The objective of the dataset is to list and categorize any participant in international environmental negotiations who identifies as a youth representative. This is a political definition of youth, which does not limit itself to the definition of youth as an age range. Indeed, our dataset methodology is not based on age data, which is, in any case, not available from the lists. Instead, it considers that the way actors define their own role through their registration data is the most relevant criteria to identify youth political activism.

As a consequence, we worked on individuals’ own description records. In particular, the data was acquired in different steps. First, we collected the lists of participants to the international negotiation processes covered by the study through the corresponding secretariat websites: the UNFCCC website; the CBD website; the HLPF website and the CSD former websites; for on-site participation exclusively. Access to all the UNFCCC lists of participants is public as they all appear on the website. It is the same situation for the HLPF and CSD. However, not all the CBD’s lists of participants were available online. In that case, the team contacted the CBD secretariat in the context of the YouthEarth project to obtain the missing lists by email. We also used the reporting online service of the Earth Negotiations’ Bulletin (ENB) that provides summaries of global environmental events, and general online searches (including a UNDESA archive: https://www.un.org/development/desa/youth/youth-at-the-sustainable-development-commission.html), to check for additional attendance and to cross-check the data.

Second, we used keywords to identify youth actors within lists and official documents. All individuals attending international negotiations must register through an organization. We therefore searched both the participants’ and organizations’ descriptions for keywords including ‘youth’, ‘young’, ‘child’, ‘jeune’, ‘joven’, ‘student’, ‘estudiant’, ‘étudiant’, ‘scout’, ‘girl’, ‘boy’, and ‘kid’ and their derivatives. This means that students from Bachelor to PhD were also registered in the dataset. However, we excluded these (students), if the main organization of the participant was based in the country where the event was taking place. Indeed, in that case, we considered students to be more inclined to be there for educational purposes than for political influence.

Third, we developed a methodology to include in the dataset all participants from youth organizations which did not necessarily have youth-related keywords in their description. To do so, we collected the lists of organizations accredited to the UNFCCC Secretariat and to UN Environment (UNEP) under the official youth constituency and of organizations members of YOUNGO in 2018/2019. We included all the participants who registered with these organizations in the dataset.

On this basis, each youth participation record was described, processed or analyzed in the dataset (see Table 1, Table 2). Finally, we anonymized the dataset, due to the sensitivity of personal data, in a domain where activists can be prosecuted for their environmental actions. In total, four different researchers worked on creating the dataset, constantly coordinating to compare coding results. Moreover, to ensure the robustness of the data, random checks were regularly performed to verify coding reliability. Finally, the team already produced several research articles using part of the dataset, meaning that it conducted verification procedures of the dataset for each of these publications.

Finally, to contribute to debates on multilateral environmental negotiations [16], we categorized country data into developing and developed in the country_type variable. To reflect existing categorization in each of the three processes, we followed the UN nomenclature of each of the processes when it was available in international treaties, as is the case for the CBD and UNFCCC processes. For the HLPF and CSD summits, as no official nomenclature existed, we processed with the most commonly used categorization in UN processes, i.e. is the IMF country classification. To simplify and harmonize our country_type variable across processes, we chose the most commonly used category for each country, with the Helper_country_type_categorization variable making this harmonization step explicit.