@HASTS_MIT: A Community & Public Scholarship Twitter Experiment

Post,Review February 17, 2015 11:36 am

Screen Shot 2015-02-16 at 4.56.31 PM

I’m a PhD candidate, prof of Japanese culture & media, acting director of Knight Science Journalism, NSF postdoc in Anthropology, member of the History Faculty, HASTS alum, 2nd year HASTS student, a historian of science, a Mellon Postdoc in Anthro, Academic Administrator in STS, an Exchange Scholar.

I study digital/social partnerships to resolve child exploitation & #trafficking; online parody as a form of social critique; laboratory-grown meat; body maps of anatomy and acupuncture in England and China; the production, practice, & circulation of evidence-based psychotherapies; the Atlantic sugar trade in the late 19th cent; history of technology, business, higher ed and “innovation”; digitization, computing, networks.

This week I’m tweeting about interesting meetings & events bearing on the future of @KSJatMIT; the history of medicine and east asian STS; science illustration, gender & science, & writing; #AAA2014; stuff happening around HASTS & MIT; what I’ve been reading lately.


The @HASTS_MIT Twitter project launched on September 1, 2014. Each week of the fall semester a different HASTS community member—including students, faculty, staff, alumni, and affiliates—hosted the account. HASTS, for those unfamiliar with it, is an interdisciplinary program at MIT that brings together researchers in history, anthropology, and science, technology, and society. As the excerpts above suggest, interests vary widely.

The @HASTS_MIT project began with twin aims: to engage in public scholarship and to serve as a HASTS community space. Along the way it has provided a structured, low-commitment space for playing with Twitter; given hosts an opportunity to share their interests with a different group of people; and showcased the HASTS community. The fall semester of the project has been nominated for a 2014 Digital Humanities Award (voting ends February 28, 2015). And on February 2, 2015, the project resumed for the spring semester.

Three sections follow; feel free to jump ahead/around as you desire: 1. Twitter, Collaboration, Community & Homophily, or some thoughts on the project, identity perception, and mediated relationships; 2. Numbers & Strategies, presenting some data from the fall and some ideas for the spring; and 3. Livetweeting, Retweeting, tweeting, & Intratweeting, exploring examples of different uses of the account. The mechanics of running the project are discussed in a separate post: Make-Your-Own Collaborative Twitter Account.


Twitter, Collaboration, Community & Homophily

Twitter is a surprisingly good tool for collaborative creation, both across accounts and within an account. The platform allows multiple user login, eschews a real name policy, supports automation, and, like most social media, institutes name–profile pic changes globally. That is, whenever the name or profile pic is changed, that change is applied to all previous instances in the Twitter ecosystem. Joy Rankin was the final host of the fall semester. Thus, during her week, every tweet from each of the previous weeks was now labeled as coming from HASTS / Joy.

As a host, this last feature means that not only do you see your name on tweets that you didn’t write as they continue to be retweeted or favorited (and they are—tweets both flash for a moment and span forever), you also know that the tweets you’re writing will soon bear the name of someone else. And then someone else. And someone else. This yields a collaboration with a particular flavor. A collaboration, I would suggest, that strongly contrasts with ideas of Twitter as for self-promotion.

The HASTS community has existed for more than 25 years; its members are already connected by various communication channels. This project adds a new communication space, in a new channel, to that mix. In doing so, it operationalizes ideas about media ecologies (Postman) and media ideologies (Gershon). That is, one of the project’s contributions is simply the act of adding a new channel—one with different associated expectations—and the subsequent changes in the larger communication context.

Also outside the Twitter account itself are the paratexts of the project—the monthly announcements of the host schedule and bios to the various HASTS-related lists, the posts on social media about the project, etc. These announcements both serve as introductions to other community members and articulate members’ desire to be connected as a community. Interestingly, HASTS community members who have neither hosted the account nor interacted with it through a Twitter account of their own have expressed to me their pleasure that the project exists. Since they’re not interacting with the project through Twitter, their main acquaintance with it is through the project’s paratexts. The paratexts are thus, in their own way, also important primary texts. (This makes me think of Jonathan Gray’s discussion of the importance of paratexts in creating anti-fans and nonfans.)

