Cooperation structures
Cooperation partners
Research collaborations with other institutions
In order to gain a better insight into the existing research collaborations, we asked the scientists of the four BUA institutions more specifically about their cooperation partners (see Figure 5). Of those who collaborated, the vast majority said they did so with other scientists from their own research unit or team (86%), or from their own university (71%). The respondents also reported considerable collaboration with other universities in Germany and abroad: 59% and 55% respectively, (see Figure 5). This means that the question of regional location plays a lesser role in collaborations with partners at other universities, since a larger proportion of respondents tended to collaborate with supraregional or international partners than with partners from other universities in Berlin (41%).
Collaborations with non-university research institutions were also considerable (47%). However, unlike university collaborations, collaborations with non-university research institutions had a much stronger regional focus. Just under a third of respondents undertook research collaborations with non-university research institutions in the Berlin research area (see Figure 5).
Somewhat less common, at 27%, were research collaborations with companies, most of them in Germany. Just under 14% of all BUA respondents collaborated with companies from the Berlin research area (see Figure 5). Proximity to research institutions is also important for firms and their innovative capacity (Rammer et al. 2019). Here, the information from the Berlin Science Survey complements the data from the Berlin Innovation Panel, which surveys companies in the Berlin Metropolitan Area (Blind 2016).
The least frequent research collaborations were with universities of applied sciences (UAS) (6%). More frequent were collaborations with civil society actors and organizations, such as foundations, associations, non-governmental organizations (NGOs), and citizens (15%) (see Figure 5).
Figure 5 Cooperation partners of BUA employees
Compared to university collaborations, the structure of scientists working at non-university research institutions – who comprise the so-called “Berlin Research 50” (BR50) group – was somewhat different (see Figure 6). As can be seen, collaborative relationships with non-university research institutions were significantly higher among respondents of the BR50: 85.2% compared to 47% among university employees, although collaborative relationships within the own institutions of the BR50 were also included in this sum. The proportions are comparable for collaborations within the own team, with companies, civil society, and universities of applied sciences. Even in the case of collaboration with universities, the proportions are similarly high: respondents from BR50 collaborated very frequently with partners at universities (almost 84%), and more than half of the respondents from BR50 even maintain collaborative relationships with universities in Berlin.
Figure 6 Cooperation partners of BR50 employees
A comparison of the four BUA institutions also revealed clear differences in collaborative relationships. These can be attributed in part to the different subject structures of the institutions, but only partially, as the data from the BSS pilot study did not allow for representative comparisons of the four BUA institutions (see Lüdtke and Ambrasat, 2022b).
Figure 7 shows that collaborations with universities as a whole (including the BUA's own institution) were roughly equally strong in all BUA institutions and were at a very high level overall. However, there were differences regarding cooperation with non-university research institutions. Employees at Humboldt-Universität zu Berlin (HU) collaborated most with non-university research institutions (52%), while the lowest percentage was found among employees at the Charité Universitätsmedizin (40%). Employees of the Technische Universität Berlin (TU) and the Charité were most involved in research collaborations with companies (39% and 37%, respectively), while HU employees were hardly involved at all in collaborations with this group (15%, see Figure 7). There were also major differences in research collaborations with civil society. While these collaborations played virtually no role for Charité employees (7%), some 21.5% of HU employees reported such collaborations with these organizations. Collaborations with universities of applied sciences were rare among all institutions but were most likely to be found at the TU, with 11% (see Figure 7).
Figure 7 Cooperation partners of BUA employees, by institutions
If we look at the distribution of partners in research collaborations by status group, we see that professors cooperate the most and predocs the least (see Figure 8). Furthermore, the structure of the collaborations was comparable across all status groups: the most frequent collaborations were with universities, the least with UAS.
A different picture emerges when comparing the subject groups (see Figure 9). Although collaboration with universities was most frequent across all subject groups, the collaboration structures differed. Scientists in the humanities and social sciences collaborated more with external partners from civil society than with companies. Collaboration with civil society organizations was particularly frequent among social scientists, at about 29%. Engineering scientists, on the other hand, collaborated more with companies and even more frequently than with non-university research institutions. The level of cooperation with UASs was also higher among engineers than in the other subject groups (see Figure 9). What is striking, though not surprising, is the high proportion of corporate collaborations (59%).
Figure 8 External research collaborations, by status groups
Figure 9 External research collaborations, by subject groups
Research collaborations by region
Since it is the aim of the Berlin University Alliance to strengthen Berlin in particular as a research area and to increase collaborations here, it is also interesting to look at whether external research collaborations differ by region. We have therefore differentiated external research collaborations according to whether they are with partners from Berlin, Germany (excluding Berlin) or abroad.
The comparison of status groups reveals diverging cooperation profiles as well as different levels of cooperation. Professors most frequently collaborated with external partners from abroad (84%), while scientists below professorship level were most likely to collaborate with partners in Germany (see Figure 10). Among non-doctoral scientists, research collaborations abroad were the least common, at just under 40% (see Figure 10). Less than one in two of this status group maintained collaborative relationships with researchers in Berlin.
Figure 10 External collaborations with partners from different regions, by status groups
A comparison of subjects (see Figure 11) revealed that the humanities participated in international collaborations the most(72.2%). In the social sciences, all regions were represented equally, and in the engineering sciences, national cooperative relationships predominated. In the life sciences, natural sciences and engineering, collaborative relationships within the Berlin research area occurred somewhat more frequently than in the humanities and social sciences.
