Article A asks whether the subject studied helps explain party support, and when that relationship develops. Article B asks how far that explanation extends: does it help explain policy beliefs, political activity, and feelings about parties equally well? The authors build on the earlier study rather than arguing against it. Together, the papers suggest that educational experiences can help shape how people see politics without explaining every way they participate in it.
Article A
Field of Education and Political Behavior: Predicting GAL/TAN Voting
Liesbet Hooghe, Gary Marks, and Jonne Kamphorst American Political Science Review · 119(2): 794–811 · DOI 10.1017/S0003055424000583
Knowing what someone studied helps explain party support even when we know their qualification. The relationship begins before study and can develop through education and work.
The question and the explanation
Why would two people with the same qualification vote differently? Research often compares their level of education: secondary school, a bachelor's degree, or a master's degree. Hooghe, Marks, and Kamphorst argue that this misses another difference. A field of education is the subject studied, such as engineering, teaching, or sociology. At college, it is often called a major. A teacher and an engineer can have the same degree level while having different educational experiences.
The authors investigate a European division over immigration, cultural diversity, authority, and cooperation across borders. GAL means green, alternative, and libertarian parties, generally favoring cultural openness, personal freedom, and European integration. Here, libertarian concerns social autonomy, rather than simply opposition to government spending. TAN means traditionalist, authoritarian, and nationalist parties, emphasizing tradition, authority, national identity, and sovereignty. These labels describe parties' positions, not every supporter's character.
The proposed explanation concerns ways of understanding problems developed through repeated study. Cultural skills include interpreting history and literature; communicative skills include teaching and understanding social interaction. Economic skills include accounting and management; technical skills include computing and production methods. Subjects can combine them. The authors expect greater emphasis on social and cultural context to connect with more GAL support and less TAN support.
In our fictional housing example, students might learn to ask how rental rules and unequal negotiating power shape hardship, rather than explain it only through individual choices. These habits can influence which solutions seem reasonable. But students might already think this way before choosing a subject. Later work can strengthen it too. The article therefore investigates a sequence: who chooses the subject, what changes during study, and what continues afterwards.
What the researchers compared
The European Social Survey comparison asks whether people studying different subjects supported different parties, even when degree level, income, gender, and other recorded characteristics are considered. This statistical adjustment tests whether subject adds information beyond those characteristics; it does not make the people identical.
The subject score comes from survey ratings of educational content. For each subject, the researchers compare the combined emphasis on cultural and communicative skills with the total across all four skill groups. A higher score means greater relative emphasis on the first two. It is not a measure of caring or intelligence. Science and mathematics score differently from engineering, so it is not simply a humanities-versus-STEM divide.
The researchers also use panels: surveys returning to the same people over time. Dutch respondents report sympathy toward parties; German respondents report whether they would consider voting Green. A further German analysis estimates their political path without higher-education attendance—a counterfactual—and compares it with the observed path. Comparing this attendance estimate across subject groups differs from making one student change major.
What they found and what it means
A higher cultural-communicative skill score was associated with more GAL and less TAN voting after accounting for degree level and other characteristics. Subject adds information that qualification alone misses. Educational backgrounds typical of an occupation also help explain party support.
Eventual subjects already predicted differences in party sympathy in high school, supporting selection through earlier influences. In Germany the relationship strengthens during study and work; the smaller Dutch analysis does not show clear growth across stages. The German counterfactual analysis supports a shift toward Green support after attendance in both subject groups, larger in the higher-score group. Earlier selection and later educational influence can therefore coexist.
Where the conclusion stops
The study supports examining educational content alongside qualifications. It does not isolate a lecturer, lesson, or peer group: these experiences can change together. The counterfactual estimate requires a credible alternative path without attendance. Group tendencies also cannot guarantee an individual's vote.
The findings do not estimate a change of major or Meridian's proposed department closure. Nor does educational influence establish a deliberate plan to train party supporters. That intention requires evidence about institutional goals and actions.
