Local education authorities and student learning: the effects of policies and practices

Abstract

This article addresses an issue that has not been well explored in empirical research, namely whether local education agencies (districts) have an impact on student learning. We assumed that local district effects on learning would be largely indirect, mediated by how teachers work together in schools (in professional communities) and the quality of instruction that is provided. Based on the literature, we also assumed that promoting data-driven decision making was an insufficient stimulus for student learning, and we therefore chose to examine another current policy strategy that is being widely adopted by local authorities: the development of networks for learning among schools. Using survey data and structural equation modeling, our results suggest that the development of networks has a positive relationship with instruction and subsequent learning, while district emphasis on learning targets and data use has a negative relationship. The discussion offers a number of interpretations of the findings, and suggests further arenas for inquiry.

Keywords:

• data-driven decision making
• leadership
• districts
• professional community
• networks

View correction statement:

Corrigendum

Notes

 1. Our research was funded by the Wallace Foundation. This article presents only a small window into the large and complex project.

 2. Throughout this article, we will use the general term, local education agency (LEA), to denote recognized units that exist between the national/state/provincial governments and have some responsibility for the management of individual schools. This could include Swedish municipalities as well as Dutch school boards, which are private but publicly funded entities that do not have a geographically defined catchment area. We will also use the term district, which is the name for LEAs in North America and some other countries.

 3. Professional community is closely associated with organizational learning, and the term professional learning communities has become a common short hand among practitioners and some scholars

 4. Others argue that the paradigm for effective teaching has fundamentally changed, in large measure because of the demand for adaptive learning capacities in the workforce and society in general (Cheng & Mok, 2008).

 5. Frequently cited investigations of leadership effects often use only one type of school (Bryk & Schneider, 2002; Cascadden, 1998; Harris, 2002). Those that use samples from all levels are based on a small number of cases (Marks & Printy, 2003), or convenience samples from a single district (Leech & Fulton, 2008; Leithwood & Jantzi, 2000).

 6. Because there were seven schools where principals did not respond to the principal survey, seven assistant principals' responses were alternatively used for the schools.

 7. According to Marsh, Hau, Balla, and Grayson (1998), it could perform best especially when a factor has relatively many items (e.g., more than 12 items).

 8. This measure was shown in a previous paper (Louis, Dretzke, & Wahlstrom, 2010) to be associated with student achievement.

 9. The correlation coefficient of language arts proficiency between 2005/6 and 2006/7 was .936 (p < .01). We conducted preliminary analyses with mathematics achievement, and similar results emerged.

10. The Sobel test is based on the assumption that the mediating effect of a x b is normally distributed. In this regard, bootstrap methods, suggested by Shrout and Bolger (2002), can be alternatively used to test mediating effects.

11. We used a SAS program to calculate this.

12. We admit that using indicators with cross loadings is not recommended in general. Nonetheless, consistent with the primary purpose of the study, they were included in the analysis after assessing the two factors' interim reliability with and without those indicators (Pett, Lackey, & Sullivan, 2003).

13. We have not presented the analyses (i.e., multiple group SEM) separately for elementary and secondary schools because of the small Ns. These remarks are intended as suggestive of the need for further investigation.