Conference Agenda

Purpose: This study investigates the possible intersection between formal academic collaborations, measured through co-authorships, and informal interactions, assessed through joint organization of conferences, events, and/or teaching, among scholars of Early Women Sociologists (EWS). Part of the broader project “Gendering Sociology: Proposal for Research and Teaching”, this paper aims to recover the intellectual contributions of EWS by mapping formal and informal academic networks in this field, verifying if informal collaborations foster structured scholarly output or remain detached from formal research.

Methods: The mapping process was structured in three phases: (1) Desk-based identification of scholars, conducted through a multilingual scoping review across academic databases and research networks to identify EWS. (2) Administration of a questionnaire to international scholars and experts, to collect two types of ego-alter data: the informal relations, and the inspiring EWS. (3) Collection of official bibliographic records from Scopus and Ebsco repositories for respondents and alters.

Hence, we analyze relational data using a multilayer network approach. Specifically, examine three layers: the “formal” academic interactions captured by the co-authorship network (i.e. Layer 1); the “informal” collaborations network, or the overall sociometric structure derived from questionnaire responses (i.e. Layer 2); and finally, the EWS layer, which determines the foundational knowledge base for these scholars (i.e. Layer 3).

Contributions: The findings reveal that, while some international EWS have influenced sociological thought, the research community remains highly fragmented, with informal collaborations largely confined to small, self-referential groups. Notably, informal scholarly interactions rarely translate into formal co-authorships, suggesting the presence of structural barriers that hinder the transformation of intellectual exchange into sustained research output.

In 1994 Gibbons [3] claimed that knowledge production shifted from its traditional form to a new synergy between academic and industrial actors. Although numerous studies have shown the strengthening of academia-industry ties, many also noted resistance to industrial logic and the effort to maintain clear boundaries [2, 6].

The field of AI, with its largely mixed composition, appears emblematic of this reconfiguration.

If initial discoveries were made in academia, many of the recent advancements are industry driven. The private sector, due to its substantial resources and its access to large databases, attracts most new talent, produces the most groundbreaking models and consequently dominates the field [1,4].

If elements such as the creation of industrial laboratories, participation in conferences, or contributions to the scientific literature highlight the rapprochement between the two actors, Raimbault [5] points out that, while some practices align, the goals and subsequent artifacts in which researchers invest remain differentiated. In other words, academics focus on papers as they produce value in their field, and industrials will focus on providing working code and models.

Considering the industrial domination in the field and the many convergences with academia, can we observe the emergence of a unified system of knowledge production and diffusion?

Using the PapersWithCode database as well as the OpenAlex and Github APIs, we collected extensive information about research projects including a code repository and an academic paper. Our findings suggest that, while some similarities emerge, industrial studies largely deviate in their topics, diffusion, and reception by scientific communities.

First, looking at the citations, we noted a slight industrial edge. Although industrial articles receive more attention, we noted that they also were significantly less likely to be fully published (12.3% when including an industrial author and 24% when fully academic). Furthermore, if published, industry papers have a longer time to publication (¯x ≈ 0.97 years for academic papers and ¯x ≈ 1.17 for industrials) and only focus on the top journals, suggesting an instrumental use of the publication system. Moreover, using Uzzi’s [7] method, we observed a higher tendency towards classical topic combinations for industry papers. Finally, employing a time series clustering method, we noted that academic articles are overrepresented in the clusters with exponential citation growth.

Regarding repositories, we noted significant differences. Industrials display higher diversity in programming language and some specificity, such as the usage of CUDA. Moreover, private researchers appear to invest more in the presentation and usability of the repository. This can be seen in several metrics such as the documentation length or the inclusion of” Code examples”. We also note a higher industrial popularity on Github as well as a faster accumulation of stars. But we paradoxically observed very little discrepancy in the maintenance of repositories, with even an academic edge in some metrics, such as the time to close issues.

Overall, our results show that, while more successful, industrial research has narrower interests and displays relatively instrumental use of the artifacts questioning the idea of a unified system of knowledge production.

