Four SUTD students who took Associate Professor Yow Wei Quin’s course 02.218 Introduction to Psychology in 2019 submitted a research paper to the Singapore Conference on Applied Psychology (SCAP) in 2020 (postponed to 2021 due to covid-19). Their paper has been accepted for presentation! They had started on the project while taking the course, and successfully applied for IDC UROP funding which enabled them to continue working on the project throughout the rest of their studies at SUTD.
Paper Title: Mining Relationships Between Personality, Emoji Usage Patterns and Interpretation
Authors: Yan Linxuan, Wu Shangjing, Chen Xiaoming, Justin Leung, & W.Quin Yow
Abstract: This paper explores the relationships between personality, the Big-5 personality in particular, emoji usage, and emoji interpretation. With the advancements of instant messaging and social media, also known as computer-mediated communication, emoji have become a part of the communication tools people use to convey their ideas and feelings. The current research aims to bridge the gap between emoji usage, emoji interpretation, and personality, the central point of interest in various fields, thus exploring and examining the possibility to build a predictive model of which the input is emoji-related data and the output is the 5 personality scores. A web-based survey was conducted involving 299 English speaking participants, all of which recruited on Amazon Mechanical Turk (www.MTurk.com), where the study was posted as an HIT(Human Intelligence Task) with a brief description of the purpose and procedure. Data regarding their emoji usage, emoji interpretation, demographics and Big-5 personality were collected, wherein for emoji usage, the participants were required to select a usage frequency from a scale of 1 to 7, 1 being “Never” and 7 being “Every Time”. Top 25 used emojis on Twitter were included in the study where usage pattern and interpretation of 13 of them displayed meaningful correlation with personality, as demonstrated in a Kendall’s Tau correlation test; the strongest correlation is between the usage of the “See-no-evil monkey” emoji and a lack of the personality trait of Conscientiousness. Demographics wise, when studying similarity between distributions of demographic groups and emoji usages in a Wilcoxon sum rank test, while gender does not affect emoji usage and emoji interpretation, the more senior age groups use emoji much less frequently than the younger age groups, and should thus be considered independently. Emoji interpretation wise, it is discovered that interpreting emojis as showing “rage” is strongly correlated with the personality trait of Extraversion. Finally, a CATA (Check-All-that-Apply) analysis similarly demonstrated moderate to high associations between emoji interpretation and personality scores. Interrelationships found in this research suggested significant predictive power of both interpretation and usage of emoji in personality scores. Positive results from this paper will potentially serve as the ground for more progressive statistical analysis and machine learning to establish a foundational link between emoji usage and personality score.
As an update, Xiaoming and Justin are working as technology consulting associates in PricewaterhouseCoopers and Ernst & Young respectively. Shangjing has just completed his master degree at Georgia Tech and is returning to Singapore to work. Linxuan about to start medical school at Duke-NUS in July 2020.