Search
Browse By Day
Browse By Time
Browse By Person
Browse By Mini-Conference
Browse By Division
Browse By Session or Event Type
Browse Sessions by Fields of Interest
Browse Papers by Fields of Interest
Search Tips
Virtual Exhibit Hall
Change Preferences / Time Zone
Sign In
X (Twitter)
Efforts to study historical public opinion and policy responsiveness in the American states have been hampered by a dearth of surveys with random samples sufficiently large to estimate subnational opinion. Recent developments in opinion simulation with Multilevel Regression and Poststratification (MRP) have improved on older methods and allow scholars to more accurately estimate state-level public opinion. However, MRP and its predecessors generally fail to recover reliable estimates when polls' respondents are the result of a cluster sample, which was ubiquitous in polling into the 1980s as a technique to produce nationally representative samples, but which generally was not designed to produce representative samples at the state level. In this paper, I adapt and extend MRP to estimate state-level opinion with cluster-sampled polls by modeling the portion of estimated opinion that is usually generated with a state-level random effect. I validate my method by using the Cooperative Congressional Election Study (CCES) to simulate cluster-sampled polls. I then use it to estimate state-level opinion on relevant policy issues and apply them to the basic substantive findings of state legislative responsiveness by Erikson, Wright, and McIver (1994).