JOURNAL OF RESEARCH ON ADOLESCENCE, 26(1), 60–75

The Role of Domain Content in Understanding Identity Development

Processes
Kate C. McLean Moin Syed
Western Washington University University of Minnesota

Alisha Yoder Andrea F. Greenhoot

Western Washington University University of Kansas

The current studies examined the importance of domain content in the processes of identity development using two
approaches—narrative and status. We examined personal narratives for identity domain content, the co-occurrence of
different contents, and the relations between content and processes, using two approaches to identity—status and nar-
rative. Across two studies, 762 participants (average age = 19 years) and 2214 narratives, traditional ideological and
interpersonal status domains were present, but so was a novel domain: existential concerns. Narrative identity pro-
cesses were more frequent in narratives with multiple contents, and relations between identity statuses processes and
narrative processes were modest. We discuss theoretical implications, the importance of examining content, and the
utility of narrative approaches for doing so.

Identity development is the major psychosocial stand the link between identity process and con-
task of adolescence and emerging adulthood (Ar- tent.
nett, 2000; Erikson, 1968). It is through the pro-
cesses of exploring one’s roles and beliefs across
Erikson’s Theory of Identity Development,
domains and time that individuals come to feel a
Contemporary Approaches, and Critiques
sense of coherent integration, and to understand
their place in society. However, we see two current In his seminal life span theory of development,
problems in the field that limit our understanding Erikson (1950) proposed that individuals negotiate
of this developmental task. First, the bulk of the qualitatively distinct tasks at each life stage. The
extant research has focused on processes of identity stage on which he spent a great deal of intellectual
development to the neglect of the content of iden- effort was identity versus role confusion. He
tity. Second, there are two prominent approaches argued that adolescence is the first time that indi-
to identity development—status (e.g., Marcia, 1966) viduals begin to consider questions of identity, as
and narrative (e.g., McAdams, 1993)—that exist lar- part of the process involves the realization that
gely in parallel, with little knowledge about how childhood identifications are no longer useful (Erik-
they differ or complement each other (McLean & son, 1968). In emerging adulthood, individuals gain
Pasupathi, 2012; Syed, 2012). In the current study, new cognitive skills that make complex self-explo-
we take on these two problems by reporting find- ration increasingly possible (Labouvie-Vief, 2006;
ings from two studies in which we adopted both Piaget, 1965). Further, cultural norms about per-
status and narrative approaches to better under- sonal exploration in contemporary American soci-
ety make emerging adulthood an optimal time for
identity work. Given a tolerance for, or even expec-
We wish to thank the McLean Laboratory at Western tation of, delayed adulthood in modern society,
Washington University and the Greenhoot Laboratory at the Erikson (1959/1980) viewed this as a time to
University of Kansas for data collection and coding of Study 1,
engage in role experimentation, and scholars have
Natasha Maloney and Stephen Esposito for coding in Study 2,
Becky Wood for the management of Study 1, and Western pointed to emerging adulthood as the primary
Washington University for funding of Study 1. stage for identity exploration, even more so than
Requests for reprints should be sent to Kate C. McLean at
Western Washington University, Department of Psychology, MS
9172, 516 High Street, Bellingham, WA 98225. E-mail: Kate.
[email protected]; or Moin Syed, University of Minnesota, © 2014 The Authors
N218 Elliott Hall, 75 E River Road, Minneapolis, MN 55455. Journal of Research on Adolescence © 2014 Society for Research on Adolescence
E-mail: [email protected]. DOI: 10.1111/jora.12169
 IDENTITY DEVELOPMENT 61

adolescence (Arnett, 2000; Kroger, Martinussen, & descriptive, with little emphasis on how exploration
Marcia, 2010). and commitment might differ by domains (cf.,
Erikson proposed that to prepare individuals for Fadjukoff, Pulkkinen, & Kokko, 2005; Fris
en &
the tasks of adulthood, a full and healthy resolu- W€angqvist, 2011; Skorikov & Vondracek, 1998).
tion of the identity crisis involves engagement at Second, researchers have consistently defined
three levels, two of which we address—ego and domains for participants a priori (Balistreri, Busch-
personal identity, which are the foci of contempo- Rossnagel, & Geisinger, 1995; Bennion & Adams,
rary research on identity development. We propose 1986; Crocetti, Rubini, & Meeus, 2008; Marcia,
that these levels have been differentially captured 1966), neglecting the personal relevance of domains
by narrative and status approaches to identity (cf. Fris
en & W€angqvist, 2011). For example, if
development, which is one reason that prior someone scores low on exploration of religious
research has found few or small relations between beliefs, it is difficult to know whether that is an
them (e.g., McLean & Pratt, 2006; Syed & Azmitia, indicator of low relevance of, or low engagement
2008). with, religion. Third, little research has examined
how identity exploration might occur as content
domains intersect (see also Grotevant, 1987; van
Status Approaches
Hoof, 1999), such as how the exploration of reli-
Personal identity centers on how one defines one’s gious beliefs may also be related to the exploration
beliefs or goals via negotiation with culturally rele- in romantic relationships. Finally, although Erikson
vant roles. This process hinges on the availability emphasized psychobiography and qualitative
of these roles and belief systems, the meaning of approaches (e.g., 1950), current-status approaches
them to the individual, and how they intersect are dominated by survey measures to assess indi-
with each other (Erikson, 1968). For example, one’s cators of exploration and commitment within
beliefs about religion, or one’s role as male or domains (Balistreri et al., 1995; Bennion & Adams,
female, will be influenced by one’s own experi- 1986; Crocetti et al., 2008; cf. Kroger, 2002, 2007), or
ences and perceptions, as well as the possible reli- as general tendencies not tied to specific domains
gious belief systems or gender norms available to (Luyckx, Goossens, & Soenens, 2006). Some have
the individual. We propose that personal identity argued that this approach voids the context and
is best captured by the identity status approach the person from the rich study of identity (van
that focuses on the processes of exploration and Hoof, 1999; Josselson & Flum, forthcoming;
commitment in various interpersonal (e.g., dating) Schwartz, 2005; Syed & Azmitia, 2008).
and ideological (e.g., religion) identity content
domains (e.g., Marcia, 1966, 1993; see also
Narrative Approach
Schwartz, 2001 for a review). Exploration entails
information-seeking behaviors about major life In contrast to the exploration of personal belief sys-
decisions, such as exploration of different religions tems, Erikson argued that ego identity focuses on
by attending various services. Commitment is personal continuity and is found when one is able
defined as adhering to a set of values or beliefs, to integrate one’s most important, basic, and inter-
such as engaging in daily prayer. Based on the nal understanding of the self to create a sense of
degree of exploration and commitment, Marcia personal sameness across time. We propose that
(1966) arrived at four different statuses as follows: this level is captured by theories that center on
achievement (exploration and commitment), mora- subjective processes of constructing a personal life
torium (current exploration without commitment), story as the critical pathway to identity develop-
foreclosure, (commitment without exploration), and ment (e.g., McAdams, 1993). Narrative approaches
diffusion (neither exploration nor commitment). emphasize the management of inevitable personal
Achievement represents the successful negotiation changes that can be a threat to personal continuity
of this psychosocial stage, which comes after the (Pasupathi, Brubaker, & Mansour, 2007): If my
move from foreclosure/diffusion to moratorium beliefs or body or personality changes, am I still
(Kroger et al., 2010). the same person? Forming a coherent story that
We raise four critiques of the status approach explains how one has changed and remained the
that we sought to address in the present studies. same can preserve a sense of personal continuity
First, while the status approach does define the through time.
content of identity work (e.g., religion, dating), As with the status approaches, the primary
research on domain content has been relatively flaws in the narrative approach are the lack of
62 MCLEAN, SYED, YODER, AND GREENHOOT

