Scholars argue that American politics has experienced a critical realignment over the last three decades. Credentialed “white collar” professionals increasingly side with the Democratic party, while non-college educated voters tend to back Republicans—reversing a trend that predominated for as long as public opinion surveys have existed (Zingher, Goldberg, Reference Goldbergn.d.). The transformation reflects long-running cultural and political trends, and is having far-reaching, ongoing implications (Grossmann and Hopkins, Reference Grossmann and Hopkins2024). Public trust in institutions, and especially in higher education, continues to reach new lows, as Republican office-holders deliver on promises to reshape those institutions in ways they promise will benefit their new political coalition (e.g., Rufo et al., Reference Rufo, Peterson, Barron, Foxx, Hanson, Ferguson, Ali, Klainerman, Haque, Rauh, Cochrane, Marinovic, Abbot, Mitchell, Swain, Thompson, Saad, Jussim, Kaufmann, Haltigan, Priou, Boghossian, Papadopoulos, Domingos, Bonevac, Castro, Warmke, Caplan, Connerly, Kolasinski, Katz, Sommers, Gold, Greene, Yenor, Piereson, Wood, Hazony, Shapiro, Lowry, Kimball, McCarthy, Bauerlein and Rieff2025; Trump, Reference Trump2025).
We report evidence that educational polarization exists in the political institutions that make policy. Our evidence comes from the White House staff, an at-will organization that American presidents assemble to be the nerve center for their political agenda (Cohen and Krause, Reference Cohen and Krause2000; Hess and Pfiffner, Reference Hess and Pfiffner2021; Gibson, Reference Gibson2026). We use these new data on educational pedigree to bring to light several important, stylized facts. Some of these have been observed in other political institutions, while others are clear in cultural spaces in American society.
Specifically, we show that Democratic presidential staffers are, on average, significantly more credentialed than their Republican counterparts. Republican presidents tended to hire staffers with more diverse educational backgrounds—people who earned degrees from public universities, those without R1 status, and those below the top 50% of ranked institutions. The baseline credential divide mirrors the American electorate, whereas the differences in “elite” educational pedigree are consistent with those found in Congress (Volden et al., Reference Volden, Wai and Wiseman2024). While about 27% of Democratic presidential staff in our data were educated at top 34 institutions, only 11% of Republicans were.
Moreover, we show college-educated staffers polarize by degree field, with Democrats more likely to specialize in sociology, anthropology, public policy, and gender and minority studies, whereas Republicans were more likely to study business, economics, and criminal justice. These differences are consistent with evidence that educational field indexes cultural differences across right and left parties, both within the United States, and outside it (e.g., Chin et al., Reference Chin, Ortega, Shaw and Yoo2025; Hooghe et al., Reference Hooghe, Marks and Kamphorst2025).
Like Volden et al. (Reference Volden, Wai and Wiseman2024), we argue these differences in educational pedigree are worthy of study, in part, because they may have implications for what political institutions do. Elite-educated members of Congress behave in distinct ways once in office. For the White House staff, we show educational divides are associated with a policy output: specifically, the targeting of universities by the second Trump administration. The Democratic lean of a university’s White House alumni is positively associated with investigations and enforcement actions by the Trump administration, along with research grant and contract terminations by the National Science Foundation (NSF) and National Institutes of Health (NIH) in 2025. For example, a standard deviation increase in Democratic alumni lean corresponds to a loss of about 15% more research funding. These associations remain after accounting for a college or university’s size, prestige, wealth, type, county-level support for President Trump, and pro-Palestinian protest activity.
These associations are particularly noteworthy in the context of fundamental changes to the way policy is typically made in the United States. Increasingly, change occurs via executive action, which means that White House staff are often the political professionals who write and monitor the implementation of policy on behalf of the American president. We show these staffers have increasingly divergent educational communities.
There are several processes that might contribute to the association between staffing and organizational behavior. Colleges, universities, and degree programs have ideological reputations. Knowing these reputations, future party staffers may have selected into different colleges and degrees based on political preferences. Another possible mechanism is that those in charge of staffing presidential administrations might use educational pedigree as a heuristic for both “loyalty” and “competence.” In this way, the reputations of colleges and degree programs might make them useful for screening potential hires with preferences that diverge from the president. Moreover, presidents have policy priorities that might require particular expertise, and so hiring managers would want to bring in those with appropriate credentials and training.
