How did politicians in the UK use Twitter during the 2015 General … · 2016. 7. 5. · How did politicians in the UK use Twitter during the 2015 General Election. Abstract. This

How did politicians in the UK use Twitter during the 2015 General Election.

Abstract.

This paper seeks to understand how politicians in the UK used Twitter during the 2015 General

Election. It analysed a sample of over 5,200 tweets collected between May 1st and May 11th 2015.

This paper employed a mixed methods approach utilising both qualitative and quantitative research

methodologies. This paper was not meant as an exhaustive study but instead aims to answer specific

questions related to Twitter usage and the size of the candidates’ party. This paper did not only

sample the winners from each sampled constituency but all of the candidates who were standing

and it includes constituencies from all over the United Kingdom (UK) including England, Scotland,

Wales and Northern Ireland. Understanding how Twitter has been used could help to shape party

social media policy in the future although it is important to remember that when it comes to

campaign techniques social media usage is just one of a multitude of methods available to

candidates.

Key Words. Social Media, Twitter, Party Politics, The United Kingdom, Politicians.

Introduction.

This paper seeks to understand how politicians in the UK used Twitter during the 2015 general

election. As noted by Park (2013) “social media have attracted much scholarly attention (p. 1641)”

over the last few years. This study will increase our understanding of how Twitter is, and can be,

used by examining how a sample of politicians in the UK used Twitter during the 2015 General

Election.

This paper will begin with an examination of the existing literature on the subjects of social media,

political engagement and Web 2.0 technologies. It will then move on to a discussion of the

methodology used to select a sample population, collect the data and finally how the data was

processed. It will then move on to a discussion of the hypotheses of this paper, why and how these

were developed. Next will be the presentation of the results of the coding presented in quantitative

format. Finally, this paper will conclude with an analysis of the implications of these results and a

discussion as to the problems with this paper and suggested further research.

Literature Review.

Twitter allows users to quickly and easily communicate with a potentially huge audience. As noted

by Jungherr (2014) social media including Twitter have been added to the toolbox used by many

politicians to spread their message. At the time of writing the TV personality Kim Kardashian West

had around 45.6million followers. If a politician was able to achieve this kind of exposure the

potential benefits could be enormous in terms of helping to spread their message. Indeed it has

been suggested that one of the drivers behind Jeremy Corbyn’s successful bid to become the Labour

leader in 2015 was his effective use of social media. But the extent to which politicians are taking

advantage of the potentiality of social media such as Twitter is questionable.

According to Adi et al (2014) “political networks often resemble closed networks, connecting only

along partisan lines.” Social media offer both politicians and the public the chance to break down

these barriers to communication as Twitter in particular allows users to see and to “follow” the

activities of others whom they would not necessarily normally come into contact with. As noted by

Adi et al social media allows users to engage in genuine dialogue to an extent that may not have

been possible in the past.

The study undertaken by Adi et al sought to analyse the Twitter communication patterns of Labour

Party Peers in the British House of Lords. They found that Politicians used Twitter in a range of

different ways and to different extents. They found that some users were much more willing to

engage in the use of social media than others. Within this they also found that different peers used

their Twitter communications differently. Some peers preferred to simply present and promote the

party line while others sought a more personalised format. According to them “evidence suggests

that for some politicians’ personalization is a clear strategy (p. 5).” On the other hand, Frame and

Brachotte, who studied female politicians in France, found that “unlike British and US politicians who

may choose to construct an online persona linked to “family values” … The French politicians

interviewed were all in overt agreement about the strict necessity to dissociate public and private

spheres.”

A similar study was undertaken by Gudnov and Redek (2014) who studied how politicians in Slovenia

were using Twitter. They found that politicians were largely using Twitter to promote party policies.

They noted that “It is often the case that statements made over Twitter become headline stories in

the media especially when politics is involved (p. 8).” They then go on to claim that politicians in

Slovenia use Twitter as a way to launch stories as their Tweets are picked up by the public and by

opinion leaders such as journalists. Interestingly this is not just restricted to politicians in Slovenia.