attention to identity contents and their intersec- personal values, and relationships (Arnett, 2000;
tions, and how content relates to process. While it Erikson, 1968). So, regardless of approach, we
is surprising that little attention has been paid to would expect interpersonal and ideological con-
the ways that individuals comingle identity con- cerns to be important (Bennion & Adams, 1986;
tents from a perspective that centers on subjectiv- McLean & Pratt, 2006; Thorne, McLean, & Law-
ity, the subjectivity of the narrative approach offers rence, 2004). However, within the narrative
the best mechanism for understanding the issues of approach, there is another common issue that indi-
content salience and intersectionality. viduals find self-defining, which is a focus on con-
In sum, we propose that these two approaches cerns of mortality (McLean & Pratt, 2006; Thorne
take on different levels of Erikson’s conceptualiza- et al., 2004). Issues of mortality are existential ques-
tion of identity, and they both favor an examina- tions that are at the heart of threat to personal con-
tion of the processes in which individuals are tinuity through time. Thus, we coded personal
engaging to construct an identity over an exami- narratives for the frequencies of status domain con-
nation of the contents in which these processes tents to see whether they were indeed personally
occur. However, consistent with Erikson (1968), salient to our informants, as well as for existential
we contend that it is imprudent to examine processes concerns to understand the relative importance of
of identity development in contents that may be per- various domains.
sonally irrelevant to the individuals one is studying. Once we established the relevant content, we then
Thus, in this study, we examined (1) the frequen- asked questions about how content domains might
cies of identity status domain contents as they nat- be differentially linked to identity processes, and if
urally occurred in individuals’ personally identity processes were heightened when content
important narratives, (2) whether these domains domains co-occur. Following Erikson’s (1959/1980,
emerged together, or whether they co-occurred, in 1968) proposal that the intersection of these contents
individuals’ narratives, and (3) how this co-occur- may create conflict, we examined a narrative process
rence was related to the processes of identity that can aid in resolution of conflict: meaning-making.
development. Meaning-making is defined as reflecting on past
events to see how individuals have changed over
time (McLean & Thorne, 2003), which is particularly
THE PRESENT STUDIES: IDENTITY CONTENT,
likely when events are conflict-laden (Bruner, 1990;
CO-OCCURRENCE, AND THE OVERLAP IN
McLean & Thorne, 2003). McLean, Pasupathi, and
APPROACHES
Pals (2007) argued that meaning-making is a critical
The primary goal of these studies was to examine narrative process for developing a coherent identity
the links between approaches by addressing the by helping individuals to make sense of challenging
lack of attention to identity content (cf., Fris
en & events, which lead us to expect that meaning-
W€ angqvist, 2011; Grotevant, 1987; Syed & Azmitia, making might be engaged to make sense of events
2010). Erikson defined content as relationships, where contents co-occur.
occupation, and ideology, the latter of which Mar-
cia defined as politics and religion. Other domains
METHOD: STUDY 1
have been added, with a current consensus on
ideological (occupation, values, religion, politics) The data were collected as part of a study on the
and interpersonal (family, friends, dating, sex roles, relation between the narration of different types of
recreation) domains (Bennion & Adams, 1986; autobiographical events, abuse history, and well-
Grotevant, Thorbecke, & Meyer, 1982). Yet, Erikson being (Greenhoot & McLean, 2013).
emphasized the importance of relevant identity
domains; thus, analysis of content should begin
Participants
with contents that are salient to one’s informants.
The a priori definition of content neglects the local, We recruited 577 participants (n = 1659 narratives)
subjective, and contextualized nature of contents through psychology subject pools at two large pub-
(Syed & Azmitia, 2008). By taking a narrative lic universities in the Midwest (n = 136) and in the
approach to domain content, we return the issue of Northwest (n = 441) of the United States (mean
salience to our informants. age = 19.5 years, SD = 2.51; 68% females). The
In terms of what might be salient to our U.S. sample at the midwestern campus was prescreened
samples, emerging adulthood is viewed as a stage for history of abuse; 211 participants, across both
of exploration of vocational and career options, samples, reported past exposure to abuse, domestic
 IDENTITY DEVELOPMENT 63

violence, or sexual trauma. Participants self- Participants were given course credit for participa-
reported ethnicity was Caucasian (77%), followed tion, which took an average of 1.5–2 hr.
by African American (4%), Asian American (6%),
Latino/a (3%), American Indian (0.3%), Mixed
Narrative Coding
(5%), Other (3%).
Domain content. The coding system captures
identity content domains that emerge in autobio-
Measures
graphical narratives, and can be obtained from the
Narrative prompts. Individuals were randomly first or second author. Following past research, we
assigned to write three personal narratives in adopted a two-level system for organizing domains:
response to one of five types of prompts: trauma broad (e.g., ideological and interpersonal) and spe-
(n = 117), transgression (n = 119), low point (n = cific (e.g., family, values). To avoid confusion, we
113), self-defining (n = 116), and turning point refer to the broad domains as domains and the spe-
(n = 114). These prompts are commonly used in cific domains as facets. We began with the content
studies of narrative identity. Traumas were defined domains and facets that status researchers have
as the most negative, stressful, or traumatic events identified: occupation, religion, politics, and per-
of one’s life (Greenhoot, Sun, Bunnell, & Lindboe, sonal values in the ideological domain, and sex
2013). Transgressions were defined as the worst roles, family, friendship, dating, and recreation in
thing one had ever done, which may have resulted the interpersonal domain. We use the term sex roles
in physical or psychological harm to another, and to be consistent with prior literature, but this term
guilt or shame (Mansfield, McLean, & Lilgendahl, really captures issues of gender, and sexuality and
2010). Low points were described as extremely sexual behaviors would be primarily captured
negative events, which could include emotions under dating. Notably, we found no instances of
such as despair, disillusionment, terror, guilt, or politics or sex roles. Narratives that did not fit with
shame (McAdams, 2006). Self-defining memories these facets were coded as mortality, drugs/alcohol,
are highly emotional, represent an enduring theme mental health, and abuse. We labeled this collection
in one’s life, and help to explain who one is (Singer of facets the existential domain. The rationale for this
& Moffitt, 1991–1992). Turning points are episodes domain is that abuse and mental health share a
in which one underwent an important change in similarity to mortality in that they are about per-
self-understanding (McAdams, 2006). For each sonal threat and continuity through time, and the
prompt, participants were asked to provide details narratives about drugs and alcohol tended to be
about where they were, whom they were with, about fear-arousing experiences that included threat
what happened, and their reaction and others’ (if to physical well-being or, more rarely, about exis-
relevant). Narratives were 187 words on average tential experiences (e.g., hallucinogens). Coders
(range = 3–905; SD = 111). After writing narratives, rated each facet as present or absent for each narra-
participants also completed several assessments tive. To be coded as present, the facet had to be
that we do not examine here: ratings of memory related to some aspect of plot, as opposed to back-
characteristics (Rubin, Schrauf, & Greenberg, 2003), ground information. For example, if the event hap-
survey items about memory telling, a memory pened at church, religion would not be coded as
about abuse, a memory of overcoming a struggle, present unless religion had a part to play in the plot
and surveys assessing well-being, depression, or emotion of the story. One way to determine pres-
health, abuse history, and post-traumatic stress ence of content was to ask whether or not exchang-
disorder. Finally, participants completed a demo- ing potential content would change the story—for
graphics form. example, could church have been the workplace? If
details could be exchanged, then they were not
counted as identity facet content. If more than one
Procedure
facet emerged, coders decided on the dominant
After providing informed consent, participants facet, by asking what the main identity issue at play
wrote their narratives on paper and then com- was. The first author conducted reliability analyses
pleted survey items on a computer program used with the third author, who completed all facet
for the remainder of the session (written with content coding for both studies. Reliability was
Media Lab v2008; Empirisoft, 2008). Once finished, acceptable for each facet and for dominant facet:
participants were debriefed, given information values kappa = .86; occupation kappa = .88; reli-
about available counseling resources, and thanked. gion kappa = 1.0; family kappa = .81; friends
64 MCLEAN, SYED, YODER, AND GREENHOOT

