This submission argues that all knowledge-related policy needs to be evaluated from the perspective of knowledge research. Because science and innovation policies are knowledge-related policies, we argue that to design such policy well and to adequately evaluate the return on investment by government in them, government needs to employ appropriate analytical frameworks that understand knowledge in economic, social and cultural terms. We point out that while highly developed evaluation instruments do not yet exist for these purposes the fundamental theory does. Furthermore, much of this fundamental (ontological) theorisation of knowledge in knowledge-based economies and societies has been done in Australia, giving Australia a competitive advantage in knowledge-related policy development. This theorisation of knowledge demonstrates that knowledge is more than fact and information, and that it is a profoundly social and cultural phenomenon. Equally important, and flowing from this observation, is that knowledge exists in networks and is systemic. Knowledge, therefore, is highly relational. We argue that it is the relational nature of knowledge that should be a focus in any analysis or evaluation of the benefits of public support for science and innovation. In this submission, we briefly discuss the nature of knowledge in knowledge-based economies and societies, and demonstrate how by imposing the intellectual discipline of research driven analytical frameworks, important evaluative questions are foregrounded (which might otherwise be overlooked). Finally, we point out the need to conduct collaborative research to determine how best to translate the theoretical discoveries about knowledge into usable analytical or evaluative frameworks for government.