Evaluating comforting messages with the sentiment analysis and social cognition engine

Abstract

The current study extended the literature on social support by examining the differences in the word content of supportive messages using a free online linguistic tool called sentiment analysis and social cognition engine (SEANCE) across multiple levels. Participants created supportive messages in response to vignettes of distressed individuals, which were evaluated with SEANCE for word content and coded by researchers for person-centeredness. Word content was significantly different by level of person-centeredness, biological sex of support provider, and the recipient of support. The results suggest that tools such as SEANCE may help to identify levels of person-centeredness beyond that of a trained coder and that future research should examine the individual differences of support provider and recipient when evaluating message content.