Summary: The factorial survey approach, which was first introduced in the social sciences around the beginning of the 1980s, constitutes an advanced method for measuring human judgements of people or social situations. At the general level, this quasi-experimental approach involves presenting respondents with vignettes (fictive descriptions), in which selected characteristics describing the vignette ‘person’ or ‘situation’ are simultaneously manipulated. The aim of this article is to present a conceptual and an analytical framework for factorial survey studies of professional judgements in social work. Findings: In the first part of the article, I develop and discuss the proposition that this approach may be used in order to study the contents of professional judgements about the diagnosis and treatment of clients. The ‘contents’ is discussed in terms of knowledge assumptions that practitioners explicitly and tacitly use as a basis for their professional judgements. Second, I outline a strategy for modelling social workers’ judgements. This modelling strategy proceeds from the possibilities afforded by multilevel regression analysis. Applications: Findings from analyses of factorial survey data may reveal both professional agreement and disagreement in practitioners’ judgements. While results that reveal high levels of disagreement in judgements about what constitutes a particular diagnosis or about which intervention is the most suitable for a particular client may raise questions as regards the ‘professionalism’ of practitioners’ judgements, results that reveal professional agreement in diagnostic and treatment assumptions may be transformed into hypotheses that can be tested further in research.