This thesis deals with evidence in social work and has the following problem formulation: "how evidence-articulated discourse among researchers and policy makers, and cast these discourses
consistent with discourses in the practical management of social work?"

The thesis has a discourse analytic approach, based on Fairclough's theory. We have analyzed a series of key publications of agents dealing with evidence, and have also completed five qualitative interviews with informants who work with or have received evidence-based deployed MST (Multi Systemic Therapy) in Herning Municipality. In addition, the thesis has also had a management technology
approach, which has meant that we, alongside the analysis have identified management practices
based on Dean’s governmentality analysis which is a development of Foucault's governmentality-idea.

We identified a variety of discourses among the selected agents occupied with evidence,
to identify the agents' way of talking about and understanding the concept of evidence. Among the agents we have ascertained conflicting discourses, but there is consensus that the evidence covers
universal and certain knowledge, yes with divergent views on which research design contributes to
this knowledge.

We have ascertained that the identified discourses of the agents occupied in evidence, and
the differences, also can be identified in practice. Furthermore a series of discourses in the
social practice is identified that are not found by the agents within the evidence-movement.

We have also reached to the conclusion that users do not articulate discourses from the evidence-movement, but primarily refers to the discourses identified in the practical social work. In addition, they articulate a user discourse, where the evidence founding of the social work plays no role. The fact that the evidence is irrelevant for the families in practice is a crucial point for this thesis, since it is in contradiction with discourses articulated by the agents occupied with evidence on a higher social level.

We have ascertained that the users are transformed from objects when they turn to social services, to subjects with an improved ability to self-governance and a greater independence of social assistance systems after the MST cycle. The aim has been to make the implicit thinking about obvious acts explicit, as the identification of management practices have demonstrated.