Contemporary global research cultures: results from a global survey about research conditions

Questionnaire

The questionnaire includes a set of questions covering the key aspects of professional environment discussed above.

Share of time spent on research and teaching

Share of time spent on research and teaching were measured by question “In the last twelve months, what percentage of your time at work did you spend on the following duties? All responses need to add up to 100: Research (carrying out or overseeing research, supervising (doctoral) students, finding & reading papers), Applying for funding (writing/co-writing grants), Publishing (writing & co-writing, peer-review, data storage, sharing), Teaching, Administrative duties, Other.

Satisfaction with the funding possibilities and research facilities and equipment

Satisfaction with the funding possibilities and research facilities and equipment for research was measured by three questions: “In the last 12 months, how satisfied have you been with the available research facilities and equipment for your research work?”; “In the last 12 months, how satisfied have you been with the funding for research projects available at your institution (excluding third−party funding)?”, “In the last 12 months, how satisfied have you been with the available third−party funding opportunities?”. Respondents used 5-point Likert scale to answer these questions: (1) Very dissatisfied, (2) Dissatisfied, (3) Neutral, (4) Satisfied, (5) Very satisfied.

Perceived academic freedom

We used the item battery of the Academic Freedom Index (AFI) (Spannagel & Kinzelbach, 2023) in order to operationalize the concept of academic freedom, which has two dimensions. The first question refers to freedom for topical choice in research and teaching “To what extent are you free to develop and pursue your own research and teaching agendas without interference by non-academic actors”. Note that non-interference does not exclusively relate to the state as one of the most relevant actors, but may also include commercial actors, or religious groups, such as the church. The second question refers to freedom for getting funding for research “To what extent are you free to pursue your personal research interests with the funding you have available?”. Respondents used 5-point Likert scale to answer these questions: (1) Completely restricted, (2) Severely restricted, (3) Moderately restricted, (4) Mostly free, (5) Fully free.

Structure of the academic labor market in terms of representation of different roles and types of contracts

In order to characterize the structure of the academic labor market, questions were asked about the current primary role/position (e.g. full professor, associate professor, postdoctoral fellow), employment status (e.g. employed full-time on a permanent contract, employed full-time on a temporary contract), and the next career milestone (e.g. tenure, a promotion, leading own research group or lab, publication in a highly reputable journal).

Importance of the international peer community

The importance of the international peer community was measured by the question “How international or local is the professional peer community that you want to advance your career in?” Respondents used 5-point Likert scale to answer this question: (1) Exclusively national/local, (2) More national/local, (3) Balanced, (4) Somewhat international, (5) Highly international.

Perceived ability to contribute to science

Perceived ability to contribute to science and also to other aspects of work and family were measured by three questions: “Looking back over the last 12 months, do you feel like you spent too much, too little, or the right amount of time on the following areas?” Areas included “My ability to contribute to science”, “My private/family life”, and “Other aspects of my work, as administrative duties or teaching”. Respondents used 3-point Likert scale to answer these questions: (1) Too little, (2) About right, (3) Too much.

Data collection

The State of the Research Community survey results from a cooperation between the German Centre for Higher Education Research and Science Studies (DZHW) and ResearchGate and was conducted from April to July 2022. The survey is based on a jointly developed questionnaire with 51 questions on the topics career prospects, current research environment, participation in scientific discourse, work satisfaction and demographic covariates. ResearchGate implemented the survey via the third-party software SurveyMonkey. Participation was facilitated via a single URL to the online questionnaire for all participants to safeguard the anonymity of the answers. This link was shared in a post on the ResearchGate blog¹ and the post was prominently mentioned in a simultaneous press release by DZHW². The openly published link allowed also scientists without a ResearchGate account to participate. In addition, a subset of ResearchGate users were made aware of the survey via a pop-up window while being logged into their ResearchGate account. This special access to respondents allowed ResearchGate to manage the quantity of respondents. Upon reaching the goal of 10 k respondents the link to the survey form was deactivated. ResearchGate checked the open text fields on the potential presence of any personal information before passing the data on to DZHW. An initial analysis on the demographic characteristics of the respondents including the calculation of the completion rate of 60% has been published on the ResearchGate blog³.

