Acknowledgements. The author wishes to thank Professor Daniel Muijs, Professor of Education at the University of Southam
CENTRE FOR THE STUDY OF MARKET REFORM OF EDUCATION
Taking a lead
Collaborative overreach: how to access the leadership premium Why collaboration probably isn’t key to the next phase of school reform
James Croft
Research report 9
Research report 7 James Croft
Centre for the Study of Market Reform of Education 23 Great Smith Street, Westminster, London, SW1P 3BL • www.cmre.org.u
CENTRE FOR THE STUDY OF MARKET REFORM OF EDUCATION i 23 Great Smith Street, Westminster, London, SW1P 3BL • www.cmre.org.uk
Taking a lead how to access the leadership premium James Croft
About the author James Croft is Executive Director of the Centre for the Study of Market Reform of Education. He has authored or co-authored a number of reports for the Centre and its partners, including most recently, ‘Collaborative overreach: why collaboration probably isn’t key to the next phase of school reform’ (2015).
Acknowledgements The author wishes to thank Professor Daniel Muijs, Professor of Education at the University of Southampton, and Director of Research at Southampton Education School, for his valuable comments and suggestions, and Gabriel Heller Sahlgren, Director of Research at The Centre for the Study of Market Reform of Education, for his rigorous editorial input. Finally, this project would not have been possible without the kind sponsorship provided by OneEducation Ltd.
CENTRE FOR THE STUDY OF MARKET REFORM OF EDUCATION First edition published 2016 by The Centre for the Study of Market Reform of Education Ltd. Copyright © The Centre for the Study of Market Reform of Education Ltd 2016 The moral right of the author has been asserted. All rights reserved. No part of this publication may be reproduced, stored in or introduced into a retrieval system, or transmitted in any form, or by any means (electronic, mechanical, photocopy, recording, or otherwise) without the prior permission of the publisher. Any person who does any unauthorised act in relation to this publication may be liable to criminal prosecution and civil claims for damages. Every care has been taken that all information was correct at the time of going to press. The publisher accepts no responsibility for any error in detail, inaccuracy or judgement whatsoever. Typeset and design by Pete Johnson, Great White Designs Ltd The Centre for the Study of Market Reform of Education 23 Great Smith Street, Westminster, London SW1P 3BL www.cmre.org.uk
[email protected] @cmr_ed
iii
Contents Executive summary
v
Introduction: the leadership premium
1
Is it really all about the leader?
2
Leadership variables only modestly related to outcomes 3 Are we asking the right question?
4
Mediators of leadership influence identified in leadership effectiveness research 5 Parallels in the economic literature on the effects of ‘de-centralised decision-making’
6
Contextual factors and reciprocal effects
9
The theory-laden nature of leadership studies and its consequences
12
A case in point: the transformational theory of leadership
14
Oppositional research vs theoretical integration: why it’s important to get the effects model right
16
Moving forward with the reciprocal effects model: challenges for research
18
Conclusion: what does this mean for policy?
20
References
23
iv
Executive summary
Executive summary • As plans for whole system structural reform have developed, much of the government’s education reform strategy has come to turn on its being able to capitalise a leadership premium. • Yet despite a massive literature that has built up around the subject, the convictions of politicians, regulators, and professional development bodies, about what leadership looks like, how it contributes to school improvement, and what really counts (personalities, traits, or practices?), on how to develop leadership, and how it is related to context, are not well-grounded. • Those engaged in school effectiveness research (SER) share with them a common set of assumptions, to mutually reinforcing effect. • The most obvious of these assumptions, bizarrely, relates to the tendency to over-state the importance of leadership as a standalone factor in improving attainment, to the neglect of understanding of how it interacts with other key school factors. • This tendency to over-statement of leadership’s direct importance to academic outcomes is largely the result of adopting theories and research strategies incapable of questioning it. • In relation to the literature as a whole, few studies have attempted to quantify the contribution of leadership. Those that do find leadership variables only modestly to weakly related to pupil outcomes. Yet this has not v
Taking a lead
checked enthusiasm for theories of leadership assuming direct effects. • The realisation is slowly dawning on leadership effectiveness researchers that they may not be asking the right question in respect of the influence of leaders on outcomes, with an increasing and substantial minority of studies positing a number of school-related factors as mediators. • Though not without design and methodological limitations of their own, estimations of mediate, indirect effects of leadership work with the theory that leaders help create the conditions under which teachers may be optimally effective. Mission and goal setting, variables related to the setting of the curriculum, and the provision of instructional guidance for teachers, are identified as important means by which they exercise their influence. Leadership matters for determining both the motivation of teachers and the quality of teaching in the classroom. • The importance of these mediators is corroborated in the economic literature by a number of studies of ‘school autonomy’ – the result of reforms to governance often alternatively referred to as ‘decentralised decisionmaking’ or ‘school-based management’. • In addition to changes to leadership and management structure, this research also indicates that the scope to shape curriculum and instructional method, and the motivation of staff (specifically through the use of appropriate pay and conditions incentives), are important for raising academic achievement. In more autonomous school contexts the importance of these mediating factors is accentuated. • Yet problems with the underlying effects model persist. This research suggests leadership decisions as the sole cause of change in organisational performance – a vi
Executive summary
conception which clearly does not do justice to the complexity of heads’ relationships and interactions with staff, pupils, parents, and others, in context. • Considering these issues, some have begun to posit that educational outcomes and school environment are equally strong determinants of principal leadership behaviours as principal leadership behaviours themselves are of educational outcomes. • This ‘recursive’ or ‘reciprocal effects’ model has clear resonance with earlier ‘contingency’ theories of how leadership works, stressing that organisational performance depends as much on the favourability of the situation for the leader as it does on the effective exercise of his/her influence over followers. Substantiating this theory becomes more viable with the development of this effects model, with the availability of statistical tools for testing the notion of leader ‘fit’, and in what should be a more receptive policy environment. • The greater part of the leadership research community however, still seems reticent to move forward with this agenda. In that the leading question for social scientists is not ‘what works?’ but rather ‘what works for whom and under what conditions?’ this is surprising. • Among more prosaic reasons relating to resources and capacity, it is undoubtedly the case that, wittingly or otherwise, underlying commitments to different effects models supporting various research theories, and reluctance to work together to develop a fuller understanding of the nature of leaders’ influence in context, have often not helped move things forward. • ‘Great man’ or heroic theories of leadership, and the ‘trait’ and ‘behaviour’-based permutations of these that followed – all bolstered by the assumption of direct vii
Taking a lead
effects – have persisted in various forms long after they should have done – most recently in the form of transformational leadership theory. • Here organisational success (and failure) is attributed to individual competence on the basis of virtually no evidence at all. With the vigorous support of their advocates, they have, in turn, exercised with their popularity, an unhelpfully normative influence on practice. • This has been encouraged by a tendency within the field to pit different aspects of leadership – hierarchical / distributed, transactional /collaborative, for example – against one another, rather than seeing them as integral to what we mean by leadership. • Thinking in terms of competing theories or paradigms is encouraged by the high stakes nature of the discourse around school effectiveness and accountability, and has led to widespread over-prescriptivity in relation to what is at best a limited research base. • While there is some evidence of theory convergence around the ‘leadership for learning’ model, and increasing acknowledgement of the adaptive nature of leadership in relation to context, there is still too often a gratuitous preference for new theory among researchers. • Meanwhile, best practice case study approaches, involving interviews and the collection of documentary evidence, and questionnaires – methods susceptible of a range of biases – continue to predominate. • Researchers are slow to collaborate and rise to the challenge of the kind of longitudinal study that would be useful in assessing reciprocal effects, and appear to have no interest in undertaking studies of quasi-experimental design, let alone full experimental studies featuring randomised assignment to treatment and control groups. viii
Executive summary
