THE BIBLIOMETRIC LENS<\/h1>\n<\/div><\/div><\/div><\/div><\/div>Table of contents<\/strong>
\nData and identi\u00adfy\u00ading mobility<\/a>
\nHow inter\u00adna\u00adtion\u00adal is the German research system?<\/a>
\nDo mobile researchers perform differently?<\/a>
\nWhat do these results mean?<\/a><\/p>\n<\/div><\/div><\/div>Mobil\u00adi\u00adty patterns and research perfor\u00admance of researchers connect\u00aded to Germany<\/h2>\n
AUTHOR: Andrey Lovakov
\nPUBLISHED: August 11, 2026
\nAFFILIATION: German Centre for Higher Educa\u00adtion Research and Science Studies (DZHW)<\/p>\n
Almost half of all researchers with ties to Germany have worked abroad at some point in their careers, and their research perfor\u00admance tends to be higher than that of their peers who remained in Germany. This blog post uses large-scale biblio\u00admet\u00adric data to examine mobil\u00adi\u00adty patterns and perfor\u00admance differ\u00adences across five mobil\u00adi\u00adty groups with impor\u00adtant impli\u00adca\u00adtions for science policy and the openness of the German research system.<\/strong><\/p>\nInter\u00adna\u00adtion\u00adal mobil\u00adi\u00adty has become a defin\u00ading feature of modern scien\u00adtif\u00adic careers. Researchers often move between countries for doctor\u00adal train\u00ading, postdoc\u00adtor\u00adal positions, and facul\u00adty jobs, and inter\u00adna\u00adtion\u00adal collab\u00ado\u00adra\u00adtion is now standard practice in many fields. As a result, nation\u00adal research systems are becom\u00ading more integrat\u00aded into the global flows of scien\u00adtif\u00adic talent. Germany is no excep\u00adtion. Univer\u00adsi\u00adties and research insti\u00adtutes attract researchers from abroad, while many schol\u00adars trained in Germany spend part of their careers in other countries. At the same time, migra\u00adtion has become one of the most debat\u00aded topics in German public discourse. Public discus\u00adsions often focus on labor markets, demograph\u00adic change, and the integra\u00adtion of migrants. Highly skilled migra\u00adtion, howev\u00ader, receives less atten\u00adtion outside of special\u00adized policy debates, even though it plays a crucial role in the function\u00ading of research systems. Against this background, it is useful to better under\u00adstand how inter\u00adna\u00adtion\u00adal mobil\u00adi\u00adty shapes the German science system. Two questions are partic\u00adu\u00adlar\u00adly relevant: First, how inter\u00adna\u00adtion\u00adal\u00adly mobile are researchers in Germany? Second, do researchers with differ\u00adent mobil\u00adi\u00adty patterns differ in their research perfor\u00admance? Using large-scale biblio\u00admet\u00adric data, this blog post examines mobil\u00adi\u00adty trajec\u00adto\u00adries and research perfor\u00admance for more than 300,000 researchers connect\u00aded to Germany.<\/p>\n<\/div>
Data and identi\u00adfy\u00ading mobility<\/h2>\n
The in-house version of the Scopus database provid\u00aded by the German Compe\u00adtence Network for Biblio\u00admet\u00adrics<\/a> (snapshot as of 10.2025) was used (Schmidt et al. 2025<\/a>)<\/span>. Elsevi\u00ader disam\u00adbiguates authors in Scopus by group\u00ading all publi\u00adca\u00adtions into profiles with a unique author IDs. The Scopus author ID was used to identi\u00adfy all publi\u00adca\u00adtions for each author. All publi\u00adca\u00adtions and affil\u00adi\u00ada\u00adtions linked to a given author ID were treat\u00aded as belong\u00ading to the same researcher. The initial sample includ\u00aded all authors who had published at least one document with a German insti\u00adtu\u00adtion\u00adal affil\u00adi\u00ada\u00adtion between 2000 and 2025.<\/p>\nTo study mobil\u00adi\u00adty, we recon\u00adstruct\u00aded individ\u00adual affil\u00adi\u00ada\u00adtion trajec\u00adto\u00adries at the country level. For each researcher and publi\u00adca\u00adtion year, the modal country of affil\u00adi\u00ada\u00adtion was identi\u00adfied, meaning the country that appeared most frequent\u00adly across that researcher\u2019s publi\u00adca\u00adtions in that year. When two or more countries were equal\u00adly frequent\u00adly in a given year, they were compared to the modal country from the previ\u00adous year. If one of the tied countries matched the previ\u00adous year\u2019s modal country, that country was assigned. If none matched, one country was select\u00aded at random. To account for short gaps in publi\u00adca\u00adtion activ\u00adi\u00adty and reduce the artifi\u00adcial fragmen\u00adta\u00adtion of mobil\u00adi\u00adty sequences, publi\u00adca\u00adtion years were extend\u00aded backward by two years. This proce\u00addure produced a longi\u00adtu\u00addi\u00adnal country-level trajec\u00adto\u00adry for each researcher<\/p>\n
