Corpus of digital interactions: systematization of techniques to collect data on WhatsApp

The collection of datasets from real interactions is an unavoidable step in many research works aiming to understand language use. In the field of digital discourse analysis, data collection is complex due to the fast-paced changes in the applications and the ethical decisions involved. This work has two goals. First, we seek to show an overview of the literature on datasets of digital exchanges by WhatsApp. Then, we aim to systematize different sampling techniques used in previous research. We thus proceeded by applying content analysis to 100 research articles and theses retrieved from open access portals. We conducted a descriptive analysis that included the amount of data collected, the technique employed in the collection of the data, the method used to contact participants, and the online access to the linguistic corpora, among other variables. The results show the existence of some corpora annotated and available in languages other than Spanish. In addition, most of the literature shows a combination of different techniques to collect a wide set of linguistic and multimodal data. Then, we systematize the main methodological alternatives for data collection from digital interactions by WhatsApp, with the participant observation method standing out.