Introduction

Within research on policy-advisory systems and knowledge utilization, attention has regularly been turned also to scholars themselves. Caplan’s two-communities theory continues to serve as a strong metaphor for the ‘separate worlds’ that researchers and policymakers populate, with their “often conflicting values, different reward systems, and different languages” (Caplan, 1979: 459). This argument serves to help explain a perceived ‘gap’ of putting existing policy knowledge from the academic sphere into policy action (see e.g., Caplan, 1979; Schmidt, 2021). Especially in times of increasing demands for evidence-based policy, and intense public discussion about the extent and manner in which scholars should engage in policy work, the interaction between these two spheres becomes more salient (and also contested).

More recently, attention has been paid to ‘pracademics’ (Posner, 2009; Brans & Pattyn, 2017), whose careers are rooted in both the academic and policy-making community, and who “serve the indispensable roles of translating, coordinating, and aligning perspectives across multiple constituencies” (Posner, 2009: 16). While different conceptualizations of pracademics exist (Powell et al., 2018), such scholars generally tend to bridge the two worlds and through this challenge earlier conceptions of a clear separation between them. Moreover, even in more confined academic settings, scholars may be more or less engaged in policy work, and to different extents be ‘boundary scholars’ (Wittrock, 1991; Hoppe, 2005) whose work becomes both academically and practically relevant. Adding to this, the broader societal impact agenda in higher education (Bandola-Gill et al., 2021) can be read as a call for scholars to cross the boundary between academia and practice. Yet, despite these trends, little empirical evidence exists on the extent to which scholars indeed engage in boundary-crossing activities, and the conditions that foster or jeopardize boundary-crossing behavior. This is the gap that this study addresses.

As it currently stands, the literature on boundary work seems biased towards the study of boundary organizations (e.g., Guston, 2001; Miller, 2001) and boundary arrangements (e.g., Dannevig & Aall, 2015; Hoppe & Wesselink, 2014), or scholars working in specific organizations established for such purposes (e.g., Hoppe, 2009). This leaves boundary-crossing work of academics in ‘mainstream’ academic institutions sparsely covered. Available empirical research tends to be case-based, and predominantly focuses on objects of research (e.g., research projects, units, results) as the starting point of analyses (Tellmann & Gulbrandsen, 2022). This makes it challenging to uncover patterns across countries, or policy fields. Other than this, in line with Gieryn’s (1983) seminal contribution, the literature mainly emphasizes ‘competitive’ boundary work, in which individuals or groups are working to create boundaries and ‘demarcate’ those to protect their status or territory (Tellmann & Gulbrandsen, 2022; Halffman, 2003). Such emphasis overshadows scholars’ active engagement in transcending or bridging boundaries by deliberately interacting with policymakers. Finally, empirical evidence is scarce when it comes to factors explaining boundary-crossing practices. Thus, conceptual frameworks on boundary work (e.g., Hoppe et al., 2014) require more back-up by systematic empirical evidence. The present study provides some of this evidence by focusing on boundary-crossing activities of individual scholars who are employed at political science departments in universities in 39 European countries. Through the multi-country comparison and survey data combined with the clear focus on boundary-crossing activities, this study complements existing knowledge both from the literature on boundary-spanning but also policy advisory work. The main research question is whether there are different types of boundary-crossing scholars, and, if so, what makes academics more or less likely to engage in boundary-crossing activities.

By studying work at and between the boundaries of science and society, we join a more recent literature strand on collaborative boundary work (Langley et al., 2019; Tellmann & Gulbrandsen, 2022). With policy advising having a central role in how we operationalize boundary-crossing behavior, our study also contributes to the growing body of literature on policy-advisory work (Brans & Timmermans, 2022; Jungblut et al., 2023; Marciano & Craft, 2023; Craft & Howlett, 2013). To account for different degrees of intensity in which scholars can be involved in boundary-crossing work, we propose a new typology that can capture different boundary-crossing profiles building on recent literature.

With its empirical and comparative focus, our study adds to scholarship on political science as a discipline (Goodin & Klingemann, 1996), which received reinvigorated attention in recent years (Capano & Verzichelli, 2023; Hendrix et al., 2023; Flinders & Eisfeld, 2021; Boncourt et al., 2020). The focus on political science may be conceived as a critical case study (Flyvbjerg, 2006), because the discipline is by nature concerned with its subject matter of ‘real life politics’, and the image of a ‘science for democracy’ has traditionally strong roots in the self-understanding of the discipline. More than for other disciplines, perhaps, “for scholars working within the social and political sciences, more specifically, issues of relevance increasingly matter” (Flinders & Eisfeld, 2021: 2). Whether political scientists conceive ‘relevance’ as engaging in policy work is less evident though and likely varies across contexts and individuals.Footnote 1

