEIFFGE-KRENKE

with the college years. These findings converged with Cohn’s (1998) meta-analysis of cross-sectional age comparison studies, suggesting that ego development primarily develops during ado- lescence followed by stabilization in emerging adulthood. In a separate meta-analysis, Cohn (1991) found that girls score higher than boys on ego development in middle and high school but that this gender difference is no longer apparent in emerging adulthood and beyond, suggesting that girls progress more quickly through the stages than boys, but ultimately they arrive at the same level.

Although Cohn’s (1998) meta-analyses provide clear evidence of more substantial development during adolescence than during emerging adulthood, the studies can be more accurately described as comparisons between longitudinal studies in adolescence and longitudinal studies in emerging adulthood. That is, none of the studies included in the meta-analysis actually assessed the same individuals across the transition. In fact, we could locate only one study that examined how ego development changed across the transition from adolescence to emerging adulthood (Westen- berg & Gjerde, 1999). That study, which covered nearly the same period of the life span (14 –23) as in the present study (14 –24), yielded two important findings that inform our expectations in this study. First, consistent with prior work (e.g., Gfellner, 1986), they found that Level 5, self-aware, served as an important develop- mental marker after which progress tended to slow. This finding is important, as it suggests that development may stabilize once individuals reach a particular level, rather than a particular age, per se, as suggested by Cohn (1998). Second, they found that vari- ability in ego development levels was greater at age 23 compared with age 14, suggesting heterogeneity in how ego development unfolds across the transition to adulthood, consistent with an individual approach to personality development.

Despite these important findings, the study was limited by having only two time points. Thus, Westenberg and Gjerde (1999) were unable to map the different trajectories that individuals followed between ages 14 and 23. Indeed, nearly all longitudinal studies of ego development are based on only two waves of data (see Cohn, 1998), leaving many unanswered questions, including: At what point are the greatest gains observed? Is there continuous progression in ego development during emerging adulthood? What proportion of youth tend to plateau at the self-aware level? And at what point does variability in ego development begin to increase? These are some of the questions that we were able to address in the present study.

Individual Change in Ego Development

There has been some research on individual trajectories of ego development. Hauser and colleagues (Billings, Hauser, & Allen, 2008; Hauser et al., 1991; Henninghausen, Hauser, Billings, Schultz, & Allen, 2004) have described six theoretical trajectory types for how individuals change in ego development over time, three of which represent stability (stable high, moderate, and low) and three represent change (increase, decrease, and fluctuating). Empirically, they have found evidence for all of the trajectory classes except for decreasing, with the stable moderate trajectory being the most common among healthy adolescents. Although their work suggests the feasibility of an individual approach to personality development through an examination of ego develop- ment trajectories, it has been limited by focusing on relatively

short time frames (e.g., 1–2 years). Accordingly, the first goal of the present study was to map the variability in individual trajec- tories of ego development from age 14 to age 24. On the basis of the existing literature, we expected that the majority of the sample would demonstrate increases in ego development from adoles- cence to emerging adulthood and that growth would be more pronounced during adolescence than emerging adulthood. How- ever, consistent with our individual approach to personality devel- opment and the work of Hauser, we expected that there would be groups who deviate from the general pattern previously observed in the literature. In particular, we hypothesized that there would be groups of youth who defied the overall increasing trend and remained relatively stable over time. This hypothesis is consistent with past research that has indicated greater variability in ego development in early adulthood compared with adolescence (Gfellner, 1986; Westenberg & Gjerde, 1999), suggesting the possibility of many different personality pathways being repre- sented.

Our analysis of individual ego development trajectories builds on existing work in four important ways. First, most studies investigated a relatively short time span and thus possibly did not capture changes across time, as ego development—the integration of inner perceptions of people and events—requires time to unfold (Loevinger, 1976). Second, most studies focused exclusively on adolescence or some segment on adulthood, meaning that they did not study change across an important life transition. Third, most research has focused on samples of college or university students, thereby neglecting other groups of same-aged individuals who are working, unemployed, or in apprenticeships (i.e., the “forgotten half,” as termed by Rosenbaum, 2001). Finally, the research on individual trajectories by Hauser and colleagues was conducted by sorting participants into categories “by hand,” rather than using statistical procedures to identify the trajectory types. In the present study, we identified trajectory groups using latent class growth analysis, a statistical method for identifying heterogeneous trajec- tories of change (Nagin & Tremblay, 2001).

In addition to correcting for the limitations of previous studies, in the present study we go beyond describing ego development trajectories from age 14 to age 24 by linking this aspect of personality development to other aspects of psychosocial develop- ment. As it is suggested that there is more than one way to traverse the developmental period from adolescence to emerging adulthood with respect to ego development (Hauser, 1991; Westenberg & Gjerde, 1999), we expand our focus to key developmental contexts during these two periods. As described in detail below, we ex- plored how parents and the family environment during adoles- cence could lead to gains in ego development. In emerging adult- hood, we examined how ego development is associated with a central developmental task of this period, developing a sense of identity (Arnett, 2000; Erikson, 1968). Thus, in the present study, we investigated developmentally appropriate antecedents and con- sequences of youths’ ego development trajectories.

Predictors in Adolescence: Family Environment and Parents’ Ego Development

Research in developmental psychology has highlighted the im- portance of the family context during adolescence for a number of positive outcomes, including well-being, emotion regulation, aca-

373TRAJECTORIES OF EGO DEVELOPMENT

demic achievement, and autonomy (Grusec & Davidov, 2010; Laursen & Collins, 2009; Steinberg, 2001). Despite this vast research, there have been relatively few studies that have examined family influences on children’s ego development trajectories. This is surprising, as Loevinger (1976) suggested that movement along the ego development path comes about through social interactions that challenge individuals and require them to think deeply about the relation between self and others (see also Thorne, 1993). Indeed, Loevinger (1976) offered the concept of pacers as impor- tant to the ego development process. Pacers are individuals who, by virtue of their higher ego development, challenge others in ways that facilitate their ego development. Given that parents are both proximal to and more developmentally advanced than their adolescent children, they likely serve as pacers for their children’s ego development.

A handful of empirical studies support the idea that parents play an important role in their children’s ego development. These studies generally indicate that adolescents have greater ego devel- opment when their parents cognitively challenge them within the context of a warm and supportive relationship (Allen, Hauser, Bell, & O’Connor, 1994; Hauser et al., 1991). In one of the only longitudinal studies on the topic, Allen et al. (1994) found that adolescents who had fathers—not mothers—who interacted in this way exhibited greater gains in ego development over a 2-year period. This may be explained by the distinctive function fathers serve in contrast to mothers, that is, by supporting an independent view and an autonomous development of their adolescent child (Shulman & Seiffge-Krenke, 1997). This finding raises an impor- tant point: the differential influence of mothers and fathers for ego development. On this point, the literature has been inconsistent. In contrast to Allen et al.’s (1994) findings, Hauser et al. (1991) found evidence for a link between adolescent ego development and both mothers’ and father’s tendency to challenge their children in a supportive way. The concept of pacers can be enacted both within the parenting and family context, such as in the studies described above, but also in terms of the individual characteristics of the parents, such as their own level of ego development. However, to our knowledge, no study has actually tested direct links between parents’ levels of ego development and their chil- dren’s levels of ego development.

Taken together, there is some evidence that a family climate that is supportive yet able to challenge and grant freedom is important for the adolescent’s progression in ego development. Accordingly, the second goal of the present study was to test the hypotheses that (a) a more supportive family climate in adolescence will be asso- ciated with more progressive ego development trajectories and (b) that mother’s and father’s ego development would be predictive of the adolescents’ ego development trajectories.