kappa = .74; dating kappa = 1.00; other = .88; provided up to three narratives, and thus, narratives
dominant facet kappa = .90. are nested within persons. We provide more detail
about these models below.
Meaning-making. Each of the memory narra-
tives was coded for sophistication of meaning on a
What Identity Content Domains and Facets Are
4-point scale (McLean & Pratt, 2006). A zero indi-
Represented in Personal Narratives?
cated no explanation of the meaning of the event.
Narratives were scored as one if there was a spe- The domains and facets of the memories were con-
cific lesson that the reporter learned from the sidered in terms of their presence, and if they were
event. A score of two was assigned to narratives the dominant theme if more than one domain/facet
that contained vague meaning; narratives of this was coded as present (Table 1). Interpersonal
sort describe some growth or change in the self, domains were the most frequently occurring, fol-
but the specifics of the change are not clear. Narra- lowed by existential domains and ideological
tives were scored as three if there was evidence domains. The rank order for dominant theme was
that the reporter gleaned specific insight from the the same: interpersonal, existential, and ideological.
event that applies to broader areas of the reporter’s Although the ideological domains were the least fre-
life. Three research assistants who had acceptable quently occurring, when they were present with
reliability with a trained graduate student coded other contents they were most likely to be dominant.
meaning (intraclass range = .83–.89). Table 1 also illustrates the variability in facets
Reliability was completed on 10% of narratives within the three broad domains. The frequencies
(54 cases, 164 narratives). Coders were blind to are notable for both what was, and was not,
condition, abuse status, and hypotheses. Once reli- included. Within the interpersonal domain, the
ability was reached, coders discussed difficult cases family facet accounted for the majority of narra-
as needed and checked 20% of their codes with tives, followed by the friendship and dating facet.
each other to prevent coder drift. Recreation was infrequently occurring, and sex
roles were not evident. For the ideological domain,
the occupational facet was the most frequent, fol-
RESULTS: STUDY 1 lowed closely by values. Religion was quite rare,
and politics was not evident. Finally, within the
Analysis Plan
existential domain, mortality narratives were the
The analyses were carried out in two steps. First, we most frequently occurring, followed by drugs,
examined the frequencies with which the identity abuse, and mental health narratives.
contents were present in the narratives, how the con-
tent co-occurred within narratives, and how con-
How Does Content Facet Co-Occur?
tents varied by narrative prompt, gender, age, and
abuse status. These analyses were conducted using Despite the large number of possible facets (11),
chi-square statistics, relying on adjusted standard- there was relatively little co-occurrence of facets
ized residuals (ASRs) and Cramer’s v. ASRs are a within the narratives. The majority of narratives
metric for evaluating the degree of discrepancy (66%; n = 914) featured only one facet, with 31%
between observed and expected values for an indi- (n = 427) including two facets, 3% (n = 44) includ-
vidual cell within a contingency table. Whereas a ing three facets, and only one narrative including
statistically significant chi-square statistic indicates four facets (none had more than four). For the pur-
global nonindependence in the contingency table, poses of subsequent analyses, we combined the
the ASR allows researchers to locate the source of narratives containing either three or four facets into
the nonindependence within individual cells. The a single category.
ASR is a standardized metric, and therefore, can be The number of facets present varied significantly
interpreted like a z-score. Cramer’s v is an effect size by dominant domain, v2(4, N = 1386) = 37.38,
index that represents the strength of the association p < .001, v = .12. Each domain had a different dis-
between two variables. In the second step, we used tribution pattern. Interpersonal domains were least
multilevel modeling (MLM) to examine how con- likely to include three or more facets (ASR = 2.8),
tents were related to meaning. MLM is a statistical but did not differ from expected on one or two fac-
technique that generalizes linear regression to situa- ets. Ideological domains were most likely to be one
tions in which data are nested, or hierarchically facet (ASR = 3.5) and less likely to be two facets
organized. We used MLM because each participant (ASR = 3.9), with no difference on three. Finally,
 IDENTITY DEVELOPMENT 65

TABLE 1
Distribution of Content Domains

Study 1 Study 2

Presenta Dominant Presenta Dominant

Domain n Percent (%) n Percent (%) n Percent (%) n Percent (%)

Interpersonal 999 74 743 53 329 76 262 59

Family 577 42 420 31 155 36 119 27
Friends 227 17 164 12 89 20 72 17
Dating 195 14 145 11 66 15 55 13
Recreation 14 1 14 1 19 4 16 4
Sex roles 0 0 0 0 0 0 0 0
Ideological 334 24 272 20 83 19 67 15
Values 145 11 108 8 21 5 18 4
Religion 25 2 18 1 3 1 2 0.5
Occupation 164 12 146 11 59 14 47 11
Politics 0 0 0 0 0 0 0 0
Existential 556 41 371 27 147 34 114 26
Mortality 307 22 206 15 103 24 78 18
Mental health 70 5 43 3 18 4 14 3
Abuse 82 6 63 5 4 1 4 1
Drugs 97 7 59 4 22 5 10 2
Total 1903 1386 559 435

a
Percentages do not add to 100 because multiple domains could be present within each narrative.

existential domains were less likely to be one facet less often mortality stories (ASR = 1.9), with no
( 5.2) and more likely to be two (ASR = 4.4) or differences on mental health or drug stories. Addi-
three (ASR = 2.4). tionally, abuse status was significantly related to
memory prompt, v2(4, N = 1386) = 16.88, p = .002,
v = .11. Participants with a history of abuse more
How Do Content Domains Vary by Memory
often had low-point prompt (ASR = 2.0) or turn-
Prompt?
ing-point prompt (ASR = 1.9), less often transgres-
Dominant domains varied by memory prompt, sion ( ASR = 2.9) and negative (ASR = 1.9)
v2(8, N = 1386) = 61.17, p < .001, v = .15. Interper- prompt (this occurred despite random assignment
sonal domains were less likely to be dominant with to conditions).
turning-point memories (ASR = 2.0), but were
evenly distributed across the four other memory
How Are Content Domains Related to Process?
types. In contrast, ideological domains were more
likely with turning-point memories (ASR = 3.8) We examined whether the three broad content
and less likely with both low-point (ASR = 2.9) domains were associated with variations in narra-
and traumatic (ASR = 4.6) memories. Existential tive meaning-making. Because participants each
domains were more likely with traumatic memo- reported up to three narratives, the subsequent
ries (ASR = 5.2) and less likely with transgression analyses were conducted using multilevel model-
memories (ASR = 3.3). ing (MLM) to account for the nesting of multiple
stories within participants. There were two focal
predictors as follows: identity domain, which was
Do Content Domains Vary by Gender, Age, or
dummy-coded with the interpersonal domain spec-
Abuse Status?
ified as the reference category, and co-occurrence,
There were no gender or age differences in the dis- which was treated as a continuous variable. Partici-
tribution of the broad domains, facets memory pant age, gender, and abuse status were included
prompt, or co-occurrence. Abuse status was not as Level 2 covariates, and word length was
related to broad domain or co-occurrence, but was included as a Level 1 covariate. All models were
related to facets. Participants with a history of specified using full-information maximum likeli-
abuse more often had abuse stories (ASR = 3.7), hood. In reporting the results, we included Cohen’s
66 MCLEAN, SYED, YODER, AND GREENHOOT