Closing a program changes the choices available to applicants: some may enter another subject or another institution. To predict the political effect of a closure, we therefore need evidence about those changed pathways, not just existing differences between graduates.
Study details and evidence anchors
What was compared?
The researchers compared party support among people who studied different subjects, while accounting for the qualification they reached and other recorded characteristics. The subject still added information. This challenges the idea that the education divide can be described only as graduates versus nongraduates.
How was the subject measured?
A score summarized how much a subject emphasized cultural and communicative skills relative to all four skill groups. Higher scores were associated with more GAL and less TAN voting. This is a measured difference in educational content, not a judgment that one kind of student is morally better.
What did the timeline reveal?
Eventual subject choice predicted party differences already in high school. In the German panel the differences strengthened during study and work, and the attendance analysis supported an additional shift toward Green support. Earlier selection and later educational influence can therefore coexist.
What comparison is still missing?
The attendance analysis compares an observed educational path with an estimated path without attendance. It does not assign the same student to two majors or isolate a lesson. It supports an argument about educational influence under assumptions, but cannot by itself show intentional partisan training.
Optional research detail
The cross-national comparison covers 15 European countries. The main European Social Survey observations are from 2004, 2006, and 2008, the waves containing educational-subject information. Respondents were at least 21 and reported their vote in the most recent national election. The paper also reports a later Dutch robustness check.
The skill measure adapts the 1998 Family Survey of the Dutch Population, which asked respondents to rate 16 skills on a five-point scale. Cultural and communicative emphasis is divided by emphasis across all four skill domains. Educational and occupational CECT scores are rescaled from zero to one. The occupational score averages the educational scores of people in that occupation.
The high-school-through-completion analysis uses 443 Dutch LISS respondents with 2,271 observations and 4,040 German SOEP respondents with 46,364 observations. These are specific analysis samples, not the entire study. LISS uses party-sympathy ratings; SOEP uses whether the respondent would consider voting for the Greens. The German analysis does not estimate TAN intentions.
The larger German attendance analysis includes people aged up to 30. It uses the IFEct counterfactual estimator, constructing an alternative without attendance from comparison observations, and repeats the analysis for groups above the median subject score and at or below it. Interpretation requires assumptions about the alternative trend; the paper examines pre-attendance patterns and notes small deviations. This is not random assignment to a major.
Read the original sections Education and the Socio-Cultural Divide; Data and Measures; Where Do Differences Between Fields Come From?; and Conclusion.
Read first: Start with the abstract and the explanation of the four skill groups. Then read the panel section to see why earlier differences and later change can both matter.
Open-access full text on the publisher page · CC BY.
Education and Political Attitudes: The Power and Limits of Field Theory
Liesbet Hooghe, Gary Marks, Jonne Kamphorst, and Julia Schulte-Cloos Perspectives on Politics · First View: 1–16 · DOI 10.1017/S1537592726104824
The subject studied is strongly related to which political positions graduates favor. It is less consistently related to how much attention, activity, or partisan passion they devote to politics.
The question and the explanation
If subjects help explain politics, what exactly do they explain? The 2026 article develops a broader argument. Here, field theory means a theory about how the content of a field of education—the subject studied—connects to political judgment. Different subjects repeatedly teach people to notice different causes of social problems. Graduates may carry these habits into politics.
The authors draw on attribution theory: how people explain events and assign responsibility. They distinguish explanations centered on personal effort or choices from explanations centered on social conditions and constraints. In our fictional housing example, does rental hardship mainly reflect spending decisions, unstable work, housing supply, or rental rules? Different answers can make different policies appear sensible. This is our teaching illustration, not an experiment in the article.
