Verbal person centeredness (VPC) is a long-dominant conceptualization of quality in supportive communication, but prior work provides limited guidance for researchers seeking to examine its impact in interaction. This study examined the extent to which overall VPC for an interaction is predicted by support providers’ specific behaviors. In a laboratory setting, pairs of friends (320 dyads) discussed a problem that one of them was currently experiencing. Transcripts of these interactions were coded by separate research teams for overall VPC of support provision and for the topic, function, and experiential focus of each discrete thought-unit produced by the support provider. Regression analyses showed that VPC was predicted by units representing specific, theoretically interpretable combinations of topical, functional, and experiential focus. The findings provide evidence of validity for procedures used to code VPC for supportive interactions while also displaying ways in which person centeredness is accomplished somewhat differently in interaction than in discrete messages.