Frame and Brachotte found that one of the primary targets of French politicians on Twitter was

journalists (p. 4). Similarly Lee (2010), in his study of the 2010 UK general election found that Twitter

interactions were often aimed at opinion leaders such as journalists.

On the other hand, as noted by Frame and Brachotte simply broadcasting is not the sole aim of

politicians Twitter usage. Twitter also facilitates an interactive feature. Users are able to contact

politicians and politicians are able to seek the opinions of members of the public in a more

immediate way than would previously have been possible. Queensland Premier Anna Bligh stated in

an interview conducted with Grant et al (2010) that “Engaging with the community online is a great

way for me, as Premier, to get feedback on the decisions and actions of my government.” Similarly,

Isabelle Attard claimed in Frame and Brachotte’s 2014 study that she uses Twitter to seek out expert

advice during parliamentary debates so that she can ask relevant and important questions. There

are, however, risks that come attached to this new discussion forum. Vergeer and Hermans (2013)

note that “It is very easy for candidates to forget their role as politicians and act as amateurs by

debating online, without the necessary professional restraint” which as they note can lead to

problems for both the party and the individual involved.

According to research carried out by Coleman (2005) there is a prevailing opinion in Britain that

politicians do not listen to the public enough. His research suggested that this lack of faith in

politicians was leading to a decline in political engagement. These findings were echoed by Zamora

Medina and Zurutuza Munoz (2014) who found that a similar phenomena had happened in Spain as

the public perceived that electoral campaigns were largely top-down campaigns aimed primarily at

grabbing the attention of the media rather than ordinary citizens.

As research suggests social media could be a tool that politicians could use to combat this malaise.

Frame and Brachotte for example found that one of the ways French politicians utilised Twitter was

as a monitoring tool. While this activity is invisible, politicians could show they are engaging in this

activity through the function of re-tweets or possibly replying to a sample of the tweets they read

but the extent to which this is actually occurring is unclear. On the other hand, groups such as the

Islamic State (ISIS) have long recognised that social media is one of a range of tools that can be used

to boost political engagement as noted by Cohen (2015). According to his article ISIS use social

media including Twitter to spread their message in a relatively sophisticated way which earns them

both sympathy and sometimes even recruits. Obviously the costs for joining ISIS are much higher

than attending a political event such as a door knocking session in the UK and yet it is unclear

whether politicians in the UK are taking advantage of this potentiality. These findings related to

increasing engagement are backed up by Junco et al (2011) whose research suggested that Twitter

could be used to boost student engagement and grades. Similary Kasmani et al (2014) suggest that

social media can help to make groups who may otherwise be marginalised to feel empowered in the

democratic process. Their study focused on the elections in Malaysia and their findings suggested

that social media tools such as Facebook and Twitter helped to “encourage the interest of young

people in politics.”

Interestingly this lesson appears to have been taken on board at an institutional level with

government agencies increasingly taking advantage of, so called, Web 2.0 technology to engage with

citizens in an experiment that has come to be termed e-government. Advocates of e-government

such as Small (2012) suggest that e-government can increase engagement with, and trust of,

government agencies as it can help to facilitate openness and accountability. E-government allows

citizens to move away from simply content consumers to content creators through the use of tools

including social media. According to Small one of the leading advocates of e-government is Canada

where many public services are taking advantage of this type of technology although he

acknowledges that the extent to which the public is aware of, and engaging with, these new

technologies is questionable. He claims that “while Canadians continue to use the internet and

government websites to access government information, few use it to engage politically.”

While the potential for online communication, especially through social media, exists there are

practical limits. As noted by Golbeck et al (2010) “personal and meaningful Twitter interactions are

only possible when a limited number of people are using Twitter.” Similarly, the benefits of social

media such as Twitter may be overstated in terms of audience. Fung et al (2013) note that

individuals online, just like any social group, tend to cluster together with like-minded people. This

can create a, so called, echo chamber effect where an idea may seem more popular than it actually

is.

