How does leadership affect student achievement? Results from a national US survey

Abstract Using survey responses from a national sample of US teachers, this paper provides insight into 2 questions: (1) Do 3 specific attributes of leadership behavior – the sharing of leadership with teachers, the development of trust relationships among professionals, and the provision of support for instructional improvement – affect teachers' work with each other and their classroom practices? and (2) Do the behaviors of school leaders contribute to student achievement? We tie this investigation of school leader behaviors to 2 additional factors that have also received increasing attention in research because they have been shown to be related to student achievement: professional community and the quality of classroom instruction. Our analysis provides an empirical test of the notion that leadership variables are positively related to student learning. It also suggests that both shared and instructionally focused leadership are complementary approaches for improving schools. Keywords: leadershiptrustinstructionstudent achievement Notes 1. This analysis was supported by a grant from the Wallace Foundation. The funding agency bears no responsibility for the contents of this paper. 2. See Ogawa and Bossert (1995) for a discussion of the relationship between broader patterns of leadership as an organizational property. 3. This view of shared leadership reflects an emerging consensus among current scholars and distinguishes our approach from scholars who blend shared leadership with instructional leadership (Marks & Printy, 2003). 4. Principal surveys and qualitative data around instructional leadership are being analyzed in other papers and reports (see Louis et al., 2010, for the broadest review). 5. While teachers are associated with principals, we do not have individual student achievement scores and are therefore unable to use a hierarchical analysis like hierarchical linear modeling. 6. Two other factors emerged on our initial efforts to examine instruction, which are described in a previous paper (Wahlstrom & Louis, 2008). While not reported here, we initially attempted to use them in this investigation. However, neither of the other instructional factors is associated with the measure of student achievement used in this paper, nor with other measures available to us. We therefore have chosen not to report these findings. 7. Because of changes in state tests and state cut scores on stable tests, and the inability to obtain tests at the student level from many states, we were unable to look at changes in test scores over the 5-year period, which was our original intention. 8. The RMSEA is .45, which is considerably higher than the suggested value of .05. 9. We have, elsewhere, drawn on our case study data to illuminate some of the differences between instructional leadership in elementary and secondary schools (Louis et al., 2010, pp. 74–90), but the interview data from teachers are relatively thin.