When the separation-individuation process goes awry

Abstract

Problematic separation-individuation has been conceptualized almost unilaterally as separation anxiety or as intolerance of being alone (i.e., dysfunctional dependence). However, as separation-individuation involves a dynamic interaction between independence and relatedness, it was argued in this study that disturbances in the separation-individuation process could manifest in at least two ways; that is, as dysfunctional dependence and as dysfunctional independence. In a sample of 232 emerging adults, we examined correlates and outcomes of the two types of dysfunctional separation-individuation. We found that both types were related in similar ways to depressive symptoms and a general measure of pathological separation-individuation. Yet they were associated differentially and in theoretically expected ways with: (a) dimensions of attachment (i.e., anxiety and avoidance); and (b) dimensions of personality that confer vulnerability to depression (i.e., dependency and self-criticism). In addition, person-centered results showed evidence for four groups of individuals with distinct profiles of separation-individuation (i.e., healthy, dysfunctional dependent, dysfunctional independent, and combined). Implications for clinical practice and future research are discussed.

Keywords

dependence dysfunctional emerging adulthood independence separation-individuation

Separation-individuation is an intrapsychic process that reverberates throughout the life cycle. It refers to the establishment of a sense of self, separate from other primary love objects (i.e., separation) and the acquisition of one’s unique individuality (i.e., individuation; Mahler, Pine, & Bergman, 1975). Its principal developmental challenges are situated in early childhood and adolescence/emerging adulthood; the first (Mahler, 1963) and second phase of separation-individuation (Blos, 1967), respectively. Throughout both phases, the child gradually reduces psychological dependence from significant others, especially parents, while trying to maintain a sense of connectedness with them. Hence separation-individuation is about the resolution of a complex dialectical interaction between independence and relatedness (Allen, Hauser, Eickholt, Bell, & O’Connor, 1994; Grotevant & Cooper, 1986).

From this notion of an interaction between independence and relatedness, it follows that disturbances with separation-individuation may manifest in at least two different ways. Specifically, dysfunctional separation-individuation may manifest as inadequate coping with the issue of independence or as inadequate coping with the issue of relatedness. In this regard we argue that research on separation-individuation can be informed by attachment theory-based research and research on personality vulnerability. In both literatures, a distinction has been made between an orientation primarily involving concerns with relatedness (i.e., attachment anxiety and dependency) and an orientation involving concerns with individuality (i.e., attachment avoidance and self-criticism; Blatt & Maroudas, 1992). Yet current diagnostic models and empirical frameworks have tended to emphasize dysfunctional dependence as the main manifestation of problematic separation-individuation. To redress this imbalance, the overarching goal of this study was to examine commonalities and differences between dysfunctional dependence and dysfunctional independence. Specifically, this goal is pursued by examining associations of the two manifestations of unsuccessful separation-individuation (a) with commonly used measures of pathological separation-individuation, and (b) with measures of dimensions of attachment and personality vulnerability to depression. These research questions were examined in a sample of emerging adults because the process of separation-individuation forms a critical turning point in their movement towards more independent functioning (Tanner, 2005).

The first and second process of separation-individuation

During the first separation-individuation process, the child emerges from the symbiotic relation with the caregiver to become an individuated toddler, experiencing a “sense of identity” for the first time (Mahler, 1963). Different subphases of this first process of separation-individuation reflect the child’s oscillation between a need to defend its recently achieved independence and a wish for reunion with the caregiver. At the final stage the child should be able to find an optimal balance between distance from and closeness to the caregiver. Such balance is achieved through identification with the caregiver and through internalization of rules and demands (Mahler et al., 1975).

Puberty marks the onset of the second process of separation-individuation (Blos, 1967, 1979). Due to both physical and cognitive development, adolescents no longer see themselves as children and caregivers too are no longer perceived as the almighty figures they once were during childhood. This process of de-idealization gives adolescents the opportunity to actively search for who they are outside the family of origin. However, similar to the first separation-individuation process, adolescents are likely to experience feelings of ambivalence over their newly gained independence as they wish to remain connected to the parents (Allen et al., 1994; Grotevant & Cooper, 1986). This ambivalence can be overcome by gradually transforming the hierarchical parent–child relationship of childhood into a mutual relationship between two equal adults (Levy-Warren, 1999).

Because the transition to adulthood is prolonged in Western societies, the redefinition of the relation between the individual and his/her parents typically continues beyond adolescence. Extant research has provided evidence for this idea, with separation-individuation processes in late adolescence and emerging adulthood being associated with psychological adjustment (Holmbeck & Leake, 1999; Holmbeck & Wandrei, 1993) and adjustment to college (Lapsley, Rice, & Shadid, 1989; Mattanah, Brand, & Hancock, 2004). These findings suggest that the transformation of the parent–child relationship, entailing a shift from parent regulation to self-regulation, is active and highly salient during emerging adulthood (Tanner, 2005). Therefore it was decided to focus on emerging adults in this study.

Unsuccessful resolution of the separation-individuation process

Individuals’ success in progressing through the separation-individuation process is thought to have implications for personal adjustment and later psychosocial functioning. Conversely, when people fail to deal adequately with the developmental task of separation-individuation, they are thought to become vulnerable to psychopathology (Blos, 1979; Mahler et al., 1975; Pine, 1979). Given that a healthy resolution of the separation-individuation process entails obtaining an optimal balance between closeness and distance in close relationships (Grotevant & Cooper, 1986), it logically follows that disturbances of the separation-individuation process may manifest either as problems of coping with the developmental task of achieving independence or as problems of coping with the task of staying connected to others. When people fail to obtain a healthy degree of separation/independence from others, they may develop a dysfunctional dependent orientation, where they need constant physical and emotional proximity to maintain their well-being. In contrast, when people fail to remain connected to other people, they may develop a dysfunctional independent orientation, where they constantly strive for self-reliance and consider any type of intervention by others as an intrusion.

Although it seems plausible to define at least two broad categories of separation-individuation disturbances, diagnostic models and empirical research have tended either to focus exclusively on dysfunctional dependence or to rely on global and undifferentiated measures of pathological separation-individuation. The Diagnostic and Statistical Manual for Mental Disorders (DSM-IV-TR; American Psychiatric Association [APA], 2000), for instance, includes the separation anxiety disorder as a mental disorder that is usually first diagnosed in infancy, early childhood, or adolescence. Symptoms of this disorder involve distress when separated from attachment figures, persistent worrying about loss, and extreme fear of being alone. Similarly, research has typically focused on the dependent type of separation-individuation pathology. Wood (2006), for instance, found that children’s separation anxiety was predicted by parental intrusiveness. Other studies have used a general and undifferentiated measure, not allowing distinction between possible subtypes of problematic separation-individuation. One often-used general measure of pathological separation-individuation is the pathological separation (PATHSEP) measure developed by Christenson and Wilson (1985). Although this scale was developed to capture a variety of disturbances in the separation-individuation process, studies have shown that all items coalesce around a single factor (Christenson & Wilson, 1985; Kins, Soenens, & Beyers, 2011; Lapsley, Aalsma, & Varshney, 2001).

The Separation-Individuation Test of Adolescence (SITA; Levine, Green, & Millon, 1986) seems to be the most differentiated measure of (un)healthy separation-individuation currently available. It has been used to examine associations between unhealthy separation-individuation and diverse developmental outcomes (e.g., anxiety; e.g., Holmbeck & Leake, 1999; McClanahan & Holmbeck, 1992). Some of the SITA scales seem to tap rather specifically into dysfunctional dependence. The nurturance–symbiosis subscale, for instance, describes individuals who have enmeshed interpersonal relationships in which they wish for a state of oneness with the other. Similarly, the separation anxiety scale is an assessment of strong fears of losing emotional and physical contact with the other following the realization of one’s increasing separateness from others. In contrast, other scales from the SITA seem to tap into dysfunctional independence. The engulfment anxiety scale, for instance, taps into fear of close interpersonal relationships and a view of closeness as a threat to one’s independence. Similarly, the dependency denial subscale describes individuals who deny or avoid their need for connectedness as a defensive way of dealing with increasing demands for separation.

