FrameAxis: Characterizing Framing Bias and Intensity with Word Embedding

Framing is a process of emphasizing a certain aspect of an issue over the others, nudging readers or listeners towards different positions on the issue even without making a biased argument. Here, we propose FrameAxis, a method for automatically characterizing framing of a given text by identifying the most relevant semantic axes ("microframes") that are overrepresented in the text using word embedding. In contrast to the traditional framing analysis, which tends to be constrained by a small number of manually annotated general frames, our unsupervised approach can be readily applied to large datasets because it does not require manual annotations. It can also provide more nuanced insights by considering a host of semantic axes. Our method is designed to quantitatively tease out two important dimensions of framing: \textit{framing bias} -- how biased an argument is -- and \textit{framing intensity} -- how much a particular aspect over another is highlighted -- from the text, offering a nuanced characterization of framing. We demonstrate how FrameAxis successfully captures framing bias and intensity in a variety of text data from restaurant reviews to news media. The existing domain knowledge can be directly incorporated into FrameAxis by guiding candidate microframes to test and fine-tune automatically discovered microframes. We propose methods for explaining the results of FrameAxis at the level of individual words and documents. Our method may accelerate scalable and nuanced computational analyses of framing across disciplines.