Additional data

The survey data were enriched with other country-level data. First, in order to establish an economic perspective on research conditions, the data was enriched with income data from the World Bank. Based on GNI per capita, countries were divided into four groups: low-income economies, lower-middle-income economies, upper-middle-income economies, high-income economies. Second, the data was also enriched with the data on the political regime. Based on the data from the V-Dem dataset (version 13, Coppedge et al., 2023) from the V-Dem Institute at the University of Gothenburg, the countries were divided into four regimes (v2x_regime): Closed autocracy, Electoral autocracy, Electoral democracy, Liberal democracy.

Analytical strategy

The analytical strategy was as follows. The dependent variables were numeric answers to the questions about main aspects of their professional environment. The main independent variables – regime and income – were transformed into n – 1 dummy variables (0 or 1). This allows the use of them in regression models. If the coefficient on the regime or income dummy is different from zero, there are differences between two regimes or income groups. Two control variables are also included in the models: gender (male, female, other), position (Full professor, Assistant/Associate Professor, PhD candidate, Postdoctoral fellow, other). The survey data are nested: individual respondents are nested within countries. This means that individual responses are not independent of each other. The linear mixed effects model was used to account for this non-independence. The linear effects mixed model (estimated using REML and the nloptwrap optimizer) was fitted by lme4 (Bates et al., 2015). Each country can have its own interception in the model.

Broad participation

Over 8,000 researchers from different disciplines and at different stages of their careers, working in countries all over the world, took part in the survey. However, only 5,597 of them indicated the country in which they were working at the time of the survey. The survey reached the broad academic community. An analysis of the distribution of participants by country shows that researchers from the Global South are more strongly represented among the respondents of the survey, compared to bibliometric retrieved from Web of Science (WoS) database in 2021. The rankings of countries based on the share of participants and based on the share of articles and reviews in Web of Science (WoS) are different (see Figure 1). Data about the location of the respondents shows the broad coverage of regions in the survey. However, the data is also biased, as respondents from Western Europe make a larger share than actual share of world population. This may also be related to regional and cultural proximity of ResearchGate to Europeans, since ResearchGate is located in Western Europe. In addition, the proportion of Chinese researchers is much smaller than China’s share on world publications as represented by WoS. Nevertheless, the participation of respondents, as regards the physical location of the respondents, is much more diverse than science-related surveys on other subjects which often focus on very few countries outside the Global West.

Figure 1. Top countries by number of survey participants (left panel) and by number of articles and reviews in WoS in 2021 (right panel).

Figure 2 shows the distribution of the participants by income and political regime groups. Participants are distributed by groups relatively equal, except low-income group with that are affiliated only 149 participants. To make income groups more comparable we merged low-income and lower-middle-income groups into one group for further analysis. Distribution of respondents by groups of countries with different income levels and political regimes are shown in Appendix (Figures A1 and A2).

Figure 2. XXX.

Differences between country groups in the share of time spent on research

Research is the most common type of duties in the sample. On average, participants spent 36% of their time on research. Teaching follows research at around 22%. However, there are significant differences between groups of countries in the proportion of time spent on research and teaching (see results of regression analysis in Table A1 in Appendix). The higher income, the higher share of time spent on research and lower share of time spent on teaching (see Figure 3). Participants from low- and lower-middle-income countries report the lowest proportion of time spent on research (34%, 95% CI [31%; 36%]) and the highest proportion of time spent on teaching (24%, 95% CI [22%; 27%]). Participants from high-income countries report the highest proportion of time spent on research (38%, 95% CI [36%; 40%]) and the lowest proportion of time spent on teaching (19%, 95% CI [17%; 21%]). Higher teaching duties can be considered as a source of inequality in the resources available for research. These time resources appear to be a product of economic scarcity and economic considerations regarding the priorities of scientific personnel. Consequently, this inequality in resources, including time resources, may lead to the differences in research output and impact that are consistently reported in studies based on bibliometric databases.