Correlates in Emerging Adulthood: Identity Commitment and Exploration

Identity has long been considered a central developmental task of adolescence and emerging adulthood (Erikson, 1968; Marcia, 1966). Identity from an Eriksonian perspective is defined as an internal sense of continuity and coherence across time and life domains. In creating the identity status model, Marcia (1966) operationalized a portion of Erikson’s theory and highlighted the interrelation of two key identity processes: exploration and com-

mitment. Identity exploration refers to individuals’ active search and consideration of various identity alternatives. Commitment indexes individuals’ selecting a particular identity to adopt, along with all that identity entails. Marcia proposed that considering both identity processes simultaneously provides insight into youth’s current developmental identity status: Those who have gone through a period of exploration and have committed to an identity are called achieved; individuals who are actively exploring their identities without making commitments are in moratorium; those who commit to an identity without engaging in the exploration process are foreclosed; and those who are neither exploring their identities nor committed to an identity are diffused. Present re- search using the identity status model uses both the four-status typology and the continuous measures of exploration and commit- ment (Bosma, 1992; Kunnen, 2009; Meeus, 2011).

Although the development of the identity status model has led to hundreds of empirical studies documenting its correlates and de- velopment trajectories (for reviews, see Meeus, 2011; van Hoof, 1999; Waterman, 1999), surprisingly little is known about how identity processes are related to ego development. We consider this fact surprising because both the identity status model and ego development draw from portions of Erikson’s (1963) life span theory of psychosocial development, and Loevinger (1976) con- ceptualized ego development as subsuming Erikson’s notions of ego-identity. Given that ego development has been referred to as the “master orchestrator” of personality (McAdams, 1998), devel- opmental gains in ego development should correspond to devel- opment gains in identity development.

There is some indirect evidence for this hypothesis in the extant literature. In an early longitudinal study, Adams and Fitch (1982) charted the changes in college students’ identity status and ego development over 1 year. Their results indicated gains in both constructs across the year, leading them to conclude that the same psychological processes underlie each. However, they did not examine how the participants’ identity status classification and ego development were explicitly linked. Bauer and McAdams (2010) recently linked indicators of narrative identity to increases in ego development in a sample of emerging adults. In particular, those with higher ego development also had narratives containing evi- dence of personal growth, in terms of deeper knowledge of oneself and a quest for knowledge. As research has indicated similar processes underlying identity status and narrative identity (McLean & Pasupathi, 2012; Syed, 2012), these findings can be taken as indirect evidence for the link between identity status and ego development. In the present study, however, we explicitly tested this association. In particular, we hypothesized that more progressive trajectories of ego development would be associated with greater likelihood of classification as identity achieved in emerging adulthood, as well as higher levels of both identity exploration and commitment.

The Present Study

Recent research has substantiated the emergence of a new developmental phase between adolescence and young adulthood, which is characterized by a postponement and delay in many developmental areas, including identity and self-exploration (Ar- nett, 2004; Kroger et al., 2010). This study sets out to investigate stability and change in a fundamental component of personality

374 SYED AND SEIFFGE-KRENKE

development, ego development. Accordingly, the purpose of the present study was to document trajectories of ego development across the transition from adolescence to emerging adulthood, investigate how the trajectories are influenced by the family con- text in adolescence (i.e., family climate and parents’ ego develop- ment), and link the trajectories to identity development in emerg- ing adulthood. Three main research questions guided the present study:

1. Through the use of latent class growth analysis (LCGA), we aimed to identify several trajectories of ego development from adolescence and emerging adulthood. On the basis of earlier lon- gitudinal studies (Hauser & Safyer, 1994), we expected that the majority of the sample would demonstrate increases over time in ego development but that some individuals would deviate from the normative path by remaining stable over time.

2. The degree to which parents serve as pacers for ego devel- opment was investigated in two ways. First, we examined the influence of an autonomy-supporting family climate on ego devel- opment trajectories. On the basis of the extant literature (Allen et al., 1994; Laursen & Hafen, 2010; Steinberg & Morris, 2001), we hypothesized that family climates characterized by high levels of cohesion and open expression of emotions as well as support for independence would be related to the child’s progressive ego development from adolescence to young adulthood. Conversely, we predicted that family climates characterized by high levels of control would be associated with a relative lack of developmental progression. Second, we examined whether the different ego de- velopment trajectories of the children are related to their parents’ ego development. We expected that parents’ higher level of ego development would be related to their child’s progressive trajec- tory.

3. We examined whether the developmental trajectories of the children were related to their identity development in emerging adulthood. We expected that more progressive trajectories of ego development would be associated with greater likelihood of clas- sification as identity achieved in emerging adulthood as well as higher levels of both identity exploration and commitment.

Method

Participants

Participants were taken from the German Longitudinal Study (Seiffge-Krenke, 2003), which received full Institutional Review Board approval from the University of Mainz, Germany. The original sample included a total of 145 adolescents (52% girls) and their families; 81% of the participants were raised in two-parent families, and 19% were raised in single-parent or divorced fami- lies. For this study, a subsample of 68% was drawn, including complete data sets of two-parent families; altogether, 98 adoles- cents (M age � 14.7 years, SD � 1.38) and their mothers and fathers participated in a longitudinal study spanning 10 years. Assessments were made in a total of four waves with varying intervals, at age 14, 15, 17, and 24 years of the child. Participants came from broad socioeconomic strata: 52% of the families were middle class. All of the fathers and 65% of the mothers were employed. Mean age of fathers at Time 1 was 44.6 years (SD � 1.9); mean age of mothers was 41.6 years (SD � 2.3). All of the participants resided in western Germany, and 92% were of German

descent. At Time 1, 51% of the children were female, and 48% were male, attending seventh grade in secondary school. At Time 4 (M age � 24.8 years, SD � 1.23), all children had finished their secondary school education; 42% were enrolled in vocational training; 39% percent were employed, 16% were studying, and 3% were homemakers or unemployed. At age 24, only 15% of the children continued to reside with their parents. As in most longi- tudinal studies, data were missing at different time points for different participants. For the present study, 7% of the adolescent data, 8% of the mother data, and 10% of the father data were missing. Comparisons of participants with and without missing data using t tests and chi-squares indicated no differences in age, gender, parents’ marital status, family’s socioeconomic status (SES), or type of school the participant attended. To minimize the bias associated with this occasional attrition, we used expectation maximization algorithm to impute missing data. Little’s (Little, 1988) missing completely at random tests indicated that missing values could be reliably estimated.

Measures

Ego development. The Washington University Sentence Completion Test (WUSCT; Hy & Loevinger, 1996) was used at Waves 1, 2, 3 (adolescent version; Form 2–77), and 4 (adult version; Form 81) for children and at Wave 2 (adult version; Form 81) for parents by using a gender-specific version for males and females. Altogether, 36 incomplete sentences (e.g., “When I am criticized . . ., ” “Men are . . .”) had to be completed. Responses to sentence stems were coded by two raters using the manual by Hy and Loevinger (1996) and was based on the model of ego devel- opment by Loevinger (1985) entailing nine ego stages (impulsive stage, self-protective stage, conformistic stage, self-aware stage, conscientious stage, individualistic stage, autonomous stage, and integrated stage). Typically, adolescents exhibit modal ego levels from self-protective (i.e., to control self and others in order to further develop one’s own interest) to conformist (i.e., attuned to the needs, expectations, and opinions of others), whereas the modal level for young adults varies between self-aware (i.e., awareness of being different from others) and conscientious (i.e., a strong sense of responsibility for one’s thoughts and values), with females scoring higher than males (Westenberg & Gjerde, 1999). Answers were transcribed completely, made anonymous, and randomized across the two independent raters. Kappa ranged from .63 to .83 for the individual items across 30 randomly selected protocols. A third rater provided the consensus scoring in case of disagreement. Furthermore, in line with Hy and Loevinger (1996), for each participant the total protocol rating (TPR) was calculated, which is based on the cumulative frequency distribu- tion of the item ratings. Higher TPR scores indicate more advanced ego development. For the present analyses, the TPR ratings of the child from Waves 1, 2, 3, and 4 were used. The 36 TPR items had an internal consistency (Cronbach’s alpha) of .75 (adolescent version) and .78 (adult version). For fathers and mothers, the TPR ratings from Wave 2 (adult version) were used, with an internal consistency of .81 for mothers and .77 for fathers.

