How Effectively did Parliamentary Candidates Use Social Media During the 2015 UK General Election Campaign

and Facebook posts (n = 616) made by a sample of candidates (n = 10), focused on 4 main post functions: broadcasting, posters/infographics, organising, and interacting. The results show that overall UK politicians mainly used social media as a unidirectional method of communication, as opposed to interacting. However, this research also found key differences between Twitter and Facebook, the former used much more interactively.

trends with more than quarter of a million posts (Robinson 2015). Parties also began to invest more resources in social media. Reports revealed that the Conservative party was spending over £100,000 per month on their Facebook page alone (Hawkins, 2015), and the Labour party hired the man behind Obama's successful social media campaign, Matthew McGregor (Wright, 2013).
Following the high expectations surrounding the significance of social media, this paper investigates the extent to which social media was used effectively for political communication and campaigning by parliamentary candidates during the 2015 UK general election campaign. It first locates the study in the key academic debates surrounding social media use and political communication, explaining what makes social media different to traditional media outlets in terms of its potential to revolutionise political communication. Fundamental differences include interactive features, and the ability to reach wide audiences at a low cost. The paper then presents primary research into ten parliamentary candidates' social media accounts in an attempt to illustrate how social media was used during the 2015 short campaign. By relating this evidence to the alleged potential effects of social media, the paper considers wider questions about the effective use of social media during the 2015 UK general election campaign.
Following the election, many commentators expressed the view that social media was not used effectively. The Guardian dubbed 2015 as 'the social media election that never was' (Fletcher, 2015, no pagination), due to lack of interactive, innovative and personalised use of optimists about just how radical this change will be. Whilst varying in detail, optimist theories can be divided into two broad categories according to how radical they believe the change will or can be. The more radical optimists are described by Anstead and Chadwick as direct democracy optimists (2009). Direct democracy optimists believe that the internet will undermine representative political institutions, and move society towards a direct democracy, where citizens directly shape policy via the interactive features of the internet. However, more applicable to British politics today is the category of optimism that Anstead and Chadwick (2009) label 'representative democracy optimism'. This category of internet optimism believes that the internet will not replace or destroy representative institutions, but that it can create positive change by reforming and rehabilitating indirect vehicles of democratic participation, such as parties and elections.
Representative democracy optimists have highlighted many features of the internet that could enhance democracy and political campaigning, as opposed to the traditional old forms of mass media usually used for political communication. Lilleker (2006) explains why mass media is now considered an unsatisfactory method of communication during campaigns. Firstly, there are drawbacks with mass media communication from the perspective of the politicians, candidates and parties. This is because the mass media follow their own agenda. They make decisions about what is news, and frame stories in certain ways to suit editorial purposes. This makes advertisements 'the only mass media form over the construction of which the politician has complete control' (McNair, 2011, p. 86). Citizens are also no longer satisfied with the way information is communicated from politicians to the public; 'they want to interact, have their say and have power' (Lilleker, 2006, p. 73). In this sense, the traditional mass media fails to meet the needs of the electorate as a form of political communication.
Jackson and Lilleker believe that Parliament is viewed as an insular community by the public. They argue that a disconnection from politics is created by the perceived proximity between the mass electorate and the political elite. This has been perpetuated by traditional forms of political communication, as 'historically it has been more about information provision and persuasive communication than public dialogue' (Jackson and Lilleker, 2009, p. 235). Political communication through the traditional mass media has been top down and hierarchical, in the direction of politicians to citizens.
With the proliferation of the world wide web in the mid-1990s, optimist academics were enthusiastic about the features of the internet that set it apart from the old media forms of communication, and the possibility of an electronically enhanced democracy. Barber et al.'s analysis of political websites as early as 1997 found that the internet could offer the following for electronically enhanced democracy: inherent interactivity; potential for lateral and horizontal communication; point-topoint and non-hierarchical modes of communication; low costs to users (once a user is set up); rapidity as a communication medium; lack of national or other boundaries; and freedom from the intrusion and monitoring of government (Barber et al., 1997, p. 8). Sparks (2001) added three more structural advantages of using the internet for political communication. These include the anonymity of social actors, discursive requirements, and search mechanisms. The option of anonymity means that members of the public may not be subject to discrimination based on factors such as accent, age, race, or gender (although of course many individuals choose not to remain anonymous on the internet). The analysis of Barber et al. (1997) and Sparks (2001) were mostly concerned with the possibilities for electronically enhanced democracy from the perspective of a citizen, as opposed to the perspective of a politician or political candidate.
One of the main appeals of using the internet for political campaigning from the perspective of a politician is that it ' offers a source-controlled form of communication' (Tedesco, 2004, p. 510). Websites enable politicians to communicate with citizens without going through 'the gatekeeping filter' (Blumler and Coleman, 2015, p. 123) of news organisations and other mainstream media outlets. These media outlets often skew, frame or edit the content or tone of the communication, either to suit their own political agenda, or to create sensationalism to generate profits. Ward and Gibson (2003) also argue that being able to bypass the mainstream media through the internet is an effective way of delivering information to voters without having to worry about media distortion. Although this advantage of the internet is mostly explained from the perspective of politicians, it is not limited as an advantage to only these actors. Citizens may also have something to gain by obtaining information from the primary source, or at least having the option to do so.
The second and perhaps most significant difference between old media and the internet is the interactivity that the internet can facilitate. This is useful for many reasons. Firstly, it gives the internet the potential to change the nature of political communication and relationships between politicians and citizens. Through old media methods, political communication largely remains a hierarchical one-way communication method from the one to the many. Interactive features of the internet can transform this into a more symmetrical relationship where the public are able to participate. Lilleker argues that with the advent of the internet, ' communication is two-way, between public and political, rather than top down and in persuasive form, and both parties are influential and influenced' (Lilleker, 2006, p. 72). This is not only attractive to citizens as they can be more actively involved in the political process, but it can also be a useful tool for MPs to receive feedback. Ward and Gibson (2003) suggest that this could be the cause of a huge change in electioneering, as the ongoing dialogue with voters and interactive features of the internet enables MPs to receive feedback almost instantly.
