Large representative surveys have become a valuable resource to inform public policy in an increasingly complex modern world. They provide authority to policy since they are considered objective, neutral and scientific. In contrast, this article conceives the production of knowledge as an interactive process. We argue that the conduct of large social surveys tends to reinforce existing world views, power relations and a narrow construction of social issues. To illustrate this, we draw on a small exploratory study which examined the experience of responding to selected survey questions of the Household Income and Labour Dynamics in Australia study (HILDA). We suggest that while more open approaches are required to capture the complexities of everyday life, these are unlikely to be implemented given the dominance of particular forms of knowledge.