Printer Friendly

Adult attachment and developmental personality styles: an empirical study.

The current study was designed to test specific hypotheses associated with W. J. kyddon and A. Sherry's (2001) attachment theory model of developmental personality styles. More specifically, 4 adult attachment dimensions were correlated with 10 personality scales on the Millon Clinical Multiaxial Inventory-Ill (T. Millon, R. Davis, & C. Millon, 1997) with a sample of 273 men and women. Findings indicated that the adult attachment dimensions were able to predict 7 of the 10 personality styles.

Researchers have tried to pinpoint the antecedents to personality development for decades. Although there is some evidence of heritability for normal personality traits, there does not appear to be sufficient evidence to support the notion that personality disorders (PDs) are genetically transmitted (McGuffin & Thapar, 1992; Nigg & Goldsmith, 1994). This leaves much of the variance in PD behavior to be explained by psychosocial and environmental factors (Vaillant, 1987). Risk factor research has identified several environmental risk factors for personality dysfunction, including family dysfunction, early separation and loss, parental neglect, childhood abuse or trauma, parental psychopathology, and social disintegration (Paris, 1996, 1997). However, none of these risk factors can consistently account for a significant amount of personality dysfunction. It has been suggested that successful approaches to the treatment and understanding of PDs need to be built on a broad etiological framework in order to understand how such disorders are developed and maintained (Paris, 1998). Although risk factor research has identified developmental antecedents that can lead to significant impairment in one's interpersonal, cognitive, and affective ways of relating (American Psychiatric Association [APA], 2000), published studies on the etiology of PDs are limited.

Research by Watson and Sinha (1995) lends support to a dimensional (rather than categorical) view of personality, even when considering PD characteristics and traits. For example, using clinical and nonclinical samples, they found similarities in the dimensional structure of PDs across measures and samples, noting that PD traits appear to be present in a variety of intensities in both clinical and nonclinical samples. This suggests that PD traits are not unique to individuals with PD diagnoses but rather indicate support for a common dimensional structure between nonclinical and clinical samples (Dyce, O'Connor, Perkins, & Janzen, 1997; Livesley & Schroeder, 1990; Livesley, Schroeder, & Jackson, 1992).

A dimensional view of PDs suggests that dysfunction may lie on a continuum. As such, all behavior may be considered adaptive behavior, and "pathological behavior" is viewed as once having important utility but is currently no longer functional or productive (Ivey, 1991). Following from Ivey, and consistent with prior work (Lyddon & Sherry, 2001), this article addresses PDs as developmental personality styles (DPSs). The word styles emphasizes that such personality traits lie on a dimensional continuum replacing the categorical word disorder. The term developmental implies that such personality traits, although no longer functional for the individual, often develop adaptively in the context of the person's life experiences and relationships.

Attachrnent Theory

In order to organize a dimensional view of DPS etiology, a comprehensive theoretical framework is needed. Such a framework should account for and explain the dysfunctional areas of cognition, affect, interpersonal functioning, and impulse control that are essential for the diagnosis. Attachment theory is a metaperspective that holds promise for explaining these aspects of DPSs. Not only does it emphasize cognitive, effective, and interpersonal dynamics, it also addresses important links between one's past learning experiences and current functioning (Lopez, 1995; Lyddon, 1995). Attachment theory effectively accounts for the inevitable developmental logic associated with even the most disturbed behavior (Ivey, 1991). Attachment theorists propose that there are distinct qualities of the caregiver-child relationship that shape the development and personality of a child (Bowlby, 1969, 1973, 1980). Attachment theory's principal concern is with the role that enduring affectional bonds between child and caregiver play in shaping one's personality and life course (Lopez, 1995).

Central to attachment theory is the concept of working models (Bowlby, 1973). Working models consist of one's expectations about (a) the accessibility and responsiveness of one's caregiver (working model of other) as well as (b) one's own ability to elicit need-meeting responses from that caregiver (working model of self; Bowlby, 1973). According to Bowlby (1973), these working models initially develop from experiences with primary caregivers and eventually generalize to global expectations about others and the world. Working models serve at least three functions regarding the development of dysfunctional personality styles. First, they lay an important foundation for why cognitive schemas become fixed and inflexible to change in environments that consistently perpetuate a specific view of the world as unsafe, unforgiving, or punitive. Second, they provide a context for the interpersonal turmoil experienced by individuals who often have a distorted view of themselves in relation to other people. Finally, both of the aforementioned aspects of working models perpetuate the affective responses associated with one's worldview and how the person feels in relation to other people.

According to Bowlby (1973), early attachment experiences and their associated working models do not exclusively determine one's specific attachment attitudes and feelings. The confirmation of these early working models through exposure to similar attachment experiences during development is believed to contribute to their persistence. In addition, the operation of assimilative, "feedforward" mechanisms are thought to influence attachment continuity. According to constructivist theory, feedforward processes are inflexible, cognitive strategies that anticipate and "fit" new experiences into existing cognitive constructions (Lyddon, 1993; Lyddon & Sherry, 2001; Mahoney, 1991). The working models of securely attached individuals generally reflect a balance between feedforward (assimilative) and feedback (accommodative) processes. However, individuals who are insecurely attached tend to have working models that are dominated by feedforward processes. Thus, these individuals are less open to new information, because they tend to selectively attend to information that confirms existing beliefs and are likely to ignore disconfirming information.

Bartholomew's (1990) theory of adult attachment extends Bowlby's (1973) concept of cognitive working models and provides a useful framework from which to conceptualize DPSs that have become fixed and inflexible (Bartholomew & Horowitz, 1991; Griffin & Bartholomew, 1994b). Like Bowlby, Bartholomew proposed that adult working models of self and others carry with them either positive or negative valences. Working models of self are representations of the degree to which one experiences self-worth and believes she or he is worth positive and fair treatment. Working models of others are representations of the degree to which others are seen as either trustworthy and available or rejecting and unreliable (Bartholomew, 1990). Using this 2 x 2 dimensional framework (Positive/Negative Model of Self x Positive/Negative Model of Other), Bartholomew has labeled these four attachment styles as secure, preoccupied, fearful, and dismissing. Secure individuals have a healthy, flexible blend of positive and negative working models and possess a sense of selfworth and an expectation of others that they are generally trustworthy, accessible, and responsive. Those with a preoccupied attachment style have a sense of personal unworthiness but a positive evaluation of others, and, as a result, they tend to be externally oriented in their self-definitions. Those who are fearfully attached exhibit both a sense of personal unworthiness and an expectation that others will be rejecting and untrustworthy. Therefore, these individuals trust neither their own internal cognitions or feelings nor other people's intentions. Finally, those with a dismissing attachment style are characterized by a positive, worthy sense of self but a low evaluation or mistrust of others.

According to Bartholomew (1997), most individuals show variation in all four of the attachment prototypes. Therefore, attachment is viewed multidimensionally. It is believed that a categorical conceptualization of attachment overlooks the meaningful variation that occurs within an attachment dynamic. As a result, a dimensional approach to attachment, similar to a dimensional approach to PDs, may address individual differences more accurately by taking into account the considerable variations within any one of the best fitting attachment categories.

