Using computational linguistics to understand near-death experiences: Concurrent validity for the Near Death Experience Scale.

Abstract

Latent semantic analysis—a technique to quantify qualitative data—was used on a large dataset of near death experience (NDE) accounts for which a sizable portion also had scores on Greyson’s NDE Scale. Given previous research with the NDE Scale showing there is a “core NDE” comprised of a probabilistic hierarchy of cognitive, affective, transcendental, and paranormal components, we hypothesized that there would be a similar hierarchy of experiential components reflected in NDErs’ verbal accounts as evidenced by a significant relation between NDE intensity and NDE content. Predictions were largely confirmed. The verbal accounts associated with True NDEs, defined by above median scores on the NDE Scale, stood out as highly structured episodes with a clear framework comprised of seven major linguistic clusters of descriptors. Four of the linguistic factors included transcendent or paranormal themes, whereas the remaining three factors tended to focus on both vague and specific references to physiological or environmental elements. The results are argued to support concurrent validity issues with Greyson’s NDE Scale and illustrate how latent semantic analysis (LSA) and related approaches can be used to study qualitative information across topics relevant to the many facets of theory-building in consciousness studies.