Artificial Extremism in Interest Group Ratings
and the Preferences versus Party Debate
Legislative Studies Quarterly XXIV:525-42

Congressional voting studies often use interest group ratings as proxies for legislator policy preferences. This paper investigates the extent to which artificial extremism in interest group ratings affects the ability of such studies to estimate accurately the impact of legislator preferences and party membership on roll-call votes. Using a sequence of Monte Carlo experiments, I find that artificial extremism does not have serious implications for understanding whether policy preferences impact legislator voting behavior. However, in many cases artificial extremism can cause analyses of roll-call votes to draw improper conclusions regarding the direction and magnitude of the impact of party membership on roll-call voting decisions.

  Go to next abstract

  Return to 1999 Titles

  Return to LSQ home page