I have made interviews with a teacher and with pupils from the lower grades in a school dominated by pupils with Swedish as their second language, about the role and effects of assessment with particular focus on the subject of mathematics. I have also studied the policy documents (“kursplanen”) for the subjects of Swedish and mathematics. The formulations of the learning goals in Swedish are differentiated, so that the demands are somewhat lower for pupils with Swedish as their second language. In mathematics, however, the learning goals are the same for all pupils, and they include the ability to communicate mathematical concepts in meaningful and relevant situations. The language of this communication is supposed to be Swedish, and the cultural code necessary to find the situation meaningful and relevant is often Swedish. The textbooks in mathematics for the early grades are oriented towards the learning goals, and contain many problems formulated in plain Swedish language, and with meaning and relevance often oriented to assumed fruit distribution habits among culturally Swedish children. This creates a problem with assessments in mathematics for many pupils with Swedish as their second language. They often feel that the assessments in mathematics are unfair, if they already know the numerical system and could have solved the problem, had it been given to them in a more simplified language. There is a risk that these pupils give up on the subject of mathematics too early, and for linguistic and cultural reasons. My research has shown that some of these pupils even question the mathematical ability of their teacher, because they can so easily check the numerical results on their pocket calculators, and at the same time find the verbal explanations of the teacher so difficult to understand. The problems resulting from the discrepancy between the learning goals in Swedish and mathematics, are aggravated by the recently introduced national tests in the third grade, witch give an official stamp on the use of summative assessment in the form of mathematical tests that includes verbally and culturally demanding problems. What the individual teacher in mathematics can do under these circumstances, is exploiting the rights and demands for individualization of learning to the outmost for pupils with Swedish as their second language. Children who in the future are going to use their mathematical skills and abilities professionally in the Swedish society certainly have to learn the appropriate words, as well as the cultural codes for the Swedish way of formulating mathematical problems. But their roads to achieving these particular goals in the national learning plan for mathematics, will in various ways be different from that of pupils with Swedish as their first language. The advent of national tests in mathematics in the third grade makes it even more important for the teacher to focus on the use of formative assessments, that make temporary allowance for the pupils present shortcomings in the linguistic and cultural understanding of mathematics, while at the same time setting up personal and individualized learning goals, aimed at overcoming these shortcomings.