Attachment Theory

What it is

Two continuous dimensions — attachment anxiety (fear of abandonment, hypervigilance to the other’s availability) and intimacy avoidance (discomfort with closeness and dependence). The four "styles" are just the name of the quadrant where the dimensions fall; the primary measure is the dimensions. It reads how you behave in relational episodes under stress — rupture, reunion, reaching out for comfort — not how you label yourself.

Where it comes from

What it does NOT measure

It isn’t the Enneagram’s existential fear (being worthless, being controlled), nor the diffuse stress sensitivity of the Big Five. Attachment is specific: the availability of the OTHER in the bond.

Evidence — moderate evidence

Adult Attachment Theory has a solid empirical base compared with most typological personality models. The ECR-R shows high internal consistency (alpha around .90 for Anxiety and Avoidance in validations) and explains roughly 30–40% of the variation in emotions reported in interactions with romantic partners, but only 5–15% in interactions with family/friends (Sibley, Fischer & Liu, 2005, PSPB, 31, 1524–1536). The AAI shows large predictive validity for infant security (d=1.06; N=854; van IJzendoorn, 1995, Psychological Bulletin) and good test–retest/inter-rater reliability (78% agreement across the 3 categories, kappa=.63; Bakermans-Kranenburg & van IJzendoorn, 1993). However, the correspondence between self-report and the AAI is trivial (mean r=.09; N=961; Roisman et al., 2007), indicating that the developmental and social-psychological traditions measure distinct aspects of attachment — a divergence reinforced by recent studies finding low-to-moderate correlations between the AAI, self-report and implicit measures (Strauss et al., 2022; Petrowski et al., 2025). Longitudinal stability is real but moderate (r=.39, with no significant stability beyond 15 years; Pinquart, Feussner & Ahnert, 2013). The categorical structure (secure/anxious/avoidant) isn’t supported taxometrically: Roisman, Fraley & Belsky (2007) showed that the variation is dimensional and that "earned security" doesn’t form a qualitatively distinct group, favoring a continuous representation. In romantic relationships, both anxiety and avoidance negatively predict relationship quality (Li & Chan, 2012; N=21,602), with avoidance more strongly associated with overall dissatisfaction.

The studies behind this

Real, verifiable references — click to check them yourself. That’s what separates a reading from a horoscope.

How it enters the crossing

Anchor tier (0.85; instrument 0.9). In the crossing, it feeds the Heart and the Tide of the Inner Map, and pulls against narratives of independence: "I don’t need anyone" may be autonomy (Values) or avoidance (Attachment) — the crossing decides by the evidence.

The five styles