Figure 3. Estimated marginal means and 95% confidence intervals for time spent on research and teaching in countries with different income level.

Differences between groups of countries in satisfaction of the funding and research facilities

The technical and financial resources provided for research are among the most relevant conditions. Particularly in several of the most competitive subjects such as bioengineering, material science or semiconductors, there is a need for particularly expensive technical infrastructures, such as laboratory equipment, machinery, or clean rooms. These technical infrastructures not only require financial resources, but also availability of components, some of which are issued or even sanctioned. But technical infrastructures also require trained personnel who can maintain, curate, and use the infrastructure and machines efficiently. Against this background, the differences in technical infrastructure are particularly striking. On average, respondents from low-income countries have less access to technical facilities and should therefore report less satisfaction with technical facilities. Our results show the differences in satisfaction with funding and research facilities and equipment between countries groups (see results of regression analysis in Table A2 in Appendix). Participants from high-income countries report significantly higher level of satisfaction with the available research facilities and equipment (3.49, 95% CI [3.37; 3.60]), the funding for research projects available at the institution (2.86, 95% CI [2.72; 2.99]), and the available third-party funding opportunities (2.91, 95% CI [2.79; 3.02]) compared to researchers from upper-middle-income (3.17, 95% CI [3.03; 3.31], 2.60, 95% CI [2.42; 2.78], and 2.68, 95% CI [2.54; 2.82], respectively) and low- + lower-middle-income countries (3.05, 95% CI [2.92; 3.18], 2.49, 95% CI [2.33; 2.65], and 2.57, 95% CI [2.43; 2.70], respectively). The difference in these measures between upper-middle-income and low- + lower-middle-income countries are not so pronounced. However, in general, the lower income, the lower satisfaction (see Figure 4). The differences here provide rather clear indication of different material conditions for the conduct of research within different global regions. Another visible pattern is that in general researchers in all countries are less satisfied with funding (means are vary between 2.49, 95% CI [2.33; 2.65]) and 2.91, 95% CI [2.79; 3.02]) than with research facilities and equipment (means are vary between 3.05, 95% CI [2.92; 3.18] and 3.49, 95% CI [3.37; 3.60]). These results suggest that researchers in all countries suffer from nonperfect funding systems. However, researchers from high-income countries on average suffer less compared to researchers from the rest of the world.

Figure 4. Estimated marginal means and 95% confidence intervals for satisfaction of the funding and research facilities and equipment in countries with different income level.

Differences between country groups in perceived academic freedom

By enriching our survey with additional data on political regimes, we were also able to compare the perceived freedom to pursue research agendas in relation to the different political regimes. The results show the differences in perceived academic freedom between groups of countries (see results of regression analysis in Table A3 in Appendix). However, the differences are less pronounced than might be expected, especially when compared with the data from (Kinzelbach et al., 2022; Spannagel & Kinzelbach, 2023). Only respondents from countries in the liberal democracy group report a slightly higher level of perceived freedom both to develop and pursue their own research and teaching agendas without interference from non-academic actors (3.85, 95% CI [3.74; 3.95]) and to pursue personal research interests with the available funding (3.49, 95% CI [3.37; 3.61]) compared to respondents from the electoral democracy (3.67, 95% CI [3.56; 3.79] and 3.35, 95% CI [3.22; 3.48], respectively) and electoral autocracy (3.71, 95% CI [3.59; 3.82] and 3.35, 95% CI [3.22; 3.48], respectively) groups (see Figure 5). However, respondents from countries with a closed autocratic regime are not statistically different from the other groups. The means in this group have wider confidence intervals, which may be due to a higher dispersion of the responses in this group. This dispersion may be an indicator of the greater inequality of working conditions in closed autocracies, where one group of researchers is given far more resources and freedom than others. Another possible explanation for the lack of substantial difference between democratic and autocratic regimes may be the differences in the respondents’ internal definition of academic freedom, and also in the perceived standards that researchers in different countries and academic systems may have. Future research could more systematically ask for specific conceptions and frameworks of academic freedom research.