To investigate the boundary-crossing work of European political scientists, we rely on an original dataset (Luca et al., 2023). We test several hypotheses at the micro-level, and at the meso-level, that can explain variation in boundary-crossing intensity. We show that European political scientists are rather frequent travelers across the border between academia and the policy world. Moreover, individual characteristics play a key role for the likelihood of being involved in boundary-crossing: scholars who publish more regularly are more likely boundary crossers as are more senior and male scholars. Having (or having had) a political or administrative position has an especially strong effect on the likelihood of boundary-crossing. Finally, we also uncover regional and sub-discipline effects. These show that those working in a more policy-relevant subfield tend to engage more in boundary-crossing work, and that political scientists from the southern European region are less likely to cross boundaries, compared to their northwestern European counterparts. Political scientists from third countries outside the EU, such as Russia, Serbia, Israel, or Turkey, turned out to be more active than northwestern Europeans.

In the following, we first develop our conceptual framework and discuss how we analytically distinguish between different types of boundary-crossing scholars. Next, we present the hypotheses that guide our study. We continue with a presentation of our methods and data, and our findings. The article concludes by discussing the findings’ implications for further theoretical and empirical advancement on boundary-crossing scholars and policy advice.

Boundary-crossing scholars

Conceptual foundations and typology

In 1979, Caplan published his seminal work on the Two-Communities Theory, and scholarly debate around knowledge utilization and policy advice has been engaging with this since (see, e.g., Newman et al., 2016; Löfgren & Bickerton, 2021). According to Caplan’s strong metaphor, policymakers and social scientists “live in separate worlds with different and often conflicting values, different reward systems, and different languages” (Caplan, 1979: 459). He presented this argument in explicit response to an ‘abounding literature’ on why social science knowledgeFootnote 2 has little impact on policy – and it is fair to say that such claims and questions have not become less relevant more than 40 years after Caplan’s observations (see e.g., Schmidt, 2021). Often connected is a negative portrayal, according to which policymakers do not make use of expert knowledge or use it only to legitimize pre-determined proposals.

A key literature strand has engaged with the ‘separate worlds’ by focusing on the boundaries that exist between them, and on where, how, and by whom these boundaries are drawn. In line with Caplan’s arguments, Gieryn (1983) argued how demarcating the boundaries between ‘science’ and ‘non-science’ is serviceable for scientists in pursuit of their professional goals and according to the scientific reward systems. Vice versa, Jasanoff (1987) showed how boundaries are drawn from within the rule-making and administrative process, where scientific expert advice is regularly deconstructed, and its weaknesses (such as scientific uncertainty) highlighted through particular uses of language.

Next to these efforts –both from academia and the policy world– to demarcate boundaries between the fields, however, attention has also been turned to the efforts and mechanisms of crossing boundaries and establishing science-policy interaction. In such processes of knowledge exchange, or even collaboration and co-production (Bandola-Gill et al., 2022), drawing clear boundaries may become less relevant (or even detrimental), while “establishing good working relations” (Turnhout et al., 2008: 229) gains in importance. Also, the very definition of ‘boundary work’ is then approached differently, namely that it does not only involve the ‘demarcation’ of science and non-science but also a coordinative function (Hoppe, 2005) or ‘knowledge brokering’ (e.g., Turnhout et al., 2013; van Enst et al., 2017). In that sense, boundary work also “defines proper ways for interaction between these practices and makes such interaction possible and conceivable” (Hoppe, 2005: 207; following Halffman, 2003).

The different conceptualizations of ‘boundary work’ (see also Van Ernst et al., 2017) may enrich the field, but challenge systematic research on the topic. The fact that the concept is used to refer to different types of actors, which can be situated at either side of the science-policy boundary, adds to the complexity. In our study, we therefore focus solely on boundary work by scientists, who are affiliated with an academic institution. As mentioned, scientists working at mainstream academic institutions are often overlooked in the literature on boundary work. We label them as boundary-crossing scholars. When zooming into the type of activities carried out by boundary workers, a plethora of approaches can be identified, in line with the many conceptual understandings and actor foci.

Typically, boundary work is associated with three functions: translation and facilitating a mutual understanding between scientists and policymakers; communication between both communities; and mediation (Van Ernst et al., 2017). Still, these functions can encompass multiple activities when operationalizing them in concrete terms. Given our focus on academics, we join previous conceptual contributions and empirical studies on boundary-crossing work (see, e.g., Hoppe & Wesselink, 2014Footnote 3; Hoppe, 2005Footnote 4) in approaching it from the lens of policy-advisory activities. As it has been put in the context of organization studies, advice networks and related advisory activities “constitute the social plumbing system that allows knowledge and experiences to be shared, interpreted, transferred” (Lomi et al., 2014: 439) across organizational boundaries. Scholars can engage in different advisory activities, which captures the idea of an active and tangible undertaking of academics to take steps from the academic to the policy world.

