This paper presents the results of the Structural Equation Model (SEM) analyses of the surveys of key staff (i.e. senior and middle school leaders) in a sample of improved and effective secondary schools in England (n=1,167) and Hong Kong (n=179). The two datasets were based on similar theoretical perspectives on school leadership. Focus The analyses aimed to identify the similarities and differences in key staff members’ perceptions of the impact of their principals on school improvement across the two countries. Analytical Framework In educational research, model building enables ‘the systematic study of underlying concepts in a particular research context and the consideration of the relationships between them’ (Silins & Mulford, 2002). In this study our intention was to explore potential direct and indirect effects of different features of leadership in predicting change in school structures and processes and pupils’ academic outcomes. Methods First, exploratory factor analysis and then confirmatory factor analysis were conducted to test the factors identified as important in the literature and to examine relationships among different dimensions of leadership practice. We then explored country-specific SEM models and linked them to school level data on pupil academic outcomes (value added measures over a three-year period in both countries). By developing models separately we were able to establish whether the relative influence of school leadership on pupil outcomes differs between the two countries. Findings A four-level tentative structural model was identified for the English sample, comprising leadership qualities and practices (Level 1), distributing leadership tasks (Level 2), school conditions and cultures (Level 3), and changes in pupils’ academic outcomes (Level 4). Similarly, a four-level tentative structural model was also identified for the Hong Kong sample, suggesting that principals’ leadership is influenced by their values, qualities and efficacy (Level 1) and that principals influence pupils’ academic outcomes (Level 4) through prioritising internal and external policies (Level 2) and improving school conditions for teaching, learning and staff development (Level 3). There are similar interrelationships on leadership practice in the two SEM models, especially those relating to setting directions for school improvement, developing teachers, and improving school conditions for teaching and learning. The findings also indicate strong similarities between the two countries in the links between leadership practices, trust and the improvement in intermediate outcomes of school improvement. Major differences relate to the influence of accountability and resource management on the processes of school improvement in the two countries.

Silins H, Mulford WR, 'Schools as learning organisations: the case for system, teacher and student learning', Journal of Educational Administration, 40, (5) pp. 425-446.