• Such undertakings are crucial if leadership effectiveness research is recover its credibility and move towards a useful understanding of the complexities of school leadership, and what makes for success. • Because of the over-theorised nature of leadership studies, and the limitations of both research design and method that have characterised the field, atheoretical re-description of actual leadership decision-making and practice, in context, and over time, is a necessary first step. If it can be established, over time, how far particular patterns are influenced by context or are constant, then there is a route to assessing the impact of those practices, and thus to a more viable strategy for replicating good practice. • There is, in short, a long way to go in leadership studies. • These findings have a number of policy implications. • First, we simply do not know enough about what particular practices are impactful, learnable and transferrable, to require participation in leadership development programmes. There is no robust evidence to support claims that professional qualifications make a difference to the quality of headship. • In that it seems likely that emerging leaders are as much shaped by the opportunities and challenges of the contexts in which they find themselves, it makes sense to shift the locus of leadership identification and development to the schools level, to be led by head-teachers, in situ. Happily, this is broadly in accord with the trajectory of recent government policy. Moves towards the process of acquiring the necessary skills becoming more practical and work-based are also sensible. • If school leaders themselves cannot identify potential leaders, and design and develop opportunities for ix
Taking a lead
leadership development in school, there is certainly no basis for believing this to be within the skill set of central government. • Incentivising them to do so is therefore the right direction for policy. The best way of doing this is to work towards a model of headship training and continuing professional development that is leader- and demand-led. • This should essentially consist of two elements: 1) a headteacher shadowing scheme based on the observation of leaders in the school context, supported by 2) course content oriented to improving research literacy and building and refreshing knowledge of what works. • Mentoring and peer-to-peer support would be provided by leaders of similar schools according to demand from those in need of advice, strategy, and support. This consultancy model would get around the misleading sense that an intervening ‘hero head’ is in some way taking responsibility for the outcomes of the advice he/she has given. • The present government, however, seems at present to be set on re-investing in hero-heads and a suite of leadership development qualifications to be provided by the National College of Teaching and Leadership. This is in accord with the government’s general counter-veiling tendency to override incentives-based strategy and take decision-making back to the centre – as witnessed by its ongoing control of the recruitment of academy sponsors, and increasing centralisation of decisionmaking in respect of national educational standards and assessments, regulation, and accountability. • In that the research suggests it is leaders’ discretion in the key areas of change to leadership and management structure, motivating staff through mission and goalsetting and appropriate incentives in pay and conditions, x
Executive summary
and curriculum and pedagogy – and consistent implementation of those decisions – that counts, centralisation policy geared to accessing a leadership premium is not well-advised. • Such reforms threaten to stymie the potential of its school-based management reforms by taking over recruitment at the top tier and reducing leaders’ scope to take decisions and effect strategies in the areas identified to have bearing on academic improvement. • While the evidence base in this respect, as in others, is limited and far from robust with respect to causation, and researchers have a long way to go before they are capable of articulating in any detail the specific decisions and practices, in context, that shape the conditions for effective learning, it does at least give us a starting point. • To progress from here, what we really need in our leaders is the ability attend to and to follow the evidence – for our leaders, in short, to become followers.
xi
Introduction: the leadership premium Leadership has long been regarded as a key factor in organisational effectiveness (Yukl 2006), but interest in educational leadership has increased more markedly in recent decades with the growth of school-based management in many countries (Fullan 2004; Hallinger 2010, 2011; Muijs 2011b; Pont, Nusche & Moorman 2008). Encouraged by a burgeoning ‘school effectiveness’ literature, coming to a point on the contribution of leadership to improving pupil outcomes (Goldstein & Woodhouse 2000; Muijs 2011b; Hallinger & Huber 2012), policy investment in school autonomy has been accompanied by increased investment in leadership development too, with schemes requiring certification and participation in ongoing programmes of support following in train wherever this reform trajectory has become established. These trends are nowhere more apparent than in England, where the emphasis has increasingly been on whole system structural reform. Accordingly, much of the government’s education reform strategy now turns on its being able to capitalise a leadership premium. In consequence, the importance of strong and effective leadership for improving pupil outcomes is now rarely questioned, despite the fact that the nature of that influence is not well-understood. Despite the massive literature that has built up around the subject, the convictions of politicians, regulators, and professional development bodies, about what leadership looks like, how it contributes to school improvement, and what really counts (personalities, traits, 1
Taking a lead
or practices?), on how to develop leadership, and how it is related to context, are not well-grounded. Reviewing the history of leadership studies in education (and indeed the same may be said of the field more broadly), the assumptions of those that make and shape leadership policy and practice seem, furthermore, rarely to have been substantively challenged. Indeed, all too often, those engaged in school effectiveness research (SER) appear to have shared a common set of assumptions, to mutually reinforcing effect.
Is it really all about the leader? Perhaps the most obvious of these assumptions relates to the tendency to over-state the importance of leadership as a standalone factor in improving attainment, to the neglect of understanding of how it interacts with other key school factors. Undergirding this tendency are issues of theory, research design and methodology, which are explored below. The tendency to over-state the importance of leadership as a standalone factor in improving attainment is largely the result of adopting theories and research strategies incapable of questioning it. Over the course of more than a century, thousands of descriptive studies of successful leaders have been conducted, while comparatively few have attempted to quantify the contribution of leadership, either generally or by type, in respect of improving academic, and other, outcomes (Hallinger & Heck 1996; Marzano, Waters & McNulty 2005).1 Granted that isolating and quantifying the relative importance of this variable in the context of the school as an organisation is difficult to do, it is not impossible. Given the 1 In their review of the quantitative research from 1980 to 1995, Hallinger & Heck (1996a) identified only 40 studies addressing the relationship. Likewise, out of over 5,000 studies addressing the topic of leadership in schools, Marzano, Waters & McNulty (2005) were able to identify only 69 in the period from 1970 that had examined the impact of building leadership on academic achievement according to a quantitative research design.
2
Is it really all about the leader?
importance generally attached to the contribution of leaders, and the resources allocated to identifying, developing, overseeing and incentivising them (a tendency undoubtedly encouraged by the qualitative nature of the greater part of the literature), this lack of interest is disturbing. Leadership variables only modestly related to outcomes Of those few that have risen to the challenge of substantiating a direct effect of leadership on achievement, almost all use some form of cross-sectional, correlational design, often employing surveys or interviews as methods of gathering information. The inherent weaknesses of these designs and methods mean that, despite improvements to aspects of the effects modelling and statistical techniques employed,2 we cannot claim to know much, with any certainty, about direct leadership effects. This makes the discovery that on balance the literature suggests ‘leadership variables are only modestly to weakly related to pupil outcomes’,3 all the more surprising. 2 Initially bivariate in design (i.e. involving simple studies of the relationship between only two variables) (Nettles & Herrington 2007), more recent efforts to do so have progressed to the multivariate, using regression (or multilevel) models to yield an estimate of the percentage of the total variation in students’ scores that is situated at the leadership level, relative to the school level (Hallinger & Heck 1996a). Structural equation modelling is now widely employed also, to correct for background characteristics (usually including prior achievement), and in-school mediating factors (such as teacher behaviour, curricular organization, etc.). 3 Muijs (2011b) refers to D’Agostino (2000); Hallinger & Heck (1998); and Van de Grift & Houtveen (1999) others finding in support are Witziers, Bosker, & Krüger (2003) and Robinson (2007), among many others. Given that some studies of this type find no relationships at all (see Creemers 1993; Leitner, 1994), Scheerens & Bosker (1997) find the evidence base inconsistent. Branch, Hanushek & Rivkin (2012) stand alone in finding marked positive effects, reporting that ‘highly effective principals raise the achievement of a typical student in their schools by between two and seven months of learning in a single school year; ineffective principals lower achievement by the same amount’ (Branch, Hanushek & Rivkin 2013:63). However, there are questions about whether the techniques adopted for estimating principal value-added were as robust as they might be, and whether therefore the effect may be exaggerated. Orr (2013) points out that following its release, several other value-added estimates of principal effectiveness studies were released or published that dealt with this issue as background. They yield somewhat more modest estimates, particularly when trying to isolate principal effects from other factors (Chiang, Lipscomb & Gill 2012; Dhuey & Smith 2014; Grissom,
3
Taking a lead
That this reality has not checked the enthusiasm for theories of leadership based on this effects model is extraordinary. In a recent state of the art review of the literature (446 articles from the key journals published between January 2005 and March 2010), Muijs (2011a) examined underlying causal models and found 55% of the output to be based on direct effects models.