Family climate. Both parents and children completed the Family Environment Scale (FES), as developed by Moos and Moos (1981). Reliability and validity data of the German version are provided by Schneewind (1987). The FES measures how the

375TRAJECTORIES OF EGO DEVELOPMENT

family climate is perceived according to the following scales: 1. cohesion, 2. expression, 3. conflict, 4. independence, 5. achievement-orientation, 6. intellectual-cultural orientation, 7. ori- entation to active leisure-time activities, 8. moral-religious orien- tation, 9. organization, and 10. control. The 10 scales can be consolidated into three general dimensions, namely, “Interpersonal Relationships” (Scales 1–3; sample item: “There is a kind of togetherness in our family”), “Personal Growth” (Scales 4 – 8; sample item: “Our family encourages independence”), and “Sys- tem Maintenance” (Scales 9 and 10; sample item: “In our family it is very important to follow certain rules”). The first dimension assesses the degree of commitment, help, and support family members provide one another. The second dimension assesses the amount of independence and autonomy experienced in the family. The third dimension assesses how much rules and structure are used to run family life. Items were rated on a scale ranging from 1 (little or not true) to 5 (mostly true). Because we consider the family as a dynamic unit, we combined the information of both parents and adolescents into one family score. For each dimension, we calculated the average of adolescent, mother, and father scores. Previous research on parent–adolescent relationships demonstrated that combining scores of multiple informants significantly reduces rater bias and Type I error rate (Holmbeck, Li, Schurman, Fried- man, & Coakley, 2002; Schwarz, Barton-Henry, & Pruzinsky, 1985). Cronbach’s alphas in the present study for the three dimen- sions were .81, .85, and .82 for the child’s report; .86, .88, and .84 for mothers’ report; and .80, 88, and .83 for fathers’ report.

Identity status. As a measure of identity status, we used the Identity Status Interview (ISI; Marcia, 1966), which contains semi- structured questions pertaining to three life domains (career, rela- tionships, view of the world), which were coded according to two main identity dimensions: exploration and commitment. On the basis of the criteria for exploration (e.g., activity directed toward gathering information, evidence of considering alternative poten- tial identity elements) and commitment (e.g., activity directed toward implementing the chosen identity element, identification with significant others, projection of one’s personal future), de- tailed in Marcia (1993) and Waterman (1993), every child at Wave 4 was given a 1– 4 rating for both exploration and commitment and subsequently assigned to one identity status: achievement, mora- torium, foreclosure, or identity diffusion. Interrater agreement (kappa) across 30 randomly selected interviews ranged from .76 to .80. Again, a third rater provided consensus scoring in case of disagreement. For this study, we used both the continuous mea- sures of exploration and commitment and the dichotomized achievement codings (0 � not achieved; 1 � yes achieved) of relational identity and global identity (mean across the three do- mains).

Procedure

The present study was based on data collected when the children were adolescents (Wave 1, age 14; Wave 2, age 15; and Wave 3, age 17, respectively) and a decade later, when the child was an emerging adult (Wave 4, age 24). When the participants were 14, 15, and 17 years old, their ego development was assed via the WUSCT. In addition, at age 24, they completed the WUSCT to assess ego development. Furthermore, at age 24, the ISI was conducted to assess the identity status. As regards parents’ ego

development, father’s and mother’s ego development were as- sessed by the WUSCT when their child was 15 (Wave 2). In addition, fathers, mothers, and children completed the FES when the child was 14, 15, and 17 years. Participants were visited at home by trained research assistants at each time point and were asked to fill out questionnaires. In addition, the Marcia Interview was conducted. All participants provided written informed con- sent. Adolescents and parents were requested to complete the questionnaires independently.

Results

The analysis was conducted in three steps to address the three objectives of the study. First, we analyzed participants’ ego devel- opment data using LCGA to identify heterogeneous trajectories of change over the 10-year period. Second, we compared the different trajectory groups identified through the LCGA model for varia- tions in family climate and in mother’s and father’s ego develop- ment at the time of the participants’ adolescence (i.e., at the beginning of the trajectory). Third, we compared the different trajectory groups for variations in identity exploration and com- mitment when the participants were emerging adults.

Descriptive Statistics and Correlations

Descriptive statistics and bivariate correlations for all study variables are reported in Table 1. In terms of the child’s ego development, the average at Times 1 and 2 was close to the conformist stage, whereas the average at Time 3 corresponded to the self-aware stage, and the average at Time 4 was between self-aware and the conscientious stage. Thus, on average, the sample gained 1.12 steps in ego development across the 10 years, similar to previous research (Westenberg & Gjerde, 1999). In contrast, both mother’s and father’s ego development was just above the conformist level.

The bivariate correlations for child’s ego development revealed modest relations between adjacent time points (rs � .42–.54). The correlation between Time 1 and Time 4 child ego development, however, was much lower (r � .24), which is suggestive of heterogeneity in the growth of ego development over time. Look- ing at the correlations between child ego development and our other study variables, the correlations between both mother’s and father’s Time 2 ego development were stronger at Time 4 than at Time 1. In contrast, the correlations between child’s ego develop- ment and the Time 1 FES family scores were stronger at Time 1 than Time 4 for the Interpersonal Relations and Personal Growth subscales, but identical for the System Maintenance subscale. Finally, Time 4 identity achievement and identity exploration were more strongly related to child’s ego development at Time 1 than at Time 4.

Although these bivariate correlations are useful for describing the sample, they do not take into account the diverse, nonlinear paths of ego development that children may follow. The associa- tions with family factors and identity that are manifest in the bivariate world may be quite different when examining the asso- ciations with different trajectories of growth over time. That is, our interest in the present study is in how family and identity are related to the trajectory of ego development across the 10 years, rather than associations at one particular point in time.

376 SYED AND SEIFFGE-KRENKE

Identifying the Trajectories of Ego Development

LCGA is a procedure for identifying multiple latent classes of change in a sample (Nagin & Tremblay, 2001). It is conceptually similar to latent growth curve analysis, in which change is modeled by specifying latent intercept and latent slope factors. Latent growth curve models derive a single intercept and slope that represents the average initial value and rate of change, respec- tively, in the sample. The variance of the two latent factors is also estimated, which are informative of the degree of variability around the average intercept and slope. LCGA is a method for building some of that variance back into the model. Rather than a single latent intercept and slope for the entire sample, LCGA allows for the specification of multiple latent intercepts and slopes, each of which describes the pattern of change for a subgroup in the sample. Importantly, within LCGA, the variance of the growth parameters (i.e., intercepts, slopes) within classes is fixed to zero (unlike in latent growth mixture modeling, in which the within- class variance is estimated).

We specified a series of models, ranging from a single class to five classes, to determine the best fitting growth model for the data. Because we had data from four time points, and nonlinear change is theoretically indicated, within each class solution we tested for whether a linear or quadratic growth model was a better

fit to the data. Model selection was guided by the commonly used fit statistics, the balance and size of each class, and past theoretical and empirical work. The three fit statistics that we consulted were the sample size-adjusted Bayesian information criterion (ssBIC), which corrects for model complexity and allows comparisons across nonnested models, with lower values indicating a better fitting model; entropy and posterior class probabilities, which index how well the model-implied classifications correspond to the observed classification, with values closest to 1.0 preferred; and bootstrap likelihood ratio test (BLRT), which tests the added value of a model with k classes over a model with k�1 classes. A significant BLRT indicates that the additional class results in a better fitting model. All models were conducted using full infor- mation maximum likelihood in Mplus 6.1 (Muthén & Muthén, 2010).

