Several students and faculty from the Claremont Graduate University (CGU) and Pomona College (PO) have collectively come together to form the Digital Innovation & Text Analysis “Fake News” Lab, a lab with the goal to find a method to evaluate truth claims, public statements, and the trustworthiness of news stories. While similar research in the past has looked at superficial features such as article headlines and article sources as reference points for a text’s trustworthiness, this lab is looking at linguistic (semantic and syntactic) features which could be more holistically indicative of a text’s trustworthiness. The project is an extension of a similar early-stage research project from a year prior, but this lab has the goal of further developing and improving upon the dataset and method of evaluation of trustworthiness that resulted the previous project’s work.
The lab meets weekly and provides students with an opportunity to work under the guidance of an experienced faculty member, Professor Hovig Tchalian, a professor of management whose research focus is language and innovation. He also teaches an introductory data analysis course and uses text analysis techniques to study social discourse. The lab is managed by a project manager, Kristina Khederlarian (CGU PhD student in computational analytics and international relations). The rest of the research team is made up of Anthony Lyons (CGU PhD student in information systems), Amin Nash (CGU Master’s student in English critical theory and analysis), Brady DeMeritt (Pomona ‘19 majoring in computer science and linguistics & cognitive science), Daniela Hinojosa Sada (PO ‘19 majoring in linguistics), Alex Ker (PO ‘22 majoring in philosophy and computer science), and Jack Weber (PO ‘22 majoring in Computer Science and Economics).
Comprising people with diverse backgrounds, the team’s skill sets range from politics and simulation modeling to linguistics and computer science. The project, which is funded by a Project Research Grant from the Digital Humanities Initiative at The Claremont Colleges (DH@CC), has given students a chance to do extensive research by not only having them find and collect usable data sources, but by also having them process that data and use advanced digital techniques such as those of natural language processing and machine learning to analyze the data to create a trustworthiness prediction model. The lab has so far worked to collect data and process that data into linguistic features, and the next step is to use machine learning to identify relevant correlations between the data’s features and trustworthiness levels. There are hopes to in the future not only incorporate this research into team member’s dissertations and theses, but to also further analyze the role that a fake news proliferator’s intent, whether backed by malevolence or ignorance, plays in the manifestation of the linguistic features in fake news.