Combining attachment theory, dimensional approaches, and personality dysfunction, Lyddon and Sherry (2001) suggested that attachment theory is a viable way in which to conceptualize DPSs. As Bolwby (1973) stated,
   Environmental pressures are due largely to the fact that the
   family environment in which a child lives and grows tends
   to remain relatively unchanged. This means that whatever
   family pressures led to the development of a child to take the
   pathway he is now on are likely to persist and so to maintain
   development on that same pathway. (p. 368)

In other words, when children encounter specific attachment experiences consistently over time, with the same relative intensity, these working models persist into adulthood. In the case of DPSs that have become dysfunctional, the hypothesis is that these attachment experiences are so detrimental in quality and intensity over time that children in these families ultimately grow to anticipate their worlds in much the same way they did as children. However, as adults, the thoughts, feelings, and behaviors that were once adaptive and functional in the context of their childhood environments are no longer functional in their adult environments and relationships. In addition, these individuals are dominated by assimilative, feedforward mechanisms, making change and adaptation to these new environments difficult. The development of such mechanisms again is adaptive in the child's dysfunctional environment to anticipate and protect him- or herself from possible emotional and/or physical pain that often accompanies these family systems.

Integration and Hypotheses

On the basis of reviews of risk factor research and family/ environmental influences on PDs, Lyddon and Sherry (2001) have developed a conceptual framework from which DPS are understood in the context of Bartholomew's (1990) adult attachment theory. Lyddon and Sherry's (2001) conceptual model also serves as an outline for the hypotheses investigated in the current study. Each DPS is associated with either a main attachment dimension or a blending of attachment dimensions based on previous literature. The dimensional nature of attachment allows for a blending of attachment constructs in the conceptualization of DPS, offering a potentially more useful conceptualization of clients' particular working models of self and others. The aim of the current study was to empirically examine the 10 DPSs (based on the 10 PDs recognized in the Diagnostic and Statistical Manual of Mental Disorders [4th ed., text rev.; DSM-IV-TR; APA, 2000]) using Bartholomew's (1990) dimensional model of adult attachment. Generally, we investigated the question of whether adult attachment was related to DPSs. More specifically, it was hypothesized that preoccupied attachment would be related to dependent, obsessive-compulsive, and histrionic DPSs. A blending of preoccupied and fearful attachments would be related to avoidant DPS. Fearful attachment would be related to paranoid DPS. A blending of fearful and dismissing attachments would be related to antisocial, narcissistic, and schizotypal DPSs. Dismissing attachment would be related to schizoid DPS. Finally, it was hypothesized that borderline DPS would be related to a disorganized attachment schema in which all three insecure attachments (fearful, dismissing, preoccupied) would be represented.



Participants were 277 undergraduates. However, four protocols were deemed invalid, resulting in 273 undergraduate students in the final analysis. Of these, 68.5% (n = 187) were female and 31.5% (n = 86) were male. They were recruited from three, midsize public universities from various regions of the United States (111 or 40.7% from the southcentral region, 45 or 16.5% from the Southeast, and 117 or 42.9% from the Pacific Northwest [percentages do not equal 100% because of rounding]). The ages ranged from 18 to 56 years (M = 22.15, SD = 6.19). For race, 63.4% of the participants identified as Caucasian (n = 173), 23.1% identified as African American (n = 63), 5.1% identified as Asian or Asian American (n = 14), 3.3% identified as Hispanic or Hispanic American (n = 9), 1.1% identified as American Indian (n = 3), and 3.7% identified as other (n = 10); 1 participant did not report ethnicity. About 83% of the participants indicated that they were never married (n = 227), 7.7% were in their first marriage (n = 21), 2.6% were divorced (n = 7), 1.9% were cohabitating (n = 5), 1.9% were remarried (n = 5), and 1.9% identified their marital status as other (n = 5); 3 participants did not report a marital status. (Percentages do not equal 100% because of rounding.)


Relationship Scales Questionnaire (RSQ," Griffin & Bartholomew, 1994b). Attachment in adulthood was assessed using the RSQ. The RSQ is a short, 30-item instrument that draws from Hazan and Shaver's (1987) attachment measure, Bartholomew and Horowitz's (1991) Relationship Questionnaire, and Collins and Read's (1990) Adult Attachment Scale. On a 5-point scale (range of 1-5 for the current study), participants rate the extent to which each statement best describes their characteristic style in close relationships, and means are calculated for each of the four attachment dimensions. There are three anchors: not at all like me (1), somewhat like me (3), and very much like me (5). Questions included items similar to "I find it difficult to depend on other people" Of the 30 items in the RSQ, 18 are used for creation of scores for the secure (5 items), fearful (4 items), dismissing (5 items), and preoccupied (4 items) attachment dimension variables. Three items are reverse scored, and 1 item is used for scoring in reverse-coded form on two variables. Because the items were drawn from several attachment questionnaires, the other items could be used to create other attachment subscales identified by other researchers (Collins & Read, 1990; Simpson, Rholes, & Nelligan, 1992). However, these subscales were not the focus of the current study.

The RSQ was designed as a continuous measure of adult attachment, providing an attachment profile for each participant rather than a categorical assignment. The higher the score in any one dimension, the more that participant endorsed items measuring that construct. Griffin and Bartholomew (1994a) noted that these items reflect the "two dimensions of anxiety (or positivity of the self model) and avoidance (or positivity of the other model)" (p. 30). These two dimensions yielded coefficient alpha reliability estimates of .70 and .74, respectively, for the current data. Griffin and Bartholomew (1994a) argued for combining these two orthogonal dimensions to yield the continuous "prototype scores" for each of the four attachment styles. It should be noted, however, that the four variables are a function of combining the two dimensions based on theoretical expectations, as opposed to existing as four unique factors. Therefore, although there are two factors for the scale, two scores per factor are derived theoretically for the purposes of measuring the four dimensions of attachment. For example, the Secure subscale consists of items related to one's comfort with self and others, which draws from both of the anxiety and avoidance dimensions, respectively. As such, internal consistency reliability estimates are not appropriate for the subscale level. The reader is referred to Griffin and Bartholomew (1994a, 1994b) for a thorough discussion of the empirical and theoretical evidence for this manner of assessing adult attachment.

Although several scales exist to measure adult attachment, the RSQ was chosen for this study because it appears to most closely represent Bowlby's (1969) original theory of attachment. Other scales have used a three-dimensional model of adult attachment. However, some studies have suggested that this model is slightly limited. In a study by Collins and Read (1990), for example, individuals were given a 21-item questionnaire designed to objectively measure the three-dimensional model developed by Hazan and Shaver (1987). Using a cluster procedure to investigate whether Hazan and Shaver's model produced clusters consistent with the three-dimensional conceptualization, Collins and Read's results suggested that a four-cluster solution may be a more appropriate description. Bartholomew's (1990) four-dimensional model not only encompasses Bowlby's (1969) original 2 x 2 self/other theory of attachment, but it may also be a more inclusive theoretical model of adult attachment and one that may account for the four-cluster solution noted by Collins and Read.

