American Public Opinion in the 1930s and 1940s: The Analysis of Quota-Controlled Sample Survey Data

Abstract

The 1930s saw the birth of mass survey research in America. Large public polling companies, such as Gallup and Roper, began surveying the public about a variety of important issues on a monthly basis. These polls contain information on public opinion ques- tions of central importance to political scientists, historians, and policy- makers, yet these data have been largely overlooked by modern researchers due to problems arising from the data collection methods. In this article I provide a strategy to properly analyze the public opinion data of the 1930s and 1940s. I first describe the quota-control methods of survey research prevalent during this time. I then detail the problems introduced through the use of quota-control techniques. Next, I describe specific strategies that researchers can employ to ameliorate these prob- lems in data analysis at both the aggregate and individual levels. Finally, I use examples from several pubic opinion studies in the early 1940s to show how the methods of analysis laid out in this article enable us to utilize historical public opinion data. The study of political behavior has flourished in recent decades with the development of long-term, time series data collected by both academic survey researchers and commercial pollsters. Our understanding of the dynamics of mass opinion and behavior prior to the 1950s has, however, been more limited. Expanding political behavior research to this earlier era is valuable ADAM J. BERINSKY is an associate professor of political science at the Massachusetts Institute of