Page 85 - Personality disorders and insecure attachment among adolescents

Supplementary material 1 gives an overview of the three-way and five-way attachment classifications distribution (AAI) over the whole sample and over the three SCID-II personality disorder groups. No relation was found between BPD, OP, and NP and the (forced) attachment classification (Fisher’s exact test 1.24, p = .921). The comparison between the CC category in the five-way attachment classifications distribution and BPD and OP showed no significant difference (p = 1.0). Also, the E category in the forced classifications distribution of the BPD as opposed to the OP group was not significant (p = 0.569). Subgroups and the AAI scales at t-1 Next, differences between BPD, OP, and NP on the paternal and maternal attachment were examined. The BPD group scored significantly higher on the ‘Devaluating father’ scale (F (2, 59) = 5.69, p = 0.006) in comparison with both other groups. Next, when comparing the BPD group (t-test) with the two other groups combined (Non-BPD), differences were found for: ‘Loving father’ (BPD M = 0.90, SD = 1.37; Non-BPD M = 1.64, SD = 1.25, t = 2.09, p = .041) and ‘Devaluing father’ (BPD M = 2.65, SD = 1.81; Non-BPD M = 1.46, SD = 1.06, t = -2.71, p = .012). To test the predictive value of the two variables (‘Loving father’ and ‘Devaluating father’) of the AAI that significantly differed between the (dichotomous dependent variable) BPD and the non- BPD group, a binary regression was performed. This model was statistically significant (χ2 (2, N = 60) = 6.75, p = .034), explaining 14.8% (Nagelkerke R square) of the variance in the Personality disorders groups, and correctly identifying 71,7% of cases (Loving father OR = 0.908 95% CI 0.462-1.279; Devaluing Father OR = 1.660 95% CI 1.052-2.484). Finally, information on whether the adolescents had a residential father (63%) or mother (93%) or not was compared with the paternal and maternal attachment scales. A significant difference was identified on the ‘Devaluing father’ (p = 0.005) and ‘Idealising father’ (p = 0.005) scale in the group with a non-residential father. 81