When, in 2013, the Danish media organisation Politikens Hus introduced a payment wall on its online site, it did so against a backdrop of the increased success of the online newspaper Pol.dk over a 15-year period, and a decline in readership, and thus advertising revenue, of the printed paper. It seems that slowly but steadily the power balance between old and new media has shifted, but without a guarantee that the new platforms can compensate for losses elsewhere in the media organisation. Professional journalism currently finds itself in a fluid position, meaning the need to study this development historically and structurally remains as strong as ever. But how can we analyse such a transition and structural changes over time? And how can we capture the relational differences between different media organisations and platforms, and changes over time, within our analytical questions? In this article we argue that operationalising Pierre Bourdieu’s Field Theory, and utilising a Principal Component Analysis (PCA) to construct a Danish “field of news”, are valuable tools in providing at least some answers. Due to the article’s explorative nature, we are not in a position to provide empirical conclusions on structural changes in Danish news production; to do so would require many more indicators and a lengthier study. However, we can identify minor changes and, more importantly, the mapping method allows us to explore the
The advantages of analysing data relationally and displaying it in map form are in sociology best known from the correspondence analysis of Pierre Bourdieu (Bourdieu 1984, 1986 and 1988). In media studies, this type of data analysis has only been developed into more advanced theories of, for example, “The Journalistic Field” (see Benson and Neveu 2005) in the last ten years. However, researchers adopting these theories have continued to focus on data relating to individual journalists, such as that gleaned from survey questionnaires (see Hovden 2008; Wiik 2010), or on the long-term development of specific types of content, for example financial stories (Slaata 2003) or a rise in health journalism (Marchetti 1997, 2005). In another branch, scholars have worked with the more anthropological studies of Bourdieu, often looking at
When ethnographic observations reveal that online journalists feel inferior to journalists working on offline platforms, but that they also feel an increasing respect for the online platform over time (see Hartley 2011), a PCA provides a way of putting such case-specific data in perspective. What has happened in the production of news that has led to this change in perception? How have the relations between different “forms of capital” changed during this period, and what does this mean for the positions of the different online newspapers and other platforms in the media organisation? This article explores changes over a relatively short period (2008–2010), and in future research the model could be expanded both in terms of the number of indicators and the historical period studied.
Following a literature review, the first part of the paper briefly maps the theoretical and methodological frameworks. The second part of the paper provides the results of the content analysis and presents the explorative maps of “the Danish news space” constructed on the basis of these results using a PCA. The content analysis consists of a coding of news production from print, online and television platforms of six major Danish media organisations. This is elaborated in the methodology section of the paper. The final part of the paper offers up conclusions and discusses ideas for further development of this analytical model as a useful tool in online journalism research, and journalism research in general.
According to David Domingo (2006), online journalism research has had, since the early days of research into this sub-field of journalism, a particular methodological focus on content analysis of the features of online newspapers. Less often, researchers have compared online and offline content, but the few studies carried out within a European context showed that much of the online production was shovel-ware from the printed newspaper (see Neuberger et al. 1997; Van der Wulf & Lauf 2005). Much of the online production was also distributed from wire agencies, and the sources were much more diverse in the printed version, as Barnhurst (2002) showed in the US and Gasher and Gabriele highlighted in Canada (2004). Lim studied four American online newspapers and revealed how similar their content was and how they seemed to prioritise different news stories in the same way (Lim 2010). Furthermore, studies from the US have considered the differences between local and national online and offline newspapers, concluding that the online newspapers had a much more local outlook than their printed counterparts (see Singer 2001; Gasher & Gabriele 2004).
In the Scandinavian context, researchers have often been interested in measuring the level of hypertextuality, interactivity and multimedia on websites from a historical perspective (Eriksen & Ihlström 2000; Engebretsen 2006). Based on a study of online news production in Spain, Domingo points to the fact that “digital utopias” often dominate researchers’ strategies, as they assume the digital possibilities inherent in the structure of the Internet will also affect journalism in a positive manner. Thus, the historical perspective consists of a measurement of how far the online newspapers have moved towards the end goal of a truly multimedia, interactive, hypertextual and immediate coverage of news – a true digital utopia. However, many studies conclude that online journalism remains far-removed from this digital utopia, and they have been criticised for failing to question the inherent assumptions, or why these ideas remain utopian (Domingo 2006; Hartley 2011). This can partly be explained by the theoretical perspective of many of the more descriptive studies of the 1990s, as they were often chiefly concerned with questions of technological innovation. It can be argued that sociological analytical perspectives have been overlooked in the analysis of media content in general, as the focus has been on the specific news article and the website affordances rather than online news production as a process.