Based on these trajec\u00adto\u00adries, the researchers were classi\u00adfied into five mutual\u00adly exclu\u00adsive mobil\u00adi\u00adty types:<\/p>\n
\n- Locals<\/strong>: The modal country of affil\u00adi\u00ada\u00adtion was Germany through\u00adout the entire observed period.<\/li>\n
- Arrivals<\/strong>: The initial modal country is not Germany and the final modal country is Germany.<\/li>\n
- Leavers<\/strong>: The initial modal country is Germany and the final modal country is not Germany.<\/li>\n
- Returnees<\/strong>: The initial modal country is Germany, subse\u00adquent affil\u00adi\u00ada\u00adtion are abroad, and the final modal country is Germany.<\/li>\n
- Visitors<\/strong>: The initial modal country is not Germany, tempo\u00adrary affil\u00adi\u00ada\u00adtion is Germany, and the final modal country is not Germany.<\/li>\n<\/ul>\n
To ensure reliable identi\u00adfi\u00adca\u00adtion of mobil\u00adi\u00adty and stable produc\u00adtiv\u00adi\u00adty indica\u00adtors, the analyt\u00adi\u00adcal sample was restrict\u00aded to researchers with at least three publi\u00adca\u00adtions and at least two years between their first and last publi\u00adca\u00adtion year. These researchers should have been active in recent years, which is defined as a first publi\u00adca\u00adtion year \u2264 2022 and a last publi\u00adca\u00adtion year \u2265 2023. After apply\u00ading these crite\u00adria, the final sample includ\u00aded 317,314 researchers.<\/p>\n<\/div>
How inter\u00adna\u00adtion\u00adal is the German research system?<\/h2>\n
The results show that most researchers connect\u00aded to Germany are Locals (Figure 1<\/em>). About 57% of the sample remained affil\u00adi\u00adat\u00aded with Germany through\u00adout the observed period. At the same time, inter\u00adna\u00adtion\u00adal mobil\u00adi\u00adty is far from margin\u00adal. Rough\u00adly 43% of researchers fall into one of the inter\u00adna\u00adtion\u00adal\u00adly mobile categories. Among these, Arrivals and Leavers each account for about 13% of the sample. Visitors repre\u00adsent 12%, while Returnees make up about 5%. These figures suggest that the German science system is deeply embed\u00added in inter\u00adna\u00adtion\u00adal mobil\u00adi\u00adty flows. Researchers both enter and leave the system, and some return after spend\u00ading time abroad.<\/p>\n<\/div>![](\"https:\/\/bibliometrie.info\/wp-content\/uploads\/2026\/09\/mobility_group.jpg\" "\\"mobility_group\\"")
<\/span> <\/h2>
Figure 1: Distri\u00adb\u00adu\u00adtion of German researchers by mobil\u00adi\u00adty group. <\/p><\/div><\/div><\/div>
Howev\u00ader, the degree of inter\u00adna\u00adtion\u00adal\u00adiza\u00adtion varies great\u00adly across research fields (Figure 2<\/em>). Highly inter\u00adna\u00adtion\u00adal fields include mathe\u00admat\u00adics, physics and astron\u00ado\u00admy, arts and human\u00adi\u00adties, earth and plane\u00adtary sciences, and econom\u00adics, econo\u00admet\u00adrics and finance. In these disci\u00adplines, less than half of researchers remain locals. By contrast, fields such as health profes\u00adsions, dentistry, nursing, medicine, and chemi\u00adcal engineer\u00ading show much higher shares of locals, exceed\u00ading 70%. These disci\u00adplines appear more strong\u00adly embed\u00added in nation\u00adal insti\u00adtu\u00adtion\u00adal and profes\u00adsion\u00adal structures.<\/p>\n<\/div>![](\"https:\/\/bibliometrie.info\/wp-content\/uploads\/2026\/09\/research_fields.png\" "\\"research_fields\\"")
<\/span>research_fields<\/h2>
Figure 2: Distri\u00adb\u00adu\u00adtion of German researchers by mobil\u00adi\u00adty group across research fields. <\/p><\/div><\/div><\/div>
Do mobile researchers perform differently?<\/h2>\n
The analy\u00adsis also examined whether researchers with differ\u00adent mobil\u00adi\u00adty patterns differ in their research perfor\u00admance. Research perfor\u00admance was measured using sever\u00adal indica\u00adtors, including:<\/p>\n
