It is well known that network coding can improve the secrecy of a network that is represented by a graph with edges tapped by an eavesdropper. However, the application of such secure network coding to emerging problems in information-theoretic secrecy in wireless networks has been limited; in particular, the graph-based wiretap network approach does not readily extend to the physical wireless network. In this paper, we consider the impact of such an extension on scaling laws for wireless networks. We first employ simple examples to illustrate how an extension of the secure network coding approach can address difficult problems that have plagued wireless network security; for example, relaxing the known eavesdropper location assumption or avoiding the use of artificial noise generation. Based on this understanding, we then add secure network coding approaches to recent constructions for secrecy scaling in large wireless networks, most notably achieving secure pernode throughput in large wireless networks in the presence of an arbitrary number of non-collaborating eavesdroppers.