incongruent with affect; stereotypic, jerky; and freezing (Main & Solomon, 1990). Disorganized attachments often emerge in the context of a caregiver who is at times frightening and also frightened (Main & Hesse, 1992).
This foundational work in attachment theory was later extended to attachment in older children and adults. Across development, our attachments become less about physical proximity and more about emotional availability. Therefore, Mary Main and her colleagues (George, Kaplan, & Main, 1985) developed the Adult Attachment Interview, which focuses on how individuals talk about their early childhood experiences (you can find a full interview at http://www.psychology.sunysb.edu/attachment/measures/content/aai_interview.pdf ). Critically, the interview bases attachment styles on the way in which caregiving experiences are described. Respondents are assigned to four groups that mirror those identified in the Strange Situation Paradigm. An autonomous (i.e., securely attached) adult is one who values attachment relationships and talks about caregiving experiences in a way that is balanced, coherent, and open. A preoccupied adult will dominate the interview with excessive attention towards caregivers and anger. A dismissing adult, on the other hand, will tend to de‐emphasize the impact of caregiving experiences. Finally, a disorganized adult’s interview will contain evidence of prior difficult caregiving experiences (like abuse or loss) that are unresolved in adulthood. Extensions of this interview have now been developed and used with school‐age children (Shmueli‐Goetz, Target, Fonagy, & Datta, 2008) and adolescents (Venta, Shmueli‐Goetz, & Sharp, 2014).
While most early research has focused on categorical attachment styles (e.g., secure), more recent research measuring attachment with self‐report questionnaires demonstrated that people’s attachment may vary continuously (Brennan, Clark, & Shaver, 1998; Fraley & Waller, 1998). In the continuous model (Figure 4.3), people are assumed to vary regarding avoidance and anxiety. The combination of an individual’s location on each dimension maps on to the categorical notion of attachment styles (Fraley, Hudson, Heffernan, & Segal, 2015). Whether attachment is categorical or dimensional remains quite controversial.
FIGURE 4.3 Two‐Dimensional Model of Attachment.
Source: Fraley et al. (2015). Reproduced with permissions of American Psychological Association.
Factors Affecting the Distribution of Attachment Styles in the General Population
The distribution of attachment styles in the general population depends somewhat on the assessment used, whether a dimensional or categorical approach was considered, and the developmental stage. Overall, about 60% of people tend to be classified as securely attached, with the remaining 40% divided among the insecure and disorganized classifications. Estimates are also affected by age, gender, socioeconomic status, and race/ethnicity, as described in the following sections. It should be noted, however, that the attachment classification estimates provided are based on community samples—that is, based upon studies that did not seek out participants with psychopathology where, as a general rule, attachment security is less common.
Gender and Age
Regarding age, rates of attachment security are similar across development, though adolescents tend to be overrepresented in the dismissing group and very young children are more often rated as disorganized. Indeed, in a sample of fifteen‐month‐old children, 60.0% were classified as secure, 13.3% avoidant, 4.2% resistant (i.e., ambivalent), and 15% disorganized (Fraley & Spieker, 2003b), whereas in childhood, 66% were classified as secure, 28% dismissing, 3% preoccupied, and 4% disorganized (Shmueli‐Goetz, Target, Fonagy, & Data, 2008); in adolescence, approximately 52% were classified as secure (with 34.44% dismissing, 5.38% preoccupied, and 2.87% disorganized; Fearon, Shmueli‐Goetz, Viding, Fonagy, & Plomin, 2014); and in adults, 58% were rated as secure, 24% dismissing, and 18% as preoccupied (van IJzendoorn & Bakermans‐Kranenburg, 1996). In terms of gender differences, Shmueli‐Goetz et al. (2008) reported a non‐significant association between attachment and gender in middle childhood and van IJzendoor and Bakermans‐Kranenburg (1996) reported comparable attachment classifications between mothers and fathers. However, age and gender differences have been more pronounced in studies of continuously rated attachment (Fraley et al., 2011).
Socioeconomic Status (SES), Race, and Ethnicity
There are mixed findings regarding relations between attachment, SES, race, and ethnicity, which are further complicated by interactions between these variables. For example, Bakermans‐Kranenburg and colleagues (2004) found that African American children had significantly lower scores on a measure of attachment security than White children but that this relation was explained by income—African American ethnicity was associated with low‐income status, which affected the quality of maternal sensitivity and therefore attachment security. Thus, in this study, it was poverty, not race/ethnicity that was truly linked to decreased attachment security. This study echoes a broader link between low income and attachment insecurity (van IJzendoorn & Bakermans‐Kranenburg, 2010).
There have been concerns that the measurement of attachment is biased against non‐Western cultures. Some studies have demonstrated the overrepresentation of insecurity among children from parts of Africa, Japan, Indonesia, and Israel (see van IJzendoorn & Bakermans‐Kranenburg, 2010; van IJzendoorn & Sagi‐Schwartz, 2008). However, most evidence suggests that secure attachments are readily detected across varied cultures (van IJzendoorn & Sagi, 1999), intracultural differences are larger than intercultural differences (van IJzendoorn & Sagi‐Schwartz, 2008), and attachment styles are similar across studies conducted with many cultures and languages (van IJzendoorn & Bakermans‐Kranenburg, 2010).
Attachment and Psychopathology
Many studies conducted in children, adolescents, and adults document a link between insecure attachment and psychopathology with, generally, attachment insecurity in infants and young children relating most to behavior problems, with the links between attachment insecurity and psychopathology becoming more diffuse across development.
Infancy
Research with high‐risk samples of infants shows significant relations between psychopathology and attachment. For example, avoidant attachment was longitudinally associated with externalizing symptoms like non‐compliance and hostility at age 4 (Erikson, Sroufe, & Egeland, 1985) and in elementary school (Renken, Egeland, Marvinney, Mengelsdorf, & Sroufe, 1989). Disorganized attachment in infancy has been linked to the later emergence of hostile and aggressive behavior (e.g., Lyons‐Ruth, Alpern, & Repacholi, 1993) and to dissociative symptoms across a 19‐year longitudinal study (Ogawa, Sroufe, Weinfield, Carlson, & Egeland, 1997). Specific links between early attachment styles and other forms of psychopathology have been more difficult to establish. For example, internalizing symptoms have been associated with resistant (Warren, Huston, Egeland, & Sroufe, 1997), disorganized (Shaw, Keenan, Vondra, Delliquandri, & Giovannelli, 1997), and avoidant (Lyons‐Ruth, Easterbrooks, & Cibelli, 1997) attachments.
Preschool and School‐Aged Children
In preschool and school‐aged children, a link between attachment insecurity and behavior problems remains with studies among preschoolers linking behavior problems to resistant‐like attachment