Script-like attachment representations in dreams containing current romantic partners

Abstract

Recent research has demonstrated parallels between romantic attachment styles and general dream content. The current study examined partner-specific attachment representations alongside dreams that contained significant others. The general prediction was that dreams would follow the “secure base script,” and a general correspondence would emerge between secure attachment cognitions in waking life and in dreams. Sixty-one undergraduate student participants in committed dating relationships of six months duration or longer completed the Secure Base Script Narrative Assessment at Time 1, and then completed a dream diary for 14 consecutive days. Blind coders scored dreams that contained significant others using the same criteria for secure base content in laboratory narratives. Results revealed a significant association between relationship-specific attachment security and the degree to which dreams about romantic partners followed the secure base script. The findings illuminate our understanding of mental representations with regards to specific attachment figures. Implications for attachment theory and clinical applications are discussed.

Keywords:

• secure base
• script
• mental representations
• dreams
• cognitions
• narratives

Acknowledgements

Special thanks to Suzanne Riela and Arthur Aron, and also to all the research assistants who helped during the course of this project: Jamie McDonald, Marissa Gomez, Mike Subrize, Stephanie Nolasco, Brittany Anderson, Corey Herth, Stephanie Nandoo, Danielle Wischenka, Monica Margulis, Greg Vosits, Melissa Willner, Sal Silinonte, and Urmi Vaghela.

Notes

1. Dimensions of attachment anxiety and avoidance were also assessed using the 36-item Experiences in Close Relationships Revised questionnaire (ECR-R; Fraley, Waller, & Brennan, 2000). These facets of attachment security did not correlate with the secure base script measure and were not central to the current research question, thus they were excluded from the main analyses. However, all associations reported in this paper remained significant while controlling for self-reported attachment styles in multiple regression equations.

2. This correlation is slightly lower than that reported in the Results section, because single coder scores are less reliable than scores aggregated across several coders (Block, 1963; Epstein, 1983; Waters, 1978). Applying the Spearman-Brown correction for attenuation yielded the value reported in the Results.

3. In addition to self-reported attachment style, participants were also assessed on interdependence (Rusbult Investment Model Scale; Rusbult, Martz, & Agnew, 1998) and closeness (IOS; Aron, Aron, & Smollan, 1992) in their relationship, as well as depression, anxiety, and stress (DASS; Lovibond & Lovibond, 1995), Big-5 traits (Rammstedt, 2007), and self-esteem (RSE; Rosenberg, 1965). These variables were used for analyses unrelated to this study, and also used to compare those who dropped out of the study and those that completed it.

4. As stated previously, the third-person stories from the original generalized script assessment (Waters & Rodrigues-Doolabh, 2001) were also included in the design, although they were not relevant to the study goals or hypotheses. Participants’ scores on this generalized attachment measure correlated significantly with the first-person relationship-specific script scores (r = .76, p < .001), but they were not significantly associated with dream content (though there was a trend in the predicted direction, it did not reach statistical significance; r = .15, ns). Thus, it was not included in the regression analyses. This point is further addressed in the Discussion section.