Trends (1995 – 2000) in the TIMSS Mathematics Performance Assessment in The Netherlands Pauline Vos a * and Wilmad Kuiper b a University of Groningen, Groningen, The Netherland; b University of Twente, Enschede, The Netherlands In TIMSS-95, participating countries could administer the TIMSS Performance Assessment consisting of practical tasks, and considered to match well with the Dutch intended curriculum. But in 1995, Dutch students did not score as expected on this test, revealing a discrepancy between intended and attained curriculum. Therefore, in 2000, the test was replicated. Results show an increased teachers’ acceptance of the test, but – still – no significant gain in Dutch students’ achievements. Additionally, if reliability is well controlled, the study revealed that there are valid mathematics assessment alternatives, which can supplement paper-and-pencil tests, not only in The Netherlands. Introduction and Problem Statement In 1995, The Netherlands participated in TIMSS, amongst others for the grade-8 students. At the level of student outcomes, the study comprised the administration of an international written test for mathematics and science. Countries participating in this study could opt to administer additionally an international performance assessment, which also covered science and mathematics. From the achievement results, it appeared that Dutch students performed relatively well on the written mathematics test (significantly above the international average, 9 th in the ranking with 41 participating countries, 4 th in the ranking of 19 countries that participated in both study components; for the latter see Table 1). This was paradoxical, as Dutch mathematics curriculum experts objected to the nature of the written test. The *Corresponding author. Faculty of Mathematics and Science, Department of Education, University of Groningen, Nijenborg 4, 9747 AG, Groningen, The Netherlands. Tel.: + 31-50-563-4941. Email: fpvos@hotmail.com Educational Research and Evaluation Vol. 11, No. 2, April 2005, pp. 141 – 154 ISSN 1380-3611 (print)/ISSN 1744-4178 (online)/05/020141–14 ª 2005 Taylor & Francis Group Ltd DOI: 10.1080/13803610500110794