Chapter 1: Introduction

The traditional category also underwent several stages of development. Initially this category was labelled Automatic and was determined by tweets which were sent out as automatic feeds from newspaper websites, involving no human process. However when testing the categories the researchers noticed that there were a large number of tweets that had been written by a journalist or web editor but were essentially a rewritten headline and being used a promotional tool for the website. There could be a number of reasons for this including a failure in the automatic feed system or journalists rewriting headlines to make them more suitable to Twitter or to enhance search engine optimisation. It was felt that these tweets formed the same function as an automatic machine processed tweet and therefore the category heading was changed to allow for these. Automatic became Traditional to reflect a traditional form of gatekeeping communication which has no informal / personal traits or interactivity. For example the tweet by @thisisleic ‘New superstore plan for Braunstone Gate, Leicester http://ow.ly/2TUx4’ had been written by the web editor but read like an automatic feed. Technically it had been altered from the original website headline ‘New superstore in Braunstone Gate would 'transform' retail park’ and under the original coding system could not be counted as Automatic. However the revised category Traditional allowed this tweet to be coded alongside automatic feeds which serve the same function.

In contrast the Personal category reflects tweets which are unrelated to their profession and may be directed to friends and family although readers can still see the tweets and respond to them. The Personal category was not an original category but during the development period the Informal unit was divided into Informal and Personal. This was to reflect the interesting ethical issues faced by journalists about their professional and personal roles and the merging of the two on social media, which will be explored further in Chapter 11. Without separating the Personal non-work related tweets from the Informal work-related tweets it would not be possible to explore these issues. During the coding the researcher had to check the links included in tweets in order to determine its category and also the profile of the @username to clarify whether they were a colleague, friend or reader. If it was unclear the researcher checked with the context unit journalist.

In the original coding design the use of the hashtag # within a tweet was classed as a sharing activity as the # is used to mark all tweets on a similar subject, for example #London2012 referring to the London Olympic Games. Users can then click on #London2012 to read all of the tweets from all users on Twitter talking about that subject. It was felt that this was a way of sharing information to a wider community. However during the testing of the categories the researcher discovered that the use and meaning of the # had changed. Many users were including # in their tweets as a way of making a flippant remark, additional comment or summarising their tweet, for example:

martin_crowson: Sepp Blatter wins FIFA vote 172 to 17. Utterly incredible. The ultimate sham. #spineless #jokeorganisation.

In this tweet #spineless and #jokeorganisation do not refer to a trend or commonality as this would be represented by #FIFA or #SeppBlatter - the subject matter being talked about. To complicate things further #SeppBlatter could also be interchanged with @SeppBlatter as a username, as they are an individual with a Twitter account as well as a trending story topic. Due to these inconsistencies it was felt that the # was not a useful determination of the content of a tweet and it was not taken into consideration. The one exception to this was the use of #ff which refers to shared content and came under the Sharing: external category. #ff stands for Follow Friday and the correct Twitter etiquette is to use it on a Friday to recommend the best Twitter users to follow.

As with the comments discussed in section 4.5.1 a dominant category system was used for coding tweets. The categories were tested three times and on the second test a dominant system was chosen over a double coding system. This ensured that the statistics were clear and there was no ambiguity. Also as there were seven categories allowing for a range of descriptors it was felt that double coding was unnecessary particularly as it was calculated that only two per cent of all tweets would be doubled coded if that system was used.

Once the final data was collected and coded it was inputted into a spreadsheet, the results of which are examined in Chapter 11.

4.5.3 Coding Facebook

The use of Facebook at the two case study sites was not as widespread as Twitter, with most journalists preferring to use Facebook for personal use rather than professional reasons. Twitter is an open social network where anyone can follow another user without permission, whereas Facebook is a private network where people are granted permission to view other’s feeds. As such both case study sites had limited Facebook profiles or pages and at both newspapers these were managed by the web team. There were also other obstacles for journalists wanting to interact with readers via Facebook, in that they are banned from accessing Facebook due to a companywide policy. Individuals could apply to have the ban lifted on their computer but this had to be agreed by the editor. As a result very few journalists could access Facebook at work and relied on the web team for this medium. Furthermore the Leicester Mercury did not introduce a Facebook page until half way through the research period.

