Audience Analytics of Online Media Organizations: A Cross-Platform and Multi-News Outlet Study Of The Factors Affecting User Engagement of Social Media Content

Abstract

This research, first, provides a four-level framework for user engagement based on the audience’s public expressiveness of engagement. Then, the study tackles five aspects affecting news content engagement: topical, sentiment, emotional, linguistic style, and cross-platform factors. The commonly used topics are then analyzed across five social media platforms at four levels of user engagement for 53 news outlets over eight months. The differences in 53 news organizations’ platform strategies are investigated by focusing on the organization’s topic variations and the corresponding effect on user engagement. Then, predicting the four levels of user engagement for each platform. The sentiments of all posts and comments for each extracted topic and across platforms are measured for analyzing sentiment’s effect on engagement. Next, the relationship between the emotion expressed in social media posts and comments made on these posts is investigated. By measuring nine emotions for eight major news outlets across four social media platforms, we then predict emotional audience reactions both before and after the publishing of the posts. The individual outlets’ linguistic styles are then studied, and the power of different style features on predicting user engagement is eval- uated between two periods separated by three years. Lastly, we investigate the cross-sharing behavior of similar news content into multiple social media platforms and its effectiveness on user engagement prediction. We find that content topics, linguistic style, and emotions differ by both news outlets and social media platforms. We predict a volume of four user engagement levels for given news content, with an 83% maximum average F1-score for the external posting of news articles from one platform to another using language and metadata features. The findings have implications in understanding the complex informational inter- play among social media content, platforms, and audiences and applications such as content delivery platforms, recommender systems, and advertising campaigns.

Date of Award	2021
Original language	American English
Awarding Institution	HBKU College of Science and Engineering