Using Bayesian Aldrich‐McKelvey Scaling to Study Citizens' Ideological Preferences and Perceptions

Abstract

Aldrich-McKelvey scaling is a powerful method that corrects for differentialitem functioning (DIF) in estimating the positions of political stimuli (e.g., parties and candidates) and survey respondents along a latent policy dimension from issue scale data. DIF arises when respondents interpret issue scales (like the standard liberal-conservative scale) differently and distort their placements of the stimuli and themselves. We develop a Bayesian implementation of the classical maximum likelihood Aldrich-McKelvey scaling method that overcomes some important shortcomings in the classical procedure. We then apply this method to study citizens' ideological preferences and perceptions using data from the 2004-2012 American National Election Studies and the 2010 Cooperative Congressional Election Study. Our findings indicate that DIF biases self-placements on the liberal-conservative scale in a way that understates the extent of polarization in the contemporary American electorate and that citizens have remarkably accurate perceptions of the ideological positions of Senators and Senate candidates.

What Makes Experts Reliable?

“…This truncation is necessary for identification purposes, and rests on the relatively safe assumption that experts code in the right direction. Note that the model also accounts for systematic biases in how experts translate perceptions into ratings-a common concern in surveys that rely on multi-rater judgment (Aldrich and McKelvey, 1977;Bakker et al, 2014;Hare et al, 2015)-through γ, a k = 1, ..., n vector of threshold parameters specific to each expert. As a result, β estimates the degree to which experts stochastically diverge from other experts who coded the same cases, conditional on her scale perception.…”

Working Class Hero? Interrogating the Social Bases of the Rise of Donald Trump

“…Baldassarri and Gelman 2008;Abramowitz 2010;Hare et al 2015). In spite of all the noise generated by the unusual Trump candidacy and the 2016 campaign, a remarkably strong partisan divide largely held to form.…”

Approaches to Studying Policy Representation

“…Capturing data on citizens' views on the same issues that legislators have taken positions on and examining a hypothesis for each issue requires exceptional effort (e.g., Gilens 2012;Lax and Phillips 2012;Lloren and W€ uest 2014;Matsusaka 2015). Constructing a "joint scale" that bridges politicians and voters may require only a few points of overlap (e.g., Bafumi and Herron 2010;Barber 2014;Jessee 2009;Shor 2013), or potentially none (Aldrich and McKelvey 1977;Hare et al 2014;Ramey 2014). Unfortunately, studying citizen's policy preferences is simply not as easy as these methods imply.…”