We demonstrate the utility of a new methodological tool, neural-network word embedding models, for large-scale text analysis, revealing how these models produce richer insights into cultural associations and categories than possible with prior methods. Word embeddings represent semantic relations between words as geometric relationships between vectors in a high-dimensional space, operationalizing a relational model of meaning consistent with contemporary theories of identity and culture. We show that dimensions induced by word differences (e.g., man -woman , rich -poor , black -white , liberal -conservative ) in these vector spaces closely correspond to dimensions of cultural meaning, and the projection of words onto these dimensions reflects widely shared cultural connotations when compared to surveyed responses and labeled historical data. We pilot a method for testing the stability of these associations, then demonstrate applications of word embeddings for macro-cultural investigation with a longitudinal analysis of the coevolution of gender and class associations in the United States over the 20th century and comparative analysis of historic distinctions between markers of gender and class in the U.S. and Britain. We argue that the success of these high-dimensional models motivates a move towards "high-dimensional theorizing" of meanings, identities and cultural processes.

No abstract

Confidence in the scientific community became politically polarized in the United States at the turn of the twenty-first century, with conservatives displaying lower confidence in scientists than liberals. Using data from the General Social Survey from 1984 to 2016, I show that moral and economic conservatives played distinct but complementary roles in producing this divide. I find that moral conservatives exhibited low confidence in scientists before any substantial division existed between self-identified political conservatives and liberals on this issue. However, as moral conservatism increasingly consolidated under the label of political conservatism, a negative association between political conservatism and confidence in the scientific community emerged. Economic conservatives, by contrast, previously held disproportionately high confidence in scientists, but this positive relationship wanes in the beginning of the twenty-first century. These findings suggest that interpreting political polarization requires attention to the multiple dimensions along which political attitudes are organized and ideological coalitions are formed.

We utilize recent data to reassess Baldassarri and Gelman’s (2008) influential characterization of American polarization. Analyzing American National Election Study (ANES) data from 1972 to 2004, Baldassarri and Gelman found that "partisan alignment" – the correlation of party identification with issue attitudes – rose substantially, while "issue alignment" – the correlation between pairs of issues – remained low. These findings suggested that Americans sorted themselves into partisan camps while remaining largely non-ideological. In this paper, we extend Baldassarri and Gelman’s analysis through 2016 with three subsequent ANES waves. First, we find that partisan alignment’s growth has accelerated in recent years. Second, we discover a surprising surge in issue alignment after 2004. While elite subpopulations show the greatest gains, we find that economic issues become more highly intercorrelated across the electorate. These findings motivate a reconsideration of both Baldassarri and Gelman’s "partisans without constraint" thesis and Philip Converse’s dictum that the public is "innocent of ideology."

No abstract