Dynamic Text: Exploring the Relation Between Text Features and Human Emotions

Abstract

Computer-mediated communication (CMC) allows people to interact with each other using various mediums such as text, images, videos, etc. Previous studies have shown that conveying one’s emotion can be difficult in virtual communication due to the lack of non-verbal cues, which often act as crucial components for human emotion during face-to-face interaction. However, Social Information Processing (SIP) theory suggests that text-based CMC has the capacity to convey emotional and social context through verbal cues by facilitating additional information for overcoming the absence of non-verbal cues presented during face-to-face communication. Based on this theory, many studies identify strategies for adapting text-based emotional cues to express feelings in such settings. There has been an ongoing effort to categorize emotional cues in CMC into verbal (e.g., direct emotional words) and nonverbal (e.g., vocal spelling) ones in a multitude of research. One study demonstrates the use of lexical surrogates and altered vocal spellings as potential strategies for emotional expression in text communication. Previous studies also showed different mappings of non-verbal textual symbols to basic human emotion to explore their potential for effective and emotionally rich communication through manipulating textual cues during CMC. Overall, there have been studies that showed supported experiments that manipulation of textual cues has the potential to facilitate effective, emotionally rich communication, leading CMC to serve as a valued medium for both information exchange and emotional sharing.However, despite the continued effort to find a correlation between emotion and text, most of the previous studies have not been able to provide a significant relationship between the two factors without supplying a concrete scientific basis. Therefore, the current research aimed to explore how the manipulation of textual cues affects human emotion and, specifically, which emotions are associated with different text features to which extent by utilizing a more refined and methodological approach that led to an objective and scientific conclusion.The experiment was designed based on the application of Kansei Engineering as the methodological framework, which aims to measure and translate perceived human emotion to quantifiable values. While applying this method, two stages of online surveys and analysis were conducted to assess participants' perceptions of text features. Key findings highlight the effects of Noteworthy font in italics on vibrancy, spacing on comfort, Times New Roman font and boldness on harmony, and its combination with letter case on happiness. The type of font is shown to affect modernness, and Arial font with spacing relates to confidence. Regarding general preferences, Times New Roman in sentence case was most preferred, while in uppercase, it was least preferred. The findings can provide clarity on the way certain text features trigger unique emotional reactions, with potential applications for enhancing the user experience in text-driven virtual environments.