The purpose of the study was to determine the relationship between superintendent instructional leadership behaviors and the fidelity of implementation of the Instructional Practices Inventory. The leadership behaviors examined by this study were identified by Marzano and Waters (2006) including Collaborative Goal Setting, Non-negotiability of Goals for Achievement and Instruction, Monitoring Progress Toward Goals, Providing Resources to Support Goal Attainment, and Providing Defined Autonomy for Principals. The Instructional Practices Inventory (IPI) was developed by Valentine and Painter and subsequently refined by Valentine (Valentine, 2007). The IPI is a research-based instructional intervention that embeds the reflective study of instructional practices and student engagement in a collaborative model of continuous school improvement. Data were collected from 43 superintendents using surveys adapted from the work of Leithwood, Aitken, & Jantzi (2006) as found in Making Schools Smarter: Leading with Evidence (MSS). Follow-up interviews were conducted with 28% of the superintendents. Data analyses produced no significant differences in superintendent instructional leadership behavior when schools were categorized by the fidelity with which the IPI was implemented. Data analyses were confounded by problematic data in one of the reporting categories. However, post-hypothesis testing produced intriguing patterns. Qualitative data analyses produced evidence of noticeable differences in leadership dispositions and behavior when fidelity of implementation groups were compared. While no statistically significant differences were found in leadership behavior across Fidelity of Implementation groups, interesting patterns in means for each leadership behavior emerged when examined across Fidelity of Implementation group when those groups were disaggregated by the presence or absence of an assistant superintendent. Poverty rates were found to be positively and significantly associated with four of the five leadership behaviors analyzed by this study. Qualitative analyses indicated that superintendents from schools classified as High Fidelity Implementers reported a sense of responsibility for outcomes as well as greater influence than subordinates in the area of change implementation. The character of collaboration varied across implementation categories with High Fidelity Implementers reporting greater preparation, respect for other stakeholders, and clearer vision for outcomes than their peers. High Fidelity Implementers were more inclined than their peers to grant freedom to principals and teachers to make decisions, formulate plans, and put those plans into action, once the superintendents were sure that the other professionals were clear on the goals and constraints. Interestingly, all superintendents were able to articulate how they implemented change but were challenged to articulate their own leadership skills. Data analyses did not allow statistically significant differences to be determined in how superintendents lead. However, post-hypothesis analyses did find a link between low SES and collaborative setting of goals, of non-negotiable goals for achievement and instruction, monitoring of progress toward those goals, and providing resources for goal attainment. Further, the presence of an assistant superintendent decreased the degree to which superintendents felt closely tied to instructional leadership. Qualitative data analyses produced findings that implementation of the Instructional Practices Inventory (IPI) process with high fidelity fostered second-order change in teaching practices and was closely aligned to instructional improvement recommendations by Marzano and Waters (2009). Superintendents who implemented the IPI with high fidelity were more likely to communicate respect for others with whom they collaborated, were more likely to thoroughly prepare when they facilitated collaboration, felt a greater sense of burden for positive student outcomes, and were more likely to empower others to act once a vision and set of boundaries were clearly established. Superintendents uniformly expressed some level of resentment at the mandated goals of No Child Left Behind. However, they did so while simultaneously requiring that those goals be adopted at the district level, even when a collaborative planning process was used, and insisting on articulation of supportive building level goals. Superintendents did not comment on the irony of this situation and uniformly expressed pride when goals were met. When asked directly about power, influence, and their own leadership skills, superintendents' responses were ambiguous. While some superintendents spoke frankly about differences in power and influence, others reported substantial equality among all of their districts stakeholders. Superintendents spoke not of their leadership skill but of personal qualities such as trustworthiness that supported their success. However, analysis of the stated actions of High Fidelity Implementing superintendents indicated that these superintendents engaged in aspects of research-based transformational leadership (Leithwood & Jantzi, 1990) and leadership through relationships (Ogawa & Bossert, 1995).