Introduction

Erikson (1968) stated that young individuals need to build a relationship with society at large through identification with an occupation, and through religious and political views. The identity formation process incorporates the negotiation of these issues and values and is important because decisions made during this time period can affect the rest of a person’s life (e.g., occupation, romantic partnerships, etc.). Erikson called identity development a psychosocial process, because individuals conduct the work of identity construction within a rich social context, including family, extended kin, peers and others. All of Erikson’s previous four stages prior to the consolidation of a person’s identity involve social components including the family. Therefore, changes in the American family (such as an increasing number of children and adolescents growing up in single or blended families) (Emery 1999) need to be examined. It is important to understand if and how family structure affects identity exploration and commitment during adolescence and young adulthood.

The essence of identity work involves exploring identity alternatives and making personally meaningful decisions or commitments. In 1966, Marcia undertook the task of operationalizing some of Erikson’s aspects of identity formation in terms of measurable exploration and commitment in three general domains: occupational choice, religious beliefs, and political ideology. By attending to the outcome of the two processes, he developed the identity status paradigm, which evaluates the occurrence or non-occurrence of the exploration of alternatives and the achievement or non-achievement of commitment, and yields four possible outcomes or statuses. There are four possible outcomes called statuses. They are: (a) achieved—individuals who have explored, and then made a commitment; (b) moratorium—individuals who are exploring, but have not yet committed (not surprisingly, this status is often found among college students); (c) foreclosed—individuals who have not explored, but nevertheless made a commitment; and (d) diffused—individuals who have neither explored nor made a commitment. The status paradigm has made significant contributions pertaining to adolescent identity development (Bluestein and Palladino 1991).

More recent measurement strategies used in identity research have increased the number of domains to include not only occupation, religion, and politics, but also values, friendship, dating, sex roles, family, and recreation (Bennion and Adams 1986; Craig-Bray and Adams 1986; Grotevant et al. 1982). These eight domains are frequently collapsed into one global identity domain or two “grand” domains—ideological and interpersonal (e.g., Adams 1985). The ideological domain includes occupation, religion, politics and values. The interpersonal domain is comprised of friendship, dating, family, and recreation. However, some research gives separate treatment to all eight domains (e.g., Pastorino et al. 1997). Results reveal that differences do occur for identity exploration and commitment level based on gender (Pastorino et al. 1997), age (Marcia 1993), and other variables (for a detailed review see Marcia 1993). However, to our knowledge the link between identity exploration and commitment and family structure has been less explored.

A critical context for understanding identity, even among late adolescents and young adults, is the family. The epigenetic framework applied by Erikson suggests that each of the five developmental stages of psychosocial development (including the consolidation of one’s young adult identity) involves the family; and each positive resolution in previous stages increases the likelihood of a positive resolution to the identity task in adolescence and early adulthood.

Demographic changes in the American family raise interesting questions about how the family context may influence identity processes among late adolescents and young adults. Increasing numbers of families reorganize their structure through divorce and stepfamily formation (Amato 2003). In addition, an impressive number of families never organize themselves into the traditional nuclear arrangement characterized by heterosexual marriage partners rearing biological or adopted offspring. Family transitions and/or reorganizations has long been thought to influence child outcomes. Many studies conducted since the mid 1970s have focused on the mostly negative consequences that youth may experience due to parental divorce. Findings included an increase in depression, a decline in school performance, and other behavior problems (Amato 2003). Approximately 25% of children from divorced families (in comparison to 10% of children from original families) exhibit behavioral and emotional problems, internalization problems, and other adjustment problems (Amato 2003; Hetherington and Kelly 2002). Nevertheless, the majority of children from divorced families remain emotionally healthy and well adjusted as they grow up. The important question this raises is whether the structure of the family within which an adolescent or young adult is raised matters directly to patterns of identity exploration and commitment. If family structure is relevant in these processes, it could matter across identity domains, or its influence could be more pronounced in some domains in comparison to others.

