This correlation study contributes to seek an alternative approach towards caring classroom by way of integration into daily practice. The possibilities of a teacher and students interaction, which occupies most classroom activities, in facilitating classroom climate around caring were examined through correlation and multiple regression analyses. This strategy is grounded in a review of current deficit-based, decontextualized programs and interventions. In particular, this study investigated the impact of a teacher’s evaluative or non-evaluative feedback on students’ perception of caring classroom, in which studies have overlooked. For that, Wells’ framework for analyzing a teacher-students interaction (Wells, 1999) was used with dichotomous categories, the IRE (initiation/response/evaluation) and the IRF (initiation/response/non-evaluative follow up). This study was conducted in one six grade classroom in South Korea, with one teacher and 28 students being participated. A teacher-students interactions were recorded through ten lessons, and students’ perceptions about their classroom climates were collected by the personal measure called ‘What is Happening In this Class? (WIHIC; Fraser, McRobbie & Fisher, 1996). The WIHIC included four aspects of caring classroom climate: Student cohesiveness, teacher support, cooperation and equity. Both data were then analyzed in order to respond the research questions below: 1. To what extent do IRE and IRF take up classroom dialogue? 2. Does IRE/F proportion of the observed dialogue correlate with student’s caring rating scale? 3. How much impact of the IRE/F pattern dialogue on students’ perception of caring classroom? The results showed that a teacher-students interaction during class has potentials in cultivating caring classroom climate. Especially, the teacher’s dialogic stance on whole classroom interaction was more related to students’ sense of teacher support and equity than the monologic interaction. Significant associations were yielded between the IRE/F and teacher support and between the IRF and equity in the classroom. However, the IRE/F were not significant predictors of student cohesiveness and cooperation in the study; thus, further study is required given the theoretical relevance and the complexity of those dimensions. In addition, the regression models in the study presented the opposite effects of the IRE/F pattern dialogue on a range of aspects of caring classroom, negative and positive respectively. The excerpts from the transcripts were used to support the statistical outcomes of the study.