We tested and compared eight different growth models: two-, three-, four-, and five-class models, each with either a linear growth term only or a linear and quadratic growth terms. A summary of the fit statistics for the different models that we tested and considered is in Table 2. On the basis of the fit statistics only, the five-class quadratic model was preferred, as it had the lowest ssBIC, highest entropy, and the BLRT indicated added value over the four-class quadratic model. Inspection of the class membership

Table 1 Descriptive Statistics and Correlations for all Study Variables

Variable M SD Observed

range 1 2 3 4 5 6 7 8 9 10 11 12 13

1. Child’s ego development T1 4.37 0.79 2–7 — 2. Child’s ego development T2 4.22 0.82 2–6 .54 — 3. Child’s ego development T3 5.00 0.91 3–7 .43 .53 — 4. Child’s ego development T4 5.49 0.94 3–8 .24 .49 .42 — 5. Mother’s ego development T2 4.17 0.38 4–5 .11 .14 .21 .28 — 6. Father’s ego development T2 4.12 0.32 4–5 .12 .13 .06 .25 .15 — 7. FES family score-Interpersonal

Relations T1 3.42 0.21 2.90–3.87 .21 .25 .19 .03 .16 �.02 — 8. FES family score-Personal

Growth T1 3.18 0.22 2.69–3.75 .31 .22 .07 .13 .07 .11 .44 — 9. FES family score-System

Maintenance T1 2.85 0.32 2.17–3.86 �.18 �.12 �.12 �.18 �.19 �.06 �.20 �.09 — 10. Child’s identity achievement-

Relationships T4 0.28 0.45 0–1 .22 .16 .05 �.02 �.16 �.13 �.06 .05 �.01 — 11. Child’s identity achievement-

Global T4 0.38 0.49 0–1 .25 .12 .09 .06 �.01 �.20 �.01 .12 .08 .80 — 12. Child’s identity exploration T4 2.38 0.97 1.0–4.0 .49 .26 .22 .17 .01 �.02 .14 .29 �.22 .53 .62 — 13. Child’s identity commitment T4 3.10 0.77 1.5–4.0 �.12 .04 �.05 �.07 �.16 �.11 .16 �.02 .30 .42 .40 �.04 —

Note. rs � .20 are significant at p � .05. T1–T4 � Time 1–Time 4; FES � Family Environment Scale.

Table 2 Fit Statistics for all LCGA Models

Variable

Two class Three class Four class Five class

Linear Quadratic Linear Quadratic Linear Quadratic Linear Quadratic

ssBIC 959.87 955.66 947.17 945.46 866.25 861.51 782.18 773.62 Entropy .67 .67 .86 .77 .97 .97 .97 .97 BLRT 57.80 59.77 16.98 15.91 85.20 89.66 86.16 93.60 BLRT p �.001 �.001 �.001 �.001 �.001 <.001 �.001 �.001

Note. Preferred model is highlighted in bold. ssBIC � sample size-adjusted Bayesian information criterion; BLRT � bootstrap likelihood ratio test.

377TRAJECTORIES OF EGO DEVELOPMENT

assignments, however, indicated that two of the five classes were extremely small (n � 2 and n � 3). Thus, we selected as our final model the four-class quadratic model, which had the same high entropy values (.97) as the five-class quadratic model and the lowest ssBIC of the remaining models. The growth parameters for the final model are reported in Table 3, and the trajctories are depicted in Figure 1.

In accordance with the nature of the trajectories, the four groups were labeled normative stable (49%; n � 48), marked by moderate initial growth during adolescence followed by decceleration into adulthood; moderate progression (33%; n � 32), who steadily increased across the 10 years; rapid progression (13%; n � 13), marked by high initial increases during adolescence followed by deccelerated but still increasing growth; stable low (5%; n � 5), who remained at low levels throughout the 10 years. Taken to- gether, the trajectories indicated much greater ego development during adolescence compared with emerging adulthood, wherein the different groups began to plateau (other than the moderate progression group). There were no gender differences in the ego development trajectories, despite previous research indicating such (see Cohn, 1991). Furthermore, there were no difference in clas- sification by SES, nor was SES related to any other study vari- ables. Finally, there were no differences in classification by em- ployment status at Time 4 (i.e., working, attending school, etc.).

Predictions in Adolescence: Family Climate and Parent’s Ego Development

We next tested our two hypotheses about the influence of family factors on youth’s ego development trajectories. Mplus does not contain an analysis of variance procedure, but equivalent models can be conducted within a regression-based path model framework by creating a set of dummy codes for the categories. A single set of dummy codes only allows for pairwise comparison. Because we were interested in comparing all four of the groups for differences on the parenting variables, we conducted three separate path mod- els, rotating the reference category, so that all pairwise compari- sons could be tested. We conducted two sets of multivariate models in this way, one set examining differences among the four classes in the three Time 1 FES family scores and one set exam- ining differences among the four classes in mother’s and father’s ego development. The path models included all three FES family scores (Interpersonal Relations, Personal Growth, and System Maintanence) in a single model so as to control for overlap in the measures. All models included gender and SES as control vari- ables.

For family climate, the rapid progression group had significantly higher Time 1 Personal Growth scores compared with the norma- tive stable group (� � .22, p � .05), and marginally significantly

higher Time 1 Personal Growth scores than the moderate progres- sion group (� � .19, p � .07). There were no other significant group differences for Personal Growth and no differences on Interpersonal Relations or System Maintanence.

We next tested for whether the four groups differed in their mother’s and father’s level of ego development during the partic- ipants’ adolescence. The rapid progression group had both mothers with significantly higher ego development scores (� � .29, p � .01) and fathers with significantly higher scores than the normative stable group (� � .22, p � .05). The rapid progression group also had mothers— but not fathers—with significantly higher ego de- velopment scores than the stable low group (� � .24, p � .05).

Correlates in Emerging Adulthood: Identity Development

Finally, we tested for whether the four groups differed in levels of identity in emerging adulthood. This was done in two ways: by examining difference in classification in the four identity sta- tuses and difference in the continuous identity processes of exploration and commitment. The former was tested using chi-square analyses, and the latter was tested in a path-analytic procedure identical to what was used for the FES analyses above. We hypothesized that progressors would be more likely to be classified as identity achieved and would report higher levels of identity exploration and commitment. Contrary to our hypothesis, the trajectory groups did not vary in their likeihood of being classified as achieved in the relational domain, �2(3) � 1.65, p � .65, or in terms of global identity, �2(3) � 2.19, p � .53.

The four trajectory groups did, however, vary in their levels of identity exploration. The moderate progression group reported significantly higher identity exploration than did the normative stable group (� � .32, p � .01). There were no group differences in identity commitment.

Discussion

The purpose of the present study was to document trajectories of ego development across the transition from adolescence to emerg- ing adulthood and to link these trajectories to the family context in adolescence and identity development in emerging adulthood. In doing so, we aimed to highlight the utility of an individual ap- proach to personality change, as well as the benefits associated with a greater integration between personality and developmental science. In brief, the results indicate (a) heterogenity in ego de- velopment trajectories from adolescence to emerging adulthood, (b) that growth in ego development is more rapid during adoles- cence and tends to taper off in emerging adulthood, (c) family contexts that promote personal growth are linked to greater ego

Table 3 Growth Parameters for the Final Four-Class Quadratic Model

Variable Normative stable Moderate progression Rapid progression Stable low

Intercept 4.02 (.11)��� 4.62 (.14)�� 4.37 (.18)��� 3.56 (.20)���

Linear slope 0.34 (.08)��� 0.23 (.09)�� 0.67 (.14)��� �0.20 (.14) Quadratic slope �0.03 (.01)�� �0.01 (01) �0.04 (.01)�� .02 (.02)

�� p � .01. ��� p � .001.

