In collaboration with Adam Berinsky of MIT, I worked for many years on a project that involved rehabilitating hundreds of older, under-utilized opinion polls. The survey data were originally collected in the 1930s-50s and are archived at the Roper Center for Public Opinion. As the project evolved, we worked closely with Devin Caughey, Sara Chatfield, Erin Hartman, and Jas Sekhon to develop an approach for re-weighting the data to correct for the problems introduced by the sampling methods used in the early years of survey research.
The original datasets, along with our recodes, the weights, and an overview of how to use the data, are available through the Roper archive ("The Berinsky Schickler American Mass Public Opinion Study.").
We published a book on our weighting approach, which has applications for addressing sampling issues in both historical and contemporary survey data, Target Estimation and Adjustment Weighting for Survey Nonresponse and Sampling Bias, Cambridge University Elements, 2020).
I have drawn upon this data in a number of articles as well as in Racial Realignment: The Transformation of American Liberalism, 1932-1965, including:
“Policy and Performance in the New Deal Realignment: Evidence from Old Data and New Methods.” Journal of Politics, 2020 (with Devin Caughey and Michael Dougal)
“New Deal Liberalism and Racial Liberalism in the Mass Public, 1937-1968.” Perspectives on Politics, 2013.
“Public Opinion, Organized Labor, and the Limits of New Deal Liberalism, 1936-1945.” Studies in American Political Development, 2011 (with Devin Caughey).