Leveraging Expertise Diversity

Cross-domain collaboration is considered an important source of creativity and innovation, particularly important to deal with complex tasks. Tackling such complex tasks requires insights from different disciplines. A key research question arises: How can efficient collaboration occur across boundaries of different expertise areas? Inspired by various reference disciplines including psychology, sociology, organization/management theory and computer science we focus on how collaboration across boundaries of different disciplines takes place and what are its consequences. Applying social network analysis perspective, we will use two-mode network methodology to examine patterns of similarity and complementarity in collaborative relationships between scientists. Our goal is to investigate patterns of knowledge sharing and utilization across boundaries.

This project draws on the results of a recent successful interdisciplinary collaboration previously funded by the Network Institute (research voucher). In the previous project we coded and classified self-reported areas of expertise applying grounded theory. We developed a simple framework for the unique labels respondents reported (open coding), then we identified common themes and categories (axial coding), generalized the labels (grouping similar labels into a more general category) and finally applied simple disambiguation techniques. We validated the coding procedure by combining its results with a pilot network survey to confirm that expertise overlap was strongly predictive of (intended) self-reported collaboration over and above the effects of formal supervisory relationships, tenure and age similarity, same gender, hierarchical position, and departmental affiliation. Results were presented at Sunbelt, an international conference on social network analysis in May 2013. Extending these findings, we now propose to assess intercoder reliability and design an approach through which experts from the relevant fields can rate the expertise areas to further advance disambiguation. To assess the impact of interdisciplinary collaboration we need to collect data on the pre- and post-survey research output for the sampled group of scientists.



  •  Melvin Matthesius
    MSc Business Administration
  • Marija Selakovic
    MSc Computer Science