The post-conflict nature of the Kosovo society and economy led to an urgent need to address educational policy, specifically to raise the quality of the reconstituted formal schooling system. To address this priority major foreign aid and government subsidies were targeted at both the demand (students) and supply (school) side. One of the major contributors, the World Bank, aimed to improve the supply side by allocating development grants to schools in order to improve student performance. In this thesis the following four research questions are addressed: how appropriate are current evaluation strategies of education policy initiatives in developing countries, what has been the impact of school development grants on student dropout, attendance and student attainment, what are determinants of pupil dropout, attendance and attainment and what are the implications of the answers to the above questions for the reform of education policies in developing economies and the evaluation of policy initiatives. This is the first study that critically reviews previous attempts at evaluating educational initiatives in Kosovo and then employs econometric methods to measure the impact of school development grants on educational outcomes. A quasi experimental approach is utilised and comparisons made between schools with treatment and schools without treatment. A similar study for Cambodia serves as a reference for our research, though we have extensively refined the approach taken in that study. The empirical evidence presented in this thesis suggests at best only a marginal positive impact of these policy initiatives on educational outcomes. More specifically there is some evidence of reduced dropout but no effect is found on student attendance and attainment. These findings are consistent with the results of recent reviews of the literature on this type of policy initiative. This study seeks to act as an example of best practice which can be followed in future evaluations of policy initiatives in countries like Kosovo. It draws important conclusions about the need at the policy design stage to formulate appropriate evaluation strategy and to address related issues about data quality, collection and analysis.

The introduction of civic education to Serbian primary and secondary schools in 2001 marked a beginning of an all-encompassing education reform that followed the country's 2000 'democratic revolution'. In the context of a socio-political shift from various authoritarian regimes, including the 1945-1990 state socialism and 1990-2000 nationalist authoritarianism, the policy set the tone for future changes that were designed to support democratisation of Serbia, and assist its return to Europe (Birzea, 1994). A part of the broader programme for democratisation of education and education for democracy in Serbia, the policy enabled various discursive elements constitutive of the desired post-2000 ideology to enter the national educational discourse. This thesis explored its formation. I approached the policy as a way to explore the beginning of Serbia's first proper post-communist reform, and analyse the actors and ideologies that had shaped it. I used Ball's notion of policy-as-discourse and conceptualised civic education policy as a part of a discursive project of creating a 'new Serbia'. Using elite interviews and documentary analysis, I explored its formation and development, its place in the wider reform, and its relation to religious education, (re)introduced at the same time. Combining the elements of Fairclough's critical discourse analysis, and elements of Bourdieu's social theory, I looked into the meaning and function of civic education as a part of the ideological construction of the future Serbia, as well as capital used to position Serbia favourably in the global field in the early days of its educational transition. Within the wider transition literature, I attempted to establish a comparison between Serbia's 'belated' post-communist transition, and educational changes happening across formerly communist countries of Central and Eastern Europe in early-to-mid 1990s. I also explored the applicability and usefulness of the recent theoretical developments in the transition literature that go against the conceptualisation of post-communist transitions as modernising projects, and argue instead for a focus on unique transformations that happen as a result of a meeting between globally dominant and desirable discursive elements and local contexts. I conclude that the discursive elements of the Serbian civic education policy were used as capital by Serbian policy actors to ensure their better positioning not only in the global, but also in the national field, as suggested by differences in the ideological construction of the policy discourse in different fields. This prompts a concern with the concept of various 'policyspeaks', as recently explored by Halász (2012) and Steiner-Khamsi (2014). I argue that as a part of a discursive project intended to construct post-2000 Serbia, civic education policy worked more towards eradicating the undesirable ideology of violent nationalist authoritarianism, than towards eradicating the ideology of communist authoritarianism. In this sense, the specificity of the context proved important for the shape and meaning of a post-communist reform and ideologies that it was designed to propagate. However, instead of rejecting modernist concepts of transition and democratisation, I advise a future focus on careful unpacking of their context-dependent ideological-discursive constructions.