Methodology

In order to answer the central research question, how do politicians in the UK use Twitter, it was

necessary to develop a coherent research methodology. I began by selecting a sample group of

politicians from the UK. This sample was chosen based on a range of factors. Firstly, I split the UK

into its constituent nations; England, Wales, Scotland and Northern Ireland. I then further sub-

divided England, being the largest constituent part geographically, into broad geographic regions;

North, South and Midlands. The sample was selected in this way because I wanted to avoid

geographic bias. This was a possibility if all of the sample had been selected from one place as the

issues affecting different regions are, naturally, different. Issues facing Londoners for example may

be less salient to voters in Newcastle or in Edinburgh.

Once this geographic decision had been made I then began to consider different types of

constituencies. I wanted to study a broad range of constituencies from relatively safe seats such as

Liverpool Wavertree, where Lucianna Berger enjoyed a large majority of over 7,000 votes at the

2010 general election, to more marginal constituencies, such as Dudley North where Ian Austin’s

majority was just 649 votes at the 2010 general election. Finally, the sample was further broken

down in terms of party hierarchies. I wanted a sample which was broadly representative of all levels

of each party from back benchers to cabinet members. Once all of these considerations had been

applied I was left with 18 constituencies with a sample of 3 from each different geographic region. A

table showing these 18 constituencies and the reasons behind their selection is included at appendix

1.

It was decided to select a sample group for two main reasons; firstly, because measuring every

politician who used Twitter in the UK, even just those standing at national level, would produce an

overwhelming amount of data. Secondly, and more importantly, valid statistical inferences can be

drawn from a relatively small sample from the population which in this case is UK politicians who use

Twitter.

Once the sample group had been selected I set up a Twitter account to monitor the activity of all of

the candidates who were standing in each constituency at the 2015 General Election. I found their

accounts using a mixture of google searches and Twitter’s own search API. I found that out of 126

potential individuals in my sample 87 had Twitter accounts.

Once I had identified these 87 accounts I then set a time frame for this study. The data was collected

by copy and pasting my timeline into a word document from May 1st to May 11th 2015. This gave me

a sample of the week before and the 4 days after the 2015 General Election. This timeframe was

deemed appropriate because it would yield significant amounts of data without becoming

overwhelming and it would provide a snapshot of how the sample pool had used Twitter in arguably

the most important time in the 2015 general election.

After the data had been collected it was decided that the best way to analyse how the sample

population had used Twitter, in the timeframe suggested, was to code the data and so I uploaded

my word documents into specialist coding software (Nvivo 11). I then began by coding the data

according to which party the individual who tweeted it was affiliated to. It was then necessary to

develop a coding sheet which would lay down the criteria for how each individual tweet would be

coded. This can be viewed below in table 1.

Table 1.

Category Description

Attack A tweet attacking another political party or person. This does not include vague attacks against outside groups such as “the media.”

Help A tweet asking for help. This could be from an individual or a group. This includes calls for relatively low commitment help such as asking for a retweet or for more substantive help such as a call to come on a door knocking session or to join the Party as a member.

Broadcast An undirected tweet whose main aim is simply to impart some information. No distinction was made in the type of information such as whether it was personal or professional.

Congratulations A tweet congratulating someone for something. Usually a fellow party member for the amount of votes they got but not exclusively.

@ A tweet which displayed a meaningful interaction with another Twitter user. This does not include retweets unless the purpose of the original tweet was to interact with the politician.

Thanks Tweets thanking people or groups for things. This includes thanking constituents for their votes, thanking colleagues for their support and thanking venues for hosting events.

Welsh Tweets in Welsh.

Gaelic Tweets in Gaelic.