But ties to an alma mater might also lead staff to exempt the institution from punishment, or lobby for its benefit. There is increasing evidence that network ties are influential in government decision-making (Hübert et al., Reference Hübert, Rezaee and Colner2023; Lee and You, Reference Lee and You2023; Garlick et al., Reference Garlick, Junk and Brown2025). The American presidency grants unequal access to special interests (Miller, Reference Miller2023). In an era where much of policy is driven by presidential initiative, the backgrounds and associations of the people who staff the presidency warrant special scrutiny.
To be clear, our evidence cannot speak to which of these causal pathways is most prominent. The facts we reveal make the case that each of these processes is worth studying. More generally, they reveal the deep, political roots of contemporary changes in higher education. No presidential administration has launched more investigations, enforcement actions, or been as aggressive with funding cuts toward colleges and universities. These actions are sometimes characterized as an indiscriminate attack on higher education. Our evidence suggests, however, that these targets are not random, but instead reflect a political divide in higher education itself.
1. Evidence
This study relies on recently collected personnel data on the White House office. Why this entity, of all others in the vast federal bureaucracy? The White House provides an important empirical context to complement existing work on the diploma divide: it is powerful, exclusive, and embodies the political coalition of the party in power. American presidents have almost complete discretion over who to hire and fire.Footnote 1 This office, unlike any other, is a direct reflection of what kinds of expertise and pedigree their governing coalition values (Lowande, Reference Lowande2024). The American presidency has overpowered other branches of government, and the White House, more than any other office, drives policy change in government. In addition, the White House is a transient, high prestige workplace, where professionals go on to work in Congress, run executive departments, advise Fortune 500 companies, and even serve on the Supreme Court.
Our White House staff data from 1995 to 2024 come from (Lowande and Birdsong, Reference Lowande and Birdsong2024). Prior to 1995, there is no official accounting of who worked in the White House. The records are assembled from annual, July snapshots of presidential staff maintained by presidential and internet archives. In total, the data contain information about 5,955 unique persons who worked in the White House or 12,874 staffer-years. These records were supplemented with hand-collected records of staffers’ terminal degrees and educational institutions at the time of hire, collected from social media pages, archival biographies, and more. Details regarding the coding procedures, inter-coder reliability, and more can be found in the referenced work.
Like all data, these records are imperfect. Most important is the fact that data on educational attainment are unknown for 23.7% of staffers. Missingness was found to be positively correlated with age of record and working for a Republican president, and negatively correlated with salary. Fortunately, each of these patterns biases against the findings we report in this study. The staffers most difficult to find tend to be in low-ranking secretarial positions for President Bush. These staffers are unlikely to possess advanced degrees on par with their senior colleagues, so it follows that we likely under-estimate the proportion of Republicans who held a B.A. or less. Also reassuring is the fact that our findings are even more stark when we only compare the Obama and Trump administrations, for which we have more complete records.
Another key feature of these data is that they only represent a snapshot of educational attainment coterminous with being hired by the President. That means that in what follows, we do not examine education or training staff may have pursued after the White House.Footnote 2 It also means that the results that follow conflate degree levels and do not take into account degrees the staffer may have earned before their last degree. Thus, our findings should be thought of as representing the highest level of education staffers had received when the White House personnel office considered hiring them.Footnote 3
In the sections that follow, we pair these records with data on institutions of higher education from the 2023 Integrated Postsecondary Education Data System (IPEDS),Footnote 4 crowd-sourced records of pro-Palestinian protest activity (Pressman et al., Reference Pressman and Chenoweth2024), and existing federal research funding (National Center for Science and Engineering Statistics and National Science Foundation, 2025). We also assemble records of contemporary actions related to higher education institutions, which come from two sources: (1) the Chronicle of Higher Education’s “Tracking Trump’s Higher-Ed Agenda” page,Footnote 5 and (2) the Grant Watch Project.Footnote 6
2. The White House diploma divide
To examine the elite implications of the diploma divide, we begin with educational attainment in the White House staff. Not surprisingly, almost all White House staffers have some college degree, with only 2.3% holding no degree. Since roughly 38% of U.S. adults have a B.A. or equivalent, this would seem to rule out the possibility that the White House would reproduce the diploma divide evident in the broader electorate. Figure 1 suggests, instead, that the diploma divide merely tracks credential inflation.
Democrats tend to hire staff with more credentials. Plots the proportion of White House staffer-years from 1995 to 2024, within party, with staffers at each level of educational attainment. Less than a bachelors includes associate’s, high school, high-school drop outs, and professional certificate programs. The right panel subsets to staffers serving between 2017 and 2024.
