Undertook a longitudinal study of 144 dating couples (mean age male = 19.4, female = 18.7) to investigate degrees of interdependence, commitment and trust as well as frequency and type of emotions and (6 months later) reaction to relationships dissolusion for the 48 couples who had broken up.

Attachment type was measured using a new 13 item, self-report scale, the Attachment Style Measure. Some calculations were performed using two dimensional (avoidant and anxious), three dimensional and descrete categories. (the attachment classifications). The general pattern of findings were found to be the same, regardless of style of measuring.  Other factors were measured using various existing self-report measures.

Comparing scores within dyads
At the dyad level, various effects were found. The following table gives the Pearson product moment correlations found between genders.

This can be reduced to three clauses which relate likely attachment style pairings:

1. Those with high scores on secure are most likely to be with partners who are also high on secure.
2. Avoidant individuals and anxious individuals tend towards each other.  Whilst still slight, this has a stronger effect where the male is avoidant and the female anxious.
3. Avoidant and anxious partners second preferance is to be with a partner with similar attachment style to themselves.

Individual Scores on Relationship Characteristics
In the following tables, comparing various relationship measures with own and partner's attachment style, positive and negative relationships are shown.  Interdependence was nbased on love for, dependency on and self-disclosure to partner. Commitment was based on commitment to, and investment in the relationships. Trust was scored on predictability and dependability of, and faith in partner.  Stronger effects (p<.01) are shown with three characters. 

 
 Similar examinations were made of the subject's attachment style the partners scores on these dimensions:
 

 
From the above, the following correspondences between attachment style and experience of relationships can be drawn:

1. Personal attachment style has a more significant effect on how relationships are experienced than partner's style.
2. Males will report lower levels of interdependence, commitment and satisfaction when with anxious females.
3. Female will report lower levels of trust and satisfaction when with avoidant males.

Individual Scores on Experience of Emotion
Found that secure individuals described more frequent positive emotions and less frequent negative emotions within their current relationships than both insecure attachment types. Regarding the effect of the partner's atrachment style, women dating secure men tended to experience more frequent positive emotion than if dating insecure men and men dating dating secure women tended to experience more frequent positive emotion than those dating anxious women.

Response to Relationship Dissolusion
They followed subjects up 6 months later and assessed affective response to relationship breakup in those subjects who were no longer together. Found that avoidant attachment reported less distress at the ending of the relationship but that this effect was only significant for men. This supports the link between attachment and affect regulation.

Discussion
Simpson notes that personal attachment style is more important in how a relationship is perceived and experienced than the attachment style of the partner.  Also, whilst relationships were found between subject and partner's attachment style, these were all mild and were apparently unaffected by duration of the relationship.  If a subject's attachment style had a major effect on the partner's style, stronger relationships should have been observed.  Whilst it is accepted that relationships can have an effect on attachment style, this study suggests that attachment style is actually pretty robust.as no significant relationships effects on attachment style were not found.