It’s day two at the Lancaster Twitter and Microblogging Conference. As I did yesterday I will be live blogging the parallel sessions I attend, the plenaries, etc. But much of the conversation around this event is to be found on Twitter on the #lutwit hashtag.
Nordiana’s research looks at the use of microblogging and the factors in that use and the complex issues around adoption and use.
My work specifically looks at the use of Twitter and particularly how Twitter can increase or enhance teaching and learning (Dunlap and Lowenthal 2009); classroom use (Junco et al 2011) student use (Wakefield et al 2010) etc. There is a qualitative study of academic tweets (G. Velatsianos 2011) found that scholars engaged education and sharing best practices, used it for information resource and media sharing, to have an online pesence etc…
But Selwyn (2011) highlights some issues around social media use: the assumptions about omnipresences/ubiquity of internet access and the digital divide; belief of “digital student”, that the students will explore and use these tools but not all will, some may struggle with use; And there is unclear discussion in terms of social media in relation to learning and teaching contexts.
The literature here is growing but evidence still lacking. See Reuben 2008 for the potential in education for Facebook and YouTube but Twitter hasn’t yet found the right niche.
My first research questions here is around the different ways that academics use Twitter. This will be investigated through a qualitative approach to obtain detailed understanding of the use of “Twitter of Academics”. The research sample will be academics in HE with a Twitter profile and regular microblogging activity (defined here as once a week). There will be a thematic analysis using the Twitter API. The theoretical framework for this work is practice theory – mapping academics ongoing interactions as revealed through recurrent practices, the concept of the “practice lens”; and academics practies of Twitter – may be conceived as a continuum in which activities dynamically change influences. The research will begin with interviews, transcribe and analyse three of these; then observation of Twitter (300 tweets); and observation of Twitter and also using sentiment analysis (of those posting).
Interviews will have thematic analysis (inductive and deductive). The observation of twitter posting qualitative content analysis, deductive approach. Both will be combined to gain a fuller understanding of use and factors. Sentiment analysis (see Pang and Lee 2008).
There are some categories of activities which academics claim they use Twitter for. Conversation for instance has been selected by them, I have defined sub themes of:question and response; for opinion; for update activity. Information and sharing breaks down into research/publication; quick information; links information/news (retweet). Engagement breaks down into student; research community; public. Connecting and networking breaks down into: professional; social; community; research collaboration; seek opportunity. Identity breaks down: professionalism/profile; online presence; self presentation. Learning and Self promotion also sub categorised here. I have also categorised tweets into status; conversation; sharing resources; social.
So the discussion is about so many differing roles and activities. There are many faces of the “new” academic – a real balance to be struck around all areas of role, public engagement and impact agenda, and of course teaching, research, and opportunities.
I have created a research model of factors influencing academic use of microblogging – academic identity; technology affordances; public engagement, etc.
Q) Monica Lalanda, doctor in US: was it hard to find that number of academics using Twitter?
A) At this point – at the beginning of the research – it was hard to find academics using Twitter. But this summer it became clear that there were far more academics on Twitter. I think this is a good time for this research. I’m not sure about percentage of usage at the moment. Some use it mainly for teaching. Some use for publishing only. Academics are starting to engage on Twitter. Lots of training and promotion around use of these tool.
Chair) There are some studies in the UK of use of Twitter by academics – see LSE Impact Blog.
Comment) I’m sure there are a lot of prejudices, many don’t see the potential, are concerned about the timing.
A) One of my interviewees is a doctor who is very happy to be able to update colleagues and patients in what he does. He has gained patients through his use of Twitter. But I was quite surprised at his usage. At the beginning I am quite surprised at this use but he is engaged in community and research community and he found organisations and media have asked him for views because of his presence.
Q) Sentiment analysis – can you say a bit more?
A) This is a new area in some ways. Analysis tends to be on content rather than sentiment. I want to explore what they are saying on Twitter and how that relates to what they actually feel, what that relationship that. Many of the academics I interviewed don’t want to enter arguements on Twitter, they don’t want to impact their own or their institution’s reputation.
And we had a diversion there about the backchannel and tone… (not appropriate to amplify but challenging) back to the questions though:
Q) Are there particular characteristics of how academics tweet compared to how others tweet.
A) Interesting question. There are real contrasts between different academic Twitter users.
Chair) Offering to share some sentiment analysis work on corporations.
This piece of work looked at the phenomenon of tweet chats on particular professions and interests, usually regular and they vary a lot. e.g. lrnchat; innochat; edchat; PhDchat. The topic is usually selected in advance and the actual tweeting is within a time limit, usually 60 to 90 minutes. All organised around the hashtag. There are some people who always attend at the core here. Some teachers described these as their main professional learning activity.
I wanted to explore how professional practices are being “talked in to being” in discussion events held in an open online environment, I particularly looked at Human Resources professionals. I looked at three events on hashtags here over three months. These vary in how many participants attend (between 54 and 72), some had 10 tweets per minute, some less. But this stuff isn’t easily constrained. Schneider and Foot (2005) describe this issue of web spheres – the bringing in of other resources, chats, tweets, etc. So looking at a blog about the event provides additional context. One participant decided to recontextualise their Twitter contributions in a blog post. And you see comments that there are poeple outside of the chat who follow up, ask questions, and a blog post has been used to address that. And there is an example here of someone correcting themself – for own happiness as much as those who may see the Tweets.
