- My second point concerns the operationalization process and labelling of the clusters
The study sample is composed of 70 keynote speakers at the historical main conference focused on computational social science and all their publications as retrieved using the Microsoft Academic Graph. Authors argue this allows for tracing back the emergence of the field. This is totally fine, but when authors state that it allows for comparing ‘more recent publication patterns to those from times when ‘computational social science’ was not yet recognized as designating an emerging field’, it would be welcome to prior specify when exactly it was recognized as an emerging field, by who, and in what context. Authors state that ‘it seems natural to assume that the choice of keynote speakers reflects an attempt to choose a representative and diverse selection of influential authors in the field’ (page 21). Here again, more context and maybe some indications about the background of these speakers would inform on the social and institutional conditions for the emergence of the computational social science.
The methodological part is quite extensive, as compared to the empirical section. Moreover, some methodological elements are displayed in the results. For instance, many elements in section 4.3 RQ2: The coherence of computational social science as a field should be put into the methodology. The result section also deserved a deeper engagement in the interpretation of the findings that support the conclusion.
I have several demands for clarification concerning the methodological part.
- The functioning of Leiden community-detection algorithm should be unpacked, and authors should explain what software they used to perform their analyses.
- Authors should explain why 168 communities. There are many ways for choosing between clusters solutions, hence this must be clarified.
- Authors need to clarify why they chose to manually inspect 20 most co-cited papers x 168 communities, instead of producing a distribution of journals disciplines in and across the clusters, for instance.
- How was the degree of triadic closure and the subsequent clustering measured?
The methodological section deserves a clear indication of how disciplines, field, and topics were operationalized. Otherwise, conclusions cannot be drawn on the empirical material. For instance:
- ‘First of all, inter-field links between works are not apparent from the clusters’ (p.18). To support this conclusion, one must understand what these clusters are and how they were labelled.
- ‘Their pairing does not seem to demonstrate the integrative idea-sharing that is held up as an ideal for knowledge-building across the branches of the social and computational sciences’ – what are branches? Do they refer to disciplines?
The study sample is composed of 70 keynote speakers at the historical main conference focused on computational social science and all their publications as retrieved using the Microsoft Academic Graph. Authors argue this allows for tracing back the emergence of the field. This is totally fine, but when authors state that it allows for comparing ‘more recent publication patterns to those from times when ‘computational social science’ was not yet recognized as designating an emerging field’, it would be welcome to prior specify when exactly it was recognized as an emerging field, by who, and in what context. Authors state that ‘it seems natural to assume that the choice of keynote speakers reflects an attempt to choose a representative and diverse selection of influential authors in the field’ (page 21). Here again, more context and maybe some indications about the background of these speakers would inform on the social and institutional conditions for the emergence of the computational social science.
The methodological part is quite extensive, as compared to the empirical section. Moreover, some methodological elements are displayed in the results. For instance, many elements in section 4.3 RQ2: The coherence of computational social science as a field should be put into the methodology. The result section also deserved a deeper engagement in the interpretation of the findings that support the conclusion.
I have several demands for clarification concerning the methodological part.
The methodological section deserves a clear indication of how disciplines, field, and topics were operationalized. Otherwise, conclusions cannot be drawn on the empirical material. For instance: