Login / Signup

Comparing Main, Goldwyn, and Hesse (Berkeley) and Crittenden (DMM) coding systems for classifying Adult Attachment Interview transcripts: an empirical report.

Franco BaldoniMattia MinghettiGiuseppe CraparoElisa FacondiniLoredana CenaAdriano Schimmenti
Published in: Attachment & human development (2018)
Few studies have compared different systems in classifying Adult Attachment Interview (AAI) transcripts. In this study, the AAI was administered to 90 Italian parents (45 couples), and the AAI transcripts were independently classified according to Main, Goldwyn, and Hesse's (Berkeley) and Crittenden's (Dynamic-Maturational Model [DMM]) criteria. The two classification systems were not significantly associated, with some limited convergent results only when the interviews resulted in organized (Berkeley) and normative (DMM) attachment classifications. Otherwise, the Berkeley system identified more secure individuals than the DMM system, and many texts judged secure on the Berkeley system were identified as insecure on the DMM system. Since the Berkeley and the DMM systems rest on remarkably different conceptualizations of the nature and functioning of the attachment behavioral system (e.g. fear is conceived as organizing in the DMM and as potentially disorganizing in the Berkeley), the attachment classifications resulting from their applications should not be considered measurements of the same phenomena.
Keyphrases
  • deep learning
  • case control