Abstract: | Compulsory community treatment for people with severe mental illness remains controversial due to conflicting research evidence. Recently, there have been challenges to the conventional view that trial‐based evidence should take precedence. This paper adds to these challenges in three ways. First, it emphasizes the need for critiques of trials to engage with conceptual and not just technical issues. Second, it develops a critique of trials centred on both how we can have knowledge and what it is we can have knowledge of. Third, it uses this critique to develop a research strategy that capitalizes on the information in large‐scale datasets. |