Not only does the internet allow many-to-many dialogue, it also allows one-to-one communications (Lilleker, 2006), enabling politicians or candidates to make personal connections with potential supporters, and, as these relationships are more symmetrical, individuals may also reach out to candidates directly in this way. This potential change in the nature of this relationship is of particular significance as it can help to bridge the perceived gap between politicians and voters. Blumler and Coleman point out that digital interactivity can in theory lead to a ' diminution of the boundaries between production and reception' (Blumler and Coleman, 2015, p. 122). In turn, people are more likely to turn out to vote, and to vote for the candidate they have built a connection with online. Therefore, the interactive features of the internet could be an invaluable tool for candidates to gain votes through means that would otherwise have been almost impossible.
The internet can also provide a relatively cheap way for a candidate to communicate when compared to costs of other electioneering methods. For example, e-mailing potential supporters can be cheaper than sending letters, making phone calls, or personal visits. It may also be cheaper to produce an online poster, rather than printing and distributing multiple posters. Therefore, the internet and the interactive features that it fosters can not only help political candidates to build relationships with their supporters, but can also 'more cheaply, effectively, and efficiently mobilize them to participate within a campaign' (Lilleker and Vedel, 2013, p. 402).
As well as reducing costs for politicians, the internet can reduce the costs for citizens seeking to participate in politics. Ward and Gibson (2003) argue that the internet lowers the costs of participating in politics and joining political organisations for citizens. Citizens can now join a political party from the comfort of their own home, and learn about the policies that parties advocate with less effort than ever before, due to the ' capacity of the internet to provide vast resources of data at any time to any wired location' (Lilleker and Vedel, 2013, p. 402).
The internet could also provide a means to tap into sections of the electorate otherwise unlikely to be engaged with politics. Throughout the optimist literature there is enthusiasm about the potential to reach young people through the internet, specifically the ' e-generation'. Ward and Gibson describe this generation as a section of the electorate that 'have grown up with technology, but are traditionally much less likely to vote or join traditional political organisations like parties' (Ward and Gibson, 2003, p. 190). The ' e-generation', being familiar with the internet as a major part of their everyday lives, are likely to 'bump into' politics via the internet, and feel much more comfortable approaching politics from this source. Having a website or internet presence can also signify to the e-generation that a political candidate is 'in touch'. If a candidate can appeal to this section of society with a traditionally low turnout, and mobilize them to vote, it could be the winning factor for their campaign. Oates (2008) supports this theory by arguing that there is evidence that the internet is effective in engaging younger people in politics, and crucially, that this would not have been possible with the old media methods.
Optimism about the potential that the internet can have on political communication was further proliferated by evidence of emerging new features of the web, known as Web 2.0. Key features of Web 2.0 facilitate greater flexibility in online communication, in terms of communicative styles, and how, when and where information is accessed (Jackson and Lilleker, 2009). These features include the ' ability to build networks that connect individuals and organizations within a community, where information is shared and adapted and updated by all members of the community who choose to participate' (Jackson and Lilleker, 2009, p. 233). Citizens or members of the electorate are therefore not restricted to consuming information from the internet broadcast by political candidates, but they can also interact in such a way that enables them to co-produce the content of sites. Possible examples of this include the use of forums enabling site visitors to instantaneously contribute towards discussions displayed openly, as opposed to guest books on Web 1.0 websites, which constitute a more limited form of participation and interactivity online.
Despite the optimism surrounding the possibilities that the internet could have for political communication, Loader and Mercea note that 'the utopian perspectives of the first generation of digital democracy were quickly replaced by findings that documented the myopia of such visions' (Loader and Mercea, 2012, p. 1). Much of the internet optimism was suppressed by early analysis of political candidates' websites. The general observation was that candidates were using the internet merely as another place to host brochures, leaflets, and other informative content. In their analysis of candidate websites in the 2001 UK General Election, dubbed the first internet election, Ward and Gibson found that 'the use of the Internet was patchy and websites often acted as little more than static on-line leaflets' (Ward and Gibson, 2003, p. 188). They also spoke of a ' consensus of evidence which suggests that it had a negligible impact on the election'. (Ward and Gibson, 2003, p. 200). Political candidates were failing to utilise the interactive features of the internet; the key features that optimists believed would set it apart from old media communication methods.
Despite disappointing previous utilisation of the internet for political campaigning, ' a fresh wave of technological optimism has more recently accompanied the advent of social media platforms such as Twitter, Facebook, YouTube, wikis and the blogosphere' (Loader and Mercea, 2012, p. 1). When discussing the potential for the internet to transform political communications and campaigning, Graham et al., argue that 'with the rise of social media, scholars have once again envisioned its potential' (Graham et al., 2013). The advent of social media, and the increasing popularity of such sites amongst the public, gives more credibility to the previously envisioned potential advantages that the internet could bring to political communication.
Social media has now become a revolutionary aspect of Web 2.0. It harnesses the features associated with Web 2.0 through easy to use sites with low barriers to entry, where users can create their own personal profiles to interact with others. Social media facilitates interactivity and networking between the electorate and political candidates with even more ease than elements of Web 2.0 already available on candidate websites. This gives the internet the potential to solve problems associated with traditional media political communications through the use of social media. Loader and Mercea (2012) recognise the potential that the networked core of social media has to reconfigure communicative power relations, as ' equipped with social media, the citizen no longer has to be a passive consumer of political party propaganda, government spin or mass media news' (Loader and Mercea, 2012, p. 3).
As discussed earlier, this interactivity is also an advantage to political candidates or MPs. McNair observes that 'the emergence of Twitter, Facebook and online social networking opened up another, and in some ways more attractive channel for campaigning politicians, and those in government, to communicate their messages 'unmediated'' (McNair, 2011, p. 86). Once again there is optimism that the internet will change the nature of relationships between citizens and politicians due to the advent and growth of social media online, as social networking sites are 'not compatible with the top-down, elite-to-mass style of political communication that is traditional of political parties' (Jackson and Lilleker, 2009, p. 233). Graham et al. (2013) acknowledge that due to its interactive nature, social media could be the key tool of the internet that could bridge the gap between politics and the public.