We deliberately apply a broad understanding of policy advice to make it applicable to any policy topic. In line with Tenbensel’s (2008) types of knowledge (modelled after Flyvbjerg, 2001), and as developed further by Brans et al. (2021), we posit that policy advice can be of scientific nature (i.e., about what is objectively true), experiential advice of what works in practice, and more normative advice that indicates what should be done. In concrete terms and anticipating our focus on political scientists as one group of academics, policy advice can encompass a wide range of engagements with policy actors, including the provision of data and facts, analysis and explanations of causes and consequences of policy problems, evaluation of existing policies and functioning of institutions, giving recommendations on alternatives, forecasting and polling, or presenting value judgements and normative arguments (Brans et al., 2021: 25).

As hinted above, scholars may engage in boundary-crossing work using different levels of intensity. These reflect distinct manifestations of boundary-crossing activities. We suggest a novel typology that distinguishes between three types, amongst which one type who remains in the academic world is complemented by two types of actively boundary-crossing scholars:

  1. 1.

    The Stay-at-home: Academics falling into this type avoid travelling into the policy world. They focus on core academic work; in any case they do not engage in policy-advisory activities. To be sure, also the work of stay-at-home academics can have impact on policymakers, but this will not be supported through any active advisory work on their behalf. In other words, impact on policymaking activities could happen through academic work being picked up by policy actors, but without academics specifically promoting it.

  2. 2.

    The Visitor: Academics falling into this type are rooted in the academic world, but occasionally pack their bags to travel across the boundaries into the policy world, make contact, and engage in policy-advisory activities.

  3. 3.

    The Commuter: For academics falling into this type, policy-advisory activities are part and parcel of their professional life. They travel easily and regularly back and forth between the academic and the policy worlds, therein often actively contributing to boundary-crossing work in the above-named coordinative and knowledge-brokering sense.

Figure 1 illustrates the idea of types of boundary-crossing scholars graphically. It is -as the terminology chosen for the types- inspired by the notion of ‘boundaries’ between worlds. It also resembles the images of three distinct arenas (academic, governmental, societal) (see: Blum & Brans, 2017). Admittedly, Fig. 1 is stylized, as in reality borders between the worlds may be less clear-cut, and different ‘regions’ within them are not portrayed (e.g., the media, or industry). In this article, however, we focus specifically on the boundary-crossing activities of academics to the policy world and thus limited the figure accordingly.

Fig. 1
figure 1

Types of scholars at the borders of the academic and the policy world Source Authors

Full size image

What makes a scholar act as stay-at-home, visitor, or commuter?

Having outlined different profiles of boundary-crossing scholars, the question is how these manifest in practice and especially how individual variation can be explained. To this end, we draw on insights from studies on the academic and political science profession and policy advisory structures.

Given our study’s focus on individuals, it is intuitive to expect that actor characteristics will determine variation in boundary-crossing behavior. To start with, having a well-established network with the policy world will foster more engagement in boundary-crossing activities, particularly since such activities are often based on demand from policymakers. Being well connected and having network ties with organizations beyond the academic sphere thus increases the chances for engagement in boundary-crossing work. As it has been shown in studies on network-based mechanisms in (business) organizations, having ties outside the organization’s boundaries increases network range, and facilitates the development of new contacts in similar communities (Lomi et al., 2014). With information often being asymmetrically distributed and given the complexity of evaluating the competence of potential new partners, actors tend to rely on referrals and reputational lock-ins offered by their existing network when looking for advice (Sytch et al., 2012). Therefore, one can assume that political science scholars who held political or administrative office before or during their academic appointment will have a relative benefit to colleagues without such experience. Similarly, given their policy work track record, they can be expected to have developed more affinity with and skills in translating knowledge to policymakers. Therefore, we posit that:

Hypothesis 1

Scholars who hold / held political or administrative offices are more likely to be commuters. Scholars without such experience will be more likely ‘stay at home’ academics.