It was decided that these codes would be appropriate because they would give a broad idea of

whether politicians in the UK were using Twitter to genuinely interact with other Twitter users, be

that colleagues, constituents or opinion leaders such as journalists, or whether they were simply

using it as a traditional top-down method of one-to-many communication. It was decided that

Tweets in Gaelic and Welsh would not be further classified partly because of the authors lack of

linguistic knowledge and partly because these Tweets are inherently exclusionary as not everyone,

even in those constituencies, can understand these Tweets. But they are interesting in themselves

because they demonstrate that politicians are seeking to appeal to a wider audience. The coding was

performed solely by the author and so coding bias should not occur. These codes allowed me to test

my hypotheses which are shown in the next section.

Finally, it was necessary to analyse the data set. This was achived using the Excel statistical add-on

package.

Hypotheses

H1. Candidates from smaller parties will Tweet more often relative to the more established parties to

make up for their lack of exposure in the mass media.

In his 2014 study Jungherr stated that “we should… expect non-traditional political actors and

supporters of new parties to use Twitter for mobilization purposes.” This paper is not directly

looking at this element of the tweets but the principle is the same. That smaller, non-traditional,

parties and their candidates will use Twitter to make up for the fact that their support networks

including membership and support will be less well established than other parties.

There are many ways in which a party can be measured: number of parliamentary seats; funding;

number of years established. For the purposes of this paper membership size will be the measure

used to determine the size of the party and according to Keen (2015) the three parties with the

largest memberships are the Conservative Party, the Labour Party and the Liberal Democrats Party.

For the purposes of this paper these three parties constitute the large, traditional parties and the

others are referred to as the smaller parties.

H2. Candidates from the larger parties will send out more broadcast style Tweets when compared to

the smaller parties.

As noted by Coleman (2005) voters in mature democracies such as Britain are coming to expect

more from their politicians. Voters want to be heard and they want their politicians to show that

they are listening. By all accounts social media, and Twitter in particular, has the potential to be

interactive and so this type of dialogue with voters can be established but there are limits. As noted

by Golbeck et al “personal and meaningful Twitter interactions are only possible when a limited

number of people are using Twitter. Too many and it becomes unmanageable (2010, p. 1619).” This

paper hypothesises that candidates from the larger parties will simply be too overwhelmed to

respond in any meaningful way.

H3. All of the candidates from Wales and Northern Ireland will tweet bi-lingually.

The official language of Great Britain and Northern Ireland is English but in Wales and Northern

Ireland, Welsh and Gaelic are still spoken. In order to appeal to the maximum number of

constituents it is expected that those candidates standing in these geographic regions will tweet at

least some of the time bi-lingually as this could appeal to some voters and seems unlikely to upset

others. In other words, candidates appear to have nothing to lose by tweeting at least sometimes in

these other languages.

Results.

H1. Candidates from smaller parties will Tweet more often relative to the more established parties to

make up for their lack of exposure in the mass media.

As mentioned Jungherr (2014) suggests that non-traditional, smaller, parties may use social media

such as Twitter to try to engage with their supporters relatively more than their larger rivals. In order

to test this assumption I had to code all of the tweets in my sample (n=5287) and assign them to the

correct party. I then calculated the average number of tweets sent out per candidate in each party.

The results are shown below. For the purposes of this hypothesis the independent candidates were

excluded as they did not represent a homogenous group and the results would not have represented

a valid sample.

This graph appears to show that there is no correlation between the size of the party and the

average number of Tweets their candidate sent out during the sample period (May 1st to May 11th

2015). In order test this initial reaction it was necessary to carry out a t-test assuming unequal

variance with a confidence of 95%. The results of which can be seen below.

020406080

100120140160180

A Graph to show the average number of Tweets sent out by prospective parliamentary candidates

in the 2015 General Election

t-Test: Two-Sample Assuming Unequal Variances

Variable

1 Variable

2

Mean 51.97167 66.8361

Variance 1270.657 2019.14

Observations 3 10

Hypothesized Mean Difference 0

df 4

t Stat -0.59436

P(T<=t) one-tail 0.29213

t Critical one-tail 2.131847

P(T<=t) two-tail 0.584261

t Critical two-tail 2.776445

Variable 1 = The averages for the Labour party, the Liberal Democrats and the Conservatives.

