The simple question of what constitutes school results is rarely asked, even though schooling has become increasingly results-oriented in the Western world in recent decades. One measure that is used while being relatively unproblematised is student grades. How students perform, as assessed by their teachers in the form of formal grades, is often treated as if it is highly informative about school quality. In this paper, we illuminate how internal assessments of pupils’ knowledge, carried out by their teachers, becomes external data on the quality of education: a learning outcome. Teacher-assigned grades have a long history in the Swedish school system, as a pedagogical tool and as information for parents and employers. In a first analysis portrays a development, from the increased emphasis on grades used for merit based admission through the rise of New Public Management thinking from the 1980s to the 2010s (Lundahl, Erixon Arreman, Holm & Lundström, 2013) where grades more and more are used for accountability purposes. The methodological approach is to systematically follow arguments concerning how grades can be used as performance data in relevant committee work, governmental reports (SOU) and propositions. In a second analysis, we follow the political argument to widen and strengthen the accountability of the Swedish grading system by references to other countries’ grading systems. The argument is that in order to achieve better PISA rankings, Sweden needs a system that formally grades children at a younger age than normally done in Sweden. We analyse the legitimisation of this policy by investigating the governments’ referral to other European countries’ grading and accountability systems and how these are represented in Eurydice, which was a principal source in the government’s legitimation. Eurydice holds the most accessible knowledge regarding educational systems in Europe. Over all we treat the development for how teachers' internal assessments become legitimate external data on the quality of education, as a process of legitimation in which grading works as a ‘quick language’: a way to reduce complexity by creating a common language to enable a smooth transfer of information in the field of education (Lundahl 2008).
