INJECTING IMMEDIACY INTO
MEDIA LOGIC
(RE)INTERPRETING THE MEDIATISATION OF POLITICS ON UK TELEVISION NEWSCASTS
1991-2013
Abstract
This study of UK evening newscasts (1991-2013) interprets the degree to which political news has become mediatised, drawing on the concept of journalistic interventionism to explore edited and live conventions. News examined generally offered little evidence of mediatisation. But when live news was isolated and interpreted over time the study found newscasts were injected with a logic of immediacy, adopting a level of interventionism apparent in instant and rolling news formats. To better understand the mediatisation of politics, future studies could experiment more by theorising different media logics and developing more format specific content indicators that reflect broader influences in
journalism.
STEPHEN CUSHION
Stephen Cushion is Senior Lecturer and Director of MA in Political Communication, Cardiff School of Journalism, Media and Cultural Studies, Cardiff University; e-mail: CushionSA@cardiff.ac.uk.
LO
Cfi m
£± CP
rn o
o rN
rN
S
A 3
a
0
£ *
1 *
(fl 0 c
s
Ch m
Introduction
In many advanced Western democracies, academic debates about the changing nature of political journalism often focus on the role of political actors, considering the extent to which they shape, inform and participate in political coverage. For it has been claimed politicians and political parties have become less significant agents in media coverage, with their voices and views increasingly downgraded or even marginalised in routine reporting. Correspondingly, it is journalists themselves - identified as "media actors" - who have apparently gained greater traction, increasingly thrust into the limelight and expected to not only report the actions of politicians but to make sense of their decisions, behaviour and motivations.
When interpreting the increasing reliance on media rather than political actors, scholars have broadly characterised this trend as the mediatisation of politics, a "process" - Strómback (2008, 241) argues - "through which the important question involving the independence of the media from politics and society concludes with the independence of politics and society from the media." Or, put another way, a mediatisation of politics is displayed when a media logic supersedes political logic in editorial decisions about reporting politics. According to Esser (2013), political logic represents the production of policy by political actors, how these are publicised and how the polity - such as electoral systems - shape the way politics is conducted and news reported. Media logic, by contrast, privileges journalistic norms and routines including professional, commercial and technological factors that influence how politics is reported. The theoretical merits of each have been the subject of recent scholarly posturing, since they can be conceptualised in a multiplicity of ways delivering competing logics rather than a singular, uniform logic (see Landerer 2013).
Informed by these conceptual debates, this study focuses on the empirical ways in which the mediatisation of politics has been operationalised. It will do so by examining UK evening television newscasts - which, despite online competition, remain y the most widely consumed format of news - on the BBC and ITV from 1991-2013 in order to ask whether political news has become mediatised according to well established content indicators. While previous mediatisation of politics studies have explored cross-national differences or between public service and market-driven systems, the UK's broadcast ecology offers a more nuanced comparative inquiry. For it has a wholesale public service broadcaster, the BBC, and commercial public service broadcasters - such as ITV - which are subject to strict regulation to ensure high standards of journalism.
The concept of journalistic interventionism is drawn upon in the study to understand how different television conventions were used to convey the voices
the relative weight granted to both actors, the wider culture of journalism and its impact on television news conventions has arguably not been central to how mediatisation indicators have been operationalised. Over the last twenty to thirty years, there has been a rise in instant, rolling and online news, promoting greater immediacy in the delivery of news and culture of journalism. Against this backdrop, the aim is to examine how far fixed-time evening newscasts have been influenced ^ by 24-hour news culture by examining the extent to which news is edited or live,
and asking whether it is just coverage of politics or news generally that is subject to greater immediacy in routine coverage.
Before explaining the methodological approach of the study, a discussion about how the mediatisation of news and politics has previously been examined is necessary. In doing so, a number of hypotheses will be outlined and interpreted in the context of ongoing debates in journalism studies and political communication.
Interpreting the Mediatisation of Politics: Editing the Views of Politicians and Journalists
To interpret the degree to which a political over a media logic is subscribed to in media coverage of politics, scholars have long examined how and to what extent politicians and political parties routinely are used as sources. This has become known as a soundbite, an onscreen source capturing the continuous view of a politician. In many advanced Western democracies, but notably in the US, studies have shown a decline in the length of soundbites, with most research focussed upon election campaigns. Hallin's (1992, 5) longitudinal study of Presidential elections (1968-1988) found political soundbites had shrunk from 43 to 9 seconds, a finding interpreted as "a general shift in the style of television news toward a more mediated, journalist-centred form of journalism." Other empirical studies have not only confirmed soundbites remain at a similar if not reduced length into the 1990s and 2000s in the US and other countries (Patterson 2000; Farnsworth and Litcher 2007; Esser 2008; Grabe and Bucy 2008), but also that a longitudinal decline in politicians' voices dates back well over a century (Ryfe and Kemmelmeier 2011).
Of course, increasingly sophisticated technology has enhanced the capabilities of producers being able to edit brief soundbites more easily within a packaged news item (Hallin 1988). As technology has improved, a different approach to studying how politicians shape coverage has been pursued, shifting debates from soundbites to imagebites, where a politician appears in an television news package even if he or she does not necessarily speak. So, for example, Grabe and Bucy (2008, 78) examined Presidential election coverage between 1992 and 2004 and discovered that while the average length of candidates' soundbites declined over time, the onscreen appearances of political actors actually increased, reflecting, in their words, an "increasingly visual and journalist-centred news environment." Esser's (2008) cross-national study of the use of soundbites and imagebites in election coverage between 2004-7 led him to conclude that the US had a strongly interventionist way of interpreting politics, a moderately interventionist Anglo-German approach and a noninterventionist French approach. This conclusion was reached based not only on the degree to which politicians' voices and visuals were mediated (by their relative length, for example), but how ostensibly active journalists were in news making (by appearing onscreen and speaking to camera).