Surprisingly, the influence of family structure on identity formation and ultimately adult life has rarely been studied. Marcia (1993) reviewed the limited literature connecting family structure to identity development. Findings revealed that father-absence and divorce were associated with a greater likelihood of diffused identities for males (Oshman and Manosevitz 1976), meaning that males from divorced or single families were less likely to explore and commit compared to females. On the other hand, early adolescent girls growing up in single-parent homes were more likely to be achieved compared to boys (Streitmatter 1987). Girls growing up in single families were more likely to explore and commit compared to boys. However, these findings were not replicated using a college sample. No overall differences in identity status were found between males and females growing up in divorced, mother-headed homes (Imbimbo 1995). Interestingly, using a domain-specific analysis, college-aged females from divorced families, compared to similar males, were more likely to be achieved in the occupational domain (Imbimbo 1995). Given the higher probability that post-divorce family structure involved living with a single mother, Imbimbo (1995) hypothesized that daughters of single working mothers have a more positive role model, whereas males often may not have a gender specific role model.

Given the limited understanding of how family structure and gender influence the formation of identity, the purpose of our study was to examine the extent and nature of identity formation among older adolescents and young adults and to investigate the relationships between identity exploration/commitment and family structure, gender, and age.

Method

A survey instrument was used to examine identity exploration and commitment among young adults between the ages of 19 and 25.

Participants

388 undergraduate students from a southeastern university in the United States participated in this study. Seventy-five percent (n = 293) were females and 25% (n = 95) were males. This gender composition is consistent with the imbalance of the authors’ department from which the participants were recruited. The average age, at the time of the study was 20 years (SD = 1.45) ranging from 19 to 25. The majority of the sample was white (94%). Most participants stated that their parent figures worked, and had at least 14 years of education.

In terms of family structure, 72% (n = 280) grew up in original families, and 28% (n = 108) grew up in other family structures. Specifically for six, the biological parents never married; for 12, one biological parent died before the subject reached 19 years of age; and for the remaining 90, a parental divorce had occurred. Non-original families shared the possibility of subsequent family reorganizations such as remarriages and divorces. For the current analysis, all non-original families were treated as one category.

Measures

Family Structure

“Original family structure” was defined as two biological/adoptive parents still married and living together at the time of data collection. This structure was contrasted with “non-original family structures” (divorced, never married, remarried, stepfamily) for the purposes of this paper. If treated separately, the other structures would create groups too small to analyze.

Age

This variable was categorized into a dichotomy; less than 20 years of age (coded 0, n = 161), and 20 years and older (coded 1, n = 227).

Identity Exploration and Commitment

The Ego Identity Process Questionnaire (EIPQ) (Balistreri et al. 1995), consists of 32 items that describe the dimensions of exploration and commitment within two global identity domains (the ideology domain consisted of items pertaining to occupation, religion, politics, and values; the interpersonal domain consisted of items pertaining to family, friendship, dating, and sex-roles). The measure was used in a continuous rather than a categorical form to maximize the variability in exploration and commitment, and thus maximize the information available for analysis (Anthis 2003). Responses were assessed using a 5-point Likert Scale ranging from “strongly disagree” to “strongly agree.” An example item for commitment in the occupational domain is: I have definitely decided on the occupation I want to pursue.

Alpha coefficients for this sample were as follows: Exploration ideology = .62, commitment ideology = .64, exploration interpersonal = .56, and commitment interpersonal = .67. Although these reliability coefficients are lower than preferred, they are consistent with other studies using this measure with young adults (Luyckx et al. 2006).

Data Analysis

Our research examines profiles of exploration and commitment in the interpersonal and ideological domains and how family structure, gender and age are associated with these profiles. Given two observations (one for exploration and one for commitment) on each of two identity domains (ideological and interpersonal), we examine patterns of observations and establish whether associations exist between these patterns and individual attributes (i.e., age, gender, family of origin). The statistical procedure used to conduct these examinations is a doubly-multivariate repeated measure ANOVA (Tabachnick and Fidell 2006). In our analysis, two within-subject (repeated) factors are identified: Factor 1 compares “Exploration–Commitment.” It aggregates scores for each type of identity work across these two domains (ideological and interpersonal). Therefore, a main effect for “Exploration–Commitment” suggests that respondents report more exploration or more commitment, ignoring domain. Factor 2, “Ideological–Interpersonal” compares the two identity domains. It aggregates scores for each domain across the two types of identity processing. A main effect for this factor, therefore, indicates more identity processing (ignoring whether it is in the form of exploration or commitment) in one domain compared to the other. A significant interaction between either main effect and an individual difference variable (age, gender, family of origin status) indicates that one type of subject (e.g., males versus females) has a different pattern than the other. These main effects and interactions are important, but they tell only part of the story of interest here. We are especially interested in the interaction between the two factors (i.e., Exploration–Commitment and Ideology–Interpersonal) because, if found, it would indicate that different patterns of exploration and commitment exist for the two domains. Given our attention to family structure effects, we have a keen interest in whether a three-way interaction exists between the within-subject factors and the family of origin status variable. If such a three-way interaction would be discovered, it would indicate that differing patterns of identity work in the ideological and interpersonal domains depend on the family origin status of our research participants.