Data was collected from Facebook mirroring the Twitter methodology to make results more comparable. A two week coding context sample was taken from the Bournemouth Daily Echo during the observation period in January 2011 and a one month sample was taken in March 2011. A sample was not taken in June, as was the case for the Twitter sample, because the digital projects co-ordinator was on maternity leave at this point and it became apparent to the researcher in June 2011 that the Facebook profile was therefore largely inactive. It was possible to go back to the active period of March on Facebook, but due to the limitations of Twitter discussed above it was not possible to do the same on Twitter. A one month sample was taken from the Leicester Mercury in June 2011 as the Facebook page did not exist during the observation period in October 2010 or in March 2011. The Bournemouth Daily Echo had two Facebook presences – one a profile and one a page. The page was set up first and was more traditional in nature, with automatic feeds from the bournemouthecho.co.uk website and some content from the web team. The profile was named Bournemouth Echo Sam and was managed by the digital projects co-ordinator Sam Shepherd. This was a more informal profile and was set up by Sam Shepherd because she felt people would be “more likely to be friends with an individual on Facebook rather than an organisation which might look uncool”. The researcher therefore decided data from the Bournemouth Echo Sam profile would provide richer data than its traditional counterpart, and this was the only Facebook presence to be captured.

The following definitions were used:

• 
  Wall: Each Facebook profile or page has a feed where all of their posts and responses to them can be viewed.
  

• 
  Post: Each individual unit on a wall and its responses. Sometimes referred to as a message.
  

• 
  Posting: The act of putting something onto a wall (message, photo, video, link, comment).
  

• 
  Comment: An external user can place a comment on a post on another user’s wall. A user can also place a comment on their own posts.
  

• 
  Like: The act of effectively ticking a box to say you agree with a post or comment.
  

• 
  Profile: A Facebook account of a user with their personal details. You must have permission to become their friend before you can view their profile wall.
  

• 
  Page: Similar to a profile but set up by organisations or events. You do not need permission to view the page wall but you must Like the page in order to view it.
  

The seven Twitter categories were used as the starting point for the Facebook analysis but were adapted for the different functions that exist on Facebook that are not replicated on Twitter. For example on Facebook it is possible to trace when a reader engages with the newspaper (external interacting with internal) whereas as discussed in section 4.5.2, on Twitter it is only possible to trace the newspaper user when they engage with readers (internal interacting with external).

The Facebook categories were defined as:

• 
  Traditional
  

Promotional link to own website competition

• 
  Informal
  

Informal news (including live updates)

Comment (inc. comment on current affairs)

Personal message to readers

• 
  Newspaper Sharing
  

Newspaper ‘Likes’ external page

• 
  Reader Response
  

• 
  Reader Interactive
  

Comment from reader

• 
  Newspaper Interactive
  

Setting up vote

Newspaper comment in response to reader

The Personal category was dropped because during testing it become apparent that the Facebook profile was not used for personal messages or non-work related comments due to all of the followers (or friends) being readers. The implications of this are discussed as part of the findings in Chapter 11.

Facebook has a more complex interface than Twitter and as a result there are more ways to engage with the content. These ‘engagements’ can include posting a message, posting a photo, posting a video, making a friend, commenting on a message, sharing a message, liking a message, liking a comment or liking a page. Making a friend and updating the profile photo were not coded as it was viewed as a routine operation rather than an active engagement between two people or the posting of information.

Unlike the Twitter content analysis where each tweet was treated as a coding unit, on Facebook each engagement was treated as a coding unit. This is due to each individual Facebook post, the equivalent of a tweet, hosting a range of other engagements within it. On Twitter these would appear as individual tweets which referred to one another via @username, but on Facebook responses are posted within the original content. An example of multiple engagements within one post is given in Figure 4.4. In this post there are five engagements: one message from the user, two comments by readers and two comment Likes by readers.

Figure 4.4: Multiple Facebook engagements

A new category of Reader Response was created due to the ability of readers to ‘Like’ a post or comment, of which there is no equivalent on Twitter. The engagement of Liking something was viewed as a minor form of interactivity like a public gauge but it was not full interactivity as no views or information were given. It could not be classed as sharing, as the engagement did not appear on the participant’s own wall. Due to the large number of comments and Likes it was felt that more valuable results would be yielded if they were defined as two separate categories Reader Response and Reader Interactive. In the second category of Reader Interactive there are also elements of sharing as when a reader posts a comment on another Facebook Profile or Page this appears on their own wall meaning their comments and a link to the original source material is shared to an even wider audience. These two categories however could be collapsed into one category for analysis where deemed appropriate. It was not possible to create a Newspaper Response category as the engagement of Liking a post or comment does not appear on the profile wall. However the engagement of Liking a page does appear on the wall and therefore this was categorised as Newspaper Sharing.

