Mossaab Bagdouri. “Journalists and Twitter: A Multidimensional Quantitative Description of Usage Patterns”
Summary:
The last couple years witness a major shift in people’s sources of news. Due to its convenient, instant and interactive nature; Twitter is arising as a major news platform, especially among the younger generations. As a result, there is a demand for understanding the behaviors and strategies of journalists on Twitter. This research aims to answers several questions:
- Do journalists build more personal engagement with their audience than organization broadcast
- Can English journalists’ behaviors be generalized to a broader population with a different cultural, linguistic, and regional background?
- Can new consumers be distinguished from journalists by their behavior on Twitter?
- Do journalists working for different types of news outlet act differently?
- Do journalists speaking the same language but live in different regions share similar behaviors on Twitter?
Thirteen million tweets of 5,358 Twitter accounts of journalists and news organizations and two billion posts from more than one million of their connections are collected. The authors extract eighteen numerical features from these data and conduct Welch and Kolmogorov-Smirnov statistical tests to analyze their distributions across different groups.
The authors discovered a few interesting patterns of behaviors across the groups:
- Organization broadcasts use more formal language and share more links than journalists while journalists prefer to build a more engaging, targeted communication with their audiences.
- The higher frequency of question mark usage also suggests that journalists use Twitter as a source of information.
- Some journalists (Arabic) are more distinguishable from news consumers than others (Arabic).
- Journalists’ behaviors differ across media types.
Reflection:
To begin with, the authors mentioned in the introduction they strike to provide observations that can be generalized to the larger population of journalists rather than a particular group like previous studies. Even though their dataset is massive, the majority of it still belongs to English speaking journalists. Only two tests involve Arabic speaking journalists. Furthermore, they never explain why they choose Arabic instead of other languages and Ireland instead of other British-English speaking regions. Hence, an interesting question is how journalists from other regions act differently to those studied in this research?
In addition, the journalists’ preference of medium varies significantly across the groups. While journalists use mobile for 54.95% of the time, news organizations use desktop and special Twitter application more often. Meanwhile, Arabic journalists use desktop twice as frequent as news consumers. However, print journalists are not fond of the ideas of posting articles via mobile phones. It might be interesting to investigate which factors affect the journalists’ preference of medium used to publish their tweets. Do younger and on-site journalists tend to post via mobile devices? What type of posts are mostly posted via mobile device: breaking news, discussion, questions, etc.? Is there a correlation between the validity of the account and its medium preference? As posting via mobile devices is more convenient and instant, it might suggest a lack of consideration and time investment in the tweets. A perfect example is the President; most of his tweets are posted from his phones. However, fake news is possibly generated and posted by an algorithm thus tweets posted via desktop might have a higher chance of being fake news.
Despite the fact that the behavior of Arabic and English speaking journalists are mostly similar, the audiences’ reactions to their tweets diverge significantly. Arabic journalists’ tweets receive much more retweets and favorites compared to that of news consumers. On the other hand, English journalists receive fewer reactions for their tweets than other news consumers’. What are the causes of this pattern? Can this be because there are more internet figures (celebrities, vloggers, etc.) in the English speaking region than that of the Arabic world?
Finally, one of the biggest pitfalls of this research is that it detects answer-seeking question tweets based solely on the existence of question marks. Even though the authors recognize the potential of misclassification, they decide to go with the simple, naïve approach without providing a justification. Since the usage of rhetorical questions as click baits has been emerging in recent years, the rate of false negative might be quite high, especially with less formal tweets by journalists and news consumers. Therefore, it possibly will provide a more accurate insight if future research can use natural language processing techniques to classify whether a tweet is an answer-seeking question. Moreover, it might be interesting to see if there is a correlation between the use of rhetorical questions and the validity of the news and the reaction of the audience?