Section three: Methodology
The previous section discusses how the unique features of social media, namely its interactivity, have the potential to lead to more effective communication between politicians and the public. In order to assess how effectively social media was used by political candidates during the 2015 UK general election campaign, the way candidates used social media will be analysed. This research will assess the extent to which candidates were using only the most basic functions available, their social media campaigning reflecting the traditional top-down communication flows of offline campaigning; or whether interactive functions central to the optimism about social media campaigning were utilised. Whether candidates held social media accounts and how active they were on these platforms will also be considered.

Sample
The sample of parliamentary candidates was selected from the list of 650 MPs on the UK parliament website (UK Parliament, no date). All MPs on this list would have been parliamentary candidates during the 2015 general election campaign. I note that there may be a sight bias here in the sense that all of these candidates were successful in winning the election, and therefore we might expect them to employ a more successful use of social media. However, a full list of candidates is not widely publically available, and most unsuccessful candidates tend to delete their campaign social media pages after losing an election. Initially every 32 nd MP was taken from the list, to end up with 20 MPs. As there was no way of knowing how long it would take to code all of the social media posts made by 20 MPs, the posts of every 2 nd MP from the list of 20 were coded first. This ensured that if the sample size of MPs needed to be reduced at a later date due to time restrictions, it could be reduced to a sample of 10 MPs, whilst remaining a random sample. Due to time restrictions for coding the social media posts, the sample resulted in being these 10 MPs.
MPs that did not hold a particular social media account during the observed time period were not removed from the sample. Not holding an account signifies the lowest possible effective use; no use. Therefore, including these MPs contributes towards measuring the overall effective use of social media by parliamentary candidates.
MPs that held a private account on at least one of the social media platforms were removed from the sample. This was the case with Laurence Robertson MP, who had a private Facebook account. In order to view Mr Robertson's use of Facebook, I would have to be accepted as a 'friend' first. In this case it would have been inaccurate to draw conclusions about his social media use as the data would be incomplete. The next MP on the list from the UK Parliament website was selected instead.
Both Twitter and Facebook social media platforms were used for the analysis, as these were the two most actively used social media platforms in the UK in 2015 (Kemp, 2015, p. 343). Additionally, both platforms take on a similar format in the sense that a user makes posts that can be located on the user's profile page, and users have similar options in terms of what they are able to post. This means that data from both platforms can be grouped together to provide an overall picture, and also compared.
Each post on Twitter and Facebook made by each of the 10 MPs in the sample from the 13 th April 2015 to the 26 th April 2015 was analysed and coded (n = 616). The data was restricted to a 14-day period in order to make the research manageable, whilst also keeping the results as representative as possible. It was considered more accurate to analyse 10 MPs' social media posts from a 2-week period, than the posts of 20 MPs from a 1-week period, as there are some post types on social media that would not be expected to occur as frequently as every week, but likely to occur once within a 2-week period. These particular dates were selected as they marked mid 2 weeks of the short campaign for the 2015 UK general election, a time period in which I would expect to see an accurate representation for the levels activity and the range of functions used during the election campaign.

