Institutions are central to the functioning of a complex society, and they play a very important role in the production and use of knowledge. The institutional representatives operating in such settings develop routines in order to handle the complexities of their work tasks. Creating routines implies simplifying things. It makes the environment predictable, and the tasks re- cognisable.One significant element of creating routines is the categorising and classifying of things, events and people in order to make sense of them. In fact, people simply cannot describe or deal with reality without invoking categories. This is one of the "most fundamental principles of human thought and action" (Edwards, 1991, p. 515) in any context. This implies that categories are instrumental in perspectivising events and objects within institutional practices. Thus, categorisation is part of human action. We do something in talk when using categories. They are not just names for things and relationships, they do a real job by signifying and informing us how to act in a specific context. Hence, by categorising, institutional representatives know how to handle the issue in question, and which steps are relevant to take. This implies, that through the use of categories in institutional practices, people become 'transformed' into entities that the organisation can recognize and process (Lipsky, 1980). In school as an institutional activity categorising is an omnipotent social practice. The problem of how to handle pupils who are considered not to meet the expectations of the school is a dilemma, which has to be solved. Categories are necessary resources in this process. The aim of this paper is to analyse the in situ use of different categories when the institutional representatives meet the parents of a child experiencing difficulties in school. In Sweden, this is an activity that occurs routinely within so-called pupil welfare conferences where the child's parents and different types of expertise, for example psychological and pedagogical, participate. Together they are making a decision of how to solve the dilemma and what institutional resources are suitable.The empirical setting in which the study has been conducted is a comprehensive school. More specifically, I have followed a child and the pupil welfare conferences related to his difficulties during two years. These meetings have been audio-recorded and later transcribed. The data presented here were generated during these meetings.The findings show that the meetings are characterized by ambiguity and competing ways of representing the child's problems. The professionals represent the alleged problems as resident within the child and caused by a medical description. The parents, on the other hand, justify the pupil's behaviour by invoking normality. For example, they say that 'he is just as us', 'the other children behave in the same way', 'my husband did also lose his temper'. Thus, reaching a common understanding among the participants in the meeting is a complex and conflicting process where different and competing ways of representing and giving meaning to events occur. This is an issue of 'politics of representation'. In this case the institutional representatives succeed in establishing the identity of the child as an ADHD-pupil. Edwards, D. (1991). Categories are for talking. On the cognitive and discursive bases of categorization. Theory & Psychology, 1(4), 515- 542.Lipsky, M. (1980). Street-level bureaucracy. Dilemmas of the individual in public services. New York, NY: Sage.