Second, boundary-crossing work is but one of a range of activities in which scholars can profile themselves in the highly competitive environment of academia (Stroebe, 2014), and scholars may wish to prioritize other work. Political science as a discipline has not escaped the publish-or-perish trend (Hesli & Lee, 2011), and it is plausible that scholars feel that boundary-crossing work may come at the cost of publication productivity. Moreover, policy-advisory work competes with publication productivity as scholars will have less time to work on publications if they dedicate more time to boundary-crossing. While policy-advisory work can certainly be ‘prestigious’, there are also debates about its potential cost for academic esteem, as too much engagement with the policy world may be associated with ‘diluted’ science. Therefore, we believe that extensive engagement in boundary-crossing work will be especially undertaken by scholars who are less active in academic publishing.

Hypothesis 2

Scholars with a high publication productivity are less likely to be a commuter and more likely to be a ‘stay-at-home’ academic.

Building on this hypothesis, one can as well expect that age is a decisive factor for strong engagement in boundary-crossing work (see, e.g., Blum & Jungblut, 2022). Two elements support this assumption. First, it will especially be younger faculty members who feel the need to publish to advance their academic career (Stroebe, 2014). Publication pressures can to some extent ease off for senior scholars, especially once they reach a certain rank (Hendrix et al., 2023). Seniority may as such come with more discretion to take on boundary-crossing work. In Smith and Stewart’s (2017) study, for example, interviewees indicated it to be riskier for earlier career academics to be involved in activities required to achieve policy impact. On top of this supply-side view comes a demand-side perspective: engagement in policy-advisory activities is often only possible upon invitation, which will as such benefit scholars with a longer career-span, a certain position, or more opportunities for networking to gain visibility in the policy world.

Hypothesis 3

Older scholars will more likely be commuters or visitors, whereas junior scholars will more likely be ‘stay-at-home’ academics.

In addition to age, gender is a relevant variable for differences in boundary-crossing work. Within political science, the ‘marginalization’ (Atchison, 2018) of women has received increasing attention (Beaulieu et al., 2017; Engeli & Mügge, 2020). Female political scientists are disadvantaged relative to their male colleagues when it comes to being published (Teele & Thelen, 2017), being cited, being invited to contribute to edited volumes, or conference panels (Beaulieu et al., 2017). Also, within policy advice, ample evidence has proven an underrepresentation of women, for instance, amongst members of expert commissions (see: Lepperhoff, 2006). Given these observations, one can expect that it will be primarily male political scientists who will act as more active boundary-crossing scholars. In parts, the underlying mechanism work similarly to those regarding age: female scholars have less visibility as well as presumably less opportunities to build a network with the policy world and thus will be invited less frequently for policy-advisory work. Relatedly, precisely because female colleagues face more challenges to reach prominence in the field, it can be assumed that they will engage less in boundary-crossing work for which career benefits are less straightforward, and which can compete, timewise, with their core academic activities. This is amplified by the persistent gender divide in caring responsibilities, which makes female academics more likely to face higher time constraints (see: Smith & Stewart, 2017).

Hypothesis 4

Male scholars are more likely to be commuters, while females are more likely to be ‘stay-at-home’ academics.

Next to these four hypotheses, our analyses will investigate broader structural environments by looking for variation across different regions and policy areas. Indeed, gender and age differences in advisory activities have not been confirmed by all empirical studies (see, e.g., Migone et al., 2022). Moreover, a recent study of internationalization efforts in European political science clearly highlighted differences between regions of the continent (Tronconi & Engeli, 2022). Following their argument, we will also investigate differences between parts of Europe. Also, as extensively discussed in public policy scholarship, policymaking takes place in multiple venues in many sectors spread across government, each having its own institutions, networks, or ideas (Cairney, 2021). The notion of policy styles (Freeman, 1985; Richardson, 1982; Howlett & Tosun, 2021) at sectoral level (Cairney, 2021) captures these non-uniform policy processes which are determined by the functioning of ‘policy subsystems’ (e.g., Sabatier, 1988; Baumgartner & Jones, 1991; Howlett et al., 2009). The fact that policy processes are influenced by the sector(s) to which a policy problem is assigned, also implies that different sources of advice may be prioritized in a given policy sector. Political science is a discipline with a varied set of sub-disciplines, some of which are inherently closer to the policy world. Therefore, we will also explore whether patterns of boundary-crossing work of political scientists vary substantially across sub-disciplines.

Methodology

Our analysis is based on data from an original survey sent out to 12,400 political scientists at universities working in 39 European countries between March and December 2018. Targeted political scientists were invited to participate up to four times (one initial invitation and three reminders). The total number of completed surveys before applying any of our restrictive sampling criteria was 2,354, with national response rates ranging from 7% in Turkey to 70% in Albania. The average response rate was 26%. Survey questions focused on respondents’ advisory activities, but also tapped into the state of political science in the respective country. The questionnaire was originally drafted in English, but was translated in several other languages (e.g., French, German, Italian, Polish, Spanish) aiming at enhancing access. Respondents could choose to complete the questionnaire in English or in one of the other languages.

