Ten attachment items at intake, read against the trauma history
- Five Traditional Chinese short forms of the Experiences in Close Relationships-Revised (18, 16, 12, 10 and 8 items) were tested in three Taiwanese samples: 380 adults recruited through universities, 402 community adults and 317 adults who met DSM-5 Criterion A for PTSD.
- The two-factor model of attachment anxiety and avoidance was superior for the 18-item and 10-item forms. For the 10-item form, Cronbach's alpha was 0.84 for anxiety and 0.87 for avoidance in the community sample, and 0.83 and 0.87 in the trauma-exposed sample.
- In 128 adults retested after four weeks, ICCs ranged from 0.76 to 0.79 for the anxiety subscale and from 0.69 to 0.75 for the avoidance subscale.
- Among trauma-exposed adults, those with interpersonal trauma (n = 183) scored higher on 10-item anxiety than those with impersonal trauma (n = 134), 4.18 against 3.75 (F = 6.48, p < .05). Avoidance did not differ: 3.53 against 3.47 (F = 0.12).
Attachment questionnaires rarely reach an intake battery, because the full Experiences in Close Relationships-Revised runs to 36 items. Researchers in Taiwan tested how far it can be shortened in a new Traditional Chinese translation. Ten items kept the two scales the instrument is built on, anxiety and avoidance, in trauma-exposed adults too.
Three samples, five lengths
Yu-Lien Huang of Chung Shan Medical University and Kuang-Hui Yeh of Academia Sinica built forms of 18, 16, 12, 10 and 8 items, each rated from 1 to 7. They recruited 380 adults through universities, 128 of whom answered again a month later, then 402 community adults and 317 adults who met DSM-5 Criterion A for PTSD.
The two-factor model fitted best for the 18-item and 10-item forms. For the 10-item form, alpha was 0.84 for anxiety and 0.87 for avoidance in the community sample, 0.83 and 0.87 in the trauma sample. Over four weeks, ICCs ran from 0.76 to 0.79 for anxiety and from 0.69 to 0.75 for avoidance.
Within the trauma sample, people hurt by another person scored higher on anxiety than people whose trauma was impersonal: 4.18 against 3.75. Avoidance stayed level, 3.53 against 3.47.
Two scores, one history
That split is the useful part at a first meeting. A client harmed by someone close may arrive with the anxiety score raised and avoidance where it would be anyway. In the room this tends to look like a client who watches for signs of being dropped: a late reply, a moved session, a holiday. Score the two scales separately, never as a total, and take the trauma history before reading either number.
One caution, in plain words. The ten items measured the same two dimensions in all three groups, but the authors warn that average scores of the trauma sample cannot be set directly against those of the other two. Compare a client with their own earlier score, not with another population's mean.
After interpersonal harm, a raised anxiety score with ordinary avoidance tells you what the client will watch for in the therapeutic relationship.
The samples were predominantly female and included no clinically diagnosed group, which limits generalisation to psychiatric patients. The design was cross-sectional, and the same sample served for both exploratory and confirmatory factor analysis. The items are a new Traditional Chinese translation; scores do not transfer to other language versions.