Over the past decade, international development agencies have begun to emphasize the improvement of the quality (rather than simply quantity) of education in developing countries. This new focus has been paralleled by a significant increase in the use of educational assessments as a way to measure gains and losses in quality. As this interest in assessment has grown, low-income countries have begun to adopt and adapt international and other assessments for a variety of uses, including the comparability of national quality with other countries, improved ways of measuring reading achievement, and further attempts to reach marginalized populations within a country. The present paper reviews a number of international, national and ‘hybrid’ assessments, and considers their merits in terms of how learning is measured, as well as their credibility, sampling and scaling methodologies. The new hybrid assessments, for example, provide innovative opportunities for early intervention for children in their local languages. They also put a premium on local validity over international comparability. The review concludes that there is no single assessment with a dominant scientific superiority, nor is strict comparability across populations or nations a requirement. Rather, different assessments have different policy and practical purposes, and can be used in important and differing ways to improve educational quality. Educational decision-makers working in developing countries have important assessment needs and priorities, and will have to choose carefully in order to address them.