The issue of arriving at agreement over outcomes in summative assessment of portfolios has been a major concern, given the complexity of the assessment task, the educational and political context, and the widespread and growing use of portfolios in higher education. This article examines research findings in this area. The discussion takes place in a philosophical and theoretical context. The first section of this article considers various approaches to portfolio assessment (e.g. positivist, interpretivist, feminist) and the assumptions underlying them. The second section examines research findings in these different traditions, pointing out the findings and gaps in each, as well as suggesting potential meeting points between them. The article does not argue that any approach offers all answers to all problems connected with agreement over outcomes in portfolio summative assessment. Rather, the purpose of this article is to clarify the choices facing assessors. The underlying issues raised here have relevance to other methods of assessment, apart from portfolios.