Figure 5. Estimated marginal means and 95% confidence intervals for perceived academic freedom in countries with different political regimes.

Differences in perception of wider working conditions

Differences between the countries become more profound when taking other aspects of working conditions into account. As regards the representation of different academic groups in the survey, it is striking that the share of postdoctoral fellows is significantly lower among respondents from the Global South (see Figure 6). If the income perspective is taken, about 15% of respondents from high-income countries describe themselves as postdoctoral fellows, while only 5% fall in that category in upper-middle-income countries or 3% in low- or lower-middle-income countries. One possible explanation is that postdocs in western high-income countries feel stronger inclined to be present on academic social networks in order to disseminate their work or to present themselves. The most common categories among the respondents from the Global South countries are Assistant/Associate professors, PhD candidates, and Lecturers. Yet, the strong presence of postdocs among the respondents in the West seems to generate a range of structural differences in the survey as we will see.

Figure 6. Proportions of different roles/positions among respondents from countries with different income level.

We also find that a larger share of the researchers in the Global North is on fixed contracts (24% in high-income countries vs 16% and 11% in other groups), whereas the share of respondents on tenured or long-term positions is significantly lower compared to respondents from other world regions (41% in high-income countries vs 53% and 55% in other groups) (see Figure 7).

Figure 7. Proportions of different employment statuses among respondents from countries with different income level.

Analysis of the next career milestones shows that respondents from all country groups have similar plans (see Figure 8). The most common milestones include publication in a highly reputable journal, finishing degree, and promotion. An interesting aspect is as to whether national or international networks or resources are considered important to achieve this goal or to reach the next career step. Here, we find that respondents from all country groups report a high importance of the international professional peer community (see Figure 9). The data seem to suggest that relying solely on national resources and networks only does not seem to be a viable option. There are also differences in importance of international community between regions and between the different regimes (see Table A4 in Appendix). International community is more important not only for researchers from high-income liberal democracies (4.02, 95% CI [3.91; 4.12]) but also for researchers from closed autocracies (4.11, 95% CI [3.95; 4.28]). The higher importance of the international community for researchers from closed autocracies may be due to crucial restrictions in national academic system.

Figure 8. Proportions of next career milestones among respondents from countries with different income level.

Figure 9. Estimated marginal means and 95% confidence intervals for importance of the international peer community in countries with different income and political regimes.

Researchers from the Global West appear to be under strong pressure, with apparently high expectations regarding the scientific relevance of the use of their research. When asked to assess their contributions to science, respondents from less economically developed countries report these observations to a lesser extent (see Table A5 in the Appendix). The mean for high-income group is 1.79, 95% CI [1.72; 1.85], the mean for the upper-middle-income group is 1.88, 95% CI [1.79; 1.96], and the mean for the low- and lower-middle-income group is 1.98, 95% CI [1.90; 2.06] (see Figure 10). Furthermore, the perception of low contribution in Western academia is not related to the effort put into academic work. There are more respondents from the Global West who report that their efforts to contribute to science would be at the expense of their private life (means are 1.58, 95% CI [1.52; 1.64] vs 1.70, 95% CI [1.63; 1.77] and 1.76, 95% CI [1.69; 1.82]). Respondents from the high-income group report that they contribute too much to other work-related tasks apart from science, such as administration. Whereas respondent from the upper-middle-income and low- and lower-middle-income groups have similar perceptions of their contribution to other work-related tasks. The mean for the high-income group is 2.21, 95% CI [2.14; 2.27], the mean for the upper-middle-income group is 2.09, 95% CI [2.01; 2.16], and the mean for the low- and lower-middle-income group is 2.03, 95% CI [1.96; 2.11]. Summarizing, respondents from the Global South have more balanced responses, with individual expectations and practice more closely aligned than respondents from the Global North. The responses show that the competitive culture and fixed-term contracts in the Global North have a more fundamental impact on perceptions of freedom to pursue research on a long-term basis.

Figure 10. Estimated marginal means and 95% confidence intervals for satisfaction with contribution to science and private life.