Jeffares Pol

Policy ideas in a digital age Stephen Jeffares

University of Birmingham Presented to Session 3 New Ideational Turns, as part of panel 84 New directions in the study of public policy, convened by Peter John, Hellmut Wollmann and Daniel A. Mazmanian, the 1st International Conference on Public Policy, Grenoble, France, June 26-28. Presented Friday 28 June, 8.30-10.30, Sciences Library Auditorium.

Abstract This paper argues that the discussion of public policy online is offering new and exciting opportunities for public policy research exploring the role of policy ideas. Although considerable work focuses on political ideas at the macro or mid-range, specific policy ideas and initiatives are overlooked, thought to be "too narrow to be interesting" (Berman, 2009, p. 21) .This paper argues that the prolific use of social media among policy communities means it is now possible to systematically study the micro-dynamics of how policy ideas are coined and fostered. Policy ideas are purposive, branded initiatives that are launched with gusto; flourish for around a thousand days; and then disappear with little trace as attention shifts to the latest and loudest. At best, media reports will document that Birmingham's Flourishing Neighbourhoods initiative has been “scrapped", "Labour's Total Place programme has been “torn up", or the Coalition's big society policy is “dead”. Save for a return to the policy termination literatures of the late 1980s, our impotence in conceptualising such death-notices reveals how little effort has been invested in understanding and theorising the lifecycle of policy ideas. In response, this paper conceptualises policy ideas, their life, death and succession. The paper draws on a case of the recent Police and Crime Commissioner elections held across England and Wales in November 2012, and the attempts of the Home Office to coin and foster the hashtag #MyPcc.

 Acknowledgement:  The  primary  research  reported  was  funded  by  British  Academy  Grant  –  SG112101  The  shape  of  ideas  to  come:  harvesting  and  analysing  online  discussion  about  public  policy.    And  University  of  Birmingham  Roberts  Fellowship  2008-­‐2013.    Heartfelt  thanks  to  the  research  team:  Gill  Plumridge,  Becky  O'Neil,  Tom  Daniels,  Pete  Redford,  Phoebe  Mau,  Diha  Mulla,  Misfa  Begum,  Sarah  Jeffries  and  Osama  Filali  Naji  for  your  empathetic  coding,  unwavering  enthusiasm  and  crowd-­‐like  wisdom.  

Work in progress. Please do not cite without permission from the author.

http://jeffar.es s.rjeffares@bham.ac.uk

Policy Ideas in a Digital Age

2

Policy ideas in a digital age Stephen Jeffares University of Birmingham

"If one end of the spectrum of political ideas is occupied by ideologies too broad to be useful, the other end is occupied by policy positions that may be too narrow to be interesting…In order to obtain useful and provocative results, therefore, I suggest concentrating on the complex ideas lying between ideologies and policy positions - the middle range of ideas" (Berman 2009: 21).

The process of policy making involves the creation of policy ideas, the purposive naming of policy initiatives, and related activity (S. Jeffares, 2014, forthcoming). The process of naming gives policy ideas an identifiable life, a birth, a parent, a lifecycle, and a death. The increasing use of social media to discuss and share information offers new platforms for policy ideas to be launched, promoted, discussed, critiqued and destroyed. With this in mind this paper asks three questions: What are policy ideas? How should we conceptualise their lifecycle? And what are the implications for policy ideas when launched and discussed on social media? In three sections, the paper first explores differing conceptions of policy ideas, before exploring how research is adapting to a digital age. The final section demonstrates how policy ideas can be monitored, visualised, captured and sifted using a variety of software tools, data analytics and Q methodology. The paper revisits previous research on Flourishing Neighbourhoods and Total Place and reports primary research exploring the policy ideas of elected PCCs (Police and Crime Commissioners), Compassionate Care and the Bedroom Tax. To conclude, the paper argues that continued advances in technology and methods will breathe new life into ideational understandings of public policy.

Policy ideas Policy ideas are a form of political idea between policy positions and middle range programmatic beliefs (Berman, 2009, p. 21). They can be attributed to a particular actor

Jeffares Pol

but they are seldom owned and controlled for their whole lifecycle. Framed as ideas they are vogue, fashionable, caught on, ideas in good currency (Schön, 1973). They are catchwords, useful yet loosely defined (Lasuen, 1969). They are magic concepts that dominate for a time, (Pollitt & Hupe, 2011) and issues rising and falling in cyclical patterns (Downs 1972). They are perennials that lie dormant before flowering again (Kingdon, 1995), they are key symbols providing a common experience (Lasswell 1949), they label populations and events (Edelman, 1977) and serve as popular keywords (Williams, 2013). They are buzzwords and fuzzwords clarifying and clouding understanding (Cornwall & Eade, 2010); they are virus-like entities (Gofas & Hay, 2012), memes that can spread across boundaries (Chesterman, 1997) and are capable of mutation (Dudley & Richardson, 2001; Richardson, 2000). While a policy idea like Prime Minister David Cameron’s “Big Society” fits these ideational descriptions of policy ideas, the same could be applied to the broader political ideas of austerity, leadership, partnership or co-production. What is it about policy ideas that makes them so widely a part of policy making, and yet so difficult to define? In response we need to consider instrumental qualities, visionary properties, capacity to accommodate demands and branded identity. We need to consider policy ideas as instruments; policy ideas are portrayed as self-consciously developed and promoted ideas (Moore in Reich, 1990, p. 88), serving as roadmaps at times of uncertainty (Goldstein & Keohane, 1993). In practice they are invariably described by policy actors as smokescreens, veneer, rhetorical cover, the product of spin or vehicles for wider agendas. We need to also consider their visionary capacity, that is the capacity to project and outline an imaginary vision of the future. How they can describe a situation in a future time (Hopkins & Zapata, 2007; Shipley & Newkirk, 1999) couched in a memorable phrase that can attract commitment and energise people, offering a bridge between the present and the future (Nanus, 1995). They are simplified mental accounts offered through slogans, rhetoric and stories), creating an understandable reality to work within and adhere to, and one loose enough to allow creative competition between narratives (van Hulst & Gerrits, 2008) discussing (Sharkansky, 2002). We also need to consider their capacity to accommodate: Policy ideas are defined by their capacity to accommodate demands, agendas, meanings and definitions. In this