For the purpose of this study, we selected two of the SITA scales that, in our view, represent the most direct and valid indicators of dysfunctional dependence and dysfunctional independence; that is, separation anxiety and dependency denial, respectively. First, inspection of the SITA items showed that, in terms of face validity, the items of the separation anxiety scale and the dependency denial scale best represent the concepts of dysfunctional dependence and dysfunctional independence, respectively. Second, previous research using the SITA (McClanahan & Holmbeck, 1992) demonstrated that, of the scales most closely linked to the concept of dysfunctional dependence and independence, separation anxiety and dependency denial demonstrated the highest predictive validity, with both scales showing the strongest and most robust associations with measures of ill-being (e.g., depression, anxiety, and loneliness).

Nomological network of associations with relevant constructs

A first aim of this study was to substantiate the distinction between dysfunctional dependence and dysfunctional independence by developing a nomological network of associations with relevant constructs. To do so, we examined how both expressions of problematic separation-individuation relate to undifferentiated and one-sided measures of pathological separation-individuation and to a measure of depression. We also considered associations with relevant constructs from attachment theory and from Blatt’s theory on personality vulnerability to depression (Blatt, 1974, 2004).

Other measures of dysfunctional separation-individuation and depression

Because both dysfunctional dependence and independence represent disturbances in the separation-individuation process, it was expected that both would relate positively to a general and undifferentiated measure of separation-individuation pathology (i.e., PATHSEP). Further, if our claim that problematic separation-individuation, as defined by the DSM-IV, is characterized by a one-sided focus on fears of being alone (i.e., dysfunctional dependence) is true, a measure of the DSM-based diagnostic criteria of the separation anxiety disorder should be mainly or even uniquely related to our measure of dysfunctional dependence (i.e., separation anxiety) and to a lesser extent or not to our measure of dysfunctional independence (i.e., dependency denial). Finally, given that problematic separation-individuation is considered a risk factor for maladjustment (Christenson & Wilson, 1985; Lapsley & Edgerton, 2002; Lapsley et al., 2001), it was examined whether the two manifestations of dysfunctional separation-individuation would relate independently to emerging adults’ ill-being and to depressive symptoms in particular.

Attachment theory

During the processes of separation and individuation the child/adolescent needs to achieve a sense of intrapsychic separateness in order to become an individuated person. This newly gained independence might at least temporarily be experienced as a threat to the relationship with the attachment figure and may as such elicit feelings of loss (Blos, 1979; Levy-Warren, 1999; Mahler, 1963; Mahler et al., 1975). In attachment theory (Bowlby, 1969), it is argued that the way people react to such (threat of) loss of close relationships depends on the representational models of attachment figures and the self that were developed early in life. Those with a secure internal working model are likely to respond adequately to events of separation. In contrast, insecurely attached individuals may feel more strongly threatened by the developmental demands of separation-individuation and may therefore develop derivative ways of coping with this process. Interestingly, in attachment theory a distinction is made between two instantiations of insecure attachment, each of which may relate differentially to the two types of dysfunctional separation-individuation (Hazan & Shaver, 1987). That is, individuals high on attachment anxiety are insecure about availability of attachment figures. They are preoccupied with social support and highly vigilant about abandonment and rejection. These individuals are particularly likely to experience the separation-individuation process as a threat to closeness in their relationships and may as such fail to achieve a healthy degree of separation. In contrast, individuals high on avoidance have a strong preference for emotional distance and feel uncomfortable with closeness or dependence on others. These individuals are likely to have difficulties balancing their strong urge for independence with the establishment of close relationships. Instead, they may display excessive strivings for self-reliance and independence, as expressed for instance in denial of dependence.

Testifying to the validity of the distinction between attachment anxiety and avoidance, research has shown convincingly that, whereas both attachment orientations relate independently to ill-being (e.g., depressive symptoms), they relate differentially to features of personality, interpersonal functioning, and emotion-regulation (for an overview, see Mikulincer & Shaver, 2002). It has been shown, for instance, that attachment anxiety is primarily related to hyperactivation of negative emotions and that attachment avoidance is primarily related to deactivation or suppression of negative emotions (Mikulincer & Shaver, 2002). In the present study it was expected that dysfunctional dependence would be most strongly associated with attachment-related anxiety, whereas dysfunctional independence would be most strongly associated with attachment-related avoidance.

Blatt’s theory of depressive personality

Similar to models of separation-individuation, Blatt’s (1974, 2004) theory on personality development and vulnerability to depression views personality development as an ongoing dialectical interaction between two developmental lines; that is, interpersonal relatedness and self-definition (Blatt, 2004; Guisinger & Blatt, 1994). Ideally, relatedness and individuality develop throughout the life cycle in an interrelated manner, with achievements in one line contributing to and being contingent upon achievements in the other line. Vulnerability to psychopathology and to depression in particular would ensue when people develop an excessive focus on one of these two developmental lines, at the expense of the other developmental line. Thus Blatt (1974, 2004) distinguishes between two major personality vulnerabilities; that is, dependency and self-criticism. A dependent personality orientation is characterized by excessive attempts to establish and maintain satisfying interpersonal relationships, resulting in a neglect of the development of a sense of self. Dependency would create a vulnerability to a type of depression characterized by feelings of loss, abandonment, and loneliness. Given their problems of coping with increasing demands for separation and independence, dependent individuals are likely to display dysfunctional dependence. Self-critical individuals, in contrast, are highly preoccupied with self-definition and personal achievement, at the expense of close relationships (Blatt & Maroudas, 1992). Self-criticism would render individuals vulnerable to a type of depression characterized by feelings of inferiority, worthlessness, and guilt (Blatt, 1974). Given their preoccupation with personal achievement at the expense of relatedness, self-critical individuals are likely to display dysfunctional independence.

Research already addressed associations between Blatt’s configurations of psychopathology and the dimensions of insecure attachment style (e.g., Zuroff & Fitzpatrick, 1995). Findings suggest that, whereas dependency is primarily related to attachment anxiety, self-criticism is primarily related to attachment avoidance. It remains to be examined, however, whether Blatt’s personality dimensions relate differentially to the two types of problematic separation-individuation. We expected that a dependent personality orientation would be predominantly related to dysfunctional dependence and that a self-critical personality orientation would be predominantly related to dysfunctional independence.

A person centered-approach to derive profiles of problematic separation-individuation

A second aim of this study was to ascertain how many separation-individuation profiles could be distinguished based on individuals’ scores on dysfunctional dependence and dysfunctional independence. These profiles were then further examined by relating them to the same nomological network of variables discussed in the preceding section.