To be part of the final sample for the analysis, respondents had to satisfy the following criteria: i) holding a PhD in political science and working at academic research institutions or being affiliated to formal organizational units within universities (e.g., departments) whose main specialization is political science or a similar field (e.g., public administration, international relations); and ii) focusing on research or teaching topics directly related to political science or other related political science subjects. Applying these two restrictive criteria left us with a sample of 1,995 respondents.

Operationalization of the dependent variable

Building on the initial survey design, our dependent variable utilizes a question inquiring about the types and frequencies of advisory activities respondents engaged in. The lead-in to the items was the following: “How often, on average, during the last three years, have you engaged in any of the following advisory activities with policy actors:”. The six survey items were then asking about the provision of:

  • data and facts about policies and political phenomena.

  • analysis and explanation of causes and consequences of policy problems.

  • evaluation of existing policies and institutions.

  • recommendations for policy initiatives.

  • forecasts and polls.

  • value judgments and normative arguments.

Survey respondents had to report how often they engaged in those activities, using the following categories: “never”, “less frequently”, “at least once a year”, “at least once a month” and “at least once a week”.

We then used the responses to construct our categorical dependent variable, with three main categories:

0 = no advisory activity whatsoever in the last three years, meaning “never” for all survey items about advisory activities.

1 = engagement in at least one type of advisory activity “once a year” or “less frequently”.

2 = engagement in at least one type of advisory activity “once a week” or “once a month”.

We finally assigned labels to those three categories. “0 = the stay-at-home”, “1 = the visitor” and “2 = the commuter”.

Operationalization of the independent variables

The survey included questions about publications and positions outside academia, we used these responses to create our first two independent variables. “Frequency of publication” is a dichotomous variable:

0 = a maximum of one publication in whatever format (articles, book chapters or books) over the last three years.

1 = at least two or more publications in whatever format over the last three years.

The variable “Frequency of international publication activity” is similarly constructed but targets more specifically publication activities in international outlets:

0 = a maximum of one publication in English and/or in a journal outside the respondents´ own country over the last three years.

1 = at least two or more such international publications over the last three years.

The independent variable “Position outside of academia” is also dichotomous:

0 = never held any positions outside academia.

1 = held/is holding at least one type of position outside academia.

We used data on respondents’ year of birth to calculate their age in years. While we also have data on respondents’ academic position, which could be a proxy for seniority, we decided to use the age measurement as both variables are highly correlated and as academic employment systems and career structures differ widely across the 39 countries, we felt that age was a more equally applicable operationalization.

We also used the categorization proposed by Tronconi and Engeli (2022) to cluster countries into regional groups within Europe to control for regional differences.Footnote 5 Finally, the independent variable ‘Sub-disciplinary background’ is a dummy variable differentiating respondents by the subfield and its relevance for advice-takers in the policy world. Fields that we considered particularly relevant for policy advice were public policy, public administration, local government, electoral behavior, and political institutions. All other fields (e.g., political theory, social science methods) were put in the other category.

Data analysis

We conducted multiple multinomial logistic regressions with a three-category dependent variable. In addition to our baseline model, we ran a multinomial logistic regression model with clustered standard errors at the country level, to control for the across-country data structure. Moreover, we ran the same kind of regressions both with a control variable for the region in which a country is and then with country-fixed effects. Finally, we tested the robustness of our results by using a finer-grained index as dependent variable (see appendix A3). A correlation matrix for the independent variables is provided in the appendix (appendix A2).

Results

Our analyses yield some interesting findings. Already the descriptive statistics (see table A1, in the Appendix) of our dependent variable deliver an interesting first view on the phenomenon of boundary-crossers. It seems to be rather common among European political scientists to travel across the border between the academic and policy worlds. Most respondents have either been a visitor or commuter in our conceptualization of boundary-crossers indicating that it is rather common to have some level of activity in both worlds. In fact, those who only stay in the academic world are the smallest of the three groups. However, being a visitor is much more common than being a regular commuter.

Looking at more detailed analyses, Table 1 shows that publishing more frequently increases the probability of falling in either the “visitor” or “commuter” category, when setting the “stay-at-home” category as baseline. Both effects are highly significant (p-value = 0.000). Looking at the regression coefficients, especially in the baseline model, if individuals publish frequently, they have a larger chance to end up in the visitor category (= 0.783) than in the commuter category (= 0.594). This could be an indication that it is difficult for scholars to maximize their activity regarding both publication and policy advice at the same time. Scholars who held a position outside of academia over the course of their career have an increased probability to cross boundaries and fall either in the visitor or commuter category (p-value = 0.000). For individuals who have had a position outside of academia, there are very large chances that they become commuters (regression coefficient 1.539). This clearly highlights the importance of contacts and networks in the policy world for engaging in policy advice.