Are we asking the right question? The obvious question to ask at this point is whether we are asking the right question in respect of the influence of leaders on outcomes. Surely it depends on what leaders do? Principal Leadership
Educational Outcomes
Figure 1: Direct effects model
Principal Leadership
Intervening Variables
Educational Outcomes
Figure 2: Indirect effects model
Kalogrides & Loeb 2012). Orr raises further issues concerning information on the quality and completeness of the data set. According to the authors, an ‘unspecified number’ of special education and limited English proficient students were exempted from the tests upon which their estimation of the variance in principal effects was based. In addition, they state that ‘about 15 percent of students do not take the tests, either because of an exemption or because of repeated absences on testing days’ (p. 7), but do not address the likelihood that these absences and exemption rates were greater in high poverty schools, thus possibly contributing to the greater variance in principal effects. Neither do they discuss the impact of student transfers and attrition rates (likely to be higher in poorer contexts). Finally, although the dataset is substantial, combining six years’ worth of staff and pupil background data, pupil achievement data, and observations of 7,420 principals (yielding a total of 28,147 observations/year), the choice of data from 1995 to 2001, when they might have chosen a panel extending to as recently as 2008, is curious, but may also have accentuated the degree of variance. The years covered by the data in this analysis precede federal, state and local assessment-based accountability policies designed to address laxity in testing conditions, alignment of instruction to assessment outcomes, and other inadequately outcome-focused practices.
4
Are we asking the right question?
For reasons that will become clear, this, at any rate, has been the tentative conclusion of those in the vanguard of efforts to quantify the importance of leadership for school effectiveness. The relationship may be indirect, but nonetheless educationally significant. Unfortunately those working to a direct effects model, for whose purposes cross-sectional data may be sufficient, are precluded from identifying potential feedback loops, recognising reverse causality, or addressing changes in the way variables interact – which could be key to understanding how leaders exercise their influence. The weaknesses of cross-sectional research design are especially relevant because the dependent variable of interest is school improvement, which implies the need to measure leadership impact over time (Hallinger & Heck 2011:153; Hallinger & Heck 1996a; Jackson 2000; Mulford & Silins 2003; Ogawa & Bossert 1995). Not only then might research into direct effects on outcomes be misguided, it may in fact be misleading.
Mediators of leadership influence identified in leadership effectiveness research Whatever influence leaders may have on outcomes is now regarded in a substantial minority of studies as, at very least, mediated by a number of school-related factors that are more proximal to the student level, if not also influenced by ongoing and interactive contextual factors (D’Agostino 2000; Hallinger & Heck 1998; Hallinger 2008; Van de Grift & Houtveen 1999; Teddlie & Stringfield 1993). Muijs (2011a) puts the proportion of studies positing mediated effects models at just under 30%. Those applied to estimation of mediate, indirect effects work with the theory that leaders help create the conditions under which teachers may be optimally effective (Hermosilla, Anderson & Mundy 2014). This represents progress at the conceptual level, even if the methodological issues remain 5
Taking a lead
generally the same. A number of studies, accordingly, have found that leadership vision and imparting a strong sense of shared mission are related to teacher effectiveness (D’Agostino 2000; Hallinger & Heck 1998; and Teddlie & Stringfield 1993). Others add to this mission and goal-setting function of leadership, variables related to the setting of the curriculum, and the provision of instructional guidance for teachers (Grissom, Loeb & Master 2012; Leithwood & Riehl 2003), as important ways in which the influence of leaders is exercised. Leadership matters for determining both the motivation of teachers and the quality of teaching in the classroom (Branch, Hanushek & Rivkin 2012, 2013; Fullan 2002; Sergiovanni 1999).
Parallels in the economic literature on the effects of ‘de-centralised decision-making’ Again, few of these studies escape (certainly not entirely) the methodological problems outlined above,4 but they are 4 D’Agostino (2000) is longitudinal (up to a point), uses hierarchical linear modelling (HLM) statistical techniques, and controls for parental socioeconomic status, but is surveybased. Teddlie & Stringfield (1993) is sufficiently longitudinal, uses multilevel modelling techniques, controls for background characteristics and prior attainment, but the generalisability of its conclusions about the importance of principal leadership and the efficacy of different improvement strategies are undermined by the sample size (being based on data from just 16 elementary schools). Leithwood & Riehl’s (2003) overview sets much store by case studies of successful schools indicating that setting ambitious goals is important, and by the conclusions of meta-meta-analyses which paint in broad brushstrokes over methodological issues and are dismissive of publication bias. Fullan (2002: 95) generalises from the literature but pays little attention to methodological issues. Grissom, Loeb & Master (2012) offer a more promising approach, drawing on detailed time-use observations undertaken over a three year period and an impressively long panel of outcomes data to assess the degree to which specific instructional leadership activities predict student achievement gains in schools. Branch, Hanushek & Rivkin (2012) also buck the trend, with their longitudinal analysis of the management of teacher transitions, and specifically, the relationship between the quality of teachers who transition out of a school and the quality of principals. Estimates of principal value-added were undertaken showing substantial variation in quality; and higher quality leadership then related to the quality of teachers (gauged on a similar basis) transitioning out of their schools. For the best principals, the rate of teacher turnover was found to be highest in grades in which teachers are least effective, which the authors providing some support for ‘the notion that a primary channel for principal
6
Parallels in the economic literature Parallels in the economic literature
corroborated, up to a point, by economic research findings on the effects of ‘school autonomy’ – the result of reforms to governance often alternatively referred to as ‘decentralised decision-making’ or ‘school-based management’. These are relevant because, working on the theory that autonomy and appropriate accountability for outcomes incentivise leaders to taking certain kinds of decisions in the interests of maximising success, researchers studying the effect of autonomy reforms might just as well be said to be studying leadership effects, albeit within this framework. Building on a study by Machin & Vernoit (2011) that uses matching methods and long panel data to show significant improvement in academic outcomes following academy school conversions in the years 1997-2010, Machin, with Eyles (Eyles & Machin 2015) examines the responses to a 2014 DfE survey of academies collecting information on changes that may have occurred following conversion, thought to have contributed to this improvement. They rank the responses in order of the percentage of their sample of 23 of them which actually made the particular changes considered in the survey. The most prominent (and most frequently linked to improved outcomes) were changes to leadership and management structure, and curriculum change. The hypothesised importance of changes to leadership and management structure, as also the overall thrust of the leadership effectiveness research positing mediating factors, is substantiated by related findings of research oriented to the analysis of ‘federation effects’. This research suggests that schools within those types of federation most expressly purposed to improving pupil attainment, within a coherent and outcomes-focused organisational structure, are more impactful than those associated under influence is the management of the teacher force’. Note, however, the issues raised by Orr (2013), detailed above, n. 4.
