Understanding retweeting on Twitter

As we try to work out how Iranian citizens, activists, journalists, new media propagators, and politically conscious folks are using Twitter to converse about the Iranian election, we need to step back and think about some of the practices that are core to what’s taking place. One of these is retweeting, or the act of spreading a message along inside Twitter. Earlier this week, Scott Golder, Gilad Lotan, and I just finished a descriptive paper on retweeting as a conversational practice:

Tweet, Tweet, Retweet: Conversational Aspects of Retweeting on Twitter

The purpose of this paper is simple. We wanted to explore retweeting as a conversational practice. In doing so, we highlight just how bloody messy retweeting is. Often, folks who are deeply embedded in the culture think that there are uniform syntax conventions, that everyone knows what they’re doing and agrees on how to do it. We found that this is blatantly untrue. When it comes to retweeting, things get messy. The 140 character constraints introduce new dynamics and people route around a potential limitation is unique ways. But this doesn’t mean that everything is honky dory. There are authorship issues and attribution issues. The fidelity of a message often gets corrupted as it spreads, revealing the ways in which retweeting has become the modern day incarnation of the “Telephone Game.”

This paper is currently under review in an academic setting, but we’re making it available for public commentary and critique. Also, given how confused folks are in the public and mainstream media, we felt that getting this out sooner rather than later might be helpful in clearing up some myths about what’s going on. Retweeting is core to information dissemination on Twitter but how it’s unfolding is more complex than many believe.

Please enjoy! And we welcome any and all feedback!

Print Friendly, PDF & Email

19 thoughts on “Understanding retweeting on Twitter

  1. donnabaratta


    As someone who is new to Twitter, I read your article with the eyes of a novice. I have a few suggestions to make the information more accessible to those unfamiliar with Twitter.

    In the abstract, I suggest you briefly define retweeting for folks unfamilar with Twitter.

    page 2 “there is no ability to comment on individual posts.” No ability implies to the reader that there isn’t an opportunity to comment. I believe what this sentence intends to say is that there is no direct comment opportunity as there is on a blog. Readers do have the ability to comment to a tweet by contacting the person who posted the tweet, or by posting a comment on their own Twitter homepage.

    I enjoyed the article and will pay more attention to RT’s! Fascinating stuff!

    I am addicted…initially I didn’t understand why folks were interested in Twitter, but I have found it to be a tremendous resource.

    Looking forward to your presentation at AASL in November!

  2. Peter Fletcher


    A wonderful article.

    My comments:

    I think there’s something unaddressed/partially addressed about re-tweeting as conversational practice. As “conversations” develop on Twitter – and you’ve pointed this out well – the context of a tweet is lost. This stems in part from Twitter not having threaded conversations and having a limit of 140 characters.

    A conversation between two people that only involves @ replies has a shared context that runs out over time. They are very much more “in the moment” conversations and often make no sense to a person’s followers. The RT convention attempts to add that context back in to a tweet and therefore the conversation. However, these attempts quickly fall foul of the 140 character limit where, to attribute authorship correctly, more and more space is needed for multiple authors, each of whom adding something to the originating tweet. I wonder if this form of retweeting might not make it into your list of ten reasons to retweet.

    A couple of further comments:

    “One interpretation of Twitter’s value derives from
    the real-time nature of the conversations it supports”…Maybe accurate but possibly a little speculative? A more critical view might be that there is little meaningful conversation taking place and a lot of ego-casting. As you’ve pointed out some RT’s are of this nature.

    “”RT @StopAhmadi Bring down Khomeini’s website” with a link to his site;
    shortly later, the site faltered”…A short time later?

    “By retweeting, she brings her audience into a conversation, helping them
    understand the context of what is being discussed before adding her own commentary”…Might it be worth critiquing the architecture of the Twitter conversation? If a restaurant had only one big table at which everyone sat would the conversational dynamics change?

    Thoroughly enjoy your in-depth analysis as always.