A one-year increment in age makes it more likely for individuals to fall in either of the two active boundary crossing profiles too. These findings provide some support for the idea that more senior scholars have an easier time traveling between the worlds, but the effect is rather cumulative as a single year increase in age does only have a limited effect. Figure 2 illustrates the effect of age (based on model 1 outputs). Individuals of all ages still are more likely to be visitors than stay at home or commuters, but there are age-specific variations. Compared to older political scientists (aged 60–65), younger political scientists (aged 25–35) are more likely to stay at home, and to some extent less likely to cross boundaries as visitors or commuters. Similarly, increasing in age decreases the likelihood for staying at-home while it increases the likelihood of visiting or commuting. Our analysis also showed that female scholars have a reduced probability to fall in the commuter category (p-value = 0.049 in the baseline model). Results for the visitor category were not significant, but the direction of the relation also indicates a lesser likelihood for female scholars, at least in our sample.

Controlling for disciplinary background and geographical location of political scientists proved to be important (see model 2). Colleagues working in a more policy-relevant subfield tend to engage more in boundary-crossing work, and especially act as ‘commuters’. Political scientists from the southern European region are less likely to cross boundaries, be it as commuters or visitors, compared to their north-western European counterparts. Political scientists from third countries outside the EU, such as Russia, Serbia, Israel, or Turkey, turned out to be more likely commuters than northwestern Europeans (p-value = 0.038)Footnote 6. In model 3, gender and age of political scientists do not seem to interact with each other to alter boundary crossing patterns in any way.

Table 1 Main results from multinomial logistic regression models
Full size table
Fig. 2
figure 2

Predicted probabilities of boundary-crossing as a function of political scientists’ age

Full size image

Finally, we ran another multinominal logistic regression model to gauge the effects, not of general publication activity, but of international publication activity as an independent variable. Table 2 below presents the results of this model. Here, patterns like the ones presented in Table 1 are visible: A higher international publication activity comes with a greater likelihood of being either ‘visitors’ or ‘commuters’, but the likelihood is even stronger to be a ‘visitor’. However, comparing the effects of general publication activity and international publication activity (in Tables 1 and Table 2), it seems that the former is associated with a somewhat stronger likelihood of boundary-crossing.

Table 2 Results from multinomial logistic regression model changing general publication activity for international publication activity
Full size table

Robustness checks

To check the robustness of our models, we ran an alternative regression model using a different operationalization of the dependent variable, namely a simple additive index, summing up responses from the six survey items about advisory activities. We ended up with a continuous variable ranging from 0 to 24, 0 indicating a complete absence of advisory involvement and 24 standing for very frequent involvement in all 6 types of advisory activities. We ran linear regression with the same predictors as in Table 1. Regression outputs for this model are displayed in Appendix A2. The results of this robustness check are generally in line with our main results, given that we observe the same direction of the effects of all our predictors and no serious drop in the effects´ significance.

Discussion

The results of our analysis highlight some interesting findings regarding our five hypotheses. Already the distribution of the three ideal types of scholars revealed that it is common for political scientists to move between the academic and the policy world. Most respondents indicate that they were active in some form of boundary-crossing activities in the last three years and the majority of respondents falls in the visitor category. At the same time, the category of those who stay at home has the lowest number of respondents. This already tells us that European political scientists seem to be frequent travelers who interact regularly with policymakers. However, our main interest was to go beyond a mere description of the status-quo as we wanted to know, in how far certain individual or contextual factors influence the likelihood of being a boundary-crosser.

In our first hypothesis, we expected that scholars who held a political or administrative office will be crossing between the academic and policy world more frequently. Our analysis supports this hypothesis as we show that those scholars who held or are holding a position outside of academia are more likely to fall in the categories of visitor and commuter. Here, the likelihood is even stronger to be in the commuter category, indicating that having expertise from the policy world is especially important for high-frequency boundary-crossing. There could be different explanations for this, as outlined above, and it is impossible with our data to distinguish between them. It could, for example, be that scholars who held an office have the right networks and necessary visibility that allow them to be more active in the policy world. At the same time, it could also be that those who worked in the policy world have the skills necessary to transmit their knowledge in a language that is easier accessible for policymakers making them more sought-after advisors.