Coding structure
Each social media post was coded in three stages. First, posts were coded as one of four functions: 1) Broadcasting; 2) Organising; 3) Posters/Infographics; or 4) Interacting. Other studies have categorised posters/infographics and organising under the broadcasting function (Ross et al., 2015;Graham et al., 2013). My decision to analyse these as four separate functions is based on recent literature identifying the significance of using social media for sharing campaign posters (Campbell and Lee, 2015) and for organising campaign events (Lilleker, 2015). However, thus far there has been no empirical analysis of UK parliamentary candidates' social media accounts exploring these areas in depth.
I have already discussed that the Interacting function signifies the most effective use of social media; the potential for interaction has fuelled social media optimism, and sets it apart from other forms of media communication. If a post was not used for this Interacting function, placing a non-interactive post down in one of the other three categories creates a clearer picture of exactly how social media was used. A post with a Broadcasting function would constitute the most basic, and therefore for the least effective use of social media.
Broadcasting can be carried out through any other media form. Organising suggests an interaction between social media campaigning and traditional campaigning. Whilst using Posters/Infographics would suggest that a political candidate is employing more of the features available on social media than a Broadcasting post, it would also signify that candidates' use of social media is simply reflecting their use of the Web 1.0.
Posts were then coded into a sub category of the function, which categorised posts based on the type of information included. Posts were finally coded as an exact post type within the sub category, giving a more detailed picture of exactly what was included within the post (see Figure 1 for visual representation of coding structure) (see Appendix 1 for full Coding Frame). The focus of the analysis was not in terms of content such as themes, ideas, or subject matter, but rather the type of posts in terms of the elements that made up each post, and the mechanisms utilised. Although a majority of posts were only coded 3 times, once for each function, subcategory, and exact type, there were a small number of more complex posts that were coded more than the standard 3 times (n = 10), as they proved difficult to categorise.

Section four: Overview of the Data
This section will briefly outline the data collected to give a picture of what the dataset looks like as a whole, before moving on in subsequent chapters to analyse in more depth the breakdown of the data in terms of the functions of social media posts.

Sample
Initial observations suggest that a parliamentary candidate was very likely to have a social media account during the short campaign, with all of the 10 MPs in the sample holding an account and being active on at least one social media platforms during the two-week period analysed. 8 out of 10 of these MPs held a Twitter account during this time, with Rosie Winterton and Luke Hall as the only two MPs without a Twitter account. A slightly higher proportion of MPs held Facebook accounts during the mid-two weeks of the short campaign with only 1 MP, Gavin Robinson, without a Facebook presence. However, it is worth noting that there are some MPs who did not make any posts on one of the social media platforms, despite holding an account for that platform; both Diana Johnson and Desmond Swayne did not post on Facebook, even though they held a Facebook account. Although a candidate is more likely to have held a Facebook account over a Twitter account, candidates are more likely to have been active on Twitter (see Table 1).

Totals and averages
The total number of social media posts made by all 10 MPs during the two weeks was 616. The average number of posts on social media (mean) for the two weeks was 61.6, and the middle value (median) was 44. This ranged from the MP with the least total number of posts at 1 (Luke Hall MP), and the MP with the most social media posts at 170 (Ian Murray MP).
Of the total 616 posts, 453 were made on Twitter, in comparison to the 163 Facebook posts. Twitter posts therefore made up 73.54% of all posts. The average amount of Twitter posts per MP (mean) was 45.3, and the median 33. The number of Twitter posts made by MPs during the two weeks ranged from 0 (Rosie Winterton and Luke Hall) to 125 (Stella Creasy). It is important to note that the two MPs with the fewest Twitter posts did not hold a Twitter account between 13 th and 26 th April. If these MPs were removed from the sample, the results would differ slightly. As explained in the Methodology, it is more useful to include these MPs in the sample and calculations for the purpose of this study.
The 163  Again, I note that Gavin Robinson did not hold a Facebook account during this time, and if he were removed from sample, these averages would be slightly different (see Table 2).
Immediately it is clear to see that the amount of social media posts differs significantly between each MP, making it hard to generalise the extent to which MPs used social media during the short campaign. These results also show that Twitter posts were much more frequent than Facebook posts for politicians during the short campaign, with 47.08% more posts on Twitter than Facebook.

