Of our sample (300,542 users, collected in May 2009), 80% are followed by or follow at least one user. By comparison, only 60 to 65% of other online social networks' members had at least one friend. This suggests that actual users understand how Twitter works.

We found that 32% of all tweets made by the most active Twitter users were generated by machine bots that posted more than 150 tweets/day. Among the most active Twitter users with more than 50,000 followers, we find singer Tyrese (@tyrese4real), actress Alyssa Milano (@alyssa_milano), celebrity Tila Tequila (@officialtila), tv host Jonathan Ross (@wossy) and evangelist Guy Kawasaki (@guykawasaki).

On average, Twitter users have five degrees of separation between each other - meaning nearly everyone within Twitter is only five steps away. Of all friendship distances, five steps is the most common (41%), while a friendship distance of four steps is the second-most common (37%).

Simply because a follower has a low Twitter authority, doesn't mean they aren't worth connecting with. The opposite holds true for followers with high authority levels; simply because they have a high Twitter authority, doesn't make them the right influencer for everything.

After analyzing information disclosed on 11.5 million Twitters accounts, we discovered that: 72.5% of all users joining during the first five months of 2009, 85.3% of all Twitter users post less than one update/day, 21% of users have never posted a Tweet, 93.6% of users have less than 100 followers, while 92.4% follow less than 100 people.

“Informers” are more likely to post messages that share information (such as news links), where “meformers” tend to focus on what the researchers call “Me Now” messages – posts that update a user’s followers about that users thoughts, location or immediate situation.

Mor Naaman and Jeffrey Boase found that 80 percent of regular users are “meformers,” people who use the platform to post updates on their everyday activities, social lives, feelings, thoughts and emotions. The remaining 20 percent – “informers” – share information, have larger social networks and are more interactive with their followers.

Pear Analytics analyzed 2,000 tweets (originating from the US and in English) over a 2-week period in August 2009 from 11:00a to 5:00p (CST) and separated them into six categories: Pointless babble ( 41%), Conversational (38%), Pass-along value (9%), Self-promotion (6%), Spam (4%), News (4%).

A research report by Danah Boyd (Microsoft Research), Scott Golder (Cornell / Microsoft Research), Gilad Lotan (Microsoft). http://www.danah.org/papers/TweetTweetRetweet.pdf

Our list of 100 tips, tools and resources for Twitter research will help you keep it all together.