«2009/ED/EFA/MRT/PI/32 Background paper prepared for the Education for All Global Monitoring Report 2009 Overcoming Inequality: why governance matters ...»
This reasoning can still be valid under two conditions. Firstly, when a country is relatively homogeneous and with little disparities, supervision as a standardization tool can have a positive impact and may not worsen disparities. In such a scenario, applying the same framework and norms throughout the country may indeed make sense. A second condition relates to the State having sufficient resources to ensure that its supervision reaches out to all schools on a fairly regular basis, as this is needed to guarantee the respect of these standard rules and regulations. While these two conditions may to some extent still be present in a few countries, they are absent in most places which work with this classical model, even in some of the most developed countries where disparities have increased and resources are insufficient for regular coverage of all schools. Several responses are then possible. In France for instance, the classical supervision services have changed little, but special programmes have been created to give additional resources and support to the most disadvantaged schools, the Zones d’éducation prioritaire. These compensatory programmes, which think along the same lines as the third model described above, tend to lead to a stigmatization of the schools which belong to these priority zones and their record is patchy.
In developing countries, the key constraint is that the mandate of supervisors, namely to supervise and support all schools, outweighs by far the resources. A rather unpromising response, which many supervision services nevertheless adopt, is to ask continuously for more resources, which however never arrive. Another option is to accept the limited level of resources and to restrict the mandate, i.e. to cover a limited number of schools. If disparities between schools would be of great concern (and in many developing countries they are of such a high level that they should be), then supervisors probably should focus their attention on the schools which need their intervention most. Unfortunately many supervision services in developing countries have made a different choice, namely to focus on the schools which are within easy reach, such as those close to their office or to the main road. The result is that more remote schools tend to get less supervision. In Zimbabwe for instance at the end of the 1990s, while a teacher was on average visited every two and a half years, those in rural areas had to wait four years (De Grauwe, 2001, p.103).
There are at least four reasons to explain this scenario. Firstly, because these schools are easier to reach, they are less expensive to cover. Secondly, because supervisors are to some extent evaluated on the number of schools and teachers supervised and the number of reports, it makes more sense to visit many schools close by even if they do not need such a visit, rather than a smaller number of isolated schools, even if they are more in need. In other words, institutional incentives are absent for these officers to change their traditional practices. Thirdly, the administrative structure is such that all districts or regions within the country are assigned a similar number of administrative or supervisory staff, without taking into account the needs and characteristics of schools for which they are responsible. Finally, few supervisory offices have developed within their own district the necessary data and indicators allowing them to carefully select schools most in need. Because of this complex set of reasons, in many developing countries the standardized supervision service is probably reinforcing disparities, because it is reaching out more to better-off schools than to schools facing difficulties. To summarize, the classical supervision model is in most developing countries no longer suitable to the existing education system. Not only does it fail to fulfil its
mandate – comprehensive administrative control – but its mandate itself has become outdated:
while some form of control is undoubtedly needed, the focus must shift to pedagogical improvement.
It could be hoped then that more support-oriented strategies, introduced precisely to make up for the over-emphasis on administrative control, would be more successful in addressing disparities by concentrating their efforts on the more disadvantaged schools. However, services which specialize in pedagogical support suffer from a somewhat comparable weakness. In many case, their advice is also benefiting mainly the schools closest to where these support services are located. Research undertaken on the role of resource centres for instance in India, Kenya, Nepal and Zambia (Khaniya, 1997; Knamiller, 1999) shows that they are generally not able to reach out to a large number of schools, and even where they succeed in doing so, they offer advice which is of little relevance to the situation of schools whose resources and context are too far away from the standard one that these services know and cater for. The interventions by these school monitoring services are reflective of the State’s intervention as a whole: because the State is incapable to fulfil its mandate, authorities tend to focus on those groups whose support is important to their survival. The politically less vociferous groups are to some extent abandoned and will at times, with their own scarce resources and with the help of non-governmental organizations, set up their own services.
The fourth model – the peer supervision model – pays little attention to disparities, mainly because, as we saw above, this model was developed in countries where such social and educational disparities are weak. The absence of these disparities is not a simple coincidence;
it is the result of socio-economic policies aimed at reducing disparities. These sociodemocratic traditions are strongly anchored in the psyche of these countries. This in itself offers an important insight: reducing disparities is probably not the key role of education, which cannot be asked to repair what socio-economic policies have failed to solve or may even have contributed to. Educational policies evidently can have an impact on disparities, but as long as they are not an integral part of an overall framework with the common objective to reduce such disparities, they will never succeed in a sustainable diminution of these disparities and will always be fighting a rearguard battle.
This also brings up the point how far this “peer-supervision” model can be transplanted from one national context to another. It relies on a distant evaluation by the centre, a strong trust in the teaching corps and the involvement of a capable and interested community. Where disparities are important, such a model may not be able to address them because there is too little external intervention in schools. Even in Finland, one of the economically most egalitarian societies, differences in the level of school evaluation are developing as a result of the actions by municipalities (De Grauwe, 2006, p.260-271). Indeed, some big urban municipalities have set up their own school evaluation units, believing that these will contribute to higher quality. In more remote rural regions, school principals and teachers are not at all supervised by any external actor, mainly because municipalities do not have the necessary financial or human resources to organize supervision. In the long run, this could lead to increased disparities for which the actual system is not prepared.