Policy Ideas in a Digital Age

4

sense they are container concepts (Baker, 2008), garbage cans (Cohen, March, & Olsen, 1972), and empty signifiers (Laclau & Mouffe, 2001). We need to consider their brand-like uniqueness, repeated and marketed, gaining meaning by reputation and context (Lasswell, 1949). Policy ideas are compounds of two or more words joined in a single lexical unit (Algeo, 1993), they are the product of neology and creative compounding (Benczes, 2006). As brands, they focus and direct behaviour, aid implementation and foster positive associations, giving meaning and value to the policy as a product (Eshuis & Edwards, 2012; Kapferer, 1994). So policy ideas are something greater than a single programme or initiative, and more specific than a midrange political idea. It has caught on and mutated, it projects a vision of the future in a memorable phrase, it is loose enough to allow for creative competition to accommodate competing demands and definitions. They are two-word compounds that can be branded, sold, launched and re-launched. In addition to these five characteristics there is one further important consideration - that is - mortality (Bardach, 1976; Boin, Kuipers, & Steenbergen, 2010; DeLeon, 1977). Policy ideas are mortal. They have a limited lifecycle. Policy actors do not like acknowledgement of termination and instead talk mostly of recent adoption (Behn, 1978), and certifying the death of policy or succession requires sensitive measurement (Parsons 1995). Traditionally, we have relied on tracking and counting mentions or activity, sometimes showing cyclical patterns (Hogwood & Peters, 1982; Peters & Hogwood, 1985). With tools like Nexis and GoogleTrends we can count fluctuation in newspaper citations and web searches (Choi & Varian, 2012; Grundmann & Krishnamurthy, 2010). Or we can shift the focus onto adoption rather than activity, measuring who is adopting the term or attaching their demands (Rogers, 2003). Where the early adopters are the think tanks and consultants - the ‘performing fleas’ (Cornford, 1990) - and the late majority are the social sector - frontline workers and the like - by the time the latter start working with a policy idea, the early adopters have already "hopped off" onto the next big thing. Or perhaps in the same way as analysts track technologies, we can track policy ideas in terms of fluctuating expectation (Fenn & LeHong, 2011): the launch of the policy idea is met by a “peak of inflated expectation…a trough of disillusionment”, a pause, a rethink and then a gradual “slope of enlightenment” levelling out in a “plateau of productivity”

Jeffares Pol

(Fenn & LeHong, 2011). The requirement to iterate, revise and re-launch several generations of a policy idea is not a failure but part of the process of modern day policy making. The three approaches laid out below are summarised as tracking policy by fluctuation in (1) activity; (2) diffusion; and (3) expectation. They all track policy over time but differ by the y-axis, counting mentions, adoption and disillusion/hopes respectively. What these three approaches do is follow the brand and not the ever-changing meaning. They do not reflect how the increased activity, adoption and expatiation is also a form of subscription, whereby policy actors, in mentioning, embracing or critiquing a policy idea, are continually adding and modifying and affecting its life course. Therefore, we also need to consider equivalence, building on previous work and informed by the work of political discourse theorists (Glynos & Howarth, 2007; Howarth, 2000; Howarth, Norval, & Stavrakakis, 2000; Laclau & Mouffe, 2001; Torfing, 1999). The figure below expresses the process outlined above in terms of an initial expression of a single demand, the articulation of equivalence, a precarious growing chain, the enjoyment of hegemonic status, and the dramatic dislocation. With it comes an alternative theory of motivation: rather than furthering interests (Goldstein & Keohane, 1993), policy ideas are seeking to fulfil unfulfilled identities (Howarth et al., 2000). And this fantasy appeal of the policy idea (Glynos & Howarth, 2007) explains much of its pull. In earlier work I presented the example of Flourishing Neighbourhoods (S. Jeffares, 2008) - a popular policy idea in the English city of Birmingham between 2001 and 2004. In its first year it was the expression of a group of neighbourhood activists; it was then adopted at a Future Planning conference; in year two it was adopted as a corporate vision for the city and other organisations and partnerships, and was increasingly adopted by the social sector. In year 3, despite a late majority continuing to label their work as Flourishing Neighbourhoods, the ruling party adopted it as an election slogan, re-particularising the meaning before losing the election. Flourishing Neighbourhoods was dislocated. This is just one example of the cycle, but as newspaper archives attest, it was seldom mentioned again.

---------Figure 1 about here --------

Policy Ideas in a Digital Age

6

While the dislocation of a policy idea is a dramatic moment upon which to focus (Hajer, 2010), it is the translation from the original demand to early articulation of equivalence by others that holds the key to understanding the fate of policy ideas. It is that “take-off point” (Kingdon, 1995). Such an argument is analogous to fundraising campaigns or political petitions (Hale & Margetts, 2012), where the likelihood of attracting widespread support is dependent on the status of early adopters. While it is possible to track the early mentions, adoption and expectations expressed towards a policy idea, following the case of Flourishing Neighbourhoods, it is also important to systematically explore the early accumulation of meaning. Q methodology can be used for the purposes of capturing and mapping the subjectivity surrounding an emergent policy idea (S. Jeffares, 2014; Willis & Jeffares, 2012). By collecting documents, and interviewing and observing policy actors involved in a policy called Total Place, it was possible to capture something of the diversity of definitions and arguments circulating during the first three months of the policy initiative. Policy actors were then given a representative sample of these arguments and asked to rank-order them into a configuration representing their point of view. These responses were correlated and factorized to reveal four distinct shared viewpoints on this emergent policy idea of Total Place. The idea was subsequently abandoned, in name at least, following the election of May 2010, but Q methodology offered a means of systematically capturing subjectivity crystallising around an emergent policy idea.

A digital age Although Q methodology offers a means of mapping subjectivity surrounding a policy idea, how this occurs in a Facebook era (Shih 2012), where policy actors use Twitter to launch policy and create hashtags, is uncertain. Online activity in terms of friends, purchases, searches, location and text is leaving a “vast digital record” and generating “new discoveries in every area of research that uses it” (Hubbard 2011). The challenge facing public policy research is capturing the debate across many online and offline platforms, appreciating the temporal nature of the discussion and doing so while maintaining a systematic and transparent approach to research methodology. In this era of social media, public policy research is now required to adapt to capture discussion of policy that is both high volume and high velocity.

