What do we do with Twitter research now?

Yesterday, the Washington Post featured a well written article about the words used to describe Hillary Rodham Clinton. The argument in the paper was sound, the methods were a bit less robust. To build a meaningful dataset, the authors had used a seed hashtag to collect a functional survey of user language. Historically, the great strength of Twitter research was that it was not a survey: Tweets were treated as something of a wild-sample of talk in the world and possibly as natural experiments.

The reasons for using a seed hashtag are fairly straight-forward. Popular hashtags are often so heavily trafficked that the data would be bulky and cumbersome. Twitter datasets older than a few days are iffy. Cleaning raw Tweets is extremely labor intensive. Basically, creating a useful dataset from the outset is a lot easier than trying to reshape sloppy reality.

Twitter shifting toward an algorithmic timeline will present new difficulties. The assumption that impressions of a status or collected data from a scrape mean that there was viewing will now be gone, just as in the case of Facebook research. On a deeper level, this speaks to the failure of the moment as an organization principle for research.

Political temporality is a tricky idea. Topologically, time comes first in the stack of perception. This is a central insight for Benjamin, taken up by Agamben. Messanic time is different than ordinary time, Edkins discusses this through the idea of trauma time. The window that Twitter offered into public opinion was radically synchronic, it was only ever the now. Now is messy. Journalists and critics struggle to characterize the right now across time and space. They simplify relationships into basic cause-effect patterns where the reality of causation is far more complex. None of the relationships between political negotiations make sense from a simple cause and effect model, food stamps were cut with no public outcry. The items in the budget bill should have caused protest, instead nothing. Everyday life offers a time and a slate of topics, these resemble little of what happens in Washington or New York. The lifeworld is alien to the system. Coal miners in a heroin crazed town with near a recently closed Walmart have a very different perspective on reality than wealthy educated people in Washington playing a political capital simulation game. They turn against the system, towards rupture because time has run out. This is the distant future, it is bleak, painful, and hopeless.

Just as in Twitter research, all political research struggles with the fleeting moment. Moving to reduce the scope of research, as in the Clinton example, is a good choice. Does this say much about the expression of rage on Twitter? Not really. It can’t. Twitter for it’s part is trying to deal with the difficulty of a pained public by cleaning the product, distilling what people might want to see.

Image for post
Image for post
This is not the collective intelligence.

On Star Trek, the collection of voices through telepathic transceivers became a powerful collective. In various episodes the strength of all voices together was described by drones as something comforting. The voice of Twitter users together is not the smooth flowing voice of the collective assimilating the Enterprise. It is something else, something harder and less organized. The collective is in constant pain.

What does this all mean for the future of research? If we are seeking the signal, or a clear relationship: good news. Twitter will likely be less messy, impression counts will offer additional relevance for understanding which Tweeters actually matter. For understanding political temporality, the spirit of the now, it will mean less. We will replace the actual experience of disorder and anger with a simulacrum of already existing politics. The algorithm will amplify existing allocations of power. In short we continue, but we do this with the idea in mind that we are doing research on mass media content, not public discourse.

Get the Medium app

A button that says 'Download on the App Store', and if clicked it will lead you to the iOS App store
A button that says 'Get it on, Google Play', and if clicked it will lead you to the Google Play store