Political science: A discipline of troublemakers or loyal agents of the state?

The origins of political science can be described in different ways, depending on the criteria that one applies to define the discipline. One widespread notion is that the ancient Greeks – especially Plato with his first typology of political systems in The Republic or Aristotle’s categorisation and empirical mapping of systems in Politics – are the founding fathers of the study of systems of government and thus political science (Almond Reference Almond, Goodin and Klingemann1998). Scholars of the Enlightenment such as Hobbes, Locke, and Montesquieu; US founding fathers such as Madison or Hamilton; and philosophers of the 18th and 19th century such as Hegel, Marx, and Comte are also commonly viewed as early contributors of relevance. However, it was the second half of the 19th century and the beginning of the 20th century that saw the emergence of political science as a separate discipline, rather than an area of study within law or philosophy. During this time, the study of politics became established at, for example, Oxford and Cambridge, albeit under the label of contemporary history, and the first departments of political science came into existence towards the end of the 19th century (Almond Reference Almond, Goodin and Klingemann1998).

In the USA, several free-standing political science departments had been established by the first decade of the 20th century, and the American Political Science Association (APSA) was formed in 1903. In most Western European countries, political science developed into a self-standing discipline around the same time, albeit with specific national paths of development (for an overview, see Ilonszki and Roux Reference Ilonszki and Roux2019). In general, though, the development of political science in North America and Western Europe was tightly linked, as individuals and ideas would regularly cross borders as well as the Atlantic. This mutual influence culminated in the creation of the European Consortium for Political Research (ECPR) in 1970 with the help of funds from the Ford Foundation (Almond Reference Almond, Goodin and Klingemann1998). Overall, one can say that political science in North America and Western Europe was, at least since 1945, institutionalised as a discipline, and this discipline was characterised by rather stable and steady growth as well as strong transatlantic links.

While the institutionalisation of political science happened first in North America and Western Europe, the discipline became more established in other countries not long after. In China, which we will elaborate on in some detail to illustrate some more general points on the role of politics in non-democratic regime contexts, the first course on politics was taught at the Capital Academy (later Peking University) as early as 1903 (Fu Reference Fu, Easton, Gunnell and Graziano1990). The Chinese Social and Political Science Association was founded in 1915, and a separate Chinese Political Science Association was established in 1932 (Fu Reference Fu, Easton, Gunnell and Graziano1990). In parallel, universities started creating separate political science departments so that, by 1949, around 40 departments existed at Chinese universities.

Following the establishment of the People’s Republic of China, the situation for political science in China changed. With the focus of the Chinese Communist party (CCP) on controlling social knowledge, and following the example of the Soviet Union, which outlawed political science as early as in the 1920s, China abolished political science as an institutionalised discipline in 1952 (Fu Reference Fu, Easton, Gunnell and Graziano1990; Wang and Guo Reference Wang and Guo2019). However, this dynamic was not entirely new: Even during imperial times, whenever the state enjoyed a stronger hegemony over civil society, it limited the space for social science knowledge (Fu Reference Fu, Easton, Gunnell and Graziano1990). During these phases, the state would prioritise the development of technical knowledge, which was seen as a necessity for societal progress, while the development of societal knowledge remained under strict control. A similar dynamic played out during Mao’s Cultural Revolution, when most social scientists were sent to the countryside for labour, and social science research came to a standstill.

After the rise of Deng Xiaoping, science received renewed recognition in China, but the natural sciences received more autonomy than the social sciences (Fu Reference Fu, Easton, Gunnell and Graziano1990; Perry Reference Perry2020; Wang and Guo Reference Wang and Guo2019). Only after some time, and as the government realised that they could be useful in solving societal challenges, the social sciences received more space. Nonetheless, political science research and teaching remained closely monitored by the state. By 1980, the Chinese Political Science Association was (re-)established, coinciding with an opening for academic studies of policies and political institutions, foreign and domestic (Fu Reference Fu, Easton, Gunnell and Graziano1990; Wang and Guo Reference Wang and Guo2019). The realisation that China required various reforms to ensure effective state functioning and decentralised decision-making created a demand for more and better trained civil servants. This, in turn, led to a growing interest in administrative research within political science among CCP officials. In 1983/84, Fudan University and Peking University started enrolling students in political science programmes, and by 1986, more than 10 Chinese universities had set up political science programmes or departments (Fu Reference Fu, Easton, Gunnell and Graziano1990). Yet, following the Tiananmen massacre in 1989, the state ordered many universities to stop enrolment in social science programmes, and many students had to undergo ‘ideological training’ courses before they were allowed to start or continue their studies (Fu Reference Fu, Easton, Gunnell and Graziano1990).

