The purpose of "Considering Consistency" is to provide conceptual and procedural guidance regarding reliability within the context of a local assessment system. Some of this document's contents echo "Measured Measures." For instance, methods for estimating the reliability of individual assessments, particularly as appropriate for constructed-response and performance-based assessments are revisited. But unlike "Measured Measures," the present document offers specific guidance regarding desirable levels of reliability and, moreover, it does so by considering the system as a whole. For example, what is meant by the "reliability" of a local assessment system? What strategies are available for establishing reliability in this regard? What is an acceptable reliability at the lowest level in the system? How does reliability change as one moves from one level in the system to the next--e.g., from the performance indicator score to the content cluster to the content area? "Considering Consistency" has been separated into conceptual guidance (Part I) [prepared by Theodore Coladarci]; and procedural guidance (Part II) [prepared by Jill Rosenblum]. The intention is that the conceptual discussion provides a helpful rationale for what follows procedurally. But it is also recognized that by better understanding the procedural, readers in turn will have a fuller appreciation of the conceptual. (Indeed, some readers may prefer to approach the document in reverse order.) The hope is that the task of becoming thoroughly acquainted with "Considering Consistency" will fall primarily on the shoulders of a small team who, in turn, will use the document as a framework for staff development, local assessment initiatives, and related activities.