To date, only a few studies have used a person-centered approach to classify individuals into different groups on the basis of separation-individuation measures (Kruse & Walper, 2008; McClanahan & Holmbeck, 1992). Using cluster analysis, we aimed to identify different subtypes of problematic separation-individuation. Given that dysfunctional dependence and dysfunctional independence were expected to represent two qualitatively different ways of disturbances in the separation-individuation process, we hypothesized that at least three different separation-individuation profiles could be identified: (1) a group that has resolved the separation-individuation process in a healthy manner and thus shows neither signs of dysfunctional dependence nor signs of dysfunctional independence; (2) a subgroup that predominantly demonstrates separation-individuation problems of the dependent type; and (3) a group exclusively presenting disturbances of the independent type. In addition, we also explored the possibility that some individuals might display the two types of dysfunctional separation-individuation simultaneously (i.e., high on both dimensions). Extant person-centered research on separation-individuation has yielded rather mixed evidence for this idea. That is, even though some evidence was obtained for a group that successfully resolved the separation-individuation process and a group showing symptoms of both dysfunctional dependence and dysfunctional independence, McClanahan and Holmbeck (1992) could not distinguish a group that demonstrates a pure dysfunctional independent type of separation-individuation pathology, whereas Kruse and Walper (2008) could not find evidence for a pure dysfunctional dependent type of problematic separation-individuation. Thus additional person-centered research on separation-individuation that adds to these inconsistent results is warranted. To obtain a purer cluster solution, the current study only included the two SITA scales that are considered the most valid indicators of dysfunctional dependence and dysfunctional independence.

Furthermore, we also aimed to examine associations between the clusters and the nomological network of variables used to differentiate between dysfunctional dependence and dysfunctional independence: that is, general separation-individuation pathology, DSM separation anxiety disorder, depressive symptoms, attachment and personality orientation. We anticipated that individuals in both the cluster high on dysfunctional dependence and the cluster high on dysfunctional independence would display more maladjustment (i.e., elevated scores on depressive symptoms and on an undifferentiated measure of pathological separation-individuation) compared to individuals in the healthy separation-individuation cluster. Individuals in the cluster high on dysfunctional dependence were expected to display particularly elevated scores on a DSM-based measure of separation anxiety, attachment anxiety, and dependency. Individuals in the cluster high on dysfunctional independence were expected to display particularly elevated scores on attachment avoidance and self-criticism. Conversely, subjects in the healthy separation-individuation cluster were expected to display a secure attachment profile (i.e., low anxiety and avoidance) as well as low scores on Blatt’s personality dimensions.

Method

Participants and procedure

Participants were 232 Belgian emerging adults aged between 21 and 26 years. Subjects were recruited through undergraduate psychology students as part of a course on developmental psychology. Each student was asked to contact two emerging adults who had to meet certain criteria with respect to age and gender. That is, the first emerging adult had to be between 21 and 23 years old and the second emerging adult had to be between 24 and 26 years old. Moreover, one emerging adult needed to be female, the other male. As a result, we obtained a sample with an equal number of men and women, covering a broad age segment of the emerging adult period. The mean age of the total sample of emerging adults was 23 years and 7 months (SD = 1 year, 9 months). Most of them were highly educated (i.e., 75% followed post-secondary education), and came from intact families (i.e., 79%). At the time of data gathering, 43% of the participants lived permanently in the parental household, whereas 31% lived fully independently. The other 26% lived in a semi-independent living situation, meaning that they lived away from the parents but returned to the parental home on a regular basis (Goldscheider & DaVanzo, 1986; Kins, Beyers, Soenens, & Vansteenkiste, 2009). More than half of the emerging adults (i.e., 62%) reported to be involved in a romantic relationship.

Emerging adults who agreed to take part in this study received questionnaires that were completed during a home visit. Participation was completely voluntary and anonymity was guaranteed. The number of missing values in this data set was small (i.e., 2%), and according to Little’s test these data were missing completely at random (MCAR), χ² (217, N = 232) = 183.60, non-significant [n.s.]. Therefore the expectation maximization (EM) algorithm was used to obtain maximum likelihood estimates for the missing values (Schafer, 1997). Hence all further analyses relied on a sample of 232 people.

Measures

Two types of disturbed separation-individuation

To measure dysfunctional dependence and dysfunctional independence, we used the separation anxiety scale and the dependency denial scale from the Separation-Individuation Test of Adolescence (SITA; Levine et al., 1986). Items are scored on five-point Likert scales ranging from 1 (does not apply at all) to 5 (totally applies). The separation anxiety subscale comprises eight items referring to a strong fear of abandonment and loss of important others. A sample item reads: “I worry about being disapproved of by others.” This subscale showed good internal consistency (Cronbach’s α = .78). The 13 items of the dependency denial subscale reflect dysfunctional strivings for independence as a kind of defensive style against feelings of anxiety associated with separation (e.g., “I don’t really need anyone”). Cronbach’s alpha was .78 in this study.

To examine whether dysfunctional dependence and independence represent two different expressions of problematic separation-individuation, we tested the factor structure of the items of the separation anxiety and dependency denial subscales of the SITA with confirmatory factor analysis (CFA) using Lisrel 8.71 with maximum likelihood estimation (Jöreskög & Sörbom, 1996). Chi-square difference testing revealed that a two-factor model represented the SITA items better, χ² (188) = 413.99; root mean square error of approximation (RMSEA) = .07; standardized root mean square residual (SRMSR) = .09; comparative fit index (CFI) = .88, than a single-factor model, χ² (189) = 994.02; RMSEA = .14; SRMSR = .13; CFI = .58: Δχ² = 230.11; d.f. = 1; p < .001. The fit of the two-factor solution was less than optimal in terms of CFI, but improved significantly when allowing an error correlation between two items of the separation anxiety subscale referring to “death” (Δχ² = 25.28; d.f. = 1; p < .001). Further, because one of the items of the dependency denial subscale had a substantial cross-loading on the dysfunctional dependence-factor, we decided to remove this item (“Often I don’t understand what people want out of a close relationship with me”) in all further analyses. In line with the criteria for model evaluation (Bentler, 1990; Hu & Bentler, 1999), this final two-factor model revealed good fit: χ² (168) = 288.78; RMSEA = .06; SRMSR = .08; CFI = .93.

Alternative measures of separation-individuation

Participants were administered two well-validated and frequently used measures of disturbances in the process of separation-individuation. First, emerging adults completed a general measure of pathological separation-individuation (PATHSEP; Christenson & Wilson, 1985; Lapsley et al., 2001). Scale construction of the PATHSEP was based on Pine’s (1979) clinical observations of disturbances in the separation-individuation process. Although the 39 items of this questionnaire refer to various expressions of pathological separation-individuation (e.g., difficulty in differentiating from others, splitting, relationship disturbances), factor analytical examination has repeatedly shown that the common variance of the items is represented by a single factor (Christenson & Wilson, 1985; Kins et al., 2011; Lapsley et al., 2001). As such, an overall scale score is computed, with higher scores being indicative of greater separation-individuation pathology. Reliability and validity of this scale were demonstrated in previous studies (Christenson & Wilson, 1985; Lapsley & Edgerton, 2002). Cronbach’s alpha was .89 in our sample.

Second, participants filled out the separation anxiety disorder subscale of the Revised Child Anxiety and Depression Scale (RCADS; Chorpita, Yim, Moffitt, Umemoto, & Francis, 2000). The seven items of this scale measure separation-individuation disturbances in terms of the DSM diagnostic criteria for separation anxiety disorder (DSM-IVT-R; APA, 2000). Respondents are asked to rate how often each of the items (e.g., “Fear of being alone at home”) applies to them on a scale ranging from 0 (never) to 3 (always). Support for reliability and validity of the full RCADS has been provided in previous studies (Chorpita, Moffitt, & Gray, 2005; Chorpita et al., 2000). Cronbach’s alpha was .67 in this study.

Depressive symptoms

Emerging adults were administered the 12-item version of the Center for Epidemiologic Studies Depression Scale (CES-D; Radloff, 1977). This scale assesses the respondent’s level of distress and depressive symptoms experienced during the past week. Items are scored on a four-point Likert scale ranging from 0 (never or seldom) to 3 (mostly or always) and cover somatic as well as emotional and cognitive symptoms of depression. Previous studies showed adequate reliability and validity of this version of the CES-D (Roberts & Sobhan, 1992). Cronbach’s alpha in this study was .82.