\n- number of publi\u00adca\u00adtions: Articles, Reviews, and Books;<\/li>\n
- number of publi\u00adca\u00adtions in high-prestige journals: Level 2 of the Norwe\u00adgian Regis\u00adter for Scien\u00adtif\u00adic Journals<\/a>;<\/li>\n
- mean field-normal\u00adized citation rates: the average number of citations per paper normal\u00adized by subject catego\u00adry and publi\u00adca\u00adtion year, calcu\u00adlat\u00aded using three citation windows: 3\u2011year window, 5\u2011year window, and full citation window;<\/li>\n
- number of top 10% most cited publi\u00adca\u00adtions: number of publi\u00adca\u00adtions ranking among the top 10% most cited papers global\u00adly within their field and publi\u00adca\u00adtion year, comput\u00aded for 3\u2011year, 5\u2011year, and full citation windows.<\/li>\n<\/ul>\n
The indica\u00adtors were calcu\u00adlat\u00aded using both whole and fraction\u00adal count\u00ading methods and with differ\u00adent citation windows. All indica\u00adtors were comput\u00aded using both whole count\u00ading, in which each co-authored publi\u00adca\u00adtion is fully counts for each author, and fraction\u00adal count\u00ading, in which publi\u00adca\u00adtion credit is divid\u00aded by the number of co-authors. To control for the influ\u00adence of large-scale collab\u00ado\u00adra\u00adtions, where individ\u00adual contri\u00adbu\u00adtions are diffi\u00adcult to assess, all indica\u00adtors were calcu\u00adlat\u00aded using two sets of publi\u00adca\u00adtions. One set includ\u00aded all publi\u00adca\u00adtions, while the other includ\u00aded only publi\u00adca\u00adtions with fewer than 20 authors.<\/p>\n
Indica\u00adtor labels follow the struc\u00adture [Metric] [variant] [team filter] ([count\u00ading])<\/strong>. Pubs<\/em> = number of publi\u00adca\u00adtions; Prest.<\/em> = publi\u00adca\u00adtions in prestige journals (Norwe\u00adgian Regis\u00adter, Level 2); Cit.rate<\/em> = mean field-normal\u00adized citation rate; Top10%<\/em> = number of top 10% most cited publi\u00adca\u00adtions. Citation windows: 3y<\/em> = 3\u2011year, 5y<\/em> = 5\u2011year, all<\/em> = full window. <20<\/em> = restrict\u00aded to publi\u00adca\u00adtions with fewer than 20 authors. (W)<\/em> = whole count\u00ading; (F)<\/em> = fraction\u00adal counting.<\/p>\nTo allow meaning\u00adful compar\u00adisons across disci\u00adplines and career stages, all indica\u00adtors were convert\u00aded into percentiles within fields and acade\u00adm\u00adic age groups. A percentile of 90 means that a researcher\u2019s indica\u00adtor value is higher than 90% of other researchers in their field and acade\u00adm\u00adic age group. Next, weight\u00aded mean percentiles were calcu\u00adlat\u00aded for each mobil\u00adi\u00adty group. Final\u00adly, the results across all indica\u00adtors were synthe\u00adsized using a random-effects meta-analysis.<\/p>\n
The results show a consis\u00adtent pattern. Locals have the lowest median perfor\u00admance across all 32 indica\u00adtors, with mean percentiles below 50 (Figure 3<\/em>). This indicates that, on average, researchers in this group are in the lower half of the perfor\u00admance distri\u00adb\u00adu\u00adtion. In contrast, Returnees and Visitors show the highest average perfor\u00admance levels. Their mean percentiles are typical\u00adly above 50, placing them in the upper half of the distri\u00adb\u00adu\u00adtion (Figure 4<\/em>). Arrivals and Leavers fall between these two extremes, with mean percentiles both above and below the median, depend\u00ading on the indica\u00adtor (Figure 5<\/em>). Howev\u00ader, it is impor\u00adtant to note that these differ\u00adences are averages. The distri\u00adb\u00adu\u00adtions overlap signif\u00adi\u00adcant\u00adly (Figure 3<\/em>). Each mobil\u00adi\u00adty group includes researchers with both very high and very low performance.<\/p>\n<\/div>![](\"https:\/\/bibliometrie.info\/wp-content\/uploads\/2026\/09\/performance_indicators.png\" "\\"performance_indicators\\"")
<\/span>performance_indicators<\/h2>
Figure 3: Distri\u00adb\u00adu\u00adtion of perfor\u00admance indica\u00adtors among differ\u00adent mobil\u00adi\u00adty groups.<\/p><\/div><\/div><\/div>
![](\"https:\/\/bibliometrie.info\/wp-content\/uploads\/2026\/09\/indiv_effects.png\" "\\"indiv_effects\\"")
<\/span>indiv_effects<\/h2>
