Constructing and Repairing Our Bridges: Statistical Considerations When Placing Agents into Legislative Preference Space

Abstract

A statistical test of hypotheses regarding the strategic interaction between legislators and third-party agents, such as lobbyists, bureaucrats, or experts, requires some ``bridging'' method to place each type of actor into preference spaces that are comparable. Current solutions to the bridging problem either attempt to place both legislators and agents into an arbitrary preference space entirely disconnected from the institutional setting, or they attempt to place agents into a legislative roll-call preference space mistakenly as if agents were themselves legislators. I propose a new method that leverages the observed behavioral hypotheses to identify a set of agent-specific bridging parameters that place agents directly into legislative roll-call preference space as agents, rather than counterfactually as legislators. I apply my method to test whether members of Congress condition their questioning of witnesses in committee hearings on preference similarity within the legislator-witness dyad, as a test of lobbying models for strategic information transmission.