The current focus upon school performance in many countries is driven by questions about the effectiveness of investments in schooling, coupled with widespread concerns about national economic competitiveness. Given the central role of human capital in the modern economy (Friedman, 2005; OECD, 1994, 1996, 2001), a nation’s schools are seen as a potential source of competitive advantage. A related worry is that the existence of substantial levels of heterogeneity in school performance together with meaningful differences in education outcomes for recognisable subgroups in the population can lead to societal strains and create economic inefficiencies, (OECD, 2008; Lucas, 1988; Romer, 1994). To properly address these issues methods for accurately measuring school performance are required to effectively evaluate investments in schools, identify best practices and to highlight areas where improvements need to be made. Such a system should adequately convey this information to illustrate how such improvements can be made to enhance the performance of all schools.

In this report, the term value-added modelling is used to denote a class of statistical models that estimate the relative contributions of schools to student progress with respect to stated or prescribed education objectives (e.g. cognitive achievement) measured at at least two points in time. To the extent that such progress is a desirable outcome of schooling, value-added modelling can therefore provide a valuable source of information. Indeed, as Part I makes clear, the output of value-added modelling might be used in many ways by both education authorities and school officials. There are many different value-added models in use today, each with its own advantages and disadvantages. Part II of this report identifies the key issues in the design of value-added models and then presents descriptions of some of the more common value-added models. Various statistical and methodological issues are then discussed to assist policy makers and administrators in the design of value-added modelling and in choosing the most appropriate model for school development and to monitor progress toward specified objectives in their education system.

Regardless of the nature of the statistical and methodological underpinnings of the value-added modelling, the impact upon policies, practices and outcomes can be negligible or even negative if an effective implementation is not undertaken. This belief was evident in a number of countries involved in the development of this project and led to more detailed analysis of the methods for implementing a system of value-added modelling. Part III of this report builds on the discussion presented in Part I and Part II to provide a guide for the implementation of a system of value-added modelling in education systems. Such a guide is not a definitive list, nor will each aspect be applicable to all education systems. Rather, it builds on the knowledge gained both in various education systems and from the expert group who have experience in implementing systems of value-added modelling in various education systems.