The authors expect education emphasizing social and cultural context to connect with more support for redistribution and inclusion. Redistribution means government action to reduce economic differences through taxes, benefits, or public services. Their human-centeredness score measures relative emphasis on cultural and communicative skills; the contrasting material-centered emphasis concerns technical processes and economic incentives. These describe educational content, not a graduate's kindness.
The theory also predicts limits. Learning to explain inequality might change whether someone supports redistribution without making them follow politics or vote more often. Political outlook means the positions and parties someone favors. Political engagement means attention, participation, confidence about acting, and emotional investment. Which party someone supports differs from whether they cast a ballot; favoring a party differs from passionately loving its supporters or disliking its opponents.
What the researchers compared
The researchers link subject skill scores with survey answers about 17 outcomes across US and Western European sources. Original surveys ask graduates to rate the skills their education emphasized. The authors build subject scores from these ratings and match subjects in other surveys to them.
They compare outlook and engagement across subjects, accounting for degree level and other recorded characteristics. They also compare graduates from subjects with higher and lower scores with nongraduates. These are cross-sectional comparisons: different people's answers, rather than each person's change through university.
Two rival explanations receive specific tests. Perhaps the pattern mainly reflects graduates' economic interests or the different proportions of women and men studying each subject. The authors compare models using the skill score, average earnings by subject, and gender composition by subject, separately and together.
What they found and what it means
Higher human-centeredness scores generally accompany more liberal or left-leaning outlooks: support for redistribution, immigration, action on racial inequality, same-sex marriage, environmental protection, and foreign aid. They accompany less support for Christian nationalism and political action outside the law. The precise questions vary, but the direction recurs across measured attitudes.
Subject differences often rival differences associated with educational level, so graduates cannot be treated as one political group. The skill-score relationship persists when average subject earnings and gender composition are considered. Those tests weaken these two accounts as complete explanations without isolating coursework as the cause.
Engagement looks different. The US results do not establish a clear subject-score relationship for seven measures of attention, participation, efficacy, and partisan emotion. Efficacy means feeling able to understand or influence politics. European relationships with party feelings are smaller. This supports a boundary: subjects relate more consistently to what graduates believe than to how actively or passionately they engage.
Where the conclusion stops
The positive conclusion is that educational content makes important political differences among graduates more intelligible. Because the study compares different people, it cannot establish when or through which experience their beliefs changed. Earlier selection, courses, peers, and work still need investigation.
The engagement result does not establish that college attendance cannot increase turnout. It concerns subject differences, primarily among graduates. A statistically unclear result also does not prove exactly zero influence. Together, the articles support a selective account of educational influence in the US and Western European settings studied.
Study details and evidence anchors
What does field theory predict?
Here, field theory is the argument that a subject's repeated forms of explanation can carry into political judgment. The researchers expect subjects emphasizing social and cultural context to connect with more support for redistribution and inclusion. That prediction can be tested; the theory is not simply the observation that graduates differ.
How were the predictions tested?
The researchers linked subject skill scores to survey answers about specific policies, preferred parties, turnout, attention, and partisan feelings. They compared subject groups while accounting for qualification and other recorded characteristics. The outlook pattern remained substantial, giving educational content information that degree level alone misses.
Where did the explanation reach a boundary?
The subject pattern was much clearer for positions such as support for redistribution than for interest, turnout, or partisan emotion. Article A's question about which party people support and article B's engagement question about whether they vote can therefore have different answers.
Could earnings or gender explain it instead?
The authors compared models using the skill score with models using average earnings or the proportion of women and men in each subject. When these accounts were considered together, the skill-score relationship generally persisted. Those tests make an earnings-only or gender-composition-only explanation less convincing; they do not isolate the effect of a course.
Optional research detail
Four sources supply the data: an original US survey in February 2024, N=6,435; an original Dutch college-graduate survey in fall 2024, N=1,000; US General Social Survey data from 2012–22, N=17,272, including 7,324 college educated; and a 2023 YouGov survey in ten Western European democracies, N=16,798. Not every outcome was asked of every respondent. The original surveys use demographic quotas.