  3. R�mi Thibert

    Hello Danah,
    I’ve just read quickly your article (which is longer than a tweet !). I spot a spelling mistake on page 2 (“visibility”).
    There is another thing you didn’t mention which can make the matter worse. I use another microblogging service (Indenti.ca), which is automatically linked to twitter (more people on twitter than on identica). My messages are sometimes retweeted on twitter but they come from identica. I sometimes retweet messages on identica, even if they come from twitter (because I think the content is worth being spread on both channels)
    It would be great if things were more stabilized between the different microblogging services.

    Thanks for your article. I’m waiting for the definitive version to cite it (I wrote an article about microblogging at school : http://tinyurl.com/cp37vb

  4. Niespika

    Hi dannah
    For I dunno what reason, I can’t seem to get you paper to dl.
    Clicking on the link does nothing. Is the link broken ?
    Can’t wait to read it tho.

  5. Niespika

    Hello again,
    Thx to you I finally read the paper. It’s great.
    A few comments tho. I find the last section to be a bit short, I would have liked a bit more discussion around this idea. Is it really a conversation ? How does it compare to other types of conversation ? A critic might have been nice. Of course you could write another paper on the subject, but just a bit more could be nice.

    And there are also a few sentences I didn’t get. But it might come from the fact, that english is not my first language.
    p. 7 section 4.4.1 : I didn’t get why you brought Milgram in the discussion. It doesn’t seem to be related to what comes before or after. Plus It seems to me there is something wrong with the next sentence “Granovetter further points out the cognitive effort required to keep track of not only alters, not to mention the ties between that set of alters, is immense”

    Otherwise, great work. I really enjoyed reading it.
    I really do see what you mean here.

    p.10 (begin of the page) “Ego retweeting is also a way publicly appreciate someone else’s attention (…)” is there a word missing ?

  6. zephoria

    Joe – it’s not about being beloved of academics.. it’s about being required by the conference. I recommend downloading the PDF.

  7. Mark Adams

    Why does your paper feel negative? Looks like you are bemoaning the mess of conversation not realising that conversation, like people, is not ordered. The only useful conclusion I can see is your paper paving the way for “an ordered set of control of Govern how we use Twitter”. Please tell me the point of the paper other than stating the obvious?

  8. Adina Levin

    An interesting followup question regarding the messiness – how bad is the problem of “the broken telephone”.

    Clearly there are some instances where meaning is distorted when retweets are passed on – the tweet that falsely suggested that @eszter’s boyfriend had broken up with her or that All green had dies.

    Out of all retweets, how many become obviously distorted as above, and how many others are misunderstood because of changes (using surveys to test understanding of the changed message).

    My guess is that the strong majority of retweets do convey the original meaning. But that would be an interesting question to answer.

  9. Thothful

    Hi Danah,
    I am reading your paper on retweet phenomenon.
    First, I must congratulate you for the blog as well as the in-depth paper.

    I must, however, point out that you may consider putting in a kind of crisper/standout definition of What exactly is a Retweet, so that its easier for the the uninitiated to get started.


  10. Zohar Manor-Abel

    Great read!!

    Just a little addition – RT can be confusing to new users of Twitter, who are not sure of the jargon they should use. This is why some of the messages have RT and not a username. Also, some of them are requesting others to retweet.

    I remember when I just started on twitter, and I wanted to forward someone else’s tweet, and I just used “retweet @user ABC”, then I saw that others are using RT, which was shorter and made much more sense.

    Also, like you said, I use “(via @user)” only at the end of a tweet, and if I changed, translated (which I do sometimes), or actually heard it directly from the person (i.e. via email or chat).

    Excellent paper, can’t wait for more entertaining writing.

  11. Marco

    Great paper, thanks very much.

    Made me rethink my RT vs. via message strategy and issues of authorship, which really hadn’t come into play for me (especially when forwarding RT @, RT @, RT @ messages!!), as I’m more generally interested in just getting the info/links to my followers. Rethinking now!