7
Taking a lead
more informal arrangements, and comparator ‘stand-alone’ schools (Chapman, Muijs & Collins 2009; Chapman, Muijs & Sammons 2010; Chapman, Muijs & MacAllister 2011; Chapman, Muijs & MacAllister 2012; Muijs, Reynolds & Chapman, forthcoming). In this research autonomy reforms incentivising leaders towards taking key decisions are enabled by structural reforms at the institution level. The extent of central steering is thought to explain chain effectiveness. It is only when they enter into formal federation – involving shared governance and leadership and integrated systems of quality control, etc. – that schools in the chain begin to make real and consistent gains (Croft 2015:26-30). Eyles & Machin’s hypothesis in regard to curriculum change is corroborated by Hanushek & Woessmann’s (2010:25-26) report on findings from cross-national studies making use of data from international assessments (and questionnaire responses giving information on school background characteristics), to test the impact of autonomy as implemented in several different decision-making areas.5 They find that the general pattern of results is that students perform significantly better in schools that have autonomy at the delivery process level [including in relation to curriculum and instructional methods], as well as in respect of personnel decisions (Fuchs and Woessmann 2007; Woessmann 2003b; Woessmann, Luedemann, Schuetz & West 2009). Students perform better if their teachers are incentivised to perform and may select appropriate teaching methods. This appears further reinforced by a pattern of results across these 5 Cross-country comparative approaches based on data from such sources have a number of advantages. They can provide richer datasets, with a much larger variation than usually available within any country; they circumvent selection issues that plague within-country identification by using system-level aggregated measures; and can reveal whether effects are systematically heterogeneous in different settings. As with other non-experimental studies, however, they cannot control for unseen contextual variables, such as cultural and political influences. In addition the available achievement data was, at least initially, mainly cross-sectional, though recent work in a few countries has built within-country follow-ups into the PISA testing.
8
Contextual factors and reciprocal effects
studies, also reported by Hanushek & Woessmann (2010), indicating that autonomy in pay and conditions is positively associated with student achievement in external exam-based accountability systems. Measuring performance against external outcomes-focused measures thus ties in with the mission- and goal/target-setting function of leaders too. Hanushek, Link & Woessmann (2013) later substantiate these patterns in relation to autonomy in regard to determining academic content and personnel decisions, while adding discretion over budget allocations as another factor positively associated with student achievement. These are important findings for leadership effectiveness researchers to digest. They suggest that the impact of leadership increases with the degree of autonomy afforded schools. But, perhaps more pertinently, they also indicate that it is not so much the leader, but the decisions he/she takes, that are important.
Contextual factors and reciprocal effects Nevertheless, studies considering mediators only tell half the story. A substantial body of work was undertaken on the assumption of direct or indirect influence on outcomes (as modelled above) in the years 1980-1995. Looking back over this period, Hallinger & Heck (1998) conclude that it had shed little light on ‘the most important theoretical and practical issues concerning the means by which principals achieve an impact on school outcomes and how contextual forces influence the exercise of leadership’. In his recent state of the art review Muijs (2011a) finds this situation largely unaltered. As with direct effect models, the unidirectional nature of the causal ordering in indirect effect models suggests leadership decisions as the sole cause of change in organisational performance (see below, Hallinger 2008; Hallinger and Heck 1996, 1998; Pitner 1988). The notions that contextual 9
Taking a lead
responses to leadership initiative might in turn influence leaders’ next moves, or that the demands of context may bring forth leadership and shape its expression, are precluded from discussion. This conceptualisation clearly does not do justice to the complexity of heads’ relationships and interactions with staff, pupils, parents, and others, in context (Bossert et al. 1982; Bridges 1982; De Maeyer et al. 2007; Hallinger & Heck 2011; Meindl 1995). As Hallinger & Heck (1998) argued – and later, more strenuously, Hallinger (2008) in the light of the discovery of the persistence of the use of the direct effects model, mounting empirical data highlighting its deficiency, and the shortcomings of indirect effects models – it is more likely that, as Muijs puts it, ‘educational outcomes and school environment are equally strong determinants of principal leadership behaviours as principal leadership behaviours themselves are of educational outcomes’. The so-called outcomes of leadership variables may also be independent, because leaders are likely to change their behaviour as a result of the performance of their followers. In other words, ‘the leader does not simply shape the organisational culture and environment leading to enhanced outcomes, as is often assumed’ (2011b). Following Pitner (1988) and Hallinger and Heck (1996, 1998), Hallinger (2008:17) models the relationships thus: Principal Leadership
Intervening Variables
Educational Outcomes
Figure 3: Recursive or reciprocal effects model
This ‘recursive’ or ‘reciprocal effects’ model is clearly supported by earlier ‘contingency’ theories of how leadership works, 10
Contextual factors and reciprocal effects
stressing the role of context (Fiedler & Chemers 1967; Fiedler 1971), and yet despite the long recognition of its promise by those in the vanguard of quantitative research on leadership effects, the wider SER community has been slow to respond. This is not because Fiedler’s ideas were discredited. Contingency theories had clear insights to offer but gained only limited traction among researchers at the time, partly for political reasons (discussed below), but mainly for want of this kind of clear conceptualisation of leadership effects and how they work, and of the statistical tools for testing the notion of leader ‘fit’, which seemed integral to assessing the theory’s utility. Proponents of the theory at the time could thus do little to substantiate and develop it, even had policymakers been more receptive. These impediments now removed, it is surprising to discover how few studies have been advanced in more recent times on the reciprocal effects model. In Muijs’ (2011a) analysis of the previous five years’ output, only 16% of papers on educational leadership in general had utilised the model, and only 9% of quantitative studies – and this against the backdrop of general interest in contingency theories of leadership tapering off dramatically over the course of the decade as a whole (Day & Antonakis 2012:7-9).6 The latter trend may be partly accounted for by the rise in the number of articles taking broader contextual approaches; nevertheless, in that the leading question for social scientists is not ‘what works?’ but rather ‘what works for whom and under what conditions?’ this apparently slow progress towards a more refined understanding of contextual and situational leadership is surprising. Researchers have barely begun with the difficult tasks of identifying potential feedback loops, recognising reverse causality, and addressing 6 Gardner et al. (2010) note that only 1% of articles published in Leadership Quarterly 2000-10 were focused on contingency theories.
11
T Taking a lead
changes in the way variables interact. Understanding why even theoretical progress towards this model has been so slow, and the implications it has had for our understanding of educational leadership and for policymaking, is critical for the progress of school reform.