In Denmark, very few studies have looked into the online media as a news provider, despite its increased popularity over the last 15 years. One study compared the two public service channels and their online content (Bang 2008), and showed that a limited amount of material was parallel-published online, which was seen as a negative consequence of the lack of collaboration between online and television staffers (ibid). Another study considered the development of the online newspaper within the structural framework of media history (Falkenberg 2009), while a third compared content across online newspapers, but did not compare this to the content production of other media platforms (Hartley 2011). The latest study of online news is following in the footsteps of the website affordance tradition, using content analysis of specific case studies to investigate the use of interactivity, hypertextuality and multimediality on Danish online news sites (Kammer 2013).
This article aims to fill this gap by exploring an analytical model where the different online and offline news sites, and their relational differences, can be analysed using a number of indicators that could be expanded almost endlessly. The analytical model provides us with ways of comparing online newspapers across the “Field of News”, and the means to establish the degree of relatedness of online and offline media platforms belonging to different media organisations. Thus, the aim of this paper is to contribute to the methodological and analytical developments in the field of online journalism research. The theoretical contribution consists of an operationalisation of some key concepts in Pierre Bourdieu’s Field Theory, and applying these to a specific study of Danish news production.
The theory behind the development of the analytical model presented in this paper is Pierre Bourdieu’s Field Theory. The field is seen as a hierarchical social field, where each position is defined by its relation to other positions (Schultz 2006, 2007). Following Bourdieu, we can take an empirical approach to the relatively open concepts of field, capital and habitus:
“(…) Concepts have no definition other than systematic ones, and are designed to be put to work empirically in systematic fashion.”
The concept of field is first of all a corrective against positivism. Fields are conceptual constructions based on the relation mode of reasoning. “To think of field is to think relationally,” Bourdieu stressed (Bourdieu & Wacquant 1992: 96). They illustrated Bourdieu’s relational logic by encouraging the researcher to seek out underlying and invisible relations that shape action, rather than properties given in common-sense categories. Bourdieu wanted to draw attention to the latent patterns of interest and struggle that shape the existence of these empirical realities. Positivist conceptions of social location, such as “milieu”, “context” or even “social background”, fail to highlight the conflictual character of social action. However, we need to operationalise the concept of field into a number of indicators to study what is at stake within a specific field, i.e. the specific conflictual oppositions of, for example, the journalistic field and how different individuals or organisations are positioned within it. Bourdieu introduced the concept of “geometric data analysis” in his studies, the idea of using many indicators to construct a multidimensional social space, where the planes marked the most important structural oppositions – or forms of capital – within this space (Lebaron 2009).
Employing Bourdieu’s ideas, Dominque Marchetti (2005) suggested a number of indicators that can be used to map the relational differences within a field in an analysis of the different forms of capital in, for example, the journalistic field, and further noted that it might be helpful to look at specialised forms of journalism as sub-fields. In our analysis, we therefore look at “online journalism” as a specialised sub-field. Marchetti argued that the sub-field’s capital can be established using various indicators, such as the number of journalists dedicated to it, the number of exclusive news stories covered, and economic factors, such as the amount of readers and subscribers and the income derived from advertising (Marchetti 2005: 79). From a field perspective, we can view the number of exclusive news stories as an indicator of the specific capital of the journalistic field –
In this way, we operationalise Bourdieu’s concept of
An Overview of the Different Variables Used in the PCA Analysis in 2008 and 2010
2008 | 2010 | |
The number of users/viewers/readers | The number of users/viewers/readers | |
+ the original media platform (newspaper, television news bulletins) | ||
The number of journalists | The number of journalists | |
The amount of exclusive news stories/own production | The amount of exclusive news stories/own production | |
External quotes (shovel-ware) | External quotes (shovel-ware) | |
Internal quotes | Internal quotes | |
Amount of stories from news agencies | Amount of stories from news agencies | |
Subject areas | Subject areas | |
+ The original media platform (newspaper, television news bulletins) |
Analytically, a PCA enables us to use the selected indicators to construct a space and thus operationalise the concept of
The content of six major Danish online newspapers were coded for one week in November 2008, providing a total of 2343 news items, and again for one week in 2010, when a total of 5200 news items were coded. In 2010, we also coded the content from the corresponding newspapers and TV bulletins within the media organisation. The coding was carried out using SPSS, and the six online media outlets chosen all belonged to major and established Danish media organisations:
The content analysis shows that some sites are more dependent on agency production than others. It also reveals that the media organisations have different strategies in terms of how much they distribute from online to print and vice versa. From a field perspective, this can be interpreted as some online media being more autonomous than others. In 2008, the percentage of news agency shovel-ware was between 29 per cent and 62 per cent on all the sites. By 2010, the main news agency in Denmark, Ritzau, appears to have increased its position in the field, as the online sites have decreased their level of shovel-ware from the other platforms (print and television), and during this period have increased their dependence on news agencies (the percentage of news agency articles in 2010 ranged from 36 per cent to 66 per cent).