Graduates rated how much their education emphasized each of 16 skills on a five-point scale. The researchers summarize relative cultural and communicative emphasis, calculate scores for educational subjects, and rescale the index from zero to one. They match these subject scores to the GSS and European categories, also checking results using the Dutch-derived scores.
The main regressions account for educational level, gender, income, ethnicity, residence, age, religiosity, and relevant country/year differences. Further models test average earnings by subject and sex composition by subject. These aggregate alternatives differ from controlling only for each respondent's income or gender.
The authors report no statistically significant subject-score relationship for seven US engagement outcomes. Smaller European associations appear for party affect and affective polarization. Party affect refers to warmth toward one's party or its supporters; affective polarization concerns warmer feelings toward one's own side than the other. A clear relationship means the data distinguish it from sampling variation under the model; it does not establish its cause.
Read the original sections Empirical Implications of Field Theory; Measurement and Data; Where Field Matters: Political Attitudes, Not Political Engagement; and Income and Sex Composition as Alternative Explanations.
Read first: Start with the theory: why ways of explaining social problems should affect policy judgments. Then compare the attitude results with the engagement results.
Open-access publisher full text · CC BY. A free author-hosted PDF is also linked.
Agustina S. Paglayan · Annual Review of Political Science 29: 525–546 · DOI 10.1146/annurev-polisci-042623-034027
Schools and universities do more than provide useful skills. They also distribute opportunities, teach values, employ teachers, and help states exercise power. Paglayan reviews research on these political roles and asks why governments sometimes expand access while providing weak teaching. The next question is how educational policy affects social order and political beliefs, as well as participation.
Try this question: If a government expands schooling, how would you investigate both who gains opportunities and what political goals the government is pursuing?
Publisher open access · CC BY-NC-ND. The full-text interface loads dynamically.
Educating for Democracy? Going to College Increases Political Participation
Andreas Videbæk Jensen · British Journal of Political Science 55: e1 · DOI 10.1017/S0007123424000486
Does going to college make young people more likely to vote? Jensen compares changes in self-reported turnout between US school graduates who attended college and those who did not, using panels of roughly 10,000 young voters. The main comparison covers the 2004 and 2008 elections; a second panel provides a replication. The study supports an attendance effect, while relying on assumptions about the comparison groups and accurate turnout reports. This differs from comparing the subjects chosen by people who already attend.
Try this question: How could college attendance increase turnout while the subject studied is more closely related to policy beliefs?
Education or Indoctrination? The Violent Origins of Public School Systems in an Era of State-Building
Agustina S. Paglayan · American Political Science Review 116(4): 1242–1257 · DOI 10.1017/S0003055422000247
Why would rulers who fear popular rebellion pay to educate the population? Paglayan examines 40 countries over 1828–2015 and investigates Chile after its 1859 civil war. Schooling expanded after conflict, and historical evidence shows elites turning to education to cultivate obedience. The argument explains why rulers built schools, rather than establishing that every student became obedient. It shows why evidence about intentions and evidence about consequences answer different questions.
Try this question: Which records would tell you why rulers wanted schools, and which different records would tell you what the schools actually changed?
Foundations of a New Democracy: Schooling, Inequality, and Voting in the Early Republic
Tine Paulsen, Kenneth Scheve, and David Stasavage · American Political Science Review 117(2): 518–536 · DOI 10.1017/S000305542200079X
In nineteenth-century New York, some neighboring towns received more public-school funding because of an earlier land-allocation boundary. The authors use that comparison to investigate whether schooling helped sustain participation. Better-funded towns later had higher earnings, lower earnings inequality, and higher turnout. They did not find more support for removing property restrictions on Black men's right to vote. Education can encourage participation by already eligible citizens without extending inclusion.
Try this question: Why should an inquiry into education and democracy measure both participation and who is allowed to participate?
