Morteza Dehghani: USC Brain and Creativity Institute, ARTIS Research Fellow

The availability of vast and seemingly insurmountable volumes of human-related data has provided an unprecedented opportunity to study human cognition with range and detail previously not imaginable. An enormous amount of such data, however, is in the form of human generated text, and cannot be analyzed directly. As a result, there has been rapid developments in automated text analysis methods focused on measuring psychological and demographic properties. In this talk, I will present a computational text analysis technique for tracking and measuring transformations in moral concerns with regards to different social-cultural issues, and for examining the moral dimensions of different debates using text. This technique uses Latent Semantic Analysis to compute the semantic similarity between concepts of interest and moral keywords taken from the Moral Foundation Dictionary (Graham, Haidt & Nosek, 2009). I will specifically focus on analyzing Twitter data regarding the 2013 federal government shutdown. Our results demonstrate that using moral loadings of tweets we can make various accurate predictions about Twitter users and communities.