In addition, from a field perspective we can take the “quoting” of a media organisation as a positive sanction by its competitors. When a story is cited, it is recognised that the cited media had the story first, and that it is a “good” story worth citing (Bourdieu 1998 [1996]: 94). The same mechanisms are seen in Shultz’s news ethnography of Danish journalists (2006) and Hartley’s study of Danish online journalists (2011). Thus, if we consider how the different online newspapers are citing others, and to what extent they publish stories from competing media, we can map their relational positions with regard to this indicator It was only possible to code the quoting where the online journalists were honest about where they got stories. The analysis thus tends to favour the less honest sites that hide the origins of stories. However, focusing on one week and considering the overlap of news agendas across the media landscape made it easier to detect non-original stories on that specific site, even when they may have initially appeared to be original.
The sites are publishing content based on quoting stories from competing online sites at levels ranging from 7 per cent to 40 per cent. Therefore, there is a vast difference between the online newspapers with regard to their original journalistic production of news. However, in 2010, most of the online newspapers have increased the amount of original production and decreased the use of news stories from competing media organisations.
Above we referred to a number of studies that have researched what might be labelled “internal distribution” – in other words, the amount of news stories directly distributed, without changes, from the offline media to the online media. In comparison with these studies, Denmark is no different. The percentage of internal distribution of news from offline to online, in other words the number of stories published directly from the “parent” platform to the online platform, varies between 47 per cent (in 2008) at the online newspaper
The Three Principal Component Analysis of the Field of News
Component 1 | Component 2 | Component 3 | Component 4 | |
---|---|---|---|---|
Eigenvalue | 2.88 | 1.74 | 0.93 | 0.43 |
Explained variance (%) | 48.0 | 29.1 | 15.5 | 7.2 |
Coordinates of active variables | ||||
Original production | −0.524 | 0.769 | −0.321 | −0.172 |
Internal shovel-ware | −0.576 | −0.645 | 0.498 | 0.051 |
External shovel-ware | 0.878 | 0.334 | 0.310 | −0.125 |
News agency articles | 0.575 | −0.467 | −0.606 | 0.290 |
Users 2008 | 0.770 | 0.452 | 0.327 | 0.301 |
Journalists 2008 | −0.760 | 0.449 | 0.076 | 0.460 |
Eigenvalue | 2.98 | 2.27 | 0.72 | 0.03 |
Explained variance (%) | 49.7 | 37.8 | 12.0 | 0.4 |
Original production | 0.869 | 0.414 | 0.262 | −0.067 |
Internal shovel-ware | −0.025 | 0.985 | 0.131 | 0.110 |
External shovel-ware | 0.514 | −0.777 | −0.357 | 0.066 |
News agency articles | −0.997 | −0.024 | 0.054 | −0.046 |
Users 2010 | 0.447 | −0.624 | 0.640 | 0.028 |
Journalists 2010 | 0.876 | 0.365 | −0.312 | −0.036 |
Eigenvalue | 3.85 | 1.26 | 0.57 | 0.34 |
Explained variance (%) | 61.6 | 21.2 | 9.6 | 5.6 |
Coordinates of active variables | ||||
Original production | −0.967 | −0.003 | 0.017 | 0.245 |
Internal shovel-ware | 0.532 | −0.606 | 0.591 | 0.016 |
External shovel-ware | 0.414 | 0.834 | 0.289 | −0.220 |
News agency articles | 0.851 | −0.326 | −0.355 | −0.188 |
Users 2010 | 0.834 | 0.318 | 0.000 | 0.423 |
Journalists 2010 | −0.941 | 0.015 | 0.122 | −0.134 |
In conducting the PCA we used these indicators to create a map of where the media organisations and platforms are situated in relation to each other. The analysis has the purpose of graphically demonstrating the improved position of online newspapers, and providing important insights into the disparities between the different media organisations and the strategies of journalistic production for their online platforms. However, as we only included six cases, the PCA should not be regarded as a statistical analysis-testing hypothesis due to the sensitivity to variations in the small sample. Rather, it should be taken as a description-oriented method of reducing dimensionality in the description offered by the content analysis by both summarising the set of variables and fitting the cases into a graphical illustration of components (Le Roux and Rouanet 2004: 133) for identifying central oppositions in the
Figure 1 shows which indicators structure the relations between the different online media in the Danish field of news. These are marked in solid black type, and the indicators that contribute most to the construction of space will generally have the longest lines. Indicators of prioritising different content, which helps our interpretation of the space, are marked with dashed lines. It is a constructed field and shows the
Towards the bottom left hand corner of the Figure 1 “map” we find the amount of users and quote stories, which indicates a correlation between these two indicators, suggesting that users – paradoxically – are attracted to the online media most tempted to use stories from other sources. These are again also placed in opposition to the internal shovel-ware (the publishing of news stories from the printed paper or the evening television news bulletin). In other words, a bigger online desk means more original online journalism and less need to re-use stories from the “parent” media platform. This form of converged news production and sharing of content negatively influences users, and to a greater degree than a high level of shovel-ware.
