(PS10-A18) Assessing Convergent Validity of a Observational Vulnerable Feelings Task

Based on the Behavioral Theory of Intimacy (Cordova & Scott, 2001), Dorian and Cordova (2004) developed a system to observe intimate exchanges in couples called the Hurt Feelings Interactions task.  Specifically, individuals are asked to share with their partners about a time when their feelings were hurt by the partner. This interaction is then coded with the Intimacy Coding System for elements of intimacy including manifest vulnerability and reinforcing or punishing responses to vulnerability. Per the Behavioral Theory of Intimacy, an interaction is “intimate” when one partner expresses vulnerability and the other partner responds supportively.  For the current study and in consultation with Cordova, we have adapted this system to be a more general Vulnerable Feelings task, wherein individuals are asked to think of a time when they felt sad, scared, ashamed, lonely, or insecure and talk about it with their partner. This is then coded with the Intimacy Coding System.    

Given this innovation, in the current study, we plan to present interrater reliability for the coding system, and evaluate a pattern of convergent validation in terms of how dimensions of the coding (e.g., vulnerability, support) relate to each other and to couples self-reports on conceptually related concepts (e.g., satisfaction, trust, and communication).     

We plan to test these patterns in a sample of 204 opposite sex community couples. The sample is 43.8% White non-Hispanic, 22.1% non-White Hispanic, 19.1% Black non-Hispanic, with the remaining 15% multiracial or other racial identifications. For the Vulnerable Feelings Task, both partners each chose a vulnerability example to share. Thus, there were two interactions per couple, with each partner having a turn as the vulnerable speaker and the supportive/punishing responder. Partners then completed self-report measures with several controls to ensure privacy of responding and encourage honest responding.    

Processing the coding data is ongoing, but will be complete in time for presentation of this work. Analysis of self-report data shows significant and logical convergence among relationship satisfaction, trust, and communication patterns, both within same reporter and between partners in the dyad.  For example, associations of these constructs between partners ranges from r = .27 to .53, all ps < .001.  Given the adequate sampling distribution across racial and ethnic groups, and the distinguishable dyads based on gender, we can also evaluate potential differential reliability and validity indices of the coding data for groups based on gender and race/ethnicity.    

Results from the current study will provide a useful overview of this adaptation of the observational system, inform psychometric adequacy and generalizability of the system to diverse groups, and provide insights into the connections of constructs such as trust and the manifest expressions of vulnerability.