The second hypothesis focused on the level of academic productivity. The results of the analysis show the opposite pattern than we expected. We found that individuals who publish more frequently are more likely to fall in the visitor or commuter category. However, the effect size is not as strong for the commuter category as for the visitor category. Thus, our results paint a mixed picture. On the one hand, publishing more actively makes it both more likely to be a visitor and a commuter. This highlights that those who interact with the policy world also tend to be active producers of academic knowledge, which could be an indicator for having a strong and up-to-date knowledge base serving as the basis for policy-advisory activities. However, our data do not allow to match the topic of publications and policy advice, thus it could be that both are de-coupled. On the other hand, the results also show that the effect of publishing more actively is also stronger regarding visitors than commuters. This could be an indicator that there is a certain trade-off regarding the available time in the sense that if a scholar is publishing more actively it might be easier to combine this with once-in-a-while visits to the policy world rather than with being a frequent commuter.

Our third hypothesis focused on academic seniority measured through the age of respondents. The results of the analysis support part of our hypothesis. Increasing in age makes it more likely to fall in the commuter and the visitor category compared to the stay-at-home group. However, even among the youngest age cohort, the likelihood to be a stay-at-home academic is lower than being a visitor (see Fig. 2). At the same time, the likelihood of being a stay-at-home academic decreases over the increasing age groups, while the likelihoods of being a visitor or commuter increase. These results highlight that academic seniority is advantageous for boundary-crossing activities. Our data do not allow to distinguish though whether this advantage is driven by supply or demand factors, namely whether more senior academics have the freedom necessary to pursue regular travels across the border (supply) or if policymakers are more likely to invite older colleagues into their world (demand), or whether it is an interaction of the two.

Finally, our fourth hypothesis that expected that men are more likely to be active boundary-crossers was also supported by the results, as they have been found to be significantly more likely to be commuters. Regarding the visitor group the results are not statistically significant but the effect points in the direction that we expected. Again, it is not possible for us to distinguish whether this effect is due to supply or demand factors or whether it is a mixture of both. It remains clear though, that both older and male political scientists are more likely to be involved in boundary-crossing activities.

Controlling for the sub-disciplinary background and regional context of respondents also produced interesting results. Having a background in more policy-relevant sub-disciplines seems to influence the likelihood of engaging in boundary-crossing work. However, it is not possible with our data to discern whether this is driven by the scholars’ own pre-disposition to crossing boundaries, whether certain forms of research are more sought after in the policy world, or whether it is an interaction of both. Besides, our analyses revealed clear regional differences between north-western and southern Europe as well as northwestern Europe and non-EU contexts. There seems to be a more restrictive border between academia and the policy world in southern Europe compared to north-western Europe (see also Pritoni & Galanti, 2023) and it seems easier to commute (but not to visit) between both worlds in non-EU contexts compared to north-western Europe. While our data do not allow us to fully explain these differences, our results clearly point to the relevance of earlier findings regarding e.g., policy-advisory systems (Craft & Howlett, 2013), different regional traditions of advice-giving (e.g., Pattyn et al., 2022; Hustedt, 2019), or the role of impact-regimes (Bandola-Gill et al., 2021) in creating regional differences in policy advice and boundary-crossing.

Conclusion

This study aimed at providing more empirical evidence regarding the question which scholars engage in boundary-crossing activities, and which individual characteristics and contextual conditions foster boundary-crossing behavior. To this end, we investigated in how far political scientists from 39 European countries are active in boundary-crossing work. To conceptualize this relationship, we developed a novel typology of boundary-crossers and formulated hypotheses regarding the relation between individual and context-level characteristics and the likelihood of being active in both the academic and policy worlds.

Most respondents indicate that they are active in both worlds and thus cross the border between them regularly. This shows that political scientists in Europe are actively reaching out to policymakers and that they are aware of the social and political relevance of their work (see also Jungblut et al., 2023). The high volume of political scientists engaging in boundary-crossing work sheds a different light on Gieryn’s work (1983) on boundary-demarcating behavior. At the same time, this confirms that the two worlds are not as disconnected (anymore) as Caplan’s (1979) two communities’ metaphor might suggest. Moreover, our results indicate that those scholars who publish more regularly are also more likely to be among the boundary-crossing academics. Thus, it seems that earlier accounts, including Gieryn’s (1983), that highlight the trade-off between academic work and policy-advisory work do not describe the relationship in an appropriate manner, as there rather seems to be a mutually supportive effect.

At the same time, our analysis also found that publishing more regularly has a stronger effect on being a visitor rather than being a commuter. This could indicate that there is a certain level of trade-off, which could be linked to the available time budget of each individual scholar. Our results also reveal that having (held) an administrative or political office has a strong positive effect on the likelihood of crossing the border between the academic and the policy worlds. Whether this effect is driven by the existence of networks, or by scholars having learned how to properly transmit their knowledge to the policy world is something our data do not allow us to unpack. However, future research should have a closer look at this relation to better understand which mechanism is behind it.