Millon Clinical Multiaxial Inventory-III (MCMI-III; Millon, Davis, & Millon, 1997). The MCMI-III is a theoretically driven personality instrument. The instrument consists of 175 forced-choice questions. It is a self-administered, self-report measure of personality and clinical functioning. The following subscales were of interest in this study: Schizoid, Avoidant, Dependent, Histrionic, Narcissistic, Antisocial, Compulsive, Schizotypal, Borderline, and Paranoid (Millon et al., 1997). The subscales used in this study pertain to the 10 PDs recognized in the DSM-IV-TR (APA, 2000). The 10 DPSs are assessed with regard to the coping styles the participant uses. This is done by assessing aspects of the self and aspects of one's perceptions of others. The instrument also investigates important facets of the participant's behavior, interpersonal conduct, and cognitive style in terms of how the participant may gain comfort or avoid emotional pain (Millon et al., 1997). Higher scores indicate more items being endorsed on that construct.

All protocols were computer scored, and the three protocols identified as invalid were not included in analysis. A score of 75 on any one of the scales is considered clinically significant according to the manual (range = 1-115; Millon et al., 1997). Internal consistency coefficient alphas for the current study for the personality subscales are as follows: .74 (Schizoid), .82 (Avoidant), .74 (Dependent), .80 (Histrionic), .71 (Narcissistic), .72 (Antisocial), .67 (Compulsive), .76 (Schizotypal), .75 (Borderline), and .79 (Paranoid).

There are two important considerations regarding the use of the MCMI-III for this study. First is the use of nonclinical participants, and second is the dimensional application of results versus using clinical cutoff scores. Regarding the former issue, nonclinical samples have been found to be highly similar to the clinical samples reported by Millon (1994). Although mean scores may be higher in clinical samples than in nonclinical samples, the pattern of associations between scores on the MCMI-III PD subscales were similar for the two groups (Dyce et al., 1997). Regarding the latter issue, data gathered by the MCMI-III were used in a dimensional manner (i.e., in continuous form) and assessed PDs in terms of patterns of responding, not clinical significance. This approach is aligned with the theory of DPSs as well as a dimensional view of personality dysfunction while still being able to capitalize on the nature in which the MCMI-III is able to capture traits associated with each PD. The purpose of this study is to investigate traits or the dimensionality of PDs, rather than the clinical manifestation of them, making the use of a nonclinical sample ideal. In addition, this approach, as well as the statistical approach explained later, allows for all scores to be used in the analysis, regardless of how high or low the score.


Participation in this study was voluntary. Participants received either extra credit or a lottery ticket for $100 for their participation. They were solicited from psychology undergraduate courses and campus dormitories. Questionnaires were administered in group format, with each person completing his or her own packet of materials independently. The procedure took approximately 1 hour to 1 1/2 hours to complete.

Statistical Analyses

Canonical correlation analysis (CCA). After careful consideration of the research questions to be addressed, it was decided that CCA would be an appropriate statistical method with which to analyze the data. A CCA using the four Bartholomew and Horowitz (1991) attachment dimensions as predictors of each of the 10 MCMI-III personality styles was conducted. These analyses examined the differential predictive validity of attachment dimensions as noted in the hypotheses. CCA is a multivariate method that subsumes other parametric methods, such as regression and multivariate analysis of variance, in the general linear model (Thompson, 1991). It has been demonstrated that all other classical univariate and multivariate methods can be conducted with CCA (Henson, 2000; Knapp, 1978). There are general advantages to this method. First, a multivariate method best honors the reality of human behavior research. Speaking specifically about attachment and personality variables like those used in the current study, it assumes that these variables lie on a continuum, may have multiple outcomes, with multiple causes and multiple effects. Second, it allows for simultaneous comparisons between multiple predictors and multiple dependent variables (Sherry & Henson, 2005; Thompson, 1991). Again considering the variables of the current study, with four attachment variables as predictors and 10 personality variables as dependent variables, CCA is able to examine all of these together, not only in terms of what predictor and criterion variables are correlated but also in terms of shared correlations within each variable set. Given the current hypotheses in which more than one attachment variable may predict more than one personality variable, CCA is an appropriate method for analyzing these data.

Distributional normality. The data were analyzed in order to ensure that the variables met the assumption of multivariate normality. As with univariate analyses, this assumption is necessary for the appropriate use of statistical significance tests and also to estimate the appropriate effects (i.e., nonnormality may attenuate the observed effect). Using a graphical procedure described by Henson (1999), we evaluated the predictor and criterion variable sets used in the CCA for multivariate normality. The procedure focuses on the Mahalanobis distances of observations from the mean vector and plots these distances with expected chi-square values assuming normality. The resulting graph is analogous to a univariate Q-Q plot, and multivariate normality is observed when a straight line forms in the scatter plot. The lines were examined by the current researcher (first author) as well as by an independent rater (third author). There was interrater agreement that multivariate normality was tenable for the present data.


Means, standard deviations, and intercorrelations for the variables of interest are noted in Table 1. A CCA was conducted using the four attachment variables as predictors of the 10 personality styles in order to analyze the multivariate shared relationship between adult attachment and personality style. This analysis yielded four canonical functions, with only the first two functions yielding interpretable squared canonical correlation ([R.sup.2.sub.c]) effect sizes of 35.9% and 20.9%, respectively. The third and fourth functions explained only 5.4% and 1.8%, respectively, of the variance between the predictor and dependent variables. As a result, they were omitted from interpretation. Table 2 presents the results for the first two functions, including the standardized canonical function coefficients and structure coefficients.

Collectively, the full model across all functions was statistically significant using the Wilks's [LAMBDA] = .470 criterion, F(40, 968.79) = 5.32, p < .001. Because Wilks's [LAMBDA] represents the variance unexplained by the model, 1 - [LAMBDA] yields the full model effect size in an [r.sup.2] metric (Henson, 2006). Thus, for the set of four canonical functions, the [r.sup.2] type effect size was .530, which indicates that the full model can explain a substantial portion, approximately 53%, of the variance shared between the variable sets.

The dimension reduction analysis allows the researcher to test the hierarchical arrangement of functions for statistical significance. As noted, the full model (Functions 1-4) was statistically significant, Wilks's [LAMBDA] = .470, F(40, 968.79) = 5.32, p < .001, [R.sup.2.sub.c] = 53.0%. The test of only Functions 2-4 was also statistically significant, Wilks's [LAMBDA] = .734, F(27, 748.29) = 3.08, p < .001, [R.sup.2] = 26.6%. However, the tests of Functions 3-4 and 4 did not explain a statistically significant amount of shared variance between the variable sets, Wilks's [LAMBDA] = .928, F(16, 514.00) = 1.21, p = .256, [R.sup.2] = 7.2%, and Wilks's [LAMBDA] = .981, F(7, 258.00) = 0.68, p = .689, [R.sup.2.sub.c] = 1.8%, respectively.

Table 2 presents the standardized canonical function coefficients and structure coefficients for Functions 1 and 2. The squared structure coefficients are also given, as well as the communalities ([h.sup.2]) across the two functions for each variable. For the Function 1 coefficients, the relevant criterion variables are primarily Avoidant, Dependent, Histrionic, Schizotypal,

Borderline, and Paranoid. This conclusion is supported by the squared structure coefficients. These personality styles also tend to have the larger canonical function coefficients. Furthermore, with the exception of Histrionic, all of these variables' structure coefficients have the same sign, indicating that they are all positively related. Histrionic is inversely related to the other personality styles.