So these events are hugely expansive, you have to cut the cloth as best you can. I focused on the event archive and then started to explore outward until I reached the point where time no longer allowed. I approached the data in terms of discourse analysis, using a division that Heracleoud (2006) came up with. This splits discourse into Communicative action; discursive structure – this is the use of shorthand etc .; generating common meaning – a shared understanding of human resources here which was required to join in, to be part of this event. A particular problem here was “what is human resource development”. It’s really not well understood at all, real split and shorthand in academia around US approach (v. corporate and profit orientated) vs European (self empowering ideas). But that doesn’t work McKendrick suggests a hologram metaphor here. But as a professional body of practice there is no standard approach, it’s fluid and contested. There is no manual. But early stage researchers, professionals, customers etc. expect there to be a standard and professional approach.
It did prove very difficult to make sense of this all. I decided to draw on ANT and translation, different perspectives on the hologram of practice, to try to develop networks of people who agreed on translation. Those network assemblages reinforce that idea of professional structure.
I actually started off using the structure of these chats, using a structure designed for unstructured group work in classrooms. So here we see an initiation, then a string of suggestions or propositions to get conversation going. One suggestion dies out fast, another gets limited interaction, another gets little, another becomes complex and connected… it becomes hard to trace. And things migrate off into discussions entirely unrelated to the chat but using the same hashtag. Indeed you sometimes see tweets asking for help in dealing with what’s going on, what that structure is. Simpson (2005/6) talks about conversational flaws. A retranslation of a topic that allows the discussion to flow. Many of these suggestions and propositions with different levels of success in the example I’ve shown actually came from a single individual. But capture of the conversational flaw has to be accepted by the audience, it’s a two way thing, not just projection of power in itself.
There is some thematic structure that comes out of these discursive networks. And some interesting behaviours. A couple of examples here of participants dismissing topics. Kirkpatrick is a widely used model in this field, 98% of businesses use this model and yet these Twitter users mock it or treat it as a drinking game. Similarly classroom off-the-job training is treated as an irrelevant, old fashioned, dehumanising practice. Is made to look ridiculous through sharing of images of Victorian classrooms.
So I came up with these three areas of discussion here. Change – and change as being in deficit because of Managament & “the business”, and Human Resouce Development. But this mix was constantly reassembled and changed depending on who was speaking, what the topic was. So, the symmetry thing. The people organising and engaged in the chat recommend use of third party app, like Tweetdeck. So you get a multi column view and the same tweet can reoccur in separate places. For participants these chats are seen as part of much wider community of practice which they are involved with. So they have a single column for event but eyes on other things… and that all starts to merge as they take part in the event. So the technology changes how that event is consumed. It appears to potentially have some impact on how that discussion is shaped, which utterances are priviledged by which users.
So, the piper and tune… it changes all the time. It gets redefined during the events and over multiple events. The hologram is both restrictive and expansive – you have to see it from one perspective but has to be seen by multiple perspectives.
Q) Norreen Dunnett: Did you get a sense around these hashtags that they felt like a community or was it more permeable than that?
A) I think there is a core group that sees themselves as a community, they see each other as different from others. e.g. in introductions they present themselves differently. And they try to capture conversations more often. Others are drawn in quite a bit. The core group are very similar and there is overlap between several hashtags (we looked at several) and different dynamics on different days and different times, some opportunity for social network analysis here. A slightly weird mix of community and permeable.
Q) I was really interested in your comments around the difficulty of sense making, of constraining data collected, of knowing how to deal with archives etc. But I don’t quite get the hologram metaphor, can you say a bit more.
A) McGoldrick talks about it as a reconciliation of contradictions around this. The idea of HRD as management and business discipline, about growing or improving the business, but at the same time it has the aspect of development, a discourse about developing employees, about learning and empowerment of individuals. The Hologram idea recognises all of these elements. Where are someone like Monica Lee says that we just shouldn’t even try to reconcile those factors. The metaphor is about looking in and seeing what you want to see, but also having capacity to see other things, that by shifting your gaze you can see those elements.
Q) Karin: Could I ask you a little bit more about the use of ANT and what it added in terms of your conclusions?
A) It’s a good question. I’m not sure what it adds to be honest. Other than giving a loose framework to hang ideas off. The idea of network assemblage works well, the idea of entry criteria (the Kirkpatrick issues) but also the symmetry of the material. I did some work on community managers and relationship between Twitter chat event and a presentation, that they moved from cynical position to enthusiast through chatter. But this work is part of a wider piece of work where ANT will be used to some extent.
Julia is introducing Greg and referring to his book, The Discourse of Blogs and Wikis – it visibly perks up her Understanding Media students apparently!
There is, btw, a real life handout! https://pbs.twimg.com/media/BHkEFg_CEAAG2Gu.jpg:large
I want to look at 8 tweets. This first is an Aprils Fools joke about Twitter. Mars Rover behaving like a celebrity on Twitter, sulking out of Twitter. So I want to talk about different kinds of Twitter feeds, different communities, different behaviours etc. Much of the quantitative research we do and cite looks at big stream of data without any reference to differences. But I know many of you here are focusing on qualitative aspects. Back when I wrote that book that Julia mentioned blogs were being treated the same, and that seemed so lacking in understanding of their distinction. So I want to basically ask… is Twitter really one genre?
Looking across the programme we have use in the Lords, tech companies, EU, academics, cricket fans etc…. can that be one genre? These are communities use Twitter for different activities. They do develop different kinds of texts, for instance, more or fewer RTs, URLs, replies. Even if they use some o fthe same feaures such as conversational particles. What are these questions and do those differences matter in our research?
