Testosterone recognitive networks were synthesized varying feed crosslinking percentage and length of bifunctional crosslinking agent to analyze the effect of manipulating structural parameters on template binding parameters such as affinity, capacity, selectivity, and transport. Crosslinking agent was varied from 50 to 90% and associated networks experienced a twofold increase in capacity and a fourfold increase in affinity, with equilibrium association constants ranging from 0.30 +/- 0.02 x 104 M-1 to 1.3 +/- 0.1 x 104 M-1, respectively. The higher concentration of crosslinking monomer increased crosslinking points available for interchain stabilization creating an increased number of stable cavities for template association. However, by increasing the length of crosslinking agent approximate to 4.5 times, the mesh size of the network increased resulting in 40% faster template PEG20ODMA) recognitive network had an association constant of (0.20 +/-0.05) x 10(4) M-1 and bound (0.72 +/- 0.04) x 10(-2) mmol testosterone/g dry polymer which was less by six and threefold, respectively, compared to a similarly crosslinked poly(NMA-co-EGDMA) recognitive network. Structural manipulation of the macromolecular architecture illustrates the programmability of recognitive networks for specific binding parameters and diffusional transport, which may lead to enhanced imprinted sensor materials and successful integration onto sensor platforms. (c) 2007 Wiley Periodicals, Inc.