KEY CONCEPT

Despite modern communications technology, collaborators will learn more from each other if they are not geographically separated, according to a study based on the analysis of collaborations of 1.7 million scholars.

---

IDEA SUMMARY

Because of new communications technologies, including a proliferation of video-conferencing options over the Internet, many have declared the “death of distance.” Geographical separation of even thousands of miles is no longer a barrier to close collaboration. Collaborators can work together on projects with the same ease and effectiveness whether they are physically located on different continents or in the same building.

Not everyone agrees. One study, based on collaborations among academics, argues that geographically separated collaborators do not benefit from the same level of “knowledge spillover” that are more likely to emerge from in-person collaborations. Knowledge spillover refers to knowledge acquired from observing, interacting with, and imitating others. The repeated, information-rich interactions of collaborations are among the best opportunities for collaborators to increase their knowledge portfolio. Geographic separation, according to the study, reduces this important benefit of collaborations.

For this study, researchers drew on the massive Microsoft Academic Graph (MAG) database of more than 60 million academic publications dating back to 1800. They focused on academic papers published between 1975 and 2015, looking for scholars who had published three papers consecutively that met certain criteria:

• The first paper published by the scholar would be single-authored and allowed the researchers to capture the scholar’s entire pre-collaboration knowledge portfolio.

• The second paper in the sequence was a collaborative effort that included new knowledge not included in the first paper.

• The third paper was a single-authored publication.

To identify knowledge spillover, the researchers compared the knowledge portfolios in the second and third papers. If the third paper incorporated new knowledge found in the second paper, the researchers concluded that knowledge spillover had occurred. If the new knowledge in the second paper was not present in the third paper, the researchers considered that no knowledge spillover had occurred.

The research team correlated knowledge spillover to collaboration distance by averaging out the distance between the affiliations (e.g., university departments) of the collaborators in the second paper. An average distance of less than 700 meters between affiliations was considered a local collaboration; anything further was non-local.

The researchers also calculated from the available publication data the learning rate the percentage of collaborators who acquired new knowledge from the collaboration and the learning premium the proportion of local to non-local collaboration learners. For example, a 100% learning premium meant that local collaborators would be twice as likely to learn from each other than non-local collaborators.

An analysis of the data yielded the following results:

• Local collaborations were more likely to lead to knowledge spillover than non-local collaborations. In addition, both the learning rates and learning premiums for collaborators were higher if the collaboration was local.

• For both the learning rate and learning premium, certain factors, such as academic discipline, career stage, team size, and affiliation ranking, further increased the advantage of local collaborations.

• For example, the researchers found that learning rates after local collaborations were highest for geography (10%), and lowest for medicine (3.9%). The greatest difference between local and non-local collaborations was in physics, where local collaboration led to learning rates 4.17% higher than non-local collaboration.

• The researchers also found that the greatest learning premiums were attained when scholars were in the early stages of their careers, worked with fewer collaborators, and were associated with lower-ranked institutions.

The learning premium for local collaborations increased from approximately 50% in the 1980s to about 80% in 2010, indicating that collaborating in person is even more important for learning today than in the past, despite the advances in technology.

---

BUSINESS APPLICATION

This study makes a persuasive case that geographical proximity enhances the chances of collaborators learning from each other. The study’s conclusions have implications in various domains.

Innovation strategy has long recognized the value of local collaboration as evidenced by firms involved in a single economic activity clustering in one geographic location to benefit from a local pool of specialized knowledge. The study supports the benefit for companies especially those seeking to target fields of knowledge beyond their knowledge portfolio to locate subsidiaries in areas with local access to knowledge.

Working from home is more prevalent, especially for knowledge workers, than ever before as a consequence of the COVID-19 pandemic health constraints. This study supports leaders who believe that employees in knowledge industries benefit from the repeated interactions and face-to-face collaboration that are lost with work-from-home arrangements.

This study also has implications for teamwork. To ensure that team members learn from each other, this study implies that companies and organizations should attempt to create smaller teams and have them meet face-to-face as much as possible.

---

---

FURTHER READING

Frank van der Wouden’s profile at University of Hong Kong

https://www.geog.hku.hk/f-vanderwouden

Hyejim Youn’s profile at Kellogg School of Management

https://www.kellogg.northwestern.edu/faculty/directory/youn_hyejin.aspx

---

REFERENCES

The impact of geographical distance on learning through collaboration. Frank van der Wouden, Hyejim Youn. Research Policy (March 2023).

https://www.sciencedirect.com/science/article/pii/S0048733322002190?via%3Dihub

•   DOWNLOAD IDEA
•   PIN TO MY IDEAS
•   PRINT IDEA
•   LIKE

---