Contemporary work on political science research in China continues to describe similar tensions. For example, research on local politics is considered less sensitive, and is thus more prevalent, than research on the central government (Reny Reference Reny2016). Similarly, there is a continuous tension between ‘westernisation’ and ‘indigenisation’ of political science in China, with debates centring on how applicable ‘western’ political science concepts and ideas are to the Chinese context, or whether a specific Chinese understanding of the discipline is more appropriate (Wang and Guo Reference Wang and Guo2019). These debates relate to the notion, suggested by Perry (Reference Perry2020), that Chinese higher education in general is characterised by ‘educated acquiescence’. This entails that higher education contributes to autocratic persistence, following the successful co-optation of academia, its structures, processes, and graduates through the state. Instead of looking at higher education as a threat, the autocratic regime can actively promote its interests through promoting higher education but then controlling, for example, the content of teaching, career paths of professors, or entry examinations. Similar arguments have been made regarding non-democratic states in general, highlighting that they may strategically use parts of the education system, including primary education, to promote core political interests (for example, Paglayan Reference Paglayan2024).

The tumultuous history of political science research and teaching in 20th-century China is a good – albeit only one – example of the complicated relationship between political science and the state, and how this relationship depends on the shifting political conditions and the perceived benefits and risks to incumbent regime elites. Similar developments can also be observed in Africa, where, following the end of colonial rule, the importance of nation building and training of a capable civil service gave a strong impetus to give political science more attention and space in several countries (Jinadu Reference Jinadu, Easton, Gunnell and Graziano1990). However, the discipline was also perceived as subversive by leading elites, and thus in need of tighter control and limited growth. As a reaction, many countries tended to create separate programmes or departments for public administration or international relations, which were perceived as less troublesome subdisciplines (Jinadu Reference Jinadu, Easton, Gunnell and Graziano1990).

In the last two decades, the discipline of political science has undergone a general debate about its societal and political relevance (see, for example, Capano and Verzichelli Reference Capano and Verzichelli2023; Flinders and Pal Reference Flinders and Pal2020). While there is a perception in parts of the discipline, and even among some outside observers, that political scientists could do a better job of being relevant when it comes to political or societal debates (see, for example, Ostrom Reference Ostrom2000; Putnam Reference Putnam2003), there is no hard proof that political science actually has become or is becoming less relevant (Flinders Reference Flinders2013). At the same time, political science, similar to other disciplines, has been exposed to an international impact agenda (Bandola-Gill, Flinders, and Brans Reference Bandola-Gill, Flinders, Brans, Eisfeld and Flinders2021) through the introduction of incentives to deliver demonstrable evidence of non-academic impact. This has increased the incentives for political scientists to focus on topics that matter for politics, and thus follow the notion of Gerring and Yesnowitz (Reference Gerring and Yesnowitz2006: 112) that ‘[s]ocial science is science for society’s sake’.