378 SYED AND SEIFFGE-KRENKE

development, (d) both mother and father ego development is instrumental in childrens’ ego gains in adolescence, and (e) that youth who demonstrate continued ego development into emerging adulthood also demonstrate heightened levels of identity explora- tion. We discuss each of these findings in greater detail below, as well as the implications of this study for the broader field of personality development.

Trajectories of Ego Development

The first goal of the present study was to document the heterogeniety in ego development trajectories across the tran- sition from adolescence to emerging adulthood. We found ev- idence for four unique trajectories that are consistent with past research (e.g., Hauser, 1991): normative stable, rapid progres- sion, moderate progression, and stable low. Congruent with past research (Gfellner, 1986; Hauser, 1991; Westernberg & Gjerde, 1999), the normative stable group made up the largest group. These youth spent most of their adolescence within the con- formist stage and began to plateau as they approached the next stage, self-aware. Thus, for the majority of individuals, the self-aware stage appears to represent a ceiling on ego develop- ment (see also Gfellner, 1986; Westernberg & Gjerde, 1999). In tandem with prior work, our findings suggest that the stabili- zation of ego development has more to do with reaching a particular level—self-aware—rather than a particular age, as was suggested by the Cohn (1998) meta-analysis. That is, age is not likely the reason why most youth do not continue to progress through the ego development sequence. Rather, as many youth experience upheavals in the professional and the relationship domain during the transition to adulthood (Osgood, Foster, Flanagan, & Ruth, 2005), they may not be in the position to face the challenges of moving into the greater

self-reflection, insight, and cognitive complexity associated with the conscientiousness stage.

An important finding of the present study, which is masked by the Cohn (1998) meta-analysis, is that a substantial proportion of young people do not follow the stabilization trend into adulthood. Indeed, the two trajectories of progression represent departures from the tendency to plateau at the self-aware stage. The moderate progressors reached the self-aware stage at about age 16 and continued to make progress toward Stage 6, conscientious, as they transitioned to emerging adulthood. The rapid progressors were quite exceptional, in that they had rapid gains in adolescence, from conformist toward conscientious, and continued to develop into emerging adulthood, ending up near Stage 7, individualistic, at age 24. These two trajectories of progression constituted nearly half the sample, indicating that a large portion of youth continue to develop beyond adolescence and into emerging adulthood. In general, these findings support present theories of development that highlight the continuation of developmental processes beyond the adolescent years (e.g., Arnett, 2000), at least for some propor- tion of youth, and, as is discussed below, under a favorable developmental context. In regard to ego development specifically, as suggested by Cohn (1998), analyses based only on group data obscure the fact that some individuals may indeed continue to develop.

The final trajectory that we observed was a stable low group, which was a small group of youth who started out lower than the others and remained much lower over time. In adolescence, these youth were between Level 3, Self-Protective, and Level 4, Conformist, and they remained near self-protective as they transitioned to emerging adulthood. Although the frequency of this group was quite small in the present sample, it is a theo- retically important group that has been previously identified

1

2

3

4

5

6

7

8

9

Age 14 Age 15 Age 16 Age 24

Eg o

D e

ve lo

p m

e n

t

Age

Stable Low

Rapid Progression

Norma�ve Stable

Moderate Progression

Figure 1. The four ego development trajectory classes.

379TRAJECTORIES OF EGO DEVELOPMENT

and examined in detail by Billings et al., 2008. Future research with larger samples is necessary to better understand the factors that inhibit personality development in this group.

Taken together, the four trajectories identified in the present study indicate a good deal of development in both adolescence and emerging adulthood. More specifically, at age 14, the four trajec- tory groups did not vary substantially in their levels of ego devel- opment (range � 1.07 steps), but began to diverge by age 15, and continued to diverge at age 16. Finally, by age 24, the difference between the highest and lowest levels of ego development was 4.08 steps. Thus, the range in emerging adulthood was 4 times that represented in mid-adolescence (see also Westenberg & Gjerde, 1999). Importantly, the present findings indicate that this devel- opmental divergence, or multifinality, is a gradual process that unfolds over time and not the result of abrupt and sudden change (see also Josselon, 2009; Syed, 2010, for simliar views of the slow progress of identity development). Moreoever, as discussed above, the findings in this study suggests that the prevailing view of ego development occurring primarily in adolescence and stabilizing into adulthood ought to be reconsidered. Although we found this to be the case for approximately half the sample, the majority of the other half continued to progress through the ego development stages. This observation underscores the value of the individual approach to personality development, as the average trajectory can, at times, provide a misleading account about the nature of developmental processes (Donnellan et al., 2007).

The Impact of the Family: What Have Parents Got to Do With It?

Our second goal in the present study was to use family climate and parents’ ego development during the youth’s adolescence to predict membership in the four different ego development classes. We hypothesized that family climate would be related to adoles- cents’ ego development due to past research demonstrating the developmental benefits of an open, supportive family environment that fosters autonomy (Laursen & Hafen, 2010; Steinberg & Mor- ris, 2001). Unexpectedly, family climate was not a strong predictor of class membership. In general, the four trajectory groups all seemed to experience an authoritative and a supportive environ- ment as they traversed adolescence. The one exception to this pattern is that those youth who were in the rapid progression group came from families who promoted greater personal growth com- pared with those in the normative stable group, and marginally greater than those is the moderate progression group. This finding dovetails with recent research (Bauer & McAdams, 2010) linking ego development gains to personal growth in other domains of individuals’ lives. Furthermore, in terms of fostering ego develop- ment, it appears that simply providing a warm and supportive environment is not sufficient. Rather, engaging in family practices that specifically target autonomy and personal growth is what has an impact on the adolescent’s ego development (Allen et al., 1994). Personal growth, in this study, was indexed by family support for such dimensions as independence, achievement, and intellectual curiosity. These dimensions dovetail with the nature of the conscientious and individualistic stages of ego development, which is the terrain that the rapid progressors were navigating. Thus, the findings suggest that those adolescents who are on the upper end of the distribution of ego development come from

families who serve as pacers, fostering associated developmental tasks within their children.

In addition, Loevinger’s concept of pacers (i.e., individuals who facilitate the ego development of others), was applied more spe- cifically in order to analyze the function of mother’s and father’s ego development in this process. Similar to the findings for family climate, mother’s and father’s ego development primarily served to discriminate between the rapid progression and normative stable groups. Importantly, and in line with the findings of Hauser et al. (1991), both mother’s and father’s ego development were unique predictors of the rapid progression group. In contrast, only moth- er’s ego development differed between the rapid progression and stable low groups. Taken together with the findings for family climate, it appears that rapid progression is fostered by a network of family support that promotes personal growth and ego devel- opment. In general, however, it appears that parents do not uni- versally serve as pacers for their children’s ego development. This is understandable, as parents’ levels of ego development, on av- erage, were lower than the child’s level. The average level of parents’ ego development was at the conformist level, which is similar to the majority of the adolescents. But nearly half the adolescents exceeded this level. Given that parents’ ego develop- ment was only related to the rapid progression group, it may be the case that the direction of influence is the other way around: Children may impact their parents’ ego development. Research on the reciprocal influences have substantiated that parental behavior is greatly influenced by their child’s behavior (Kerr & Stattin, 2003) and that parents, for example, alter their coping behavior in response to their children’s’ coping (Hauser et al., 1991; Seiffge- Krenke & Pakalniskiene, 2011). Such a reciprocal view of ego development has rarely been investigated or discussed in the literature (for an exception, see King, Scollon, Ramsey, & Wil- liams, 2000), but could be an important direction for future re- search. Of course, like most dimensions of personality, ego devel- opment has been shown to be strongly heritable (Newman, Tellegen, & Bouchard, 1998), and therefore genetics cannot be ruled out as an explanation for the observed findings.