Regarding the predictor variable set in Function 1, Secure and Preoccupied attachment variables are the primary contributors to the predictor synthetic variable, with a secondary contribution by Fearful. Because the structure coefficient for Secure is positive, it is negatively related to all of the personality styles except for Histrionic. Preoccupied and Fearful attachments are positively related to the PDs, again except for Histrionic. These results are generally supportive of the theoretically expected relationships between adaptive and maladaptive adult attachment and PDs.

With respect to Function 2, the coefficients in Table 2 suggest that the only criterion variable of relevance is Schizoid. As for attachment, Dismissing is now the dominant predictor, along with Preoccupied again. These attachment variables are also inversely related. Noting the structure coefficients for the entire function, one can see that Dismissing is positively related to Schizoid. Preoccupied attachment has the opposite pattern.


The purpose of this study was to examine the relationship between adult attachment and dysfunctional personality styles. Adult attachment dimensions were assessed using the RSQ (Griffin & Bartholomew, 1994b), in which each participant received a score for each of the four attachment dimensions: secure, preoccupied, dismissing, and fearful. These four scores were used together to create an attachment profile for each participant. Personality styles were assessed using the MCMI-III (Millon et al., 1997). Individual scores were viewed on a continuum as personality styles rather than categorized into specific PDs using clinical cutoffs. The data were then analyzed in multivariate context to preserve this dimensional framework. The following discussion is organized by DPS.

Dependent Personality Style

It was hypothesized that people with dependent personality features would primarily exhibit a preoccupied working model of attachment due to a view of themselves as personally inadequate and a core belief that others are in their life to take care of them. However, findings suggest a role for both preoccupied and fearful attachment. It is interesting that fearfully attached persons exhibit this same sense of unworthiness coupled with an expectation that other people will be rejecting and untrustworthy. A blending of these two attachment dimensions is consistent with some of the diagnostic characteristics of this particular personality style. Persons with a dependent personality style are described as lacking self-confidence and assertiveness, which in turn is believed to be manifested in interpersonal behavior that is usually clinging, pleasing, compliant, and self-sacrificing. Their family histories are often characterized by caregivers who consistently convey to them the belief that they cannot be independent (Bornstein, 1992). Individuals with a dependent personality style take an ecologically passive stance toward life. They consistently lean on others to guide their lives, encourage others to take the initiative, and depend on others to nurture and protect them (Millon, 1996).

With so little personal control over their own lives, it is highly possible that others' attempts at guidance and nurturance may fail to meet the expectations of people with dependent personality styles. In addition, this lack of balance between getting needs met independently and getting needs met through others may eventually exhaust others around them, ultimately leading to avoidance or rejection by others toward the dependent individual. Eventually, a dichotomy of both positive and negative views of others may be created--positive in the sense that the dependent person desperately needs these others to guide and direct their lives, but negative in the sense that people will not always be able or even willing to follow through with the dependent person's expectations, leading to distrust and disappointment.

Avoidant Personality Style

It was hypothesized that the working models of persons with avoidant features would exhibit a blend of preoccupied and fearful attachment dimensions. This was supported by the current analyses. Persons with an avoidant personality style often initially avoid people and appear shy. They possess a desire to be liked and accepted by others coupled with the fear of abandonment and rejection. Developmentally, their family histories are characterized by one of two extremes--either an engulfing or avoidant parenting style (Sperry & Mosak, 1996). Although these individuals seek acceptance and nurturance from others, they are unwilling to get involved unless they are certain to be liked. In addition, they are often hypervigilant about others' evaluations of them and are more apt to distort potentially positive feedback from others to feedback that confirms their negative self-image (Millon, 1996). This further alienates these individuals from others, even though the desire to feel connected with others (positive view of others) still remains.

Paranoid Personality Style

It was hypothesized that the working models of people with paranoid personality features would be most consistent with the fearful attachment dimension. Although this hypothesis was supported, what was also found was that these individuals exhibited high levels of preoccupied attachment as well. One possible explanation for this finding may involve the paranoid person's relationship with others. Interpersonally, these individuals tend to be provocative in their interactions with others. They often bear grudges and are unforgiving to those with whom they have related in the past and are also likely to display a quarrelsome, fractious attitude toward recent social acquaintances. They often test others' loyalty and are preoccupied with the possible hidden motives of others (Millon, 1996). They resist external influence, are chronically tense, and often present with restricted affect, likely in an effort to control the outward expression of emotions that might convey internal emotional information to others. Family histories often include some form of active rejection and persecution by caregivers coupled with a critical and vigilant approach to parenting (Sperry & Mosak, 1996). It is possible that the combination between a rejecting and persecuting family history and a more instinctual human desire to connect with others could account for the association between preoccupied attachment and the paranoid personality style. There clearly remains an investment in others as noted by the extreme reaction these individuals have toward the perceived slights of others. It is possible that, on some level, these individuals still seek out the needs of nurturance and support that were never met during their development. Hence, underneath a myriad of defenses, a need, and thus positive view, of other people is present. However, their personality structure may have become so fixed that they are unable to allow themselves to believe in the positive motives of others and, as a result, default back to the safety of their negative view of others in their world.

Borderline Personality Style

Persons with a borderline personality style are reactive to their environment, which contributes to interpersonal consequences such as a series of intense, unstable relationships. This reactivity often presents emotionally as labile, resentful, impulsive, intense, helpless, and empty. Because of these characteristics, it was hypothesized that the working models of people with borderline personality features would exhibit dismissing, preoccupied, and fearful attachment. However, results indicated that these individuals primarily exhibited significantly high levels of preoccupied and fearful attachment only. Whereas other personality styles use a variety of defenses to guard against the perceived threats of others, people with borderline personality features appear to be much more invested in others, often displaying an intense separation anxiety. It is likely that this strong investment in others, similar to the dependent personality style, accounts for its strong relationship to the preoccupied attachment dimension. Regarding the fearful aspect of their working models of attachment, people with borderline personality features often have personal histories that reinforce a notion that they can never fully trust others nor hope to gain all of the affection and support they need (Millon, 1996). Family histories of these individuals are often marked with neglect or abuse (Laporte & Guttman, 1996; Sabo, 1997). Thus, the current data suggest these two attachment themes (overinvestment in others and history of interpersonal disappointment or abuse) may primarily account for the preoccupied/fearful attachment presentation of the borderline personality style.

Schizotypal Personality Style

Although it was hypothesized that the working models of individuals with schizotypal features would greatly display both fearful and dismissing dimensions of attachment, results instead suggested a primary role of fearful attachment. This finding is consistent with the fact that persons with schizotypal personality styles are not only socially detached from others but also estranged from themselves in the form of depersonalization or derealization. Individuals with schizotypal personality styles are socially isolative and experience intense social anxiety and apprehension. Their developmental histories are often marked with parents who possess a cold and derogatory parenting style (Sperry & Mosak, 1996). School and employment histories of these individuals show that they consistently fail to progress toward normal social attainments. Overall, they prefer isolation, drawing increasingly into themselves (Millon, 1996). Their ability to depersonalize likely speaks to this fearful attachment dimension in that in order to adapt to their own experience, at times, they need to cut themselves off from both their estranged self-image and other people because of the negative working models that permeate their world.