Variable 2 = The averages for all of the other sampled parties.

These results suggest that there is indeed a statistically significant difference between the amount of

Tweets sent out by the Labour party, the Liberal Democrats and the Conservative parties which have

been classified as the three largest parties when compared with all of the others with a critical t

value of 2.78 which is significantly higher than the t statistic of -0.59.

This data also suggests that the initial hypothesis is correct with the mean number of Tweets send

out by the largest three parties as 51.97 compared to 66.83 for the rest of the parties.

H2. Candidates from the larger parties will send out more broadcast style Tweets when compared to

the smaller parties.

As suggested by Golbeck meaningful interaction is only possible of social media when it is of a

limited scope and on a limited scale. In other words, it is only possible to reply to so many tweets a

day. Conversely it is relatively easy to simply broadcast tweets in the traditional top-down format of

past elections. This hypothesis will test this assumption. Again I started off with a simple bar chart to

see if there was any immediately obvious correlation between party size and the average amount of

broadcast tweets the candidates sent out. The results of which can be seen below.

This graph would appear to indicate that there is no correlation between the size of the party and

the average percentage of broadcast tweets sent out by each party but in order to test this

hypothesis further it was necessary to carry out a t-test assuming unequal variance with a

confidence of 95%. The results can be seen below.

t-Test: Two-Sample Assuming Unequal Variances

Variable

1 Variable

2

Mean 62.72727 57.33333

Variance 220.6182 82.33333

Observations 11 3

Hypothesized Mean Difference 0

df 5

t Stat 0.78263

P(T<=t) one-tail 0.234635

t Critical one-tail 2.015048

P(T<=t) two-tail 0.46927

t Critical two-tail 2.570582

Variable 1 = The averages for the Labour party, the Liberal Democrats and the Conservatives.

Variable 2 = The averages for all of the other sampled parties.

The results show that, contrary to my initial assumption after seeing the bar chart, there was a

statistically significant difference between the amount of Tweets sent out by the Labour party, the

Liberal Democrats and the Conservative parties which have been classified as the three largest

parties when compared with all of the others. This is shown by the critical two tailed figure of 2.57

being significantly larger than the t- statistic measured as 0.78.

The data also suggests that I have failed to prove my hypothesis. Variable 1 which represents the

broadcast tweets of the smaller parties has a mean of 62.72 compared to Variable 2 which

represents the average number of Tweets by the candidates of the three largest parties which

stands at 57.33.

0102030405060708090

100

A Graph to show the average number of broadcast Tweets sent out by the sample PPCs at

the 2015 UK General Election

H3. All of the candidates from Wales and Northern Ireland will tweet bi-lingually.

This hypothesis ultimately proved incorrect. The DUP, SDLP, UUP, Alliance Party and Sinn Fein all

stand exclusively in Northern Ireland and yet only Sinn Fein used the Gaelic language on Twitter (in

22% of their sampled Tweets). Superficially this may be surprising as the other parties appear to be

sacrificing these potential voters. However there may be cultural factors which this paper does not

include which may be affecting the outcome in this case. Linked to this another surprise was the

relative lack of tweets in Welsh, excluding Plaid Cymru 23% of whose sampled Tweets were in the

Welsh language. Out of a total of 1,376 tweets in the sample from Labour candidates only 2 were in

Welsh and the Liberal Democrats tweeted just 3 times in Welsh. Despite the fact that all three of the

sampled Conservatives in Wales had active Twitter accounts none of them registered a single Tweet

in Welsh.

Conclusion.

This paper began with a brief summary of the relevant literature concerning the current trends and

themes in social media research in terms of how it can, and is, used in a political context. It was

shown that social media and Twitter in particular have the potential to be a useful campaign tool in

an increasingly contested political world. Adi (2014) for example noted that social media has the

potential to break down former barriers to communication. This review suggested that social media

can be used in a multitude of ways but the main three ways were; to communicate with the

electorate and with other opinion leaders (Adi, 2014; Gudnov and Redek, 2014; Lee, 2010; Frame

and Brachotte, 2014); to create a more personalised campaign (Grant et al, 2010) and to boost the

political engagement of previously marginalised groups such as young people in Malaysia (Kasami,

2014). This paper decided to focus on this issue of political engagement and the dilemma faced by

politicians between broadcasting generalised messages and genuinely interacting with the

electorate.