So, why study Science Twitter? Well it’s one community: research scientists (there are other people who tweet about science of course). They are a great community as they have been good at networking since the 1660s. But they are an odd community for Twitter as they work in a system that rewards formal publication system, there is a divergence there. Networks of texts, also samples, people, skills, equipment, methods, money (see ANT). And Scientometrics links made by citations – impact comes from the idea that science is measurable in these ways.
There are two themes from Science and Technology Studies I’d like to bring in. Firstly the heterogenity if scientific networks – texts, materials, equipment, skills, publics, money (ANT, Callon, Latour, Law). But at the same time Rhetorical tensions between empiricist repertoire, impersonal and timeless claims in the formal literature <missed ref>.
But there is a huge volume of prior work on scientific texts, those working in science and technology studies would say what I’m saying about Twitter is really not new. So if we look at a quote from Henry de la Beche to Adam Sedgewick in 1834, published by Rudwick in 1985. It’s a very tweet-like exchange… BUT it takes place in letters and only becomes public when published 100 years later. And that is difference.
So in my handout you will see sample tweets. I have a corpus of tweets from Scientists, and a cohort of comparison tweets. See thelanguageofblogs.typepad.com. You might do concordance analyse such a corpus for topics – and here you will find keywords around science, research papers etc. But there are also other keywords that are telling. e.g.
- But also I (less elision) – subjectivity indicated compared to other tweets)
- of (more complex NPs) – more like other science writing
- but (concession structure) – perhaps this is what makes an academic an academic!
- may, maybe, some (hedged statements)
- and a negative keyword: love – they just don’t use those sorts of terms, they evaluate in different ways.
So there are empirical grounds for seeing this community as a distinct community, but they also present themselves as a distinct community as well.
So, onto Tweet two! These tweets are at 10.30 at night – a “solidarity check in” to ask if anyone else is still working. Gets very playful responses. So this is “phatic communication” – that is communication solely for the sake of contact, not sharing information:
- “still” implies this is late
- “#GoTeam” parallels this to other shared projects (Merkhofer, Zappavigna)
- “#ThursdayNightScience” invented for their shared activity
- “Woot” – online gaming term
- But where is “here” – well they are not in the same geographic place. The obvious inference being is everyone online, but also the less obvious inference is “in the lab” and this idea of solidarity.
And now to turn to Tweet Three. This is a series of tweets about a fieldtrip. Two kinds of time in these tweets. From inside “headed out” (current action), “my spring break” (current period defined in terms of work), projected future contacts. And from outside “early to mid Pleistocene” etc. Once he goes out he has limited mobile reception. We see a few images shared but of his campside at dawn, not his work but a sense of that experience.
Now to his return… tweets about a late flight. So more complex time. Present moment looking back to immediate past (nap) and forward to immediate future (this week’s lecture) and with very few verbs.
So, why do time references matter? Well they present themseleves as a community sharing norms, focusing on work, they have the shared ideas of terms, of routine, of publications. This is what everyone else does on Twitter but these references to time are a different representation of science work from that in journal articles or popularisations.
But there are more unique things here. Science tweets link A LOT and comment on them a lot. So here we have an example of something of a takedown of @drphilhammond’s tweet about children and screen usage. A response comes back with citation. response back from experimental psychologist criticising sampling basis and link to blog post. And others join in and say it was a mistake that that original cited article was published.
So the first post takes for granted that stating those two facts will be uncontroversial. Response uses irony to criticise. The response with textual citation (not a link). Response questions the cited work using a link. So reference does not settle the matter (unlike Wikipedia). And a real sense of reaction to this media person representing mediacal science (representation regress? – Collins).
Onto tweet 7, a chat about bafflement. It’s a kind of criticism. Sort of self-criticism of her own bafflement with technical term and time of day. But suggests something else to criticise there. See also LOLCAT on term “Thermodynamics” also representing confusion.
And our final tweet to the tag #chemophobia. This is about expertise. So an urban ecologist and science outreach person in Ohio. She doesn’t know about organic chemistry but has read something on a food blog about chemicals. Used #chemophobia tag, used for fear of chemistry going wrong etc. rather than toxicology. Asked for someone to fact check. Delayed response but eventually gets a response with whether this reaction occurs, does it exist, etc. Comes back with paper. Then she responds saying “but both chemicals in apples, naturally occurring”. SO I’m interested both in the intersection here between science and non science, but also science and science in another area to your own. So the original questionner is not a chemist and raises a question that acknowledges that. A (non-addressed) chemist responds with brief unsupported evaluation but she comes back with her own analysis to that.
So… two kinds of evaluation across these samples. There is personal stance (e.g. Yup, 7am, wildly wrong). And there is impersonal reference to shared norms of methods, citations, rhetoric, publication – the idea that critiquing an article that breaches norms settles the issues. And also hierachies of authority worked out for present purposes in the exchanges – this is where specific mentions become so interesting here, the idea of certain individuals as authoritative sources.
The implications for science stufies cover two themes: embedding of science in everyday life; rhetorical application of norms of evaluation to texts. Non-scientists would get useful (and entertaining) view of science by reading these tweets. It contextualises science in everyday life and work. But that message probably doesn’t get out. I was struck yesterday by the idea of a bubble around the European Commission. I think some of the same here, the tweets tend to go to scientists, science communicators and science educators. There is a large and somewhere sealed off world here. Almost no replies from outside. And many of the tweets are concerned with boundary work. They both open science up but also maintaining it’s distinctness, it’s inaccessibility.
