Do What You Say—Computing Personal Values Associated with Professions Based on the Words They Use

Abstract

Members of a profession frequently show similar personality characteristics. In this research, we leverage recent advances in NLP to compute personal values using a moral values framework, distinguishing between four different personas that assist in categorizing different professions by personal values: “fatherlanders”—valuing tradition and authority, “nerds”—valuing scientific achievements, “spiritualists”—valuing compassion and non-monetary achievements, and “treehuggers”—valuing sustainability and the environment. We collected 200 YouTube videos and podcasts for each professional category of lawyers, academics, athletes, engineers, creatives, managers, and accountants, converting their audio to text. We also categorize these professions by team player personas into “bees”—collaborative creative team players, “ants”—competitive hard workers, and “leeches”—selfish egoists using pre-trained models. We find distinctive personal value profiles for each of our seven professions computed from the words that members of each profession use.