Screen Media Use Among Children and Adolescents – Applications of Supervised and Unsupervised Machine Learning and Sentiment Analysis

Abstract

Screen media has become increasingly pervasive in our everyday lives and has profoundly changed the way people communicate and interact with each other. However, we are still unclear about the long-term influence of screen media use on our physical health, mental health, and social wellbeing. Children and adolescents are in an important stage of brain development and are susceptible to the environmental influence that screen media possess. This dissertation pursued three aims to address research gaps related to screen media use among children and adolescents: 1) identify topics and knowledge gaps in screen media use research among children and adolescents in a large corpus of scientific literature using topic modeling 2) leverage machine learning techniques to assist systematic literature review on the association between screen media use and executive functions among children and adolescents. 3) analyze and classify the public sentiments expressed in the text referring to screen media use and screen media parenting on the Twitter platform. In the first aim, we identified some highly researched topics such as the use of social media platforms for health information and communication, screen media use and health behaviors (e.g., physical activity, sleep, smoking, eating, etc.), and computer-assisted intervention for disease diagnosis and treatment. The less explored topics were dietary management, healthcare-related screen media use for injury rehabilitation, and web information and body image concerns. The second aim found that overall screen media activity was inversely associated with executive functions among children and adolescents. However, the results for television and video games were mixed with positive, negative, and null findings when different content and context in these screen media activities and specific executive functions were examined. This work indicated that the association was nuanced by the type of device used and the nature of the interaction. In addition to the first two aims; the third study found that there was slightly more positive sentiment than negative sentiment related to screen media use and screen media parenting. Among the tweets with negative sentiment, video game violence and gun violence had been extendedly discussed and came out as the top concerns for screen media use among children and adolescents.