Title: Understanding visual messages: visual framing and the Bag of Visual Words
Abstract: 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. In this project, I implement computer vision and image retrieval techniques to measure and understand messages conveyed in pictures. More specifically, the article focuses on the analysis of the content and structure of images of Black Lives Matter movement (BLM) protests. For this purpose, the article presents and details the implementation of a Bag of (Visual) Words (BoVW). This method drawn from the field of Computer Science allows researchers to build an Image-Visual Word matrix that emulates the Document-Term matrix in text analysis in order to feed models and classifiers that can provide insights about the content of visual material. Preliminary results from the application of a Structural Topic Model to a corpus of images posted by U.S. newspapers show that conservative outlets tend to include “darker" elements in their depictions of protests: they show more nocturnal events and features like smoke, fire and police patrols than liberal outlets. Overall, the article sheds light on the characteristics and consequences of visual means of communication and persuasion, and provides a useful technique for an accurate analysis and measurement of messages in pictures.
Hi everyone!This week at the Applied Statistics Worksho
p we will be welcoming Michelle Torres, a graduate student in Political Science and Statistics at Washington University in St Louis. She will be presenting work entitled Understanding visual messages: visual framing and the Bag of Visual Words. Please find the abstract below and on the Applied Stats website here.
As usual, we will meet at noon in CGIS Knafel Room 354 and lunch will be provided. See you all there!
-- Dana Higgins
Title: Understanding visual messages: visual framing and the Bag of Visual Words
Abstract: How should one perform matching in observational studies when the units are text documents? The lack of randomized assignment of documents into treatment and control groups may lead to systematic differences between groups on high-dimensional and latent features of text such as topical content and sentiment. Standard balance metrics, used to measure the quality of a matching method, fail in this setting. We present a framework for matching documents that decomposes matching methods into two parts: (1) a text representation, and (2) a distance metric. We consider various methods that can be used at each step and conduct a systematic multifactor evaluation experiment using human subjects to identify the methods that dominate. We also show that our framework can be used to produce matches with higher subjective match quality than current state-of-the-art techniques. We then apply our chosen method to a substantive debate in the study of media bias using a novel data set of front page news articles from thirteen news sources. Media bias is composed of topic selection bias and presentation bias; using our matching method to control for topic selection, we find that both components contribute significantly to media bias, though some news sources rely on one component more than the other.