Exploring Similarity and Stability of Differentiation in Relationships: A Dyadic Study of Bowen's Theory

Abstract

Bowen family systems theory (BFST) identifies differentiation of self as a crucial characteristic that relates to one's individual and relational maturity. Bowen theorizes that an individual's level of differentiation typically remains static over time and that individuals select and pair in relationships with others who have similar levels of differentiation. This study aimed to test the hypotheses of BFST by using components of differentiation of self, emotional reactivity, and emotional cutoff, in dyadic structural equation modeling. Specifically, a longitudinal dyadic confirmatory factor analysis and structural equation longitudinal panel model were two analyses utilized to statistically test the dyadic nonindependence and stability of emotional reactivity and emotional cutoff across time. Results indicated mixed support for Bowen's assumptions regarding similarity (nonindependence) and stability. That is, small levels of nonindependence and significant stability paths across time were found. Theoretical implications and considerations for future research are discussed.

Differentiation of self and relationship attachment, quality, and stability: A path analysis of dyadic and longitudinal data from Spanish and U.S. couples

“…Conversely, the solid self is characterized by more consistent conviction and more stability over time, and individuals with higher levels of DoS are postulated to operate from a place of the solid self [ 14 , 29 ]. Handley and colleagues found more stability (less change) in DoS over time with their sample [ 26 ], while the current study found a significant increase in DoS over time in both men and women and in both cultures. Notably, Handley and colleagues used yearly intervals in their assessment of changes in DoS [ 26 ], while the current study had a much longer interval between the two time points assessed (four years in the U.S. sample, seven years in the Spanish sample).…”

“…Handley and colleagues found more stability (less change) in DoS over time with their sample [ 26 ], while the current study found a significant increase in DoS over time in both men and women and in both cultures. Notably, Handley and colleagues used yearly intervals in their assessment of changes in DoS [ 26 ], while the current study had a much longer interval between the two time points assessed (four years in the U.S. sample, seven years in the Spanish sample). It is possible that in the current sample, the greater amount of time between assessments of DoS was an important factor in forming a more solid self, particularly during a time of stressful events.…”

“…Simultaneously, we focused on extending this examination of potentially co-occurring interactions by analyzing relationship stability. While some researchers have conceptualized (and assessed) relationship stability as a component of relationship quality [ 11 ], other convincing evidence suggests that the stability of and satisfaction in a relationship represent two distinct constructs [ 26 , 27 ]. More specifically, an individual may remain in a relationship (i.e., stability) and not find satisfaction in it and, on the other hand, satisfaction may be relatively high, yet the individual remains open to finding a “better” partner.…”

“…We also controlled for stressful life events and explored gender differences. Moreover, we examined the stability of DoS, a hypothesis with mixed empirical longitudinal support [ 26 ]. Specifically, we sought to answer the following research questions:…”

Psychometric evaluation of the Swedish Quality of Dyadic Relationships scale – homogeneity and construct validity

“…There is thus a potential risk that members of dyads have influenced each other's answers without the research team knowing about it. Secondly, even if the participants answered the questionnaires completely independently, there is a risk that the partner's scores are related because individuals select and pair with others who are similar to themselves in how to form relations (35). However, this was accounted for in the mixed-effects logistic regression analyses.…”