In this study, the authors conceptualize teaching episodes such as an integrated set of observable student-teacher interactions. Instructional interactions that take place between teachers and students around critical academic content are a defining characteristic of classroom instruction and a component carefully defined in many education interventions (Cohen, Raudenbush, & Ball, 2003; Pianta & Hamre, 2009). This presentation focuses on observational data collected during two separate IES-funded randomized controlled trials. The first study, Project Enhancing Core Reading Instruction (ECRI), is a four-year project designed to study the efficacy of a multi-tiered, first-grade intervention system. The second study, the Early Learning in Mathematics (ELM) project, investigated the efficacy of a core (Tier 1) kindergarten mathematics curriculum. There are two primary purposes of this presentation. The first is to establish classroom stability of instructional interactions during early literacy and beginning mathematics instruction. The second purpose is to examine the association between these instructional interactions and academic achievement. Three research questions were posed to guide the research on instructional interactions: (1) How stable are instructional interactions across the academic year? (2) How reliable are classroom means of instructional interactions (averaged across observations)? (3) Is the rate of instructional interactions associated with student achievement? Standardized observation protocols have demonstrated the capacity to measure core elements of intervention trials, including implementation fidelity and intervention impact (Snyder et al., 2006). This study used a direct observation tool to systematically investigate the explicit instructional interactions that occur during early literacy and beginning mathematics instruction. Preliminary analysis of the Early Learning in Mathematics (ELM) and Project Enhancing Core Reading Instruction (ECRI) data reveals that many of the studied instructional variables were unrelated to student achievement. One explanation for these results is that the low stability of the instructional variables attenuated the relationships. The present study highlights the importance of obtaining stability of observed teaching practices (Shavelson & Dempsey, 1976). The implications of the present study warrant further investigation of methods to improve stability of these instructional variables, such as controlling for the type of academic content observed and observing consecutive days of instruction. Tables are appended.