In November, Dan presented the TakePart team’s ongoing work on emotional cues in protest images at the monthly ‘Research in Progress’ Seminar of the Department of Sociology & Criminology, City St George’s, University of London. Based on a dataset of protest images associated with the Romanian #rezist protests that were shared on Twitter, the research probes the relation between the multimodal expression protest images encapsulate on social media (as a combination of facial expressions, textual messages and textual messages embedded in images as in the case of placards held up by protestors) and the engagement they generate. To that end, the team compared protest images (i.e. captured at the site of a protest) and non-protest images. Using a combination of human coding, machine classification and statistical analyses, the team was able to show that protest images accompanied by positive messages increased the visibility of Twitter posts when compared to non-protest images and negatively valanced or neutral messages. As such, the research suggest that protest images combined with positive messages can act as a pull factor for protest engagement online. Other research (Kim & Elisson, 2021) has already show that such online engagement is integral to a social learning process that culminates with offline political participation.