What Prompts Reader Participation in News Articles? The Relationship Between Reader Comments and Exposition, Topic and Columnist
M. Hour, P. Papatla, N. Jalali
The number of comments received by articles is an issue of significant importance to news publishers because it relates to audience engagement and an increase in engagement can increase both advertising and subscription revenues. An increase in comments also results in several additional tactical and strategic payoffs for news publishers. We investigate the relationship between the number of comments on a news article, its expositional structure, and two additional characteristics: (1) how is the article written (the structure)? (2) what is the article about? (3) who is the article written by? Empirically, we investigate these questions on a corpus of 41,210 articles written by 2,239 unique authors on eight digital news sites whose audiences vary in geographic scope (national vs. local), size (the largest of the eight sites has 263 million visits in May 2022 while the smallest has 47 thousand visits during the same month), and interests that varied across 25 distinct topics.
We use a combination of rule-based (Emolex dictionaries) and machine learning methods (Word2Vec) to measure the level of different emotions in each article, its concreteness, certainty, tone, and its verbosity, Google’s NLP API to identify the topic of each article, and Twitter to assess each article’s author(s)’ audience following. Empirically, we use multiple specifications of a Hierarchical Bayesian Poisson Lognormal model to investigate the relationship between the number of comments received by each article in the primary dataset and its associated how, what, and who characteristics.
Our results reveal several relationships between the volume of comments received by a news article and its expositional structure, its topic, and its author. Specifically, the volume is negatively related to two emotions, anticipation and joy, but positively related to surprise and trust. With regard to verbosity, the volume has a positive relationship with more sentences in the article. Volume is also negatively related to two overall expositional characteristics of concreteness and positive tone. Finally, in terms of the author characteristics, perhaps not surprisingly, articles written by authors whose posts on Twitter have been more engaging receive more comments than those by other authors. Also, not surprisingly, volume is higher for some topics like politics, health, and government than other topics.
Keywords: Audience Engagement, News Consumption, Emotions, Natural Language Processing
The Effect of Size of Mobile Apps on User Satisfaction
(Target: Journal of Business Research)
M. Hour, A. Bhatnagar
When app developers add features to or improve the quality of an app, its size increases. While consumers appreciate more features and higher quality, they are likely to dislike that the installation requires additional space and more time. This study investigated about 14,000 applications in the Apple App Store to find out how user satisfaction and rating volume change as app size increases. A downward concave quadratic relationship was found for each model that confirms the existence of an optimum size. We found that this relationship was moderated by two factors, type of app and time. First, this optimum size was smaller for utilitarian apps, which relate to functionality, compared to hedonic apps, which are associated with feelings. Second, we collected and analyzed the data in two waves to determine if the optimum size changes with time. While the relationships were similar, the optimum sizes increased after a year. As a robustness check, we investigated reviews for these apps. Sentiment analysis for reviews regarding the size of the application confirmed our findings for hedonic apps but not for utilitarian ones.
Keywords: Internet of Things (IoT), Apple App Store, file size, rating volume, user satisfaction, sentiment analysis