Skip to content


We use the positions advocated in amicus curiae (“friend of the Court”) briefs filed in the 1953 through 2013 Supreme Court Terms to estimate the ideal points of organized interests and governments (i.e., the states and the federal government, as represented by the solicitor general) in the Supreme Court’s legal policy space. We treat these amicus brief-based “votes” on cases as analogous to the votes cast by the justices in these cases, which lets us estimate the locations of these actors and the justices in the same policy space. Using these “votes” by interests, governments, and justices, we estimate item response theory (IRT) models that treat the ideal points of these actors as a latent, unobservable trait to be estimated via Bayesian Markov chain Monte Carlo methods.

Because organized interests and governments can choose the cases in which to “vote” (i.e., file amicus briefs) it is not safe to assume that abstentions can be treated as missing-at-random (MAR). The logic of the spatial voting model underlying ideal point estimation models typical approach implies that these abstentions should be a function of the location of the actors, which suggests that these abstentions are not random. We therefore employ a recent extension of the IRT ideal point estimation model designed to account for nonresponses or abstentions (developed by Rosas, Shomer, and Haptonstahl).* This abstention-allowing IRT model also allows actors to have different baseline rates of voting that are unrelated to spatial considerations. We use Martin and Quinn’s (2002) informative priors for a handful of the justices, which orient our ideal point estimates so that smaller values (often negative) correspond with what might be viewed as liberal positions and larger values correspond with what might be viewed as conservative positions.**

Using the above approach, we estimate static ideal points for the 600 most active organized interests, the U.S. states (with ideal points that are allowed to vary from attorney general to attorney general), and U.S. solicitors general (including several acting solicitors general), and the justices.

For more details on the methodology, see our working papers.

* Rosas, Guillermo, Yael Shomer, and Stephen R. Haptonstahl. 2015. “No News is News: Non-Ignorable Non-Response in Roll-Call Data Analysis.” American Journal of Political Science 59(2):511-528. The specific version of the model we use is presented in the Supplemental Information for this article.

** Martin, Andrew D., and Kevin M. Quinn. 2002. “Dynamic Ideal Point Estimation via Markov Chain Monte Carlo for the U.S. Supreme Court, 1953–1999.” Political Analysis 10:134-53.




This material is based upon work supported by the National Science Foundation under Grant No. SES-1351922. Any opinions, findings, and conclusions or recommendations expressed in this material are those of the authors and do not necessarily reflect the views of the National Science Foundation.