Specifically, Democratic staffers were significantly more likely to have earned any advanced degree (an M.A., J.D., or Ph.D.) at their time of hire, whereas Republicans were more likely to come in with less than a Bachelors. The magnitude of this divide is instructive. About 41% of Democratic staffers were hired with advanced degrees, relative to 30.5% of Republicans.
One relevant question is whether these differences reflect differences in the kind of jobs Democrats and Republicans create. For example, the high rate of law degrees among Democrats might reflect their desire for more legal counsel within the White House. We find this is only partly the case. Democrats are slightly more likely to employ legal counsel (8.2% of staff, relative to 6.7%; 
$p \lt 0.001$), but this difference is not sufficient to explain the 10 percentage point gap in the broader pool. One illustrative example is to compare Chief of Staff for President Biden with that of the second Trump administration. Biden hired Ron Klain, who held a law degree and was educated at Georgetown and Harvard, while Trump hired Susie Wiles, who holds a B.A. from the University of Maryland, College Park. In short, Democratic presidents tended to hire people with law degrees whose job title suggests they worked in the White House doing something other than practicing law. Finally, the gap in educational attainment is starkest for the Biden and first Trump presidencies, which suggests it is widening over time.
Another question is whether the education gap is driven by differences in the demographic composition of White House staffs across parties. For example, if Republican presidents tended to hire more White staffers, and non-White staffers tended to have more credentials, the pattern above could be related to differences in political coalitions across parties. As it turns out, only the former is true. Republican presidents did tend to recruit fewer people of color (11.2% vs. 29.8%, 
$p \lt 0.001$). But, among White House staff, there are no significant differences in educational attainment across race and ethnicity. Democratic presidents hired more women (55.4% vs. 50.6%, 
$p \lt 0.001$), and women tended to have fewer degrees, but this actually works against the educational attainment gap. After accounting for salary, year of record, sex, and race and ethnicity, party remains a strong predictor of educational attainment among White House staff (see Table A1).Footnote 7
Other studies find that, even among credentials, partisan support varies systematically across fields for voters (Hooghe et al., Reference Hooghe, Marks and Kamphorst2025) and faculty (Chin et al., Reference Chin, Ortega, Shaw and Yoo2025). To investigate this in our data, we classified the hundreds of unique degree field names into one of 36 areas of study and used them to predict whether a staffer worked for a Democratic president. Staffer-year is the unit of analysis, and year is included as a covariate to account for trends in the prevalence of particular majors. Figure A1 of the Online Appendix reports coefficients and 95% confidence intervals constructed from standard errors clustered at the staffer level. Accordingly, coefficients less than zero indicate the major field is more closely associated with Republicans, and greater than zero indicates a Democratic tilt.
Democratic staffers were more likely to earn degrees in gender studies, sociology, area studies related to American cultural minorities (e.g., African American, Latin American, and Native American studies), and the arts (e.g., theater, dance, fine arts, music composition, and design). Republicans, on the other hand, were more likely to earn degrees in business and economics, criminology, and criminal justice. The White House diploma divide is not just about attainment, it is also evident in the skills and expertise these political professionals bring to their jobs.
Alumni in the White House. We next examine where White House staffers were trained.Footnote 8 We first subtract the number of Republican staffer alumni-years from the total count of Democratic staffer alumni-years, so that positive values indicate a Democratic tilt, and negative indicate Republican tilt.Footnote 9 Zero thus indicates an even split (or no alumni in the White House).
We then merge this measure with 2023 IPEDS data on endowments, R1 status, public (or private), and student enrollment. In addition, we normalize an institution’s U.S. News rank from 1988 to 2024 and account for proportion of county-level 2024 presidential election votes President Trump won. Naturally, this analysis excludes staffers with no college experience, and those whose degrees came from specialized military schools not included in IPEDS.Footnote 10
We report the results of this linear model in Table A2 and provide a visualization of the basic findings in Figure 2. Both illustrate key points about educational polarization among political elites. First, at baseline, local support for the Democratic party is associated with institutional prestige and Democratic alumni. As Figure 2 illustrates, there is no college or university that has both consistently been ranked in the top 25% and also sits in a county President Trump won in 2024. Notably, nearly without exception, these institutions have more Democratic alumni staff than Republican.Footnote 11 On the other end of the spectrum, there are institutions like Baylor University, Hillsdale College, and Texas A&M, which all sit in conservative counties, and tend to train more Republican staffers. The results also suggest that Democratic tilt is positively associated with R1 status.
