Uses, applications and problems of educational value-added models

The use of complex value-added models (VAM) that attempt to isolate the contribution of schools or teachers to student development is increasing. This paper presents an overview of several aspects related to the state of art of VAMs. The documents used for the review were peer-referees articles, books and reports from different institutions interested in VAMs. Firstly, it presents some VAMs currently used operationally and very well received by stakeholders. Secondly, the paper provides an overview of the main uses of VAMs. Two broad objectives are identified that can benefit from the use of results: school improvement and school accountability with references to school choice. Finally, the current active research is revised. It provides an overview of the issues related to the use of VAM and leads to caution in the interpretation of results. The main issues are related to three topics: statistical, psychometric and practical issues. The statistical issues are: sensitivity of estimated effects to model assumptions, model structure, missing data, volatility of estimated effects, causal attribution and uncertainty of estimates. The main psychometric issues are related to the construction of vertical scales to measure educational progress. The practical issues are related to the timing of measures, choice of measures of results and some frequent critics such as the narrowing of curriculum and the teaching to the test. We conclude that a diligent application of VAM, with transparency and adequate information to stakeholders, holds considerable promise for school assessment, especially for diagnostic and school improvement, and as aid in the planning of educational reforms.