Activity rates
An average social media activity rate per MP per day was calculated by taking the number of posts across all MPs and from both Twitter and Facebook for the two-week time period, finding the average per MP, and dividing this by the 14 days. Overall, the social media activity rate for the average MP was 4.4 p/d (posts per day). The social   Table 3).
These average activity rates point towards effective use of social media, as they suggest that a majority of parliamentary candidates are likely to have posted on social media each day during the short campaign. However, there is a large range in the activity rates between each individual MP, so it is difficult to say with certainty that a candidate will have been as active on social media as the average rate of 4.3 p/d. These activity rates also suggest that Twitter was used more frequently per day than Facebook, despite fewer MPs holding a Twitter account, with Twitter activity on average nearly three times higher than Facebook activity. A majority of the MPs in the sample individually support this observation, with only two MPs showing a higher activity rate on Facebook.

Initial findings
This research shows that a candidate is very likely to have held a social media account, with all MPs in the sample holding at least either a Facebook or Twitter account. Merely holding a social media account is an important first step towards the effective use of social media for political campaigning. Whilst the data suggests that candidates were more likely to have held a Facebook account than a Twitter account, all of the results show that they were likely to be more active on Twitter than on Facebook, even when including MPs that didn't have a Twitter account in the calculations.
Although it has been possible to use the data to work out average activity rates per day, it is not easy to infer how much any individual parliamentary candidate is likely to have posted on social media during the short campaign using these averages. This is because the range is large between the MP(s) with the highest number of posts and the MP(s) with the lowest, so the data is not tightly clustered around the mean. Therefore, it will be more useful to look further into the data for similarities in the way that MPs used social media, to see if there were any trends in which functions were either used effectively or neglected, and whether MPs utilised the features central to the optimism about the use of social media for political communication; namely interactivity.
The figure below (Figure 2) shows the broad breakdown of posts across social media into the four main function categories. The majority of posts across the social media platforms combined were not utilising the interacting function, with only 27.92% of posts categorised as interacting. However, the amount of interaction was much higher on Twitter than on Facebook. I will go on to analyse these functions further in subsequent chapters.   The first of the four functions that posts were coded into was broadcasting (see Figure 3 for broadcasting coding structure). This is defined by Graham et al., as 'unidirectional communication' (Graham et el 2003, p. 707). Broadcasting is the most basic function, mirroring the traditional form of political communication. Therefore, it can be clearly set apart from any social media posts with an interactive element. The overall number of posts categorised as broadcasting across both social media platforms came to 391 out of a total of 616 (63.47%), making it the most common post function by a large margin, with 35.5% more posts than the second most common function.
Whilst broadcasting was the most popular function on each social media platform, it made up a noticeably higher proportion of posts on Facebook at 79.14% (129 posts) compared to 57.84% (262) on Twitter. These percentages are consistent with Graham et al.'s analysis of MPs using Twitter during the 2010 UK general election short campaign, with broadcasting making up 68% of all Twitter posts (Graham et al., 2013, p. 703). Tweets with an organising function, or with posters/infographics, were not coded under broadcasting in this study, but were for Graham et al.'s research, perhaps explaining this 10% difference.
This difference between the main two social networking platforms was also consistent with each individual MP's posts, with Facebook having a significantly higher proportion of broadcasting posts than Twitter for each of the MPs holding an account on both. Despite this difference, a majority of MPs still had broadcasting as their most common social media post function on both Twitter and on Facebook (see Table 4).
The broadcasting posts were further coded into subcategories. These consisted of a) own material, b) shared links, and c) retweets or shares (of organisations).
On Twitter the most common overall subcategory within broadcasting was own material with 59.59% of with 66.41% of broadcasting tweets (see Figure 4). The own material sub category describes posts that did not   include any links to other sources, and encompasses the most basic of the post types. On Facebook shared links was the most common post subcategory within broadcasting, and the most common subcategory of Facebook posts overall, with 42.33% of all Facebook posts including a link. This is consistent with Ross et al.'s research on MPs using Facebook in New Zealand, which found that 43% of Facebook posts with text included links (2015, p. 261). These findings support the observation that 'politicians use Facebook as a signposting platform, providing short teasers to fulllength policy documents or longer commentaries' (Ross et al., 2015, p. 261). Despite this, own material still made up just under half of all broadcasting posts on Facebook.
Own material included three different exact post types: standard (just text), own image, and own video. Of these, the most common type found on Twitter was standard (just text) (Figure 5). This was also the most common exact post type within the broadcasting function on Twitter. However, standard posts were the second most popular post type on Twitter overall, after conversation with citizens, with only just over a 1% difference between the two.
On Facebook, the most common type of own material post, as well as the most common exact post type on Facebook overall, was own image. Many MPs were using Facebook to update the public on aspects of the campaign trail with attached photographs of campaigners or door knocking (see Figure 6), or sharing pictures of events they had attended both in parliament and in constituencies (see Figure 7).
The finding that broadcasting posts made up over half of all posts on social media is consistent with expectations and other studies (Graham et al., 2013;Ross et al., 2015). This suggests that overall candidates were overwhelmingly using social media to broadcast information, rather than to organise, share posters or infographics, or to interact with others. There were also few differences between each individual MP and their proportion of social media posts that had a broadcasting function, making this an accurate description of the way in which candidates used social media during the short campaign. However, the way that candidates broadcast information differed on Twitter and on Facebook, with Facebook being a much more visual form of social media, and hence the higher percentage of images used alongside a broadcasting post. Although broadcasting was the most common post function of Twitter overall, it is worth noting that the most common exact Twitter post type was not of a broadcasting function, but an interacting function.

