This past Friday marked the 60th day of the 2011 Regular Session of the Texas House of Representatives. During this period the House held roll call votes on several procedural and internal housekeeping matters as well as on HB 15 (requiring a sonogram be done before an abortion) and a variety of amendments related to HB 15. In all a total of 50 votes were at least minimally contested during this time frame, providing a sufficient number of votes to conduct a preliminary analysis of the location of Texas House members on the Liberal-Conservative ideological dimension along which most votes in the Texas House take place.
Before proceeding, two caveats are in order. First, the number of votes thus far is very small, with a representative’s location on the Liberal-Conservative dimension heavily influenced by their position vis-à-vis the issue of abortion. Therefore, it is quite possible that some representatives will not find themselves in a similar location once votes on other issues are taken into account.
The second caveat is that given the relatively small number of votes, in contrast to analysis conducted last year where the DW-NOMINATE procedure was utilized, for this preliminary analysis of the 2011 roll call vote data I employ a Bayesian estimation procedure (IDEAL) developed by Stanford University professor Simon Jackman. Due to the small number of votes, in addition to plotting each representative’s location on the Liberal-Conservative dimension, I also provide a 95 percent credible interval (CI) for this point estimate. Only when a representative’s CI does not overlap with that of another representative can we say with any real certainty that their respective locations on the Liberal-Conservative dimension are credibly distinct.
These three downloadable documents provide the representatives’ locations (and 95 percent CIs) on the Liberal-Conservative dimension for: all representatives, Republican representatives and Democratic representatives. Based on a review of these figures and the data used to create them, three principal points stand out regarding legislator floor behavior over these first sixty days.
First, unsurprisingly, to date there exists a high level of partisan polarization in floor voting behavior in the Texas House. With the sole exception of Republican Sarah Davis (West University Place), the mean ideological location of every Republican (red dots) is more conservative than that of every Democrat (blue dots). Furthermore, even taking in account the representatives’ respective CIs, there is little partisan overlap. The CIs of only 12 of the 100 Republicans (Speaker Joe Straus normally does not vote) overlap those of one or more Democrats, and, with the exception of Sarah Davis, the maximum number of Democrats with whom any of these House Republican overlaps is five. By the same token (excluding Sarah Davis), the CIs of only five of the 49 Democrats overlap those of one or more Republicans.
Second, the data allow us to roughly divide the Republican delegation into conservative, moderate and liberal wings utilizing each representative’s mean ideological location and 95 percent CI. Given the limited number of votes and consequent uncertainty regarding the representatives’ true ideological location, to differentiate among the members of the delegation I employ a conservative (albeit still arbitrary) approach based on the presence or absence of CI overlap among the delegation members.
In the conservative wing of the Republican Party are those representatives located at the right end of the ideological continuum whose CI does not overlap with a fifth or more of their fellow party members. Moving from most conservative to least conservative, the members of this group are (nine total): Charlie Howard (Sugar Land), Leo Berman (Tyler), Bill Zedler (Arlington), Debbie Riddle (Tomball), Tan Parker (Flower Mound), Wayne Christian (Center), Randy Weber (Pearland), Ken Legler (Pasadena), and Jim Landtroop (Plainview). In the Liberal wing of the Republican Party are those representatives located at the left end of the ideological continuum (for members of the Republican delegation) whose CI does not overlap with a fifth or more of their copartisans. Moving from least conservative to most conservative, the members of this group are (six total): Sarah Davis (West University Place), David Simpson (Longview), former Democrat Allan Ritter (Nederland), Beverly Woolley (Houston), Paul Workman (Austin), and John Davis (Houston). The remaining 85 Republican representatives are at present located in the Moderate category.
It is quite possible that at least some of the Republicans with current ideological locations at the liberal extreme of the Republican delegation will move significantly back toward the center of the party distribution as future votes are cast. For example, the voting behavior in prior sessions of Beverly Woolley, and to a lesser extent John Davis, would lead me to expect to find them (particularly Woolley) closer to the party median by the end of the session.
Third, the data allow for a similar division of the Democratic delegation into conservative, moderate and liberal wings. Using the same approach as above, three Democrats are located within the party’s Liberal wing. Moving from most liberal to least liberal, they are: Roland Gutierrez (San Antonio), Lon Burnam (Fort Worth), and Borris Miles (Houston). At the other end of the spectrum (in the Conservative wing) are six Democrats. Moving from most conservative to least conservative, they are: Chente Quintanilla (Tornillo), Joe Pickett (El Paso), Jose Manuel Lozano (Kingsville), Sergio Muñoz Jr. (Palmview), Ryan Guillen (Rio Grande City), and Tracy King (Batesville). The remaining 40 Democratic representatives are at present located in the moderate category.
After only sixty days, and a relatively small number of roll call votes in a very limited number of thematic areas, these data on ideological location should be viewed much in the same way as college football rankings in September. Some useful information is provided, but at the same time significant changes are likely to occur as the session/season progresses. In this case, the large number of votes which will be cast between now and the end of May will allow for both a much more accurate measure of each representative’s true location on the Liberal-Conservative continuum, as well as for a much more precise differentiation among the representatives than is presently possible.
Mark P. Jones is the James A. Baker III Institute for Public Policy’s Fellow in Political Science as well as the Joseph D. Jamail Chair in Latin American Studies and Chair of the Department of Political Science at Rice University.