d for the pairwise comparisons of the dummy- For a direct test of the relation between different
coded content predictors and unstandardized approaches to identity, we added a measure of
regression coefficient (b) for the continuous predic- identity statuses and ego development. We should
tors. not be entirely surprised to see a lack of substantial
overlap in status and narrative processes, given
Meaning. Both identity domain and co-occur- our proposal of the different focus of personal and
rence were significantly associated with meaning ego identity. Indeed, findings on the relations
(Table 2). Narratives within the ideological domain between narrative and status identity processes
included significantly more meaning than both the have been modest, with the most basic finding
interpersonal (d = .14, p < .05) and existential being that markers of autobiographical reasoning,
domains (d = .25, p < .001), and the interpersonal including meaning-making, are associated with
domain had marginally greater meaning than the more mature identity status development, but with
existential domain (d = .09, p = .09). Co-occurrence small effect sizes (Alisat & Pratt, 2012; Josselson,
was positively associated with meaning (b = .15, 1982; McLean & Pratt, 2006; Neimeyer & Rares-
p < .05). There were no significant interactions hide, 1991; Orlofsky & Frank, 1986; Pasupathi,
between identity domains and any other variables Wainryb, & Twali, 2012; Syed & Azmitia, 2008,
in the model. 2010). Thus, we expected identity achievement to
In sum, across various types of memories, the be modestly associated with meaning because each
domains and facets that are viewed as central to of these constructs is viewed as a marker of pro-
identity development from a status perspective do cesses of identity development. Novel to the cur-
occur in personally important narratives, with rent study, we also expected achievement to be
some notable exceptions (i.e., sex roles, politics). related to co-occurrence of content domains. We
We also saw that there is another domain that has conceptualized co-occurrence as representing Erik-
not been captured in status approaches—existential son’s (1950, 1968) concept of identity synthesis—
concerns. The ideological domain was especially the beginning of the process of piecing one’s iden-
likely to have meaning, followed by interpersonal, tity together, which should be related to achieving
and then existential. We also saw that co-occur- a coherent sense of one’s self.
rence, while relatively rare, was related to more Given its foundation in psychodynamic theories
meaning in the narratives. and the emphasis on the ego as the synthesizer of
experiences, or identity contents (McAdams, 1998),
we examined ego development in relation to pro-
STUDY 2
cesses of identity development. For example, Loevin-
We aimed to replicate our initial findings, and to ger (1976) articulated the ego as a process of
add several other factors, while simplifying our “selfing,” existing to make meaning out of one’s own
focus by concentrating on only one type of prompt experiences (McAdams, 1998). Because ego level is
to simplify our analyses, self-defining memories. presumed to be a marker of increasing identity syn-
thesis, we expected ego development to be associated
TABLE 2 with meaning-making and identity achievement.
Multilevel Models Predicting Meaning Indeed, those higher on ego development construct
more integrative personal narratives (Bauer &
Study 1 Study 2
McAdams, 2004; McLean & Fournier, 2008).
b SE b SE
STUDY 2 METHOD
Intercept .56*** .11 .82*** .10
Covariates Data from this study were also reported in McLean
Age .04 .08 .08 .06
(2005), which focused on an entirely different topic
Abused .16 .12 — —
Male .13 .13 .06 .13 of the functions of memory telling.
Word count .26*** .05 .11 .06
Predictors
Ideological versus interpersonal .26* .11 .03 .16 Participants
Existential versus interpersonal .17 .10 .28* .14
The sample consisted of 185 participants (n = 555
Co-occurrence .15* .08 .12 .11
narratives), collected from a participant pool at a
Note. Coefficients for interaction terms not shown. public university in Northern California. Age ran-
*p < .05; ***p < .001. ged from 16 to 27 years (mean age = 18.75;
 IDENTITY DEVELOPMENT 67

SD = 1.19; 47% male). Sixty-two percent of the par- responses are coded and computed to arrive at an
ticipants described themselves as Caucasian, 17% ego level, with higher levels representing more
Asian, 6% Latino, 1% African American, 14% complex thinking about self and others. The possi-
mixed race, and 4% of the participants were cate- ble range is 2–9, and the range in this study was 3–
gorized as other. Two percent of the participants 7, with an average of Level 5, consistent with
did not report ethnicity. research on this age group (Loevinger et al., 1985).
Ego level was rated based on the automatic ogive
rules described by Hy and Loevinger (1996). The
Procedure
first author and an undergraduate research assis-
After providing informed consent, participants tant, who were blind to the study hypotheses, con-
completed a paper-based questionnaire in a room ducted reliability on 24 cases, after training with
alone. The questionnaire included narrative the ego development manual. Reliabilities were
prompts, and several surveys, of which relevant acceptable using a linear variable (interclass
ones are described below. Participants were given r = .92), or examining each level as a category
1 hr to complete the study. (overall kappa = .86). The research assistant coded
the rest of the cases.
Tasks and Measures
Coding Self-Defining Memory Narratives
Self-defining memory questionnaire. The first
page of the questionnaire (adapted from Singer & Content domain and facet coding were identical to
Moffitt, 1991–1992) elicited demographics (gender, Study 1 and were performed by the same coder.
age, ethnicity) and described features of a self- Coding for meaning-making was identical to Study
defining memory, defined nearly identically to 1. A research assistant, who was blind to the
Study 1. The survey also elicited information on hypotheses, rated 30% of the narratives for reliabil-
whether and how the event had been told to oth- ity (intraclass r = .87). The first author coded all
ers, which was not analyzed here (see McLean, narratives for meaning-making.
2005). Narratives were 104 words on average
(range = 4–321; SD = 48).
STUDY 2 RESULTS
Eo-MEIS-2. Participants completed the 64-item
What Identity Content Domains Are Represented
Extended Objective Measure of Ego Identity Status
in Personal Narratives?
(Bennion & Adams, 1986). Each item is rated on a 6-
point scale (strongly agree to strongly disagree) and Domains were considered in terms of their pres-
focuses on exploration and commitment in various ence in any of the narratives as well as whether
facets (occupation, religion, politics, sex roles, fam- they were dominant if more than one domain was
ily, friends, dating, recreation). Continuous scores coded. The distribution of frequencies was similar
are calculated for each of the four identity statuses to Study 1 (Table 1). Interpersonal domains were
(achieved, moratorium, foreclosed, diffused). Alphas the most frequent, followed by existential domains
for the scale items are as follows, with comparisons and ideological domains. The rank order for domi-
to Bennion and Adams (1986) in parentheses: ideo- nant theme was the same: interpersonal, existential,
logical diffusion = .55 (.62); ideological foreclo- and ideological. The proportion of dominance rela-
sure = .79 (.75); ideological moratorium = .66 (.75); tive to presence was similar across the three
ideological achievement = .63 (.62); interpersonal domains: Ideological was dominant in 81% of the
diffusion = .63 (.64); interpersonal foreclosure = .82 narratives in which it was present, compared with
(.80); interpersonal moratorium = .52 (.58); interper- 80% for interpersonal and 78% for existential
sonal achievement = .59 (.60). domains.

Ego development. Participants completed the

How Do Content Domains and Facets Co-Occur?
36-item Washington University Sentence Comple-
tion Test (Hy & Loevinger, 1996). Participants are Despite the large number of possible facets (11),
asked to complete sentence stems (e.g., “What gets there was relatively little co-occurrence of facets
me into trouble is. . .”). This well-validated assess- within the narratives. Similar to Study 1, the major-
ment is a projective test designed to assess ego ity of narratives (73%; n = 325) featured only one
level (Hy & Loevinger, 1996). Participants’ facet, with 23% (n = 100) including two facets, 2%
68 MCLEAN, SYED, YODER, AND GREENHOOT

(n = 9) including three facets, and only two narra- tionally, there was a significant interaction between
tives including four facets (none had more than ego development and identity domain (b = .34,
four). p < .05). As shown in Figure 1, while ego develop-
The number of facets present (combining three ment was associated with greater meaning within
and four) marginally varied significantly by the the interpersonal domains, the opposite pattern
dominant domain, v2(4, N = 431) = 7.83, p = .09, was observed for the ideological domain: Greater
v = .10. However, the effect size was the same ego development was associated with less meaning.
magnitude as in Study 1, and the distributional None of the measures of identity status, either
pattern was similar: Interpersonal domains were ideological or interpersonal, significantly predicted
least likely to include three or more facets ( 2.3). meaning.
The ideological and existential domains did not In sum, the distribution of domains and likeli-
have any significant deviations, but all differences hood of co-occurrence were similar to Study 1.
were in the same direction as Study 1. While the association between co-occurrence and
There were no gender or age differences in the meaning was not significant, the effect was of a
distribution of the broad domains, facets, or co- similar magnitude to Study 1. Finally, meaning
occurrence. was associated with ego development, particularly
in the interpersonal domain, and there were no
relations between statuses and meaning.
How Are Domains Related to Process?
The subsequent analyses are identical to those
GENERAL DISCUSSION
reported in Study 1. We used multilevel modeling
(MLM) to account for the nesting of multiple sto- This study was motivated by the lack of attention
ries within participants. There were two focal pre- to how the two most common approaches to iden-
dictors: identity domain, which was dummy-coded tity development relate to each other, and by the
with the interpersonal domain specified as the ref- neglect of content in both status and narrative
erence category, and co-occurrence, which was approaches, even though content was central to
treated as a continuous variable. Participant age Erikson’s theorizing. When considering personally
and gender were included as Level 2 covariates, relevant content, there are three salient domains,
and narrative length was included as a Level 1 co- and that while rare, co-occurrence of content was
variate. All models were specified using full-infor- generally related to processes of identity develop-
mation maximum likelihood. In reporting the ment—meaning-making. Consistent with the exist-
results, we included Cohen’s d for the pairwise ing literature, we saw relatively modest
comparisons of the dummy-coded content predic- associations between narrative and status
tors and unstandardized regression coefficient (b) approaches, supporting our proposal that they may
for the continuous predictors. capture different levels of identity development—
personal and ego identity. Finally, these results
Meaning. Identity domain was significantly were relatively consistent across age, gender, peo-
associated with meaning (Table 2). Narratives ple with and without histories of trauma, and
within the interpersonal domain included signifi- memory prompt.
cantly more meaning than the existential domains
(d = .28, p < .05). The difference between the ideo- 1.2
logical and existential domain was of similar mag-
1
nitude to Study 1 (d = .25), but nonsignificant. Co-
occurrence was not significantly associated with 0.8
Meaning