Schizoid Personality Style

The hypothesis that the working models of persons with schizoid personality features would be highly related to the dismissing attachment dimension was supported by the current data. This suggests that these individuals operate their world with a negative working model of others and a positive working model of themselves. Individuals with a schizoid personality style rarely respond to the feelings or actions of others and remain aloof and indifferent in interpersonal relationships. The developmental histories of these persons often consist of rigid, emotionally unresponsive caregivers and undersocialized interpersonal interactions (Thompson-Pope & Turkat, 1993). Possibly through a consistent process of not getting their emotional needs met by their caregivers, it is likely that these individuals have given up on the possibility of this ever happening and have decided instead to be aloof to the presence of others. Schizoid personality features were also negatively correlated with preoccupied attachment, again highlighting the lack of a positive view of other people and subsequent lack of investment in interpersonal relationships often exhibited by persons with this personality style.

Histrionic Personality Style

The finding that histrionic personality features were positively related to working models of secure attachment was unexpected. Millon (1996) suggested two prototypal variants of the histrionic personality style that he considered to be "normal" variants. These individuals are often described as dramatic, outgoing, friendly, and gregarious. They have confidence in their social abilities and can readily influence and charm others in a way that makes people like them. He argued that such behaviors in U.S. culture are rewarded and reinforced. It is highly possible that this normal variant of histrionic personality style is overrepresented in nonclinical samples.

Another possible explanation for this finding is that there is something unique about this particular personality style that needs to be investigated further. For example, it may be important to examine the various defense mechanisms used in the presentation of this style versus other styles. Overall, persons with this personality style initially appear charming and exciting. Their interpersonal behavior is marked by attention-getting, flirtatious, and even exhibitionistic behaviors. Caregivers are usually enmeshed and engulfing and usually only reinforce the person when the caregiver's needs or desires are being met. Ultimately, they view others favorably as long as they can get their attention and affection and the alliances they form with others are based on the idea that the histrionic person is the center of attention (Beck & Freeman, 1990). In addition, Leary (1957) characterized this type of personality as the cooperative-overconventional personality. They possess misperceptions of their social reality, are unable to see hostility or power in themselves, avoid feelings of depression, possess rigid overoptimism, and saturate their social exchanges with affiliative motifs. It is possible that the histrionic personality style is so dominated by defense mechanisms that promote this affiliative stance that on measures such as the RSQ, which primarily rely on self-report of one's internal affective experience, people with a histrionic personality features report as securely attached.

Results indicated that histrionic personality features were also correlated with the preoccupied attachment dimension, but to a much lesser extent. It is possible that the RSQ was only able to assess the hypothesized primary internal attachment working models of these individuals to a limited extent. Theoretically, the preoccupied attachment dimension is more consistent with the diagnostic presentation of this personality style, where the primary focus is on attracting attention from others in order to help compensate for a negative working model of the self. However, more likely is the use of the MCMI-III with a nonclinical sample. Histrionic personality as measured by the MCMI-III has been shown to be positively correlated with extroverted traits and may measure a healthy histrionic style rather than an unhealthy disorder, particularly when administered to nonclinical samples (Craig, 1993, 1997). This may explain its relation to secure attachment in the current study.

The Role of Secure Attachment

One consistent finding was that secure attachment dimension was negatively related to all of the seven personality styles noted except for histrionic. This is a theoretically consistent finding. The working models of securely attached individuals are purported to reflect a generally positive view of self and others. The secure attachment system is flexible and adaptive to most situations encountered by these individuals. As personality becomes more dysfunctional, it also becomes less flexible and not as open to new information. Therefore, a negative relationship between a flexible, secure system and a less flexible personality structure would be expected.

Personality Styles Not Accounted For

The three personality styles that could not be accounted for in this analysis were antisocial, narcissistic, and compulsive. One possible explanation is that these styles are not as developmentally anchored as the other seven discussed earlier. Although there is support in the developmental literature for developmental antecedents to these personality styles (Ingram, 1982; Ivey & Ivey, 1998; Sperry & Mosak, 1996), Watson and Sinha (1995) suggested that compulsive PD is qualitatively different from other PDs. For example, in their research, Watson and Sinha found compulsive PD to be less associated with the other nine PDs as measured by the MCMI-I (Millon, 1983) and the MCMI-II (Millon, 1987). This difference may extend to the attachment mechanisms that possibly account for the etiology of personality dysfunction. Another possibility is that the models of self and others associated with these three personality styles are not as well differentiated as those associated with other personality styles. Research on the MCMI-III PD subscales have indicated that elevations on the Compulsive and Narcissistic subscales in nonclinical samples are often positively correlated with measures of good mental health and negatively correlated with measures of emotional maladjustment (Craig, 1993, 1997). Still, a final possibility is that the attachment-personality relationship may only be measurable as the personality style becomes more dysfunctional, and the use of a nonclinical population was unlikely to include individuals with this level of dysfunction. This is likely particularly true for the antisocial personality style. Therefore, it is possible that compulsive, narcissistic, and antisocial personality styles were not well measured by the MCMI-III in this nonclinical sample. Certainly, more research is needed to investigate attachment relationships with these styles. It may be found that these particular styles develop independently of attachment experiences and may be better accounted for by either other risk factors in one's environment or biological determinants.

Implications of Findings

The findings from this study have at least three important implications for counseling. First, it is important to note that many of the personality styles accounted for in the analysis displayed a negative view of self. This finding might serve to guide counselors in the counseling process, noting that a negative core sense of self is fundamental to the presenting dysfunction. In addition to using cognitive behavioral strategies supported in the literature for this population (Bailey, 1998; Ball, Kearney, Wilhelm, Dewhurst-Savellis, & Barton, 2001; Cottraux & Blackburn, 2001; Freeman, 2002; Reisch, Thommen, Tschacher, & Hirsbrunner, 2001; Springer, Lohr,

Buchtel, & Silk, 1995; Swenson, Sanderson, Dulit, & Linehan, 2001), constructing a therapeutic alliance that facilitates the client's exploration into her or his ability to appropriately elicit need-meeting responses from the counselor may prove beneficial. With this approach, it is hoped that learning new interpersonal behaviors will generalize to the client's relationships outside of counseling and gradually begin to facilitate the development of a more positive sense of self. An example of an attachment theory approach to the treatment of borderline personality is outlined by Sherry (2007).

Second, the idea that working models of attachment can predict DPSs was supported, suggesting that attachment theory may be a viable theoretical base for the conceptualization of personality dysfunction. Although the DSM-IV-TR (APA, 2000) is a useful tool in clinical settings, its lack of a conceptual basis for understanding different diagnoses may be a contributor in the low reliability among professionals regarding the diagnoses they give their clients (Marshall & Serin, 1997). Although the DSM-1V-TR describes personality dysfunction as having developmental antecedents, and is thus coded on Axis II, it does not provide the practitioner with either a theoretical or practical guide related to the nature and possible etiology of such dysfunction. The current findings at least preliminarily indicate that attachment theory may provide this guide. The theory is able to encompass many of the fundamental aspects of personality dysfunction, such as cognitive, affective, and interpersonal dysfunction. Currently, the atheoretical stance of the DSM-IV-TR clusters personality diagnoses as being in either Cluster A, B, or C. However, if one considers how the personality styles cluster using the theoretical stance of attachment theory, this theoretical approach changes how these styles cluster and, ultimately, the understanding of how they may be related to each other. In turn, this gives the practitioner a guide from which to formulate hypotheses not only about diagnosis but also about etiology and treatment.

