1 TESTING IN BILINGUAL EDUCATION PROJECTS: LESSONS LEARNT FROM THE ABLE PROJECT Elize Koch i Dedication: Neville Alexander for the profound impact his life and work had on us Abstract This paper discusses the topic of tests, bilingual tests, and the role played by the language of tests in home-language-based bilingual education projects such as the Additive BiLingual Education (ABLE) project. More specifically, we highlight how national, macro-contextual issues have impacted on the project team’s decision to move away from an experimental approach in the research, necessitating the use of tests, to a more participatory action research approach. The implications of these macro-contextual factors for projects of a similar nature, and suggestions for engagement, are given. In addition to discussing the impact of discrepancies between the additive bilingual language in education policy of the DoE of SA (LiEP) and practices around the language of tests such as in the Annual National Assessments (ANA), the paper also engages with the concept of bilingual educational tests, what the concept means in practice, and how we dealt with it in the project. An argument in support of a consistent approach to bi-or multilingual tests is provided. Some practical issues related to tests in general, such as validity, and bilingual tests specifically, such as equivalence, are discussed. Research on the equivalence of bilingual tests used in the ABLE project is synthesised, and the implications of the results discussed. It is argued that bilingual tests could contribute to a radical new way of approaching language in education in the context of post-colonial South Africa. Background As discussed in Koch, Landon, Jackson, and Foli (2009), the primary aim of the ABLE project was to assist teachers at a rural Eastern Cape school with the translation of the 1997 South African Language in Education Policy into a workable model of implementation. The school was formed during the late 1990s out of the merger, on a single site, of several farm schools, and caters for pupils from Reception level (grade R) to grade 9. The majority of learners and educators at the school are native speakers of isiXhosa. For the most part, the learners and educators are neither exposed to, nor required or motivated to, use their first additional language, English, in the community beyond the classroom (Stanfield, 2003).