Next this paper addressed the methods to be used to carry out this research. This was done through

a mixture of non-random sampling techniques, data collection and coding and finally statistical

analysis.

This paper then set out three hypotheses to be tested. H1. Candidates from smaller parties will

Tweet more often relative to the more established parties to make up for their lack of exposure in the

mass media; H2. Candidates from the larger parties will send out more broadcast style Tweets when

compared to the smaller parties And H3. All of the candidates from Wales and Northern Ireland will

tweet bi-lingually.

Finally, these hypotheses were tested and the results presented. The results suggest that smaller

parties do indeed Tweet more regularly. Although it is impossible to prove why at this stage it can be

speculated that this may be to help compensate for their lack of exposure in the mass media

including TV, Radio and Print form. The results also suggest that H2 may not be correct. The results

suggest that, contrary to the initial assumption, smaller parties broadcast more than their larger

rivals. This was surprising for a number of reasons. Firstly, it was assumed that the candidates from

the larger parties would be too overwhelmed with interactions and so would stick to broadcasting

very generalised messages. It was also assumed that candidates from smaller groups may copy the

model of marginalised groups in society such as ISIS by using Twitter to engage with followers. It

clearly costs more to the individuals to go and fight for an insurgent group than it does for an

individual to go out leafletting for their local party but it appears that candidates still struggle to fully

grasp the potentiality of Twitter. Finally, the results also showed that despite standing in relevant

constituencies not all of the candidates used the local languages. This was a surprise because it

seems like an easy and politically costless measure to take, to try to interact with people in their

chosen language. But on the other hand there may be cultural reasons which are outside the remit

of this paper to investigate.

This paper is by no means meant to be an exhaustive study nor are the findings necessarily

generalizable across all times and situations. The evidence presented was collected during a specific

and short time period and more sampling is required over an extended time period to fully answer

the question of how politicians in the UK use Twitter.

Acknowledgements

This paper was made possible thanks to KPES, Palacky University, Olomouc. Thank you to all the staff

and colleagues for their help and advice.

Bibliography

Adi, A; Erickson, K; Lilleker, D (2014) Elite Tweets: Analysing the Twitter Communication Patterns of

Labour Party Peers in the House of Lords. Policy and Internet. (6), 1. Pp. 1- 27

Cohen, J (2015) Digital Counterinsurgency: How to marginalize the Islamic State Online. Foreign

Affairs (94), 6.

Coleman, S (2005) Direct Representation. Towards a conversational democracy. IPPR exchange

Frame, A; Brachotte, G (2014) Le tweet stratégique: Use of Twitter as a PR tool by French Politicians.

Public Relations Review.

Fung, A; Gilman, H. R; Shkabutur, J (2013) Six Models for the Internet and Politics. International

Studies Review (15), 1. Pp. 30- 47

Golbeck, J; Grimes, J; Rogers, A (2010) Twitter use by the US congress. Journal of the American

Society for information science and technology (61), 8. Pp. 1612- 1621

Graham, T; Broersma, M; Hazelhoff, K; van t’ Haar, G (2013) Between broadcasting political

messages and interacting with voters. Information, Communication and Society (16), 5. Pp. 692- 716

Grant, W; Moon, B; Busby Grant, J (2010) Digital Dialogue? Australian Politicians’ use of the social

network tool Twitter. Australian Journal of Political Science (45), 4. Pp. 579- 604

Gudnov, U; Redek, T (2014) The use of Twitter for political purposes in Slovenia. Romanian Journal of

Political Science (14), 1. Pp. 4- 34

Hridina, M; Karaščáková, Z (2014) Parties, Pirates and Politicians: The 2014 European Parliamentary

Elections on Czech Twitter. Human Affairs (24) pp. 437- 451

Junco, R; Heibergert, G; Loken, E (2011) The effect of Twitter on college student engagement and

grades. Journal of computer assisted learning (27), 2. Pp. 119- 132

Jungherr, A (2014) The logic of Political Convergence of Twitter: Temporal Dynamics and Content.