To be more specific, data from a large survey of European political scientists, for example, show that around 80% of respondents were engaged in some form of policy advisory activity in the past three years (Brans and Timmermans Reference Brans and Timmermans2022), and studies on the USA highlight the transfer of academics into government and think tanks as a key transmission belt for political science knowledge into government (Flinders Reference Flinders2013). So, it seems that, at the individual level, the interest of political scientists when it comes to providing knowledge that is of relevance for politics is very much present (see also Capano and Verzichelli Reference Capano and Verzichelli2023). Therefore, it is reasonable to assume that political science as a discipline produces knowledge that is intended to ‘improve politics’. Politicians, in turn, should presumably be interested in engaging with this knowledge. At the same time, political science has a moral foundation that is clearly pro-democracy (Capano and Verzichelli Reference Capano and Verzichelli2023; Flinders and Pal Reference Flinders and Pal2020). This can give rise to tensions in contexts where these values are not embedded in the current regime or shared by key political decision makers. Thus, the interest in the expertise offered by political scientists should depend on the political environment, and the framing of the issue at hand that might make it more or less compatible with existing values (Brint Reference Brint1990).

Overall, the historical development of political science reflects two general characteristics of the discipline. On the one hand, political scientists are often described as handmaidens to power and the state (Goodin Reference Goodin2011; Mansbridge Reference Mansbridge2014). They train bureaucrats and diplomats, act as counsellors to rulers or parties, and investigate how power is most effectively employed through a well-functioning state. Especially in situations where the state can co-opt higher education into educated acquiescence (Perry Reference Perry2020), political leaders can ensure that higher education serves their needs without encountering any regime-related risks. On the other hand, higher education, in general, and political science education, more specifically, can promote more critical stances if it is not perfectly controlled. For political science teaching, these critical stances could pertain to lectures and readings that focus on, for example, normatively good and just social arrangements (which may deviate from what students can observe in their country), why executive power should be constrained, or why marginalised voices should be heard (Capano Reference Capano2025; Goodin Reference Goodin2011; Mansbridge Reference Mansbridge2014). Hence, politicians might sometimes have strong reasons to be sceptical of political science being taught to young citizens in their country. At the same time, tolerating this type of critical and ‘dangerous’ academic teaching and research in nondemocratic contexts, might be used by political leaders as a signal of their openness and tolerance, potentially enhancing their legitimacy domestically and internationally.

The two, at times competing, roots of the discipline can create opposing incentives regarding government support for the institutionalisation of political science, and the relative weight of these incentives may depend on the political-regime context. Thus, when trying to hypothesise about the cross-country distribution of political science departments, existing work on the discipline points us in somewhat contradictory directions. The argument that political science helps run a more effective state suggests that any country, no matter the political regime, should have an interest in having political science departments. In contrast, the argument highlighting political science’s democratic and normative core would suggest that autocratic regimes will be more careful in establishing such departments, given that the content that is taught or researched in them could potentially undermine the regime. In the following, we will expand on this tension and develop specific hypotheses regarding how the existence and other features of political science departments relate to regime type.

Why political science may (or may not) be contingent on regime type

The empirical examples outlined in the previous section point towards a more general observation made by Newton and Vallès (Reference Newton and Vallès1991): The development of political science teaching and research requires organised settings, which, in turn, hinge on the state providing the necessary framework conditions (access to funding, absence of bans and censorship, etc.). When those governing the state are unwilling to provide these conditions, the study of politics will not flourish. However, we have already indicated that the literature offers two somewhat competing arguments about the potential gains and risks associated with the institutionalisation of political science, especially in nondemocratic contexts. Let us detail these two arguments in turn.

The first argument, which sees political science as a handmaiden of the state and a source for training effective bureaucrats and diplomats, highlights gains that come from having political science departments and which are independent of the regime type. The research and graduates produced by political science departments provide knowledge and expertise in various areas that are required for effective governance in different realms – from steering the economy to information control and domestic security to international relations – with the resulting improvement in performance potentially contributing to increased government and regime stability. Even under autocratic regimes, co-optation of higher education by the regime can create framework conditions that turn universities into anchors of authoritarian stability (Perry Reference Perry2020). Following this logic, one would expect that incumbent politicians in democracies and autocracies are interested in establishing political science departments to reap the benefits that the discipline provides, suggesting few systematic differences by regime type. In the extreme – that is, if these concerns are dominant and similar across regimes – the likelihood that universities have political science departments should be independent of regime type.

