Image

Donald Snyder, a doctoral student in political science and graduate fellow with the UMass Amherst Poll, was recently announced as the winner of the New England Association of Public Opinion Research (NEAAPOR) Student Paper Competition. Winning the regional competition earns Snyder the opportunity to present the paper at the 81^st Annual American Association of Public Opinion Research (AAPOR) Conference, to be held May 13-15 in Los Angeles, with up to $750 of his travel expenses paid by NEAAPOR.

Snyder’s paper “There’s a Lot to Unpack There: Personalization in Political Language and Affective Polarization” introduces an “Ideological Personalization” scale – a measurement of political thinking using previously defined orientations, placing “vague, symbolic, abstract politics” on one end of the model opposite from “grounded, concrete, personal political considerations.”

The paper, currently under review, is based on his project that used large language models (LLMs), validated against human raters, to capture levels of personalization in a dataset of over 60,000 verbatim responses collected from 1984-2020 by the American National Election Studies collaboration, in which respondents described their “likes” and “dislikes” about the Democratic and Republican parties.

“Ideological Personalization acts as a complement to the political sophistication literature by leveraging open-ended text as a rich data source for detecting when political reasoning is framed in terms of abstract principles, party-cued content, or concrete personal stakes,” Snyder writes in the abstract of his paper. “Ideological Personalization is not a direct operationalization of factual knowledge or ideological constraint, but rather, a crucial supplement to this puzzle, measuring an individual’s personal attachment to the political landscape. Using text methods, the construct can be measured at scale within tens of thousands of responses, capturing political frameworks that might otherwise be lost to time using standard survey batteries.

“Across time, Ideological Personalization does more than merely predict polarization: it illuminates the changing psychological distance between citizens and the political world that they inhabit,” he concludes. “Ultimately, these findings affirm that political behavior is not only a matter of what citizens believe, but also their personal relationship with politics itself; sometimes understood as abstract, distant philosophies, and sometimes understood as a deeply integral part of their everyday lives.”