The theory-laden nature of leadership studies and its consequences The reasons for the lack of progress in education leadership studies generally, and school leadership in particular, are clearly complex, but as outlined above, have to do primarily with the theory-driven nature of the choice of effects model, and accordingly the research methods adopted. To be sure, limited access to sufficient quality data; the initial lack of appropriate tools for statistical analysis (De Maeyer et al. 2007; Kyriakides et al. 2010; Reynolds et al. 2014; Sammons & Luyten 2009); and the cost associated with properly longitudinal research (Muijs 2011a) have all played a role in inhibiting the development of deeper understanding of what constitutes effective leadership and how it works. The rapid growth in the number of researchers and studies over the past 30 years has also made it a challenge for researchers to adopt cumulative approaches to the knowledge base (Reynolds et al. 2014:202). Nevertheless, wittingly or otherwise, underlying commitments to different effects models supporting various research theories, and reluctance to work together to develop a fuller understanding of the nature of leaders’ influence in context, have often not helped move things forward. It has been widely recognised that for much of the last century the assumption of a direct relationship between leadership and outcomes that supported successive ‘great man’ or heroic theories of leadership from Thomas Carlyle on, also slowed progress through its ‘trait’ and ‘behaviourist’ permutations such that the leadership 12
The theory-laden nature of leadership studies The theory-laden nature of leadership studies
research community only reluctantly conceded the role of context and contingency in shaping leadership practice in the early seventies. Trait theories persisted long after it became apparent that identified associations with measures of leadership effectiveness could not be replicated, and have obstinately recurred ever since, in various forms (as exemplified by the rise of transformative theories discussed below). Likewise researchers seeking to identify generically effective leader behaviours were slow to come to terms with contradictory findings in respect of the efficacy of different behavioural ‘styles’ of leadership for outcomes, and slow to recognise that the reasons lay in the susceptibility of field observations, surveys and questionnaires to confirmation bias (Grissom, Loeb & Master 2012; Muijs 2011a). Both trait and behaviourist researchers ‘gave little thought to the specific role demands of leaders, the context in which they functioned, or to differences in the dispositions of leaders or followers’ (Day & Antonakis 2012; House & Aditya 1997), and as a result were dealt what should have been a critical blow by Fiedler’s argument (Fiedler & Chemers 1967;, Fiedler 1971) that organisational performance depends as much on the favourability of the situation for the leader as it does on the effective exercise of a leader’s influence over his/ her followers. As mentioned above, the theory was in a sense ahead of its time in that it suggested not only an effects model that was beyond the reach of researchers to test, but also a direction in policy that entailed significant devolution of responsibility for identifying and developing leaders to the school level – which to date had enjoyed little traction in central government. The emphasis on leader ‘fit’ suggested to school effectiveness researchers and policymakers alike little cause for optimism in respect of whether governments’ might themselves find the leverage to realise any premium that might be associated with the contribution of leadership. 13
Taking a lead
A case in point: the transformational theory of leadership Little wonder perhaps then that the bracing theories of Bass and his associates (Bass, 1985, 1998; Bass & Avolio, 1994; Hater & Bass, 1988), and others promoting visionary and charismatic understandings of leadership, gained such ready reception. While previous theories had essentially been about the mutual satisfaction of leader– follower transactional (i.e., social exchange) obligations, ‘transformational’ theories posited that ‘idealized and inspiring leader behaviours’ enabled followers ‘to transcend their interests’ in pursuit of the greater good (Day & Antonakis 2012:11). Though transformational theory, in its emphases on leadership personality and intrinsic leader qualities, has clear resonances with the ‘heroic’ model, and in many ways represented a step backwards in respect of its simplistic conception of the relationships with followers and to outcomes (Grint 2011), perhaps unsurprisingly it has proved popular in education, where a strong sense of moral purpose is seen as important. Unsurprisingly, the theory has been no less popular among researchers, developing into the single most dominant leadership paradigm of the period 1990 to 2010 (Gardener et al. 2010). This despite the fact that the qualities described in the literature on transformational leadership are decontextualized, ahistorical, and merely understood to be correlated, rather than determinative for outcomes, and that ‘organisational success’ is attributed ‘to individual competence on the basis of virtually no evidence at all’ (Grint 2011:9). The influence of the theory proved to be a great stimulus to what had previously been a flagging leadership consultancy market, and exercised a marked influence on policymakers’ expectations and upon the allocation of resources to 14
A case in point: the transformational theory A case in point: the transformational theory
programmes. But it was of course, far from a panacea. While leadership behaviours can be developed through CPD activities to a certain extent, the charismatic elements in the traditional definitions of transformational leadership have proved hard to engender if not already present in the personality of the leader, leading to a limited pool of applicants who are capable of leading organisations in this fashion. In a further blow to the efficacy of this theory, even where strongly charismatic personalities are involved, failure may be highly likely if organisational ‘fit’ is lacking – as witnessed by the mixed results of successive government ‘Super Head’-type initiatives. Considering the influence of transformational theories of leadership, Muijs (2011a) remarks that while ‘managing change, and, where necessary, instigating change are important’, the preoccupation with change in the research has ultimately led to ‘a situation where every new head feels s/he has to make changes whether they are necessary or not, purely to demonstrate leadership’. By focusing overmuch on one aspect of what leadership involves, many heads may also have overlooked the importance of a steady hand – that there is a right time for stability too. Arguably against a background of increasingly centralised accountability and overweening regulatory demands, it’s now more important than ever that we get robust and independently minded heads capable of building institutional identity. Suffice to say that this protracted theoretical debate has not helped the development of useful models of what consistently effective leadership looks like and how it works, and thus neither has it helped progress understanding in respect of the key questions of how effective leadership practices may be learned and/or transferred. Those who have worked on surveys of this history have been unable to distil more than a very limited, and indeed flawed, consensus answer to 15
Taking a lead
these questions. Definition of ‘leadership’ essentially boils down to ‘the capacity to influence others’, and ‘leading’ as ‘a process of influence’. The reliance of these definitions on direct and indirect effects models is not difficult to discern. Nor, for that matter, is their failure to accommodate the more awkward insights of contingency theory and, latterly, of the ‘recursive’/‘reciprocal effects’ model.7
Oppositional research vs theoretical integration why it’s important to get the effects model right There is evidently a relationship between the theoretical paradigm adopted and the underlying effects model that fuels misunderstanding of how leadership works and so the tendency to over-state the leader’s role in organisational success. Muijs (2011a) identifies a clear tendency among qualitative researchers committed to the transformational view of leadership to adopt a direct effects model and research methods likely to validate the theory (such as, for example, case studies relying largely on head-teacher self-report). Likewise, among quantitative researchers interested in distributed leadership models, there’s a clear preference for the mediated model, according to which leaders affect outcomes through, for example, influencing teacher behaviours. Finally he comments on a third type of research, which he describes essentially as ‘positional’ – where proponents of one or another theory make their case, based neither on empirical research nor systematic review of the literature. ‘Position papers’ of this kind made up fully 16.5% of the 446 articles reviewed in the course of his study. Due to their often ideological or evangelical stance, these generally assume a direct effects model. 7 Since this model was commended to the mainstream of school leadership research by Hallinger & Heck (1996a; 1996b), there have been not more than a handful of empirical studies that have employed reciprocal-influence models to the study of leadership and school improvement (e.g., Hallinger & Heck, 2010; Heck & Hallinger, 2010; Mulford & Silins, 2009; Rowan & Denk, 1984).
16
Oppositional research vs theoretical integration Oppositional research vs theoretical integration
The theories outlined above are not without their insights, of course, but as models, with the vigorous support of their advocates, they have, in turn, exercised with their popularity, an unhelpfully normative influence on practice (Bush & Glover 2003), and indeed continue to do so (Muijs 2011a). Muijs attributes this to a pervasive dualism in the field that pits different aspects of leadership – hierarchical / distributed, transactional /collaborative, for example – against one another, rather than seeing them as integral to what we mean by leadership. Picking up on this point, Grint speculates that though the fashion for distributed and collaborative forms of leadership (implying a ‘stretching’ of the leadership function with and across the organisation) may in some measure have been a necessary corrective to the romance of individual leadership, we may now be in just as much danger of adopting the same attitude to these more recent theories (Grint 2011:9-10; Leonard 2007). For Muijs, this tendency to think in terms of competing theories or paradigms is exacerbated by the political pressures associated with the growth of high-stakes accountability, and encourages over-prescriptivity in relation to what is at best a limited research base. Reviewing progress in the field, Hallinger (2011) feels that ‘the fervor of debates over which model offers the greatest leverage for understanding how school leaders contribute to learning’ has ‘reduced’. He observes that the term ‘leadership for learning’ has subsumed features of instructional leadership, transformational leadership and shared leadership in the work of a number of leading researchers in the field, who are, moreover, keen to stress the ‘adaptive’ nature of the process by which it exerts its influence (Hallinger & Heck 2010). Cognisant of these developments, Muijs’ (2011a) research nevertheless clearly demonstrates that ‘there is still too often a tendency [among researchers] to coin new phrases and so-called types of leadership’, rather than to do the hard work of identifying, 17
Taking a lead
testing and quantifying the effects of specific practices associated with what we think is effective leadership according to existing theories.