Figure 4: Forest plot with individ\u00adual effects. <\/p><\/div><\/div><\/div>
![](\"https:\/\/bibliometrie.info\/wp-content\/uploads\/2026\/09\/pooled_mean.png\" "\\"pooled_mean\\"")
<\/span>pooled_mean<\/h2>
Figure 5: Pooled mean percentile in mobil\u00adi\u00adty groups. <\/p><\/div><\/div><\/div>
The results above are based on all researchers in the sample. To explore the same figures for individ\u00adual research fields, an inter\u00adac\u00adtive compan\u00adion app is avail\u00adable here<\/a>.<\/em><\/p>\n<\/div>What do these results mean?<\/h2>\n
The results are not entire\u00adly surpris\u00ading. Previ\u00adous studies have already shown a positive relation\u00adship between inter\u00adna\u00adtion\u00adal mobil\u00adi\u00adty and individ\u00adual research perfor\u00admance. This relation\u00adship is likely bidirec\u00adtion\u00adal. On the one hand, mobil\u00adi\u00adty may improve research perfor\u00admance. Working in differ\u00adent countries can provide access to valuable resources, expand profes\u00adsion\u00adal networks, and expose researchers to new research infra\u00adstruc\u00adtures, insti\u00adtu\u00adtion\u00adal environ\u00adments, and funding oppor\u00adtu\u00adni\u00adties. These experi\u00adences may help researchers devel\u00adop new collab\u00ado\u00adra\u00adtions and increase the visibil\u00adi\u00adty of their work. On the other hand, mobil\u00adi\u00adty may also reflect selec\u00adtion process\u00ades. Researchers with strong publi\u00adca\u00adtion records and promis\u00ading career prospects are more likely to receive oppor\u00adtu\u00adni\u00adties to work abroad. In this sense, inter\u00adna\u00adtion\u00adal\u00adly mobile researchers may already belong to the more success\u00adful members of the scien\u00adtif\u00adic workforce before they move.<\/p>\n
It is also impor\u00adtant to keep in mind that the groups being compared here may differ not only in their mobil\u00adi\u00adty patterns, but poten\u00adtial\u00adly also in their compo\u00adsi\u00adtion. Although all indica\u00adtors were normal\u00adized by field and acade\u00adm\u00adic age, the groups may still differ in ways that these adjust\u00adments do not capture. The Locals group is the largest and is likely to be the most hetero\u00adge\u00adneous. For example, it may include researchers in predom\u00adi\u00adnant\u00adly teach\u00ading-orient\u00aded positions, at univer\u00adsi\u00adties of applied sciences, or working part-time. All of these charac\u00adter\u00adis\u00adtics can limit publi\u00adca\u00adtion output without reflect\u00ading lower research abili\u00adty. By contrast, inter\u00adna\u00adtion\u00adal\u00adly mobile researchers may repre\u00adsent a more selec\u00adtive group within the research workforce. These compo\u00adsi\u00adtion\u00adal differ\u00adences should be kept in mind when inter\u00adpret\u00ading the perfor\u00admance gaps between groups.<\/p>\n
From a science policy perspec\u00adtive, the findings never\u00adthe\u00adless highlight the impor\u00adtance of inter\u00adna\u00adtion\u00adal mobil\u00adi\u00adty as a struc\u00adtur\u00adal feature of the German research system. Inter\u00adna\u00adtion\u00adal mobil\u00adi\u00adty is not margin\u00adal but a central compo\u00adnent of how the system operates. Germany is a highly inter\u00adna\u00adtion\u00adal\u00adized research environ\u00adment that both attracts researchers from abroad and sends researchers to other countries. Scien\u00adtif\u00adic careers connect\u00aded to Germany are there\u00adfore often shaped by inter\u00adna\u00adtion\u00adal mobil\u00adi\u00adty. The results suggest that Germany benefits from this global circu\u00adla\u00adtion of researchers. Inter\u00adna\u00adtion\u00adal\u00adly mobile groups tend to show higher average perfor\u00admance levels than the Locals group, though this pattern may partly reflect the compo\u00adsi\u00adtion\u00adal differ\u00adences discussed above. This under\u00adlines the impor\u00adtance of maintain\u00ading openness to inter\u00adna\u00adtion\u00adal researchers. At the same time, the finding that researchers who leave Germany also tend to perform above the overall average may raise questions about how attrac\u00adtive the German research system is for retain\u00ading produc\u00adtive scien\u00adtists. Final\u00adly, the findings highlight the poten\u00adtial impor\u00adtance of return mobil\u00adi\u00adty. Researchers who spent time abroad and later returned to Germany show the highest average perfor\u00admance levels among all mobil\u00adi\u00adty groups. Tempo\u00adrary mobil\u00adi\u00adty may there\u00adfore allow researchers to accumu\u00adlate inter\u00adna\u00adtion\u00adal experi\u00adence, networks, and knowl\u00adedge that they later bring back into the German research system.<\/p>\n<\/div>