Higher-ranked institutions in liberal counties tend to have more staffers go on to work for Democratic presidents. Plots county-level vote share for President Trump in 2024, normalized U.S. News and World Report rank since rankings began, and the relative proportion of Democratic versus Republican staffers by college. Omits universities never ranked in either the “national universities” or “liberal arts” category.
Figure 2 Long description
The vertical axis label reads, County level Trump 24 Vote. Tick labels shown are 0.25, 0.50 and 0.75. The horizontal axis label reads, U. S. News Percentage Rank (1988 to 2024). Tick labels shown are 0.00, 0.25, 0.50, 0.75 and 1.00. A legend is labeled, Alumni Lean. Tick labels shown are 1.0, 0.5, 0.0, negative 0.5 and negative 1.0. The plot contains many circular markers. A dense vertical band of markers is located at the horizontal value 1.00 and spans much of the vertical range. Additional markers are distributed across the horizontal range from 0.00 to 1.00, with many markers between 0.50 and 1.00. Markers appear across the vertical range, including points below 0.25, around 0.50 and above 0.75. The marker distribution includes a concentration at the horizontal value 1.00 compared with the rest of the horizontal range.
White House alumni and targeting. We next consider whether educational divides in the White House map onto government behavior. Following Volden et al. (Reference Volden, Wai and Wiseman2024), we argue these differences in backgrounds are worth studying because they have implications for how political organizations operate. In 2025, the second Trump administration began the most dramatic government intervention in higher education in recent memory, targeting individual colleges and universities with investigations, lawsuits, and funding cuts. These actions were driven by the White House, with targets picked by political appointees at the behest of activists urging what Deneen (Reference Deneen2023) calls “Machiavellian means” for higher education reform. Was the historic White House diploma divide associated with any of these actions?
To examine this, we collect publicly available information on letters, visits, investigations, and enforcement actions taken against colleges and universities since the start of the Trump administration. In total, this represents 227 distinct actions taken by executive departments, as tracked by Chronicle of Higher Education.Footnote 12 We then predict whether an institution was singled out for action, using normalized Democratic alumni lean, while controlling for the institutional characteristics we have already examined. We also include a control for pro-Palestinian protests between October 7, 2023 and December 2024 (Pressman et al., Reference Pressman and Chenoweth2024).
Table A3 reports the results, while Figure 3 depicts the likelihood of being targeted, based on whether alumni were all Democrats, all Republicans, or split.Footnote 13 Other political factors, like local support for Trump, and pro-Palestinian protests, have intuitive associations with investigations and enforcement actions. But the association between alumni lean and Trump administration actions is obvious, even after accounting for what we have already shown predicts staff lean. Notably, universities with Republican staff lean have not been exempt from scrutiny, but a strong Democratic staff lean makes action more likely.
Democratic lean in alumni who go on to work in the White House (1995–2024) was associated with investigations and enforcement actions by the Trump administration in 2025. Plots the predicted probability of a (negative) action taken against a college or university, based on its alumni lean. Full results reported in column 2 of Table A3 of the Online Appendix.
These data results are based on what is publicly known, which leads to a potential limitation. Conservative institutions might not announce, and the Trump administration might not publicize, policy decisions they make that negatively impact conservatives. Likewise, both belligerents have incentives to announce when a left-leaning institution has been subject to the administration’s agenda. Therefore, we supplement the previous analysis with grant and contract terminations by NSF and the NIH, based on what was known by the Grant Watch Project as of July 2025.Footnote 14 The advantage of these data is that they come from both researcher self-reports and government data. It is easier to imagine a university counsel keeping a cease-and-desist letter secret, than it is to imagine the sworn silence of all staff and co-PIs on a terminated multi-million dollar research study.
We then repeat the analysis above, using Democratic alumni lean as a predictor, accounting for observables. To ensure that these results are not driven by pre-existing reliance on federal research funding, we include a measure of federal research funding as of 2024 (National Center for Science and Engineering Statistics and National Science Foundation, 2025). The results are reported in Table 1.
