Comparing two Methods of Sociological Analysis of Political Internet Users

While some researchers suggest giving up the sociological analysis of the producers of online messages (Boullier, 2015), we defend the methodological importance of this approach in the study of the political web. Even so, the sociological characterization of the producers of online political messages constitutes a major methodological challenge, which this paper seeks to address. More specifically, we compare the effectiveness and the respective biases of two methods of sociological analysis of the producers of online political messages: - online collection by the researcher of data left by Internet users unaware they are the object of a scientific inquiry - the collection of data directly through answers to an online questionnaire, supplied by Internet users who know the situation of scientific inquiry. To this end, we use data collected for our Ph.D. thesis to constitute a panel of 3509 French political bloggers, whose demographic and sociological characteristics were established by collecting directly, without interaction with them, the sociodemographic information left online by the members of the panel. We then compare these data to those collected via an online questionnaire submitted to the same individuals (745 respondents). Our first results show that the online data posted by the political bloggers are very similar to the answers provided by them to the questionnaire, once they know the survey situation. In addition to these first results, this paper will also attempt to discuss the advantages and inconveniences of both methods of sociodemographic data collection, specifically in terms of occupancy rate of the data or of representativeness of the individuals.