Refer\u00adences<\/h3>\n
Schmidt, Marion, Chris\u00adtine Rimmert, Dimity Stephen, Christo\u00adpher Lenke, Paul Donner, Simone G\u00e4rtner, Niels Taubert, Thomas Bausen\u00adwein, and Stephan Stahlschmidt. 2025. \u201cThe Data Infra\u00adstruc\u00adture of the German Kompe\u00adten\u00adznet\u00adzw\u00aderk Bibliome\u00adtrie: An Enabling Inter\u00adme\u00addi\u00adary Between Raw Data and Analy\u00adsis.\u201d Quanti\u00adta\u00adtive Science Studies<\/em> 6: 1129\u201346. https:\/\/doi.org\/10.1162\/qss.a.20<\/a>.<\/p>\n<\/div><\/div><\/div>
Table of contents<\/strong> AUTHOR: Andrey Lovakov Almost half of all researchers with ties to Germany have worked abroad at some point in their careers, and their research perfor\u00admance tends to be higher than that of their peers who remained in Germany. This blog post uses large-scale biblio\u00admet\u00adric data to examine mobil\u00adi\u00adty patterns and perfor\u00admance differ\u00adences across five mobil\u00adi\u00adty groups with impor\u00adtant impli\u00adca\u00adtions for science policy and the openness of the German research system.<\/strong><\/p>\n Inter\u00adna\u00adtion\u00adal mobil\u00adi\u00adty has become a defin\u00ading feature of modern scien\u00adtif\u00adic careers. Researchers often move between countries for doctor\u00adal train\u00ading, postdoc\u00adtor\u00adal positions, and facul\u00adty jobs, and inter\u00adna\u00adtion\u00adal collab\u00ado\u00adra\u00adtion is now standard practice in many fields. As a result, nation\u00adal research systems are becom\u00ading more integrat\u00aded into the global flows of scien\u00adtif\u00adic talent. Germany is no excep\u00adtion. Univer\u00adsi\u00adties and research insti\u00adtutes attract researchers from abroad, while many schol\u00adars trained in Germany spend part of their careers in other countries. At the same time, migra\u00adtion has become one of the most debat\u00aded topics in German public discourse. Public discus\u00adsions often focus on labor markets, demograph\u00adic change, and the integra\u00adtion of migrants. Highly skilled migra\u00adtion, howev\u00ader, receives less atten\u00adtion outside of special\u00adized policy debates, even though it plays a crucial role in the function\u00ading of research systems. Against this background, it is useful to better under\u00adstand how inter\u00adna\u00adtion\u00adal mobil\u00adi\u00adty shapes the German science system. Two questions are partic\u00adu\u00adlar\u00adly relevant: First, how inter\u00adna\u00adtion\u00adal\u00adly mobile are researchers in Germany? Second, do researchers with differ\u00adent mobil\u00adi\u00adty patterns differ in their research perfor\u00admance? Using large-scale biblio\u00admet\u00adric data, this blog post examines mobil\u00adi\u00adty trajec\u00adto\u00adries and research perfor\u00admance for more than 300,000 researchers connect\u00aded to Germany.<\/p>\n<\/div> The in-house version of the Scopus database provid\u00aded by the German Compe\u00adtence Network for Biblio\u00admet\u00adrics<\/a> (snapshot as of 10.2025) was used (Schmidt et al. 2025<\/a>)<\/span>. Elsevi\u00ader disam\u00adbiguates authors in Scopus by group\u00ading all publi\u00adca\u00adtions into profiles with a unique author IDs. The Scopus author ID was used to identi\u00adfy all publi\u00adca\u00adtions for each author. All publi\u00adca\u00adtions and affil\u00adi\u00ada\u00adtions linked to a given author ID were treat\u00aded as belong\u00ading to the same researcher. The initial sample includ\u00aded all authors who had published at least one document with a German insti\u00adtu\u00adtion\u00adal affil\u00adi\u00ada\u00adtion between 2000 and 2025.