Abstract
The purpose of this paper is to analyze the effect of primary voting systems on the ability of agent-representatives to deviate within the median voter model. While conclusions are consistent with the results found in Gerber and Morton (1998), this paper extends their analysis by including the role played by electoral security and the extant incentives that accompany it. The results presented herein make 2 important contributions to the literature. First, while the report has been consistent with the literature in maintaining that deviation is driven by ideology, these results reflect that institutional arrangements also allow for deviation. Second, the measurement combining ADA and ACU scores reduces the potential of liberal bias or conservative bias by combining these scores into the determination of the dependent variable. Results suggest that more open primaries produce candidates with positions that are closer to those of the median voter. If blanket primaries encourage broader electoral participation by voters, the resulting median voter in such primaries will more closely resemble the median voter in the entire constituency.