Third, the dimensional view of DPSs seems to offer a more complete understanding of personality that relies on both taxonomy and the client's qualitative experience rather than on the solely atheoretical DSM-IV-TR (APA, 2000) taxonomy. In addition, the current classification system for psychological disorders (the DSM-IV-TR) uses a categorical approach to diagnosis. From this perspective, either people carry the diagnosis or they do not, and dimensional differences between individuals are not noted. However, a diagnosis is only as helpful as it is reliable and valid regarding a client's treatment. The lack of dimensionality may actually decrease the reliability and validity of the diagnostic classification. In an effort to increase the utility of diagnostic descriptions, counselors are beginning to explore and support a more dimensional approach (Marshall & Serin, 1997; Widiger, 1992; Widiger & Frances, 1985). Such a dimensional approach would likely be particularly useful in the diagnosis of PDs.

These authors have argued that only a truly dichotomous variable would result in better reliability, which is impossible with issues of individual personality. Many aspects of personality are derived qualitatively, and a clear distinction between the presence or absence of a disorder is often not possible. In addition, because PDs are often understood as "pathological amplifications" of normal personality traits, a dimensional system would likely present a more realistic picture of dysfunction, including clients who are in need of services who may otherwise go undiagnosed (Paris, 1998). Finally, not all individuals with a PD diagnosis are alike with respect to their symptomatology (Widiger, 1992). For example, there are 149,495,616 diagnostic combinations of antisocial PD, and many of the differences may be extremely important regarding clinical practice and research (Widiger, 1992). In addition, a person who lacks only one of the diagnostic criteria may still experience clinically significant distress or impairment. However, these individuals would not warrant the diagnosis, failing to communicate clinically relevant information to other professionals (Widiger, 1992). As the results of this study illustrate, the use of a dimensional approach to both attachment and DPS also adds to the practitioner's ability to convey information to other professionals.

One of the limitations of this study is that all variables were assessed at the same point in time. Therefore, although relationships appear to exist, implications about how these relationships may be causally related cannot be addressed. Second, and related, the measure of attachment used is based on the participants' current working model of attachment rather than on their recollections of actual attachment experiences from caregivers. Although current attachment provides an important "window" into persons' working models of self and others, any strong causal link to early childhood attachment cannot necessarily be made. Third, self-report measures are vulnerable to reactivity, response bias, and response sets. Fourth, although it has been demonstrated that attachment is a universal human experience, there are some limitations associated with the exclusive use of undergraduate student samples, such as a lack of variability of experiences and restricted age range. Fifth, economic data were not collected in the current sample. Although the attachment research to date suggests that attachment is a universal construct, it is not known to what extent economic background may have influenced the relation between attachment and personality style. Finally, although finding adequate representation for each of the 10 PDs would have been difficult, the use of a clinical population would have potentially made results more applicable to inpatient or more severe clinical settings (i.e., community mental health settings that serve chronically ill or cyclically acute clients). Expansion of the current study to the general population as well as a clinical population is warranted.

Several researchers have examined risk factors for developing particular PDs and dysfunctional personality traits (Born stein, 1992; Brennan, Shaver, & Tobey, 1991; Frank & Paris, 1981; Laporte & Guttman, 1996; Zanarini & Frankenburg, 1997). A logical blend of research approaches would involve a study of the potential interaction between risk factors, attachment antecedents from childhood, and adult attachment working models. Although many people are exposed to a variety of documented risk factors to developing personality dysfunction, only a small percentage of individuals exposed to these factors actually develop such dysfunction. In addition, a little more than 50% of surveyed individuals report feeling insecure in their attachment experiences (Bartholomew & Horowitz, 1991; Duggan & Brennan, 1994). However, incidence data indicate that not all of these people progress to the point of developing a disorder. Therefore, future longitudinal research could address the interaction over time between risk factors, insecure attachment, and the development of dysfunctional personality styles. These possibilities notwithstanding, the current study does empirically support a noteworthy relationship between attachment and DPSs and may serve as a possible launching point into future DPS research.


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

Bailey, G. (1998) Cognitive-behavioral treatment of obsessive-compulsive personality disorder. Journal of Psychological Practice, 4, 51-59.

Ball, J., Kearney, B., Wilhelm, K., Dewhurst-Savellis, J., & Barton, B. (2001). Cognitive behaviour therapy and assertion training groups for patients with depression and comorbid personality disorders. Behavioural and Cognitive Psychotherapy, 28, 71-85.

Bartholomew, K. (1990). Avoidance of intimacy: An attachment perspective. Journal of Social and Personal Relationships, 7, 147-178.

Bartholomew, K. (1997). Adult attachment processes: Individual and couple perspectives. British Journal of Medical Psychology, 70, 249-263.

Bartholomew, K., & 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.

Beck, A. T., & Freeman, A. (1990). Cognitive therapy of personality disorders. New York: Guilford Press.

Bornstein, R. E (1992). The dependent personality: Developmental, social and clinical perspectives. Psychological Bulletin, 112, 3-23.

Bowlby, J. (1969). Attachment and loss: Vol. 1. Attachment. New York: Basic Books.

Bowlby, J. (1973). Attachment and loss: Vol. 2. Separation:Anxiety and anger. New York: Basic Books.

Bowlby, J. (1980). Attachment and loss: Vol. 3. Loss: Sadness and depression. New York: Basic Books.

Brennan, K. A., Shaver, P. R., & Tobey, A. E. (1991). Attachment styles, gender and parental problem drinking. Journal of Social and Personal Relationships, 8, 451-466.

Collins, N. L., & Read, S. J. (1990). Adult attachment, working models, and relationship quality in dating couples. Journal of Personality and Social Psychology, 58, 644-663.

Cottraux, J., & Blackburn, I. (2001). Cognitive therapy. In W J. Livesley (Ed.), Handbook of personality disorders: Theory, research, and treatment (pp. 377-399). New York: Guilford Press.

Craig, R. J. (Ed.). (1993). The Millon Clinical Multiaxial Inventory: A clinical research information synthesis. Hillsdale, NJ: Edbaum.

Craig, R. J. (1997). A selected review of the MCMI empirical literature. In T. Millon (Ed.), The Millon inventories: Clinical and personality assessment (pp. 303-326). New York: Guilford Press.

Duggan, E. S., & Brennan, K. A. (1994). Social avoidance and its relation to Bartholomew's adult attachment typology. Journal of Social and Personal Relationships, 11, 147-153.

Dyce, J. A., O'Connor, B. E, Parkins, S.Y., & Janzen, H. L. (1997). Correlational structure of the MCM-III Personality Disorder Scales and comparisons with other data sets. Journal of Personality Assessment, 69, 568-582.