<\/p>\n To study mobil\u00adi\u00adty, we recon\u00adstruct\u00aded individ\u00adual affil\u00adi\u00ada\u00adtion trajec\u00adto\u00adries at the country level. For each researcher and publi\u00adca\u00adtion year, the modal country of affil\u00adi\u00ada\u00adtion was identi\u00adfied, meaning the country that appeared most frequent\u00adly across that researcher\u2019s publi\u00adca\u00adtions in that year. When two or more countries were equal\u00adly frequent\u00adly in a given year, they were compared to the modal country from the previ\u00adous year. If one of the tied countries matched the previ\u00adous year\u2019s modal country, that country was assigned. If none matched, one country was select\u00aded at random. To account for short gaps in publi\u00adca\u00adtion activ\u00adi\u00adty and reduce the artifi\u00adcial fragmen\u00adta\u00adtion of mobil\u00adi\u00adty sequences, publi\u00adca\u00adtion years were extend\u00aded backward by two years. This proce\u00addure produced a longi\u00adtu\u00addi\u00adnal country-level trajec\u00adto\u00adry for each researcher<\/p>\n Based on these trajec\u00adto\u00adries, the researchers were classi\u00adfied into five mutual\u00adly exclu\u00adsive mobil\u00adi\u00adty types:<\/p>\n To ensure reliable identi\u00adfi\u00adca\u00adtion of mobil\u00adi\u00adty and stable produc\u00adtiv\u00adi\u00adty indica\u00adtors, the analyt\u00adi\u00adcal sample was restrict\u00aded to researchers with at least three publi\u00adca\u00adtions and at least two years between their first and last publi\u00adca\u00adtion year. These researchers should have been active in recent years, which is defined as a first publi\u00adca\u00adtion year \u2264 2022 and a last publi\u00adca\u00adtion year \u2265 2023. After apply\u00ading these crite\u00adria, the final sample includ\u00aded 317,314 researchers.<\/p>\n<\/div> The results show that most researchers connect\u00aded to Germany are Locals (Figure 1<\/em>). About 57% of the sample remained affil\u00adi\u00adat\u00aded with Germany through\u00adout the observed period. At the same time, inter\u00adna\u00adtion\u00adal mobil\u00adi\u00adty is far from margin\u00adal. Rough\u00adly 43% of researchers fall into one of the inter\u00adna\u00adtion\u00adal\u00adly mobile categories. Among these, Arrivals and Leavers each account for about 13% of the sample. Visitors repre\u00adsent 12%, while Returnees make up about 5%. These figures suggest that the German science system is deeply embed\u00added in inter\u00adna\u00adtion\u00adal mobil\u00adi\u00adty flows. Researchers both enter and leave the system, and some return after spend\u00ading time abroad.<\/p>\n<\/div> Figure 1: Distri\u00adb\u00adu\u00adtion of German researchers by mobil\u00adi\u00adty group. <\/p><\/div><\/div><\/div> Howev\u00ader, the degree of inter\u00adna\u00adtion\u00adal\u00adiza\u00adtion varies great\u00adly across research fields (Figure 2<\/em>). Highly inter\u00adna\u00adtion\u00adal fields include mathe\u00admat\u00adics, physics and astron\u00ado\u00admy, arts and human\u00adi\u00adties, earth and plane\u00adtary sciences, and econom\u00adics, econo\u00admet\u00adrics and finance. In these disci\u00adplines, less than half of researchers remain locals. By contrast, fields such as health profes\u00adsions, dentistry, nursing, medicine, and chemi\u00adcal engineer\u00ading show much higher shares of locals, exceed\u00ading 70%. These disci\u00adplines appear more strong\u00adly embed\u00added in nation\u00adal insti\u00adtu\u00adtion\u00adal and profes\u00adsion\u00adal structures.<\/p>\n<\/div> Figure 2: Distri\u00adb\u00adu\u00adtion of German researchers by mobil\u00adi\u00adty group across research fields. <\/p><\/div><\/div><\/div> The analy\u00adsis also examined whether researchers with differ\u00adent mobil\u00adi\u00adty patterns differ in their research perfor\u00admance. Research perfor\u00admance was measured using sever\u00adal indica\u00adtors, including:<\/p>\n The indica\u00adtors were calcu\u00adlat\u00aded using both whole and fraction\u00adal count\u00ading methods and with differ\u00adent citation windows. All indica\u00adtors were comput\u00aded using both whole count\u00ading, in which each co-authored publi\u00adca\u00adtion is fully counts for each author, and fraction\u00adal count\u00ading, in which publi\u00adca\u00adtion credit is divid\u00aded by the number of co-authors. To control for the influ\u00adence of large-scale collab\u00ado\u00adra\u00adtions, where individ\u00adual contri\u00adbu\u00adtions are diffi\u00adcult to assess, all indica\u00adtors were calcu\u00adlat\u00aded using two sets of publi\u00adca\u00adtions. One set includ\u00aded all publi\u00adca\u00adtions, while the other includ\u00aded only publi\u00adca\u00adtions with fewer than 20 authors.