The project also complicates ideas of homophily in social networks. Homophily is basically the tendency for people to associate with other people perceived as similar to themselves. This has led to moral panics about social media functioning as pre-fab architecture for filter bubbles and echo chambers. (“Social network” here doesn’t just refer to social network services like Facebook, it refers to our larger fabrics of sociality, the people we are tied to in different ways. Social networks were an object of study in fields like linguistics, sociology, etc. long before Cyworld or Friendster arrived on the scene.)

Studies of homophily in social media are usually grounded in assumptions of one account one person. However, when an account run by an individual follows a collaboratively run account, balanced comparison is disrupted. Of course, the @HASTS_MIT account’s hosts are highly similar in at least one way: every host participates in MIT’s academic life. But in addition to numerous personal differences, hosts also varied in terms of disciplines, interests, and roles.

The account retained almost all of its followers during the fall semester, despite weekly shifts in interest and style. (Followers that dropped out tended to be ones who seem to have followed the account out of a desire for self-promotion rather than interest in the account.) The weekly shifting is the primary design feature of the project, of course, so if someone initially found the project interesting to follow, they’re likely to continue to follow it. However, if followers follow the account because they’re interested in the tweets of a particular account host, once the host shifts they may no longer be interested.

If thinking about homophily is relevant here, the fact that the account is, so far, retaining its followers could point to a number of things. It could point to follower inertia. It could point to perceptions of overall similarity among the hosts. It could also point to a greater willingness to embrace difference than has previously been supposed, due to a lack of collaborative entities in studies of homophily. Collaborations as the black swans of social networks, anyone?


Screen Shot 2015-02-16 at 5.46.43 PM

Numbers & Strategies

To preface this: @HASTS_MIT wasn’t set up as an experiment with specific variables to track or measure. I created the project because it seemed a worthwhile thing to do. What follows is an example of learning from semi-structured play.

So how did the fall semester go?

There are many ways to assess a project like this one: Impact, products, skills, understanding, pleasure, relationship bonds, etc. Of course, only certain types of data are readily available. And the data that is available comes with constraints—embedded norms, collection inconsistences, questionable abstractions, etc.—which limit the objective-omniscient picture of reality that can be painted. (Objective omniscient: 3rd person point of view favored by scientific gods.) This is all pretty standard STS critique. That said, readily available data can also speak to multiple categories in interesting ways and provide insight for future work.

So, here’s a smattering of data from the fall semester:

  • The account received more than 240k impressions (via Twitter Analytics)
  • The account averaged 1.4% engagement rate per day (engagement rate = engagements (retweets, favorites, etc.) / impressions (more Twitter Analytics))
  • The account tweeted 904 times…
  • …and favorited tweets 130 times
  • The account was added as a member to 18 Twitter lists, ranging from general awesomeness to @MIT’s MIT-affiliates, as well as an interesting mix of disciplinary focuses: HistSci/HistMed, Anthropolity, and Digital, Technology, STS
  • 16 people agreed to do this; no one withdrew; the project generated a (short!) waiting list
  • 80% of the project hosts who didn’t have personal Twitter accounts at the start of the project have since started personal accounts
  • Hosts averaged a total of 56.5 tweets, twice the guideline total (see Make-Your-Own Collaborative Twitter Account for guideline info)
  • The account saw an almost continual increase in followers to 486; only one week ended with fewer followers than it began

Don’t be distracted by these minimally contextualized numbers into thinking the bullet points mean more than they do. (Can’t help it? Pause to read some of this or this.) To a certain degree, this smattering is me saying, look, shiny data points! I hope to explore the specifics of Twitter Analytics and other ways to assess Twitter accounts at greater depth in a subsequent post.

