Public involvement in research has evolved over the last two decades in a culture dominated by the principles of evidence-based medicine. It is therefore unsurprising that some researchers have applied the same thinking to involvement, particularly to involvement in research projects. This may explain why they tend to conceptualise involvement as an intervention, seek to evaluate its impact in the same way that treatments are tested, highlight the need for an evidence-base for involvement, and use the language of research to describe its practice and report its outcomes. In this article we explore why this thinking may be unhelpful. We suggest an alternative approach that conceptualises involvement as ‘conversations that support two-way learning’.