Results

Table 1 shows the multivariate test results of the doubly-multivariate ANOVA. The results display two main effects for Exploration–Commitment and Ideology–Interpersonal, and for their interaction. Exploration–Commitment shows a significant interaction with age. Ideology–Interpersonal reveals an interaction with gender. To interpret these findings, the means in Table 2 are important. The significant main effect for Exploration–Commitment indicates that across participants (and ignoring the identity domain) commitment scores are greater than exploration scores. The interaction between Exploration–Commitment with age indicates that the lowest score is for exploration among the participants under age 20. Exploration scores among older participants are significantly greater than that of the younger group, but significantly lower than either of the commitment scores. Finally, although both commitment scores are greater than either exploration score, commitment scores do not differ by age category.

Table 1 Mulitvariate test results of doubly multivariate ANOVA for exploration–commitment identity dimensions and ideological–interpersonal identity domains (within-subjects) and gender, family of origin status, and age (between subjects)
Full size table
Table 2 Means and standard errors for main effects of Ex–Com and Ideo–Interp and their significant interactions
Full size table

The main effect for Ideology–Interpersonal indicates that across all participants (and ignoring the levels of exploration and commitment) participants have higher scores in the interpersonal domain. The significant interaction between Ideology–Interpersonal and gender indicates that (a) scores in the ideological domain did not differ by gender, (b) scores in the interpersonal domain, compared to the ideological domain, are greater for both males and females, and (c) the highest scores, by a considerable margin, are for females in the interpersonal domain.

The means needed to interpret the interaction between Exploration–Commitment and Ideology–Interpersonal are shown in Table 3. These means represent the general profile for the full sample. By a substantial margin, exploration in the ideological domain is reported less than any other identity processing activity. A significantly greater amount of exploration occurs in the interpersonal domain, but the level of commitment in both domains is significantly greater. Commitment levels in the two domains are not significantly different. The difference in the reported level of exploration in the two domains is nearly half a standard deviation. Interestingly, the exploration–commitment difference is twice as large in the ideological domain as in the interpersonal domain.

Table 3 Means, standard deviations, standard errors, and mean difference scores for general profile for exploration and commitment in the ideological and interpersonal domains
Full size table

The interaction of Exploration–Commitment, Ideology–Interpersonal and family of origin status was the only three-way interaction that was statistically significant. Table 4 presents the relevant means for interpreting this interaction. The largest standard error for these means was SE = 0.45, so means differing by more than 0.90 exceeded the 95% confidence interval. Along with the relevant means, standard errors of the means and standard deviations, Table 4 presents two sets of discrepancies. The first are direct comparisons between original and non-original family status in each of the four combinations of the table (e.g., ideology with exploration). These discrepancies are presented beneath the compared means. In the ideological domain, young adults living in non-original families revealed significantly more exploration, but equal levels of commitment compared to those living in original families. In the interpersonal domain, family structure was not linked to either level of exploration or commitment.

Table 4 Means, standard deviations, and difference scores for exploration and commitment in the ideological and interpersonal domain by family structure
Full size table

The second set of discrepancies compares amounts of exploration and commitment in the two domains within each category of family status (see right-most column of Table 4). Young adults living in original families reported considerably more exploration in the interpersonal domain than they did in the ideological domain, but, interestingly, this greater exploration did not convert to higher levels of commitment in that domain. Indeed, their commitment levels for both ideological and interpersonal domains were statistically identical even though their exploration in the interpersonal domain was higher by a half standard deviation.

For the young adults living in non-original families, greater exploration appeared to convert directly to greater commitment. The significant discrepancy between exploration in these two domains matches almost perfectly the discrepancy between the reported levels of commitment in these two domains.

Discussion

The purposes of our study were to examine the influences of family structure, gender, and age on identity processes, using profiles, among young adults. The general profile for the sample indicated a tendency for commitment levels to exceed exploration levels, and for the interpersonal domain to receive greater identity-focused attention than the ideological domain. Drilling into this profile more deeply, we observed a significant interaction between these two tendencies. Specifically, the exploration–commitment difference was more pronounced in the ideological domain than in the interpersonal domain. In the ideological domain, which includes career, religious and political values, and one’s philosophy of life, commitments appeared to be crystallizing without much exploration of alternatives.