Moving forward with the reciprocal effects model: challenges for research Likewise, while Day & Antonakis (2012), Hallinger (2008), Leithwood et al (2004), and Liden & Antonakis (2009) all point to growing interest in contextual factors, and the need for leader flexibility, few have sought to gauge their importance on the more theoretically rich recursive/ reciprocal effects model outlined above. Case study approaches, involving interviews and the collection of documentary evidence, and questionnaires, continue to predominate – methods that are susceptible of a range of biases (attributional,8 self-presentation,9 interviewer expectancy,10 etc.) (Muijs 2011a). There seems not only a lack of resource, but of willingness, to rise to the challenge of the kind of longitudinal research needed for analysis of reciprocal effects. Muijs (2011a) wonders why, for example, Creemers & Kyriakides’ (2007) promising work on a dynamic theory of educational effectiveness, according to which factors at different levels are seen as having both direct and indirect effects upon student outcomes (Reynolds et al. 2014:203), was so slowly followed by projects seeking to test the dynamics between the different aspects of this effects model. This, again, despite the fact that advances in statistical analyses opened it to testing some time ago (Reynolds et al 2014:199). This kind of research is ‘essential’ 8 Attributional bias refers to the tendency to attribute to ourselves positive outcomes, while negative outcomes are externally attributed. 9 Self-presentation bias refers to the tendency to present oneself in an overly favourable manner due to the need to maintain and enhance self-esteem. 10 Interviewer expectancy bias refers to the tendency of interviewees to react to the social style and personality of interviewers, or to their presentation of particular questions, in a way that distorts results.
18
Moving forward with reciprocal effects Moving forward with reciprocal effects
given that the model assumes that leaders at moment 1 can influence achievement at moment 2, which can in its turn influence leadership at another moment, etc. (De Maeyer et al. 2007: 128). This reticence to consider and to adopt the appropriate model and methods to gain understanding of how leadership works, and how therefore its effect may be enhanced, is further suggested by Muijs’ (2011a) finding that less than 1 per cent of the papers he analysed were of quasi-experimental design and that not a single experimental study featured randomised assignment to treatment and control groups. This ought to concern us given that an increasing number of methodologists are of the view that it is impossible to establish causality in organisational behaviour studies outside of an experimental research design (Coe and FitzGibbon 1998; Goldstein & Woodhouse 2000; Stone-Romero 2009). So, how might leadership research recover its credibility and move towards a useful understanding of the complexities of school leadership in context, and what makes for success? First, it’s important to recognise that given the over-theorised nature of leadership studies, and the limitations of both research design and method that characterise the field, much of what leaders actually do, and potentially what makes them effective in their different contexts, may have been missed. The description of actual leadership decision-making and practice, in context, and over time, ought therefore to be the starting point. Then if it can be established, over time, how far particular patterns are influenced by context or are constant, there may be a route to assessing the impact of those practices, and working out better strategies for replicating the ones that work. To date, however, early promising efforts by Dooley & Lichtenstein (2008) and Wood & Ladkin (2008) to develop 19
Taking a lead
the kind of microscopic methodological approaches required to identify those patterns, and calls for investment in ethnographic (or otherwise overtly atheoretical, descriptive) approaches fitted to their interrogation (Bryman et al. 2011:26; Muijs 2011a) appear to have fallen largely on deaf ears. Muijs’ (2011a) suggestion that there should be field trials of leadership strategies devised on the back of this exercise in pattern identification seems beyond imagining. There is, in short, a long way to go in leadership studies, before we have a basis for confident replication of leadership decisions and practices that are known to be impactful.
Conclusion: What does this mean for policy? What does this mean for policy? First, we simply do not know enough about what particular practices are impactful, learnable and transferrable, to require participation in leadership development programmes. Reflecting the wider literature on leadership effectiveness, there is no robust evidence to support claims that professional qualifications make a difference to the quality of headship. The Coalition, and subsequent Conservative governments, have therefore been entirely justified in their decision to lift the requirement for new heads to take the National Professional Qualification for Headship (NPQH) (as also extension the requirement for Qualified Teacher Status). In that it seems likely that emerging leaders are as much shaped by the opportunities and challenges of the contexts in which they find themselves, it is also right for the locus of leadership identification and development to have been shifted to the schools level, to be led by head-teachers, in situ. Moves towards the process of acquiring the necessary skills becoming more practical and work-based are also sensible. This ‘informed and deferential agnosticism’ in respect of how to address the challenges of leadership 20
Conclusion: What does this mean for policy?
recruitment and development, undergirded by faith that the incentives inherent in its academy reforms will liquidise leadership capital, is a promising attitude. If school leaders themselves cannot identify potential leaders, and design and develop opportunities for leadership development in school, there is certainly no basis for believing this to be within the skill set of central government. Incentivising leaders to invest in leadership development in school is therefore the right direction for policy. The best way of doing this is to work towards a model of headship training and continuing professional development that is research-informed, leader- and demand-led. This should essentially consist of two elements: 1) a head-teacher shadowing scheme based on the observation of leaders in the school context, supported by 2) course content oriented to improving research literacy and building and refreshing knowledge of what works. Mentoring and peer-to-peer support would be provided by leaders of similar schools according to demand from those in need of advice, strategy, and support. It would be up to them do determine from such observations what could be useful and applicable to thier own contexts. This consultancy model would get around the misleading sense that an intervening ‘hero head’ is in some way taking responsibility for the outcomes of the advice he/ she has given. The present government, however, seems to be set on reinvesting in hero-heads and a suite of leadership development qualifications to be provided by the National College of Teaching and Leadership. This is in accord with its general counter-veiling tendency to override incentives-based strategy and take decision-making back to the centre – as witnessed by its ongoing control of the recruitment of academy sponsors, and increasing centralisation of decision-making in respect of national educational standards and assessments, regulation, 21
Taking a lead
and accountability. In that the research suggests it is leaders’ discretion in the key areas of change to leadership and management structure, motivating staff through mission and goal-setting and appropriate incentives in pay and conditions, and curriculum and pedagogy – and consistent implementation of those decisions – that counts, centralisation of policy geared to accessing a leadership premium is not welladvised. Such reforms threaten to stymie the potential of its school-based management reforms by taking over recruitment at the top tier and reducing leaders’ scope to take decisions and effect strategies in the areas identified to have bearing on academic improvement. While the evidence base in this respect, as in others, is limited and far from robust with respect to causation, and researchers have a long way to go before they are capable of articulating in any detail the specific decisions and practices, in context, that shape the conditions for effective learning, it does at least give us a starting point. To progress from here, what we really need in our leaders is the ability attend to and to follow the evidence – for our leaders, in short, to become followers.
