The Politics of Problem Definition: Applying and Testing Threshold Models


  • Thanks to James E. Anderson, Judith Baer, Kim Q. Hill, Nicki Van Hightower, Bryan Jones, Timur Kuran, Jan Leighley, Kenneth Meier, and anonymous reviewers for helpful comments. All errors are the responsibility of the authors.

B. Dan Wood is a University Faculty Fellow and Professor of Political Science, Texas A&M University, College Station, TX 77843-4348 ( Alesha Doan is Assistant Professor of Political Science, California Polytechnic Institute, San Luis Obispo, CA 93407 (


The key to understanding agenda setting is how and why conditions become defined as public problems. This process has been dubbed problem definition by past literature, yet past work offers little more than descriptive accounts of the process. We apply a threshold model to explain how problem definition can exhibit both incremental and punctuated change. We then demonstrate the model by explaining redefinition of the sexual harassment issue following the Clarence Thomas Supreme Court nomination hearings. Using Box-Tiao (1975) and nonlinear time varying parameter regression methods, we show that the hearings permanently altered individual and social nonacceptance of sexual harassment, as well as initiated a process of positive feedback. Given that there are many conditions similar to sexual harassment, and given that disturbances are quite frequent, sudden change should be a common feature of the American system.