Relational Maintenance and Inclusion of the Other in the Self: Measure Development and Dyadic Test of a Self-Expansion Theory Approach

Abstract

This article reports a series of two studies that develop and validate a measure of relational maintenance behavior informed by self-expansion theory. The first study evaluated an initial item pool, identified 11 dimensions of maintenance behavior, and established concurrent/divergent validity with theoretically related constructs. The second study further tested the instrument in a sample of 123 romantic dyads, demonstrating that inclusion of the other in the self (Aron, Mashek, & Aron, 2004) predicts frequency of maintenance behavior in a communally oriented fashion. Additional analysis identified four maintenance dimensions corresponding to dimensions of self-other inclusion in self-expansion theory. These results commend self-expansion theory as a promising direction for future maintenance research and offer a maintenance measure for theoretical development and practical application.

Notes

¹In the final four-category confirmatory factor analysis (Study 2), physical affection and shared networks were modeled as separate latent constructs, given their different patterns of association with IOS. Physical affection and shared networks are presented in the same factor here for theoretical coherency with self-expansion theory (Aron et al., 2004), although a researcher may find it empirically (or theoretically) necessary to model them separately in some cases.

Note. Means, standard deviations, and Cronbach's alpha obtained from manifest-level composite scores. Correlation coefficients obtained from the confirmatory model.

*p < .05. **p < .01.

Note. IOS = inclusion of the other in the self.

*p < .05. **p < .01.

Note. Change in chi-square values indicate loss in model fit when constraining the relevant parameter (means or standard deviations) to equality across partners. IOS = inclusion of the other in the self.

^a Because IOS uses one indicator, the baseline confirmatory model is saturated (i.e., zero degrees of freedom).

^b Correlation coefficient between the male construct and female construct.

^* p < .05. **p < .01.

Note. APIM = actor-partner interdependence model; IOS = inclusion of the other in the self; B_a,p = regression parameter for the actor effect and the partner effect, constrained to equality; B_r = regression parameter for the relationship (i.e., interaction) effect.

^a Test for communal orientation, that is, that each person's actor effect is equivalent to his or her partner's effect (Kenny, Kashy, & Cook, 2006), as compared to the unconstrained model.

^b Test for communal orientation and for equality of actor and partner effects across men and women, as compared to the unconstrained model.

^† p < .09. *p < .05. **p < .01.

Additional information

Notes on contributors

Andrew M. Ledbetter

Andrew M. Ledbetter, Department of Communication Studies, Texas Christian University.