Jeffares Pol

Policy actors use social media for a variety of purposes. For politicians it is a platform to receive messages, field complaints, respond to constituents and answer questions. Politicians use social media to spread positive messages, praise or vitriol; they can label and tag opinion and create themes and spaces for discussion. Government departments can share praise about their work, respond to enquiries, cascade corporate messages and news, share validation or issue reminders of events or activities. Think tanks can receive ideas, listen to gossip or publicise their products. Broadcast media can recruit interviewees, publicise programmes or invite discussion from viewers. Citizens can pose questions, share or endorse policy related messages, label events, express opinion and attempt to mobilise action. Never has it been easier, or cheaper, to publish on the internet, and social media usage is growing with users increasingly accessing it via mobile devices. Social science research exploring prominent social media platforms trails behind the immediacy of the blogosphere and grey literature, but four literatures are emerging. The largest and most established is commercially focused, interested in audience and consumer insight (Barwise & Meehan, 2010; Culnan, McHugh, & Zubillaga, 2010; Hanna, Rohm, & Crittenden, 2011; Kaplan & Haenlein, 2010, 2012; Parent, Plangger, & Bal, 2011; Wigley & Lewis, 2012). The literature is concerned with “return on investment” (ROI) of social media campaigns, how engagement can increase brand awareness and reduce negativity. At the core of this work is automated sentiment analysis (Thelwall, Buckley, & Paltoglou, 2012), offering real-time early warning alerts of emerging negativity towards a brand or, on the converse, popular endorsement worth exploiting. An example is Apple's response to negativity surrounding its new maps application in 2012. Also of interest is the role of influencers or high profile evangelists, whose opinion or endorsement shapes the views and activities of the remaining 98% (Cha, Benevenuto, Haddadi, & Gummadi, 2012). While the commercially focused literature continues to grow, it represents just the visible tip of a largely invisible iceberg obscured by commercial secrets and firewalls. The methods and, for that matter, validity of many analytics providers remains shrouded in the mystery of “white papers”, those carefully worded PDFs available for download from websites but with methods seldom subject to peer review. At the opposite end of the spectrum is a second literature where commercial potential is supplanted by a democratic ideal. Social movements, revolutions and protests united by hashtags are deconstructed by this literature (Bratich, 2011; Lotan et al., 2011; Mungiu-Pippidi & Munteanu, 2009; Papacharissi & Oliveira, 2012; Youmans & York, 2012).

Policy Ideas in a Digital Age

8

Social media offers an accessible and mobile platform for organising, mobilising, developing common language, defining situations and formulating a shared vision (Bosch, 2012; Lindgren & Lundstrom, 2011). Aside from major events and uprisings, social media is credited with building a sense of local community and fostering social awareness in localities (Gruzd, Wellman, & Takhteyev, 2011; Naaman, Becker, & Gravano, 2011), and there is much interest in contagion (Lotan et al., 2011) and the role of “spreaders” (Gonzalez-Bailon, Borge-Holthoefer, Rivero, & Moreno, 2011). Moreover, there are questions of whether social media is democratising or merely mirrors off-line hierarchies and divisions of economics or gender (Armstrong & Gao, 2011; Auer, 2011; Page, 2012). Beyond the commercial and democratic are two further literatures: practice and political. The practice literature is perhaps better expressed as a fragmented collection of literatures characterised by a discussion of whether social media is transforming the professional practice of the likes of journalists, medics, lawyers, law enforcers or academics. It explores how journalists are sourcing stories, developing personal brands, sharing information, increasing readership, creating “viewertariats” and fostering discussion (Anstead & O'Loughlin, 2011; Lasorsa, Lewis, & Holton, 2012; Waters, Tindall, & Morton, 2010). It explores how universities are engaging with their students (Junco, Heiberger, & Loken, 2011); how medics are interacting with patients or promoting healthy behaviour (Hawn, 2009; Prochaska, Pechmann, Kim, & Leonhardt, 2012); and how public services disseminate information during natural disasters (Liu, Liu, & Li, 2012). It examines how charities promote their work and build communities (Lovejoy & Saxton, 2012) and how academics are now amassing massified datasets (Neuhaus & Webmoor, 2012) - challenging existing ethical principles (Boyd & Crawford, 2012) but creating excitement about the potential to discover, collaborate and communicate research (Rowlands, Nicholas, Russell, Canty, & Watkinson, 2011). A fourth literature is characterised by psephology and a desire to predict election outcomes - the role of social media in campaigns (Larsson & Moe, 2012; Tumasjan, Sprenger, Sandner, & Welpe, 2011) and how politicians are fostering their micro-celebrity (Golbeck, Grimes, & Rogers, 2010; Lassen & Brown, 2011; Marwick & Boyd, 2011). A key reflection is the predominance of broadcasting rather than conversing (Grant, Moon, & Grant, 2010). It explores how social media can both enhance (Rui & Whinston, 2012) and destabilise reputation (Chadwick, 2011).

Jeffares Pol

For these four literatures, the means to capture, or at least understand, what is taking place across multiple social media platforms, takes two approaches: "monitor and visualise" or "capture and sift". Market researchers want to monitor brand discussions and track reviews and emerging sentiment. Sociologists want to monitor hashtag use by location or type of user. Journalists want to monitor current issues. Public health researchers want to know how cigarettes are discussed. Police want to monitor riots. Pollsters want to follow how campaigns are discussed. The myriad of tools from Radian6, Sysmos, Clarabridge and Datasift and many others talk of ‘big data’, ‘voice of the consumer’, ‘customer experience management’, ‘social intelligence’, ‘social listening’, ‘activity listening’, ‘complete signal analysis’, ‘behaviour metrics’, ‘sentiment indexes’. Most are paid subscription services, offering companies and organisations software tools to keep real-time track of how they, their brands or their communities of interest are discussed online. I used one monitor and visualise service, Topsy.com, to track discussion around the build up on Twitter to the Police and Crime Commissioner elections in November 2012. The tool offered a means to graphically plot total activity of particular hashtags, such as the official Home Office hashtag #MyPCC, revealing that mentions varied between 150 and 600 per day, peaking at 1700 on the day of the election. It showed that spikes were driven by intervention of high profile users (such as the Prime Minister tweeting a pledge of support). It offered lists of related terms and hashtags, such as ‘criminals’, ‘polling stations’ and #therestissilence. It offered a list of "experts" ranked by a measure of their influence. It offered insight into the lifecycle of individual tweets in terms of momentum or acceleration. In addition to monitoring activity, it was also possible to understand exposure, in terms of appearance in user timelines: exposure showed which articles or attempts by the Home Office had the greatest reach beyond their immediate followership. It showed how exposure dropped from 6 million to 156,000 in the day after the results of the election on 16th November, and how by the 23rd November exposure hit zero. By the 23rd November #MyPCC was “dead” online. In addition to tracking activity and exposure, “monitor and visualise” tools also offer measures of sentiment, opinion and attitude. Some sentiment meters also index sentiment on a scale of 0-100, with 50 being neutral, offering a means of tracking sentiment akin to monitoring fluctuations in share prices on the stock market. In the case of Topsy.com the text is parsed into positive, negative and neutral expression, thereby allowing for a graph representing fluctuation between positive and negative, and the possibility of identifying

Policy Ideas in a Digital Age

10

the top negative and positive tweets each day. But interpreting this is by no means clear-cut, and automated sentiment meters can get meanings wrong and miss nuances. For example, on 6th October 2012 the top ‘negative’ tweet was a tweet from the Home Office, the lead department for the implementation of the PCC election, reminding users to watch the new TV advertisement for the PCC election. Similarly, a candidate’s tweet of "no privatisation of Kent police" was classified as a drop in sentiment, whereas the message was in fact supportive of the election and encouraging a vote for independent candidates, rather than a critique of the policy to hold elections per se. This kind of tracking also masks fluctuations in the volume of activity, which, in the case of the PCC elections, varied between 39 "opinionated" tweets (17th November) and 1718 (15th November). While real-time, sophisticated and seductive, validating these automated sentiment meters is somewhat troublesome. As Shulman argues, with such automated tools "it is very hard to know what one ambiguous tweet means, much less what they all add up to” (Shulman, 2012).

