Under the advisement of Dr. Chris Callison-Burch with the School of Engineering and Applied Science and Dr. Emily Falk with the Annenberg School for Communication at the University of Pennsylvania, my PhD research project is a cross discipline collaboration between Natural Language Processing (NLP) and Cognitive Science called Motivational Boost.
Whereas prior work has looked at features of naturally occurring language and how persuasive that language turns out to be, Motivational Boost is novel in looking for casual explanations behind what makes text persuasive by manipulating linguistic features in content messages and measuring the perception of persuasion, motivation, relevance, intention, and confidence on the ability to positively change behavior. As an extension to work done by Lillian Lee in “Winning Arguments: Interaction Dynamics and Persuasion Strategies in Good-faith Online Discussions,” (Tan et al, 2016) we take the linguistic features observed in a subreddit called “Change my View” (CMV) to manipulate motivational messages on parenting. CMV is an argumentation platform on Reddit where a stance on a topic is posted by an original poster (OP) and responders can make reasonable objections through the civil discourse this platform provides. At the end of an argumentation thread, the original poster rates which responder’s viewpoint changed their mind, giving researchers a structured dataset that includes language used to persuade and ratings of success.