<\/p>\n Indica\u00adtor labels follow the struc\u00adture [Metric] [variant] [team filter] ([count\u00ading])<\/strong>. Pubs<\/em> = number of publi\u00adca\u00adtions; Prest.<\/em> = publi\u00adca\u00adtions in prestige journals (Norwe\u00adgian Regis\u00adter, Level 2); Cit.rate<\/em> = mean field-normal\u00adized citation rate; Top10%<\/em> = number of top 10% most cited publi\u00adca\u00adtions. Citation windows: 3y<\/em> = 3\u2011year, 5y<\/em> = 5\u2011year, all<\/em> = full window. <20<\/em> = restrict\u00aded to publi\u00adca\u00adtions with fewer than 20 authors. (W)<\/em> = whole count\u00ading; (F)<\/em> = fraction\u00adal counting.<\/p>\n To allow meaning\u00adful compar\u00adisons across disci\u00adplines and career stages, all indica\u00adtors were convert\u00aded into percentiles within fields and acade\u00adm\u00adic age groups. A percentile of 90 means that a researcher\u2019s indica\u00adtor value is higher than 90% of other researchers in their field and acade\u00adm\u00adic age group. Next, weight\u00aded mean percentiles were calcu\u00adlat\u00aded for each mobil\u00adi\u00adty group. Final\u00adly, the results across all indica\u00adtors were synthe\u00adsized using a random-effects meta-analysis.<\/p>\n The results show a consis\u00adtent pattern. Locals have the lowest median perfor\u00admance across all 32 indica\u00adtors, with mean percentiles below 50 (Figure 3<\/em>). This indicates that, on average, researchers in this group are in the lower half of the perfor\u00admance distri\u00adb\u00adu\u00adtion. In contrast, Returnees and Visitors show the highest average perfor\u00admance levels. Their mean percentiles are typical\u00adly above 50, placing them in the upper half of the distri\u00adb\u00adu\u00adtion (Figure 4<\/em>). Arrivals and Leavers fall between these two extremes, with mean percentiles both above and below the median, depend\u00ading on the indica\u00adtor (Figure 5<\/em>). Howev\u00ader, it is impor\u00adtant to note that these differ\u00adences are averages. The distri\u00adb\u00adu\u00adtions overlap signif\u00adi\u00adcant\u00adly (Figure 3<\/em>). Each mobil\u00adi\u00adty group includes researchers with both very high and very low performance.<\/p>\n<\/div> Figure 3: Distri\u00adb\u00adu\u00adtion of perfor\u00admance indica\u00adtors among differ\u00adent mobil\u00adi\u00adty groups.<\/p><\/div><\/div><\/div> Figure 4: Forest plot with individ\u00adual effects. <\/p><\/div><\/div><\/div> Figure 5: Pooled mean percentile in mobil\u00adi\u00adty groups. <\/p><\/div><\/div><\/div> The results above are based on all researchers in the sample. To explore the same figures for individ\u00adual research fields, an inter\u00adac\u00adtive compan\u00adion app is avail\u00adable here<\/a>.<\/em><\/p>\n<\/div> The results are not entire\u00adly surpris\u00ading. Previ\u00adous studies have already shown a positive relation\u00adship between inter\u00adna\u00adtion\u00adal mobil\u00adi\u00adty and individ\u00adual research perfor\u00admance. This relation\u00adship is likely bidirec\u00adtion\u00adal. On the one hand, mobil\u00adi\u00adty may improve research perfor\u00admance. Working in differ\u00adent countries can provide access to valuable resources, expand profes\u00adsion\u00adal networks, and expose researchers to new research infra\u00adstruc\u00adtures, insti\u00adtu\u00adtion\u00adal environ\u00adments, and funding oppor\u00adtu\u00adni\u00adties. These experi\u00adences may help researchers devel\u00adop new collab\u00ado\u00adra\u00adtions and increase the visibil\u00adi\u00adty of their work. On the other hand, mobil\u00adi\u00adty may also reflect selec\u00adtion process\u00ades. Researchers with strong publi\u00adca\u00adtion records and promis\u00ading career prospects are more likely to receive oppor\u00adtu\u00adni\u00adties to work abroad. In this sense, inter\u00adna\u00adtion\u00adal\u00adly mobile researchers may already belong to the more success\u00adful members of the scien\u00adtif\u00adic workforce before they move.