This study, one of four cohort studies, was designed to determine the relationship between student performance and leadership practices as perceived by principals and selected site-based decision making committee members of middle schools in the Region 5 Education Service Center area of Texas. Using the Leadership Practices Inventory developed by Kouzes and Posner, the study compared the perceptions of middle school principals and selected observers regarding leadership practices. These leadership ratings were compared to student achievement for each campus in the study. In addition, the study examined if selected demographic variables impact the perception of leaders and observers in regard to leadership. For schools in this particular study, there was no direct correlation between perceived leadership practices of the principals and student achievement as measured by the Texas Assessment of Knowledge and Skills (TAKS) tests. This was true of total LPI scores and also of each individual practice measured by the LPI. The data revealed that principals in the study rated themselves higher as a group than their observers rated them on the LPI. This trend was consistent for the total instrument and for each leadership practice. There were also similarities, however, between the two groups. The practice Model the Way had the highest mean for both groups, while Inspire a Shared Vision and Enable Others to Act were rated lower by both groups of respondents. A researcher-developed questionnaire was used to gather demographic information about each respondent. Years of experience, age group, gender, and ethnicity were all studied to determine if they had any effect on responses. The results indicate that there were some differences when principals and observers were studied separately, but these differences were minimized when the two groups were combined. Of particular interest was the fact that younger principals and less experienced principals rated themselves more conservatively than their older and more experienced colleagues. Younger observers and less experienced observers, however, had a propensity to rate their leader high when compared to older observers and more experienced observers. Neither ethnicity nor gender had an impact on leader ratings.