Muijs, D. (2011). Leadership and organisational performance: From research to prescription? The International Journal of Educational Management, 25(1), 45-60.
The purpose of this literature review and informal meta-analysis was to examine the current state of research regarding the impact of leadership on student outcomes, the prominent leadership activities correlated to those outcomes, and the general state of the research base upon which leaders are leading and leadership programs are teaching. The study was done because the interest among political leaders and researchers in school leadership has grown without sufficient examination of the quality of the data used in making prescriptions for schools and their leaders.
The informal meta-analysis resulted in four categories: the extent of leadership effects on student outcomes, the types of leadership that affect student outcomes, leadership development, and the limitations of the current research base. Muijs identified the massive increase in government expenditure for school leadership development around the globe and pointed out that
though some research suggested a correlation between leadership and school performance, that link was always mediated (usually by shared vision and goals).
In most studies, student outcomes were narrowly defined as standardized test results and methods were limited (regression models) and tended to find only modest to weak (or no) relationships. Meta-analysis (Hattie, 2005) showed an average effect size of 0.52 – a significantly lower effect size than those associated with direct instruction, student feedback, or student training in cognitive strategies. Leadership was consistently seen to have indirect not direct impact on student learning. Muijs showed that leadership accounted for 3 to 5 percent of school achievement variance, but 25 percent of the variation of all school factors. Leadership effects were significant, but only indirectly so on student achievement. Context, especially transnationally, was also identified as important.
The study also found that the types of leadership identified in the research suffered from two main problems: over-reliance on dualism and too readily moving from limited research to prescriptive solutions. Key examples were the juxtaposition of transactional and transformational leadership, distributed and charismatic leadership, and instructional and administrative leadership. Though evidence existed that highlights the benefits of transformational, distributed and instructional leadership, evidence also shows the necessity of transactional, charismatic, and administrative leadership in the running of successful schools.
The current research on leadership development insisted that leadership is a purely learned quality and relies on nothing innate. Professional development (PD) and leadership development have expanded greatly in response to this assumption, an assumption, the researcher pointed out, that ignores findings in psychological research that suggests at least some genetic determinism in personality traits.
Recent data on PD suggested no link between leadership development and school performance.
The only clear finding in the research (and the one most likely to be ignored) on leadership development was that leaders overwhelmingly prefer hands-on, activity-based PD programs than generic courses or in-service days.
Finally, Muijs found that the current base of research on school leadership was much weaker than the claims usually made about it. Educational research in general, suggested the researcher, too readily jumped to prescription and the reliance upon dualistic models of leadership only exacerbated this situation. This literature review suggested a need to: develop non-dualistic, more nuanced models of leadership, strengthen and diversify research methods and practices, and spread research on school leadership transnationally to account for context.