Peer-Reviewed Work
The Rise of Affective Polarization: Is It Who We Are, or What We Think?
Conditionally accepted, American Journal of Political Science
Scholars express alarm at rising affective polarization. Many blame rising demographic sorting, arguing voters’ social characteristics (“who we are”) have increasingly aligned with their party. Others point to partisan divides around ideology, policy preferences, and group sentiments (“what we think”). Unfortunately, past research lacks comparable over-time metrics of demographic and attitudinal sorting—and often conflates the two. Using measures that are comparable across time and domains, I find that demographic sorting has not meaningfully increased, nor has its relationship with polarization strengthened. Even with generous causal assumptions, demographic sorting can only explain about 2.2% of the rise in affective polarization. By contrast, attitudinal sorting has increased substantially and can account for a substantial share of polarization’s rise, with panel and meta-analyses suggesting that attitudes fuel polarization while polarization, in turn, reshapes attitudes. Thus, the rise of affective polarization is bound up in what we think, not who we are.
Work in Progress
The Political Evolution of Business Elites: Longitudinal Evidence from the Technology Industry
Manuscript under review With Olivier Bergeron-Boutin, Christian Caballero, Pia Deshpande, Greg Ferenstein, Amir Goldberg, and Neil Malhotra
Wealthy industries—and the individuals that control them—exert disproportionate influence over policymaking. However, systematic, industry-specific data on how economic elites change their political views over time remains scarce. This paper investigates the political preferences of U.S. technology entrepreneurs, a powerful interest group that many claim shifted right, contributed to President Trump’s 2024 victory, and held sway over the Administration’s policymaking. Theories of political behavior—self-interest, partisan sorting, and the stability of core beliefs—offer competing expectations regarding how tech elites’ views may have changed. To investigate this question, we conduct a novel survey of venture-backed founders, combining a fresh cross-section with reinterviews of entrepreneurs originally contacted in 2017. Our data suggests a more complex answer than a uniform rightward shift: Technology elites have moved modestly to the right on variables such as party identification and support for redistribution. However, they have moved leftward on regulation, and in other domains, their preferences have remained stable. These findings suggest that the political alignment of economic elites cannot be inferred from material interests alone; it is instead produced through the interaction of material interests, social values, partisan identities, and industry-specific political cultures.
The Unmoved Mover? A Reassessment of Party ID’s Long-Term Durability
Scholars have traditionally viewed partisanship as an identity that crystallizes in young adulthood and thereafter serves as the “unmoved mover” of American politics. However, data limitations have prevented a thorough study of how party ID changes over a lifetime—available panels have either been too short or included too few waves. Using a massive public health panel dataset long overlooked by political science scholarship, I track how voters’ party ID changes over the course of their lives with an unprecedented scope and granularity. I find the traditionalist view of partisanship is overstated: Although party ID becomes far more stable after young adulthood, this primarily reflects a decline in short-term fluctuations and does not mark the end of meaningful change. After age 30, party ID still durably moves at roughly half the rate of young adulthood, and this finding holds across multiple approaches to correcting for measurement error. I conclude by demonstrating that voters change their partisanship due to programmatic considerations such as ideology and spending preferences, even as those views are also shaped by party.
Ideological Resentment: How Substantive Grievances Fuel Negative Political Identities
Many scholars and observers explain right-wing polarization as a consequence of nativism and racial animus. Others point to partisan teamsmanship resulting from natural tendencies toward tribalism. However, conservatives often frame their grievances in primarily ideological terms. In light of both racial depolarization and surging conservative hostility toward cultural institutions such as academia, this rhetoric warrants investigation as a motivator of conservatives that is distinct from demographic resentments—and distinct from partisan social identity. To this end, I develop a theory of ideological resentment, defined as the belief that one’s political opponents have dominant cultural power and a sense of marginalization stemming from that belief. I validate robust measures of this construct and demonstrate its strong relationship with partisan animus.
Have Americans Polarized, or Just Sorted? Using Prediction Models to Revisit an Old Debate
For twenty years, political scientists have debated whether American voters ideologically “polarized”, or just “sorted”? In other words, have voters become more ideologically extreme over time, or have conservatives and liberals just moved into the Republican and Democratic parties without developing stronger views? This debate has reached no clear resolution after 20 years. But in this paper, I use predictive models to 1) benchmark the degree of over-time change that has actually occurred and 2) assess how much extremity or constraint has increased even when the distribution of attitudes within parties is held constant over time. In doing so, I find that almost all of the increase in partisan divisions through 2000 can be attributed to sorting. However, by 2024 over 20% of rising partisan divisions reflect polarization, not just sorting. Thus, there appears to have been an increase in genuine polarization, but it has only emerged in the past two decades.