Indeed, the visibility of journalists in political news has become a measure to interpret how far politics is mediatised in comparative research. Stromback and Dimitrova (2011) examined US and Swedish election newscasts and identified a far higher level of mediatisation present in American journalism. This was explained by the more commercialised media system in the US - which encourages greater involvement from journalists and editorialising of content - than in Sweden, where coverage is regulated more closely under public service safeguards. Nevertheless,
as Esser (2008) identified in his cross-national study, irrespective of different media systems between countries, there is evidence of a transnational convergence - of shrinking soundbites, enhanced imagebites and greater journalistic autonomy on display - across the US and European countries. Based on a review of previous literature, in this study it is anticipated that the following hypotheses will be confirmed:
H1a: The length of soundbites (onscreen sources) will decrease over time in political news.
H1b: The use of onscreen soundbites and offscreen sources (sources journalists refer to but do not appear onscreen) will decrease over time in political news.
H2a: The length of imagebites (when politicians appear visually) will increase over time in political news.
H2b: The use of imagebites (when politicians appear visually) will increase over time in political news.
H3: The visibility of journalists (onscreen) will increase over time in political news.
Interpreting the Mediatisation of News:
Enhancing the Speed and Liveness of Political Reporting
In Stromback and Dimitrova's (2011) study of US and Swedish election news coverage, they used various measures - of soundbites, journalists talking over politicians, framing politics as a strategic game and the role and visibility of journalists - to measure the degree to which political coverage is mediatised in each country. They concluded, however, by acknowledging that these "are not by any means the only possible indicators, and through further theorising more indicators should be identified, integrated, operationalised, and tested. To do so is an important task for the future research on the mediatisation of politics" (Stromback and Dimitrova 2011, 44). The aim of this study is to do precisely this by considering how the conventions used in television journalism shape the degree to which political news is mediatised. At first glance, interpreting the mediatisation of media might sound a somewhat tautological proposition. But what is being analysed here is how far fixed time evening newscasts are being influenced by the wider culture of news delivery, such as dedicated news channels or instantly accessed rolling news J websites and social media platforms. In other words, the character of journalism has changed, but how has this affected fixed-time newscasts?
Of course, the culture of journalism has always been preoccupied with delivering the latest news with pressures - from print to broadcast media - on time and space. Over the last twenty to thirty years, however, many scholars have argued that the pace of broadcast journalism has been accelerated by the growth of 24hour news channels and, more recently, online news and social media platforms. ^ Put differently, whereas once audiences had to wait for news to be delivered - in a newspaper, say, or an evening newscast - today it is instantly available making it editorially important to bring news "as it happens." Of course, technology that delivers immediacy has improved dramatically in recent decades, making it far easier for broadcasters to "go live" and report on location. As a consequence of the 24-hour news culture, political actors in particular have had to respond to events -as the CNN-effect posited in the 1990s - and the news value of speed has become increasingly central to contemporary political journalism.
0 £
i
C >
However, there is little longitudinal evidence about the degree to which fixed-time evening newscasts - and news about politics specifically - have actually PO been affected by the broader culture of 24-hour news over time. A 1999 study of local US television news identified more live than edited news in newscasts on 24 stations, while a cross-national study in 2012 discovered the US and UK especially had a greater volume of live news than in Norway, where dedicated 24-hour news channels are in their relative infancy (Cushion et al 2014a). But how far have newscasts changed over time remains to be seen. So, for example, while scholars have observed that the pace of journalism has increased (Gitlin 1987; Hallin 1994), there is no systematic evidence about whether this has resulted in shorter news items being dealt with in less depth.
As scholars have examined the changing culture of journalism over time, a shift towards more interpretive political news has been evidenced. Rather than simply describing or relying on external sources, interpretive news means journalists being more actively involved in making political sense of events, issues and debates (Salgado and Strömbäck 2012). Several studies have shown this is most strikingly on display at election time, such as Steele and Barnhurst's (1996) study of US network Presidential coverage between 1968-1998, which identified that factual reporting declined from close to a quarter of all coverage to just 2.4 percent. Conversely, journalists analysing election news in several countries has increased over recent campaigns, challenging the spin of politicians and interpreting their statements and behaviour (Semetko et al 1991; Deacon and Wring 2011). Television journalists, in short, appear to be more central actors in political reporting, notably in live news, with less space for politicians to air their views. Based on a review of previous literature, this study thus expects the following hypotheses to be supported:
H4a: Both political news and non-political news items will have more live reporting over time, but the percentage increase for live political news will be greater.
H4b: Both political news items and non-political news items will become shorter over time, but political news items will become shorter than non-political items.
H5: Fewer sources will appear in live political items than in edited political news items, and this gap will grow over time.
H6: Over time political reporting will become less descriptive and more interpretive in live political news items.
Finally, having reviewed a wide range empirical studies cross-nationally, many of the conclusions reached suggested market forces exacerbated the degree to which political news is mediatised (Strömbäck and Dimitrova 2011). In other words, the type of media logic this study anticipates to find in UK newscasts over time - of shorter but faster news reporting, enhanced live and more interpretive political news, with less external sources shaping coverage - have been broadly understood by scholars as a consequence of commercialisation. Since this study is comparative - comparing commercial and public service media systems - it might thus be expected that the latter resists market influence whereas the former more readily succumbs to the characteristics of mediatisation. The last hypothesis of the study overall predicts that:
H7: The commercial broadcaster will support to a greater degree H1-H6 compared to the public service broadcaster.