22
References Bass, B. M., & Avolio, B. J. (1994). ‘Transformational leadership and organizational culture’, The International Journal of Public Administration, 17(3-4): 541-554. Bass, B. (1985). Leadership beyond expectations. New York: Free Press. Bass, B. M. (1998). Transformational Leadership: Industrial, Military, and Educational Impact. Mahwah, NJ: Lawrence Erlbaum Associates. Bossert, S. T., Dwyer, D. C., Rowan, B., & Lee, G. V. (1982). ‘The instructional management role of the principal’, Educational administration quarterly, 18(3): 34-64. Branch, G. F., Hanushek, E. A., & Rivkin, S. G. (2012). ‘Estimating the effect of leaders on public sector productivity: The case of school principals’ (No. w17803). National Bureau of Economic Research. Branch, G. F., Hanushek, E. A., & Rivkin, S. G. (2013). ‘School leaders matter’, Education Next, 13(1). Bridges, E. M. (1982). ‘Research on the school administrator: The state of the art, 1967-1980’, Educational Administration Quarterly, 18(3): 12-33. Bryman, A. (2011). ‘Research methods in the study of leadership’ in The SAGE handbook of leadership, A. Bryman, D. L. Collinson, K. Grint, B. Jackson, M. Uhl-Bien (eds) (SAGE Publications): 15-28. Bush, T., & Glover, D. (2003). ‘School leadership: Concepts and evidence.’ Chapman, C., Muijs, D. & Collins, A. (2009). ‘The impact of federations on student outcomes.’ Nottingham: National College for School Leadership. Chapman, C., Muijs, D. & MacAllister, J. (2012). ‘Collaborative school turnaround: A study of the impact of federation on student outcomes.’ 23
Taking a lead
Paper presented at the annual meeting of the American Educational Research Association. Vancouver, Canada, April 2012. Chapman, C., Muijs, D. & Sammons, P. (2010). ‘Federations and student outcomes: A study of the impact of school-to-school collaboration on school improvement.’ Paper presented at the annual meeting of the American Educational Research Association. Denver, CO, April 2010. Chapman, C., Muijs, D., & MacAllister, J. (2011). ‘A study of the impact of school federation on student outcomes.’ Nottingham: National College for School Leadership Chiang, H., Lipscomb, S., & Gill, B. (2012). Is school value-added indicative of principal quality? Cambridge, MA: Mathematica Policy Research. Coe, R. & FitzGibbon, C. T. (1998). ‘School effectiveness research: criticisms and recommendations’, Oxford Review of Education 24: 421438. Creemers, B. & Kyriakides, L. (2007). The Dynamics of Educational Effectiveness. Abingdon: Routledge. Creemers, B. P. (1994). The effective classroom. London: Cassell. Croft, J. (2015). ‘Collaborative overreach: why collaboration probably isn’t key to the next phase of school reform.’ London: Centre for the Study of Market Reform of Education. D’Agostino, J. V. (2000). ‘Instructional and school effects on students’ longitudinal reading and mathematics achievements’, School Effectiveness and School Improvement 11 (2): 197–235 Day, D. V., & Antonakis, J. (2012). ‘Leadership: Past, present, and future’ in The nature of leadership. Sage Publications. 3-25. De Maeyer, S., Rymenans, R., Van Petegem, P., van den Bergh, H., & Rijlaarsdam, G. (2007). ‘Educational leadership and pupil achievement: The choice of a valid conceptual model to test effects in school effectiveness research’, School effectiveness and school improvement, 18(2): 125-145. Dhuey, E., & Smith, J. (2014). ‘How important are school principals in the production of student achievement?’, Canadian Journal of Economics/Revue canadienne d’économique, 47(2): 634-663. 24
References
Dooley, K. J., & Lichtenstein, B. (2008). ‘Research methods for studying the complexity dynamics of leadership’ in. Complexity Leadership. Part I: Conceptual Foundation, 269-290. Eyles, A., & Machin, S. J. (2015). ‘The Introduction of Academy Schools to England’s Education.’ Fiedler, F. E., & Chemers, M. M. (1967). ‘A theory of leadership effectiveness.’ Fiedler, F. E. (1971). ‘Validation and extension of the contingency model of leadership effectiveness: A review of empirical findings’, Psychological bulletin, 76(2): 128. Fuchs, T., & Wößmann, L. (2008). ‘What accounts for international differences in student performance? A re-examination using PISA data’ in The economics of education and training. Physica-Verlag HD. 209240. Fullan, M. (2002). ‘The role of leadership in the promotion of knowledge management in schools’, Teachers and Teaching: theory and practice, 8(3): 409-419. Fullan, M. (2004). Leadership and sustainability: System thinkers in action. Corwin Press. Gardner, W. L., Lowe, K. B., Moss, T. W., Mahoney, K. T., & Cogliser, C. C. (2010). ‘Scholarly leadership of the study of leadership: A review of The Leadership Quarterly’s second decade, 2000–2009’, The Leadership Quarterly, 21(6): 922-958. Goldstein, H. & Woodhouse, G. (2000). ‘School effectiveness research and Educational Policy’, Oxford Review of Education, Vol. 26, No. 3/4, The Relevance of Educational Research (September– December 2000): 353-363. Grint, K. (2011). ‘A history of leadership’ in The SAGE handbook of leadership, A. Bryman, D. L. Collinson, K. Grint, B. Jackson, M. Uhl-Bien (eds). SAGE Publications. 3-14. Grissom, J. A., Kalogrides, D., & Loeb, S. (2012). Using student test scores to measure principal performance. Nashville, TN: Vanderbilt University.
25
Taking a lead
Grissom, J. A., Loeb, S., & Master, B. (2012). ‘What is effective instructional leadership? Longitudinal evidence from observations of principals’ in Association for Public Policy Analysis and Management annual meeting, Baltimore, MD. Hallinger, P., & Heck, R. H. (1996a). Reassessing the principal’s role in school effectiveness: A review of empirical research, 19801995. Educational administration quarterly, 32(1): 5-44. Hallinger, P., & Heck, R. H. (1996b). ‘The principal’s role in school effectiveness: An assessment of methodological progress, 1980– 1995’ in International handbook of educational leadership and administration. Springer Netherlands. 723-783. Hallinger, P., & Heck, R. H. (1998). ‘Exploring the principal’s contribution to school effectiveness: 1980-1995’, School Effectiveness and School Improvement, 9(2): 157-191. Hallinger, P., & Heck, R. H. (2011). ‘Conceptual and methodological issues in studying school leadership effects as a reciprocal process’, School Effectiveness and School Improvement, 22(2), 149-173. Hallinger, P. & Heck, R. (2010). ‘Collaborative leadership and school improvement: understanding the impact on school capacity and student learning’, School Leadership and Management, 30(2): 95-110. Hallinger, P., & Huber, S. (2012). ‘School leadership that makes a difference: international perspectives’, School effectiveness and school improvement, 23(4): 359-367. Hallinger, P. (2008). ‘Methodologies for Studying School Leadership: A Review of 25 Years of Research Using the Principal Instructional Management Rating Scale.’ Paper presented at the Annual Meeting of the American Educational Research Association, New York, NY, April 2008. Hallinger, P. (2010). ‘A review of three decades of doctoral studies using the Principal Instructional Management Rating Scale: A lens on methodological progress in educational leadership’, Educational Administration Quarterly, 0013161X10383412. Hallinger, P. (2011). ‘Leadership for learning: Lessons from 40 years of empirical research’, Journal of educational administration, 49(2): 125-142. 26
References