The social life of policy ideas The alternative to “monitor and visualise” is to “capture and sift”. Here researchers access social data through the Application Programme Interface (API) of individual social media platforms. Often this represents a sample of the overall discussion, and therefore some pay third party mediators with access to the full transaction (such as GNIP), often expressed in terms of a hosepipe metaphor, whereby this unrestricted access equals the full "firehose", rather than the partial "garden hose". Although the verbs vary between scraping, trawling, recording and harvesting, the common characteristic of “capture and sift” is collecting data in a machine-readable format for the purposes of sifting and exploring it with software. Most are primarily concerned with the metadata, socio-demographic information about location, tags, links, friends and so on, rather than the substantive content of the post itself. Those that do attempt to analyse the tweets are faced with the challenge of sifting though content to disambiguate homonyms; for example, if public health researchers are interested in cigarette smoking, they are required to disambiguate posts discussing smoking cannabis, smoking hot men/women and smoking barbecues (Kim, 2013). Case study 1: Police and Crime Commissioner elections I used a “capture and sift” application called DiscoverText to capture tweets referring to the police and crime commissioner elections, collecting those containing PCC (including

Jeffares Pol

hashtaged #pcc or the official tag of #mypcc). In total I collected 196,000 tweets, 52,000 of these in the build up to the election. The tool offered a means of isolating and focusing on particular users in the dataset -such as the official feed from the Home Office Twitter handle “@UKHomeOffice”, exploring how they used Twitter to disseminate and communicate the PCC policy, and how others attempted to interact with the Home Office on Twitter. Inspection of the Home Office feed showed that it was used mainly for reminding and broadcasting rather than responding to questions. It was possible to see which tweets were retweeted or quoted and modified, and to see if any seemed to dominate or steer the discussion. It was possible to filter the results by ‘Klout’1 score, to focus on just the most influential users that have, for instance, 5000 or more followers and a Klout score of, say, 65+. It was possible to focus on tweets over a specified period and identify commonly used words, display them as word clouds and show how and when phrases like "farce" came to feature. Importantly, by capturing discussion in this way, it makes it possible to subject the data to coding and classification. The DiscoverText software has both a facility to create coding schemes, to allocate coding tasks to peers or other DiscoverText users, and such coding could then be used as training data to inform and customise a Bayesian classifier (Maron, 1961; Sebastiani, 2002). Although such a classifier does not offer real-time “monitor and visualise” approaches, it allows researchers to maintain control over the analysis, and ensure inter-coder reliability and validity. I sought to explore what proportion of the tweets related to the build up to the PCC election were expressing some form of sentiment about the policy idea of electing a police and crime commissioner in 41 areas of England. Working with a dataset of 46,000 tweets expressed during the build up to the election, I recruited 7 students from the University of Birmingham, gave them background information about the policy area, and provided DiscoverText analyst accounts through which they could each access remotely from their personal computers. Starting with four coders, I randomly generated a set of 50 tweets and distributed it to each of the coders, asking the simple question "does this tweet express an opinion about the PCC elections?" Their coding could be checked for inter-coder agreement and kappa scores, and so validity could be adjudicated. A codebook was created with examples of what is considered opinion and not. A further two sets of 250 tweets were then issued, and the codebook refined. Some students made dramatic improvements in their accuracy from 80% to 92% validity and overall 1 Klout is measured on a scale of 0-100, based on followership and influence (Edwards, Spence, Gentile, Edwards, & Edwards, 2013).

Policy Ideas in a Digital Age

12

agreement rose to kappa 0.79 with five coders. All coders who joined the project subsequently were expected to complete the first three datasets. Once the coders were initiated and agreement sufficient2, each were allotted their own datasets of between 1000 and 2500 at a time. A total of 14,000 were coded. This was then used as training data to classify the full set of 46,000. In total only a fifth of the tweets were found to be expressing opinion about the PCC elections. The majority were giving information about the election or the candidates, or featured conversation between users about PCC without directly expressing opinion. Around 10% were deemed as spam or opportunistic activity using the PCC hashtag. With the opinionated tweets identified, we then subjected these to further rounds of coding, this time exploring themes. Starting with a random selection of 250, we started iteratively coding, focusing on the role of story-lines in tweets. From this process of free coding emerged a long list of 22 potential themes. These were then consolidated into five common themes, plus an additional three codes of "Other", "Not valid", "Not Sure". Given that it was impossible to assure 100% validity in the previous round of coding, it was expected between 10 and 20% of tweets would be invalid. The “not sure” code was used to speed up the coding process, and “other” to allow new themes to be captured. The five main themes were tweets regarding information, politicisation, organisation, quality of candidates and voting. Three coders were issued with a further set of 250 tweets, randomly generated. Reliability and validity were checked, and the codeset and codebook modified; a further dataset of 100 was issued, to do one further trial of the codeset, and after this I then issued larger sets of 2000. Coders were given detailed feedback on their coding performance and how to improve their validity scores. The coding from the 2000 could then be used to train a custom classifier to work on the wider set of 9000 opinionated tweets. Isolating and thematically coding opinionated social media data, as described above, can then be used for policy analysis in three ways: chronological insight, comparison of sources, and to sample statements for Q methodology. In terms of chronological insight the tweets can be mapped over the duration dataset, showing the fluctuation of opinionated tweets over the duration of a campaign, but it is also possible to trace the consistent, albeit fluctuating, subscription to particular themes and story-lines. It is also

Jeffares Pol

possible to see how those coded as “other” accumulate over the course of the campaign, as new meanings and themes are attached to the policy idea. Displayed cumulatively it would represent something close to the stylised depiction of Flourishing Neighbourhoods in Figure 1.