Journalistic Interventionism and Operationalising Mediatisation: Method and Sample
To interpret the mediatisation of political news in UK newscasts over recent decades, the concept of journalistic interventionism was drawn upon. It has been used mostly in election research, where the "discretionary power" (Semetko et al 1991, 4) of the media is interpreted by whether news outlets follow the party agendas or journalistically pursue their own agenda. Or, put differently, it conveys how interventionist a journalist is in reporting politics and public affairs. However, this study's analytical approach moves beyond election campaigns and considers how news generally and political news specifically was routinely reported. It does it by treating television news conventions as journalistic interventions because they can be interpreted as conscious editorial decisions about how to report a news story. Based on extensive piloting in other projects (Cushion and Thomas 2013; Cushion et al 2014a, 2014b), four types of journalistic interventions were classified: (1) edited packages, where a reporter films on location and sends back a package to be edited;
(2)	anchors presenting an item, often with a still or moving image in the backdrop;
(3)	a studio discussion, where the anchor and reporter discuss a news story within the studio; (4) and finally, a reporter live on location speaking just to camera or in a live two-way exchange with the anchor. Taken together, the interventions reflect, on the one hand, a closely edited and scripted format (1-2) and, on the other hand, a more live and improvised approach to news, with journalists playing a more active role in reporting (3-4). While the latter reflects the logic inherent in fixed-time newscasts - where, in its original formation, it was designed to be a service that edited and considered the whole day's news - the latter can be seen to embody a logic more consistent with rolling news, delivering live news "as it happens" (as many 24-hour news channels or online blogs claim).
A quantitative content analysis of BBC and ITV early evening newscasts was carried out over three constructed weeks (Monday-Friday) in 1991/2, 1999, 2004 and 2013.1 The BBC is a wholesale public service broadcaster (in its UK operations), funded by a license fee whereas ITV is a commercial public service broadcaster, with a license agreement that legally obliges them to regularly schedule news programming and adhere to strict regulatory guidelines. This comparative dimen-J sion to the study thus asks whether the more commercialised newscast exhibits a greater degree of mediatisation - as the prevailing literature suggested - than the wholesale public broadcaster.
Using the four types of journalistic interventions as the unit of analysis, 1484 items were examined overall. But this N was split into two subsamples of non-political items (N = 1117) and political items (N = 367). The operational definition of "political" items included not only parliamentary news (in Westminster, or other ^ political institutions) but also international news (wars and diplomatic events) when political actors were involved. "Non-political" items were defined by anything other than political items. This allowed the type and changing lengths of political journalistic interventions - the primary focus of this study - to be compared with how news generally is reported. In other words, how interventionist is political reporting compared to all news and has television journalism changed over time? However, the subsample of political items was also analysed in more detail, using ^ previous mediatisation indicators outlined in the literature review. This included
0 £
i
C >
quantifying the volume of onscreen sources, the length of these soundbites,2 as well as the frequency of offscreen sources (e.g. references made by journalists to LH external sources). In addition, the volume and length of imagebites were recorded, measuring if and for how long a politician appeared onscreen in edited news. To explore the role of journalists in more detail, the study first asked whether a reporter was onscreen in a news item or not. It then - in all live news and political news -examined whether the primary purpose of going live was to provide an interpretive or more descriptive approach to reporting. To do so a previously piloted study was drawn upon that explored the value of live reporting (Cushion and Thomas 2013). It coded if a live news item's purpose was to 1) supply interpretation; 2) deliver the latest developments 3) report from a specific location; 4) to introduce an edited package. While there is some overlap in these different approaches, the primary value of live reporting was interpreted.
The content analysis was coded by two UK researchers and according to clearly defined operational definitions that were regularly discussed in team meetings to ensure consistency. Approximately 10 percent of the sample was recoded and Cohen's Kappa (k) - a relatively conservative inter-coder reliability test - was used to evaluate the consistency of coding. Cohen (1960) interprets Kappa co-efficients in the following ways: < 0 indicate less than chance agreement, 0.01-0.20 Slight agreement, 0.21-0.40 Fair agreement, 0.41-0.60 Moderate agreement, 0.61-0.80 Substantial agreement and 0.81-0.99 Almost perfect agreement. All measures used in the study reached substantial or almost perfect agreement between coders.
The study posed three overall research questions:
1.	Has political news become more interventionist than non-political news between 1991 and 2013?
2.	To what extent has political news become mediatised in edited or live television news coverage between 1991 and 2013?
3.	To what extent is political and non-political news interventionist and political news mediatised on the wholesale public service broadcaster compared to the commercial public service broadcaster between 1991 and 2013?
Look Who's Talking: Politicians or Journalists?
Contrary to the prevailing trends in previous academic studies, Table 1 shows that while there were fluctuations over time (notably on ITV where soundbites dipped to 9 seconds in 2004), the average length of soundbites on both channels remained steady and was at its highest peak in 2013 (16 seconds on the BBC compared to 14 seconds on ITV).
