Recommendations appear out of step with ideals of assessment for learning Among the goals of education reforms launched in 2000 was the reduction in testing and more varied teaching, learning and assessment strategies. Is the progress to date providing a less pressurised learning environment for primary school students? The Education Commission's consultation document on medium of instruction and secondary school places allocation (SSPA) seems to neglect in-depth discussion of the latter in favour of the more emotive topic of language. The recommendations on SSPA seem out of line with the reform's emphasis on assessment for learning - an assessment process which serves primarily to promote students' learning. It contrasts with the more traditional assessment of learning which has the primary purpose of measuring what children have learnt. Ideally, the two should be complementary, if an assessment task leads to solid learning and can be used to measure performance. Often this does not happen, for example, when students memorise facts for a test and then forget them soon after. SSPA is a selection device, based on internal primary school examination scores in Primary Five and Six, scaled against results of the school in previous sittings of the now defunct Academic Aptitude Test (AAT). Basically, the internal school results evaluate the individual pupil, whilst the AAT results compare the general performance of the students with those in other schools. The consultation document essentially offers two options. Firstly, to continue with the status quo, despite the fact that AAT results are out-of-date, thus penalising newer or fast-improving schools. The alternative is to scale the results against another exam, the Pre-Secondary One Attainment Test taken by Primary Six graduates before they join secondary school. This is presumably the option favoured by most of the working group. Irrespective of which option is chosen, it is hard to see how the assessment processes that will be used to generate students' scores will bear much resemblance to assessment for learning. Primary school tests are invariably summative in nature and there are a lot of them. Continuing to use these school results will reinforce the testing culture. Added to this assessment conundrum are the government Basic Competency Assessments (BCAs) taken in Primary Three and Six and used to identify territory-wide standards in the three key subjects, Chinese, maths and English. Positive aspects of BCAs include the provision of web-based assessments and feedback to schools on pupil performance. This provides data to promote further learning, so BCAs can fulfil dual purposes of assessment for and of learning. But teachers report difficulties in making productive use of BCA results. As schools are already drilling students for BCAs, it is worth considering whether they might be developed further to play a role in the SSPA. These tests are not designed for that purpose, but if the proposal leads to two exams - BCA and pre-S1 HKAT (both viewed as high-stakes by stakeholders), the policy aim of reduction in tests does not seem to be fulfilled. Tests in themselves are neither good nor bad. The extent to which they provide desirable or undesirable outcomes depends on a number of factors, including their design; how teachers prepare students for them; and how the results are used. High-stakes exams can motivate students to study for a specific goal and provide incentives, particularly for more academic students. They can develop attributes such as determination and drive to succeed. For less academic students goals may seem out of reach and then disillusionment frequently sets in. There are positive assessment for learning strategies being promoted through teacher education and starting to be adopted in schools. One technique involves the identification and discussion of 'success criteria', in other words what is entailed in good performance. For example, before doing an oral presentation in class, students discuss and identify characteristics of a good presentation. Another positive technique involves peer-assessment and self-assessment, whereby students themselves try to evaluate their progress rather than relying on the teacher as the sole provider of feedback. Peer and self-assessment are not primarily about students awarding marks but identifying what good performance is. In this way, students can begin to take responsibility for their own learning, a crucial lifelong learning skill. A longstanding assessment problem is the inordinate amount of time teachers spend on marking, but is this helping students to learn better? The evidence suggests not. It would be more productive for teachers to be directed to spend their valuable time on planning productive lessons or interacting with their students. The various assessment issues are challenging and solutions do not come readily. The recent coverage of Finnish education provides two insights which may carry resonance for the Hong Kong context; the fact that teachers are qualified to Master degree level; and that very few formal tests are carried out. This prompts two conclusions of possible relevance to Hong Kong. Firstly, it is the quality of learning experiences not the quantity of tests taken which has the most impact on student learning. Secondly, it reinforces the value of initiatives which support the professional development of teachers. If assessment for learning is to take root, we need a clearer, more integrated, direction from government and more support for our hard-working and over-pressed teachers and students. Otherwise, it may be an impossible dream. The Hong Kong Institute of Education (HKIEd) is holding a one-day conference on assessment for learning on June 10. For details, visit http://www.ied.edu.hk/loap/loac.html David Carless is a senior lecturer at the HKIEd.