A discussion with the second researcher took place around whether to create a Spam category for people leaving promotional messages on the Bournemouth Echo Sam wall. It was decided to include spam as a Reader Interactive due to the nature of the posts being community orientated. These could include discount offers for local services or information about events (see Figure 4.5). Some were presented informally via comments and others were messages posted on the wall. The newspaper could have removed these posts if they viewed them as spam but chose not to, therefore it was decided they were valid to count.

Figure 4.5: An example of community spam

Figure 4.6: Triangulated methodology design to answer multiple research questions, copyright Lily Canter

Interview x2 = journalists and audience members Interview x1 = journalists

Figure 4.7: Stages of a case study research project, copyright Lily Canter

As explained in this chapter the design of this research project involved fairly lengthy and detailed ethnographic fieldwork at the two case study sites. This resulted in a large amount of rich data collection. The findings drawn from these results are now discussed in the next eight chapters (Chapters 5, 6, 7, 8, 9, 10, 11, 12). These chapters each follow the same structure with an introduction, methods recap, display of results and finally a discussion of the findings with a concluding paragraph. As outlined in Chapter 1, section 1.5.2 the findings chapters follow the sequence of the research questions. The next chapter therefore starts with the results and discussion surrounding RQ1a which examines the changing nature of audience participation.

Chapter 5: The changing nature of participation

5.1 Introduction

For audiences the process of participating in the news has historically been a slow one, with several gatekeepers stationed along the way. Sending a letter in the post meant taking time to put pen to paper, finding an envelope, buying a stamp and walking to the post box. There was also no guarantee that your letter would ever make it into the printed newspaper as it could be spiked by journalists. With the arrival of the Internet, email made the process almost instantaneous once the send button had been clicked, but it still did not bypass the gatekeeping process and could just as easily be spiked as a letter. However Web 2.0 has introduced the potential to change this process and audiences can now publish to the web directly and in many cases can upload comments and even photographs and stories direct to newspaper websites without moderation. With the use of mobile digital technology this can happen in real time, anywhere in the world. As discussed in Chapter 2 some scholars argue that these technological advances have enabled and encouraged an increasingly active audience (Lewis, 2011; Reich, 2011; Jenkins, 2008) and furthermore audiences expect to be able to participate (Banks and Humphreys, 2008; Jenkins, 2008; Deuze, 2006; Rosen, 2006; Bowman and Willis, 2003). This is reflected in the growth of internet usage outlined in Chapter 1 and the popularity of participatory forums online such as comments on news stories, social media networks and user generated content (Heinonen, 2011; Greer and Yan, 2010).

The first research question in this study sets out to examine how the nature of participation has been changing since the introduction of Web 2.0 circa 2005 in the context of the two case studies. It seeks to explore whether participation, as set out in Table 2.2 in Chapter 2, has increased and how it has changed in nature. Drawing on the current literature the researcher expected to find an increase in activity that was also more diverse. It was also anticipated that audiences would expect to be able to participate and would expect to access their news quickly. This chapter therefore aims to explore RQ1a: How does Web 2.0 change the nature of audience participation in British local newspapers?

5.2 Methods

The methods utilised to address this research question, as outlined in Chapter 4, involved an online reader questionnaire, observation within the news rooms of the two case study sites and interviews with journalists and readers.

The online questionnaire at the Leicester Mercury received 177 responses with a 67 per cent completion rate. The response rate at the Bournemouth Daily Echo was 328 with a 72 per cent completion rate. In relation to RQ1a the questionnaire asked five initial questions on background information relating to gender, age, location, income, education and website usage. This was followed by a further six questions about changes and preferences in the respondent’s news consumption. The full questionnaire can be viewed in Appendix 1a.

