Modelling public opinion, namely in what regards political beliefs or judgments, became current in Digital Humanities due to the progress in domains such as Network Theory, AI and cognitive modelling, Machine Learning, and Natural Language Processing. The interdisciplinary approach for studying public opinion dynamics is strongly influenced by the modeller choices containing a diversity of ethical and political dimensions. For instance, the model is induced by the kind of data being used, and by how good and complete the data is considered to be. Also, models are designed within some political context, thus influencing the decision of which specific data to analyse. An ongoing research will be presented that depicts the diversity of choices during the modelling process. The case study regards news published by British newspapers during the period leading to Brexit. News are originated on three distinct political areas: far-right, right-wing and left-wing. Political narratives transmitted to the population by each of these political areas are identified and compared. This case study illustrates how decisive can be ethical and political implications during the modelling process.

--