Cross-National Similarities and Differences between Legacy and Digital-Born News Media Audiences

The decline of the news business model for print newspapers in manyWestern countries and the digital disruption caused by the Internet have influenced the rise of digital-born news media. These new media are different from legacy brands in terms of businessmodels, distribution strategies, corporate organisation, and editorial priorities. It would be expected that the different nature of both legacy and digital-born newsmedia has driven to two types of significantly different audiences. This article aims to analyse whether there are significant differences between the users of these two types of media, by comparing the online audiences of five European countries’ (United Kingdom, Germany, France, Spain, and Italy) legacy and digital-born media brands in 2015 and 2019. The article will focus on four aspects: demographic and socioeconomics profiles (sex, age, income and level of education); interest in news; payment for online news; and media trust.


Introduction
The rise of the Internet as a content distributor is at the epicentre of the crisis of the media business model. Technological advances have reduced the entry barriers to content creation and distribution by making professional-quality production tools widely accessible at low price (Küng, 2008). The emergence of new news providers and technological distribution platforms has changed the way audiences get news (Curran, Fenton, & Freedman, 2012;Küng, Picard, & Towse, 2008;McDowell, 2011), making the general audience's news 'diet' more abundant and diverse than ever. According to the Digital News Report (2015 and 2019), 62% of respondents use five or more news sources weekly (65% in 2019), and 42% of them read seven or more different sources (45% in 2019), from both online and/or offline outlets (Newman, Fletcher, Kalogeropoulos, & Nielsen, 2019;Newman, Levy, & Nielsen, 2015).
Specifically, this technological difference between online and offline products has guided research on news consumption trends, mostly from the uses and gratifications field. According to this framework, users make a rational decision, evaluating utilities, features, and rewards provided by every kind of media (Lin, Salwen, & Abdulla, 2005). Many of these studies have focused on analysing the displacement or complementary effect of print editions by digital newspapers (Althaus & Tewksbury, 2000;Dutta-Bergman, 2004;Flavian & Gurrea, 2009;Newell, Pilotta, & Thomas, 2008;Westlund & Färdigh, 2012), with mixed results, largely due to the different methodologies applied (Chyi & Lee, 2013). Some authors suggest that the Internet has had a competitive displacement effect on traditional media (Dimmick, Chen, & Li, 2004;Filistrucchi, 2005;Ha & Fang, 2012). On the contrary, Westlund and Fardigh (2015) find emerging patterns of complementary news consumption.
Probably, the difficulty in reaching clear conclusions has its roots in the fact that the majority of studies have a technological focus: they tend to look at displacement and complementarity through platforms. However, few studies have focused on this issue from a media brands perspective. Given the increased competition and multichannel media consumption, it seems crucial for media companies to build up and strengthen their brands.

Theoretical Framework
Different media distributors and media formats used to be clearly different media products with distinctive production, marketing, and consumption patterns (McDowell, 2011). However, the Internet has led to the foundation of the so-called new pure players or digitalborn news media. These new media outlets are different from legacy brands in terms of business models, distribution strategies, corporate organisation, and editorial priorities (Nicholls, Shabbir, & Nielsen, 2016). They only operate online and are carving out a profitable and sustainable business on the net, using new business models as the so-called niche journalism (Cook & Sirkkunen, 2013). Meanwhile, legacy media usually manages both kinds of sources, offline (print or broadcast) and online. Although there is not a unique definition of legacy firms, they present some common traits: Their brand heritage is anchored in the quality of their customer relations, as well as in the quality of their products (McDowell, 2011;Tungate, 2005), and their consumers value legacy brand identities (Lowe & Stavitsky, 2016). For legacy brands, it is economically rational to cover news in a quality-oriented manner. By including quality as a part of their brand identity, media outlets find an audience that is ready to pay money, or at least attention, for this sort of coverage (Siegert, Gerth, & Rademacher, 2011). From the managerial point of view, the majority of full-time jobs for journalists are in this shrinking legacy media sector (McChesney, 2012), a sector with high costs that is vulnerable to downturns in the economy (Organization for Economic Co-operation and Development, 2010).
It would be natural to expect two different types of audience as a result of the difference between legacy and digital-born news outlets. This hypothesis is based on previous research, suggesting that users/audience perceive and use different types of news differently. Ots (2010) asserted that media brands create value for audiences wanting specific content and advertisers wanting to reach specific audiences. As a result, brand equity creates a different brand image in people's mind and so far, a different response in consumers. A strong brand will foster stronger attitudes and behaviours than those fostered by a weak or anonymous brand (Siegert, Förster, Chan-Olmsted, & Ots, 2015). In communicating the key characteristics of legacy brands, media managers could differentiate their outlets from those of their competitors. As a result, they prevent imitation, stabilise, and increase their audience in the long term. However, there is no guarantee that audience perception of the legacy brands would be in line with media managers' intentions. In fact, contrary to expectations, legacy outlets and digital born media are not getting substantially different audiences (Arrese & Kaufmann, 2016).
The analysis of the two kinds of audiences proposed in this article seeks to show the relevance of brands in news markets. This study complements others with a more media-centric, technological approach (Siegert et al., 2015), following the path of other scholars (Benson, Blach-Ørsten, Powers, Willig, & Vera Zambrano, 2012;Humprecht & Büchel, 2013;Stetka & Örnebring, 2013).
This article focuses on three areas of research that have been analysed in previous studies of the differences and similarities between legacy media and digital-born new media: audience segmentation; revenue models; and media trust. The conclusions of this study could be of interest not only from a journalistic point of view, but from a managerial perspective.