In contrast, the second argument, which highlights the democratic and normative roots of political science, would lead one to expect differences across regimes. More specifically, without democratic freedoms – notably free speech and academic freedom to select topics of study without political interference – organised political science units may find it hard to thrive. Studies of higher education in general find that some illiberal regimes reduce access to universities especially in humanities and social sciences (Schofer, Lerch, and Meyer Reference Schofer, Lerch and Meyer2022). Political science may be especially prone to such limitations: ‘Perhaps more than any other academic social science, political science is a threat to anti-democratic regimes, which invariably react by trying to close it down’ (Newton and Vallès Reference Newton and Vallès1991: 227). Examples mentioned in the prior section such as China under Mao or Stalin’s Soviet Union are cases in point. The logic of the argument is straightforward: Political science may be seen as a dangerous subject – in terms of questioning regime legitimacy or otherwise destabilising the current regime – by regime elites in undemocratic societies.Footnote ¹ The costs of restricting funding or even banning the discipline may thus be viewed as lower than the expected gains from, for example, having well-trained bureaucrats. Since institutionalising a discipline by dedicating a department to it presupposes the (at least tacit) support of public authorities, any ideological aversion against the discipline will likely lead to resistance to its institutionalisation. In particular, the discipline’s history of playing a key role in legitimising democracy after successful transitions from authoritarian rule, as, for example, in Germany (Blum and Jungblut Reference Blum, Jungblut, Brans and Timmermans2022; Newton and Vallès Reference Newton and Vallès1991), further contributes to its ‘image problem’ among many non-democratic governments. Hence, one might expect that universities are more likely to have political science departments in more democratic countries.

The argument above notwithstanding, limiting political science as an institutionalised discipline (to ensure regime stability) does come with costs and trade-offs for the regime, even in non-democratic contexts. As alluded to in the first argument above, the discipline potentially plays important roles in training civil servants and diplomats and delivering knowledge about the functioning of the state and international relations. These outcomes may benefit also authoritarian governments (Fu Reference Fu, Easton, Gunnell and Graziano1990; Jinadu Reference Jinadu, Easton, Gunnell and Graziano1990). Insofar as the latter concerns are deemed as more important by at least several authoritarian governments, we may expect the difference between democracies and autocracies to be relatively small. This could especially be the case if authoritarian governments adopt various other strategies that allow them to have (and thus reap most of the benefits of) political science departments while at the same time mitigating their political risks. Some such strategies could pertain to influencing where political science departments are located, and in the following we discuss different locational aspects of relevance.

One relevant feature is the spatial location of political science departments. Autocratic regimes could locate political science departments strategically within their territories to mitigate the above-described trade-off and profit from the benefits of having some form of institutionalised political science education and research while trying to limit the threat to regime stability. Studies on autocratic systems highlight that leaders often put a high price on recruiting loyalists to bureaucratic positions, even when these loyalists are less competent (for example, Zakharov Reference Zakharov2016). Loyalists may be distributed unevenly across the territory, for instance, being concentrated in the capital area (in part because autocrats have strong incentives to relocate their capitals to cities populated by loyalists, with correspondingly lower risks of coups and revolutions, see Knutsen, Morgenbesser, and Wig Reference Knutsen, Morgenbesser and Wig2024). Indeed, the capital area may be particularly advantageous for locating the training of political science students also for other reasons than recruitment of loyalists into the bureaucracy and diplomacy: Insofar as security agencies and secret services that perform key monitoring and repression tasks for the regime are concentrated in the capital, this may help the regime in screening and constraining potentially dangerous political science students or faculty members tempted to teach ‘dangerous’ ideas in the classroom. Hence, we might expect that political science departments, all else equal, would be located closer to the capital of the country in autocracies than in democracies.

Yet, we discussed above how these concerns related to political scientists being opportunistic troublemakers may be far less relevant to many governments, including those presiding over very autocratic regimes, than the potential governance benefits accompanying the research and teaching of political science. Insofar as political science is primarily viewed by politicians – democratic and autocratic ones alike – as a handmaiden of the state and a discipline that is mainly useful for supporting the effective exercise of power, institutions that provide political science research and education should be located close to the centre of political power. In other words, we should expect that universities with political science departments would be located closer to the capital of the country than universities without political science departments, and this relationship should be visible (and fairly similar) regardless of regime type.