--------------Figure 2 about here -------- Case study 2: Compassionate Care The second use for this thematic data offers a means to compare sources. It responds to the increasingly common criticism that drawing insight from a single platform, particularly Twitter, is misleading. In the case of Twitter it has been widely argued that bias in user profiles means it does not proportionately represent wider public opinion (Mitchell & Hitlin, 2013). Thematic data of this kind can therefore serve to compare discussion of the same policy idea across several platforms. Working again with DiscoverText we sampled discussion around the policy idea of Compassionate Care, specifically, the widespread media discussion of whether the training of nurses meant too few were willing to offer compassionate care, that training was too focused on academic capacity, expressed in a single story-line that nurses were "too posh to wash". The Health Secretary had announced that nurses should spend a significant proportion of time working as care assistants. Alongside discussion from Twitter, we sampled newspaper comments from four that covered the story and allowed for reader comments: Daily Telegraph, Independent, Guardian and Daily Mail. We also sampled Facebook page comments from the main two professional newspapers - Nursing Times and Nursing Standard - as well as the Facebook page of the Royal College of Nursing. 6 coders coded first for posts expressing opinion specifically about the policy and then by theme. Thematic codes were suggested by the coders, 31 were consolidated into 9, plus a tenth code for "other". The results of the Compassionate Care study example show that certain themes are more common among certain sources. For example, the two Government-supporting

Policy Ideas in a Digital Age

14

newspapers were more likely to report it as a good or required response and were less concerned by implementation. In contrast, the Nursing Times and RCN raised it as a risk to the profession in terms of recruitment and reputation. There was also evidence of a meme-like theme coded as "apply to others" that suggested managers, doctors or politicians would also benefit from spending time as care assistants. Although this is just one example, it goes some way to suggesting that although the proportions vary, Twitter is more comprehensive than the user-generated content published by single newspapers, in that it offers diversity of debate. It is also worth adding that Twitter is in large part a mediator or referrer for other media sources, and therefore this finding is perhaps unsurprising. Nevertheless, the argument that Twitter does not represent public opinion continues to rumble on. Perhaps it is best to say it reflects public opinion rather than represents it.

-------Figure 3 about here ------- Case study 3: The Bedroom Tax A third potential use is to crowdsource viewpoints surrounding the policy issue, by combining thematically coded data with Q methodology. For example, using thematic data from an archive of 8155 tweets on the recent Spare room Subsidy/Bedroom Tax’ debate in the UK. (In April 2013 tenants considered to be under-occupying their property would face between a 14% and 25% reduction in housing benefit. This was framed by the Government as a “Spare room Subsidy.” The campaign against the move framed it as a Bedroom Tax). IN total 983 statements of opinion were isolated. These were coded against ten themes, and then randomly sampled to produce a set of 25, drawing as evenly as possible from each of the 10 themes. To complete the Q sort, respondents were recruited using email, targeting bedroom tax "experts" on twitter as suggested by Topsy.com, and Facebook advertisement. In total 26 responses were collected, with each respondent taking 19 minutes on average to complete the online sort. The results are summarised in Figure 4. Viewpoint 1 argued that there was a lack of housing supply, nowhere to go, no choice. Viewpoint 2 was that the bedroom tax was a good policy mired by myth and hypocrisy. Viewpoint 3 was that this was not a tax but rather a poor idea, lacking clarity. The difference between this and the Total Place study discussed above is the speed and systematic nature by which argument was sampled. Albeit just from one source, it is possible to collect with relative ease 983 opinions about a policy issue, thematically code them and crowdsource the viewpoints

Jeffares Pol

using online recruitment. This simple example is a demonstration of how this approach could be applied to a wide range of policy ideas or repeated with one policy issue longitudinally across a policy idea's lifecycle. A fuller write up of both this and the compassionate care study can be found in (Ch 7S. Jeffares, 2014).

----------------Figure 4 about here -----------

Conclusion This paper began with a focus on policy ideas as something greater than a specific initiative but more specific than a mid-range initiative, something resembling an idea whose time has come, projecting a vision, encased in a memorable phrase or label, that is loose enough to accommodate competing demands, additional projects and definitions. Importantly, as uniquely named entities they can be branded, sold, launched and critiqued. Branding them so gives them a life and ultimately a death; policy ideas are mortal. The paper argued that although it is possible to track frequency of mentions, diffusion or expectation, lurking beneath the outer shell of a policy idea is meaning. Throughout its life, a policy idea accumulates meaning that in the end leads to a dislocation and demise. The paper gave the example of Flourishing Neighbourhoods, that in thousand days accumulated meaning, before being adopted as an election slogan, and thereby dislocating its broad appeal. The example also highlighted the importance of the formative days and months of a policy idea, where policy actors begin to attach meaning, and themes and viewpoints crystallise. The paper demonstrated how Q methodology may be used to systematically sample opinion and identity viewpoints from policy actors engaged in the policy idea. The example of Total Place revealed four viewpoints emerging around this policy in the first three months since its launch. The paper then explored the implication of the growing use of social media platforms to launch and discuss policy ideas and what this meant in terms of lifecycle and analytical techniques. A review of literature revealed growing interest in how social media is shaping the commercial, democratic, professional and political spheres. However, much of this work focuses on metadata rather than a qualitative focus on the content itself. For those seeking real-time insights, most turn to "monitor and visualise" tools, where practitioners use indexed collections of social media transactions to plot activity, identify

Policy Ideas in a Digital Age

16

related terms in use, identify experts and influencers, momentum and exposure. Of critical importance is sentiment – the desire to gain stock-market-like insight into the fluctuating opinion directed towards a brand name. But as the example of the police and crime commissioner elections showed, it is difficult to verify the reliability and validity of these tools. They might be of interest to investors or marketers requiring real-time insight, but are of limited use to public policy research. The alternative to “monitor and visualise” is described as "capture and sift", involving the capture of social media data via APIs or third party power track providers. Once in a machine readable format, the challenge begins with disambiguation of homonyms and duplicates. It is then possible to focus on particular users and interactions with others, to cluster near-duplicates, to track commonly used words, but most importantly it is possible to subject the data to coding and machine classification. Rather than rely on a one-size-fits-all sentiment classifier, “capturing and sifting” the data means it is possible to train custom classifiers specific to the project and research questions. The paper showed how a team of coders were trained through a process of codebook creation, clarification, agreement comparison and validity adjudication. Coding preferences become training data for a machine classifier. The example identified that a fifth of tweets in the build up to the PCC election were opinionated, and from these five themes emerged over the three week period, along with additional themes growing in prominence as the volume of tweets increased. The paper suggested how the thematic data could also be used to compare opinion across different sources and also to inform a Q study. Two projects around the policy ideas of Compassionate Care and Bedroom Tax were offered by means of illustration. In response to Berman's argument that specific policy initiatives are too narrow to be interesting, I argue otherwise. For too long policy ideas have been a core element of policymaking and have been largely overlooked. This paper has offered something of a corrective by synthesising a definition, conceptualising their lifecycle and showing how the digital transaction left behind when policy ideas are discussed online offers an exciting new chapter for public policy research. Some policy researchers will continue to make great strides by focusing and visualising policy making through metadata. The exciting potential of peer coding, customised machine classification and Q methodology mean we can and should focus on meaning too. This paper has shown that advances in technology and methods in the last five years and their continued and rapid development will breathe new life into ideational approaches to public policy research.

