A Visual Political World: Determinants and Effects of Visual Content

Political communication is a central element of several political dynamics. Its visual component is crucial in understanding the origin, characteristics and consequences of the messages sent between political figures, media and citizens. However, visual features have been largely overlooked in Political Science. Thus, in this dissertation, I introduce, describe and apply computer vision techniques for the analysis and processing of visual material, in order to not only improve data collection and visual content extraction, but also the understanding of the effects that visual components have on relevant political variables. In the first main chapter of this project, I implement computer vision and image retrieval techniques to measure and understand messages conveyed in pictures. The chapter presents and details the implementation of a Bag of Visual Words (BoVW), an intuitive and accessible technique for the extraction and quantification of visual features that allows researchers to build an Image-Visual Word matrix that emulates the Document-Term matrix in text analysis. For the purposes of this chapter, I validate the BoVW approach using a structural topic model to identify relevant political features of images of the migrant caravan. The second main chapter introduces and describes the implementation of a popular tool in Computer Vision, Convolutional Neural Networks (CNN), for the processing and extraction of political information. I apply the CNN for the extraction of handwritten numbers in electoral tallies, and enumerate the benefits and drawbacks that this technique has for the study of political events. Finally, the third main chapter studies the factors behind the generation of visual frames, and the impact that these have on political attitudes. By focusing on the depictions and visual framing of protests, I find that conservative newspapers depict the protests in darker and nocturnal settings more often than liberal outlets. Further, the framing of the mood of the environment with conflict-related elements have an impact on the opinions and attitudes towards social movements: depictions of violence negatively affect identification and engagement with the movement, and these effects are moderated by who is the actor behind the violent events. Overall, the dissertation focuses on the importance that visuals have on the way that citizens engage with political information, and provides a framework that allows researchers to have a better understanding of several political dynamics