Another locational aspect of political science departments pertains to the type of university rather than geography. More specifically, who owns and operates the university could be relevant in terms of incentives to facilitate the existence of a political science department. If politicians perceive political science predominantly as a discipline that serves as a handmaiden of the state, one might expect that – independent of regime type – governments would have an interest in offering political science mainly in public universities. This would enable the state to have a more direct say about the curriculum as well as the focus of political science research so that it fits state demands. If so, we should observe, in autocracies and democracies alike, that public universities are more likely than private universities to have a political science department.

However, if political science is perceived, at least by many autocratic leaders, mainly as a ‘dangerous’ discipline in need of control, then public universities, which are directly linked to the state, should be even more likely to have political science departments (compared with private universities) in autocracies. In addition to using loyalty signals as recruitment mechanisms for students and academics, and thus limiting the risks from public universities (Haass Reference Haass2025), public ownership, funding, and management of education bodies may enable better control over curricular content and indoctrination efforts (see, for example, Neundorf, Nazrullaeva, Northmore-Ball et al. Reference Neundorf, Nazrullaeva, Northmore-Ball, Tertytchnaya and Kim2024). Additionally, several western universities have expanded their operations into authoritarian countries with the conviction that: ‘Private universities are one of the counter-majoritarian institutions that helps keep people free’ (Ignatieff Reference Ignatieff2024: 196). While allowing these private universities to operate in an autocratic context might hold several advantages for the regime, including better access to well-trained graduates, their greater independence from state control might make the regime less willing to tolerate political science education in these institutions. Therefore, it would be reasonable to anticipate that the (positive) difference in the likelihood of having a political science department between public and private universities would be larger in autocracies than in democracies. In the following section, we will present the data and operationalisation of our key variables before we proceed to test the various expectations outlined above.

Data and methods

To study the global distribution of political science departments, we draw on the WHED, which has been created and updated by the IAU in collaboration with United Nations Educational, Scientific and Cultural Organization (UNESCO) since the 1950s (IAU 2023). The WHED is the most comprehensive and authoritative dataset on universities worldwide. The data provided are only for the most recent year (2022 for our study) and thus only allow for cross-sectional analyses. This creates limitations for our study, as it prevents us from observing temporal dynamics, visible, for example, in the openings or closures of departments that could be linked to processes of democratisation or autocratisation. This also means that we cannot say with certainty how recent a phenomenon the global distribution of political science that we observe is, besides that it is the status quo in 2022. The original dataset comprises information on more than 20,000 higher education institutions, which have more than 500,000 divisions (for example, departments or schools), in total. The information is provided to the IAU by official public sources as well as the universities themselves, and it is subsequently quality-controlled by IAU staff.

To identify political science departments, we relied on earlier coding efforts by Weidmann and Rulis (Reference Weidmann and Rulis2025). The university divisions in the WHED include a subject field, where divisions are assigned 1 of around 700 subjects. Weidmann and Rulis (Reference Weidmann and Rulis2025) assigned these subjects to 26 narrow fields, following where possible the UNESCO International Standard Classification of Education (UNESCO 2013) (although the WHED categories are slightly different). From these 26 fields, we considered the following eight disciplines as ‘political science’: civics, comparative politics, government, human rights, international relations and diplomacy, international studies, peace and disarmament, and political sciences. This is a conservative measure since we excluded departments that had a more general social science denomination and included political science. For example, the University of Luxembourg has a Department of Social Science that has a dedicated ‘Institute for Political Science’, but since the department itself has a more general label, it was not counted as a political science department. Similarly, the Sultan Qaboos University in Oman has a College of Economics and Political Science that was also not counted because of its more interdisciplinary structure. Therefore, we provide analyses with variations of this measure below. To account for the possibility that university-specific differences in the ways in which universities are organised influence the likelihood of having a registered political science department, we controlled for (log) number of all divisions at the institutions as well as its age in the different regressions.