The news room observation took place at the Leicester Mercury during a three week period in October 2010. Observation at the Bournemouth Daily Echo took place during one week in November 2010 and one week during January 2011. These were arranged at the convenience of the editor and editorial staff hence the fragmented time frames. The second week of observation at both case study sites was intersected with interviews with journalists. A total of 37 professional journalists were interviewed with 19 at the Leicester Mercury and 18 at the Bournemouth Daily Echo. This represented a third of the Leicester Mercury editorial staff and about 35 per cent of the Bournemouth Daily Echo editorial staff. The semi-structured interviews were based on an interview guide displayed in Appendix 2a which included five questions on the changing nature of audience participation, termed as 'interaction’ as discussed in Chapter 2. These questions covered how interaction is changing, whether it had increased under Web 2.0, who was interacting, how many were interacting and how it compared to printed newspaper interaction. These questions were supported by a further checklist which included topics such as expectations, breaking news, user generated content, comments, email and social media.

As explained in Chapter 4 interviewees were given three options regarding their anonymity status. Those interviewees who opted for complete anonymity are referred to in this study by the generic term reporter. Interviewees who opted for their name to be anonymous but not their job title will be referred to by their job title for example feature writer. Those that opted for no anonymity are either referred to by their job title for example deputy editor or by their name and job title where appropriate for example editor, Keith Perch. The participant code of the interviewee whether anonymous or not has been included next to the name/job title of the journalist and at the end of quotes for clarification in particular to indicate which case study site the participant is from. For example L3 or B13 - L corresponds to the Leicester Mercury and B to the Bournemouth Daily Echo. A list of the participants and their corresponding code can be found in Appendix 4. All of the reader interviewees were anonymous so they are referred to in the text by their code only, for example LR3 or BR3 – R standing for reader. As explained in Chapter 4 there was a smaller sample of reader interviewees and they were asked a series of probing questions in relation to their questionnaire responses to add qualitative depth to the data.

5.3 Results

In order to provide a triangulated context to the research the results in this chapter have been drawn from three methods: a reader questionnaire, journalist and reader interviews and news room observation. The results will first explore the responses of journalists together with observations and reader interview responses where relevant, under the following sub headings: rise in participation, how participation is changing, organisational restrictions, one way participation, citizen production. The second section looks in detail at the findings of the online reader questionnaire under the sub headings: audience demographics, audience consumption, audience expectation, The Web 2.0 factor. Both case study sites will be analysed simultaneously.

The statistics discussed in Chapter 1 show an exponential rise in website users at both of the case study sites during the past year, a trend which has not seen a decline since the websites were launched. But an increase in traffic does not automatically translate into an increase in audience participation. Journalists at both case study sites were therefore asked an open question on whether they thought there had been an increase in audience participation post Web 2.0. Based on the interviewees' answers the researcher was able to identify four categories of response: significant increase, slight increase, no increase and unclassified, as shown in Table 5.1. The most popular answer was significant increase (43%), but the second most popular response was no increase at all (22%), revealing that there are still disparate perceptions amongst journalists. David MacLean, politics correspondent at the Leicester Mercury (L10) said participation had increased dramatically even in the past year alone:

I think in five years of reporting there was probably 10 letters in the paper that I would see directly referring to me saying oh I didn’t like what you wrote there, it would be a big occasion if there was a letter in the paper with your name on it, but now you can just stick in a search term on Twitter for the story that you’ve done, see all the responses to it and you also get a feel for it (L10).

However the editor of the Leicester Mercury, Keith Perch (L3), who has worked in the industry for three decades, was slightly more cautious suggesting that the readers who had always been active, were simply being even more active, and new participants were still a minority.

I think there is probably more than there was and the people that are doing it are doing more of it than they used to be in the past. I just don’t think we are yet in a position where you can say everybody wants to do it compared with the old days when nobody wanted to do it (L3).

Yet a reporter at the Leicester Mercury was adamant that nothing had changed in terms of participation levels, it was simply the tools that were different, describing the internet as the modern day equivalent of a “carrier pigeon” (L1).

Table 5.1: Journalist perceptive of participation increase

Company	Significant increase in participation	Slight increase in participation	No increase in participation	Unclassified
Bournemouth Daily Echo	9 (50%)	1 (5%)	3 (17%)	5 (28%)
Leicester Mercury	7 (37%)	4 (21%)	5 (26%)	3 (16%)
Total	16 (43%)	5 (13%)	8 (22%)	8 (22%)