Attachment and personality

First, the Experiences in Close Relationships Revised (ECR-R; Fraley, Waller, & Brennan, 2000) was administered to capture attachment anxiety and avoidance in the relationship with a romantic partner, or—for those who did not have partner—in the relationship with a close friend (Bartholomew & Horowitz, 1991). The ECR-R is a 36-item self-report measure with half of the items referring to anxiety or fear of rejection and abandonment (e.g., “I often worry that the other won’t care about me as much as I care about him/her”), and the other half reflecting discomfort with closeness (e.g., “I tell the other just about everything”; reverse coded). Items are scored on a seven-point Likert scale ranging from 1 (totally disagree) to 7 (totally agree). Cronbach’s alpha was .91 for the anxiety scale and .92 for the avoidance scale.

Second, Blatt’s personality dimensions were measured with the 19-item version (Bagby, Parker, Joffe, & Buis, 1994) of the Depressive Experiences Questionnaire (DEQ; Blatt, D’Afflitti, & Quinlan, 1976). The dependent personality orientation was measured with 10 items (e.g., “After an argument, I feel very lonely”); the self-critical orientation with nine items (e.g., “There is a considerable difference between how I am now and how I would like to be”). All items were scored on a five-point Likert scale ranging from 1 (does not apply at all) to 5 (totally applies). This version of the DEQ has shown good reliability and validity (Bagby et al., 1994). Cronbach’s alpha was .70 for the dependency scale and .85 for the self-criticism scale.

Results

Descriptive statistics

Before examining the two main research aims of this study, we conducted a multivariate analysis of covariance (MANCOVA) to explore associations between the background variables and the study variables. All study variables were entered as dependent variables. Independent variables were gender, family structure (i.e., intact vs. non-intact), relationship status (i.e., having a partner or not), and residential status (i.e., with parents, semi-independent, independent). Age was entered as a covariate. Significant multivariate main effects emerged for gender, F(9, 217) = 3.70, p < .001, η² = .13, and relationship status, F(9, 217) = 9.44, p < .001, η² = .28. Follow-up univariate analyses are displayed in Table 1. First, gender differences were found for the two types of dysfunctional separation-individuation. That is, whereas emerging adult women had significantly higher scores on dysfunctional dependence, men scored significantly higher on dysfunctional independence. Further, women also scored higher on the symptom-specific measure of the DSM separation anxiety disorder in comparison with men. Second, emerging adults involved in a partner relationship reported significantly lower scores on a number of study variables than their single peers. Specifically, they scored lower with respect to dysfunctional independence and on a general measure of separation-individuation pathology. Emerging adults with a partner also reported less depressive symptoms and both lower attachment-related anxiety and avoidance than those who were not involved in a romantic relationship. Finally, emerging adults with a partner scored significantly lower on Blatt’s self-critical personality orientation. In conclusion, because both gender and relationship status appeared to be significantly related to some of the study variables, we decided to control for these background variables in all subsequent analyses.

Table 1.

Univariate follow-up analyses of background variables with a significant multivariate effect on the study variables

	Gender						Relationship status
	Male		Female				No partner		Partner
Variables	M	SD	M	SD	F(1, 225)	η²	M	SD	M	SD	F(1, 225)	η²
Dysfunctional dependence	2.65	.66	2.85	.66	4.73*	.02	2.82	.64	2.71	.68	1.08	.01
Dysfunctional independence	1.84	.45	1.64	.41	14.42***	.06	1.87	.45	1.66	.41	17.13***	.07
PATHSEP	2.23	.42	2.18	.42	.81	.00	2.28	.42	2.16	.41	4.29*	.02
DSM-SAD	.29	.27	.43	.38	9.47**	.04	.37	.31	.36	.35	.27	.00
CES-D	.62	.44	.64	.39	.12	.00	.74	.48	.56	.36	11.64**	.05
ECR-R – Anxiety	2.97	.97	2.85	.89	.28	.00	3.37	.91	2.63	.83	30.39***	.12
ECR-R – Avoidance	2.75	.87	2.59	.82	1.56	.01	3.19	.75	2.35	.74	64.03***	.22
DEQ – Dependency	2.96	.54	3.03	.53	1.40	.01	3.02	.55	2.98	.53	.02	.00
DEQ – Self-criticism	2.32	.64	2.30	.61	.21	.00	2.52	.66	2.18	.57	11.07**	.05

Notes. PATHSEP = Pathological Separation, DSM-SAD = DSM Separation Anxiety Disorder, CES-D = Centre for Epidemiologic Studies Depression Scale, ECR-R = Experiences in Close Relationships Revised, DEQ = Depressive Experiences Questionnaire. All variables were scored on five-point Likert scales, except for DSM-SAD and CES-D, which were scored on four-point Likert scales ranging from 0 (never or seldom) to 3 (mostly or always), and ECR-R, which was scored on a seven-point Likert scale.

* p < .05; ** p < .01; *** p < .001.

The nomological network of the types of separation-individuation disturbances

The first aim of this study was to examine how both types of dysfunctional separation-individuation would relate to a nomological network of relevant variables. First, we examined how dysfunctional dependence and dysfunctional independence were associated with a general measure of pathological separation (i.e., PATHSEP), with a DSM-based measure of separation anxiety disorder, and with a measure of depressive symptoms. Second, we examined correlations between measures of dysfunctional dependence and dysfunctional independence and measures of attachment (i.e., anxiety and avoidance) and personality vulnerability (i.e., dependency and self-criticism). Correlations are displayed in Table 2.

Table 2.

Nomological network of two types of dysfunctional separation-individuation (N = 232)

	Dysfunctional dependence	Dysfunctional independence	Comparing correlations
	r	r	Z
PATHSEP	.54***	.41***	1.79
DSM-SAD	.47***	.19**	3.48***
CES-D	.38***	.27***	1.34
Raw scores
ECR-R
Anxiety	.56***	.30***	3.47***
Avoidance	.18**	.45***	−3.32***
DEQ
Dependency	.65***	.08	7.69***
Self-criticism	.60***	.40***	2.84**
Residual scores
ECR-R
Anxiety	.52***	.10
Avoidance	−.12	.32***
DEQ
Dependency	.38***	−.18**
Self-criticism	.26***	.44***

** p < .01; *** p < .001.

Associations with other measures of disturbed separation-individuation and with depressive symptoms

Both dysfunctional dependence and independence were strongly related to PATHSEP; that is, the general and undifferentiated measure of separation-individuation pathology. Direct comparison of the size of both correlations (see right column in Table 2) showed that both types of dysfunctional separation-individuation were equally strongly related to PATHSEP. Similarly, both types of problematic separation-individuation were equally strongly and positively related to depressive symptoms. In contrast and in line with expectations, only dysfunctional dependence showed a strong association with a DSM-based measure of separation anxiety disorder. Although dysfunctional independence also showed a positive and significant association with the latter measure, the size of this association was significantly smaller and was reduced to non-significance when controlling for the small amount of variance shared between dysfunctional dependence and independence (partial r = .10, n.s.).

Associations with measures of attachment and personality orientation

Because we found substantial correlations among the two attachment dimensions as measured with the ECR-R (r = .48, p < .001) and among the two dimensions of depressive personality as measured with the DEQ (r = .58, p < .001), we calculated residual scores for each of these subscales, thereby controlling for the variance they share with their counterpart subscale from the same measure. For instance, a residual score for attachment anxiety was computed by regressing attachment anxiety on attachment avoidance and by saving the unstandardized residual score obtained in this regression analysis. This residual score reflects participants’ attachment anxiety net of attachment avoidance. Table 2 shows correlations with both the original scale scores and these residual scores.