Finally, it is important to note the differences between the group-level and individual level data. Inspection of Table 1 shows that personal growth is most strongly correlated with Time 1 ego development, suggesting the possibility that the trajectory group differences between the rapid progression and normative stable groups may simply be capturing the correlations between personal growth and ego development at Time 1. However, at Time 1, the moderate progressors—not rapid progressors—actually had the highest levels of ego development. Thus, if the group differences reflected only the Time 1 correlations, we would expect to see the moderate progression and normative stable (or stable low) groups to differ. But this was not the case. This discrepancy highlights the type of insight that can be gained from adopting an individual approach to personality change, as the implications of the groups when considered longitudinally do not mirror the cross-sectional findings at a single time point.

Links Between Ego Development Trajectories and Identity Status

Our third and final research question guided our examination of the associations between the four ego development trajectories and

380 SYED AND SEIFFGE-KRENKE

a key developmental task of emerging adulthood: identity. There is a surprising dearth of research in which explicit links between ego development and identity development are tested, given the strong theoretical links between the two constructs (Erikson, 1968; Lo- evinger, 1976). On the basis of the theoretical underpinning, as well as indirect empirical evidence of a link (e.g., Adams & Fitch, 1982), we hypothesized that greater ego development progression would be associated with greater identity development, both in terms of classification in the achieved identity status and higher levels of identity exploration and commitment. Although there were no differences in identity status classification across the ego development trajectory groups, the moderate progression group reported higher levels of identity exploration than the normative stable group. This suggests that the moderate progressors were particularly embroiled in the process of learning and exploring more about different identity alternatives that they may adopt. Such behaviors are consistent with their steady increases in ego development over time. It is surprising, however, that the rapid progression group did not also evidence greater levels of explora- tion. It may be that this group, who is exceptionally high in ego development at age 24, has slowed their exploratory behaviors and have begun the processes of settling down into a more stable adulthood, which should be marked by a gain in commitment. We did not, however, find group differences in identity commitment, but this may have been due to the fact that most emerging adults reported relatively high levels of commitment. Interestingly, the bivariate correlations indicated that the strongest link was between Time 4 identity exploration and Time 1 ego development—not Time 4 ego development. This findings is due, in part, to the fact that the moderate progression group, which showed the clearest links to identity exploration, had the highest levels of ego devel- opment at Time 1 but the second highest at Time 4, again high- lighting the different information gleaned in the longitudinal ver- sus cross-sectional analyses. However, it also suggests that ego development may precede identity development. Although intrigu- ing, we did not have sufficient data to be able to properly test this causal claim. In sum, although the links we found were modest, this is the first study that we are aware of that explicitly tested for a connection between ego development and identity development.

Implications for the Study of Personality Development

Although the present study was focused on ego development, the conceptualization and analysis have several implications for the broader study of personality development. In contrast to the majority of studies examining personality change over time that analyze mean levels or differential continuity (Roberts & DelVec- chio, 2000; Roberts et al., 2006), in the present study we adopted an individual approach to personality development to understand how some trajectories may deviate from the normative pattern. Although this approach has been used in personality psychology (Donnellan et al., 2007; Lönnqvist et al., 2008; Ozer & Gjerde, 1989; Roberts et al., 2001), the vast majority of studies have consisted of only two waves of data collection, which limits the types of analyses that can be conducted (see Johnson et al., 2007, for a notable exception). Furthermore, our approach brings the mostly autonomous fields of personality and developmental psy- chology into better alignment (at least with respect to adolescence and adults). As described by McAdams and Olson (2010), recog-

nizing the different personality pathways that individuals traverse opens the door for new and interesting questions about social- cognitive and contextual factors that can predict, and be predicted by, membership in the trajectory groups. By adopting a develop- mental approach, we examined developmental contexts that are particularly salient at the beginning (family) and end (identity) of the trajectory period examined. The field of personality develop- ment would greatly benefit from using an individual, developmen- tal approach, particularly as the field moves beyond description of personality change over time and into explanation for why the changes occur and how they matter for people’s lives.

Limitations and Future Research

The present study had several strengths over previous research, such as the inclusion of multiple time points over a 10-year period, examination of change over an important life transition, mapping trajectories of ego development statistically, and linking ego de- velopment to the adolescent family contexts and emerging adult- hood task of identity. In addition, the sample includes “the forgotten half” and multiple informants. These strengths not- withstanding, there are limitations that must be considered when interpreting the findings. First, the sample size was somewhat small, which may have precluded our ability to detect certain ego development trajectories that are less frequently occurring (e.g., regressive). This limitation is mitigated, however, by the four time points and long time span, but it remains a limitation nevertheless. Second, it would have been optimal to have more time points of assessment, particularly between the ages of 16 and 24. There may be some fluctuations in ego development during this time that we were unable to detect. Studies with annual or biennial assessments that cover this same period (and beyond) would be particularly useful. Third, we only had measures of identity development at Wave 4, precluding the ability to understand the causal relations with ego development. Similarly, having information about par- ents’ ego development at one time point only was a limitation. Having parent ego development data that were aligned with child ego development data would have allowed us to ask additional questions, such as how parents’ and children’s ego trajectories codevelop and whether parents have a greater influence on their children or vice versa. Future studies that examine how parents benefit from their children’s ego development would be particu- larly interesting, as such a viewpoint has been largely neglected in the extant literature. Another important direction for future re- search is to examine other factors beyond family that may be related to the different ego development trajectories. For example, exploring the role of friendship and partnership for ego develop- ment trajectories would be valuable. Additionally, probing deeper into individuals’ own agency and motivation may provide insights into how youth create life contexts that will foster their ego development. In general, future research would be well served by the study of how ego development is related to the wide variety of life contexts that youth inhabit.

Conclusions

The present study builds on and extends prior research on the heterogeneity of ego development trajectories from adolescence to emerging adulthood. It is the first study of ego development to

381TRAJECTORIES OF EGO DEVELOPMENT

include more than two waves of data collection and follow the same individuals across an important life transition: the transition to adulthood. The findings indicate that, although ego development stabilizes into adulthood for many young people, a substantial number continues to progress. The present study also fills gaps in the literature by linking parents’ and their children’s levels of ego development and testing for linkages between ego development and identity development—two closely related constructs. Taken together, the integration of personality and development science reflected in the present study suggests promising new avenues for furthering our understanding of the nature, causes, and conse- quences of personality development.

References

Adams, G. A., & Fitch, S. A. (1982). Ego stage and identity status development: A cross-sectional analysis. Journal of Personality and Social Psychology, 43, 574 –583. doi:10.1037/0022-3514.43.3.574

Allen, J. P., Hauser, S. T., Bell, K. L., & O’Connor, T. G. (1994). Longitudinal assessment of autonomy and relatedness in adolescent- family interactions as predictors of adolescent ego development and self-esteem. Child Development, 65, 179 –194. doi:10.2307/1131374

Arnett, J. J. (2000). Emerging adulthood: A theory of development from the late teens through the twenties. American Psychologist, 55, 469 – 480. doi:10.1037/0003-066X.55.5.469

Arnett, J. J. (2004). Emerging adulthood: The winding road from the late teens through the twenties. New York, NY: Oxford University Press.

Bauer, J. J., & McAdams, D. P. (2010). Eudaimonic growth: Narrative growth goals predict increases in ego development and subjective well- being 3 years later. Developmental Psychology, 46, 761–772. doi: 10.1037/a0019654

Billings, R. L., Hauser, S. T., & Allen, J. P. (2008). Continuity and change from adolescence to emerging adulthood: Adolescence-limited vs. life- course persistent profound ego development arrests. Journal of Youth and Adolescence, 37, 1178 –1192. doi:10.1007/s10964-008-9317-4

Blasi, A. (1998). Loevinger’s theory of ego development and its relation- ship to the cognitive developmental approach. In P. M. Westenberg, A. Blasi, & L. D. Cohn (Eds.), Personality development: Theoretical, empirical, and clinical investigations of Loevinger’s conception of ego development (pp. 13–25). Mahwah, NJ: Erlbaum.

