1. See Converse, Philip E., “The Nature of Belief Systems in Mass Publics,” in Ideology and Discontent, ed. Apter, David (New York: Free Press, 1964)Google Scholar; Poole, Keith T. and Rosenthal, Howard, Ideology and Congress (New Brunswick, NJ: Transaction Publishers, 2011)Google Scholar.

2. McCarty, Nolan, Poole, Keith T., and Rosenthal, Howard, “The Hunt for Party Discipline in Congress,” American Political Science Review 95 (2001): 673–87Google Scholar.

3. Poole, Keith T., “Changing Minds? Not in Congress!” Public Choice 131, no. 3 (2007): 435–51Google Scholar. I do agree with Bateman-Lapinski and Caughey-Schickler that the cases of meaningful ideal point change can be historical and substantively important. But these changes don't negate my general argument about ideo-lite as intellectuals also occasionally change their theory-seminar ideologies.

4. These are the same-party pairs from KY, MA, MN, SC, TX, UT, and WV. There would be eight if we included the difference between “socialist” Bernie Sanders and Democrat Patrick Leahy in Vermont.

5. See Poole, Keith T. and Romer, Thomas, “‘Ideology,’ ‘Shirking,’ and Representation,” Public Choice 77, no. 1 (1993): 185–96Google Scholar.

6. For discussions of parties as “long” coalitions, see Thomas Schwartz, “Why Parties?” (Unpublished manuscript, Department of Political Science, University of California, Los Angeles, 1989); Aldrich, John H., Why Parties? The Origin and Transformation of Party Politics in America (New York: Cambridge University Press, 1995)CrossRefGoogle Scholar; Bawn, Kathleen et al. , “A Theory of Political Parties: Groups, Policy Demands and Nominations in American Politics,” Perspectives on Politics 10, no. 03 (2012): 571–97Google Scholar.

7. These ideal points are estimated using W-NOMINATE for R. See Poole, Keith T. et al. , “Scaling Roll Call Votes with W-NOMINATE in R,” Journal of Statistical Software 42, no. 14 (2011): 1–21 CrossRefGoogle Scholar.

8. Nokken, Timothy P. and Poole, Keith T., “Congressional Party Defection in American History,” Legislative Studies Quarterly 29, no. 4 (2004): 545–68Google Scholar. These scores are discussed in more detail in Section 4.

9. Nokken and Poole, “Congressional Party Defection in American History.”

10. Clinton, Joshua, Jackman, Simon, and Rivers, Douglas, “The Statistical Analysis of Roll Call Data,” American Political Science Review 98, no. 02 (2004): 355–70Google Scholar.

11. Groseclose, Tim, Levitt, Steven D., and Snyder, James M., “Comparing Interest Group Scores Across Time and Chambers: Adjusted ADA Scores for the US Congress,” American Political Science Review 93, no. 1 (1999): 33–50 Google Scholar.

12. Thus, it is itself susceptible to the criticisms that Bateman and Lapinski levied in Section 2 of their article in this issue.

13. Martin, Andrew D. and Quinn, Kevin M., “Dynamic Ideal Point estimation via Markov Chain Monte Carlo for the US Supreme Court, 1953–1999,” Political Analysis 10, no. 2 (2002): 134–53Google Scholar.

14. The difference between the Martin-Quinn model and GLS can be subtle in that each can produce any sequence of ideal points for a legislator. But the main difference concerns how each procedure adjusts ideal point estimates to ensure comparability over time. When the Martin-Quinn model estimates the scale for period t, it relies upon the estimates of x _t−1. GLS relies upon the overall average ideal points $\bar x$ .

15. In more technical language, the parameter σ ² is not identified.

16. Unfortunately, Caughey and Schickler do not report the value of σ ² or describe the necessary robustness checks.

17. Nokken and Poole, “Congressional Party Defection in American History.” Another newer alternative is Adam Bonica's dynamic extension of Keith Poole's nonparametric optimal classification (OC). See Bonica, Adam, “The Punctuated Origins of Senate Polarization,” Legislative Studies Quarterly 39, no. 1 (2014): 5–26 Google Scholar; Poole, Keith T., “Non-parametric Unfolding of Binary Choice Data,” Political Analysis 8 (2000): 211–37Google Scholar. The OC model estimates ordinal rankings of legislator ideal points instead of cardinal measures. Since rank orderings are independent of changes to the scale, OC need not worry about bridging across time. Movement can be assessed, for example, by noting that a member has shifted from the 85th percentile of conservatism to the 90th percentile. Since there is no issue of comparability over time, Bonica suggests estimating ideal points over fixed intervals of the agenda, such as every 200 votes. To minimize the likelihood of idiosyncratic shifts, Bonica suggests a smoothing process where a legislator's ideal point as on vote t is based on the votes from t − 200 to t − 1, while the ideal point at time t + 1 is based on the votes from t − 199 to t. Thus, Bonica's procedure produces a new ideal point ranking at each vote without the strong assumptions required of the GLS model (which are then violated if a legislator does in fact change position).

18. The difference may be the result of the fact that Caughey and Schickler estimated their model only on explicitly economic votes.

19. McCarty, Nolan, Poole, Keith, and Rosenthal, Howard, Polarized American: The Dance of Ideology and Unequal Riches (Cambridge, MA: MIT Press, 2006)Google Scholar.

20. Ibid., Figures 2.19a–c.

21. For each roll call, DW-NOMINATE estimates of directed distances d ₁ and d ₂ and the dimensional weight w. If a legislator's two-dimension ideal point is $(x_1,x_2)$ , then her projected ideal point is ${\textstyle{{d_1x_1 + wd_2x_2} \over {\sqrt {d_1^2 + w^2d_2^2}}}} $ .

22. Mian, Atif, Sufi, Amir, and Trebbi, Francesco, “The Political Economy of the US Mortgage Default Crisis,” American Economic Review 100, no. 5 (2010): 1967–98CrossRefGoogle Scholar.

23. These results further underscore that DW-NOMINATE scores are not simply measures of partisanship. DW-NOMINATE explains significant within-Republican party variation on a set of measures proposed by a Republican administration.

24. For a nontechnical discussion, see McCarty, Nolan, “Measuring Legislative Preferences,” in Oxford Handbook of the American Congress, ed. Lee, Frances and Schickler, Eric (New York: Oxford University Press, 2011), p. 82 Google Scholar.

25. Of course, this feat is accomplished because the projection procedure leverages information from all of the other votes and the relationship between the vote in question and the basic DW-NOMINATE model.

26. These ideal points are estimated using W-NOMINATE for R. See Poole et al., “Scaling Roll Call Votes with W-NOMINATE in R.”.

27. Recall from Section 4 that the second dimension may also explain regional differences on votes when the connection to race relations is not explicit.