    Really great overview and thoughts/questions of an emerging tool.


  12. Anon

    What about “PRT @A: blah” for a partial retweet when some of the original tweet is cut for brevity? Less frequently used, but somewhat standard.

  13. Isabel Walcott Hilborn

    Thanks for the read. At issue for me: I think there should be more explanation of why people bother to type RT and credit the orig author, rather than just tweeting what they just learned. For example, in ordinary (non-twitter) conversation, sometimes people say “Rick told me Laura told him the Iranian election was rigged” but most of the time people just say “the Iranian election was rigged” without bothering to tell people the source(s) of this idea.

    You touch on this when you bring up the shortened URL which, as you point out, proves the user’s access to your article was via the person who shortened that URL.

    There are certainly lots of cases in Twitter where people learn something from one person and decide to pick up the thread without using the text “RT” or any of the other attributions.

    So, an interesting point to me here is why the RT text is used. It affects all the cases you cover. If someone were really just retweeting to show their own audience something new that they thought was cool, why not just point friends to the link? There’s more going on here – people are driven to credit the person who pointed them this direction, and they are driven to do that for a reason. Since characters are at a premium, this drive must be quite strong.

    The reason may be because people don’t want to “plagiarize” – but that wouldn’t apply in cases where users are paraphrasing or altering the original tweet. The reason may be because people really want to give credit where credit is due, a sort of altruism. But in many of the cases you quoted, such as a user retweeting another user’s point to your blog post, the credit isn’t going to a person who actually did the original thinking (i.e. you), it’s just going to someone who read your post (maybe takes some brains to read but come on!)

    Given this, I think that a hypothesis about what is driving people not to retweet (everyone passes on interesting tidbits in conversation) but to credit the person they are retweeting and admit they are not being totally original, should be a key point in this paper. You could search for shortened URLs in much-retweeted tweets and see how many times those shortened URLs came up in other tweets without crediting the source of the link. Then you could make your own judgment whether the non-creditors were complete boors who were pretending to have discovered something on their own, or whether something else was at work.

    Personally I can see how using RT when passing along an idea may be linked to the idea you explored of attention-seeking, wanting to prove that one is part of an “in” crowd, wanting to amplify one’s own voice and therefore amplifying others’ in anticipation of reciprocation, and so forth. Or maybe it’s just a new sensibility that people who are “finders” of the new conversationally useful tidbits play a valuable role in the ecosystem and should be politely recognized? Or a desire to prove one’s authenticity by not claiming credit for a discovery one didn’t make oneself (but how often is any discovery made completely independently? How did that user learn about your blog in the first place? Probably not by running into you on the street, but whoever pointed him there originally isn’t getting any credit…)

    The most positive take I can come up with is that with the rising popularity of open source and the increased connectivity between people who have never met face to face comes a stronger recognition that everything we do online is thanks to someone unrecognized working away, that whatever we can do to build connections will only strengthen us as a community, and that only jerks take credit for other people’s work. On the cynical side, I think there’s a strong recognition among the Twitterati that in the early days of blogging links between the popular blogs solidified the elite bloggers and led to increased search recognition. Twitter is still in early enough days that it’s land grab time — the users who become part of the “in crowd” and who grab the biggest audience share now will be the new cyberelite. Is using RT when passing along an idea, instead of just passing it along, a way to become part of the “in crowd”?

  14. Barry Wellman

    danah, et al.

    I think it is “hunky dory” instead of “honky dory” — let’s not get into racist nomenclature;-)

    Looking forward to reading your article.

    Quick feedback:

    I tweet because I am.
    I retweet because I am networked —
    and want to get good ideas/findings circulated.

    Best, Barry Wellman

  15. Pingback: Consanguinité bloguesque, ou le phénomène de reblogging | Journal de Guillaume Paumier

Comments are closed.