Jeffares Pol

References Algeo, J. (1993). Fifty Years Among the New Words: A Dictionary of Neologisms 1941-1991:

Cambridge University Press. Anstead, N., & O'Loughlin, B. (2011). The Emerging Viewertariat and BBC Question Time:

Television Debate and Real-Time Commenting Online. International Journal of Press-Politics, 16(4), 440-462. doi: 10.1177/1940161211415519

Armstrong, C. L., & Gao, F. F. (2011). GENDER, TWITTER AND NEWS CONTENT An examination across platforms and coverage areas. Journalism Studies, 12(4), 490-505. doi: 10.1080/1461670x.2010.527548

Auer, M. R. (2011). The Policy Sciences of Social Media. Policy Studies Journal, 39(4), 709-736. doi: 10.1111/j.1541-0072.2011.00428.x

Baker, Keith. (2008). Strategic Service Partnerships ad Boundary-Spanning Behaviour" A Study of Muliple Cascading Policy Windows. (PhD), University of Birmingham

Bardach, Eugene. (1976). Policy termination as a political process. Policy Sciences, 7(2), 123-131. Barwise, P., & Meehan, S. (2010). The One Thing You Must Get Right When Building a Brand.

Harvard Business Review, 88(12), 80-+. Behn, Robert D. (1978). How to terminate a public policy: A dozen hints for the would-be

terminator. Policy Analysis, 4(3), 393-413. Benczes, Réka. (2006). Creative compounding in English: the semantics of metaphorical and

metonymical noun-noun combinations (Vol. 19): John Benjamins Publishing. Berman, S. (2009). The Social Democratic Moment: ideas and politics in the making of interwar

Europe: Harvard University Press. Boin, Arjen, Kuipers, Sanneke, & Steenbergen, Marco. (2010). The life and death of public

organizations: A question of institutional design? Governance, 23(3), 385-410. Bosch, T. (2012). Blogging and tweeting climate change in South Africa. Ecquid Novi-African

Journalism Studies, 33(1), 44-53. doi: 10.1080/02560054.2011.636825 Boyd, D., & Crawford, K. (2012). CRITICAL QUESTIONS FOR BIG DATA Provocations for a

cultural, technological, and scholarly phenomenon. Information Communication & Society, 15(5), 662-679. doi: 10.1080/1369118x.2012.678878

Bratich, J. (2011). User-Generated Discontent: Convergence, polemology and dissent. Cultural Studies, 25(4-5), 621-640. doi: 10.1080/09502386.2011.600552

Policy Ideas in a Digital Age

18

Cha, M., Benevenuto, F., Haddadi, H., & Gummadi, K. (2012). The World of Connections and Information Flow in Twitter. Ieee Transactions on Systems Man and Cybernetics Part a-Systems and Humans, 42(4), 991-998. doi: 10.1109/tsmca.2012.2183359

Chadwick, A. (2011). The Political Information Cycle in a Hybrid News System: The British Prime Minister and the "Bullygate" Affair. International Journal of Press-Politics, 16(1), 3-29. doi: 10.1177/1940161210384730

Chesterman, A. (1997). Memes of Translation: The Spread of Ideas in Translation Theory: John Benjamins Publishing Company.

Choi, Hyunyoung, & Varian, Hal. (2012). Predicting the present with google trends. Economic Record, 88(s1), 2-9.

Cohen, Michael D., March, James G., & Olsen, Johan P. (1972). A garbage can model of organizational choice. Administrative science quarterly, 1-25.

Cornford, James. (1990). Performing fleas: Reflections from a think tank. Policy Studies, 11(4), 22-30.

Cornwall, Andrea, & Eade, Deborah. (2010). Deconstructing Development Discourse: Buzzwords and Fuzzwords: Practical Action Publishing.

Culnan, M. J., McHugh, P. J., & Zubillaga, J. I. (2010). How Large U.S. Companies Can Use Twitter And Other Social Media To Gain Business Value. Mis Quarterly Executive, 9(4), 243-259.

DeLeon, Peter. (1977). Public policy termination: An end and a beginning. Dudley, G., & Richardson, J. (2001). Why Does Policy Change?: Lessons from British Transport

Policy 1945-99: Taylor & Francis. Edelman, M. J. (1977). Political language: words that succeed and policies that fail: Academic

Press. Edwards, Chad, Spence, Patric R., Gentile, Christina J., Edwards, America, & Edwards, Autumn.

(2013). How much Klout do you have… A test of system generated cues on source credibility. Computers in Human Behavior.

Eshuis, Jasper, & Edwards, Arthur. (2012). Branding the city: the democratic legitimacy of a new mode of governance. Urban Studies.

Fenn, Jackie, & LeHong, Hung. (2011). Hype cycle for emerging technologies, 2011. Gartner, July.

Glynos, Jason, & Howarth, David Robert. (2007). Logics of critical explanation in social and political theory: Routledge.

Gofas, Andreas, & Hay, Colin. (2012). The Role of Ideas in Political Analysis: A Portrait of Contemporary Debates: Taylor & Francis Group.

Golbeck, J., Grimes, J. M., & Rogers, A. (2010). Twitter Use by the US Congress. Journal of the American Society for Information Science and Technology, 61(8), 1612-1621. doi: 10.1002/asi.21344

Goldstein, J., & Keohane, R. O. (1993). Ideas and Foreign Policy: Beliefs, Institutions, and Political Change: Cornell University Press.

Gonzalez-Bailon, S., Borge-Holthoefer, J., Rivero, A., & Moreno, Y. (2011). The Dynamics of Protest Recruitment through an Online Network. Scientific Reports, 1, 7. doi: 10.1038/srep00197

Grant, W. J., Moon, B., & Grant, J. B. (2010). Digital Dialogue? Australian Politicians' use of the Social Network Tool Twitter. Australian Journal of Political Science, 45(4), 579-604. doi: 10.1080/10361146.2010.517176

Grundmann, Reiner, & Krishnamurthy, Ramesh. (2010). The discourse of climate change: A corpus-based approach. Critical approaches to discourse analysis across disciplines, 4(2), 125-146.

Gruzd, A., Wellman, B., & Takhteyev, Y. (2011). Imagining Twitter as an Imagined Community. American Behavioral Scientist, 55(10), 1294-1318. doi: 10.1177/0002764211409378

Jeffares Pol

Hajer, Maarten A. (2010). Authoritative Governance:Policy Making in the Age of Mediatization: Oxford University Press.

Hale, Scott, & Margetts, Helen. (2012). Understanding the Mechanics of Online Collective Action Using'Big Data'. Available at SSRN 2041856.