Figure 1 illustrates how each variable, through the factor loadings of the PCA with 2008 data, projects each online newspaper onto a map of online media, as shown in Figure 2. This means the oppositions in forms of capital in the constructed field stand out more clearly, and provides the positions of the online newspapers, i.e. how much and which combination of the two forms of capital they have. On the vertical component (Axis 1) we find a contrast between on the one hand external shovel-ware, number of users and proportion of news agency articles, and on the other hand number of journalists and proportion of internal shovel-ware and original production. We can identify a contrast between strength in users gained through external news channels and journalistic strength, which signifies that the different online news sites have different strategies of production. However, this contradiction is modified on the horizontal component (Axis 2), which primarily contrasts original production with internal shovel-ware and news agency articles. Here, the number of users and journalists is close to the pole of original production, thus creating a two-dimensional field – shown in Figure 2 – vertically distinguished based on popularity (+user strength vs. –user strength) and horizontally on the strength of journalistic profile.
In the following section we will present findings regarding the relational differences between the online newspapers based on the content analysis and the indicators summarised above. Figure 2 can be seen as a “map” of the online newspapers in the field of news based on the variables related to economic and journalistic capital (Table 1).
The horizontal component in Figure 2 is constructed by the relational opposition between “users” and “quote stories” vs. “distribution for other internal platforms” and “wire service stories”. This places
Figure 2 (based on the 2008 analysis) shows how the number of stories cited from competing media relates to reader numbers, and that these indicators are in opposition to the indicator “original production” “Original production” is the coding category that remains when the other news items have been coded as either “quote stories” (external and internal) or “wire stories”.
In Figure 1, the indicators “number of journalists” and “original production” were placed almost diametrically opposite the amount of “news agency articles”, which shows they have a strong negative relation. The online site
This places them in opposition
The map (figure 2) indicates that we can differentiate between two strong main axes, one based on the relation between users and internal shovel-ware, and the other on the oppositional relation between internal shovel-ware and the number of journalists and original production of news. The two axes above can be interpreted as user strength (the ability to attract a high number of readers and users to the site) and journalistic strength (the ability to produce original journalism). The user dimension pulls online media sites towards the right of the map, and the journalistic strength component pulls them towards the left. The sharing of content across platforms in the most converged media houses pulls the newspapers towards the bottom, and so do a high percentage of wire agency stories to an extent.
Another interesting result from the mapping is that it seems shovel-ware from the printed newspapers does not necessarily please the users. These two indicators are in opposition to each other, which places online newspapers with a great deal of shovel-ware from the printed paper at the bottom of the map, and those with a lower percentage of articles from television or print platforms in the upper part of the map. If we compare Figure 1 and Figure 3 (below), we can see how the strength of the two axes changes from 2008 to 2010. The difference between the “journalistic axis” and the “user axis” is even stronger – the sharper the angle between the different indicators placed in the map, the greater their differentiation.