Using Marcia’s status language, these results may indicate that these young adults are making commitments in the ideological domain with relatively little exploration, suggesting a tendency toward foreclosure in the ideological domain. Consistent with this view are studies that suggest many young people take a more passive approach to the identity process, either letting developments occur (a characteristic of diffusion) or following their parents’ advice (a characteristic of foreclosure) (Côté 2006; Kroger 2000; van Hoof 1999).

Also consistent with this interpretation is the interaction between age and the exploration–commitment dimension. Younger participants reported significantly less exploration than older participants, suggesting a possible developmental progression (although not proving it since the present data are cross-sectional).

The tendency for levels of exploration in the interpersonal domain to exceed those in the ideological domain offers a clear indication of the relative salience of these two domains to the participants of this study. Among these young adults, the interpersonal identity domain (organized around gender roles) and relationships with family, friends and dating partners, appear more salient than the ideological domain. The fact that the sample is so heavily weighted toward female participants suggests that these results may partly reflect the sample’s composition. It would be consistent with feminine sex role orientation to emphasize expressive elements of identity domains (Grotevant et al. 1982). The significant interaction between the domain-linked repeated measure and gender in the results is fully consistent with this view. Although males and females did not differ in their identity-linked attention to the ideological domain, females were significantly more invested in the interpersonal domain than were males. The relatively high level of commitment among both genders in both domains suggests a sense of certainty. A prolonged state of moratorium (individuals are exploring but have not made a commitment) does not appear to be characteristic of these young adults.

Our findings in this study also contribute to a better understanding of how family structure may influence identity processes. In this study, young adults living in non-original families explored more in the ideological domain than those living in original families. This pattern was not found in the interpersonal domain. In fact, in this domain, family structure had no effect on the level of exploration, but young adults living in non-original families showed higher commitment compared to those living in original families. Furthermore, even though young adults from original families reported more exploration in the interpersonal domain than they did in the ideological domain, this increased exploration did not result in higher commitment. On the other hand, for young adults living in non-original families, greater exploration resulted in greater commitment, and matched the greater level of commitment as compared to the ideological domain. In a sense, the exploration undertaken by young adults from non-original family structures seemed to be more productive.

These findings are noteworthy for several reasons. The higher exploration in the ideological domain reported by participants from non-original families might be an indication of an increased capability in perspective taking (Kogos and Snarey 1995). Additionally, studies have shown that individuals from non-original families have more opportunities to take on and observe different roles, have more autonomy at an earlier age, and show a higher stage of moral development than students from homes with two biological parents (Hetherington and Kelly 2002; Kogos and Snarey 1995). Current results suggest a similar trend might apply to identity formation.

Our findings seems to suggest that, when those individuals raised for a period of time in a non-original family structure, they may put more effort into issues of self development related to the ideological identity domain, and thus, their choices pertaining to occupation, politics, religion, and values. This findings are interesting because research findings often focus on negative outcomes (Amato and Booth 1996; Emery 1999) and overlook the strengths that can be demonstrated by young individuals growing up in non-original families.

The limitations of our study are shared by many studies undertaken in the field of identity research. The data collection strategy is cross-sectional, which strictly limits our ability to think in causal terms. Additionally, the history of exploration and commitment for participants was unavailable. Perhaps exploration had been undertaken and completed in years past. Our measure of exploration and commitment was designed to tap identity processes as Marcia (1993) and colleagues have conceptualized them. We did not tap exploration in depth or identification with commitment (Luyckx et al. 2006). It is therefore possible that these young adults may have moved beyond broad explorations and orienting commitments to deeper, more serious identity work than our measure assessed.

A more demographically balanced sample would increase statistical power, and a more representative sample would enhance generalizability. Among the participants with experience in a non-original family structure, there is great variability in the nature of that experience that is not factored into the current analysis. It is impressive that the interesting pattern of results emerged without including the many complexities that make various non-original family forms so diverse.

Future research would benefit from a closer inspection of the identity domains. We only addressed two grand domains, and more attention needs to be given to specific domains. It would be beneficial to include factors such as time spent in such family structures, the number of structural reorganizations experienced, and the individuals’ attitudes about the changes in further investigations. Longitudinal designs would also help to examine the factors mentioned above. Furthermore, aspects of family functioning as a moderator of structural influences on identity processing should be explored in future research.