meaning, although the effect was of similar magni- 0.6

tude to Study 1 (b = .12). There were no significant Existential
0.4 Ideological
interactions between domains and any other vari-
Interpersonal
ables in the model. 0.2

0
Connections to Ego Development and Identity -1 1
Ego Development (-1 = one SD below the mean and 1 = one SD
Status above)

Meaning. Ego development was positively FIGURE 1 Interaction between identity domain and ego
associated with meaning (b = .20, p < .05). Addi- development for meaning in Study 2.
 IDENTITY DEVELOPMENT 69

understanding of events by considering others’

Content Matters
perspectives, and they also provide powerful struc-
We propose that it does not matter only that indi- tures within which to locate oneself. Emerging
viduals are exploring their identities, but where adulthood also generally marks the beginning of
they are exploring these identities. Content tells us the end of cohabitation with the family (Arnett,
about the identities of our participants, as well as 2000; Goldscheider, Goldscheider, Clair, & Hodges,
the context in which they are developing. Ideology 1999), an important transition in Western culture
—one of the first domains identified by Erikson— marked by an improvement in the parent–child
was the least common in both studies. Within the relationship (Lahelma & Gordon, 2003). This transi-
ideological domain, we saw no instances—in over tion may also come with a renegotiation of roles,
1800 narratives—of politics and only 28 narratives making these family stories especially salient dur-
that included religion. While we collected data at ing this developmental stage.
liberal secular universities, the low base rate of reli- These findings are especially important because
gion is notable given that 80% of college students family is not particularly emphasized at the theo-
report being affiliated with a specific religion or retical level and is narrowly operationalized, from
denomination (Higher Education Research Institute the status perspective (see Crocetti et al., 2008; Me-
(HERI), 2004). The low base rate of narratives eus, 2011). For example, status assessments focus
about religion and politics may reflect a greater dif- on family in reference to other domains (e.g., one’s
ficulty with recalling specific instances of such agreement with parental religious beliefs; Bennion
events, or a lesser concern with these issues at this & Adams, 1986), or on how much someone has
life stage. For example, the lack of religion and considered the importance of family or one’s role
politics, in favor of vocation, is consistent with in the family (Balistreri et al., 1995). These ratings
work that has shown vocational concerns to be may not induce much variability, particularly for
primary to other ideological contents (Skorikov & importance, and are also rather vague; that is, the
Vondracek, 1998). From a developmental perspec- degree to which one has considered one’s role in
tive, it makes sense that occupation would be the the family does not tell us much about what that
most pressing concern for those in university. With consideration actually entails. However, reflecting
more committed relationships, children, and the on family stories to find connection to and distinc-
possibility of paying local taxes, religion and tion between oneself and one’s family is an identity
politics may become more salient concerns. process in which we see individual variation, and
Across both studies, the interpersonal domain which can be more fully described via narrative
was the most common. As we would expect from assessments (McLean, 2013; Zaman & Fivush,
research on the increasing importance of peer rela- 2011).
tionships over the course of adolescence and The original domains accounted for 60% of the
emerging adulthood (e.g., Youniss & Smollar, content of personal narratives in both studies. That
1985), dating and friendships were present, but we the existential domain is missing in the status
saw no instance of sex roles. We have seen similar approach may be a similar issue to the lack of
findings in a longitudinal study of college students attention given to the family domain—it is difficult
(Azmitia, Syed, & Radmacher, 2008), such that gen- to imagine what it looks like to explore and com-
der was not a facet of identity that participants mit to an existential identity. These data lead us to
could easily discuss, whether asked directly about the interpretation that the status approach exam-
it or not. One possibility is that there is an empha- ines identity from an agentic perspective with a
sis on an “equality” master narrative in the United focus on options one can actively explore (e.g., type
States, which might diminish experiences (literally of religion; see Waterman[forthcoming] for a differ-
or figuratively) in which gender is made salient. ent approach to identity that considers issues of
The great majority of narratives in the interper- eudaimonia and existentialism). Existential con-
sonal domain concerned family, consistent with cerns may be concerns that one has to wrestle to
theorizing within the narrative field, as well as in understand, not to choose between. This focus on
developmental psychology more broadly. For agentic exploration may leave out other arenas that
example, family stories are an important content of one considers identity-relevant, and which are
adolescent and emerging adult identity (e.g., Fiv- managed via different processes, such as meaning-
ush, Bohanek, & Duke, 2008; Pratt & Fiese, 2004). making. Existential concerns also tended to be
Fivush et al. (2008) have argued that family stories more likely to co-occur with other identity contents
provide opportunities for enriching one’s own (Study 1), suggesting that one way to manage
70 MCLEAN, SYED, YODER, AND GREENHOOT

existential concerns is to understand them as con- values is only understood as those values are
nected to, and in the context of, other identity con- linked to experiences with mortality and her rela-
cerns (see LaVoie & Vries, 2003), consistent with tionships, and her identity is represented by the
findings on co-occurrence. integration of relationships, values, and mortality—
this is the story that she is building that makes her
make sense to herself and to others. This is the re-
Identity Integration: Co-Occurrence Matters
configuration of relatively na€ıve childhood identifi-
Using Erikson’s terminology, we see co-occurrence cations into the reality of adult life that Erikson
as representing identity synthesis—the beginning emphasized.
of the process of piecing one’s identity together. When we examined the narratives of those who
When an individual sees facets as related to each reported co-occurrence without meaning, we gener-
other, co-occurrence, and is able to make meaning ally saw narratives that were more like “event
of events in which these facets are both relevant, descriptions.” Some of the same topics that we saw
this may be what Erikson referred to as identity above—mortality and relationships—were also
integration. While somewhat rare, when we did present but were listed almost as a chronology.
see co-occurrence in Study 1 (and in similar magni- Thus, they were not necessarily incoherent; that is,
tude in Study 2), we saw evidence of identity inte- they were temporally coherent, and sometimes
gration—meaning-making—supporting our idea even evaluatively coherent, but they lacked a kind
that when identity contents come in contact with of causal coherence, or meaning-making, which
each other it may prompt individuals to engage provides a sense of personal integration. For exam-
identity processes, as in the following narrative ple, another 19-year-old female reported the follow-
from a 19-year-old female: ing self-defining memory:

November 4th of my freshman year in high I had known that X was going to move away
school, my friend’s older brother, who was for a while, but the day had come. I went
the same class as my sister, was killed in a over to his house to play, & his dad
freak accident. He was in his truck while he answered, “He’s at his grandma’s; he’s gone.”
was camping and the bank collapsed, killing I turned and ran down the hill to my house.
him. Before this in my life, both of my grand- Once inside, I went into the tent of blankets
pas had passed away and of course it is so that we had made in my basement, and
sad, but before X [died], I never realized or remained there for I don’t know how long.
thought about how easy life can be taken Later that week, my grandpa died, so my
away. He was 23, on his way to be a fire mom was preoccupied with her loss. When I
fighter, and seemed like he had such a happy told her that I was sad that I lost my best
life. It made me realize that life is so fragile and friend, she got very angry because I wasn’t
we can never count on tomorrow. X’s death sad about her father.
occurred when my sister was abroad, study-
ing in India. So many people who I loved Here we see the contents of friendships, mortal-
were hurting so suddenly, and so deeply. It ity, and family. There is a clear temporal order
taught me the importance of reminding your par- here, but there is no explanation of how this per-
ents you love them, or anyone for that matter. I son thinks these events are connected. The story
have been fortunate enough not to have had makes basic sense, but it does not (yet) make sense
anything absolutely traumatic or horrible hap- as something that is self-defining. Given adult age-
pen, for that I am thankful. His death taught related increases in meaning-making (Pasupathi &
me how important it is to love the people you care Mansour, 2006), if this event continues to be self-
about and always remind them. defining for this person, she will have to construct
a thread that explains the connections between
In this narrative, the experience of death is nar- these losses and what they mean to herself.
rated as inextricably related to her relationships Consistent with other studies on the intersectional-
and her newly clarified values about relationships. ity, co-occurrence was rare in this age group (Azmitia
To disentangle these contents disentangles her et al., 2008; Syed, 2010). We suspect that across young
identity, which is beginning to cohere around these adulthood, as roles increasingly consolidate (Diehl &
themes via her meaning-making processes (itali- Hay, 2011), co-occurrence will become increasingly
cized). That is, her emerging conception of her common (Syed, 2010). Emerging adults are likely just
 IDENTITY DEVELOPMENT 71

beginning to see the intersection of their identity development reported less meaning. This is interest-
content domains and then only just likely to be able to ing because wrestling with values and roles and the
make sense of them via meaning-making processes. contradictions within those are a part of higher ego
This is likely due to cognitive abilities, which are development (Loevinger, 1976). This finding might
still developing during emerging adulthood, as well represent people who are still wrestling and have
as the social and cultural presses to provide a more not yet made meaning of these issues. It may also be
coherent story about the connections between the that these issues are not of concern to those at higher
various identity contents that become salient levels of ego development at this developmental
during this period. Our results point to the impor- stage. As we have noted, the ideological domain
tance of considering the complexity of the processes was less frequently reported, and those who are
used to integrate various content domains at a time wrestling with identity issues at higher levels of
when the demands for, and the abilities to do so, both ego development may be putting their attention
increase. elsewhere.
At a larger theoretical level, we have made a
claim that received some empirical support, which
Issues of Measurement and Theory: Convergence
is that these approaches target different levels of
and Divergence
identity work. This claim and these data should
In using both of the contemporary approaches to provide a more inclusive and coherent framework
identity, we saw some overlap in identity content for the study of identity development. For example,
and, consistent with previous studies, little overlap differentiating levels in terms of personal beliefs
with identity processes. Unexpectedly, we did not and roles (personal identity) and continuity
see an association between status achievement and through time (ego identity), as well as the degree
meaning. There are at least two explanations for of agency that one has in exploring various aspects
this. First, this may be methodological. The EO- of the self, provides a clearer map for interpreting
MEIS assessment has low reliabilities (Bennion & the meaning of various identity processes. Further,
Adams, 1986), which may compromise our abilities the emphasis on personally salient content reveals
to pick up these relationships, although other stud- the contextual nature of identity development. The
ies with the same measure have seen modest rela- study of content not only brings our participants’
tions (Alisat & Pratt, 2012; McLean & Pratt, 2006). identities into relief, but also tells us about the con-
Part of the explanation for low reliability is that text and culture in which they are developing.
process scores are averaged across content These claims return us to the heart of Erikson’s
domains. Second, as we have argued, we contend theory: The study of identity development should
that these approaches are really targeting different be multidimensional and contextually and cultur-
levels of Erikson’s conceptualization of identity, ally sensitive.
personal and ego.
In contrast to the status findings, ego develop-
LIMITATIONS AND CONCLUSIONS
ment was positively associated with meaning-mak-
ing, which was consistent with our expectations in We raise two main limitations of the current study.
which we conceptualized ego development as syn- First, we had relatively homogenous samples in
thesizing one’s experiences, an aspect of ego iden- regard to ethnicity, and completely homogenous
tity. This appears to be particularly likely when samples with regard to educational status. Ethnic
reflecting on relationships, which makes sense identity is a central domain of identity develop-
because those with higher levels of ego develop- ment (Phinney, 1990) and is also one that intersects
ment have a more mature and richer understand- with other identities (Azmitia et al., 2008), so exam-
ing of themselves and others. For example, the ining ethnic identity aids in understanding the
movement from the conscientious level (where intersectionality of identity contents. In terms of
most adult development occurs) through the higher education, while the majority of emerging adults
stages of ego development involves concerns shift- report attending some form of higher education
ing from a self-focus, particularly concerning dif- (Bureau of Labor and Statistics, http://www.bls.
ferentiation from others, to an understanding of gov/news.release/hsgec.nr0.htm), and our samples
the self in a social context, and from dependence as were drawn from large public universities and
a problem to an understanding of the importance classes that fulfill general education requirements,
of interdependence (Loevinger, 1976). In terms of we did not sample a fully generalizable group of
the ideological domain, those with higher ego emerging adults. We suspect that identity facets
72 MCLEAN, SYED, YODER, AND GREENHOOT

may vary more than identity processes in samples This is not to say that status approaches cannot
with different educational backgrounds, but only adequately target content, but we argue that cur-
future studies will answer that. Finally, we had rent approaches that rely on survey-rating assess-
limited variability in age to determine whether ments do not allow for the person to voice what is
some findings, notably co-occurrence, show age- most important to him or herself (see also Josselson
related increases across emerging adulthood. & Flum, forthcoming). The initial identity status
Second, we have raised the issue of the mea- interview (Marcia, 1966), on which many current
surement of identity statuses, and we hope that surveys are based, was a rich source of information
researchers will use more reliable assessments of that could be used to better understand salient
identity statuses in the future. We also note that identity contents, especially if informants are able
in asking people for momentous events, existen- to bring up issues that they find personally impor-
tial issues may be particularly salient because of tant as opposed to only asking questions about
the emotion and disruption they entail. Issues that predefined contents (see Fris
en & W€angqvist,
are still “in progress” may also be less amenable 2011). In sum, following Erikson and others (e.g.,
to requests for important past events. Researchers Josselson & Flum, forthcoming; McAdams, 1993;
may want to use different techniques, such as Thorne, 2000), we call for a refocusing on the sub-
asking directly for narrative accounts of facets jective voice of the person who can best tell us
and then assessing narrative quality. If individu- about the salient identity contents and how they
als are still unable to come up with narratives employ various processes to engage those contents.
about politics and religion, for example, or if In conclusion, we reiterate Erikson’s emphasis
those narratives are less elaborated than other fac- on the importance of paying attention to content
ets, then we can be more confident that these are and how it interacts with process, a distinction that
less salient. identity researchers, on the whole, have largely
In attempting to bring attention to the impor- neglected. We also conclude that current perspec-
tance of identity domain content, we have made tives on identity target different aspects of this
the argument that narrative is a more useful developmental process and that they are both
approach than are contemporary status approaches. important to consider in understanding this com-
First, we make this argument because of the subjec- plex phenomenon. We hope that researchers will
tivity of narrative, which allows an assessment of begin to acknowledge the variety of approaches
personal salience. Second, McLean and Pasupathi one might take to investigating identity develop-
(2012) argued that an elaborated story can help to ment, and to continue to investigate the similarities
maintain, and perhaps strengthen, the commit- and differences in these approaches (Syed, 2012).
ments one has made, from a status perspective. For Understanding the issues of identity content, espe-
example, Cox and McAdams (2010) found that stu- cially in reference to identity processes, is critical in
dents who told stories about experiences of self- gaining a fuller understanding of this complex
transformation after a volunteer trip to work with developmental task.
impoverished Nicaraguans were more likely to vol-
unteer after the trip, controlling for previous volun-
teer experience. Third, there is a substantial REFERENCES
literature on narrative development (e.g., Fivush, Alisat, S., & Pratt, M. W. (2012). Characteristics of young
Haden, & Reese, 2006; McLean et al., 2007), which adults’ personal religious narratives and their relation
provides a theoretical foundation for the study of with the identity status model: A longitudinal, mixed
how identity develops in these cultural valued set- methods study. Identity: An International Journal of
tings. For example, future studies may examine the Theory and Research, 12, 29–52. doi:10.1080/15283488.
degree to which conflict between domain contents 2012.632392
or facets is related to meaning-making, a hypothe- Arnett, J. (2000). Emerging adulthood: A theory of devel-
sis supported by Erikson’s original writings. opment from the late teens through the twenties.
American Psychologist, 55, 469–480. doi:10.1037/0003-06
Finally, narration is something that people actually
6X.55.5.469
do in their everyday lives (e.g., Fivush et al., 2006; Azmitia, M., Syed, M., & Radmacher, K. (2008). On the
McLean et al., 2007; Nelson & Fivush, 2004; Pasu- intersection of personal and social identities:
pathi, 2001; Thorne, 2000). The act of narrating is a Introduction and evidence from a longitudinal study
culturally valued and common activity, which of emerging adults. In M. Azmitia, M. Syed, &
brings the study of important psychosocial issues K. Radmacher (Eds.), The intersections of personal and
into the contexts in which they actually happen. social identities: New directions for child and adolescent
 IDENTITY DEVELOPMENT 73