Although dysfunctional dependence was related positively to both attachment anxiety and avoidance, the strongest correlation was with attachment anxiety. This difference in the size of association with anxiety and avoidance became even more pronounced after controlling for the variance shared between both attachment dimensions. Conversely, dysfunctional independence was related to both attachment anxiety and avoidance, yet showed its strongest correlation with avoidance. After the variance shared between both attachment dimensions was controlled for, dysfunctional independence was even uniquely related to attachment avoidance.

Further, dysfunctional dependence was related to both dependency and self-criticism. Although, after controlling for the variance between both dimensions of personality vulnerability, the strongest association was with dependency, dysfunctional dependence was still substantially related to self-criticism as well. Dysfunctional independence was only related to self-criticism. After controlling for the variance between the two dimensions of personality vulnerability, dysfunctional independence was even related negatively to dependency.

Profiles of problematic separation-individuation: A person-centered approach

To identify profiles of dysfunctional separation-individuation, we performed a cluster analysis with dysfunctional dependence and independence as clustering variables. We followed a two-step procedure (Gore, 2000), including a hierarchical clustering using Ward’s method followed by a k-means cluster analysis. Initial cluster centers, derived from the hierarchical cluster solution in the first step, were used as non-random starting points in a subsequent k-means clustering procedure (Gore, 2000).

Prior to the cluster analysis, scores for dysfunctional dependence and independence were standardized and data were inspected for outliers. Univariate outliers were identified as subjects scoring higher than three standard deviations above or below the mean (i.e., absolute z-score > 3). Multivariate outliers were identified using Mahalanobis distance measure. Only one univariate outlier was removed, resulting in a sample of 231 participants. In the first step of the clustering procedure, hierarchical clustering (Ward’s method) was used to determine the number of clusters. On the basis of the scree plot of the combined cluster distances, and taking into account the stepsize criterion (Milligan & Cooper, 1985), we decided to consider three- to five-cluster solutions. In a second step, cluster centers derived from Ward’s method were used as non-random starting points in an iterative k-means cluster analysis. This step was applied to the three-, four-, and five-cluster solution. In the end, the four-cluster solution was retained based on parsimony, theoretical interpretability, and explanatory power. The final four-cluster solution, explaining 64% and 73% of the variance in dysfunctional dependence and dysfunctional independence respectively, is depicted in Figure 1. To interpret the distinguished clusters, z-scores on both expressions of problematic separation-individuation (presented on the y-axis) were inspected within each of the clusters. These z-scores, which indicate the distances between the cluster means and the total sample standardized mean in terms of standard deviation units, can be interpreted as effect sizes. Analogous to Cohen’s (1988) d, 0.2 SD is a small effect, 0.5 SD is a medium or moderate effect, and 0.8 SD is a large effect.

Figure 1.

Z-scores for the dependent and independent type of dysfunctional separation-individuation in the four-cluster solution.

Cluster 1 (n = 85, 37%) included participants who scored low on both dysfunctional dependence (z = −.90) and dysfunctional independence (z = −.55). Participants in this cluster will be denoted as healthy, because these participants seem to show no signs of either type of separation-individuation disturbances. Cluster 2 (n = 58, 25%) comprised participants who scored high on dysfunctional dependence (z = .72) but low on dysfunctional independence (z = −.73). Therefore participants in this cluster will be referred to as dysfunctional dependent. Cluster 3 (n = 41, 18%) consisted of participants scoring relatively low on dysfunctional dependence (z = −.27), but very high on dysfunctional independence (z = 1.55). As a result, participants in this cluster will be referred to as dysfunctional independent. Finally, cluster 4 (n = 47, 20%) comprised participants with elevated scores on both dysfunctional dependence (z = .91) and dysfunctional independence (z = .54). Because these participants are anxious about being alone but at the same time avoid intimacy and close relationships, they will be referred to as combined dysfunctional dependent and independent or simply combined.

The stability of this four-cluster solution was examined with a double-split cross-validation procedure (Breckenridge, 2000). For this purpose, we divided our sample randomly into two halves (subsample A and B). The two-step clustering procedure (i.e., Ward, followed by k-means) was applied within each of the subsamples. Next, participants of each subsample were assigned to new clusters on the basis of their Euclidean distance to the final cluster centers of the other subsample. These new clusters were compared with the original cluster-solution by means of Cohen’s kappa (κ). The two resulting kappas, one from each of the subsamples, were then averaged. An average kappa value of .60 is considered acceptable (Breckenridge, 2000). For the present study the average kappa across subsamples was .82, suggesting that the four-cluster solution is highly stable.

Next, to examine how the four distinguished separation-individuation profiles differ with respect to the nomological network of variables discussed in the first research aim, we conducted a MANCOVA with cluster membership as the independent variable and emerging adults’ scores on the alternative measures of problematic separation-individuation (i.e., PATHSEP and RCADS-SAD), depressive symptoms (i.e., CES-D), attachment (i.e., ECR-R), and personality vulnerability (i.e., DEQ) as dependent variables. For attachment and personality vulnerability, we used the residual scores discussed earlier. We controlled for the effects of gender and relationship status by entering them as covariates. Results revealed a significant multivariate effect of cluster membership, F(21, 629) = 8.86, p < .001, η ² = .22. The F values and effect sizes of the follow-up univariate analyses, along with the means and standard deviations of the nomological network of variables across each of the clusters, are displayed in Table 3. Each of the univariate ANOVAs was statistically significant and effect sizes ranged from .10 to .29. Post hoc Tukey comparisons indicated that the dysfunctional dependent, dysfunctional independent, and combined cluster did not differ in terms of overall separation-individuation pathology (i.e., PATHSEP), DSM separation anxiety disorder, and depressive symptomatology. In line with expectations, individuals in the healthy cluster scored significantly lower on all these variables. Moreover, the healthy subgroup also showed the lowest scores for attachment anxiety and low scores for attachment avoidance and Blatt’s personality orientations. In line with expectations, participants in the dysfunctional dependent cluster demonstrated elevated scores on attachment anxiety and dependency, whereas the dysfunctional independent cluster showed high scores on attachment avoidance and self-criticism. Although the overall mean score for attachment anxiety was low in the dysfunctional independent cluster, it was still significantly higher in comparison to the healthy cluster. Finally, participants in the combined cluster showed the highest scores for attachment anxiety along with high scores on both dependency and self-criticism. Contrary to expectations, the overall mean score for attachment avoidance was low in the combined cluster.

Table 3.

Means and standard deviations of the nomological network of variables by cluster

	Cluster 1:		Cluster 2:		Cluster 3:		Cluster 4:
	Healthy		Dysfunctional dependent		Dysfunctional independent		Combined
Variables	M	SD	M	SD	M	SD	M	SD	F(3, 225)	η²
PATHSEP	1.91_b	.31	2.27_a	.36	2.43_a	.36	2.43_a	.39	30.67***	.29
DSM-SAD	.20_b	.23	.47_a	.34	.37_a	.30	.49_a	.38	13.31***	.15
CES-D	.42_b	.31	.68_a	.35	.75_a	.47	.84_a	.45	12.99***	.15
ECR-R
Anxiety	−.47_c	.61	.28_ab	.68	.01_b	.86	.46_a	.79	20.53***	.22
Avoidance	−.07_b	.59	−.27_b	.67	.52_a	.78	.04_b	.80	8.55***	.10
DEQ
Dependency	−.11_b	.40	.23_a	.46	−.23_b	.34	.09_a	.39	12.81***	.17
Self-criticism	−.23_b	.46	−.10_b	.47	.33_a	.38	.24_a	.50	15.21***	.17

Notes. Means that do not share subscripts differ significantly (p < .05; Tukey contrasts). Items of PATHSEP were scored on a five-point Likert scale, whereas those of CES-D and DSM-SAD were scored on four-point Likert-scales ranging from 0 (never or seldom) to 3 (mostly or always). For both scales of the ECR-R and DEQ, residual scores were used, controlling for the shared variance with the other subscale of their respective measure.

