Measuring Absolute and Relative Levels of Policy Support using Conjoint Choice Experiments

Abstract

Public support is a crucial, often necessary, component of political feasibility for policy proposals. Conjoint experiments are commonly utilised to assess support for policies, particularly how public opinion varies by specific policy instruments, and packaged designs. Yet, while robust methodologies have been identified to compare relative levels of support between policy instruments, current strategies often limit substantive interpretations of the absolute level of support. Given the importance of majority support thresholds across democratic settings, identifying absolute levels of support can provide critical information when assessing the feasibility of policy proposals. Here, we empirically explore a simple methodological advancement -- how do estimations of relative and absolute levels of support vary by discrete choice, proposal vote and binary ratings response methods? Drawing upon evaluations of support for carbon taxation in the United States as a case study, we find that similar levels of relative support can be found across response methods, but that the absolute level varies substantially by method and whether abstention is allowed. We further evaluate response methods by efficiency and consistency, to develop a set of recommendations towards utilising multiple response items to simultaneously assess relative and absolute levels of support for policies in conjoint-experimental designs.