development, pp. 120, 1–16). New York, NY: John Greenhoot, A. F., & McLean, K. C. (2013). The costs and
Wiley. doi:10.1002/cd.212 benefits of reflection in emerging adults’ narratives.
Balistreri, E., Busch-Rossnagel, N. A., & Geisinger, K. F. Manuscript in preparation.
(1995). Development and preliminary validation of the Greenhoot, A. F., Sun, S., Bunnell, S. L., & Lindboe, K.
ego identity process questionnaire. Journal of Adoles- (2013). Making sense of traumatic memories: Memory
cence, 18, 179–192. doi:10.1006/jado.1995.1012 qualities and psychological symptoms in emerging
Bauer, J. J., & McAdams, D. P. (2004). Personal growth in adults with and without abuse histories. Memory, 21,
adults’ stories of life transitions. Journal of Personality, 125–142. doi:10.1080/09658211.2012.712975
72, 573–602. doi:10.1111/j.0022-3506.2004.00273.x Grotevant, H. D. (1987). Toward a process model of iden-
Bennion, L. D., & Adams, G. R. (1986). A revision of the tity formation. Journal of Adolescent Research, 3, 203–222.
extended version of the Objective Measure of Ego Iden- doi:10.1177/074355488723003
tity Status: An identity instrument for use with late Grotevant, H. D., Thorbecke, W. L., & Meyer, M. L.
adolescents. Journal of Adolescent Research, 1, 183–197. (1982). An extension of Marcia’s Identity Status Inter-
Bruner, J. S. (1990). Acts of meaning. Cambridge, MA: view into the interpersonal domain. Journal of Youth
Harvard University Press. and Adolescence, 11, 33–47. doi:10.1007/BF01537815
Cox, K., & McAdams, D. P. (2010). The transforming self: Higher Education Research Institute (HERI). (2004). The
Service narratives and identity change in emerging spiritual lives of college students: A national study of col-
adulthood. Journal of Adolescent Research, 25, 1–26. lege students’ search for meaning and purpose. Los
doi:10.1111/j.0022-3506.2004.00273.x Angeles, CA: UCLA.
Crocetti, E., Rubini, M., & Meeus, W. (2008). Capturing van Hoof, A. (1999). The identity status field re-reviewed:
the dynamics of identity formation in various ethnic An update of unresolved and neglected issues with a
groups: Development and validation of a three-dimen- view on some alternative approaches. Developmental
sional model. Journal of Adolescence, 31, 207–222. doi:10. Review, 19, 497–556. doi:10.1006/drev.1999.0484
1016/j.adolescence.2007.09.002 Hy, L. X., & Loevinger, J. (1996). Measuring ego develop-
Diehl, M., & Hay, E. L. (2011). Self-concept differentiation ment (2nd ed.) Mahwah, NJ: Lawrence Erlbaum.
and self-concept clarity across adulthood: Associations Josselson, R. (1982). Personality structure and identity
with age and psychological well-being. The Interna- status in women as viewed through early memories.
tional Journal of Aging and Human Development, 73, Journal of Youth and Adolescence, 11, 293–299. doi:10.
125–152. doi:10.2190/AG.73.2.b 1016/BF01537171
Empirisoft. (2008) Media Lab v2008 [computer software]. Josselson, R., & Flum, H. (forthcoming). Identity status:
http://www.empirisoft.com. On refinding the people. In K. C. McLean & M. Syed
Erikson, E. H. (1950). Childhood and society. New York, NY: (Eds.), The Oxford handbook of identity development. New
W. W. Norton. York: Oxford University Press.
Erikson, E. H. (1959/1980). Identity and the life cycle; 5 Kroger, J. (2002). Identity processes and contents through
Selected papers, with a historical introduction. New York, the years of late adulthood. Identity: An International
NY: International Universities Press. Journal of Theory and Research, 2, 81–99. doi:10.1207/S15
Erikson, E. (1968). Identity, youth and crisis. New York, 32706XID0201_05
NY: W. W. Norton. Kroger, J. (2007). Why is identity achievement so elusive?
Fadjukoff, P., Pulkkinen, L., & Kokko, K. (2005). Identity Identity: An International Journal of Theory and Research,
processes in adulthood: Diverging domains. Identity: 7, 331–348. doi:10.1080/15283480701600793
An International Journal of Theory and Research, 5, 1–20. Kroger, J., Martinussen, M., & Marcia, J. E. (2010). Iden-
Fivush, R., Bohanek, J. G., & Duke, M. (2008). The inter- tity stat change during adolescence and young adult-
generational self: Subjective perspective and family hood: A meta-analysis. Journal of Adolescence, 12,
history. In F. Sani (Ed.), Individual and collective self-con- 683–698. doi:10.1016/j.adolescence.2009.11.002
tinuity (pp. 131–144). Mahwah, NJ: Lawrence Erlbaum. Labouvie-Vief, G. (2006). Emerging structures of adult
Fivush, R., Haden, C. A., & Reese, E. (2006). Elaborating thought. In J. J. Arnett & J. L. Tanner (Eds)., Emerging
on elaborations: Role of maternal reminiscing style in adults in America: Coming of age in the 21st century
cognitive and socioemotional development. Child (pp. 59–84). Washington, DC: American Psychological
Development, 77, 1568–1588. doi:10.1111/j.1467-8624.20 Association.
06.00960.x Lahelma, E. & Gordon, T. (2003). Home as a physical,
Fris
en, A., & W€angqvist, M. (2011). Emerging adults in social and mental space: Young people’s reflections on
Sweden: Identity formation in the light of love, work, leaving home. Journal of Youth Studies, 6, 377–390.
and family. Journal of Adolescent Research, 26, 200–221. LaVoie, J., & Vries, B. (2003). Identity and death: An
doi:10.1177/0743558410376829 empirical investigation. Omega: Journal of Death and
Goldscheider, F., Goldscheider, C., Clair, P. T., & Hod- Dying, 48, 223–243. doi:10.2190/H2K9-15Q7-U7Q5-3C
ges, J. (1999). Changes in returning home in the United QA
States, 1925–1985. Social Forces, 78, 695–721. doi:10. Loevinger, J. (1976). Ego development: Conceptions and theo-
1093/sf/78.2.695 ries. San Francisco, CA: Jossey-Bass.
74 MCLEAN, SYED, YODER, AND GREENHOOT