*** p < .001.

Discussion

Two types of dysfunctional separation-individuation

Although separation-individuation is a well-documented developmental process, research on disturbances of separation-individuation is comparatively sparser. Given that successful resolution of the separation-individuation process is about finding a balance between individuality and connectedness, it seems logical to assume that problems can emerge when people develop an unhealthy preoccupation with either self-definition (i.e., dysfunctional independence) or with interpersonal relatedness (i.e., dysfunctional dependence; Grotevant & Cooper, 1986; Guisinger & Blatt, 1994). We obtained strong evidence for the validity of this distinction between two types of problematic separation-individuation. Dysfunctional dependence and independence were found to share some common characteristics, but also to relate to indices of personality and interpersonal functioning in a differentiated manner. Both dysfunctional dependence and independence were related substantially to a general measure of separation-individuation pathology, suggesting that they can both be interpreted as expressions of problematic separation-individuation. However, as expected, only dysfunctional dependence showed a strong resemblance with the DSM-based separation anxiety disorder. This finding illustrates the unilateral approach of disturbances in the separation-individuation process in the DSM, with diagnostic criteria of separation anxiety disorder referring only to excessive levels of anxiety over being separated from a person or place. Dysfunctional independence has not yet been recognized in the DSM as a qualitatively different expression of disturbances with separation-individuation. Because Western societies highly value qualities such as independence and individuality, this type of disturbed separation-individuation might be considered as less problematic than dysfunctional dependence in a Western context (Kağıtçıbaşı, 2005). However, the fact that both dysfunctional dependence and independence were positively associated with depressive symptoms indicates that practitioners should be alert for each of the two types of problematic separation-individuation as they both seem to have negative ramifications for personal well-being.

The distinction between dysfunctional dependence and independence was further supported by the finding that both types of dysfunctional separation-individuation were associated differentially with two major dimensions of attachment (i.e., anxiety and avoidance) and Blatt’s dimensions of personality vulnerability to psychopathology (i.e., dependency and self-criticism). That is, whereas dysfunctional dependence was uniquely related to feelings of worry of not being loved by others (i.e., attachment anxiety), dysfunctional independence showed the strongest association with feelings of discomfort with closeness and dependency on others (i.e., attachment avoidance). In line with expectations, dysfunctional dependence also showed the strongest association with Blatt’s dependent personality orientation, while dysfunctional independence was associated uniquely with self-criticism. Dysfunctional dependence showed a substantial relation with self-criticism as well, although this relationship was less pronounced than the association between dysfunctional independence and self-criticism. Possibly, at least in some people, self-criticism may be undergirded by fear of being disapproved and critiqued, and of losing the approval and acceptance of others (Bagby et al., 1994; Blatt, 1974).

Although not the primary focus of this study, some interesting associations between background variables and the two types of dysfunctional separation-individuation were obtained. Consistent with developmental theories on gender stereotypes claiming that socialization typically highlights the importance of independence for men and the importance of relatedness for women (Geuzaine, Debry, & Liessens, 2000; Gilligan, 1982), women appeared more likely to express disturbances with separation-individuation in a dysfunctional dependent manner, whereas men expressed these problems rather in a dysfunctional independent manner. Further, emerging adults not involved in a partner relationship reported more dysfunctional independence. Individuals high on dysfunctional independence are indeed likely to be wary of committing themselves to an enduring relationship because such a relationship might be considered as a threat to their excessively valued independence. In contrast, no association was found between relationship status and dysfunctional dependence. Because individuals high on dysfunctional dependence have an extreme need for unity with others, at least some of them may manage to build and maintain an enduring relationship. One may wonder, however, about the quality of the relationships of people high on dysfunctional dependence. An interesting avenue for future research is to explore associations between the two types of dysfunctional separation-individuation and relationship quality and interpersonal problems.

Profiles of problematic separation-individuation

Given that dysfunctional dependence and dysfunctional independence were found to be distinct expressions of problematic separation-individuation, our second aim was to examine how many profiles of separation-individuation could be identified on the basis of these two dimensions. We identified four interpretable clusters of separation-individuation. The first cluster, characterized by low scores on dysfunctional dependence and independence, was labeled healthy. As these individuals do not overly seek or dismiss close contact with others, they seem to have resolved the separation-individuation process in a healthy way, achieving a balance between closeness and distance (Allen et al., 1994; Grotevant & Cooper, 1986). In line with expectations, it was found that individuals in the healthy cluster displayed the least signs of psychopathology. Specifically, a profile of low dysfunctional dependence and independence corresponded with low overall separation-individuation pathology, low separation anxiety, and low depressive symptompatology. Emerging adults in the healthy cluster also showed a pattern of secure attachment, with low anxiety and low avoidance, and they did not match Blatt’s (1974, 2004) dependent or self-critical personality orientations.

The other three clusters had high scores on one or both of the dimensions of dysfunctional separation-individuation. The second cluster was characterized by exclusively elevated scores on dysfunctional dependence (i.e., the dysfunctional dependent), whereas the third cluster was characterized by high scores on dysfunctional independence (i.e., the dysfunctional independent). The final cluster included individuals with high scores on both dysfunctional dependence and independence and was labeled as combined. Contrary to expectations, the combined group did not demonstrate the most signs of ill-being, as individuals with a dysfunctional dependent, dysfunctional independent, and combined separation-individuation profile showed equally elevated symptoms of general separation-individuation pathology: the DSM separation anxiety disorder and depression. The non-clinical nature of our sample might account for these results, as mean scores on all variables related to ill-being were fairly low. Scores in a clinical sample would possibly be higher and more heterogeneous. In such a sample with more variation in the ill-being outcomes, between-cluster differences might be even more pronounced and the combined group in particular might display an even more maladaptive profile of outcomes compared to the other clusters.

Still, with respect to attachment and personality orientation, some meaningful differences did emerge between the three profiles of dysfunctional separation-individuation. For instance, emerging adults in the dysfunctional dependent cluster showed a pattern of high attachment-anxiety and a dependent personality, whereas those in the dysfunctional independent cluster showed high discomfort with closeness and dependency on others (i.e., attachment avoidance) in combination with a self-critical personality. Although dysfunctional independent emerging adults demonstrated low attachment anxiety, their scores were still significantly higher than those in the healthy cluster, which may suggest that dysfunctional independent expressions of problematic separation-individuation too are, to some extent, provoked by underlying feelings of anxiety when confronted with separation and loss (Årseth, Kroger, Martinussen, & Bakken, 2009; Levitz-Jones & Orlofsky, 1985). Finally, individuals in the combined cluster demonstrated among the highest levels of fear of abandonment and rejection (i.e., attachment anxiety) and they had elevated scores on both the dependent and self-critical personality dimensions. Past research has shown that people who combine high levels of both dimensions of personality vulnerability are most at risk for psychopathology, and for depression in particular (Blatt, Quinlan, Chevron, McDonald, & Zuroff, 1982). As such, this finding suggests at least indirectly that a combination of dysfunctional dependence and dysfunctional independence creates a stronger vulnerability to psychopathology compared to the presence of one of the types of dysfunctional separation-individuation alone. Contrary to expectations, emerging adults with a combined profile of dysfunctional separation-individuation did not show elevated scores on attachment avoidance. However, the fact that this cluster comprises individuals with high scores on dysfunctional dependence and only moderate scores on dysfunctional independence might possibly explain why avoidance of closeness is less pronounced in this group.

