**2. Materials and methods**

The 2016 JCR® Social Sciences Edition [37] was retrieved on June 1, 2017, to find out the name and number of the journals within the categories of *Demography* and *Urban Studies*. For the former, 26 journals were found, and 38 for the latter.

The time interval covered in this study is from 2000 to 2016. The procedure to obtain the data consisted in analysing the information contained in the SSCI, for which all the records were searched using the parameters: *Publication Name* [name of each journal in the chosen category] and *Year Published* [2000–2016]. In order to extract information only from citable documents, these were filtered once again by their categorisation as *Article* or *Review* (from now on, we are to refer them as documents). The category of *Demography* produced 11,361 documents whereas *Urban Studies* produced 24,010. Out of those documents, those in which the author was anonymous, or the author field was blank, were discarded. Lastly, 11,361 entries were considered for *Demography* and 23,998 for *Urban Studies*, all of which constitute the sample of this study.

All the information was uploaded to an *ad hoc* Microsoft® Access® 2016 relational database (version 1801) for the treatment and normalisation of data, as well as to produce the different graphs. The data were collected by year and collaboration was analysed into two levels. The first level was authorship, looking at collaboration in relation with the number of signatory authors; the number of authors in each document was full-counted, calculating a particular Collaboration Index (CI) and Degree of Collaboration (DC). The second level was established in relation with international collaboration, identifying the countries of each of the authors' institutions.

With a view to count the authors of each document, we opted for the complete counting system, as suggested by Cronin and Overfeld [38], attributing full authorship to each co-author, considering them equally. The same procedure was applied in the case of countries. The documents were grouped according to collaboration by country, as has been done in other similar studies [39]. Given that documents can be signed by authors from different countries, the sum of the percentages is greater than 100%.

To analyse, treat and visualise collaborative networks, we have used the Pajek software [40].