Table 1: Mean Average Length (M) and Standard Deviation (SD) of Soundbites and Imagebites Used in Political News on UK Evening Television Newscasts 1991-2103
	1991/2	1999	2004	2013
BBC soundbites	M = 15 secs, SD = 7, N = 141	M = 12 secs, SD = 7, N = 73	M= 16 secs, SD = 10, N = 36	M = 16 secs, SD = 18, N = 115
ITV soundbites	M = 13 secs, SD = 8, N = 76	M = 11 secs, SD = 4, N = 77	M = 9 secs, SD = 6, N = 64	M = 14 secs, SD = 8, N = 108
				
BBC imagebites	M= 15 secs, SD = 9	M = 10 secs, SD = 8	M = 17 secs, SD = 9	M = 16 secs, SD = 9
ITV imagebites	M = 13 secs, SD = 9	M = 13 secs, SD = 7	M = 9 secs, SD = 4	M = 13 secs, SD = 9
H1a was clearly rejected as the length of onscreen sources had not declined over time. While there was lukewarm support for H7 - that onscreen sources decreased more on the commercial broadcaster - since it dipped to 9 seconds in 2004, by 2013 the differences were less striking (2 seconds).
Because the volume of political news items was different each year (see Table 3 below), it is important to interpret the ratio of sources per item. In this respect, Table 2 shows that while the BBC has gradually reduced the volume of sources per item over time, ITV's pattern is more mixed, with a sharp drop recorded in 2004, but a rise in 2013 to the same level as 1991/2.
Table 2: Ratio of Sources (Onscreen) in Political News per Average News Item on
UK Evening Television Newscasts 1991-2013
	1991/2	1999	2004	2013
BBC	2.4	1.8	1.7	1.9
ITV	1.9	2.1	1.2	1.9
Thus, H1b was confirmed to some degree on BBC newscasts, but on ITV a decline was only recorded in 2004. Contrary to H7, it can be concluded that it is the public service broadcaster - not a commercial competitor - that witnessed a decline in sources per item in political news. However, ITV had the lowest level of sources per item (in 2004) and by 2013 had the same ratio as the BBC.
Since there was a marginal decline in the use of sources, it might be expected that the length of imagebites - where political sources visually appear onscreen - could have increased as more airtime would be available. However, according to Table 1, imagebites followed no uniform pattern and in 1991/2 and 2013 they were the same average length on both channels (again, with the biggest dip on ITV in 2004). Where a clearer pattern emerged is in the average use of imagebites in political news items over time. On the BBC in 1991/2 71.2 percent of political items contained an y imagebite, compared to 26.8 percent in 1999, 38.1 percent in 2004 and 35.6 percent in 2013. Precisely half of political news items on ITV in 1991/2 had an imagebite, by contrast, dropping to 32.4 percent in 1999, 32.7 percent in 2004 and rising again to 48.3 percent in 2013. It can be concluded, then, that H2a was not confirmed because the length of imagebites did not increase over time in political news. Further still, H2b was also rejected as the use of imagebites did not increase - in fact its use as a proportion of all political news reduced over time, notably on the BBC when
0
close to three quarters of coverage contained an imagebite. Finally, there was little evidence to support H7, with the use of imagebites similar on both broadcasters.
A measure used to indicate journalists becoming more central in television news reporting is their onscreen visibility. The BBC featured a journalist onscreen in just over half of all political news items (55.9 percent) in 1991/2, dropping to 40 percent in 1999 but increasing to 85 percent and 76.1 percent in 2004 and 2013 respectively. On ITV, by contrast, close to two thirds of political news items in 1991/2 had a visible journalist (65 percent), with their onscreen appearances then increasing steadily (72.7 percent in 1999, 75.5 percent in 2004 and 82.5 percent in 2013). Overall, then, H3 was broadly confirmed in that the visibility of journalists in newscasts had increased over time. However, the BBC's increase was somewhat unidirectional
c >
\D
(with a small reduction from 2004 to 2013). While the visibility of ITV's journalists had uniformly increased over time, the visibility of journalists on both broadcasters post-millennium was remarkably similar, offering little support for H7.
The focus now turns to exploring the changing nature of journalism more generally in order to examine the differences between political and non-political news reporting. As previously explained, television conventions were interpreted as "journalistic interventions," editorial decisions that show the degree to which news is edited and scripted rather than live and more improvised. Or, put differently, examining the type of journalistic interventions pursued over time provided an insight into whether the media logic of newscasts had changed, from editing the day's news to adopting more live news "as it happens."
Live vs. Edited Media Logic: Towards More Journalistic Interventionism?
In political and non-political items edited television news coverage was overwhelmingly dominant on both channels pre-millennium (see Tables 3 and 4). On the BBC, for example, 97.9 percent of political news was made up of reporter packages in 1991/2, slightly higher than non-political items (93.4 percent). By 1999 packaged political news declined by over 10 percent (85.7 percent) compared to a much smaller drop in non-political news (89.9 percent). ITV, by contrast, had a broadly similar level of packaged news in 1991/2 and 1999 in both political and non-political news (89-91.1 percent). The role of anchors, meanwhile, barely featured in the presentation of politics on either channels between 1991-1999 (0.4-1.9 percent).
Coverage changed most strikingly after the new millennium, when the proportion of live news - notably in political reporting - increased substantially on both channels. So, for example, in live BBC reporting political news beyond the studio accounted for 28.3 percent of all coverage in 2004 and 19.6 percent in 2013, well above non-political news (11.7 percent and 9.3 percent respectively). If live political news included a discussion within a studio format - a relatively new BBC format - the proportion of live news in 2013 was not far behind its peak in 2004 (23.8 percent). Meanwhile ITV's live on location reporting of politics also increased substantially in 2004 (27.7 percent), although this dropped (15.1 percent) in 2013. Once again, if live studio discussion was included, the proportion of live news rose to almost 30 percent in 2004 and 19.5 percent in 2013. In both years - most strikingly in 2004 - political news was reported live more than non-political news. It can thus be concluded that while live political news declined between 2004 and 2013, between 1991-2013 H4a was confirmed because live political news not only increased over time, it had to a greater degree than in non-political news. H7, however, was rejected, with the rise of live news similar on both media systems.
