This is a blog to accompany a lecture I gave to the Altervilles Masters Programme in September 2014. In my talk I laid out the following argument: the Urban policy actors are engaging in practices of hashtag politics. That is, they are coining memorable and uniquely named policies, and these names give such policy a life and ultimately a death. I argue that social media is aiding the process but playing a role in their demise. As a public policy academic I think we are well placed to explain these transitions but we must continue to adapt our repertoire of analytical tools to include new forms of data.
My interest in this comes from ten years of encountering policy ideas – something I first saw when exploring Birmingham’s renaissance, how those I interviewed talked without hesitation of how Birmingham during the 1990s had returned to its former glory. I argued at the time that such a discourse had to be continually reiterated to maintain its efficacy. In the early 2000s one such mechanism was a policy idea called “Flourishing Neighbourhoods” – here I encountered tens of different ways of expressing the idea, with associations made with everything from civic pride to feeling safe, involvement and sustainable communities. Such ideas often have vague beginnings and few can remember from where they came. Many will claim it was their idea – name times and places where it was first adopted, stolen, adapted, coopered, corrupted. Few however can remember when it was forgotten. Such is the way with policy ideas – policy actors prefer to discuss the latest and loudest.
In my first project as a post-doc I visited Rotterdam and Copenhagen, there too policy actors framed their work in terms of policy ideas, such as city citizenship and integrated Copenhagen. I am sure if I returned there some 7 years on, many of the practices would be familiar but the branded policy ideas different. Every project I have been involved in has followed similar patterns – Transforming Birmingham, Total place, the big society. They operate at different scales and degrees of scope but they share common properties – they are branded ideas and visions with instrumental intent and container-like elasticity.
Around the time of studying the big society – 2010 or thereabouts – something else was happening. Watch the tour de France and you will see it on the roadside – compare audiences of the Pope in 2005 and 2013 and you will see something different. Every month websites will inform you that 1.19billion people are active on Facebook, 1.3 billion on YouTube and Google-plus, 232 million on Twitter, with growing numbers on the likes of Instagram, Pintrest and Tumblr. As smart phone ownership grows so do these numbers. Particular social networks will come and go, but mobile internet is here to stay.
Not every country is the same – in the UK Twitter is extremely popular in policy circles; in France, for now, less so, but it is growing. Facebook is extremely strong in France and DJs and footballers have followerships of over 15 million. French President Francois Holland has almost 700k followers and has tweeted over 4000 times – a similar number to David Cameron, although his office has 2.8 million. It is where news is announced – and it is where people react. Some call this the biggest focus group in the world, with people retweeting messages and replying to authors or mentioning users.
Cities too are now on Twitter: although mainly remaining in the communications department, cities are working out how to communicate messages and interact with residents. Who has control of the account remains a massive learning curve, just as airlines, train companies and brands have found when trying to promote their brands and defend their reputations. Government departments launch hashtag campaigns – like the Home Office in the UK promoting the police commissioner elections with #MyPCC and photos to make it seem personal. As the NYPD found, these campaigns can easily backfire, with users hijacking the #MyNYPD hashtag to reveal photos of police brutality. Similarly the highly effective #BringBackOurGirls spearheaded by the likes of Michelle Obama clearly put pressure on governments to respond to the plight of kidnapped Nigerian schoolgirls – and yet opened up those who got involved to harsh scrutiny of the Twitterarti.
So thinking about the big society back in 2010 led me down the route of asking two questions: What does all this mean for policy practice, and what does this mean for policy research?
Starting with practice, it leads to questions like how is Twitter changing how policy is made, or named, or implemented; how is it helping or hindering those who promote policy? When you delve into the literature it is somewhat fragmented. The biggest chunk is unsurprisingly focused on the profit motive – how can brands launch successful campaigns, how much does dialogue matter, how much does sentiment matter, can we predict sales or market performance or measure return on investment, and understand the needs of our customers? These questions have a policy relevance because businesses are the early adopters of Twitter – they have realised that two-way communication is advantageous yet difficult, but possible. They have realised that automated sentiment meters are unreliable, that there are many kinds of twitter user, that the mass media continue to exercise a hold on whether brands sink or swim. Furthermore they have sophisticated metrics that can understand which Tweets get more shares and our motivations to do so. But the literature is not all about profits and bottom lines. Much is made of the democratic potential of Twitter – the ability to mobilise and organise mass movements for or against a cause. Here they are interested in the contagion, the virility of a tweet and how these can redress democratic inequality. These literatures give many examples of countries where revolutions were fuelled by Twitter, where governments tried to switch it off or mount astroturf movements. It is changing so many ways in which professions operate – journalists sourcing stories, broadcasters creating viewertariats, emergency planners maintaining lines of communication, intelligence agencies listening in, health officials dissuading smoking, charities promoting donations. Political actors are also discussed in this literature – cultivating their micro-celebrity, self promoting, interacting, hoping to improve election outcomes, while others use Twitter data to predict the outcome, or exercise a Trial by Twitter, pointing out scandals and malpractice.
But Twitter has implications beyond just the practice of psephologists: it has wider implications for policy research. It begs questions like – what are these data – what is the role they play in the rise and demise of contemporary policy ideas – the modern day flourishing neighbourhoods, Total Place or indeed the Big Society (an idea we do not mention any more)? What can we do beyond our existing methods of counting newspaper articles and such? One of the main implications is a clamour for bigger and bigger datasets. This is comfort-data. Researchers have a classic fear of missing out – so try to collect everything, try to capture the full transaction of searchable talk. All of this brings with it issues of ethics and challenges of data management. Some choose to subscribe to companies that index social data, others essentially record it from Twitter through the APIs, and increasing numbers are buying packages of historic tweets from resellers.
Beyond the challenges there are some obvious exciting potentials – alternatives to costly case studies, alternatives to interview and document analysis, an opportunity to get to where debate is taking place, and permit the introduction of innovative methods and automated qualitative processes.
In the talk I set out a range of techniques that might be used, ranging from the most hands off to the most hands on. At the most macro levels, it is possible to count frequencies with relative ease – to track the fluctuation of mentions of a given policy and draw conclusions as to its vitality. We can drill down and explore the influencers, including a new generation of Twitter celebrities and mega-users who have tweeted over 120,000 times. We can apply algorithms to divide messages or parse parts of messages into positive and negative sentiment without the need for human input. If you are dissatisfied or worried about the reliability of such automation then you can capture the tweets and sift through them manually. If desired you can work in small teams coding and classifying items in order to isolate just those tweets expressing an opinion on the policy idea in question and then explore how the debate is changing shape over time. At perhaps the most hands on is to take samples of tweets and feed them back into Twitter itself, asking users to rank order tweets into order of preference and then subjecting these rankings to factor analysis. This Q-methodology reveals shared viewpoints that structure the data and in turn can inform further research – be it qualitative or quantitative by design.
What I conclude is that Twitter data offers a wealth of research possibilities. Whilst user numbers continue to grow and how it is used continues to evolve, so must policy research to keep up. It might be early days for some countries, but all signs suggest more and more policy actors across the world are taking to Twitter to launch and foster policy ideas. I am confident we are now well placed to be able to explain what is happening using a variety of hands on and automated tools. Returning to where we began – although social media has a relatively recent history, the techniques of policy making using memorable branded policy ideas has a much longer past; what is different are the tools that make and break them.