Bosma, H. A. (1992). Identity in adolescence: Managing commitments. In G. R. Adams, T. P. Gullotta, & R. Montemayor (Eds.), Adolescent identity formation: Advances in adolescent development (Vol. 4, pp. 91–121). Newbury Park, CA: Sage.

Caspi, A., & Roberts, B. W. (1999). Personality change and continuity across the life course. In L. A. Pervin & O. P. John (Eds.), Handbook of personality theory and research (Vol. 2, pp. 300 –326). New York, NY: Guilford Press.

Caspi, A., Roberts, B. W., & Shiner, R. L. (2005). Personality develop- ment: Stability and change. Annual Review of Psychology, 56, 453– 484. doi:10.1146/annurev.psych.55.090902.141913

Cohn, L. D. (1991). Sex differences in the course of personality develop- ment: A meta-analysis. Psychological Bulletin, 109, 252–266. doi: 10.1037/0033-2909.109.2.252

Cohn, L. D. (1998). Age trends in personality development: A quantitative review. In P. M. Westenberg, A. Blasi, & L. D. Cohn (Eds.), Personality development: Theoretical, empirical, and clinical investigations of Lo- evinger’s conception of ego development (pp. 133–143). Mahwah, NJ: Erlbaum.

Côté, J. E. (2006). Emerging adulthood as an institutionalized moratorium: Risks and benefits to identity formation. In J. J. Arnett & J. L. Tanner (Eds.), Emerging adults in America: Coming of age in the 21st century (pp. 85–116). Washington, DC: American Psycholgical Association. doi:10.1037/11381-004

Donnellan, M. B., Conger, R. D., & Burzette, R. G. (2007). Personality development from late adolescence to young adulthood: Differential stability, normative maturity, and evidence for the maturity-stability hypothesis. Journal of Personality, 75, 237–264. doi:10.1111/j.1467- 6494.2007.00438.x

Erikson, E. H. (1963). Childhood and society. New York, NY: Norton. Erikson, E. H. (1968). Identity, youth, and crisis. New York, NY: Norton. Gfellner, B. M. (1986). Changes in ego and moral development in adoles-

cents: A longitudinal study. Journal of Youth and Adolescence, 15, 147–163. doi:10.1007/BF02141735

Grusec, J. E., & Davidov, M. (2010). Integrating different perspectives on socialization theory and research: A domain-specific approach. Child Development, 81, 687–709. doi:10.1111/j.1467-8624.2010.01426.x

Hauser, S. T. (1991). Adolescents and their families. New York, NY: Free Press.

Hauser, S. T., Borman, E. H., Jacobson, A. M., Powers, S. I., & Noam, G. G. (1991). Understanding family contexts of adolescent coping: A study of parental ego development and adolescent coping strategies. Journal of Early Adolescence, 11, 96 –124. doi:10.1177/ 0272431691111005

Hauser, S. T., & Safyer, A. W. (1994). Ego development and adolescent emotions. Journal of Research on Adolescence, 4, 487–502. doi: 10.1207/s15327795jra0404_3

Hennighausen, K. H., Hauser, S. T., Billings, R. L., Schultz, L. H., & Allen, J. P. (2004). Adolescent ego-development trajectories and young adult relationship outcomes. Journal of Early Adolescence, 24, 29 – 44. doi:10.1177/0272431603260920

Holmbeck, G. N., Li, S. T., Schurman, J. V., Friedman, D., & Coakley, R. M. (2002). Collecting and managing multisource and multimethod data in studies of pediatric populations. Journal of Pediatric Psychology, 27, 5–18. doi:10.1093/jpepsy/27.1.5

Hy, L. X., & Loevinger, J. (1996). Measuring ego development. Mahwah, NJ: Erlbaum.

Johnson, W., Hicks, B. M., McGue, M., & Iacono, W. G. (2007). Most of the girls are alright, but some aren’t: Personality trajectory groups from ages 14 to 24 and some associations with outcomes. Journal of Person- ality and Social Psychology, 93, 266 –284. doi:10.1037/0022-3514.93.2 .266

Josselson, R. (2009). The present of the past: Dialogues with memory over time. Journal of Personality, 77, 647– 668. doi:10.1111/j.1467-6494 .2009.00560.x

Kerr, M., & Stattin, H. (2003). Parenting of adolescents: Action or reac- tion? In A. C. Crouter & A. Booth (Eds.), Children’s influence on family dynamics: The neglected side of family relationships (pp. 121–152). Mahwah, NJ: Erlbaum.

King, L. A., Scollon, C. K., Ramsey, C., & Williams, T. (2000). Stories of life transition: Subjective well-being and ego development in parents of children with Down Syndrome. Journal of Research in Personality, 34, 509 –536. doi:10.1006/jrpe.2000.2285

Kroger, J., Martinussen, M., & Marcia, J. E. (2010). Identity status change during adolescence and young adulthood: A meta-analysis. Journal of Adolescence, 33, 683– 698. doi:10.1016/j.adolescence.2009.11.002

Kunnen, E. S. (2009). Qualitative and quantitative aspects of commitment development in psychology students. Journal of Adolescence, 32, 567– 584. doi:10.1016/j.adolescence.2008.06.002

Lanning, K., Colucci, J., & Edwards, J. A. (2007). Changes in ego devel- opment in the wake of September 11. Journal of Research in Person- ality, 41, 197–202. doi:10.1016/j.jrp.2005.12.002

Laursen, B., & Collins, W. A. (2009). Parent-child relationships during adolescence. In R. M. Lerner & L. Steinberg (Eds.), Handbook of adolescent psychology (3rd ed.). Vol. 2: Contextual influences on ado- lescent development (pp. 3– 42). New York, NY: Wiley.

382 SYED AND SEIFFGE-KRENKE

Laursen, B., & Hafen, C. A. (2010). Future directions in the study of close relationships: Conflict is bad (except when it’s not). Social Develop- ment, 19, 858 – 872. doi:10.1111/j.1467-9507.2009.00546.x

Little, R. J. A. (1988). A test of missing completely at random for multivariate data with missing values. Journal of the American Statis- tical Association, 83, 1198 –1202. doi:10.1080/01621459.1988 .10478722

Loevinger, J. (1976). Ego development: Conceptions and theories. San Francisco, CA: Jossey-Bass.

Loevinger, J. (1985). Revisions of the sentence completion test for ego development. Journal of Personality and Social Psychology, 48, 420 – 427. doi:10.1037/0022-3514.48.2.420

Loevinger, J., Cohn, L. D., & Bonneville, L. P. (1985). Ego development in college. Journal of Personality and Social Psychology, 48, 947–962. doi:10.1037/0022-3514.48.4.947

Lönnqvist, J.-E., Mäkinen, S., Paunonen, S. V., Henriksson, M., & Verkasalo, M. (2008). Psychosocial functioning in young men predicts their personality stability over 15 years. Journal of Research in Person- ality, 42, 599 – 621. doi:10.1016/j.jrp.2007.08.006

Marcia, J. E. (1966). Development and validation of ego-identity status. Journal of Personality and Social Psychology, 3, 551–558. doi:10.1037/ h0023281

Marcia, J. E. (1993). The ego identity status approach to ego identity. In J. E. Marcia, D. R. Mattson, A. S. Waterman, S. A. Archer, & J. L. Orlofsky (Eds.), Ego identity: A handbook for psychosocial research (pp. 3–21). New York, NY: Springer.