Function 2 -Posters/Infographics
Posters/infographics was the second function that posts could be coded as (see Figure 8 for posters/infographics coding structure). Although it may be argued that this would simply fall under the definition for the broadcasting function, recent academic discussions regarding campaign posters creates a particular interest in analysing posters/infographics separately. Whilst some labelled the 2015 election as the death of the campaign poster (Wheeler, 2015), others made the observation that the campaign poster 're-emerged online, particularly through party presences on social media' (Campbell and Lee, 2015,    p. 44), and highlight the potential for a campaign poster to go viral as a particular interest. Across both social media platforms, posters/infographics made up a total of 8.12% of all posts, with a frequency of 50. Although this number seems relatively small, this means that just under 1 in every 13 posts included a poster or infographic; quite a large amount for such a specific type of social media post. This makes posters/infographics the third most common function of social media posts of the four functions identified (see Table 5).
When investigating the high percentage of posters/infographics, it became apparent that Rosie Winterton did not manage her own social media presence, which existed only on Facebook. Her number of posts was relatively low at 14, and all but one post looked to be complete copies of emails likely sent out by the Labour Party Headquarters, including either a suggested online poster/infographic, a link related to the Labour party, or both. It seemed that these had been picked up routinely by staff, with the spiel of information sent out by the central party copied and posted directly to Rosie Winterton's Facebook, including any link or poster sent (see Figure 9).   The one exception of Rosie Winterton's posts that did not include posters or shared links further supports this theory. Although this was a broadcasting own material post including an image, the caption was written in the third person, suggesting it was written by staff (see Figure 10). This explains why the data for this individual MP did not fit the general trend.
In isolation, Twitter also followed the overall trend of posters/infographics as the third most common post function, making up 4.86% of Twitter posts. The percentage on Facebook was higher at 17.18%, making this the second highest of the four functions on Facebook.
All posts coded as posters/infographics were further coded into two subcategories of a) party political and b) other, other including infographics from organisations such as charities, think tanks, and industry related. It's not surprising, especially given that the data was collected in the middle of an election campaign, that a significant majority of posters were party political, with only one poster across both social media networks coded as other. This particular post still comes across as partisan, and looks to be an attack on an opposing party, despite being produced by an independent source (see Figure 11).
Party political posters/infographics were further coded to distinguish whether they were promoting the MP's own party (positive), or criticising an opposing party (negative). It could be argued that the use of social media for presenting negative campaign posters is an ineffective use, as 'this kind of behaviour discourages citizens instead of engaging them with politics' (Graham et al., 2013, p. 708). Most MPs did not post any negative campaign posters on any of the two social media sites, with only Ian Murray posting one negative campaign poster on Facebook (see Figure 12), and Desmond Swayne accounting for all seven negative campaign posters on Twitter (see Figure 13).
In one respect, 8.12% is quite a high proportion of posts, as posters/infographics are a very select type of social media post, but given the fact that the data was taken from the middle of the short campaign, overall the proportion of posters/infographics is lower than expected.    The few MPs in the sample with a high number of posters or infographics demonstrate that there was scope for MPs in the sample to have posted more of these. It is interesting to note that posters/infographics made up a slightly higher proportion of posts on Facebook than on Twitter. This may be because Facebook is a much more visual social media platform, with Twitter users having to click to expand an image on Twitter, making them more likely to see a full image on Facebook. However, the difference between the two social media platforms was not significant enough to determine that Facebook was likely used more for posters and infographics by all candidates during the short campaign.