Hanna, R., Rohm, A., & Crittenden, V. L. (2011). We're all connected: The power of the social media ecosystem. Business Horizons, 54(3), 265-273. doi: 10.1016/j.bushor.2011.01.007

Hawn, C. (2009). Take Two Aspirin And Tweet Me In The Morning: How Twitter, Facebook, And Other Social Media Are Reshaping Health Care. Health Affairs, 28(2), 361-368. doi: 10.1377/hlthaff.28.2.361

Hogwood, Brian W., & Peters, B. Guy. (1982). The dynamics of policy change: Policy succession (Vol. 14).

Hopkins, Lewis D., & Zapata, Marisa A. (2007). Engaging the future: forecasts, scenarios, plans, and projects: Lincoln Inst of Land Policy.

Howarth, David. (2000). Discourse: Open University Press. Howarth, David, Norval, Aletta., & Stavrakakis, Yannis. (2000). Discourse Theory and Political

Analysis: Identities, Hegemonies and Social Change: Manchester University Press. Jeffares, S. (2008). Why public policy ideas catch on: empty signifiers and flourishing

neighbourhoods. (Ph.D. ), University of Birmingham. Jeffares, S. . (2014). Policy Making and Social Media: Interpreting Hashtag Politics.

Basingstoke: Palgrave MacMillan. Junco, R., Heiberger, G., & Loken, E. (2011). The effect of Twitter on college student

engagement and grades. Journal of Computer Assisted Learning, 27(2), 119-132. doi: 10.1111/j.1365-2729.2010.00387.x

Kapferer, J. N. (1994). Strategic Brand Management: New Approaches to Creating and Evaluating Brand Equity: Free Press “[u.a.]”.

Kaplan, A. M., & Haenlein, M. (2010). Users of the world, unite! The challenges and opportunities of Social Media. Business Horizons, 53(1), 59-68. doi: 10.1016/j.bushor.2009.09.003

Kaplan, A. M., & Haenlein, M. (2012). The Britney Spears universe: Social media and viral marketing at its best. Business Horizons, 55(1), 27-31. doi: 10.1016/j.bushor.2011.08.009

Kim, Y. ;Huang, J. ;Emery, S. . (2013). From the Known Knowns to the Unknown Unknowns: Precision and Relevance with Social Data. Paper presented at the Big Data in Public Health Conference, Philadelphia, Pa. . http://www.healthmediacollaboratory.org/what-weve-learned/presentations/ - sthash.48PSNR8r.dpuf

Kingdon, John W. (1995). Agendas, Alternatives and Public Policies: HarperCollins College Publishers.

Laclau, Ernesto, & Mouffe, Chantal. (2001). Hegemony and socialist strategy: towards a radical democratic politics: Verso.

Larsson, A. O., & Moe, H. (2012). Studying political microblogging: Twitter users in the 2010 Swedish election campaign. New Media & Society, 14(5), 729-747. doi: 10.1177/1461444811422894

Lasorsa, D. L., Lewis, S. C., & Holton, A. E. (2012). NORMALIZING TWITTER Journalism practice in an emerging communication space. Journalism Studies, 13(1), 19-36. doi: 10.1080/1461670x.2011.571825

Lassen, D. S., & Brown, A. R. (2011). Twitter: The Electoral Connection? Social Science Computer Review, 29(4), 419-436. doi: 10.1177/0894439310382749

Lasswell, Harold D. (1949). Style in the Language of Politics. Language of Politics: Studies in Quantitative Semantics (New York, 1949), 29, 31.

Lasuen, José Ramón. (1969). On growth poles. Urban Studies, 6(2), 137-161.

Policy Ideas in a Digital Age

20

Lindgren, S., & Lundstrom, R. (2011). Pirate culture and hacktivist mobilization: The cultural and social protocols of #WikiLeaks on Twitter. New Media & Society, 13(6), 999-1018. doi: 10.1177/1461444811414833

Liu, Z. M., Liu, L., & Li, H. (2012). Determinants of information retweeting in microblogging. Internet Research, 22(4), 443-466. doi: 10.1108/10662241211250980

Lotan, G., Graeff, E., Ananny, M., Gaffney, D., Pearce, I., & Boyd, D. (2011). The Revolutions Were Tweeted: Information Flows During the 2011 Tunisian and Egyptian Revolutions. International Journal of Communication, 5, 1375-1405.

Lovejoy, K., & Saxton, G. D. (2012). Information, Community, and Action: How Nonprofit Organizations Use Social Media. Journal of Computer-Mediated Communication, 17(3), 337-353. doi: 10.1111/j.1083-6101.2012.01576.x

Maron, M. E. (1961). Automatic indexing: an experimental inquiry. Journal of the ACM (JACM), 8(3), 404-417.

Marwick, A. E., & Boyd, D. (2011). I tweet honestly, I tweet passionately: Twitter users, context collapse, and the imagined audience. New Media & Society, 13(1), 114-133. doi: 10.1177/1461444810365313

Mitchell, Amy, & Hitlin, Paul. (2013). Twitter Reaction to Events Often at Odds with Overall Public Opinion. from http://www.pewresearch.org/2013/03/04/twitter-reaction-to-events-often-at-odds-with-overall-public-opinion/

Mungiu-Pippidi, A., & Munteanu, I. (2009). MOLDOVA'S "TWITTER REVOLUTION". Journal of Democracy, 20(3), 136-142.

Naaman, M., Becker, H., & Gravano, L. (2011). Hip and Trendy: Characterizing Emerging Trends on Twitter. Journal of the American Society for Information Science and Technology, 62(5), 902-918. doi: 10.1002/asi.21489

Nanus, Burt. (1995). Visionary leadership: creating a compelling sense of direction for your organization: Jossey-Bass.

Neuhaus, F., & Webmoor, T. (2012). AGILE ETHICS FOR MASSIFIED RESEARCH AND VISUALIZATION. Information Communication & Society, 15(1), 43-65. doi: 10.1080/1369118x.2011.616519

Page, R. (2012). The linguistics of self-branding and micro-celebrity in Twitter: The role of hashtags. Discourse & Communication, 6(2), 181-201. doi: 10.1177/1750481312437441

Papacharissi, Z., & Oliveira, M. D. (2012). Affective News and Networked Publics: The Rhythms of News Storytelling on #Egypt. Journal of Communication, 62(2), 266-282. doi: 10.1111/j.1460-2466.2012.01630.x

Parent, M., Plangger, K., & Bal, A. (2011). The new WTP: Willingness to participate. Business Horizons, 54(3), 219-229. doi: 10.1016/j.bushor.2011.01.003

Peters, B. Guy, & Hogwood, Brian W. (1985). In search of the issue-attention cycle. The Journal of Politics, 47(01), 238-253.

Pollitt, Christopher, & Hupe, Peter. (2011). Talking About Government: The role of magic concepts. Public Management Review, 13(5), 641-658.

