The purpose of this report is to present our process and results for establishing validity and reliability of an observation tool used to investigate teaching practices that high school mathematics teachers use to engage students. We developed our tool using established practices, such as reviewing literature to develop a framework for instruction and piloting the tool to design descriptive levels for rubrics. After validating externally by consulting experts, additional rubrics regarding teaching mathematics for equity were added to the tool. We conducted a reliability study of 149 episodes of classroom instruction (equivalent to 447 10-minute segments of instruction in all), two raters per episode, to investigate the nature of coding disagreements. Most disagreements occurred due to raters noticing different evidence rather than different interpretations of rubrics, which suggested the value of two raters and resolution meetings. [For the complete proceedings, see ED630060.]