Study 1 - Adult Attachment Scale

Collins and Read developed a multi-item scale, the Adult Attachment Scale, to measure the dimensions which are believed to undely attachment styles. They developed and tested a 21 item scale of which 15 items were based on Hazan and Shaver’s single item measure (1987) and a further 6 items added to consider confidence in the dependability of others and reactions to separation. Following factor analysis, 3 of the 21 items were found to be non-significant and 3 factors emerged as the most important, these were labelled Close, Depend and Anxiety. Of these dimensions, the Close and Depend dimensions correlated closely and Collins and Read experimented with a two dimensional system but found it only classified 73% of the sample correctly. These dimensions did not correspond with the attachment styles but could be related to them by relative weighting.

The proportionality of individuals classified using these factors was different to the basic attachment styles. Detailed examination showed that some Secure had been reclassified as Anxious and some Anxious/Ambivalent as Avoidant. Cluster analysis of all 21 factors showed three distinct clusters formed which appeared to correspond well to the attachment styles. These were then examined against the three dimensions. It was found that the clusters corresponded well with Hazan and Shaver’s categories (1987) but gave stronger relative scores on the three dimensions of Close, Anxiety and Depend. Collins and Read concluded that “We believe the dimensions measured by the Adult Attachment Scale capture much of the core structures that are thought to underly differences in attachment styles.” (p. 650).  

Study 2 - Attachment classification and history, beliefs about the self and the world and romantic couple behaviour

Following on from Hazan & Shaver (1987), Collins and Read explored the relationship between attachment history and beliefs about self and others.  Secure subjects had a more positive self-image than insecure types. They were more trusting in general and likely to believe in people’s altruism and capacity and willingness and to adapt and control the outcomes of their lives. Their views of love were more romantic and less practical. They tended to report warm relationships with caregivers.

Insecure subjects showed lower self-worth and confidence.  They believed human nature was complex and difficult to understand, that others were less altruistic and more likely to conform to social pressures. Love style was related to obsession/dependency. They tended to report cold or inconsistent caregiving. Collins and Read conclude that they have presented evidence that “differences in attachment are linked to different in beliefs about self and others in ways that are consistent with attachment theory.” (p. 654).

Study 3 - Relationship between attachment classification and choice of romantic partner

Collins and Read examined dating couples (Dyads = 71, mean age = 22) to investigate the relationship between (1) each partner’s attachment dimensions, (2) the subject’s opposite sex-parent and their partner’s attachment dimensions and (3) attachment dimensions and relationship quality. Examination of the interactions between partner’s attachment dimension indicated the following relationships. Those high on close and depend tended to be paired with others similarly high on these factors, especially Close. Those high on anxious tended to be paired with partner’s low on close and low on depend (avoidant). This work was followed up by Simpson (1990), Kirkpatrick & Davis (1994) and Collins & Read (1994).

“For both men and women, descriptions of the opposite sex parent predicted the attachment dimensions of their partner, but descriptions of the same-sex parent did not.” (p. 660). For men, relationship with mother mainly predicted partner’s Anxious score.  For women, relationship with Father mainly predicted partner’s Depend and (less strong) Close scores. This suggests the [relationship with the] opposite sex parent is used as a [working] model for future romantic relationships but the nature of this is different between the genders.

The main attachment dimensions that were predictive of relationship quality reflects the gender difference in the dimensions carried forward from the opposite-sex parent relationship.  For men, the most important dimension was partner’s Anxious and own Close, for women it was partner’s Close score and own Depend and Anxious scores.  The following table shows the relationship evaluations based on subject’s own and partner’s attachment dimensions. (Strong relationships (p < .001) are in bold, others are p < .01):  Collins and Read suggest these gender differences are consistent with gender stereotyping and are due to women being socialized to develop closeness and men being socialized to develop personal identity. This would lead to men being sensitive to women’s anxiety and dependance and women being sensitive to men’s closeness.    

Correlations between Relationship Quality and Own and Partner's  Attachment Dimensions 

They concluded that “scores on [attachment] dimensions are likely to influence who one chooses as a dating partner and may play an important role in organizing behaviours, perceptions and expectations within dating relationships.” (p. 662).