Tracking moral divergence with DDR in presidential debates over 60 years

Televised presidential debates, a communication form specifically designed to evoke meaningful clash of issue viewpoints, have been criticized for the lack of real clash and issue discussion for decades. Have the debaters made any improvement? This study investigates the evolution of this perennial paradox through the lens of mediatization using an instrument grounded in Moral Foundation Theory. As an outcome of mediatization, politicians have been seeking publicity to achieve authority through media, and therefore they have prioritized self-image building over issue discussion in their social actions. This study quantitatively describes this mediatization process by examining the moral divergence between each pair of presidential debaters with moral loading, an indicator for quantifying moral foundations via DDR, a computational method based on distributed representation. Our results reflect the mediatization process in politics, showing that Democrat and Republican candidates have been increasingly focusing on different moral judgments, and therefore their moral divergence has widened. This study sheds light on the development of ways to encourage more effective political communication by discovering mediatization as a potential determinant of a major challenge faced by televised presidential debates. Accordingly, it provides quantitative empirical evidence for mediatization theory. Moreover, it shows the potential of the distributed representation method, a milestone of machine learning, in future communication explorations.