This appendix presents the results of analyses of ideology and spatial voting using the 2004 and 2008 American National Election Studies (ANES) Time Series Studies. Because the ANES is the most commonly used data source in American political behavior research, readers may be curious whether the results in the main text of this book can be “replicated” by analyzing the ANES. As discussed in the main text (see Chapter 3 in particular), the measures used in the ANES are not ideally suited to an investigation of spatial voting. Accordingly, to “replicate” the book's main analyses using ANES data necessarily entails strong (I would argue heroic) assumptions. This exercise is still likely to be useful, however, both because it highlights the benefits of the main measurement approaches used in the book in contrast to the standard approach in the field and also because of the centrality of the ANES to American political research.

The main ANES measure of ideology is a simple ordered seven-point scale on which respondents are asked to place themselves from “extremely liberal” to “extremely conservative.” For better or worse, these items have provided the most commonly used measures of the ideological positions of survey respondents in both the ANES and other political surveys. The ANES also asks respondents to place presidential candidates on this same scale. The most common method for attempting to compare the positions of candidates and voters on such self-placement scales is to assume that the means of respondent placements of each candidate are accurate.