Who Listens to Whom? Assessing Inequalities in Representation

In a democratic system, elected officials are expected to represent their constituents' interests and preferences when shaping public policy. But in a diverse society such as the U.S., where people often have divergent or conflicting views, a question arises about whose preferences actually prevail when lawmakers shape policy. Recent work by Larry Bartels and Martin Gilens, sponsored by the Russell Sage Foundation, suggests that policymaking at the national level primarily reflects the interests of those who are more affluent, while tending to ignore the preferences of those in the middle or lower portion of the income distribution. Professors Justin Phillips and Jeffrey Lax at Columbia University propose to extend on this work by carrying out similar analyses at the state level. First, they intend to address the question of whether and how policy enacted at the state level reflects the public preferences for those policies across different income groups. And second, rather than looking at specific policy outcomes, Lax and Phillips will examine the association between the roll-call votes of individual lawmakers and their constituents' policy preferences by income levels. To carry out these analyses, Lax and Phillips will draw upon a wide variety of data from a number of different sources. Information on current and historical state policies will be obtained from a number of different secondary sources, including advocacy groups, policy foundations, and research organizations that have compiled such information over time. The polling data used to estimate state-level public preferences from 1970 to the present comes from the iPoll database and survey modules designed and implemented by Lax and Phillips for the 2010, 2011, and 2012 Cooperative Congressional Election Study. In their first set of analyses, Lax and Phillips propose to examine the relationship between state policy and public preferences by income terciles for the 54 different policies on which they will generate preference estimates. They start by exploring the association between the preferences of the different income groups and policy outcomes using a variety of fixed or random effects logistic regression models. They then propose to examine differential responsiveness of individual members of Congress to constituent preferences, aggregated by state or by congressional district, using roll-call votes on specified policy issues. The focus of this analysis will be on congressional votes from the past four legislative sessions and will focus on votes by members of Congress on economic, social, and foreign policy issues such as healthcare reform, the stimulus bill, and the Federal Marriage Amendment.