Frank, H., & Paris, J. (1981). Recollections of family experience in bor-derline patients. Archives of General Psychiatry, 38, 1031-1034.

Freeman, A. (2002). Cognitive-behavioral therapy for severe personality disorders. In S. G. Hofmann & M. C. Thompson (Eds.), Treating chronic and severe mental disorders: A handbook of empirically supported interventions (pp. 382-402). New York: Guilford Press.

Griffin, D., & Bartholomew, K. (1994a). The metaphysics of measurement: The case of adult attachment. In K. Bartholomew & D. Perlman (Eds.), Advances in personal relationships: Vol. 5. Attachment processes in adulthood (pp. 17-52). London: Kingsley.

Griffin, D., & Bartholomew, K. (1994b). Models of the self and other: Fundamental dimensions underlying measures of adult attachment. Journal of Personality and Social Psychology, 67, 430-445.

Hazan, C., & Shaver, P. (1987). Romantic love conceptualized as an attachment process. Journal of Personality and Social Psychology, 52, 511-524.

Henson, R. K. (1999). Multivariate normality: What is it and how is it assessed? In B. Thompson (Ed.), Advances in social science methodology (Vol. 5, pp. 193-211). Stamford, CT: JAI Press.

Henson, R. K. (2000). Demystifying parametric analyses: Illustrating canonical correlation as the multivariate general linear model. Multiple Linear Regression Viewpoints, 26, 11-19.

Henson, R. K. (2006). Effect size measures and meta-analytic thinking in counseling psychology research. The Counseling Psychologist, 34, 601-629.

Ingram, D. H. (1982). Compulsive personality disorder. The American Journal of Psychoanalysis, 42, 189-198.

Ivey, A. E. (1991). Developmental strategies for helpers: Individual, family and network interventions. Belmont, CA: Brooks/Cole.

Ivey, A. E., & Ivey, M. B. (1998). Reframing DSM-IV: Positive strategies from developmental counseling and therapy. Journal of Counseling & Development, 76, 334-350.

Knapp, T. R. (1978). Canonical correlation analysis: A general parametric significance testing system. Psychological Bulletin, 85, 410-416.

Laporte, L., & Guttman, H. (1996). Traumatic childhood experiences as risk factors for borderline and other personality disorders. Journal of Personality Disorders, 10, 247-259.

Leary T. (1957). Interpersonal diagnosis of personality. New York: Ronald Press.

Livesley, W., & Schroeder, M. L. (1990). Dimensions of personality disorder: The DSM-III-R Cluster A diagnoses. Journal of Nervous and Mental Disease, 178, 627-635.

Livesley, W., Schroeder, M. L., & Jackson, D. N. (1992). Factorial structure of traits delineating personality disorders in clinical and general population samples. Journal of Abnormal Psychology, 101, 432-440.

Lopez, E G. (1995). Contemporary attachment theory: An introduction with implications for counseling psychology. The Counseling Psychologist, 23, 395-415.

Lyddon, W. J. (1993). Contrast, contradiction and change in psychotherapy. Psychotherapy, 30, 383-390.

Lyddon, W. J. (1995). Attachment theory: A metaperspective for counseling psychology? The Counseling Psychologist, 23, 479-483.

Lyddon, W. J., & Sherry, A. (2001). Developmental personality styles: An attachment theory conceptualization of personality disorders. Journal of Counseling & Development, 79, 405-414.

Mahoney, M. J. (1991). Human change processes. NewYork: Basic Books.

Marshall, W. L., & Sefin, R. (1997). Personality disorders. In S. M. Turner & M. Hersen (Eds.), Adultpsychopathology and diagnosis (pp. 508-543). New York: Wiley.

McGuffin, P., & Thapar, A. (1992). The genetics of personality disorder. British Journal of Psychiatry, 160, 12-23.

Millon, T. (1983). Millon Clinical Multiaxial Inventory (MCMI) manual (3rd ed.). Minneapolis, MN: NCS Pearson.

Millon, T. (1987). Manual for the Millon Clinical Multiaxial Inventory-II (MCMI-II). (2nd ed.). Minneapolis, MN: NCS Pearson.

Millon, T. (1994). Personality disorders: Conceptual distinctions and classification issues. In P. T. Costa & T. A. Widiger (Eds.), Personality disorders and the five factor model of personality (pp. 279-301). Washington, DC: American Psychological Association.

Millon, T. (1996). Disorders of personality, DSM-IV and beyond. New York: Wiley.

Millon, T., Davis, R., & Millon, C. (1997). MCMI-III manual (2nd ed.). Minneapolis, MN: National Computer Systems.

Nigg, J. T., & Goldsmith, H. H. (1994). Genetics of personality disorders: Perspectives from personality and psychopathology research. Psychological Bulletin, 115, 346-380.

Paris, J. P. (1996). Social factors in the personality disorders: A biopsychosocial approach to etiology and treatment. New York: Cambridge University Press.

Paris, J. P. (1997). Childhood trauma as an etiological factor in the personality disorders. Journal of Personality Disorders, 11, 34-49.

Pads, J. (1998). Does childhood trauma cause personality disorders in adults? Canadian Journal of Psychiatry, 43, 148-153.

Reisch, T., Thommen, M., Tschacher, W., & Hirsbrunner, H.-E (2001). Outcomes of a cognitive-behavioral day treatment program for a heterogeneous patient group. Psychiatric Services, 52, 970-972.

Sabo, A. N. (1997). Etiological significance of associations between childhood trauma and borderline personality disorder: Conceptual and clinical implications. Journal of Personality Disorders, 11, 50-70.

Sherry, A. (2007). An attachment theory approach to the short-term treatment of a woman with borderline personality disorder and comorbid diagnoses. Clinical Case Studies, 6, 103-118.

Sherry, A., & Henson, R. (2005). Conducting and interpreting canonical correlation analysis in personality research: A user-friendly primer. Journal of Personality Assessment, 84, 35-46.

Simpson, J. A., Rholes, W. S., & Nelligan, J. S. (1992). Support seeking and support giving within couples in an anxiety-provoking situation: The role of attachment styles. Journal of Personality and Social Psychology, 62, 434-446.

Sperry, L., & Mosak, H. H. (1996). Personality disorders. In L. Sperry & J. Carlson (Eds.), Psychopathology and psychotherapy (pp. 279-335). Washington, DC: Accelerated Development. Springer, T., Lohr, N. E., Buchtel, H. A., & Silk, K. R. (1995). A preliminary report of short-term cognitive-behavioral group therapy for inpatients with personality disorders. Journal of Psychotherapy Practice and Research, 5, 57-71.

Swenson, C. R., Sanderson, C., Dulit, R. A., & Linehan, M. M. (2001). The application of dialectical behavior therapy for patients with borderline personality disorder on inpatient units. Psychiatric Quarterly, 72, 307-324.

Thompson, B. (1991). A primer on the logic and use of canonical correlation analysis. Measurement and Evaluation in Counseling and Development, 24, 80-93.

