The Effects of Remote Working on Scientific Collaboration and Impact

TL;DR

This study investigates how remote working during the COVID-19 era affected scientific collaboration and impact. It uses a longitudinal bibliometric analysis of OpenAlex and ArXiv preprints across before, during, and after the pandemic to track geography, institutions, and citation outcomes. The key findings show that cross-border collaborations increased significantly after 2020 and networks expanded in geographic reach, but average citation impact declined, suggesting remote interactions may dilute research quality. A parsimonious, institution-distance-aware model explains the evolution of intra- and inter-institution collaborations and the average team distance, with a parameter alpha capturing distance effects that rise toward 2021 and subsequently fall, consistent with a partial relaxation of geographic constraints in the hybrid era.

Abstract

The COVID-19 pandemic shifted academic collaboration from in-person to remote interactions. This study explores, for the first time, the effects on scientific collaborations and impact of such a shift, comparing research output before, during, and after the pandemic. Using large-scale bibliometric data, we track the evolution of collaboration networks and the resulting impact of research over time. Our findings are twofold: first, the geographic distribution of collaborations significantly shifted, with a notable increase in cross-border partnerships after 2020, indicating a reduction in the constraints of geographic proximity. Second, despite the expansion of collaboration networks, there was a concerning decline in citation impact, suggesting that the absence of spontaneous in-person interactions-which traditionally foster deep discussions and idea exchange-negatively affected research quality. As hybrid work models in academia gain traction, this study highlights the need for universities and research organizations to carefully consider the balance between remote and in-person engagement.

Figures (12)

• Figure 1: Monthly trends in inter-institution collaboration and geographic team distance. Monthly average of the fraction of inter-institution collaborations for: (a) all preprints, and (b) COVID-19-related and non-COVID-19-related preprints (in magenta and green, respectively). The fraction of intra-institution collaboration shows the opposite trend. (c) and (d) show the monthly average team geographic distance (in kilometers) with standard error bars for the same groups. Vertical dotted lines mark the onset of the COVID-19 pandemic (March 2020). A piecewise polynomial regression with two breakpoints is fitted for all preprints and non-COVID-19-related preprints, and a single polynomial model without breakpoints is fitted for COVID-19-related preprints, minimizing squared error.
• Figure 2: Temporal patterns in Microsoft Teams usage and distance-dependent collaboration.(a) Daily (green) and monthly (red) active users on Microsoft Teams from July 2019 to October 2023, based on public earnings reports. The onset of the COVID-19 pandemic (March 2020) is indicated by a vertical dotted line. (b) Collaboration strength between institutions, normalized by institution productivity, decreases with geographic distance and follows a gravity law. Colors correspond to different time periods, each fitted with a power-law decay; $\alpha_g$ denotes the distance decay exponent. (c) Temporal evolution of the optimized geographical proximity parameter $\alpha$ in the collaboration network, according to our model. A piecewise polynomial with two breakpoints is fitted by minimizing the squared error. The vertical dotted line marks the onset of the COVID-19 pandemic (March 2020).
• Figure 3: One-year impact factor and its relationship with team distance in preprints.(a) Monthly average number of citations ($IF$) received by preprints within one year from publication, with standard error bars. (b) Same as (a), restricted to non-COVID-19-related preprints. (c) Monthly counts of COVID-19-related preprints (gray bars) and their average one-year citation counts (magenta, logarithmic scale) with standard error bars. Vertical dotted lines indicate the onset of the COVID-19 pandemic (March 2020). (d) Relationship between average team distance ($ATD$) and $IF$ for preprints, split into five consecutive time periods. The loosely dash-dotted vertical line marks the $ATD$ of the authors of this study ($3474.61$ km).
• Figure A1: Modeling self-edges: institutional strength and collaboration weight Relationship between the institutional strength $s_i$ and the weight of self-edges $w_ii$ in the institutional collaboration network, grouped into five consecutive time periods as indicated by the legend. The model assumes that the probability of intra-institution collaboration is proportional to the institution’s strength, which is consistent with the patterns observed in this plot.
• Figure A2: Temporal evolution of geographical proximity parameter $\alpha$. Temporal evolution of the optimized geographical proximity parameter $\alpha$ in the collaboration network. (a) Model using overall preprints with pairwise co-authorship connections. (b) Model using preprints with three authors, represented as hyperedges. A piecewise polynomial model with two breakpoints is fitted by minimizing the squared error. The vertical dotted line marks the onset of the COVID-19 pandemic (March 2020).