The replication problem and its implications for policy studies

Abstract

In this article we re-examine the epistemological basis for claims to ‘best evidence’ and ‘best practice’ in policy studies by tracing these to a Popperian model of theory development and testing. This model outlines how only scientific explanations that survive repeated testing count as good theory. Yet repeated testing (replication) is scarce across the social sciences – this is the ‘replication problem’. More specifically, the lack of replications in policy studies undermines the epistemological basis for policy transfer based on ‘best practice’ and ‘best evidence’. To resolve this, we offer an innovative explanation of the replication problem drawing on Foucault's concept of Episteme. In doing so, we outline two respectively different accounts of replication: ‘scientific project,’ and ‘aesthetic object.’ The latter offers alternative bases from which to pursue ‘best practice’ and ‘best evidence’.