Fake news is an increasingly frequent problem and has a negative effect on readers. Horne and Adali used three data sets to try to find distinguishing characteristics of fake news, real news and satire. They looked to answer the question, “Is there any systematic stylistic and other content differences between fake and real news.” The work they did is important since misleading or incorrect information has been found to have a higher potential to become viral. It is important to be able to identify fake news and understand what characteristics set it apart. The authors had the following findings.
-The content of fake and real news is substantially different.
-Titles are a strong differentiating factor.
-Fake content is more closely related to satire than to real.
-Real news persuades through arguments, while fake news persuades through heuristics.
I thought that the paper was well done. The most surprising findings to me were that real news persuades with arguments and fake news persuades through heuristics. The Elaboration Likelihood Model was interesting and future work could be done to further study the effect of fake news on readers and how fake news is perceived. I found it interesting that so much can be determined from the titles of articles. If the intent of fake news publishers is to spread incorrect information, could they change their titles and structure to more closely match real news in order to spread misinformation more effectively? In addition, I thought that the data sets used had limitations and that future work would benefit from more comprehensive data. The authors also acknowledged this. A big challenge seems to be deciding how to classify fake news and real news consistently. Future work could be done to define this further.