But other questions here… If you do see this as a community, how does this compare to other communities studied here. This is a community that thinks of itself as a community. How distinct is the genre? Do they use affordances differently to others? Do they have different practices or simply a different register, different works because they talk about the same thing. And how does that relate to other kinds of practices. No point to study in detail unless it relates backs to other things. The people I look at are teaching, writing articles, outreach. Time patterns fit that. Other communities have different time cycles to them. In my sample I chose widely followed people. Between 3000 and 10,000 followers. And for most of us that sounds pretty good, particularly if you are tweeting about obscure aspects of astrophysics. Their authority on Twitter is about what they do there. Some are authoritative in their field but many are not as influential outside of Twitter. And finally how permeable are the boundaries of the community? Not very perhaps. Some other communities may reach out more, particularly in terms of followers and retweets.
And finally a picture… A fluffy toy and his genetic biologist look alike.
Q) Ruth: My question is simple and small. You spoke about the corpus, what was your reference corpus?
A) I chose a corpus from another specific group, rather than all of Twitter. I chose 10 scientists and I chose reference corpus of 10 others tweeting on very specific topics (similar number of followers but all tweeting on differing topics). Roughly conmparable. For most recent samples I used the same people.
Q) Me: You talked about the exclusivity
A) For some they are. One is into women in science and feminism. Another is into hip hop (and science), so lots on music. So they tweet on other topics but they seem to have an exclusive type of engagement and response on their science tweets. I suspect that many of them have lots of followers because of those other interests but that mixture of interest isn’t represented by crossover of audiences responding.
Q) How do you define “community” if at all?
A) These individuals refer to it as a community, e.g. “ThursdayNightScience”. Now at conferences they are very specific but the community they refer to communally here it is all about belonging to this giant scientific community – which doesn’t really exist elsewhere and doesn’t respect traditional hierachies. Quantitatively you could see the links between them to see a nice graph.
Q) Sean, Lancaster: You talked about the boundary work around expertise. But could you see these as breaking a boundary, reaching beyond expertise to others, as in TV shows and spoilers in Jenkin’s Convergence Culture?
A) Maybe not about hierachy. It’s not about this person has a right to speak, and this one doesn’t. Quite often people ask for help from anyone on Twitter, and get a fast response/advice. You are calling this a boundary hybridity, that’s probably fair enough. Thing you see in ANT all the time. Boundary object that means different things to different actors but both can use it in an interesting way.
Q) Monica: I belong to scientist group which is more active than most. I wanted to ask your opinion about what is happening in Britain, the GMC has new social media guidelines for doctors which does not allow them to have anonymous accounts because of the privacy of information they hold. There is protest around that, a petition about that. And that is reasonable on commentary about economic cuts etc.
A) I haven’t studied doctors so not aware of professional structures but I think it’s a shame not to have the possibility of anonymity. Only one of those I am looking at is anonymous @scicurious but that allows for lots of playfulness. There have been a lot of articles where Twitter has acted as critique of peer review process, taken down research in hours. No less than Dorothy Fisher has said, ok Twitter may have a role here. Anonymity does seem to have a useful role here.
And now for lunch…!
And we’re back…
I want to introduce a cricket journalist called Jonathan Agnew (@Aggerscricket). My work draws on David Barton’s work of “technobiography”, a very socio-cultural dynamic view. Agnew is migrating from the role of journalist and public person. It is framed as a personal Twitter account, views his own, but his website is the BBC and there are lots of pointers that this isn’t a regular Twitter user.
I’m taking a media ecology view (Barton 2007 – ecological view of literacy) here but I’m quite critical of it at the same time. Postman 1970 sees it as a sort of moral issue. Nystrom 1973 sees this as a study of complex communications systems and environments, interested in interactions, technology, technique and human process of emotion. And <another key ref missed>.
In 2006 the UK Parliament decided that television coverage of international cricket should be on free to air TV. But the rights don’t always go that way. Sky have the rights for many of the live TV coverage leaving the BBC with other means, predominently radio. A key thing to say about cricket. I am talking only about Test Cricket, which takes place over five days, frequently ends in a draw, and much of the time very little is happening and there are unscheduled breaks. And that could mean dead air… so there is a great traditional of literary coverage of cricket, it’s about much more than the sport.
The cricket media ecology… start with The Economist and a comment about surprise that the Test Match Cricket on the radio was still running, and that it is now on the web. But I wanted to dig more. To see what attitudes Agnew displays towards Twitter including relations to other communications, I’ll be talking about my methodology as well.
So I have taken a sensitive ethnographic type of approach. I started in March 2010. There were interactions on the website. There were some sample tweets. And then in 2011 Agnew wrote a book and he was also enthusiastic about Twitter so I analysed his media coverage. So I did some intense data collection. Starting on 10th August 2011 I collected all tweets in real time, who he replied to etc. preserved in a word document. Also looked across other media. Some other samples here and there. In part my approach was flexible and did change as I collected through it. For example… Agnew tweeted “15 mins to live chat” so I went and joined that and recorded that to see what that was like. And one of the things that was interesting there was seeing attitudes towards Twitter expressed in other media, and how those media related to it. So he receives a direct question about Twitter. He says he enjoys it, he likes comments during his radio commentary. But also implicit references for those also on Twitter. Makes a comment about his dog in the sign off… if you only interact in other media it’s not noteworthy. But if you follow him on Twitter you know he plays with his dog on Twitter – interacting with an account someone else set up for his dog.