Journal of Communication (64), 2. Pp. 239- 259

Kasmani, M, F; Sabran, R; Ramle, N (2014) Can Twitter be an effective platform for political discourse

in Malaysia? A study of #PRU13. Procedia- Social and Behavioural Sciences (155). Pp. 348- 355

Keen, R (2015) Membership of UK political parties. House of Commons Briefing Paper. (SN05125)

Lee, B (2014) Window Dressing 2.0: Constituency level web campaigns in the 2010 UK General

Election. Political Studies Association. Politics (34), 1. Pp. 45- 57

Park, C. S (2013) Does Twitter motivate involvement in politics? Tweeting, opinion leadership and

political engagement. Computers in Human Behaviour (29), 4. Pp. 1641- 1648

Small, T (2012) e-Government in the age of social media: An analysis of the Canadian Government’s

use of Twitter. Policy and Internet (4), 4. Pp. 91- 111

Vergeer, M; Hermans, L (2013) Campaigning on Twitter: Microblogging and Online social networking

as campaign tools in the 2010 General Elections in the Netherlands. Journal of Computer- mediated

communication (18), 4. Pp. 399- 419

Zamora Medina, R; Zurutuza Munoz, C (2014) Campaigning on Twitter: Towards the “personal style”

campaign to activate the political engagement during the 2011 Spanish General Elections.

Communication and Society (27), 1. Pp. 83- 106.

Appendix 1. Constituency No. Candidates No. with Twitter Type of Constituency Location 2015 Result

Dudley North 7 5 Marginal Middle Labour hold

Stafford 6 5 Tory safe Middle Tory Hold

Bromsgrove 5 4 Tory cabinet Middle Tory Hold

Barnsley Central 7 3 Labour Shadow Cabinet North Labour hold

Liverpool, Wavertree 7 5 Labour Safe North Labour hold

Hexham 5 4 Tory safe North Tory Hold

South Thanet 11 10 Marginal South Tory Switch

Portsmouth South 8 6 Independent South Tory Switch

Eastleigh 7 7 Lib Dem South Tory Switch

Perth and North Perthshire 7 3 SNP Scotland SNP Hold

Edinburgh East 7 5 Safe Labour Scotland SNP Hold

Dumfriesshire, Clydesdale and Tweeddale 6 6 Tory safe Scotland Tory Hold

Dwyfor Meirionnydd 7 5 Plaid Cymru safe Wales Plaid Cymru Hold

Cardiff North 8 7 Marginal Wales Tory Hold

Cardiff West 7 5 Labour Wales Labour hold

Belfast East 6 6 Alliance Party N. Ireland DUP Switch

Ulster 9 4 Sinn Fein safe N. Ireland Sinn Fein Hold

Foyle 6 4 SDLP N. Ireland SDLP Hold

126 94

% of politicians on Twitter

1.26 74.6031746

Appendix 2.

Broadcast Welsh Irish

Party Total Tweets

per Candidate Actual % Actual % Actual %

TUSC 45 15 20 44

DUP 14 14 9 64

UUP 94 47 67 71

SDLP 115 38.333 49 43

Alliance 108 54 51 47

UKIP 183 95.875 684 89

Green 1376 57.153 525 71

Plaid Cymru 537 158 230 49 110 23

Independent 767 180 68

SNP 743 78 165 71

Sinn Fein 474 111 244 73 22 7

Lib Dem 266 22.875 118 64 3 2

Labour 234 91.733 840 61 2 0

Conservative 333 41.307 251 47

Average Per Candidate 337.79

Total 5289