Tweeting is Believing? Understanding Microblog Credibility Perceptions

Nowadays, Twitter is not only considered as  a social network, but also as a source of news, and thus increasingly is being incorporated into general search engine results. Although searching Twitter for news update can provide users with real time information that are not yet available in the media,  the credibility of the information encountered online is a primary concern.

Authors cite that researchers have specified systems heuristics that influences users perceptions such as assessing the accuracy, authority, objectivity, currency, and coverage of the presented materials. In addition, authors note that according to Pal et al. study, user names of the organisations with more followers had received higher ratings. On the other hand, some other researchers begun building systems to automatically or semiautomatically classify tweet credibility.

This paper presents key elements of the information interface for the impact on credibility judgments by conducting several surveys. The first step was a pilot study that participants were asked to think aloud while reviewing the retrieved tweets about US Senate election. Furthermore, they continued their study by sampling array of Twitter users, and in order to have diverse participants, they sent email to the social media users of Microsoft and a message on the alumni board of Carnegie Mellon University. Results showed that responds were least concerned with credibility for celebrity news and gossip related tweets, and secondly for movie and restaurant reviews. News, political, emergency, and consumer oriented tweets caused greatest concern about credibility.

In order to better understand the impact of various features on creditability perception, two online experiments were carried out in which they systematically altered several properties of tweets. They also had done a t- test to test the impact of their experimental manipulations on creditability ratings plus another similar study to the first one using different image types. So, their results showed that users are concerned about the credibility of content when that content does not come from user followers.

Finally, authors demonstrated that users had difficulty in determining the truthfulness of content and that their judgments were often based on heuristics and biased systematically. Besides, they highlighted that pieces of information deemed helpful to creditability judgments that typically are buried in the interface.

The paper I have selected is entitled ” Why We Twitter: Understanding Microblogging Usage and Communities”.