Thompson-Pope, S. K., & Turkat, I. D. (1993). Schizotypal, schizoid, paranoid, and avoidant personality disorders. In P. B. Sutker & H. E. Adams (Eds.), Comprehensive handbook of psychopathology (pp. 411-434). New York: Plenum Press.

Vaillant, G. E. (1987).A developmental view of old and new perspectives of personality disorders. Journal of Personality Disorders, 1, 146-156.

Watson, D. C., & Sinha, B. K. (1995). Dimensional structure of personality disorder inventories: A comparison of normal and clinical populations. Personality and Individual Differences, 19, 817-826. Widiger, T. A. (1992). Categorical versus dimensional classification: Implications from and for research. Journal of Personality Disorders, 6, 287-300.

Widiger, T. A., & Frances, A. (1985). Axis II personality disorders: Diagnostic and treatment issues. Hospital and Community Psychiatry, 36, 619-627.

Zanarini, M. C., & Frankenburg, F. R. (1997). Pathways to the development of borderline personality disorder. Journal of Personality Disorders, 11, 93-104.

Aliasa Sherry, Counseling Psychology Program, Department of Educational Psychology at the University of Texas at Austin; William J. Lyddon, Counseling Psychology Program, Department of Psychology at The University of Southern Mississippi; Robin K. Henson, Educational Research, Department of Technology and Cognition at the University of North Texas. Correspondence concerning this article should be addressed to Alissa Sherry, The University of Texas at Austin, Department of Educational Psychology, 1 University Station D5800, Austin, TX 78712-0383 (e-mail:
Intercorrelation Matrix With Descriptive Statistics for Study Variables

Variable               1          2          3          4          5

 1. Preoccupied       --
 2. Dismissing     -.299         --
 3. Fearful         .140       .127 *       --
 4. Secure         -.085      -.194 *    -.240 *       --
 5. Narcissistic   -.100       .005      -.136 *     .205 **      --
 6. Histrionic     -.105      -.159 **   -.268 **    .384 **    .539 **
 7. Antisocial      .163 **    .023       .020      -.063       .174 **
 8. Dependent       .454 **   -.117       .175 **   -.238 **   -.435 **
 9. Schizotypal     .149 *     .131 *     .194 **   -.296 **   -.098
10. Borderline      .320 **    .080       .201 **   -.211 **   -.124 *
11. Paranoid        .188 **    .204 **    .227 **   -.368 **   -.041
12. Avoidant        .219 **    .094       .292 **   -.405 **   -.546 **
13. Schizoid       -.030       .272 **    .276 **   -.258 **   -.224 **
14. Compulsive     -.177 **    .066       .010       .027       .031

M                  2.77       3.15       2.67       3.27       67.98
SD                 0.77       0.70       0.88       0.61       20.92

Variable               6          7          8          9         10

 1. Preoccupied
 2. Dismissing
 3. Fearful
 4. Secure
 5. Narcissistic
 6. Histrionic        --
 7. Antisocial      .004         --
 8. Dependent      -.340 **    .143 *       --
 9. Schizotypal    -.436 **    .334 **   -.400 **      --
10. Borderline     -.327 **    .596 **    .497 **    .562 **      --
11. Paranoid       -.378 **    .272 **    .346 **    .569 **    .503 **
12. Avoidant       -.749 **    .030       .497 **    .468 **    .377 **
13. Schizoid       -.724 **    .127 *     .178 **    .470 **    .334 **
14. Compulsive      .147 *    -.674 **   -.239 **   -.332 **   -.574 **

M                  62.97      44.80      49.63      42.70      37.40
SD                 22.37      22.95      26.69      26.42      26.86

Variable              11         12         13         14

 1. Preoccupied
 2. Dismissing
 3. Fearful
 4. Secure
 5. Narcissistic
 6. Histrionic
 7. Antisocial
 8. Dependent
 9. Schizotypal
10. Borderline
11. Paranoid          --
12. Avoidant        .487 **      --
13. Schizoid        .466 **    .568 **      --
14. Compulsive     -.224 **   -.119      -.109   --

M                  42.85      41.26      42.14   56.49
SD                 28.29      29.67      26.12   17.44

* p < .05. ** p < .01. All other items were not statistically

Canonical Solution for Attachment Predicting Developmental Personality
Styles (N= 273)

                               Function 1

Variable          Coefficient   [r.sub.s]    % [r.sub.s.sup.2]

Schizoid              .354        -.398             15.84
Avoidant             -.299        -.722             52.13
Dependent            -.420        -.764             58.37
Histrionic            .479         .637              4.58
Narcissistic         -.157         .363             13.18
Antisocial           -.184        -.254              6.45
Compulsive           -.272         .198              3.92
Schizotypal           .064        -.561             31.47
Borderline           -.225        -.653             42.64
Paranoid             -.335        -.705             49.70
 (personality)                                      32.18
[R.sub.c.sup.2]                                     35.90
 (attachment)                                       30.97
Secure                .594         .733             53.73
Fearful              -.225        -.476             22.66
Dismissing           -.155        -.099              0.98
Preoccupied          -.649        -.682             46.51

                               Function 2

Variable          Coefficient   [r.sub.s]    % [r.sub.s.sup.2]

Schizoid              .522         .679             46.10
Avoidant             -.084         .263              6.91
Dependent            -.778        -.445             19.80
Histrionic           -.074        -.453             20.52
Narcissistic         -.282        -.081              0.66
Antisocial            .018        -.130              1.69
Compulsive            .342         .288              8.29
Schizotypal           .218         .262              6.86
Borderline            .030        -.056              0.31
Paranoid              .318         .356             12.67
 (personality)                                      12.38
[R.sub.c.sup.2]                                     20.90
 (attachment)                                       33.42
Secure               -.267        -.394             15.52
Fearful               .317         .362             13.10
Dismissing            .502         .774             59.91
Preoccupied          -.582        -.672             45.16

Variable          % [h.sup.2]

Schizoid             61.94
Avoidant             59.04
Dependent            78.17
Histrionic           61.10
Narcissistic         13.84
Antisocial            8.14
Compulsive           12.21
Schizotypal          38.33
Borderline           42.95
Paranoid             62.37
Secure               69.25
Fearful              35.76
Dismissing           60.89
Preoccupied          91.67

Note. Coefficient = standardized canonical function coefficient;
s = structure coefficient; 52 = structure coefficient squared or
variance explained. An adequacy coefficient indicates how adequately
the synthetic scores on a function do at reproducing the variance in a
set of variables. It is the mean of the squared structure coefficients
on the variable. Noteworthy structure coefficients are in bold type.
COPYRIGHT 2007 American Counseling Association
No portion of this article can be reproduced without the express written permission from the copyright holder.
Copyright 2007 Gale, Cengage Learning. All rights reserved.

Article Details
Printer friendly Cite/link Email Feedback
Title Annotation:Research
Author:Sherry, Alissa; Lyddon, William J.; Henson, Robin K.
Publication:Journal of Counseling and Development
Geographic Code:1USA
Date:Jun 22, 2007
Previous Article:Predictors of counselor educators' cultural discriminatory behaviors.
Next Article:So why aren't counselors reporting n = 1 research designs?

Terms of use | Privacy policy | Copyright © 2019 Farlex, Inc. | Feedback | For webmasters