Audience Segmentation
As stated above, one of the relevant topics in the studies of online and offline media has been the differences in the uses and gratifications perceived by the two types of audiences (Filistrucchi, 2005;Newell et al., 2008;Westlund & Färdigh, 2011). New and different types of media formats should create different kinds of audiences in relation to their needs and motivations to use media and produce more audience segmentation. In that sense, the new digital landscape contributes to a greater heterogeneity in markets (Mitchelstein & Boczkowski, 2010), due to the abundance of digital media, which incentivises the specialisation and, therefore, the targeting of smaller audiences defined by multiple and segmented interests (Arrese & Kaufmann, 2016;Fortunati, Deuze, & de Luca, 2014). Along the same lines, previous research has identified attitudinal variables as key factors driving the media selection process when compared with traditional media (Chyi & Chadha, 2011;Chyi & Lasorsa, 2002;de Waal, Schönbach, & Lauf, 2005).
By contrast, Webster and Ksiazek (2012) assert that the Internet has concentrated a vast amount of the audience attention around the leading news organisations, which are quite undifferentiated and difficult to be segmented and predictable. According to Arrese and Kaufmann (2016), these two apparently contradictory views are not necessarily incompatible, and other reasonings must be studied to understand and find answers to the homogenisation versus segmentation debate. This article adds a new focus to this question, studying whether different types of media (legacy brands or digital-born) create significantly different readership segments.

Revenue Models
Over the last few years, online news organisations all over the world have erected paywalls (Arrese, 2016;Sjøvaag, 2016). For decades, legacy newspapers have been oriented towards protecting their current markets, serving existing customers, and reacting to innovations in media markets. They were more focused on defending their flagship brand and customer base than on being proactive (Herbert & Thurman, 2007;Holm, 2016). However, the success of certain media firms, such as The Wall Street Journal and The New York Times (Mensing, 2007), and the need to compensate for losses in the traditional advertising market (Sjøvaag, 2016) have changed the revenue strategies of media firms, from free to pay.
The study of revenue models from a brand perspective provides new insight into this area of research. In the online news market, with an abundant supply of news available, news has been perceived as a highly substitutable commodity, and the reluctance to pay for digital news is widespread among the public (Gundlach & Hofmann, 2017). On the other hand, the success of certain firms is attributed to the quality and exclusiveness of the content (Vara-Miguel, Sanjurjo-San Martín, & Díaz-Espina, 2014) and the presence of strong brands (Bleyen & van Hove, 2010;McDowell, 2011;Mensing, 2007;Sjøvaag, 2016). The question is whether the use of legacy media versus digital-born media turns into a greater commitment to paying for digital news.

Media Trust
In the current media scenario, with the emergence of new alternative channels of information, some questions have arisen regarding the relationship between trust and media. Although research on media trust is abundant, the literature lacks consensus, not only on the notion of media trust, but also on the elements or dimensions that comprise it, probably due to the variety of disciplines and methodologies applied (Kiousis, 2001;Kohring & Matthes, 2007). While early studies focus on the trust generated by sources with a clear persuasive aim (Berlo, Lemert, & Mertz, 1969;Hovland & Weiss, 1951), in subsequent years scholars focused on the credibility of channels and media outlets (Johnson & Kaye, 1998;Westley & Severin, 1964) and the relationship between trust and media use (Abel & Wirth, 1977;Gaziano & McGrath, 1986). Most of these studies show a modest relation between media trust and media consumption, and those who trust media the most are more likely to use traditional outlets, while sceptics pay more attention to alternative sources (Ardèvol-Abreu, Hooker, & Gil de Zúñiga, 2018;Fletcher & Park, 2017;Kiousis, 2001;Tsfati, 2010;Tsfati & Cappella, 2003. Finally, recent worries about fake news, misinformation, and the often-low trust in the news have highlighted the value of legacy brands as trusted media. According to the Digital News Report (Newman et al., 2019), over a quarter (26%) of the respondents have started to rely on more reputable news sources (40% in the US), and a further quarter have stopped using sources of dubious reputation. The analysis of media trust from this brand perspective could help to understand whether trust is perceived as a significant asset of legacy media when compared with digital-born media.

