By Cristina Leston-Bandeira and Viktoria Spaiser
Hashtag conversations over Twitter are common place. They are used to comment on TV programmes, conferences, general themes and now petitions. The new Petitions Committee of the House of Commons has been using hashtags to support the development of discussions associated with the themes of the petitions being debated in parliament. Are these discussions on Twitter just a lot of hot air, come and gone, or can they help us understand the different purposes of petitioning? In this blog piece, we find that, instead of just noise, these Twitter discussions help to identify themes linked to petitions, different levels of sentiment associated to petitions, varying levels of polarisation, but also those petitions that despite achieving very high numbers of signatures, actually have little traction.
In the House of Commons, citizens can submit their petitions directly to Parliament and government through a collaborative e-petitions site. This recently-introduced system establishes two key thresholds: if a petition gets to 10,000 signatures, it receives a response from the government; if it reaches 100,000 signatures, it is considered by the Petitions Committee as to whether it should be debated in Parliament. For those chosen to be debated, the Petitions Committee creates a specific Twitter hashtag that is then used by those following and contributing to the debate – either in or outside parliament. This enables people to follow comments made about the debate, gather remarks on the actual debate, and promote discussions between people with an interest in the topic. For the purpose of our pilot study, we chose to listen in into the twitter conversations of the debates of the petitions on Meningitis B (#MenB), BMA contract negotiations (#juniordoctors) and Government’s EU referendum leaflet (#EUreferendumLeaflet). We also listened into the Twitter conversations of an oral evidence session the Petitions Committee held as part of its inquiry on Meningitis B, as a result of its respective petition.
To get a sense of the debate, we performed an ngram analysis and topic modelling of the Twitter data (all tweets with a hashtag representing the petition on the day of the debate, including the preceding and following day) and visualised the results with semantic networks. In the ngram analysis we calculate probabilities for most frequent words appearing together. With topic modelling we identify topic clusters in the data, a bag of collocated words that point to a common theme. Semantic network analysis have their origin in theories of meaning. In the semantic network the nodes are various, frequently appearing words from the tweets data and the edges are connecting two words, that co-occurred in the data, with the thickness of the edge, the weight, representing the association score size: the thicker the edge, the higher the association score and the more significant this co-occurrence in the data. The colours of the nodes (and edges) signify various topics. We also used the location data of the Twitter users who were participating in the debate and plotted these on a UK map to show the geographical representation of the debate in the UK.
Before comparing different debates, we explored whether there were any differences between the reaction to a debate and to an oral evidence session, on the same petition (Meningitis B). The comparison is quite insightful (see Figure 1). The evidence session led to two large topic clusters: one focused on the medical background (e.g. symptoms, long-term effects) of the disease, explaining why a free vaccination is demanded in the petition (because affected people suffer from severe disabilities, risk of dying, etc.); and another summarising the evidence session in Parliament (who organised and chaired it, who the speakers were and what was debated). Besides those two larger topic clusters we have three smaller topic clusters: one on the Byrne plan, another one on families and their testimonies, and finally, one representing a mobilisation theme, asking for support for the petition and its campaign. A simple sentiment analysis of the data shows that these discussions did not represent particularly strong positive or negative feelings; rather, the overall tone seemed to be pragmatic and factual.
This, however, seems to change in the actual petition debate. This data shows a whole theme around feelings of disappointment about the parliamentary response to the petition. The debate also included a medical background theme and a parliamentary process theme, but here these themes are emotionally framed in negative terms (e.g. Westminster Hall MPs lack vision and foresight). We have also again a family-stories/testimonies topic cluster, with a side-theme pleading for a vaccination for children and a mobilisation /petition theme. Overall the analysis of the Twitter conversations suggests that Parliament responded negatively to the petition due to cost considerations. However, the proponents of the petitions don’t accept that reason and show disappointment. One reaction is even indignation that the government is spending money on methadone programmes but no money on vaccinating children against meningitis. Overall, the debate appears calm and fact-based, even though a compromise is clearly not achieved and it is more emotionally charged.
Things are different in the Twitter debate on the #juniordoctors petition (see Figure 2). Here we see good evidence of a strong polarisation and implacability on both sides involved in the debate, Parliament/government, on the one hand, and the petition proponents, on the other. Discussions show that the offer made by the government is rejected by the petition supporters. In reaction to this, the government calls the junior doctors militants and radicals, which evokes feelings of indignation among junior doctors, who “feel utterly crushed” and betrayed by the government, accusing it of being stubborn and intransigent, risking the viability and effectiveness of the NHS and British health system and in consequence people’s life. The NHS/British health system theme moreover summarises the grievances and frustrations that brought junior doctors to the streets. The proponents do not see themselves as radicals rather as saviours of the NHS and British health system, which is threatened by the government. The semantic network finally includes also a mobilisation and a NHS whistle blower theme.
A comparison between the #MenB debate and the #juniordoctors one shows that the latter was much more incensed and polarising, even though in both cases the outcome of the debate was for all intents and purposes the same, a rejection of the points being raised by the original petition. The geographical distribution of those involved in the Twitter discussions shows, moreover, that both debates resonated across the UK (see Figure 3). Both debates led to genuine discussions, which involved and mobilised a good number of different people (around 92 for #MenB and 234 for #juniordoctors). This is not always the case, as shown by the last case study in our pilot: the #EUreferendumLeaflet debate. Despite being set-up in exactly the same format as the other two, this petition debate did in fact not lead to an actual discussion on Twitter. Although the hashtag was used, it quickly becomes clear that this was solely due to the petition committee’s activity. This shows that reaching a high number of signatures does not mean that petitioners and supporters are actually passionate about the topic or care about the parliamentary outcome.
This is a very brief analysis and we don’t claim to take substantive conclusions from this; for our purposes it is simply a pilot study to understand what Twitter conversations can tell us about petitioning. But it does also suggest the following:
- The different reaction to an evidence session and to a debate, suggests that either activity may potentially lead to a differentiated type of impact. There is usually an assumption that the highlight of a petition is for it to lead to a parliamentary debate. However, other petitions system show that evidence taking sessions can be of far more value in terms of ensuring the petition has an impact. Our Twitter analysis confirms that a debate leads to a much more polarised type of reaction, which is less likely to lead to different parts coming together and resolving an issue; although of course it may raise the petition’s profile in terms of publicity.
- The difference in level of polarisation and sentiment expressed between the Meningitis B petition and the one on junior doctors’ contracts suggests again that different types of petitions may lead to different levels of usefulness to have them presented as petitions. More politicised issues may achieve little, other than reinforcement of either sides’ position, as was the case of the junior doctors’ petitions debate.
- Finally, the lack of engagement with the EU referendum leaflet exemplifies well the case of petitions that are not about deep and long term issues that truly affect the public, which may well achieve the 100,000 signatures threshold but may in fact have little importance for those who sign it. It indicates those petitions that “take off” because they’re populist issues, easy to grasp but not necessarily about actual, heartfelt problems.
These are all good clues to explore further in our planned project about petitions in the UK Parliament.
About the authors
Professor Cristina Leston-Bandeira is Professor of Politics at the University of Leeds and co-convener of the PSA Specialist Group on Parliaments and Legislatures.
Dr Viktoria Spaiser is a University Academic Fellow in Political Science Informatics at the University of Leeds, School of Politics and International Studies. For further information, see her website.