Limitations

Although this study revealed some interesting findings on the nature of dysfunctional separation-individuation, a number of limitations should be noted. First, findings of this study are based on a non-representative sample of emerging adults. Although our sampling procedure resulted in a group of participants covering a broad age range of the emerging adult period and with substantial variability in gender and residential status, we mainly sampled emerging adults from White, well-educated, and intact families. As such, it remains to be examined whether our results can be generalized to a broader population of emerging adults. In addition, it would be meaningful to examine whether our findings could be replicated in non-Western societies. Given that relatedness is valued over independence in these cultures (Kağıtçıbaşı, 2005), it could be the case that excessive dependence on others is not considered as problematic and is less strongly associated with maladjustment as compared to the West.

Second, we used a non-clinical sample of emerging adults. Consequently, the means on all study variables were rather low in the present study. However, as we mainly aimed to demonstrate that dysfunctional dependence and dysfunctional independence are related differentially to other relevant constructs, these low scores did not hinder us from providing evidence for the distinction between two types of problematic separation-individuation. Yet future research in clinical samples would allow us to explore: (1) whether this typology of dysfunctional separation-individuation can be distinguished in a clinical population; (2) how dysfunctional dependence and independence are manifested in clinical samples; and (3) whether these dimensions of problematic separation-individuation display comorbidity with particular mental disorders.

Third, findings are exclusively based on emerging adults’ self-reports. Due to the disadvantages associated with self-report methodology (e.g., shared method variance and reporting bias), the observed strength of relationships between variables may have been artificially elevated. Future studies can try to overcome these problems by including multiple informants (e.g., parents, partners, friends, therapists) to report on emerging adults’ interpersonal functioning and personality or by using more diverse methods (e.g., observation, interview) to measure the concepts in this study. Besides questionnaire data, future research could use observational methods to measure issues of separation-individuation. The autonomy and relatedness coding system, for instance, which measures promotion or inhibition of autonomy and relatedness during a family interaction task (Allen et al., 1994; Allen, Hauser, O’Connor, & Bell, 2002), might be an appropriate alternative for questionnaire data.

Fourth, we made a couple of small adjustments to the SITA scales used in this study (i.e., separation anxiety and dependency denial) that were mainly data-driven. Future research should confirm these adjustments. Furthermore, including other scales of the SITA might be instructive when studying manifestations of problematic separation-individuation and when comparing groups demonstrating different profiles of such problematic separation-individuation. Including the healthy separation-individuation scale might be particularly important to validate the label we used for the group of participants scoring low on both types of dysfunctional separation-individuation (i.e., the healthy group).

Finally, the cross-sectional study design does not allow us to examine the developmental trajectory leading to problematic separation-individuation. Longitudinal research is needed when we want to draw definite conclusions about possible antecedents and outcomes of disturbances in the separation-individuation process. Such a longitudinal design would also allow us to determine the direction of effects in the hypothesized relationships between our study variables. For example, should an insecure attachment style or a dimension of personality vulnerability be considered as a precursor, or rather as a consequence of a pathological separation-individuation process?

Conclusion

In this study, we argued and found that problematic separation-individuation may manifest in at least two qualitatively distinct ways; that is, as a tendency to excessively seek closeness to others (i.e., dysfunctional dependence) or as a tendency to be strongly preoccupied with individuality and to avoid any kind of connectedness (i.e., dysfunctional independence). For both clinical practice and future research it is important to become aware that, besides manifestations of dysfunctional dependence, dysfunctional independence can also point to disturbances in the process of separation-individuation. As both types of disturbed separation-individuation were found to be differentially related with personality features and aspects of interpersonal functioning, both might have their own specific developmental trajectory, requiring other diagnostic criteria and therapeutic interventions.

Footnotes

Funding

This research received no specific grant from any funding agency in the public, commercial, or not-for-profit sectors.

References

Allen

J. P.

Hauser

S. T.

Eickholt

Bell

K. L.

O’Connor

T. G.

(1994). Autonomy and relatedness in family interactions as predictors of expressions of negative adolescent affect. Journal of Research on Adolescence, 4, 535–552. Retrieved from http://onlinelibrary.wiley.com/

Allen

J. P.

Hauser

S. T.

O’Connor

T. G.

Bell

K. L.

(2002). Prediction of peer-rated adult hostility from autonomy struggles in adolescent–family interactions. Development and Psychopathology, 14, 123–137. doi: 10.1017/S0954579402001074.

American Psychiatric Association. (2000). Diagnostic and statistical manual of mental disorders (4th ed., text rev.). Washington, DC: Author.

Årseth

A. K.

Kroger

Martinussen

Bakken

(2009). Intimacy status, attachment, separation-individuation patterns, and identity status in female university students. Journal of Social and Personal Relationships, 26, 697–712. doi: 10.1177/0265407509347927

Bagby

R. M.

Parker

J. D. A.

Joffe

R. T.

Buis

(1994). Reconstruction and validation of the depressive experiences questionnaire. Assessment, 1, 59–68. doi: 10.1177/ 1073191194001001005

Bartholomew

Horowitz

L. M.

(1991). Attachment styles among young adults: A test of a four-category model. Journal of Personality and Social Psychology, 61, 226–244. doi: 10.1037/0022-3514.61.2.226

Bentler

P. M.

(1990). Comparative fit indexes in structural models. Psychological Bulletin, 107, 238–246. doi: 10.1037/0033-2909.107.2.238

Blatt

S. J.

(1974). Levels of object representation in anaclitic and introjective depression. Psychoanalytic Study of the Child, 24, 107–157.

Blatt

S. J.

(2004). Experiences of depression: Theoretical, clinical, and research perspectives. Washington, DC: American Psychological Association Press.

10.

Blatt

S. J.

D’Afflitti

J. P.

Quinlan

D. M.

(1976). Experiences of depression in normal young adults. Journal of Abnormal Psychology, 85, 383–389. doi: 10.1037/0021-843X.85.4.383

11.

Blatt

S. J.

Maroudas

(1992). Convergence of psychoanalytic and cognitive behavioral theories of depression. Psychoanalytic Psychology, 9, 157–190. doi: 10.1037/h0079351

12.

Blatt

S. J.

Quinlan

D. M.

Chevron

E. S.

McDonald

Zuroff

(1982). Dependency and self-criticism: Psychological dimensions of depression. Journal of Consulting and Clinical Psychology, 50, 113–124. doi: 10.1037/0022-006X.50.1.113

13.

Blos

(1967). The second individuation process of adolescence. Psychoanalytic Study of the Child, 22, 162–186.

14.

Blos

(1979). The adolescent passage. Madison, CT: International Universities Press.

15.

Bowlby

(1969) Attachment and loss. Volume 1: Attachment: New York, NY: Basic Books.

16.

Breckenridge

J. N.

(2000). Validating cluster analysis: Consistent replication and symmetry. Multivariate Behavioral Research, 35, 261–285. doi: 10.1207/S15327906MBR3502_5

17.

Chorpita

B. F.

Moffitt

C. E.

Gray

(2005). Psychometric properties of the revised child anxiety and depression scale in a clinical sample. Behaviour Research and Therapy, 43, 309–322. doi: 10.1016/j.brat.2004.02.004

18.

Chorpita

B. F.

Yim

Moffitt

Umemoto

L. A.

Francis

S. E.

(2000). Assessment of symptoms of DSM-IV anxiety and depression in children: A revisited child anxiety and depression scale. Behaviour Research and Therapy, 38, 835–855. doi: 10.1016/S0005-7967(99)00130-8

19.

Christenson

R. M.

Wilson

W. P.

(1985). Assessing pathology in the separation-individuation process by an inventory: A preliminary report. Journal of Nervous and Mental Disease, 173, 561–565. Retrieved from http://journals.lww.com/jonmd/

20.