The map of Figure 3 shows that having many journalists is polarised from the indicator “news agency articles” to a greater extent, and it also reveals polarisation between the external shovel-ware and internal shovel-ware indicators. The PCA is thus able to show how the focus of the online newspapers differentiated during this two year period, and how the choices they make on a daily basis – for example, whether to quote a competitor or publish their own take on the story by, for instance, calling another source – has a structural influence on their relational position in the Danish field of news. Taken together, most of the online newspapers have become more
Figure 4 shows that the two axes remain the same (but as in Figure 3, more polarised), with
During the period,
To compare the production details of online and offline platforms we chose to analyse the content of the printed papers and the main evening news bulletins of the two television channels included in the study. This illustrates first how the axes are even stronger (the differences between the different media platforms are much more pronounced), and second how this places the online newspapers at the bottom of the map and their corresponding offline platforms at the top. The online newspapers thus have more in common (are close together on the map) and differentiate themselves from the “parent” platforms based on the chosen indicators. Figure 5 below shows the strength of the axes.
Component 1 (vertical) is now the opposition between originally produced journalism and wire service stories, and between users and both types of quote (internal and external). The construction of component 2 is based on an opposition between users and external quotes vs. internal quotes and wire service stories. We see that the axes are the same, but now the journalistic component has become stronger than the user component (which transposes them within the map).
More journalists thus equal fewer wire stories, and this becomes even clearer when the “parent media” are included in the analysis. A content analysis would be able to indicate this simply by comparing the number of original stories with the number of journalist employed. However, the PCA allows the researcher to consider many indicators taken together and to visually present the closely related indicators, and the online newspapers that are most similar and most different, based on all the chosen indicators. The next step is to place them in relation to each other in the expanded field of news.
The different media organisations and their platforms place themselves along the two axes. Component 1 (vertical) represents an opposition between the original journalistic production and a high number of journalists (the journalistic strength) vs. desks and organisations with higher dependence on news agency stories and internal and externally cited stories (and to an extent, user numbers). The horizontal component 2 places the media platforms along an opposition between external quotes to the left and internal quotes and news agency stories to the right.
Figure 6 shows us that the older media platforms are more closely related with regard to the journalistic and user strength, and that the online newspapers are equally related. We also see that both television channels (TV 2 and DR) are closer to the online newspapers than the printed papers in their “journalistic strength”. The visual presentation of their positions in relation to journalistic and user strength provides a differentiation between new and old media and between tabloid and broadsheet media. Old media has more journalistic strength than new online media platforms, and both new and old platforms take different positions in the Danish field of news when it comes to user strength. Theoretically, we can understand this as a field of relational differences in journalistic capital and economic capital, with the old media being dominant in both forms. However, some online desks appear more similar to some old media platforms. For example, we see that broadcast news bulletins are “closer” to online news sites in the production hierarchy, and a hypothesis could be, that the closer online and offline desk are in the maps, or at least the closer in the same side of the map, the less tension over strategies of production.
In this article, we have revealed how online media differentiate themselves according to their different strategies for online news production, and thus take different relational positions in a constructed Danish field of news. The analytical model used in this study has, in an explorative manner, emphasised how production indicators can be analysed quantitatively, and revealed how a number of online and offline media platforms differentiate themselves. It has also shown how different platforms are connected in the ecology of news production, and revealed certain “hierarchies of production” and how different news sites are positioned in relation to two major, dominant forms of capital in the field. Furthermore, the model has traced how a number of Danish online and offline media platforms have evolved over time.
From a field perspective, we can view the development from 2008 to 2010 as the result of the increased autonomy of the online platforms within the media organisations. The question is whether this increased autonomy means a less converged newsroom, and whether that goes against a European tendency towards more convergence as media organisations attempt to increase efficiency. The answer is both yes and no. The fact that the online platforms are producing a greater amount of original material does not mean there is less convergence. It means the positions and the strategies of production inside the media organisation are altered, thus changing the “production hierarchies”. Where the online platform historically often took the distribution role, they seem to have adjusted their position to became more
On the other hand, it seems that Danish newsrooms at least are moving away from media convergence towards separate production processes for online and print platforms, and thus consideration of this depends on what one understands by convergence. In future studies, we must try to conceptualise this term further to avoid a diluting of its explanatory power in the research on online journalism production. It is too easy for media organisations to proclaim “media convergence” and for researchers to take this term literally without actually analysing what it means for the organisational structuring of news production. How should we account for the different ways of telling a story for online, print and broadcast news? Does telling one story on several platforms require more in-depth journalism than telling the same story in the same way on the different platforms? And how do different ways of storytelling and presenting news relate to audiences and advertising revenue?
Theoretically, the contribution of this article has been to emphasise how we can operationalise Bourdieu’s concepts of