Hanushek, E., Link, S., & Woessmann, L. (2012). ‘Does School Autonomy Make Sense Everywhere? Panel estimates from PISA’, Journal of Development Economics 104 (September 2013): 212-232. Hater, J. J., & Bass, B. M. (1988). ‘Superiors’ evaluations and subordinates’ perceptions of transformational and transactional leadership’, Journal of Applied psychology, 73(4), 695. Heck, R. H., & Hallinger, P. (2010). ‘Collaborative leadership effects on school improvement: Integrating unidirectional-and reciprocal-effects models’, The Elementary School Journal, 111(2): 226-252. Hermosilla, M. F., Anderson, S. & Mundy, K. (2014). ‘Education Management and Leadership: A Rapid Review of the Literature’. Strengthening Education Systems in East Africa (SESEA) project, sponsored by Aga Khan Foundation Canada and the Canadian Department of Foreign Affairs, Trade and Development (formerly CIDA). House, R. J., & Aditya, R. N. (1997). ‘The social scientific study of leadership: Quo vadis?’, Journal of management, 23(3): 409-473. Jackson, D. S. (2000). ‘The school improvement journey: Perspectives on leadership’, School Leadership & Management, 20(1): 61-78. Kyriakides, L., Creemers, B., Antoniou, P., & Demetriou, D. (2010). ‘A synthesis of studies searching for school factors: Implications for theory and research’, British Educational Research Journal, 36(5): 807-830. Leithwood, K. A., Riehl, C. (2003). ‚What do we already know about successful school leadership?’ AERA Paper Task Force on Developing Research in Educational Leadership. Leithwood, K., Seashore Louis, K., Anderson, S., & Wahlstrom, K. (2004). Review of research: How leadership influences student learning. University of Minnesota, Center for Applied Research and Educational Improvement Leitner, D. (1994). ‘Do principals affect student outcomes?’, School Effectiveness and School Improvement, 5(3): 219-238. Leonard, P. (2007). ‘Moral literacy for teacher and school leadership education: a matter of attitude’, Journal of Educational Administration, 45(4), 413-426. 27
Taking a lead
Liden, R. C., & Antonakis, J. (2009). ‘Considering context in psychological leadership research’, Human Relations, 62(11): 1587-1605. Gardner, W. L., Lowe, K. B., Moss, T. W., Mahoney, K. T., & Cogliser, C. C. (2010). ‘Scholarly leadership of the study of leadership: A review of The Leadership Quarterly’s second decade, 2000–2009’, The Leadership Quarterly, 21(6): 922-958. Machin, S., & Vernoit, J. (2011). ‘Changing School Autonomy: Academy Schools and Their Introduction to England’s Education’, CEE DP 123. Centre for the Economics of Education (NJ1). Marzano, R. J., Waters, T., & McNulty, B. A. (2005). School leadership that works: From research to results. ASCD. Meindl, J. R. (1995). ‘The romance of leadership as a followercentric theory: A social constructionist approach’, The Leadership Quarterly, 6(3): 329-341. Muijs, D. (2011a). ‘Researching leadership: Towards a new paradigm’ in International handbook of leadership for learning. Springer Netherlands. 115-125. Muijs, D. (2011b). ‘Leadership and organisational performance: from research to prescription?’, International Journal of Educational Management, 25(1): 45-60. Muijs, D., Reynolds, D., & Chapman, C. (forthcoming). ‘Centralisation and scale in academy chains.’ Southampton: Centre for Educational Effectiveness, Improvement and Policy, University of Southampton. Mulford, B., & Silins, H. (2003). ‘Leadership for organisational learning and improved student outcomes – What do we know?’, Cambridge Journal of Education, 33(2): 175-195. Mulford, W. R., & Silins, H. (2009). ‘Transformational leadership and organisational learning.’ Nettles, S. M. & Herrington, C. (2007). ‘Revisiting the Importance of the Direct Effects of School Leadership on Student Achievement: The Implications for School Improvement Policy’, Peabody Journal of Education, 82(4): 724-736
28
References
Ogawa, R. T., & Bossert, S. T. (1995). ‘Leadership as an organizational quality’, Educational Administration Quarterly, 31(2), 224-243. Orr, M. T. (2013). ‘Review of two articles on principal effects.’ National Education Policy Center. Pitner, N. (1988). ‘The study of administrator effects and effectiveness’ in Handbook of research in educational administration. 99-122. Pont, B., Nusche, D., & Moorman, H. (2008). School leadership development: Policy and practice. Improving school leadership (Vol. 1). Reynolds, D., Sammons, P., De Fraine, B., Van Damme, J., Townsend, T., Teddlie, C., & Stringfield, S. (2014). ‘Educational effectiveness research (EER): a state-of-the-art review’, School Effectiveness and School Improvement, 25(2): 197-230. Robinson, V. M. (2007). School leadership and student outcomes: Identifying what works and why (Vol. 41). Melbourne: Australian Council for Educational Leaders. Rowan, B., & Denk, C. E. (1984). ‘Management succession, school socioeconomic context, and basic skills achievement’, American Educational Research Journal, 21(3): 517-537. Sammons, P., & Luyten, H. (2009). ‘Editorial article for special issue on alternative methods for assessing school effects and schooling effects’, School Effectiveness and School Improvement, 20(2): 133-143. Scheerens, J., & Bosker, R. J. (1997). The foundations of educational effectiveness. Oxford: Pergamon. Sergiovanni, T. J. (1999). ‘Refocusing Leadership to Build Community’, High School Magazine, 7(1): 10-15. Stone-Romero, E. F. (2009). ‘Implications of research design options for the validity of inferences derived from organizational research’ in The Sage Handbook of Organizational Research Methods. SAGE Publications. 302-327. Teddlie, C., & Stringfield, S. (1993). Schools Make a Difference: Lessons Learned from a 10-Year Study of School Effects. New York: Teachers College Press.
29
Taking a lead
Van de Grift, W., & Houtveen, A. A. M. (1999). ‘Educational leadership and pupil achievement in primary education’, School Effectiveness and School Improvement, 10(4): 373-389. Witziers, B., Bosker, R. J., & Krüger, M. L. (2003). ‘Educational leadership and student achievement: The elusive search for an association’, Educational Administration Quarterly, 39(3): 398-425. Wößmann, L. (2003). ‘Familiärer Hintergrund, Schulsystem und Schülerleistungen im internationalen Vergleich’, Aus Politik und Zeitgeschichte, 21(22): 33-38. Woessmann, L., Luedemann, E., Schuetz, G., & West, M. R. (2009). ‘School accountability, autonomy and choice around the world’. Wood, M., & Ladkin, D. (2008). ‘The event’s the thing: brief encounters with the leaderful moment’ in Leadership perspectives. Palgrave Macmillan. 15-28. Yukl, G. (2006), Leadership in Organizations, 6th edn. Upper Saddle River, NJ: Pearson Prentice-Hall.
30
Kieran McDermott, Chief Executive of One Education: The field of school leadership studies is characterised by a prevalence of committed and competing theories about the contribution of leadership to student outcomes. What, if anything, do we know that is supported by quantitative evidence rather than faddish assumption? James Croft’s review of the literature on school leadership provides a probing and insightful analysis of the state of the evidence base and a clear indication of the where the future of leadership studies, policy, and effective practice, lies. Effective education system design and policy formation can and should be built on hard, high quality evidence from research. Where such evidence is lacking, we need to be transparent about the limitations of what we know. Research internationally suggests that mission and goal setting, decisions related to the setting of the curriculum and pedagogy, and the provision of instructional guidance for teachers, are important means by which leaders exercise their influence. How leaders motivate staff, including through the use of appropriate pay and conditions incentives, also appears important for raising academic achievement. And in more autonomous school contexts the importance of leadership in these areas is accentuated. But researchers have a long way to go before clarity is achieved about what specific decisions and practices are impactful, and in what contexts. In this connection, Croft’s proposals for a model of headship training and continuing professional development that is research-informed, school-based, and leaderand demand-led offers a promising way forward for policy.
£8.99
CENTRE FOR THE STUDY OF MARKET REFORM OF EDUCATION 23 Great Smith Street, Westminster, London, SW1P 3BL www.cmre.org.uk
[email protected] 