Scientific collaboration, widely recognized as a key driver of research progress and innovation, has grown significantly over the years across all academic disciplines. This trend has been further strengthened by government policies at both national and international levels, which actively promote collaborative research initiatives. In this context, co-authorship serves as a concrete indicator of collaborative behavior among scholars and is commonly used as a proxy for measuring and analyzing research collaboration

While research topics and methodological approaches often differ between disciplines, there are, for example, communities of scholars that share common ground and have a distinctive pattern in their research interests.

While research topics and methodological approaches often vary across disciplines, there are also communities of scholars that can share common ground and exhibit distinctive patterns in their research interests.

A notable example of this can be found in Italy, where Economics—along with related fields such as Business, Management, and others—coexists with Statistics within the same macro research area (designated as "Area 13" by the Italian Ministry of University and Research). In many Italian universities, scholars from these disciplines are often hired within the same department and often have teaching duties within the same degree/PhD programs. This proximity reflects shared characteristics in departmental and institutional environments, as well as alignment with national strategies and policies on scientific production and research quality.

However, key questions arise regarding the potential convergence of scientific production mechanisms between these two large communities. Specifically, does this shared environment influence coauthorship behavior, shaping coauthorship structures, publication style, and productivity over time?

To address these questions, this studycontribution aims to conduct a comparative analysis of co-authorship networks in Management and Statistics, starting withfrom their network topology and examining similarities and differences in co-authorship dynamics

How can scientific collaboration be explained within a research centre? While much of the social network literature on scientific networks tends to focus on academic papers as a means to explore scientific relationships (Bellotti & Espinosa-Rada, 2025), early studies in the sociology of science and knowledge have highlighted the importance of social interactions in knowledge diffusion and production (e.g., Coleman et al., 1957; Crane, 1972; Mullins, 1972; Collins, 1998; White, 2004). To better understand the complexities of social relationships in scientific settings, this study examines how informal communication, scientific interests, specialization, and structural opportunities influence and constrain scientific collaboration.

This presentation presents a case study of scientific collaboration within a research centre focused on sustainability science. The data was collected by combining surveys (i.e., sociometric) and bibliometric data. We investigate the various mechanisms driving collaboration, including friendship, regular interactions, informal communication, and shared "lazy" time spent together. We also account for the influence of factors such as institutional affiliation, interests in different specialities, and conflicts between researchers. For our analysis, we employ stationary actor-oriented models to analyze multilayer networks (i.e., multiplex and multilevel). The primary objective of this study is to uncover how cross-layer effects shape the relationships perceived as scientific collaborations.

How do new academic organizations emerge in an institutional landscape that favors stability? This study explores the role of networks in shaping the rise of Computational Social Science Labs (CSS Labs)—interdisciplinary research hubs that bridge computer science and social science. Drawing on institutional entrepreneurship and network theories, this research examines how scholars positioned at the intersection of different academic communities (i.e., multivocal actors) play a key role in creating and leading these new organizational forms. The methodology employs two comprehensive datasets: (1) a large-scale co-authorship network mapping collaborations between social and computational scientists using Web of Science publication data, and (2) a newly constructed global dataset detailing the introduction, membership information, and research agendas of CSS Labs. The Bayesian Hierarchical Network Autocorrelation Model will be used to model on the relationship between collaboration network position and the likelihood of being a founder of a CSS Lab. Through network analysis, this study assess how increased cohesion between disciplines fosters the emergence of CSS Labs and how these labs, in turn, reshape collaboration patterns. The findings will contribute to scholarly understanding about the paradox of embedded agency by demonstrating how network positioning enables actors to recombine diverse institutional elements to create novel organizations. This research also engages with the concept of "cultural holes," integrating a knowledge-based perspective into brokerage theory. By revealing the relational and structural mechanisms underlying interdisciplinary organizational emergence, the study offers insights into how new academic institutions take shape amid evolving research landscapes in social science fields.