Cohen

(1988). Statistical power analysis for the behavioral sciences (2nd ed.). Hillsdale, NJ: Erlbaum.

21.

Fraley

R. C.

Waller

N. G.

Brennan

K. A.

(2000). An item response theory analysis of self-report measures of adult attachment. Journal of Personality and Social Psychology, 78, 350–365. doi: 10.1037/0022-3514.78.2.350

22.

Geuzaine

Debry

Liesens

(2000). Separation from parents in late adolescence: The same for boys and girls? Journal of Youth and Adolescence, 29, 79–91. Retrieved from http://www.springerlink.com/content/wp7k1r8162321r38/

23.

Gilligan

(1982). In a different voice: Psychological theory and women’s development. Cambridge, MA: Harvard University Press.

24.

Goldscheider

DaVanzo

(1986). Semiautonomy and leaving home during early adulthood. Social Forces, 65, 187–201.

25.

Gore

(2000) Cluster analysis. In Tinsley

Brown

(Eds.), Handbook of applied multivariate statistics and mathematical modeling (pp. 297–321). San Diego, CA: Academic Press.

26.

Grotevant

H. D.

Cooper

C. R.

(1986). Individuation in family relationships: A perspective on individual differences in the development of identity and role-taking skill in adolescence. Human Development, 29, 82–100. doi: 10.1159/000273025

27.

Guisinger

Blatt

S. J.

(1994). Individuality and relatedness: Evolution of a fundamental dialectic. American Psychologist, 49, 104–111. doi: 10.1037/0003-066X.49.2.104

28.

Hazan

Shaver

(1987). Romantic love conceptualized as an attachment process. Journal of Personality and Social Psychology, 52, 511–524. doi: 10.1037/0022-3514.52.3.511

29.

Holmbeck

G. N.

Leake

(1999). Separation-individuation and psychological adjustment in late adolescence. Journal of Youth and Adolescence, 28, 563–581. doi: 10.1023/A:1021654626328

30.

Holmbeck

G. N.

Wandrei

M. L.

(1993). Individual and relational predictors of adjustment in first-year college students. Journal of Counseling Psychology, 40, 73–78. doi: 10.1037//0022-0167.40.1.73

31.

Bentler

P. M.

(1999). Cutoff criteria for fit indexes in covariance structure analysis: Conventional criteria versus new alternatives. Structural Equation Modeling, 6, 1–55. doi: 10.1080/10705519909540118

32.

Jöreskog

K. G.

Sörbom

(1996). LISREL 8: User’s reference guide. Lincolnwood, IL: Scientific Software International.

33.

Kağıtçıbaşı

(2005). Autonomy and relatedness in cultural context: Implications for self and family. Journal of Cross-Cultural Psychology, 36, 403–422. doi: 10.1177/0022022105275959

34.

Kins

Beyers

Soenens

Vansteenkiste

(2009). Patterns of home leaving and subjective well-being in emerging adulthood: The role of motivational processes and parental autonomy support. Developmental Psychology, 45, 1416–1429. doi: 10.1037/a0015580

35.

Kins

Soenens

Beyers

(2011). “Why do they have to grow up so fast?”: Parental separation anxiety and emerging adults’ pathology of separation-individuation. Journal of Clinical Psychology, 67, 647–664. doi: 10.1002/jclp.20786

36.

Kruse

Walper

(2008). Types of individuation in relation to parents: Predictors and outcomes. International Journal of Behavioral Development, 32, 390–400. doi: 10.1177/0165025408093657

37.

Lapsley

D. K.

Aalsma

M. C.

Varshney

N. M.

(2001). A factor analytic and psychometric examination of pathology of separation-individuation. Journal of Clinical Psychology, 57, 915–932. doi: 10.1002/jclp.1059

38.

Lapsley

D. K.

Edgerton

(2002). Separation-individuation, adult attachment style and college adjustment. Journal of Counseling and Development, 80, 484–493. Retrieved from http://www.counseling.org/

39.

Lapsley

D. K.

Rice

K. G.

Shadid

G. E.

(1989). Psychological separation and adjustment to college. Journal of Counseling Psychology, 36, 286–294. doi:10.1037/0022-0167.36.3.286

40.

Levine

J. B.

Green

C. J.

Millon

(1986). The separation-individuation test of adolescence. Journal of Personality Assessment, 50, 123–137. Retrieved from http://www.informaworld.com/

41.

Levitz-Jones

E. M.

Orlofsky

J. L.

(1985). Separation-individuation and intimacy capacity in college women. Journal of Personality and Social Psychology, 49, 156–169. doi: 10.1037/0022-3514.49.1.156

42.

Levy-Warren

M. H.

(1999). I am, you are, and so are we: A current perspective on adolescent separation-individuation theory. In Esman

A. H.

Flaherty

L. T.

Horowitz

H. A.

(Eds.), Adolescent psychiatry: Developmental and clinical studies (Vol. 24, pp. 3–24). Hillsdale, NJ: Analytic Press.

43.

Mahler

Pine

Bergman

(1975). The psychological birth of the human infant. New York, NY: Basic Books.

44.

Mahler

M. S.

(1963). Thoughts about development and individuation. Psychoanalytic Study of the Child, 18, 307–324.

45.

Mattanah

J. F.

Brand

B. L.

Hancock

G. R.

(2004). Parental attachment, separation-individuation and college student adjustment: A structural equation analysis of mediational effects. Journal of Counseling Psychology, 51, 213–226. doi: 10.1037/0022-0167.51.2.213

46.

McClanahan

Holmbeck

G. N.

(1992). Separation-individuation, family functioning and psychological adjustment in college students: A construct validity study of the separation-individuation test of adolescence. Journal of Personality Assessment, 59, 468–485. Retrieved from http://www.informaworld.com/

47.

Mikulincer

Shaver

P. R.

(2002). Attachment in adulthood: Structure, dynamics, and change. New York, NY: Guilford Press.

48.

Milligan

G. W.

Cooper

M. C.

(1985). An examination of procedures for determining the number of clusters in a data set. Psychometrika, 50, 159–179. Retrieved from http://www.psychometrika.org/journal/Psychometrika.html

49.

Pine

(1979). On the pathology of the separation-individuation process as manifested in later clinical work: An attempt at delineation. International Journal of Psychoanalysis, 40, 225–242.

50.

Radloff

L. S.

(1977). The CES-D scale: A self-report depression scale for research in the general population. Applied Psychological Measurement, 1, 385–401. doi: 10.1177/014662167700100306

51.

Roberts

R. E.

Sobhan

(1992). Symptoms of depression in adolescence: A comparison of Anglo, African, and Hispanic Americans. Journal of Youth and Adolescence, 21, 639–651. doi: 10.1007/BF01538736

52.

Schafer

J. L.

(1997). Analysis of incomplete multivariate data. Monographs on statistics and applied probability, 72. London, UK: Chapman & Hall.

53.

Tanner

J. L.

(2005). Recentering during emerging adulthood: A critical turning point in life span human development. In Arnett

J. J.

Tanner

J. L.

(Eds.), Emerging adults in America: Coming of age in the 21st century (pp. 21–55). Washington, DC: American Psychological Association.

54.

Wood

J. J.

(2006). Parental intrusiveness and children’s separation anxiety in a clinical sample. Child Psychiatry and Human Development, 37, 73–87. doi: 10.1007/s10578-006-0021-x

55.

Zuroff

D. C.

Fitzpatrick

D. K.

(1995). Depressive personality styles: Implications for adult attachment. Personality and Individual Differences, 18, 253–265. doi: 10.1016/0191-8869(94)00136-G