So I’ve coded all of Agnew mentions on media in the book. It’s ostensibly about cricket but he is very interested in media. Radio gets the most mentions but, notably, Twitter gets the second most mentions. At the time “Aggers’ Ashes” was promoted via what he called “Twitter tour”. Related to that the @theashes follower situation arises and he melds that offline persona to the online persona of that Tweeter. He also playfully recommends follow another Twitter user who does a better job than him. He doesn’t do that. When I followed Agnew on Twitter in real time he went from 7am to 8pm (he barely stopped for lunch). So I followed him on 10th August 2011, the first day of a test match but the day after the riots. One of his followers tweets about the weather, also about the riots. His response *just* addresses the weather. Listening to the radio commentary it becomes clear they have been told not to mention the riots. But they elude to it, you wouldn’t understand comments like “I can see smoke in the distance”, for instance, without that context. And you see him respond in an authoritative way about the ECB confirming test match goes ahead as planned (meaning both weather and riots).
He does engage in arguement sometimes but, for instance, shuts down someone who tweets what seems to be a homophobic insults. Bourdieu (1999) takls about “difficult spots”: “difficult to describe and think about”. Only looking at a much larger quantity of data did I realise that this hadn’t been a homophobic comment but a reference to a co-commentators criticism of another team.
There are many ways Agnew involves others in his construction of stories. So pulling out the key tweets around “Moussaka Special”. He comments “treating the wife to Moussaka Surprise. Theory being devastation in the kitchen means I won’t be asked to cook again for at least 6 months” – refers to specific type of comedy. “The Wife” is a term he doesn’t usually use so that’s a reference to a type of comedy. The narrative builds. tension mounts… then two surprise tweets, the moussaka is good, and she tweets too and joins that narrative!
So really interesting narrative construction; and merging of online and offline.
Agnew did leave tweet for a while. He gets lots of abuse but attack by another author seems to be what drives him off Twitter for a while. I’ve used this idea of “change agent” (Mullins, Kozlowski, Schmitt Howell 2008) as he is quite influential in this rather traditional cricket world, and he has that trust to be that change agent, something you see more widely in adopting in Twitter. And you see these realms of onstage and offstage personae and performances on Twitter and intersections between them.
Q) Did you tell him you were doing this?
A) I did, he didn’t acknowledge it – sort of said a few times but I’m there in a huge (200k) audience. I’m still planning to send fully written up version to him for comment.
Q) You talked about Agnew moving from journalist to public figures. As you went through data collection did you see marked difference in tone of Tweets as follower numbers grow.
A) I’d say no actually. But by beginning of 2010 he was already well on the way, his role attracts a certain amount of attention anyway. He was already a skilled user of Twitter, building stories, orientating to audiences, some moments of tension, the abandonment. But interplay generally there throughout. But will think about that more carefully.
Q) One of the things I thought between celebrity and “real people’s” use of Twitter was how much they retweet them. One thing you could look at is how much they retweet, distribution phenomenon. Also terms of address that vary between celebrities and others. Celebrities address a collective group of followers.
A) I’m not sure he is really a celebrity here. He’s more a personality.
Comment) I’d say microcelebrity eg Alice Marwick
A) Certainly he has a lot of interactions, requests to retweets, lots of iteraction. Your comment does give me an angle. But I’m not sure celebrity is the right work.
Comment) Perhaps about tipping point here, not about celebrity status but number of followers.
A) I think Agnew is about driving change etc. He did start encouraging web 2.0 use because you could speak to everyone but actually he has become someone who extends the broadcast models.
Comment) He is followed by far far more people than he follows, classic broadcast/transmit model. Reach is the wrong term or concept but that broadcast behaviour is something that is different.
A) yes, I do make that comment in my conclusions but that’s a fair observation.
And now its me so the liveblogging stops!
The Prezi is here and I highly welcome comments!
But the in-person version sparked lots of questions so I think it went well and hope some attendees pass on their own use cases to the Twitter Workbench team (I’m happy to pass comments on!).
Like a Boss is parody rap. Everything if followed by “Like a Boss”. Massively popular, huge amounts of tweets about this. A tweet evergreen if you like. Refers to finesse or authority. Or just their way. People, animals, objects. e.g. Many presenters have presented Like a Boss at this conference!
Leadership discourse. From management and business studies we know that leadership is about meaning. There is the transformational leader, the charasmatic leader, visionary leader. Often quite gendered as well. There is a lovely paper by Baxter in Journal of Social Linguistics in how female leaders shift in and out of various “role traps”. The Like a Boss phenonemon ties into these aspects.