McAdams, D. P. (1998). Ego, trait, identity. In P. M. Westenberg, A. Blasi, & L. D. Cohn (Eds.), Personality development: Theoretical, empirical, and clinical investigations of Loevinger’s conception of ego develop- ment (pp. 27–38). Mahwah, NJ: Erlbaum.

McAdams, D. P. (2001). The psychology of life stories. Review of General Psychology, 5, 100 –122. doi:10.1037/1089-2680.5.2.100

McAdams, D. P., & Olson, B. D. (2010). Personality development: Con- tinuity and change over the life course. Annual Review of Psychology, 61, 517–542. doi:10.1146/annurev.psych.093008.100507

McLean, K. C., & Pasupathi, M. (2012). Processes of identity develop- ment: Where I am, and how I got there. Identity: An International Journal of Theory and Research, 12, 8 –28. doi:10.1080/15283488.2011 .632363

Meeus, W. (2011). The study of adolescent identity formation 2000 –2010: A review of longitudinal research. Journal of Research on Adolescence, 21, 75–94. doi:10.1111/j.1532-7795.2010.00716.x

Moos, R. H., & Moos, B. S. (1981). Family Environment Scale manual. Palo Alto, CA: Consulting Psychologists Press.

Muthén, L. K., & Muthén, B. O. (2010). Mplus user’s guide (6th ed.). Los Angeles, CA: Author.

Nagin, D. S., & Tremblay, R. E. (2001). Analyzing developmental trajec- tories of distinct but related behaviors: A group-based method. Psycho- logical Methods, 6, 18 –34. doi:10.1037/1082-989X.6.1.18

Newman, D. L., Tellegen, A., & Bouchard, T. J., Jr. (1998). Individual differences in adult ego development: Sources of influence in twins reared apart. Journal of Personality and Social Psychology, 74, 985– 995. doi:10.1037/0022-3514.74.4.985

Neyer, F., & Asendorf, J. (2001). Personality—Relationship transaction in young adulthood. Journal of Personality and Social Psychology, 81, 1190 –1204. doi:10.1037/0022-3514.81.6.1190

Osgood, D. W., Foster, E. M., Flanagan, C., & Ruth, G. R. (2005). On your own without a net: The transition to adulthood for vulnerable popula- tions. Chicago, IL: University of Chicago Press.

Ozer, D. J., & Gjerde, P. F. (1989). Patterns of personality consistency and change from childhood through adolescence. Journal of Personality, 57, 483–507. doi:10.1111/j.1467-6494.1989.tb00490.x

Roberts, B. W., & Caspi, A. (2003). The cumulative continuity model of personality development: Striking a balance between continuity and

change in personality traits across the life course. R. M. Staudinger & U. Lindenberger (Eds.), Understanding human development: Lifespan psy- chology in exchange with other disciplines (pp. 183–214). Dordrecht, the Netherlands: Kluwer Academic Publishers.

Roberts, B. W., Caspi, A., & Moffitt, T. (2001). The kids are alright: Growth and stability in personality development from adolescence to adulthood. Journal of Personality and Social Psychology, 81, 670 – 683. doi:10.1037/0022-3514.81.4.670

Roberts, B. W., & DelVecchio, W. F. (2000). The rank-order consistency of personality from childhood to old age: A quantitative review of longitudinal studies. Psychological Bulletin, 126, 3–25. doi:10.1037/ 0033-2909.126.1.3

Roberts, B. W., Walton, K., & Viechtbauer, W. (2006). Personality changes in adulthood: Reply to Costa and McCrae (2006). Psychological Bulletin, 132, 29 –32. doi:10.1037/0033-2909.132.1.29

Rosenbaum, J. (2001). Beyond college for all: Career paths for the forgotten half. New York, NY: Russell Sage Foundation.

Schneewind, K. A. (1987). Die Familienklima-Skalen [The Family Climate Scales]. In M. Cierpa (Ed.), Familiendiagnostik (pp. 232–255). Berlin, Germany: Springer.

Schwarz, J. C., Barton-Henry, M. L., & Pruzinsky, T. (1985). Assessing child-rearing behaviors: A comparison of ratings made by mother, father, child, and sibling on the CRPBI. Child Development, 56, 462– 479. doi:10.2307/1129734

Seiffge-Krenke, I. (2003). Testing theories of romantic development from adolescence to young adulthood: Evidence of a developmental sequence. International Journal of Behavioral Development, 27, 519 –531. doi: 10.1080/01650250344000145

Seiffge-Krenke, I., & Pakalniskiene, V. (2011). Who shapes whom in the family: Reciprocal links between autonomy support in the family and parents’ and adolescents’ coping behaviors. Journal of Youth and Ado- lescence, 40, 983–995. doi:10.1007/s10964-010-9603-9

Shulman, S., & Seiffge-Krenke, I. (1997). Fathers and adolescents: De- velopmental and clinical perspectives. London, England: Routledge.

Smith, M. B. (2005). “Personality and social psychology”: Retrospections and aspirations. Personality and Social Psychology Review, 9, 334 –340. doi:10.1207/s15327957pspr0904_3

Steinberg, L. (2001). We know some things: Parent-adolescent relation- ships in retrospect and prospect. Journal of Research on Adolescence, 11, 1–19. doi:10.1111/1532-7795.00001

Steinberg, L., & Morris, A. S. (2001). Adolescent development. Annual Review of Psychology, 52, 83–110. doi:10.1146/annurev.psych.52.1.83

Syed, M. (2010). Developing an integrated self: Academic and ethnic identities among ethnically diverse college students. Developmental Psychology, 46, 1590 –1604. doi:10.1037/a0020738

Syed, M. (2012). The past, present, and future of Eriksonian identity research: Introduction to the special issue. Identity: An International Journal of Theory and Research, 12, 1–7. doi:10.1080/15283488.2012 .632362

Thorne, A. (1993). On contextualizing Loevinger’s stages of ego develop- ment. Psychological Inquiry, 4, 53–55. doi:10.1207/s15327965pli0401_11

Thorne, A. (2004). Putting the person into social identity. Human Devel- opment, 47, 361–365. doi:10.1159/000081038

van Hoof, A. (1999). The identity status field re-reviewed: An update of unresolved and neglected issues with a view on some alternative ap- proaches. Developmental Review, 19, 497–556. doi:10.1006/drev.1999 .0484

Waterman, A. S. (1993). Developmental perspectives on identity forma- tion: From adolescence to adulthood. In J. E. Marcia, A. S. Waterman, D. R. Matteson, S. L. Archer, & J. L. Orlofsky (Eds.), Ego identity: A handbook for psychosocial research (pp. 42– 68). New York, NY: Springer-Verlag.

383TRAJECTORIES OF EGO DEVELOPMENT

Waterman, A. S. (1999). Identity, the identity statuses, and identity status development: A contemporary statement. Developmental Review, 19, 591– 621. doi:10.1006/drev.1999.0493

Westenberg, P. M., Blasi, A., & Cohn, L. D. (1998). Introduction: Con- tributions and controversies. In P. M. Westenberg, A. Blasi, & L. D. Cohn (Eds.), Personality development: Theoretical, empirical, and clin- ical investigations of Loevinger’s conception of ego development (pp. 1–9). Mahwah, NJ: Erlbaum.

Westenberg, P. M., & Gjerde, P. F. (1999). Ego development during the transition from adolescence to young adulthood: A nine-year longitudi-

nal study. Journal of Research in Personality, 33, 233–252. doi:10.1006/ jrpe.1999.2248

Wohlwill, J. F. (1973). Beyond age-group comparison: A programmatic view of the task of developmental psychology. The study of behavioral development (pp. 36 – 46). New York, NY: Academic Press.

Received December 12, 2011 Revision received July 23, 2012

Accepted August 20, 2012 �

384 SYED AND SEIFFGE-KRENKE