The central control model does pay attention to the issue of disparities. It puts the blame for differences in results between schools squarely on the school’s internal functioning. It is based on the conviction that as long as schools follow a certain set of principles, based on a rather simplified interpretation of school effectiveness research, they will perform with success. The supervision service imposes these principles upon all schools and therefore uses the same approach systematically. The school’s socio-economic context is disregarded.
Studies on the UK (Smith, 2000) and New Zealand (Fiske and Ladd, 2001) which implemented fairly pure examples of this model, emphasize this lack of attention to the school’s context, while at the same time pointing at the possible linkage. Smith’s study demonstrates a strong association between the socio-economic context and the evaluation made by OFSTED: among schools based in the highest socio-economic areas, 91 % received a “favourable” evaluation and only 3% a “not favourable” one; among schools in the lowest socio-economic areas, only 2 % received a “favourable” evaluation and 89 % a “not favourable” one. When OFSTED made efforts to take more into account the school’s context in its inspection visit, there was resistance from among political authorities who felt that such a move could be interpreted as a lowering of expectations and a loss of pressure on schools to perform.
More preoccupying however than this neglect of the context, is the fact that the supervision strategy followed by the central control model tends to enlarge disparities between schools. A quantitative analysis undertaken by Shaw et al (2003) indicates that in only three types of schools (girls only schools, selective schools and grant-maintained schools, all of which are schools catering for a certain elite public) inspection was associated with a light improvement of exam results. Elsewhere, there is no impact or results seem to decrease. A similar conclusion was arrived at by Cullingford and Daniels (1999) on a less representative sample.
Chapman (2002) went in some more detail and examined what changes schools in difficult circumstances undertake following an inspection visit. These changes have little to do with what takes place in the classroom but mainly concern administrative reporting or management procedures. Chapman draws the conclusion that a monitoring approach such as OFSTED’s is better at influencing management practices than classroom interactions. Unfortunately, what such weak schools need above all is a change in classroom behaviour. Research in New Zealand (Fiske and Ladd, 2001) has also led to the conclusion that the central control model tends to lead to increasing rather than reducing disparities.
There are at least two reasons for this. Firstly, schools and teachers receive too little support.
Many teachers make the point that recommendations for improvement are of little help if they are not accompanied by advice and guidance on how to implement these recommendations.
The best performing schools evidently suffer much less from this lack of support than the weaker schools. Schools in a situation of crisis feel de-motivated by a process which highlights their weaknesses (and turns them into public knowledge) without offering any solutions (Fiske and Ladd, 2001; Case et al, 2000; Shaw et al., 2003; Chapman, 2002;
Cullingford and Daniels, 1999). Moreover, the competition between schools which this model promotes, has led to a weakening of school networks, to the detriment of the weakest schools, as Harold (1999, p.11-12) demonstrates for New Zealand. Secondly, this central control model puts a heavy emphasis on the responsibility of schools and teachers. The school principals in particular suffer under this stress, when they are made responsible for the school’s performance, even if certain factors explaining this performance are beyond their control. Because of the emphasis on accountability, principals also tend to spend much of their time on administrative and managerial tasks, cutting in their time available for supporting their teachers. Again, this is more of a concern in weaker schools. It is not surprising that in South-Africa, where attempts were made to introduce a similar model, the only schools who have easily accepted to undergo a “whole school review” are those who recruit from among the richest parents and who have the best results – the previously “white” schools. The positive evaluation reports they get, are helpful in their public relations campaigns.
A further negative effect of the central control model is that in a number of cases it has delegitimized the internal school evaluation process. Indeed, teachers who have gone through a school audit, which internal evaluation helped to feed and which led to a critical public report, feel to some extent “cheated”. They become somewhat suspicious about any evaluation process, including those led by internal actors and aimed at school improvement (Webb and Vullimay, 1998, p.548). It also makes collaboration within the school more intricate. This is deplorable because all schools, including those with the poorest results, need such an internally led process in order to improve. While an external encouragement is in many cases necessary, it should not pervert the objectives of the internal process. It is because of these various criticisms that OFSTED itself has taken a number of steps which have moved it away from this pure model: there is more space foreseen during the supervision visits for discussion with teachers; reports contain some helpful advice; the school’s environment receives some attention.
Behind this central control model, we discover a model of a State, which has as main preoccupation concerning the public services, not to guarantee equity, but to ensure that public services respond effectively to the demand of the individual consumers. This has different implications. Market mechanisms are promoted as they are believed firmly to serve both individual and common interests. For this market to function well, the State needs to provide the consumers with the necessary information. The game of choice and competition will lead to an overall improvement of schools, and to the disappearance of those schools who fail to improve. This demands that these services are subject to public accountability and that their functioning becomes more transparent. We can note such a discourse in Anglo-Saxon countries (Berman, 1999; Walford, 2001, p.23 or McInerney, 2003, p. 60-62) but some authors also find traces of this reasoning in China (Chan and Mok, 2001, p.26-30) and even in the Scandinavian countries (Johannesson et al., 2002, p.326-337).