But how has the shape and character of these different journalistic interventions over successive decades changed and is political news distinctive from all news? With the exception of ITV in 2004 (by just 3 seconds), political news was routinely longer in length than non-political items. Excluding ITV's bulletin in 1991/2 (as it only lasted 15 minutes), the average length of political news items on both channels between 1999-2013 did not dramatically change (increasing by 10 seconds on ITV and decreasing by 16 seconds on the BBC).
Table 3: Percent of Time Spent on Journalistic Interventions, Average Mean Length (M) and Standard Deviation (SD) in BBC Television Newscasts in Political and Non-political News Items 1991-2013
	1991/92	1999	2004	2013
Pre edited				
Reporter package (politics)	97.9 %, M = 2 minutes and 29 seconds, SD = 49 seconds	85.7 %, M = 2 minutes and 7 seconds, SD = 32 seconds	71.2 %, M = 2 minutes and 48 seconds, SD = 1 minute and 2 seconds	74.2 %, M = 2 mins 54 secs, SD = 1 minute and 10 seconds
Reporter package (non- politics)	93.4 %, M = 2 minutes and 9 seconds, SD = 23 seconds	89.9 %, M = 1 minute and 57 seconds, SD = 35 seconds	80.5 %, M = 2 minutes and 15 seconds, SD = 38 seconds	83.2 %, M = 2 mins 28 secs, SD = 34
Combined anchor only, image and package (politics)	0.6 %, M = 25 seconds, SD = 7 seconds	1.9 %, M = 16 seconds, SD = 4 seconds	0.5 %, 12 seconds	1.4 %, M = 26 secs, 15 SD
Combined anchor only, image and package (non-politics)	4 %, M = 27 seconds, SD = 23 seconds	3.9 %, M = 23 seconds, SD = 14 seconds	7.8 %, M = 25 seconds, SD = 27 seconds	5.8 %, M = 22 seconds, SD = 7
Live news				
Combined Reporter/ anchor 2-way and reporter live (politics)	1.5 %, M = 2 minutes and 7 seconds	12.4 %, M = 1m 16 seconds, SD = 31 seconds	28.3 %, M = 1 minute and 22 seconds, SD = 22 seconds	19.6 %, M = 1 minute and 13 seconds, SD = 49
Combined Reporter/ anchor 2-way and reporter live (non-politics)	2.6 %, M = 1 minute and 40 seconds, SD = 43 seconds	5.8 %, M = 1 minute and 14 seconds, SD = 26 seconds	11.7 %, M = 49 seconds, SD = 28 seconds	9.3 %, M = 47 seconds, SD = 27
Anchor reporter discussion/ studio discussion (politics)	/	/	/	4.2 %, M = 1 minute and 45 seconds, SD = 17
Anchor reporter discussion/ studio discussion (non-politics)	/	0.4 %, 1 minute and 4 seconds	/	1.6 %, M = 61 seconds, SD = 32
Total politics N	59	41	21	59
Total non-politics N	140	182	212	118
I *
(I) 0 c
I 00
But while, on the face of it, the structure of political news may appeared to have altered little over time, it is important to interpret the type of edited and live journalistic interventions used to routinely report politics in television newscasts. Or, put differently, what specific media logic drives political coverage over time? Tables 3 and 4 showed that the type of intervention used to report a news story - whether an anchor only item, an edited package or a live two-way - brought considerable differences in terms of its length. On both broadcasters, for example, the role of anchors in story-telling was consistently short (12-32 seconds long). Moreover, political items presented by anchors appeared much less frequently than non-political news items and tended to be shorter. Edited packages, however, played
Table 4: Percent of Time Spent on Journalistic Interventions, Average Mean Length (M) and Standard Deviation (SD) in ITV Television Newscasts in Political q\ and Non-political News Items 1991-2013
	1991/92	1999	2004	2013
Pre edited				
Reporter package (politics)	89 %, M = 1 minute and 48 seconds, SD = 36 seconds	90.1 %, M = 2 minutes and 9 seconds, SD = 31 seconds	69.8 %, M = 2 minute and 30 seconds, SD = 41 seconds	80.6 %, M = 2 minute and 32 secs, SD = 40
Reporter package (non- politics)	89.5 %, M = 1 minute and 39 seconds, SD = 32 seconds	89.7 %, M = 1 minute and 57 seconds, SD = 40 seconds	75 %, M = 2 minute and 8 seconds, SD = 44 seconds	75.8 %, M = 2 minutes and 17 seconds, SD = 51
Combined anchor only, image and package (politics)	/	0.4 %, M = 19 seconds	0.3 %, M = 14 seconds	/
Combined anchor only, image and package (non-politics)	1.9 %, M = 23 seconds, SD = 16 seconds	5.3 %, M = 24 seconds, SD = 22 seconds	0.7 %, M = 32 seconds, SD = 9 seconds	1.5 %, M = 19 secs, SD = 6
Live news				
Combined Reporter/ anchor 2-way and reporter live (politics)	11 %, M = 1 minute and 3 seconds, SD = 14 seconds	8.9 %, M = 1 minute and 3 seconds, SD = 15 seconds	27.7 %, M = 47 seconds, SD = 25 seconds	15.1 %, M = 1 minute and 9 seconds, SD = 22
Combined Reporter/ anchor 2-way and reporter live (non-politics)	8.6 %, M = 1 minute and 7 seconds, SD = 20 seconds	4.2 %, M = 1 minute and 11 seconds, SD = 23 seconds	14.4 %, M = 43 seconds, SD = 22 seconds	10 %, M = 54 secs, SD = 19
Anchor reporter discussion/ studio discussion (politics)	/	/	2.2 %, M = 1 minute and 43 seconds	4.4 %, M = 1 minute and 47 seconds, SD = 8
Anchor reporter discussion/ studio discussion (non-politics)	/	0.8 %, 1 minute and 35 seconds	9.9 %, M = 1 minute and 19 seconds, SD = 49 seconds	6.9 %, M = 1 minute and 28 seconds, SD = 26
Total politics N	40	37	52	58
Total non-politics N	96	125	153	91
a more central role. But the average political news and non-political edited package news item on the BBC has remained remarkably consistent between 1991-2013 (approximately 2.07-2.54 minutes long (although the SD post-millennium is higher) and 1.57-2.28 minutes long respectively. Excluding ITV's 1991/2 edited packages coverage was broadly similar from 1999-2013 (for political news items 2.09-2.32 minutes long and non-political news items 1.57-2.17). Most striking on both channels was that edited political news was consistently longer than non-political news.