Vernacular writing is an area of huge interest, this is inter personal in nature, spontaneous and unplanned, dialogical by default. The Like a Boss hashtag and tweets displays some of these. So an example here – a micromeme here. But we see lots of sharing of quite mundane achievements. Also identity as heritage discourses (around gold star stickers). Also performances of being a boss, judgements of enoughness. Now vernacular writing, why are we drawn to it? Well the rational is pretty straight forward. More people write than ever before. They write more. Digital media enables that. And as we all saw yesterday during weird Q&A session writing goes public. And finally my interest here is about identity practices. Here we see really diverse identity practices and the use of “emblematic resources” <ref?>, and an enormous range of features that could see you doing things “Like a Boss”. From buying shoes, getting out of bed, high-fiving a shark. But the meme does cohere and what makes that happen is that these are forms of self-presentation and performance – identity-as-heritage (cultural capital – like the star stickers). And identity-as-creative-play – about subverting the norms, creative play. Cue discussion of a dog on a wingback chair – why is it funny? It’s anthropomorphised, it’s a superior expression somehow, pokes fun at self-styled ways. But I’m really interested in how that meta commentary, how far that ranges. Some of these self-performances don’t work. What is the benchmark for being or doing Like a Boss does. About “enoughness” – the resources recognised as successful. Not fully developed but… online, on YouTube, views and comments etc. index what it takes to be a Boss. That measure is policed online constantly. Ideally we want an indexical range of how you perform this successfully.
In conclusion. We approach digital communication in an ethnographic perspective. We look at vernacular cultures and cultural politics. We look at leadership discourse – management of meaning and language game. We look at vernacular writing – networked writing, scale. and we look at identity practices – enoughness, self-presentation.
Q) Isn’t self-tagging yourself “Like a Boss” a form of Index. So the measure is whether accept it as such?
A) Yes, the tweets identify successes in being “Like a Boss”. There are wannabees. The data we have… the notion of enoughness divides those who can (e.g. retweets) and those who want to.
Q) In what way is Like a Boss different from Like a Pro? In terms of semantic features.
A) Like a Boss calls on more cultural toughness etc.
Q) Can I ask about “enoughness”… if you use rankings, popularity, as a measure…?
A) As one measure…
Q cont.) So how do you measure that they really like that. And secondly what other measures
A) No answer yet.
Q) RT as a measure of goodness for a tweet. About variation of users. theoretically scale free in Twitter. 30 RTs would be the best day of my life for me, for some that’s every day. How do you account on that?
A) We don’t want to look at celebrities. Only normal folk.
Q cont) But a continuum there. Some random people have 20k followers you know?
A) Imagine we will, and others have done, look only at a particular range of followership.
Me) Compare RTs for other stuff – so does the Like a Boss stuff
Q) Vernacular use of Boss is very different isn’t it? Gaming connotations? Slang connotations? Not just leadership here? Maybe why more sticky as a meme
A) Sure, those come in.
Comment) Like a Pro just not as grabbing, those other uses include “that’s Boss!” a very 70s and 80s style.
Discussion breaking out around whether “Like a Boss” is a valid tag for some tweets in Tom’s data. And then on Pro vs Boss. Boss being more contentious, declaring power over others, superiority, dominance etc. And that others don’t care how their seen – cockiness, single mindedness, self-belief…. and now into the “great man” theory of leadership. Also discussion of ironic use.
Tom: I have been able to collect different things here. Initially tweets but that is harder with the API. Collected some
There are three players in this twitter research: politicians, media and citizens. I’m going to specifically look at how politicians behave in social media. Today I’ll look at the UK General Election but we’ve also done work on Dutch elections and on-election periods.
We took a sample from a 2 week period (April 26th to polling day, May 6th 2010). We focused on main three parties. Any candidate who tweeted in this time. 19% of conservative candidates; 22% of labour candidates; 26% of Lib Dem candidates. And a large cohort of tweets collected distributed unevenly. We did content analysis with a team of 6 trained coders. The unit of the individual tweet. The context unit of analysis was the Twitter page – that conversation. You needed that context to code them correctly, that was crucial. We had a dozen or so categories but I will focus on 4 main ones. Type of tweet; interaction with; tweet function; tweet topic. See Graham Broersma Hazelhoff and van’t Haar 2013 for the statistical analysis. I’ll be talking qualitative analysis.
Firstly to say something of the three Prime Minister candidates. Looking at frequency of mentions. 22nd April and 29th April see a big spike. Those are the last two TV debates. They had a substantial influence on tweeting. Also an increase in tweets towards polling day (less sharp).
In terms of tweet types this was basic: normal post (48.2%); reply (?%) ; retweet (?%); retweet with comment (?%). Huge difference between conservatives vs lib dems and labour. Latter two parties interacted far more. Who were they interacting with? The public lagely, politicians (mainly own party but some debate from labour candidates), party activists (hardly ever conservatives), media, etc.
We looked at tweet topics. 80% of all tweets were about Campaign and Party Affairs. Very minimal policy talk by comparison. Some “other” chat as well. The function varied more broadly. Lots of tweets about the campaign trail (23.1%); campaign promotion (20.9%); criticism and arguing (22.9); acknowledgement (9.7%); other functions less substantial. So a typology here. We saw Broadcasting with 5 behaviours: updating; promoting; party stance; etc.
Updating accounted for 23.1% of tweets. e.g. those shown from @Jeremy_Hunt and @andrew4mk. Perhaps this isn’t an unusual thing to track in the media. But some politicians did this in a more novel way. Some gave a sense of closeness, of being part of the campaign, of knocking on doors with them…
Promoting was around 20.9% of tweets. This was largely about promoting the party or the politician, their success, their performance etc. But the Lib Dems were quite interesting. They promoted the most, but they were also really creative.@CllrDaisyBenson called out for non-celebrity endorsements and got great personal endorsements and responses.
Critiquing (17%) tended to be very superficial, about style, performance, rarely any substantive critiques of policies. Created polarised Twittersphere. Really party orientated so particularly polarised during the debates for instance. Politicians complain about this BUT they do this. And a number of followers and members of the public called them on this. Stephen Coleman and Dan Jackson’s work finds that this sort of discourse puts citizens off politics.
