Online abuse is a phenomenon that has been widely studied in the context of individuals and activists. Very little scholarly attention has been paid to its occurrence among politicians, even though the implications online abuse has – especially the silencing of their online voices – are possibly very similar.

The aim of this thesis was to contribute to filling the gap in research about online abuse against politicians. To this end, I raise four hypotheses concerning ideology and Virtual Manhood Acts. To test these hypotheses, I collect about 500000 tweets and classify them using a combination of word embeddings and manual coding.

The results show that about 3.4% of all received tweets are abusive. It is present among all ideologies and genders, with a slightly higher probability to receive abusive tweets for male politicians on the left and politicians with a very traditional, authoritarian ideology.