Function 3 -Organising
The third possible function for posts was organising. (See Figure 14 for organising coding structure). Again, most organisational social media posts could also fit into the definition of broadcasting. However, these were coded separately due to their distinct purpose. Looking at organisational posts separately also provides the means to investigate whether, and to what extent, online social media campaigning supported and interacted with the more traditional offline campaigning activities. Lilleker (2015) argues that the triumphs of social media came down to its ability to mobilise activists through organisational posts.
Organising posts made up a total of nine posts across both social media platforms, making it the least common function with a share of 1.46% of posts (see Table 6).
The number of organisational posts on Twitter and the number on Facebook were similar, with five in total on Twitter, and four on Facebook. Due to the high volume of Twitter posts, organisational posts made up a higher percentage of posts on Facebook with 2.45%, compared with 1.10% on Twitter. However, this was skewed by Ian Murray's five organisational posts on Facebook, with all other MPs having no organisational posts on Facebook, and either one or zero on Twitter. Therefore, due to numbers being so low, it is hard to ascertain a distinct difference between Facebook or Twitter in terms of which was used more frequently for organising.
Organising posts were further coded into two subcategories; campaign related, and other. This purpose of this was to check whether the organising posts were actually campaign related, and therefore a more effective use of social media for political campaigning than promoting events not linked to the campaign. As expected when analysing posts from the short campaign period, the majority of events were campaign related, with only two organising posts coded as other, both on Twitter.
Campaign related organising posts were further coded into exact types of post, specifically whether they included links to event advertisements on other sites. These were coded to check for possible interaction between different social media platforms and between social media and candidate or party websites. Rose et al., refer to this as 'signposting', and argue that this can increase exposure for candidates through directing the audience to their own websites, where their characters are not limited (2015, p. 261). Organising posts coded as other were not further categorised, as these posts, being unrelated to the campaign, could already be disregarded as ineffective. Of the campaign related organising posts, the most popular post type was a standard post, including no links. This was also the most common of all posts coded as organising, with five posts. (see Figures 15 and 16) Stella Creasy was the only MP to share a link when posting with an organising function on Twitter, which she did once (see Figure 17), and the only MP to do the same on Facebook was Ian Murray MP.  I can conclude quite confidently that MPs were not using Twitter or Facebook to organise, whether for campaign related events or other. Only one MP posted more than a single organising post, and the data in terms of the proportion of posts was pretty consistent across all MPs in the sample, with organising posts making up less than 3% of all social media posts in each case. Candidates seemed to have under-utilised a useful function of social media; lowcost advertising for events. Instead, the MPs in the sample seemed to document events after they had occurred, rather than try to provoke interest in these events beforehand.
Due to such low numbers of organising posts, it is difficult to be certain of any differences between the use of Twitter or Facebook to organise. In order to do this, a much larger sample of social media posts, and perhaps even the whole population of parliamentary candidates, would need to be analysed. However, based on my research, I would expect any differences to be very minor.

Function 4 -Interacting
The interacting function is the final of the four functions (see Figure 18 for interacting coding structure). As discussed in the second section, much of the effectiveness of social media use for political communication and campaigning can be measured by looking at the extent to which it is used in an interactive way. Therefore, comparing this function to the other three functions of posts will enable me to draw conclusions about how effectively social media was used during the 2015 UK general election campaign.
Across both social media platforms there was a total of 172 posts categorised as interacting, making up 27.92% of all posts. This makes interacting the second most common function of posts made by candidates on social media during the middle two weeks of the short campaign.
Most MPs had interacting as their second most common social media post function. However, the data for each individual MP was not tightly clustered around the average percentage. It ranged between 0%, with four MPs not using social media interactively at all, and Gavin Robinson's 75.47%. The MP with the highest frequency of interacting posts was Stella Creasy at 65 posts (see Table 7).
There was a striking difference between Facebook and Twitter and the amount of      Swayne). Two MPs had interacting as their most common function on Twitter, making up 75.47% of Gavin Robinson's Twitter posts, and 52% of Stella Creasy's Twitter posts. Aside from these two high values, and the two MPs that had no interacting posts, the rest of the data for interacting posts on Twitter ranged from 6.45% of posts, to 37.89%. This means the rest of the data was fairly spread out, and not tightly clustered around either the highest, lowest or average value. The overall range for interacting posts on Twitter was between 0% and 75.47%.
Interacting posts were further coded into subcategories: a) requesting interaction, b) conversation, and c) retweet/shares (individuals).
Conversation was the most common sub category of interacting posts for both social media platforms, making up 76.83% of all interacting posts on Twitter, and 87.5% of interacting posts on Facebook (see Figure 19). Within the subcategory conversation, social media posts were further coded based on who the conversation was with. There were six different types of conversation posts, coded for the different possible types of actors involved in the conversation. The purpose of this coding was to check whether conversation posts were actually with ordinary citizens, and therefore enhancing political communication, or if conversations on social media mimicked the kinds of professional interactions a parliamentary candidate would be likely to have offline on a day-today basis, for example with journalists, organisations and other MPs. The most common exact post type within the conversation subcategory was conversation with a citizen (see  With interacting as the second most common function for all social media posts, these results may initially appear to support optimist theories. However, in this case the overall trend may not be that useful in describing the way that social media was used during the short campaign, due to the substantial difference between the trends in the data for Twitter compared to Facebook. We can conclude from the data analysed that parliamentary candidates used Twitter more interactively than Facebook, with candidates very likely to have posted an interactive post on Twitter during the 2015 short campaign, and unlikely to have done so on Facebook. It is also promising to observe that on Twitter a post that involves a conversation with a citizen was the most common post type, which was even more frequent than a standard broadcasting post with just text; the standard Twitter post that we might expect to see most often. It is important to consider that although most MPs used Twitter at least once for the interacting function, the amount of interactive posts made on Twitter differs significantly between each individual MP. It is therefore impossible to predict how much any individual MP will have utilised Twitter in an interactive way.

