In this talk I will present a new measure for systematic classroom observation, which focuses on aspects of the classroom climate hypothesized to be of importance to the development of students’ mathematics self-efficacy beliefs. Going beyond contemporary observation instruments the current instrument focused particularly on teachers’ efficacy-supportive practices. Observations were carried out in nine classrooms (181 students) during a sequence of four lessons in which a new mathematical topic was introduced or revisited (see Street et al, 2022), by two observers. In order to provide support for the new observation measure aspects of reliability (factor structure, internal consistency), and convergent and divergent validity with two established measures for classroom observation (CLASS and ISTOF) were investigated. Then the associations between students’ self-efficacy beliefs and mastery experiences were investigated. Developing high-quality classroom observation measures is important to enable insights into teacher-student interactions and their relationship with, e.g., student outcomes such as learning and motivation. The new measure presented in this talk can help inform teachers about how different classroom practices are related with the development of students’ mathematics self-efficacy and mastery experiences over time.