Adult attachment styles in 354 heterosexual couples in serious dating relationships were examined. Principal findings included the following: (a) male and female attachment styles were nonrandomly paired, for example, no anxious-anxious or avoidant-avoidant pairs were found; (b) male and female styles related to concurrent relationship ratings of both partners in different but theoretically meaningful ways; (c) male and female styles contributed significantly to longitudinal prediction of relationship stability and status, even when prior duration and commitment to the relationship were statistically controlled; (d) specifically, relationships of avoidant men and of anxious women were surprisingly stable over 3 years, particularly in light of the relatively poor ratings of these relationships by both partners at Time 1. Discussion focuses on the need to integrate gender role considerations and relationship dynamics and processes into theorizing on adult attachment.

Hazan and Shaver (1987; Shaver, Hazan, & Bradshaw, 1988) have proposed that Bowlby's (1969, 1973, 1980) formulation of attachment theory could serve as the basis for a theory of romantic love that (a) offers an explanation of how individuals develop different styles of relating to romantic partners and (b) offers an account of how such components of love as attachment, caregiving, and sex are related to each other over the life cycle. A number of studies have found attachment style to be related in theoretically consistent ways to one's own attachment history and beliefs about relationships (Feeney & Noller, 1990; Hazan & Shaver, 1987; Levy & Davis, 1988). Research with heterosexual couples has shown attachment styles to be related to both partners' levels of trust, love, satisfaction, and commitment in the relationship, including some indication of gender differences in these associations (Collins & Read, 1990; Simpson, 1990