Since the volume of live political news increased over time, it is of course difficult to interpret how the mean length of interventions such as two-ways have changed. So, for example, the one BBC live political item in 1991/2 lasted 2 minutes
and 7 seconds and since then it averaged similar lengths (from 1.13-1.22 minutes long). Non-political BBC news, however, was shorter in length (from 47 seconds to 1.40 minutes). ITV's average live political news dipped considerably to just 47 seconds in 2004 compared to over a minute (1.03-1.09) every other year. Its live non-political news, however, from 1999-2013 was comparatively shorter in length (43 seconds to 1.11 minutes). With the exception of ITV in 1999, live political news was consistently longer than non-political news on both channels.
Overall, then, H4b can be rejected because it was not political items that became shorter in length over time, but non-political news (in 2013 to less than a minute in live two-ways on both broadcasters). At the same time, however, it is important to interpret the changing nature of political news in the context of coverage overall. For while the data suggested little had changed in terms of the average length of edited or live political news, Tables 3 and 4 also showed that news generally - and political news particularly - was increasingly going live. In other words, political news has become shorter over time simply because more live journalistic interventions are being used to report routine politics on the evening bulletins.
To make sense of the editorial consequences of selecting live over edited journalistic interventions in political news reporting, the analysis now examines the use of sources and role of reporters in political news alone. For the sources - both onscreen and offscreen - used to inform live reporting as opposed to edited news have not been compared. Table 5 indicates the ratio of sources per items in the two dominant journalistic interventions - edited news packages and live reporter/two-ways - between 1991-2013.
Table 5: Ratio of Onscreen and Offscreen Sources to Edited Packages and Two-
way/reporter Live on BBC and ITV Evening Television Newscasts 1991-2013
£ 3
a
0
£ *
1 *
(I) 0 c
! o Ln
	BBC		ITV	
	Edited package	Two-way/ reporter live	Edited package	Two-way/ reporter live
1991/2	4.1	2	3.7	1.4
1999	4.1	0.9	5.2	1
2004	4.3	0.3	5.3	1
2013	5.6	1.6	3.5	1.6
While there was a little variation in the use of sources on both broadcasters over recent decades, there was consistently close to 4 sources or more (3.5-5.6) in reporter packages. Two-ways/reporter live interventions, by contrast, had considerably less sources on average, typically no more than 1 per item and considerably less on the BBC in 2004. H5 is thus clearly confirmed with less sources drawn upon in live rather than edited political news. However, H7 was not supported as the public and commercial media systems follow a similar sourcing pattern between 1991-2013.
Since live political news was not reliant on external sources in evening bulletins, the final part of the study examined the purpose - according to criteria explained in the method section - of reporting live during a newscast. Of the few live political items in 1991/2, both broadcasters used the live two-way to exclusively interpret politics. In 1999 the BBC interpreted live political news to the same degree as intro-
ducing an edited packaged (42.9 percent), whereas on ITV 60 percent of political news was interpretive. In 2004 the BBC enhanced its interpretation to almost all of its live items (88.9 percent) while ITV used two-ways to introduce political items to a greater extent (55.2 percent) than interpreting politics (34.5 percent). 2013, by contrast, saw interpretation as the primary purpose on both channels (52.2 percent on the BBC and 63.2 percent on ITV) with updates occupying a greater part in live coverage (21-22 percent on both channels). Overall, H6 was largely supported (with the exception of ITV in 2004) because live news was primarily used to interpret political action, rather than used for the location of the reporter, for introducing edited packages or supplying live updates. Although, in 2013, the primary purpose of over a fifth of live news reporting related to bringing the latest developments to a story, far greater than any previous year.
Injecting Immediacy into Media Logic: (Re)interpreting the Mediatisation of Politics
Overall, the initial headline findings of this study - that, over recent decades, the length of onscreen sources had not declined in television newscasts (rejecting H1a), that there had been a subtle decline in the use of sources informing coverage (mildly supporting H1b), that imagebites had not increased in length or use (rejecting H2a and H2b) or that the visibility of journalists had steadily grown (supporting H3) - provided, according to previous theorising, only lukewarm evidence to support the proposition that a mediatisation of political content had occurred (Stromback and Dimitrova 2011).