<\/p>\n It is also impor\u00adtant to keep in mind that the groups being compared here may differ not only in their mobil\u00adi\u00adty patterns, but poten\u00adtial\u00adly also in their compo\u00adsi\u00adtion. Although all indica\u00adtors were normal\u00adized by field and acade\u00adm\u00adic age, the groups may still differ in ways that these adjust\u00adments do not capture. The Locals group is the largest and is likely to be the most hetero\u00adge\u00adneous. For example, it may include researchers in predom\u00adi\u00adnant\u00adly teach\u00ading-orient\u00aded positions, at univer\u00adsi\u00adties of applied sciences, or working part-time. All of these charac\u00adter\u00adis\u00adtics can limit publi\u00adca\u00adtion output without reflect\u00ading lower research abili\u00adty. By contrast, inter\u00adna\u00adtion\u00adal\u00adly mobile researchers may repre\u00adsent a more selec\u00adtive group within the research workforce. These compo\u00adsi\u00adtion\u00adal differ\u00adences should be kept in mind when inter\u00adpret\u00ading the perfor\u00admance gaps between groups.<\/p>\n From a science policy perspec\u00adtive, the findings never\u00adthe\u00adless highlight the impor\u00adtance of inter\u00adna\u00adtion\u00adal mobil\u00adi\u00adty as a struc\u00adtur\u00adal feature of the German research system. Inter\u00adna\u00adtion\u00adal mobil\u00adi\u00adty is not margin\u00adal but a central compo\u00adnent of how the system operates. Germany is a highly inter\u00adna\u00adtion\u00adal\u00adized research environ\u00adment that both attracts researchers from abroad and sends researchers to other countries. Scien\u00adtif\u00adic careers connect\u00aded to Germany are there\u00adfore often shaped by inter\u00adna\u00adtion\u00adal mobil\u00adi\u00adty. The results suggest that Germany benefits from this global circu\u00adla\u00adtion of researchers. Inter\u00adna\u00adtion\u00adal\u00adly mobile groups tend to show higher average perfor\u00admance levels than the Locals group, though this pattern may partly reflect the compo\u00adsi\u00adtion\u00adal differ\u00adences discussed above. This under\u00adlines the impor\u00adtance of maintain\u00ading openness to inter\u00adna\u00adtion\u00adal researchers. At the same time, the finding that researchers who leave Germany also tend to perform above the overall average may raise questions about how attrac\u00adtive the German research system is for retain\u00ading produc\u00adtive scien\u00adtists. Final\u00adly, the findings highlight the poten\u00adtial impor\u00adtance of return mobil\u00adi\u00adty. Researchers who spent time abroad and later returned to Germany show the highest average perfor\u00admance levels among all mobil\u00adi\u00adty groups. Tempo\u00adrary mobil\u00adi\u00adty may there\u00adfore allow researchers to accumu\u00adlate inter\u00adna\u00adtion\u00adal experi\u00adence, networks, and knowl\u00adedge that they later bring back into the German research system.<\/p>\n<\/div> Schmidt, Marion, Chris\u00adtine Rimmert, Dimity Stephen, Christo\u00adpher Lenke, Paul Donner, Simone G\u00e4rtner, Niels Taubert, Thomas Bausen\u00adwein, and Stephan Stahlschmidt. 2025. \u201cThe Data Infra\u00adstruc\u00adture of the German Kompe\u00adten\u00adznet\u00adzw\u00aderk Bibliome\u00adtrie: An Enabling Inter\u00adme\u00addi\u00adary Between Raw Data and Analy\u00adsis.\u201d Quanti\u00adta\u00adtive Science Studies<\/em> 6: 1129\u201346. https:\/\/doi.org\/10.1162\/qss.a.20<\/a>.<\/p>\n<\/div><\/div><\/div>
\nData and identi\u00adfy\u00ading mobility<\/a>
\nHow inter\u00adna\u00adtion\u00adal is the German research system?<\/a>
\nDo mobile researchers perform differently?<\/a>
\nWhat do these results mean?<\/a><\/p>\n<\/div><\/div><\/div>Mobil\u00adi\u00adty patterns and research perfor\u00admance of researchers connect\u00aded to Germany<\/h2>\n
\nPUBLISHED: August 11, 2026
\nAFFILIATION: German Centre for Higher Educa\u00adtion Research and Science Studies (DZHW)<\/p>\nData and identi\u00adfy\u00ading mobility<\/h2>\n
\n
How inter\u00adna\u00adtion\u00adal is the German research system?<\/h2>\n
<\/span> <\/h2>
<\/span>research_fields<\/h2>
Do mobile researchers perform differently?<\/h2>\n
\n
<\/span>performance_indicators<\/h2>
<\/span>indiv_effects<\/h2>
<\/span>pooled_mean<\/h2>
What do these results mean?<\/h2>\n
Refer\u00adences<\/h3>\n
![](\"https:\/\/bibliometrie.info\/wp-content\/uploads\/2022\/11\/back-to-top.svg\")
![](\"https:\/\/bibliometrie.info\/wp-content\/uploads\/2022\/11\/back-to-top_hover.svg\")
<\/a><\/div><\/div><\/div><\/div>