Prochaska, J. J., Pechmann, C., Kim, R., & Leonhardt, J. M. (2012). Twitter=quitter? An analysis of Twitter quit smoking social networks. Tobacco Control, 21(4), 447-449. doi: 10.1136/tc.2010.042507

Reich, R. B. (1990). The Power of the Public Ideas: Harvard University Press. Richardson, Jeremy. (2000). Government, interest groups and policy change. Political studies,

48(5), 1006-1025. Rogers, E. M. (2003). Diffusion of Innovations, 5th Edition: Free Press. Rowlands, I., Nicholas, D., Russell, B., Canty, N., & Watkinson, A. (2011). Social media use in

the research workflow. Learned Publishing, 24(3), 183-195. doi: 10.1087/20110306

Jeffares Pol

Rui, H. X., & Whinston, A. (2012). Information or attention? An empirical study of user contribution on Twitter. Information Systems and E-Business Management, 10(3), 309-324. doi: 10.1007/s10257-011-0164-6

Schön, Donald A. (1973). Beyond the stable state: W. W. Norton, Incorporated. Sebastiani, Fabrizio. (2002). Machine learning in automated text categorization. ACM computing

surveys (CSUR), 34(1), 1-47. Sharkansky, Ira. (2002). Politics and policymaking: in search of simplicity: Lynne Rienner

Publishers. Shipley, Robert, & Newkirk, Ross. (1999). Vision and visioning in planning: what do these terms

really mean? Environment and Planning B, 26, 573-592. Shulman, Stuart (Producer). (2012, 17 June 2013). DiscoverText An Introduction to the key tools

and methods. [Video] Retrieved from http://www.screencast.com/t/BtBDsBTtw Thelwall, M., Buckley, K., & Paltoglou, G. (2012). Sentiment Strength Detection for the Social

Web. Journal of the American Society for Information Science and Technology, 63(1), 163-173. doi: 10.1002/asi.21662

Torfing, Jacob. (1999). New Theories of Discourse: Laclau, Mouffe and Zizek: Wiley. Tumasjan, A., Sprenger, T. O., Sandner, P. G., & Welpe, I. M. (2011). Election Forecasts With

Twitter: How 140 Characters Reflect the Political Landscape. Social Science Computer Review, 29(4), 402-418. doi: 10.1177/0894439310386557

van Hulst, Merlijn., & Gerrits, Lasse. (2008). Dream On: The use of the vision concept for governing networks. Paper presented at the EGPA, Rotterdam.

Waters, R. D., Tindall, N. T. J., & Morton, T. S. (2010). Media Catching and the Journalist-Public Relations Practitioner Relationship: How Social Media are Changing the Practice of Media Relations. Journal of Public Relations Research, 22(3), 241-264. doi: 10.1080/10627261003799202

Wigley, S., & Lewis, B. K. (2012). Rules of engagement: Practice what you tweet. Public Relations Review, 38(1), 165-167. doi: 10.1016/j.pubrev.2011.08.020

Williams, Raymond. (2013). Keywords (Routledge Revivals): A Vocabulary of Culture and Society: Routledge.

Willis, Martin, & Jeffares, Stephen. (2012). Four viewpoints of whole area public partnerships. Local Government Studies, 38(5), 539-556.

Youmans, W. L., & York, J. C. (2012). Social Media and the Activist Toolkit: User Agreements, Corporate Interests, and the Information Infrastructure of Modern Social Movements. Journal of Communication, 62(2), 315-329. doi: 10.1111/j.1460-2466.2012.01636.x

Policy Ideas in a Digital Age

22

Figures Figure 1 Lifecycle of Flourishing Neighbourhoods, (Jeffares 2008)

Jeffares Pol

Figure 2 Stacked Area of PCC Themes

Policy Ideas in a Digital Age

24

Figure 3: Nurse training reform comments by theme across eight media sources

0%  

10%  

20%  

30%  

40%  

50%  

60%  

70%  

80%  

90%  

100%  

Bad  idea  

Politicisation  

Not  the  answer  

Risk  to  profession  

Implementation  worries  

Already  done  

Other  themes  

Apply  to  others  

Required  Response  

Good  idea  

Jeffares Pol

Table 1: Factor Arrays of Three shared viewpoints on Bedroom Tax debate

Statement     F1   F2   F3  1. Bedroom tax is a waste of time 1   -­‐2   0  2. Bedroom tax is doomed to fail 0   -­‐1   1  3. Bedroom tax is such an ill-thought-out decision 2   -­‐1   3  4. The real impact of welfare reform is on the people 2   2   -­‐1  5. Let us not forget that a third of ex-council homes are owned by rich landlords 1   1   -­‐1  6. If we hadn't sold off all the council houses we wouldn't have this problem 2   -­‐2   -­‐2  7. Enough of talk about Cameron paying bedroom tax on chequers -he doesn't claim housing benefit 0   1   1  8. It is disgusting that MPs can claim thousands in travel expenses while levying this tax on the people 1   -­‐2   2  9. There isn't enough social housing for the under occupiers to move into - so it unfair 3   0   -­‐1  10. They initiate a bedroom tax with no homes to put the people into. 3   0   0  11. The bedroom tax was never a good idea. what about fragile pensioners & the visually impaired? it clearly wasn't thoroughly discussed 1   -­‐1   2  12. This policy will lead to rafts of unemployed under 25s homeless as parents faced with a tax they can't pay. 0   -­‐3   -­‐1  13. This is part of Cameron's Government destroying the poor 1   -­‐3   0  14. This TAX THE RICH / hate the "bedroom tax" contradictory nonsense, in the same vein as pro coal minding anti coal power -­‐1   0   -­‐1  

Policy Ideas in a Digital Age

26

15. The UK having the equivalent of three empty Cities the size of Northampton funded by other peoples money remains unjustified -­‐1   1   0  16. There is nothing socialist about refusing to downgrade to smaller homes -­‐1   3   -­‐3  17. People affected by bedroom tax should get sofa beds, that's what net contributors do. -­‐3   -­‐1   0  18. The spare room subsidy is not a tax -­‐1   2   3  19. It is good to see people making the first payments - long overdue -­‐2   0   1  20. The campaign against the Bedroom tax is riddled with misinformation 0   3   1  21. The spare room subsidy is one of many instances in our society where a good idea fails because of a lack of common sense and sensitive implementation 0   1   1  22. Why the big hoo-haa on #bedroom tax? People in council houses should count themselves lucky for the cheap rent they pay in the first place. -­‐3   -­‐1   -­‐3  23. This is nothing but bedroom tax hysteria -­‐2   0   2  24. The BBC are helping to fuel the myth that the spare room subsidy is a tax by only referring to bedroom tax. -­‐1   2   -­‐2  25. It will free up homes for thousands of families living in overcrowded conditions -­‐2   1   -­‐2