However, if one scratches below the surface of these indicators - of sourcing, imagebites and journalistic visibility - it can be observed that the changing trends in television news coverage can be principally explained - in the second part of the study - by whether political news is edited or live in format. As H4a confirmed, live news steadily increased into the news millennium, notably in the reporting of politics (supporting H4b). Since live two-ways became a more widely used convention in newscasts, the visibility of journalists onscreen also substantially shifted. When live news was excluded, for example, the proportion of visible journalists did not change as dramatically into the new millennium (featuring in approximately two thirds of coverage on both channels, 10-20 percent less in 2004 and 2013). Correspondingly, as onscreen sources were almost always featured in edited news as imagebites exclusively were, the findings suggested that while edited political news had largely remained the same, live news had a greater influence on coverage.
However, while live news items became shorter over time, live political news did not - with the exception of ITV in 2004 - rejecting H4b. But as was pointed out, political news has become shorter over time because more of it is reported live, which is shorter in length than edited news. Another key difference established by the type of journalistic interventions shaping coverage of politics was - as H5 predicted - that far less sources were used in live rather than edited news. Moreover, it was journalists themselves who acted as sources more in live news, delivering, above all, interpretation (apart from ITV in 2004) rather than a more descriptive style of reporting (supporting H6).
While the prevailing literature suggested commercialisation was a likely cause of greater mediatisation, this study found little evidence to support any major
differences over time in the nature of political reporting or the type of journalistic interventions between the public service broadcaster and the more commercially driven channel (rejecting H7). As ITV is a commercial public service broadcaster - closely regulated and in competition with the BBC, a wholesale public service broadcaster - it could be that the channel does not fall readily (by cross-national standards) into a "commercial" channel. In other words, unlike the US's laissez faire broadcast model, the UK's regulatory culture and overarching public service framework may - compared to other countries - put a buffer on commercial influences. Commercialisation as a precursor to mediatisation is further questioned below, but for now the wider significance of the study is considered in the context of debates about the mediatisation of politics.
When political news was analysed generally, there was little support for well-established indicators of mediatisation such as soundbites and imagebites. However, when live news was isolated and analysed over time, the degree of journalistic interventionism generally appeared to increase, with less political voices present, greater journalistic visibility and reliance on interpretation from reporters. This suggests that in order to understand where and why political news has become mediatised, it is important to more broadly interpret the changing nature and culture of journalism. For live news has been enhanced in newscasts generally, indicating that the media logic shaping its coverage has been informed by conventions and values familiar to rolling news rather than fixed-time programming. Or, put another way, the level of journalistic interventionism - what Stromback and Esser (2009, 219) called the "engine of mediatisation" - has increased on newscasts in recent decades due to an injection of immediacy in the values of television journalism more generally.
In light of these findings, it could be argued that when theorising new indicators of mediatisation the specific format of media - whether fixed-time newscasts, online blogs or rolling news coverage - should be more carefully considered to interpret how its logic has changed over time. In other words, rather than there being a single and uniform media logic shaping the mediatisation of politics, there y are competing and multiple logics at work. Needless to say, it is not just media logics that change over time, but wider social, political and economic forces. But while acknowledging the whiff of tautology at play, interpreting the mediatisation of news media should arguably play a greater role in understanding the changing nature of political reporting.
A common explanation for why political news has become more mediatised is due to journalists responding to a new professionalised class of "on-message" politicians. The rise of the interventionist live two-way convention in this study arguably reinforces this proposition. For live political news increased to a greater extent than non-political news, indicating that editorially speaking politics is a more interventionist genre of news. However, the study also found live non-political news was shorter than live political news, suggesting that political reporters were
c >
granted greater time to react and interpret the world of politics.
Indeed, political journalists were routinely asked to analyse rather than simply describe political news. So, for example, after the Chancellor of the Exchequer's
2004 annual Budget the anchor asked the then political editor, Andrew Marr, to interpret the politics behind it: "Now, Andrew, do you think this budget sharpens y^ the dividing line between Labour and the Conservatives?" In other words, jour-
nalists appear to be increasingly invited to act as a key source, rather than relying on politicians or political parties. It is a point similarly identified by Lundell and PO Ekstrom (2013) in a qualitative assessment of Swedish television news between 1982 ^ and 2012. They concluded that "journalism has gradually chiseled out a position for itself where journalists positioned as interviewees are enabled to act in the role of authoritative and confident news sources" (Lundell and Ekstrom 2013, 528).
While more sophisticated technology, of course, allows journalists to report beyond the confines of the studio, this study found the location of a reporter did not appear to be a significant reason for a live two-way. By 2013, for example - when the ability to go live was no longer a novelty having been an established convention for well over a decade - journalists were being increasingly asked - notably on the BBC - to comment and interpret upon stories within the studio. This can thus be seen to reinforce the importance of journalists as sources, rather than roving reporters out on location. Of course, reporters talking onscreen in well-established locations - outside Downing Street or Westminster - can be easier to produce, cheaper, and less resource intensive than gathering and editing a news item. Live news can also be shared between broadcasters (from news agencies, say) leaving it to competing journalistic personalities to offer a unique perspective/interpretation. But since live political news was most apparent on the public service broadcaster, this more interventionist approach to television news journalism does not appear to be consistent with a commercial strategy, but arguably one that is deliberately deployed to inform viewers, to enlighten rather than simply entertain. In other words, a greater mediatisation of political news does not necessarily mean adopting a market logic, a reflection of commercial decision making. At the same time, the study found live political reporting contained far less sources per item than edited news, placing media actors - or, more specifically, political editors and correspondents - at the centre of the narrative. This raises important questions that go beyond the scope of this study. For how politics is routinely communicated by television journalists - exploring more qualitatively, the type of analysis and contextualisa-tion in coverage - and understood, engaged with and interpreted by audiences is needed to be able to assess the impact of this type of interventionism in newscasts.