Interacting also had 5 behaviours: a range here so:
Attacking/Debating (9.9% of tweets): these were again superficial attacks. Typically one-off exchanges. Extended debate was rare. a yell each way but no further. Ironic as the debates – which triggered many of these – was supposed to open up debate but rather shut down here.
Mobilising/Organising (3.7% of tweets): Labour and Lib Dems really led this. @DrEvanHarris tweeted about 1400 tweets in this time period. He was chat up new followers and then ask for their help in campaigning. Candidates also shared behind the scenes type tweets.
Advice giving, helping and consulting (3% of tweets): about connecting with the citizen. Consulting with the public was about 1%. Conservatives not tweeting in this way. There are maybe 7 or 8 candidates represented in these advice and consulting tweets.
In terms of Twittering about their personal life. We had a code that just marked tweets as “personal”. e.g. Louise Mensch tweeting about running. But we are recoding those as political and personal often overlap. We also want to code the personal. But looking at some samples we saw use of personal in combination with updating – combining campaign trail with personal life note (cats and chocolate + envelope stuffing). Using personal to promote. Very common with those with young children. Tom Watson tweets about his 2 year old. Using personal to attack and critique – using kids or pets to raise funny attacks. Using life experience to draw attention to a particular issue – eg “My aunty tells me”… or use of personal experience to support arguement over several tweet discussions, e.g. on Trains.
This work is still progressing. Analysis ongoing. And looking at that mix of personal and political.
Q) What did you do about tweets with links, e.g. to policy documents etc.
A) I have a whole codebook to deal with these. so lots of politicians tweeted links to newspapers with an attacking headline – so we coded it as attack. Or links to policy on blog – coded for that function. We basically clicked on the links and coded that up.
Q) Two questions. How did you separate whom category? People act in different roles. Also for functions categories – literature theory or your own coding?
A) Coding scheme is a combination of factors. Reading the literature. Inductive coding, also Darren’s work on politicians on Twitter influenced our coding. And we had four steps in that process of checking role of the people being interacted with. Context helped, the coders clicked through to profiles, sometimes Googling that person was required, we made some rules and procedures for that.
Q) Did the party train/constrain/orchestrate these for them? If not it seems really interesting to be outside the party machine as other publications are?
A) Labour and Lib Dems had a campaign coordinator for social media. And they had keen early adopters amongst their candidates. Conservatives not so much. Found suggestions they didn’t have much.
Me) Conservatives did hire someone for social media some time back. I’ll see if I can dig that out.
A) There was a piece in Wired… may be same one…
Q) What about images and coding?
A) We did code images in tweets but there were not many of them.
Q cont) Often they mess up that personal thing…
A) We coded them as critiques on the whole (one critiquing own party leader!)
Q) Political tweets are domain specific data. Can you develop something automatic based on your work?
A) We are doing that now, have multidisciplinary team. We are working with people in linguistics. Taking out 60k tweets and having programme learn from those tweets to try and reuse it for the next election. But context would be the tricky thing here. It was a big part of the coding process.
Comment) Programmes great for explicit texts. Much harder to deal with irony. But it’s machine learning and you need a small training set hand coded, then software learns from that and can be applied automatically to wider data set.
This is our final session of the day and is about Twitter use by those with “between 1000 and 10,000″ followers on Twitter. Pre conference tweets suggest that’s a slightly controversial grouping but we shall see shortly…
Participation was sought in advance for this session and there have been two tweeters coming forward. One a Gothic Literature academic, the other @scicurious who is a prominent science blogger.
So this will work in two streams. Already have discussion on #lutwitrc up here, we’ll also keep an eye on #lutwit tweets.
I started by asking about when you find time to tweet. If you don’t tweet much it seems there is never time. If you tweet a lot it seems like a non issue. @scicurious just tweets regularly as part of the day – it runs in the background. Johnny is commenting that he tweets intensively at conferences like this, but less so at other times. But do retweet then etc. It varies a lot. When reading or writing tends to be less so than when doing things like marking and want a diversion. Several other comments. Me, tend to have in background but real morning/evening rythems. Comment that Twitter has made tweet archive of your own twitter archive to see clear patterns there. Greg: I asked Dorothy Bishop, Professor of Psychology about why so many tweets are at 7am and she said “well academics are filling in forms the rest of the day!”. Penny: comments on extreme unevenness over course of day and over longer term. Comment: also about how long you are on Twitter – count has different meaning. A measure of total tweets varies radically. Maybe come back to that as very active tweets mean something differently. Julia: we are assuming one person is one account but you may be tweeting to many accounts and time intersects with identity, not neccassarily 1:1. David: I think Penny was suggesting that we treat Twitter like other communication – we don’t measure how many chats we have in a day. Me: but the account is there, so tempting. But presence is inferred from tweeting as well, so not tweeting suggests not present.
