Bradley Spahn is a Ph.D. Candidate in Stanford's Department of Political Science and a dissertation fellow at Stanford's Institute for Research in the Social Sciences, where he studies American Politics and Political Methodology. His research uses voter file data to describe structural features of the American Electorate and improve survey methodology.
His book-style dissertation, "Before The American Voter," draws on 60 years of voter file data from the California Great Registers, spanning the period 1908 to 1968. The data, developed in collaboration with Ancestry.com, is the first representative data set covering politics before the New Deal. Consisting of over 57 million voter records matched to the census, the data follows the partisan dynamics of millions of individual California voters over time, exploring political behavior before the advent of surveys.
He uses this data to describe unique features of mass politics from 1908 to 1968. He finds that fundamental features of American political behavior like partisan stability and differentiated partisanship on the basis of class and ethnicity are historically contingent, emerging out of the New Deal Realignment. His research marshals the first direct evidence that the New Deal realignment was caused by the mass conversion of Republicans into Democrats, settling a long debate about whether the mobilization of new voters or the conversion of existing partisans accounted for the Democrats' huge gains in the 1930's.
In other work, he shows how commercial voter files used by campaigns directs voter mobilization activities away from historically underrepresented groups, exacerbating long-standing political inequalities. He has also made contributions to survey methodology, decomposing the turnout bias in the ANES and Pew American Trends panel into its component sources of bias. These studies mark the first decompositions of turnout measurement error in surveys into its constituent parts and the only causally-identified evidence of a mobilization effect for survey-takers.