Figure 11 External collaborations with partners from different regions, by subject groups
Intradisciplinarity, interdisciplinarity, and transdisciplinarity
Another dimension of collaboration concerns the crossing of disciplinary boundaries. Science policy often calls for interdisciplinarity and, increasingly, transdisciplinarity. While interdisciplinarity refers to cooperation with partners from other disciplines within science, transdisciplinary research is understood as collaboration with partners outside science. As expected, the majority of the scientists surveyed in the Berlin Research Area (80.7%) worked regularly, i.e., "often", "very often" or "always", with colleagues from their own disciplines (see Figure 12). Interdisciplinary work was a common practice for 49.7% of the respondents. Transdisciplinary research, on the other hand, was only reported as a regular practice by 18.6% of respondents. For more than one-third of the respondents, however, transdisciplinarity never occurred in their daily work (see Figure 12). These figures do not necessarily indicate a need to increase transdisciplinary research. The content of some research may mean that collaboration is only appropriate disciplinarily. On the other hand, there may be individual research contexts that would benefit from transdisciplinary research. These groups cannot be differentiated within the available data. In order to be able to assess the potential for an increase in transdisciplinary work, it is necessary to look at the extent to which the respective concrete research contexts would benefit from both transdisciplinarity and interdisciplinarity. This requires a focus on these individual contexts.
Figure 12 Intra-, inter-, and transdisciplinarity
No remarkable differences emerged when comparing forms of collaboration according to status groups. Only the generally higher level of collaboration among professors was notable (see Figure 13).
There were, however, some differences between the subject groups (see Figure 14). At 31%, transdisciplinarity was particularly pronounced in the engineering sciences, which can be attributed to the high number of corporate collaborations (see Fig. 9). By contrast, transdisciplinarity played a rather subordinate role in the life and natural sciences disciplines, at less than 13%.
When comparing the organizational forms, namely the non-university research institutions of the BR50 and the universities of the Berlin University Alliance, there was only a small difference in interdisciplinary research, which occurred somewhat more frequently among the respondents of the BR50, (54.2% compared to 48.6%, respectively; see Fig. 15).
Figure 13 Regular intra-, inter-, and transdisciplinary research collaborations by status groups
Figure 14 Regular intra-, inter-, and transdisciplinary research collaborations by subject groups
Figure 15 Regular intra-, inter-, and transdisciplinary research collaborations by organizational forms
Initiation and stability of collaborations
To get an impression of existing collaborative relationships, we also looked at the initiation of research collaborations and the stability of collaborative relationships.
When asked about how collaborative relationships usually come about, 29% of the respondents indicated that their collaborations were mostly or more likely to come about on their own initiative. On the other hand, around 17% stated that they mainly or rather received requests for collaborations. The majority (54%), however, indicated that their research collaborations occurred on their own initiative and from requests equally (not shown). Furthermore, we asked whether collaborative relationships remained stable, i.e., whether they remained with the same partners over several projects, or whether the partners tended to change on a project-by-project basis. Here we found that about one-third of the respondents enjoyed rather stable collaborative relationships (not shown), while project-related changes dominated for 69% of the respondents. More stable collaborative relationships refer to collaborations in longer-term research programs and the use of larger infrastructures.
If we look at the distribution according to status groups, we see that the relationship between personal initiative and requests becomes increasingly balanced as a research career progresses (see Figure 16). At the level of professors, 68% stated that collaborations occurred equally as a result of their own initiative and requests from others. On the other hand, there is a strong imbalance, particularly among predocs: 21% of respondents stated that collaborations were the result of external requests, while over 31% of respondents stated that collaborations were mainly the result of their own initiative (see Figure 16). By contrast, there are hardly any differences between the status groups with regard to the stability of collaborative relationships (see Figure 17).
Figure 16 Initiation of collaborative relationships by status groups
Figure 17 Stability of collaborative relationships by status groups
When comparing subject groups, it was the engineering and natural sciences where collaborative relationships occurred somewhat more frequently through external requests, at around 23% in each case (see Figure 18). These figures were remarkably lower in the humanities (6.5%) and in the life sciences (14%) (see Figure 18).
Changing project partners predominated in all subject groups but were most pronounced in the humanities at 82 %. By contrast, the proportion of stable project relationships was somewhat higher in engineering and the natural sciences (34% each) and, above all, in the life sciences (36%, see Figure 19). By comparison, only 18.3% of collaborative relationships in the humanities exhibited this stability (see Figure 19).
These differences in the engineering sciences can again be explained by corporate collaborations. In addition, it can be assumed that infrastructure-based research areas are in greater demand.
Figure 18 Initiation of collaborative relationships by subject groups
Figure 19 Stability of collaborative relationships by subject groups
A comparison between the gender of researchers shows no differences in the stability of collaborative relationships (see Figure 21). When it comes to initiating collaborations, however, women were approached somewhat less frequently than men (33.5% vs. 25.8%) but more often acted on their own initiative to instigate collaborations (see Figure 20).
Figure 20 Initiation of collaborative relationships by gender
Figure 21 Stability of collaborative relationships by gender