Of the 531,389 divisions in the dataset, 9096 (1.7%) were coded as political science departments, following the operationalisation laid out above. For each higher education institution in our sample, we coded a binary variable registering whether (1) or not (0) it had a political science department. After removing micro-states and institutions outside the borders of states, we ended up with 19,428 institutions, of which 25% had a political science department. From the WHED, we also obtained information about whether a higher education institution was a public or private organisation. To measure the distance from the capital, each institution’s city was geolocated using the GeoNames database (https://geonames.org). The distance was then computed as a straight line on the spherical earth from the country’s capital, the coordinates of which were obtained from the CShapes 2.0 dataset (Schvitz, Girardin, Rüegger et al. Reference Schvitz, Girardin, Rüegger, Weidmann, Cederman and Gleditsch2022). Since countries have vastly different sizes, we rank-transformed the distance by country so that institutions located in the national capital were assigned the lowest value (0) and those farthest away from the capital the highest value (1).

We measured regime type by drawing on data from Version 14 of the Varieties of Democracy (V-Dem) dataset (Coppedge, Gerring, Knutsen et al. Reference Coppedge, Gerring, Knutsen, Lindberg, Teorell, Altman, Angiolillo, Bernhard, Borella, Cornell, Fish, Fox, Gastaldi, Gjerløw, Glynn, God, Grahn, Hicken, Kinzelbach, Marquardt, McMann, Mechkova, Neundorf, Paxton, Pemstein, Ryden, von Römer, Seim, Sigman, Skaaning, Staton, Sundström, Tzelgov, Uberti, Wang, Wig and Ziblatt2024). For our cross-country analysis, we used the continuous Electoral Democracy Index (or Polyarchy), which ranges theoretically from 0 to 1 (highest level of democracy). This index focuses on the existence of multiparty elections and extension of voting rights but also the protection of certain rights and other conditions (for example, absence of violence or voter fraud) that enable these elections to be free and fair. More specifically, the index is composed of five subindices pertaining to, respectively, key officials being elected, the integrity of these elections, suffrage, freedom of speech and access to alternative information, and freedom of association.

We added several control variables to our models. Ansell (Reference Ansell2008; Reference Ansell2010) argues that, beyond a country’s regime type, its economic situation influences policy preferences regarding education. More specifically, countries that are better integrated in the global economy and have higher levels of economic development have a greater need for highly educated citizens to sustain their economic performance. This makes such countries more likely to increase access to education, which in autocracies is especially focused on higher education, inter alia because these regimes try to target education spending more to the elites (Ansell Reference Ansell2010). There are also other studies indicating that political science education flourishes in more affluent societies with wider access to higher education (Almond Reference Almond, Goodin and Klingemann1998; Newton and Vallès Reference Newton and Vallès1991). On the basis of these arguments, we added GDP per capita (in current USD, log-transformed) as a control variable to our cross-national models, using data obtained from the World Bank (World Bank 2025). To account for differences in country size affecting the likelihood of having political science departments, we also controlled for the (log-transformed) population of the country in these models.

Our benchmark specifications were ordinary least squares (OLS) regressions with standard errors clustered by country. In analyses that focus on within-country variation, we always included country-fixed effects to account for unmeasured country-specific confounding factors (including level of development, population level and settlement patterns, national educational cultures, and so forth) that may simultaneously influence the propensity for universities to have a political science department and the explanatory factors of interest (for example, public versus private university or distance of the university from the capital). Insofar as we hypothesised differences across democratic and autocratic regimes, we ran all within-country analysis on split samples. Rather than relying on a categorical definition of regimes, we distinguished between autocratic countries, hybrid regimes, and democracies by assigning countries to the first, second, and third tercile of the distribution of V-Dem’s Polyarchy index. This mirrors the strategy proposed by Hainmueller, Mummolo, and Xu (Reference Hainmueller, Mummolo and Xu2019) for testing for nonlinear interaction effects, which potentially also exist in our data. While less efficient than a standard interaction specification, this approach allows us to test interaction effects with a moderator (level of democracy) at the country-level, while retaining country-fixed effects in our models.