Section six: Conclusions
This research has shown that in the run up to the 2015 UK general election, most parliamentary candidates recognised the opportunities that social media could provide for campaigning. All MPs in the sample held at least one account on either Facebook and Twitter during the mid-two weeks of the short campaign, and each MP was active on at least one of the social media platforms. However, there were some slight differences between Twitter and Facebook; although candidates were more likely to hold a Facebook account, activity rates on Twitter were much higher.
Candidates were overwhelmingly using social media for broadcasting, making up over 60% of posts. When including poster/infographics and organising posts in this total, 72.08% of all posts were unidirectional forms of communication, as opposed to interactive forms, which made up 27.92% of posts. This suggests that the growth in the use of social media ahead of the 2015 campaign failed to change the nature of political campaigning through creating symmetrical relationships between citizens and candidates. The majority of communication was oneway, from the elite to the masses, mirroring traditional communication methods. Furthermore, candidates failed to utilise the potential that social networking sites have to help organise and publicise offline campaigning events through cheap and instant advertising to mass audiences; only nine out of the 616 posts had an organising function. This is an area where online support and connections could have been transformed into tangible support, by mobilizing volunteers to help with offline campaigning activities.
The lack of campaign posters/infographics found on candidates' social media pages was also surprising. There seemed to be a huge range of posters available to share, as demonstrated by Desmond Swayne, who shared 14 posters/infographics, and Ian Murray who shared 17, whilst most other MPs shared less than five. It is interesting to note that no single poster was shared more than once by an individual MP, and no single poster was posted by more than one different MP. Parliamentary parties were likely creating more posters because of the lower cost in producing and sharing posters via social media, without considering how this would limit the potential for any single poster to become iconic, memorable, or go 'viral'.
There are many good reasons why candidates may have chosen not to fully embrace all of these social media features, specifically interactive functions. Candidates focus their efforts on constituents when campaigning, and replying to non-constituents online could be a costly waste of time (Ward and Gibson, 2003). Jackson and Lilleker (2009) suggest that politicians may be even more cautious on social media during the short campaign than in-between elections, due to intense media scrutiny during this time. Interacting on social media could also create a ' communication chaos' (McNair 2006(McNair , 2011 as  (Harris, 2015, no pagination) (Robinson, 2015, no pagination).

Figure 21:
Conversation with citizen on Facebook (Reid, 2015, no pagination) (Murray, 2015, no pagination). the message sent out by the parliamentary party can be confused by individual candidates responding with ambiguous and sometimes conflicting replies to citizens' questions. Furthermore, thus far there has been 'little evidence that social media tools can deliver real wins at the ballot box in and of themselves' (Ross et al., 2015, p. 267). Although trends highlighted initially seem disappointing, the data for Twitter alone shows more promise. Twitter was used significantly more interactively than Facebook, with over a third of posts on Twitter having an interactive function. Tweets that showed a conversation with a citizen were also the most common exact type of post found on Twitter. However, the data for each individual MP and the amount of interactive posts on Twitter was extremely dispersed. Therefore, the extent to which a candidate used Twitter for interactive purposes is likely to have varied significantly between each candidate.
Despite most social media posts reflecting offline campaigning behaviours, it can be argued that 27.92% of posts interacting with the public is a relatively substantial level of interaction when compared to other forms of political communication that often foster no level of interaction whatsoever. Furthermore, this is almost a 10% improvement on the levels of interaction found by Graham et al. (2013) in their analysis of parliamentary candidates' Twitter during the 2010 UK general election short campaign. Additionally, it is important not to dismiss other functions of posts on social media as ineffective, as not all advantages of social media stem from the potential for interactivity. Low costs and avoiding media spin are key reasons why a candidate may choose to use social media to broadcast information, as well as potentially reaching younger audiences.
Although social media could have been used more interactively by parliamentary candidates during the 2015 UK general election campaign, and much of the potentially useful functions were not utilised, it is accurate to say that as social media has become more mainstream, candidates are starting to recognise the potential that it harnesses, and are using social media more and more effectively.