That said, this list includes some direct products, markers of impact, and hints at community bonding. The data shows that the project generated enthusiasm in both hosts and followers, that hosts committed to the project (the fall semester’s hosts were awesome indeed, you can check them out here), and that the project has already had a variety of external effects. The data also points to listening practices, disciplinary boundary crossing, and sustained engagement.

There’s additional data on the project’s effects on skills, understanding, and pleasure: When reflecting on their weeks, hosts highlighted improved technical skills, new knowledge about the importance timing of tweets, and the surprising fun of hosting/Twitter. Multiple hosts volunteered, unsolicited, to host the project again. Of course, perhaps the most telling data point for assessing the project may simply be this: It will continue through the spring 2015 semester.

That said, not all feedback was positive. In their reflections, hosts also repeatedly mentioned a desire for more conversation, for meaningful, responsive exchange. With that in mind, during the spring semester of the project, I’m suggesting hosts optionally explore various conversation strategies. (Have ideas? Let me know!) Here are some current possibilities:

Talk to the list. ‪The account has a Twitter list of HASTS-related accounts. As a first step, hosts can explore responding to these tweets.

Matchmaking. Hosts draw on their personal knowledge—of followers, accounts followed, and the HASTS community—to put other accounts into conversation. So, for example, a host might solicit thoughts on a topic or article from two other people who may or may not already know each other.

Follower roulette. Hosts are directed to five follower accounts selected by harnessing Twitter’s random (?) display algorithm.

Keyword buddies. Hosts search the account’s follower and following lists from the fall semester for keywords to discover others with resonant interests. These other accounts can then be approached directly via @mentions.

Host solidarity. Semester hosts who also have personal Twitter accounts are asked to respond to the @HASTS_MIT account once a week. (This excellent suggestion comes from one of last semester’s hosts.)

#RSVP. Tagging tweets to explicitly solicit response. Right now there’s no marker for this. (Except possibly “Q” for question  to a limited degree.) Twitter’s iconic features like the RT, @mention, and hashtag all arose from user practices, later given platform functionality by Twitter. We can create new ones. Will they work? Unknown.


Screen Shot 2015-02-16 at 5.23.03 PM

Livetweeting, Retweeting, tweeting & Intratweeting

How did hosts use the account?

Several hosts engaged in livetweeting, or tweeting real-time observations of an event as it occurred. Thus we had livetweets of EmTech MIT 2014 from Wade Roush, JASmed 2014 from Lan Li, and AAA 2014 from Maria Vidart-Delgado. (Note that these are embedded tweets, so you can see the reattribution aspect of the platform at work here, reflecting the current name associated with the account.)

Hosts also explored retweeting tweets from other accounts. Beth Semel, in particular, used retweeting to showcase the power and value of curation itself. Many others sprinkled retweets in with other types of tweets. The first example here comes from Beth, the second from Harriet Ritvo.

And then there was more classic tweeting: tweeting thoughts on the host’s mind. (Examples from Joy Rankin, Diane Greco Josefowicz, and Benjamin Aldes Wurgaft respectively.)

Similarly classic, hosts engaged in discussion with others on Twitter. (Examples from Mitali Thakor and Shreeharsh Kelkar.)

Hosts also engaged in what I’ll call intratweeting, or tweets that involve other HASTS community members, through @mentioning, retweets, and direct discussion. So, similar structures as to what we’ve seen in previous tweets, but relevant as a category here because one of the aims of the project was to create a community space.

A tweet by me directed to Mitali Thakor and Jennifer Mnookin:

A retweet by Karen Gardner of Jia Hui Lee:

A tweet by David Singerman in discussion with Benjamin Aldes Wurgaft:


So those are some early thoughts on the @HASTS_MIT project. Come check out its spring semester! You know the deal: each week, different host. 

~Amy (@shrapnelofme)

Leave a reply