Democratic alumni lean is associated with grant and contract terminations, after accounting for other factors
Table 1 Long description
Regression results for 5,769 colleges and universities relate the logged dollar amount of terminated grants and contracts to alumni partisan lean and multiple institutional controls, reported separately for all terminations, NSF terminations, and NIH terminations. Alumni Democratic lean is positive and statistically strong in all three models, with the largest association for all cuts and the smallest for NIH cuts. Trump vote share is negative and statistically strong across all three outcomes. R1 research status, being ranked, public status, enrollment, prior federal research funding, and protest count are all positive and statistically strong predictors in each model. Ranking tier effects vary: the top ten percent tier is strongly negative in all three models, while the top five percent tier is only clearly negative for NIH cuts; other tiers are mixed, with several positive effects for all cuts and NSF cuts and weaker or non-significant effects for NIH cuts. Total staff is not clearly related to all cuts but is positive and statistically strong for NSF and NIH cuts. Model fit is moderate, with explained variance around the mid-forties for all cuts and NSF cuts and about one-half for NIH cuts; results are associations and should not be read as causal.
Notes: Unit of analysis is a college or university. Dependent variable is the log-transformed dollar amount of grants and contracts terminated as of July 14, 2025. Key independent variable is the number of Democratic staffer-years minus the number of Republican staffer-years, by institution, so that negative values indicate Republican-lean, positive values indicate Democratic lean. Values normalized between 
$-$1 and 1. Rankings standardized based on all available U.S. News and World Report Rankings since 1988, with effects relative to never-ranked institutions. *
$p \lt $0.1; **
$p \lt $0.05; ***
$p \lt $0.01.
Democratic staff lean is associated with research funding cuts for grants and contracts from both agencies. The estimates in column (1) suggest that a standard deviation increase in Democratic lean is associated with about 15% more funds lost. Notably, this political factor remains significant, despite accounting for local support for President Trump, existing research funding, and pro-Palestinian protests, which were often cited by the Trump administration as the impetus for various enforcement actions.
3. Discussion
There are several processes that might lead to the associations we find between White House staffs’ make-up and their policy choices. Schools’ reputations surely impact the students that matriculate, which might lead to political sorting on the basis of student body. Moreover, presidents staff their administrations with personnel who are politically proximate and competent, and educational pedigree might serve as a useful signal of both. Finally, once in the White House, staffers’ alumni networks might lead them to privilege some institutions over others. These causal pathways are not mutually exclusive, and we cannot rule out any particular pathway with our data. Instead, the stylized facts we report help motivate the investigation of those processes and speak directly to existing work on educational realignment in American politics.
First, our findings are broadly consistent with arguments that the educational realignment began as an “elite” phenomenon. Schickler and Rodriguez (Reference Schickler and Rodriguez2026) find that party elites altered political messaging related to higher education prior to major shifts in voters, while Volden et al. (Reference Volden, Wai and Wiseman2024) show that the elite education gap emerged in both chambers of Congress by the mid-1990s. This timing lines up with what we report in this article. Put differently, though the education gap in the White House is most extreme when comparing Biden and Trump, we have also shown it existed prior to both presidencies.
Our findings also align with evidence that the educational divide is fueled by non-economic concerns, namely those related to identity-related political issues. Degree fields of Democrats and Republicans in the White House differ substantially, with Democrats consistently more likely to have interests in gender, race and ethnicity, sociology, and anthropology. Schickler and Rodriguez (Reference Schickler and Rodriguez2026) show these issues intersected and animated Republican party platforms that criticize higher education generally. Moreover, there is evidence that reactions to these identity issues have driven the rightward shift of non-college educated voters, both in the United States (Marble, Reference Marble2024) and abroad (Gethin et al., Reference Gethin, Martínez-Toledano and Piketty2022).
Finally, our findings speak to the emerging importance of studying the backgrounds of executive branch personnel. Today, as the separation of powers imposes few constraints on executive branch activities, individuals who serve at the pleasure of the president hold more concentrated power. Policy selection and enforcement, which used to include layers of veto-points, in some cases reduces to the preferences of bureaucrats. In this environment, understanding the network connections, preferences, and expertise of presidential agents may become more important for understanding government action.
Acknowledgements
For research assistance, we thank Julian Barnard, Noah Fasczewski, and Madelyn Yake. A previous version of this project was presented at the 2024 annual meeting of the Midwest Political Science Association, Chicago, IL. Claude Sonnet 5.5 was used to manage citations, reformat tables, and identify grammatical and typographical errors, which were then fixed manually by the authors. We are responsible for all errors.
Data availability statement
Data and replication code needed to reproduce the results in this study can be found at https://doi.org/10.7910/DVN/HKETWU.
Declaration of interests
The authors have no competing interests to declare.
Supplementary material
The supplementary material for this article can be found at https://doi.org/10.1017/psrm.2026.10146.
Open access