The different boundary-crossing types distinguished here are not understood in any normative sense of some being more desirable than others. If we should make any normative argument at all, then it would be for a plurality of roles that may complement one another. Yet on the more worrying side, our results clearly show that not every scholar has the same chance to be involved in boundary-crossing work. More senior scholars are more likely to be frequently active in both the academic and policy world. The same applies to male scholars, which echoes concerns about marginalization and inequalities affecting female political science scholars (Atchison, 2018; Engeli & Mügge, 2020). Thus, the diversity of boundary-crossers is more limited compared to the population of our respondents.

Our analysis also uncovered both regional and sub-disciplinary differences. Similar to findings about the internationalization of political science in Europe (Tronconi & Engeli, 2022), our study highlights that working in a specific regional context has an effect on the likelihood of being active in both the academic and the policy world (see also: Brans & Timmermans, 2022; Pritoni & Galanti, 2023). Here, it seems that especially colleagues in southern Europe face challenges in crossing the boundary, a finding that deserves more in-depth follow-up research (see also Pritoni & Galanti, 2023). Also, the sub-disciplinary background of respondents was found to be relevant for their level of boundary-crossing activities. Those colleagues who work in areas that could be described as more relevant for advice-takers also seem to be more active in crossing the boundary towards the policy world.

Our results show that political scientists in Europe are active in providing policy advice. However, the effectiveness of this provision is somewhat hindered by the fact that we do observe inequalities among sub-groups. This has practical implications for the discipline. If we want to ensure that the breadth and diversity of the discipline is represented in those providing policy advice, we need to be mindful of some of the findings and consider ways in which we can balance out inequalities. The gender- as well as age-gap in the provision of advice could, for example, be partly addressed by more explicitly valuing policy advisory activities in promotions or hiring processes. While it is obviously necessary that advice is provided based on relevant research, a greater valuation of advisory activities could still help to make the trade-off between publishing and provision of advice less difficult especially for younger and female colleagues who still face greater challenges in the discipline. At the same time, our results can also be seen as a call to those colleagues working in sub-disciplines that are less likely to provide policy advice to become more active. While some work within political science seems easier to transfer to the policy world just by the virtue of its research focus, we are convinced that in general all sub-disciplines produce knowledge that can be relevant for policymaking. Some research might need more translation work to speak to the policy world, but as a discipline we should aim to make the diverse breadth of our work relevant for as well as seen by the policy world. Our study has some limitations that need to be taken into consideration when assessing the results. First, the varying response rate of the survey as well as the focus on political scientists provide some limitations regarding generalization of our findings. We would expect some disciplinary effects being at play here. As highlighted above, political science is a discipline that can be perceived as being by nature concerned with ‘real life politics’. Moreover, the image of a ‘science for democracy’ has traditionally strong roots in the self-understanding of the discipline. Thus, it could be a more-likely case of boundary-crossing compared to some more ‘isolationist’ disciplines. Follow-up research could replicate the survey in a different disciplinary setting. Second, our survey does not include data regarding the question whether there is a match between the topic of the respondents’ recent publications and the themes on which they provide policy advice. While our results regarding the link between publishing regularly and providing advice are encouraging regarding the knowledge base for the advice, there is a need for more studies that explicitly link the research focus of academics and the topics for which they provide advice. Third, while we probe into the relationship between publication activities and cross-boundary work, our study does not allow to explain the relationship or potential trade-offs with teaching activities as another core dimension of academic work.

Finally, our study focuses solely on the supply-side of policy advice. However, an important aspect is also the question in how far the demand-side, for example politicians or civil servants, actively open the border for academics to travel to the policy world. Some existing research (e.g. using surveys, or focus groups) has shed light on this question and which selection mechanisms are at play at the border when deciding which academics will be invited in (see, e.g., Löfgren & Bickerton, 2021; Newman et al., 2016). Future research would profit from integrating both perspectives. For instance, our study focused on the question how active political scientists are in their boundary-crossing work, but it is only in connection with the demand-side that we can understand how effective they are in this work as high frequency and high effectiveness do not have to go hand-in-hand. This links to the nature of the borders, and the travelling activities between the worlds – all of which deserve further conceptualization. For this study, we distinguished the stay at home, the visitor, and the commuter, as types that were also covered by our survey of political scientists working at the moment at a university. Yet for future conceptualization, we could also imagine additional types of scientists who have (temporarily) left academia, thereby extending the continuum to expats, who might keep a return option into academia, or even émigrés, who have moved to the policy world for good.