Abstract. Moral exemplars contribute significantly to human morality but are often unavailable for direct study. Computational semantic analysis of exemplar interviews and autobiographical text enhances manual coding for longer texts and comparison across multiple theories. We investigated the underlying prevalence of moral values and ethical orientations within transcribed interviews. Using latent semantic analysis (LSA), we develop probe documents based upon the Portrait Value Questionnaire (PVQ) and triune ethics meta-theory (TEM) to analyze transcribed interviews of those who rescued Jewish people and others from the Nazi Holocaust. The resulting analysis measures latent presence of moral values (e.g., security, universalism) and moral orientations (e.g., self-protectionism, communal imagination). Preliminary results demonstrate promise for identifying and relating moral constructs. 

Full poster (pdf)