Rhuomeng Zhao will defend her master's project, "A Classification of Adolescent and Young-Adult Twitter Users’ Approaches to Identity Development," today, Wednesday, Oct. 26, at 1 p.m. in room 112 of the Schorr Center. Her advisor was Dr. Lisong Xu and committee members were Dr. Ying Lu and Dr. Hongfeng Yu.
The major motivation of this project is the need to develop a classification system to systematically examine adolescents’ and young adults’ behavioral patterns of using Twitter for identity exploration and expression. Once available, this system can be automated and used to classify a user’s identity-relevant behaviors on social websites. Also, this system can be used to (1) extract behavioral data as an objective measure of the process of online identity formation, and (2) process big data available on social network websites.
This project focused on constructing a classification system consisting of three dimensions of approaches to identity exploration and expression (“self disclosure”, “topic”, and “expressiveness”), which emerged from exploration of user-authored content on Twitter. A sample of 500 adolescent and young-adult Twitter users was selected and manually classified on the three dimensions. Several linguistic features were programmably extracted from the user-authored content (e.g. tweets) of the sample and used with the manual classification results to train machine learning models for automated classification. Several statistical analyses were performed on the classification results and features extracted to (1) describe Twitter users’ identity-relevant behaviors and (2) examine the relationships among these behaviors. It was found users in the sample could be meaningfully classified on the three dimensions and the classification models trained demonstrated very good prediction accuracy. The statistical analyses on classification results and extracted features revealed salient patterns of identity-relevant behaviors on Twitter.
The work of this project contributes to existing research in a number of ways. First, this project proposes a new approach to studying identity development in adolescence and early adulthood by examining the process of identity development in a virtual context. Second, this project utilizes a new methodology to examine the process of identity development by conducting objective observations of actual online behaviors. Last, this project bridges the literature on identity development in the field of psychology and the literature on social network application in the field of computer science.
More details at: https://events.unl.edu/cse/2016/10/26/114848/