Loevinger, J., Cohn, L. D., Bonneville, L. P., Redmore, C. McLean, K. C., & Thorne, A. (2003). Adolescents’ self-
D., Streich, D. D., & Sargent, M. (1985). Ego develop- defining memories about relationships. Developmental
ment in college. Journal of Personality and Social Psychol- Psychology, 39, 635–645. doi:10.1037/0012-1649.39.4.635
ogy, 48, 947–962. doi:10.1037/0022-3514.48.4.947 Meeus, W. (2011). The study of adolescent identity for-
Luyckx, K., Goossens, L., & Soenens, B. (2006). A devel- mation 2000–2010: A review of longitudinal research.
opmental contextual perspective on identity construc- Journal of Research on Adolescence, 21, 75–94. doi:10.
tion in emerging adulthood: Change dynamics in 1037/0012-1649.39.4.635
commitment formation and commitment evaluation. Neimeyer, G. J., & Rareshide, M. B. (1991). Personal
Developmental Psychology, 42, 366–380. doi:10.1037/ memories and personal identity: The impact of ego
0022-3514.48.4.947 identity development on autobiographical memory
Mansfield, C., McLean, K. C., & Lilgendahl, J. P. (2010). recall. Journal of Personality and Social Psychology, 60,
Narrating traumas and transgressions: Links between 562–569. doi:10.1037/0012-1649.39.4.635
narrative processing, wisdom, and well-being. Narra- Nelson, K., & Fivush, R. (2004). The emergence of auto-
tive Inquiry, 20, 246–273. biographical memory: A social cultural developmental
Marcia, J. E. (1966). Development and validation of ego- theory. Psychological Review, 111, 486–511. doi:10.1037/
identity status. Journal of Personality and Social Psychol- 0012-1649.39.4.635
ogy, 5, 551–558. doi:10.1037/h0023281 Orlofsky, J. F., & Frank, M. (1986). Personality structure
Marcia, J. E. (1993). The ego identity status approach to as viewed through early memories and identity status
ego identity. In J. E. Marcia, A. S. Waterman, D. R. in college men and women. Journal of Personality and
Matteson, S. L. Archer, & J. L. Orlofsky (Eds.), Ego Social Psychology, 50, 580–586. doi:10.1037/0022-3514.50.
identity: A handbook for psychosocial research (pp. 1–21). 3.580
New York, NY: Springer-Verlag. Pasupathi, M. (2001). The social construction of the per-
McAdams, D. P. (1993). The stories we live by: Personal sonal past and its implications for adult development.
myths and the making of the self. New York, NY: Guil- Psychological Bulletin, 127, 651–672. doi:10.1037/0033-29
ford Press. 09.127.5.651
McAdams, D. P. (1998). Ego, trait, identity. In P. M. We- Pasupathi, M., Brubaker, J., & Mansour, E. (2007). Devel-
stenberg, A. Blasi, & L. D. Cohn (Eds.), Personality oping a life story: Constructing relations between self
development: Theoretical, empirical, and clinical investiga- and experience in autobiographical narratives. Human
tions of Loevinger’s conception of ego development (pp. Development, 50, 85–110. doi:10.1159/000100939
27–38). Mahwah, NJ: Lawrence Erlbaum. Pasupathi, M., & Mansour, E. (2006). Adult age differ-
McAdams, D. P. (2006). Foley Center, Guided Autobiogra- ences in autobiographical reasoning in narratives.
phy. Retrieved November 20, 2006, from http://www. Developmental Psychology, 42, 798–808. doi:10.1037/00
sesp.northwestern.edu/foley/instruments/guided 12-1649.42.5.798
McLean, K. C. (2005). Late adolescent identity develop- Pasupathi, M., Wainryb, C., & Twali, M. (2012). Narrative
ment: Narrative meaning-making and memory telling. construction of group-based differential treatment:
Developmental Psychology, 41, 683–691. doi:10.1037/ Implications for ethnic identity development. Identity:
0012-1649.41.4.683 An International Journal of Theory and Research, 12, 53–
McLean, K. C. (2013, January). Finding my own way: 73. doi:10.1080/15283488.2012.632393
Negotiating personal identity from family identity. Phinney, J. S. (1990). Ethnic identity in adolescents and
Paper presented at the meeting of the Society for adults: A review of research. Psychological Bulletin, 108,
Personality and Social Psychology, New Orleans, LA. 499–514. doi:10.1037/0033-2909.108.3.499
McLean, K. C., & Fournier, M. (2008). The content and Piaget, J. (1965). The moral judgment of the child. New
processes of autobiographical reasoning in narrative York, NY: Free Press.
identity. Journal of Research in Personality, 42, 527–545. Pratt, M. W., & Fiese, B. H. (2004). Family stories and the
doi:10.1016/j.jrp.2007.08.003 life course: Across time and generations. Mahwah, NJ:
McLean, K. C., & Pasupathi, M. (2012). Processes of iden- Lawrence Erlbaum.
tity development: Where I am and how I got there. Rubin, D. C., Schrauf, R. W., & Greenberg, D. L. (2003).
Identity: An international journal of theory and research, Belief and recollection of autobiographical memories.
12, 8–28. doi:10.1016/j.jrp.2007.08.003 Memory and Cognition, 31, 887–901. doi:10.3758/BF03
McLean, K. C., Pasupathi, M., & Pals, J. P. (2007). Selves 196443
creating stories creating selves: A process model of Schwartz, S. J. (2001). The evolution of Eriksonian and
self-development. Personality and Social Psychology neo-Eriksonian identity theory and research: A review
Review, 11, 262–280. doi:10.1177/1088868307301034 and integration. Identity: An International Journal of The-
McLean, K. C., & Pratt, M. W. (2006). Life’s little (and ory and Research, 1, 7–58.
big) lessons: Identity statuses and meaning-making in Schwartz, S. J. (2005). A new identity for identity
the turning point narratives of emerging adults. Devel- research: Recommendations for expanding and refo-
opmental Psychology, 42, 714–722. doi:10.1037/0012-16 cusing the identity literature. Journal of Adolescent
49.42.4.714 Research, 20, 293–308. doi:10.1177/0743558405274890
 IDENTITY DEVELOPMENT 75

Singer, J. A., & Moffitt, K. H. (1991–1992). An experimen- emerging adults’ ethnicity-related experiences. Develop-
tal investigation of specificity and generality in mem- mental Psychology, 46, 208–219. doi:10.1037/a0017825
ory narratives. Imagination, Cognition, and Personality, Thorne, A. (2000). Personal memory telling and personal-
11, 233–257. doi:10.2190/72A3-8UPY-GDB9-GX9K ity development. Personality and Social Psychology
Skorikov, V., & Vondracek, F. W. (1998). Vocational iden- Review, 4, 45–56. doi:10.1207/S15327957PSPR0401_5
tity development: Its relationship to other identity Thorne, A., McLean, K. C., & Lawrence, A. M. (2004).
domains and to overall identity development. Journal When remembering is not enough: Reflecting on
of Career Assessment, 6, 13–35. doi:10.1177/10690727980 self-defining memories in late adolescence. Journal of
0600102 Personality, 72, 513–541. doi:10.1207/S15327957PSPR0
Syed, M. (2010). Developing an integrated self: 401_5
Academic and ethnic identities among ethnically- Waterman, A. L. (forthcoming). Identity as internal
diverse college students. Developmental Psychology, 46, processes: How the “I” comes to define the “me.” In
1590–1604. doi:10.1037/a0020738 K. C. McLean & M. Syed (Eds.), The Oxford handbook of
Syed, M. (2012). The past, present, and future of Erikso- identity development. New York: Oxford University
nian identity research: Introduction to the special Press.
issue. Identity: An International Journal of Theory and Youniss, J., & Smollar, J. (1985). Adolescent relations with
Research, 12, 1–7. doi:10.1080/15283488.2012.632362 mothers, fathers, and friends. Chicago, IL: University of
Syed, M., & Azmitia, M. (2008). A narrative approach to Chicago Press.
ethnic identity in emerging adulthood: Bringing life to Zaman, W., & Fivush, R. (2011). When my mom was a
the identity status model. Developmental Psychology, 44, little girl. . .: Gender differences in adolescents’ inter-
1012–1027. doi:10.1080/15283488.2012.632362 generational and personal stories. Journal of Research on
Syed, M., & Azmitia, M. (2010). Narrative and ethnic Adolescence, 21, 703–716. doi:10.1207/S15327957PSPR0
identity exploration: A longitudinal account of 401_5