In closing, then, the evidence in this study has suggested that newscasts have been injected by a logic of immediacy over recent decades, adopting - it was theorised and empirically confirmed - the kind of urgency and interventionism apparent in instant and rolling news formats. How far future newscasts will be distinctive from, or more complicit with, the editorial direction of rolling news, of course, remains an open question. An emergent feature of 2013, for instance, was the growth of latest updates in live reporting, further evidence perhaps of yet more immediacy in fixed-time newscasts. This arguably makes it necessary to routinely (re)interpret the mediatisation of politics theoretically and empirically, and importantly according to the different formats, media systems and political cultures that shape competing logics and levels of interventionism.
Notes:
1. The years and weeks of the longitudinal study were largely shaped by the availability of TV news footage in the 1990s and early 2000s. The sample of news in 1991/2 sample, for example, was not over one year because a full three week could not be located. However, in order to ensure coverage was not skewed, no major newsworthy stories dominated the sampling period. I am
grateful to Greg Philo and Colin Macpherson for helping me to access the television news footage from the Glasgow Media Group's archive. I would also like to thank Cardiff's Undergraduate Research Opportunities Programme (CUROP) for funding Rachel Lewis and Hugh Roger over eight weeks to help code the TV news footage.
2. Rather than only measuring the volume and length of politicians' soundbites, the study included all onscreen sources (citizens, business leaders, police etc.), since they can also be seen as important actors in the reporting of politics.
References:
Cushion, Stephen, and Thomas, Richard. 2013. The Mediatization of Politics: Interpreting the Value of Live vs. Edited Journalistic Interventions in UK Television News Bulletins. The International Journal of Press/Politics 18, 3, 360-380. Cushion, Stephen, Toril Aalberg and Richard Thomas. 2014a. Towards a Rolling News Logic in Fixed Time Bulletins? A Comparative Analysis of Journalistic Interventions in the US, UK and Norway. European Journal of Communication 29, 1, 100-109. Cushion, Stephen, Hugh Rodger and Rachel Lewis. 2014b. Comparing Levels of Mediatization in Television Journalism: An Analysis of Political Reporting on US and UK Evening News Bulletins. International Communication Gazette 76, 6, 443-463. Deacon, David, and Dominic Wring. 2011. Reporting the 2010 General Election: Old Media, New Media - Old Politics, New Politics. In D. Wring, R. Mortimore and S. Atkinson (eds.), Political Communication in Britain, 281 -303. Basingstoke: Palgrave McMillan. Esser, Frank. 2008. Dimensions of Political News Cultures: Sound Bite and Image Bite News in France, Germany, Great Britain and the United States. International Journal of Press/Politics 13, 4, 401-428.
Esser, Frank. 2013. Mediatization as a Challenge. In H. Kriesi, S. Lavenex, F. Esser, M. Jorg, M. Buhlmann and D. Bochsler (eds.), Democracy in the Age of Globalization and Mediatization, 155-176. Basingstoke: Palgrave. Farnsworth, Stephen, and Robert Lichter. 2007. The Nightly News Nightmare: Network Television's
Coverage of US Presidential Elections, 1988-2004. Lanham, MD: Rowman and Littlefield. Gitlin, Todd. 1987. Watching Television. New York: Pantheon.
Grabe, Maria Elisabeth, and Eric Bucy. 2007. Image Bite Politics: News and the Visual Framing of
Elections. New York: Oxford University Press. Hallin, Daniel. 1992. Sound Bite News: Television Coverage of Elections, 1968-1988, Journal of
Communication 42, 2, 5-24. Hallin, Daniel. 1994. We Keep American On Top of the World. New York: Routledge. Landerer, Nino. 2013. Rethinking the Logics: A Conceptual Framework for the Mediatization of
Politics. Communication Theory 23, 3, 329-258. Lundell, Asa Kroon, and Mats Ekstrom. 2013. Interpreting the News: Swedish Correspondents as
Expert Sources 1982-2012. Journalism Practice 7, 4, 517-532. Patterson, Thomas. 1993. Out of Order. New York: Vintage.
ORyfe, David. M., and Markus Kemmelmeier. 2011. Quoting Practices, Path Dependency and the
Birth of Modern Democracy. Journalism Studies 12, 1, 10-26. Salgado, Susan, and Jesper Stromback. 2012. Interpretive Journalism: Review of Concepts,
Operationalizations and Key Findings. Journalism 13, 2, 144-161. Semetko, Holly, Jay Blumler, Michael Gurevitch and David Weaver. 1991. The Formation of Campaign Agendas: A Comparative Analysis of Party and Media Roles in Recent American and British Elections. Hillsdale, NJ: Erlbaum. Steele, Catherine, and Kevin Barnhurst. 1996. The Journalism of Opinion: Network Coverage in US
C	Presidential Campaigns, 1968-1988. Critical Studies in Mass Communication 13, 3, 187-209.
Stromback, Jesper. 2013. The Four Phases of Mediatization: An Analysis of the Mediatization of
Politics. The International Journal of Press/Politics 13, 3, 228-246. Stromback, Jesper, and Daniela Dimitrova. 2011. Mediatization and Media Interventionism: A Comparative Analysis of Sweden and the United States. The International Journal of Press/ ^	Politics 16, 1, 30-49.
Ln