- Only 36% of rated tweets were considered "Worth Reading". 25% were rated "Not Worth Reading" and the remaining 39% were considered to be "OK".
- Questions to Followers, Information Sharing, and Self Promotion were the content types most likely to receive "Worth Reading" ratings.
- Presence Maintenance (brief, salutary postings), Conversation, and Me Now (the tweeter’s current status) were the content categories most likely to receive "Not Worth Reading" ratings
- Raters almost universally disliked overuse of location check-in tweets from services like Foursquare, inclusion of @mentions in tweets when a direct reply would have worked, and presence of multiple hashtags that obscured content
- "Informative" was the most widely selected positive adjective, although both "Funny" and "Useful" were also frequently selected. On the negative side, "Boring" seemed to be the reason behind most "Not Worth Reading" ratings (81%)
The researchers concluded that the raters in their study valued Twitter as an information medium -
Tweets worth reading were often informative (48%) or funny (24%), as seen in (the figure above). These tags had very little overlap: a tweet was often one or the other, but not both. Information links were valued for novelty or an appealing description: “interesting perspective on something I know nothing about”, “makes you want to know more.” Humor was a successful way to share random thoughts or opinions especially: “it’s witty and snarky. worth the read.” In keeping with Twitter’s focus on short messages, followers appreciated conciseness: “few words to say much, very clear.” A human aspect was also appreciated: “personal, honest and transparent.”Sources - Who Gives a Tweet? Evaluating Microblog Content Value, report at Journalist's Resource blog.
Who Gives a Tweet? Evaluating Microblog Content Value, research paper by Paul Andre, Michael S. Bernstein, and Kurt Luther.