Discussion of time to write a tweet. Rebekkah Kills says it takes no time. Comment that reading takes time, not tweeting. Takes time but not much. Greg: interested in issues of timing… e.g. cultural norms around eating and talking at the same time. Me: it’s about context here, what are expectations. For a personal tweet I may have hours to reply, but for a professional tweet or an enquiry to a service twitter account expects fast replies. Greg: I am constantly surprised, but shouldn’t be, about the speed of student email replies. So fast. Four hours can be slow. Comment: my girlfriend tweets and uses facebook and kind of sits down to “do facebook” and I think people do do that…
Greg: so that takes us into other media use. Johnny, do you have the same posts going to Facebook and Twitter? Johnny: no, I used to, then used to unlink now totally unlinked… Comment: I’m wondering about Facebook and Twitter and the idea on Putnum on strong and weak ties. greg: your practices change depends on speed of tweeting but also how many people you follow. Julia: I do find the temporal aspect changes. Facebook bring different time zones in at different times. Synchronicity on Twitter puts me out of touch with those in other timezones. Greg: I take part of Flickr365 and we do a daily update but we all tend to post at the same time, it’s a daily rythem to keep up with. Different to always on. Me on ties: I think strong and weak ties exist in Facebook and Twitter, not a clear distinction. And also timewise there are timezones of interest/habit beyond location – e.g. those staying up late in UK vs daytime in US.
Onto anonimity and pseudonimity. Me: recommending Violet Blue’s writing on Google+. Greg: BitPHd is a blog I read and it is very successful but wouldn’t be possible with real names for personal comments could damage tenureship chances. Any other opinions? We are all academics pretty much, do we assume freedom with our employers here? Comment: if you google me you find an old account which is an issue. Johnny: I feel ambivelent for adding “views are my own” on my profile. I think there is no need for there to be an issue there. Karen: don’t think that “all views are my own” thing, doesn’t cover you legally. Law hasn’t caught up. Look at UCU’s advice on social media. There have been cases no matter the disclaimers. Tony: social media used as brand management and reputation management, tricksiness around that. It’s problematic. See link tweeted to my study. So disclaimer not a bad idea from personal perspective, distance self from institutional or departmental position. I use it on my Twitter account. Me: its about setting expectations, perceptions. Comment: I’m from China and think pseudonyms are safer. greg: my research showed that people like to see a face, a person, not a blank or object avatar.
So, moving onto impact… I’ve thought about number of followers and of retweets. Is that impact? Is something else? Me: well Klout may be silly but that idea that followers, activity, replies, retweets and the presence of influential people in your network and their engagement with you has some merits. Klout have interest in numbers for marketing reasons… David: Why would you do that? Same concerns as impact in any area of academia Greg: I have a practical reason, for sampling. For journals you would check impact factor in choices. But maybe an issue about impact and influence being perceived as about marketing is an issue for any sorts of measures here and how they could be perceived. Todd: we’ve been interviewing dutch politicians, for them getting tweets picked up by news media is key. For political leads and celebrities the pick up by mainstream media might be a measure. Comment: opinion leaders matter here too, not about counts. Greg: indeed, concepts move beyond Twitter, to concepts from before Twitter around influence and impact here. Comment: impact in my work is about reaching sources, spreading news etc. in crisis organisations. Organisations mobilise resources on the ground through very influential presences and accounts.
Final topic… We had a question from, I think outside the institution, am I OK collecting Twitter data without institutional data? Apart from all the Terms of Service, Copyright, what are the ethics of it? Start with gut ethical issues and then move onto legal issue. David: yes, it’s public but designed for Twitter. The same as saying… would you have ethical issue with using voxpop in a newspaper and then reprinting it. Greg: I don’t see problem with that. David: an ethical issue… it’s recontextualising it. You need ethical approval for closed Facebook posts, but media does use and attribute that. Greg: do we all agree that Facebook is clearly private and needs ethical approval? Tony: issue is expectation. Facebook has expectation of privacy. Issue for Twitter is whether people quoted assume the audience is public. People get it wrong. It can be ticklish. My feeling is that hashtags are somehow signals active participation in a conversation. David: On some sites you might feel you are posting to that site but you are reposting to Twitter. Similar issue to local newspaper vs pick up more widely. Could feel very different. So continuuation of same ethical issue. Comment: Really good article about this by danah boyd about levels of publicness in social media. Applies here. Looking at hashtag corpora some feel private, single post. Some massively public hashtags with lots of responses. I tried to anonimise usernames but near impossible to anonimise tweets themselves. Have to be savvy consumer but can we expect that of everyone. Julia: ethics matter in social media, just as anywhere else. And getting to know Terms of Service matters. Looking through Twitter not all ToS are as clear as they could be but you can communicate and ask questions. Really different attitudes to text made public and screenshots. Many layers here. You as researcher in institution, the wider space and world expectation, and the legal side of things. So many layers. Can be tempted to think we are insulated from these kinds of things, a world you should be just as careful as. Johnny: the Twitter ToS directly conflict with the idea of anonimity. Twitter makes it hard for researchers. Comment: need to separate ethical and legal issues here. We should consider the ethical issues for ourselves. ToS is about brand not users. Researching, taking data for analysis, and then publishing are two different things. And publishing may be at a conference (temporary, closed) but on slideshare say that’s difference, or in a book or paper that’s different again. Comment: can we separate legal and ethical issues here really. Issues of good faith, of relationships with data providers etc. Greg: I deliberately left this to the end as I think it would be concerning if, because it easier ethically, all analysis was quantative, but there are many issues to discuss… and those discussions will hopefully carry on at #lutwitRC and #lutwit.
And with that – and my apologies if the notes are a little hard to follow, I’ll mull a better format and may update accordingly – the formal sessions ended and we moved into the evening with the launch of Language Online: Investigating Digital Texts and Practices by David Barton and Carmen Lee followed by a lovely conference dinner for continuing those (and other) discussions… More from the final day of the conference appearing on the blog tomorrow.