Research Questions and Hypothesis
As stated above, the aim of this article is to investigate whether the nature of legacy media brands or digitalborn media is relevant in explaining audience differentiation in online news markets, considering the issues discussed in the previous section. More specifically, this article analyses whether or not there are significant differences between online users of these two types of media firms by comparing the digital audiences of the main legacy and digital-born media brands of five European countries (United Kingdom, Germany, France, Spain, and Italy) in 2015 and 2019. The specific research questions are as follows: RQ1. Do online audiences of legacy media and digitalborn media have significantly different demographic and socioeconomic profiles?
RQ2. Are online audiences of legacy media and digitalborn media significantly different in their interest in news and in their frequency of news consumption?
RQ3. Do online audiences of legacy media and digitalborn media differ significantly in their behaviour and attitudes toward payment for online news content?
RQ4. Do online audiences of legacy media and digitalborn media differ significantly in their media trust perception?
The general hypothesis is that significant differences exist in all of the questions under research, something that seems reasonable considering the different natures of these two types of media firms. A priori, it could be stated that legacy brands have a similar audience, in terms of profile and behaviour, to that of traditional media outlets (higher age, income and level of education; RQ1), they declare more interest in news and, con-sequently, access news more frequently (RQ2) and are more likely to pay for digital news (RQ3). Finally, legacy users are more inclined to trust mass media and are more sceptical about social media (RQ4), in contrast to digitalborn media audiences.

Sample, Variables, and Measurement
The analysis is based on data corresponding to the United Kingdom, Germany, France, Spain, and Italy from the survey carried out for the Digital News Report 2015 and 2019, in which some questions directly related to this article were included. YouGov, using an online questionnaire in late January-early February 2014 and 2019, conducted the survey fieldwork, commissioned by the Reuters Institute for the Study of Journalism. The data were weighted to targets based on census/industryaccepted data, such as age, gender, region, newspaper readership, and social grade, to reflect the population of each country. The sample is reflective of the adult population (18+) with access to the Internet. As the survey deals with news consumption, it filtered out anyone who said that they had not consumed any news in the past month (average around 3%) in order to ensure that irrelevant responses did not adversely affect data quality (see Table 1).
In order to get the two types of readers (legacy users and digital-born users), a subsample was generated using the responses to the question: 'Which, if any, of the following have you used to access news in the last week via online platforms (web, mobile, tablet, e-reader)? Please select all that apply.' Respondents have to select options from a list of each country's main legacy and digital-born outlets. According to their responses, two types of readers were obtained: those who use more legacy media than digital-born media and those who use more digital-born media than legacy outlets (N 2015 = 6,677, 65.8% of the total sample; N 2019 = 6,532, 64.9% of the total sample). Additionally, those who use exactly the same number of legacy and digital-born media have been excluded from the subsample in order to have two clear-cut groups.
Once the grouping variable is established, we select those survey questions used as variables to examine the hypotheses related to the RQs. All the variables, as explained below, should be considered as reasonable-not exhaustive-proxy measures for the overarching characteristics under investigation.
Interest in news and frequency of news consumption (RQ2) were measured through the following questions: 'How interested, if at all, would you say you are in the news?' (extremely interested/very interested/somewhat interested/not very interested/not at all interested) and 'Typically, how often do you access news? By news we mean national, international, regional/local news and other topical events accessed via any platform (radio, TV, newspaper, or online)' (less often than once a week/once a week to six times a week/once a day to five times a day/six times a day or more).
Attitudes towards payment for digital news were used as proxy indicators of the preference for revenue models more dependent on free or paid content (RQ3). The exact question asked was 'Have you paid for online news content, or accessed a paid online news service in the last year? (this could be digital subscription, combined digital/print subscription or one-off payment for an article or app)' (yes/no).
Finally, we measured the media trust (RQ4) of the audience through two questions. The first, 'Thinking about news in general, do you agree or disagree with the following statement? "I think you can trust most news most of the time,"' focused on the credibility of the media in general, but the second asked about the audience's trust in the specific media outlets they used: 'Thinking specifically about news sources that you use, do you agree or disagree with the following statement? "I think I can trust most of the news that I use most of the time."' The possible answers to both questions were strongly disagree/tend to disagree/neither agree nor disagree/tend to agree/strongly agree.

Statistical Analysis
We decided to use the chi-squared test to analyse categorical variables (gender, income, and payment for online news) and the Mann-Whitney U-test for metric variables (education, age, interest in news, frequency of news consumption, media trust, and trust in your own media). The selection of non-parametric tests, including metric variables, was due to the lack of normality in the distribution of values. An examination of the standardised skewness coefficient and the standardised kurtosis coefficient revealed serious departures from normality for all of the metric variables. Table 2 shows the distribution and evolution of the two audience groups. Most of the respondents used more legacy media than digital-born media as a source of news in all of the countries studied. Five years later, this trend was stronger and the percentage of respondents getting news from legacy media grew in every country except Germany. In order to test if a significant difference exists between the groups of readers in terms of sociodemographic variables (RQ1), a chi-squared was used to analyse gender and income and a Mann-Whitney U-test was applied for age and level of education. The data revealed a statistically significant difference in 2015 between groups in gender (X 2 (1, N = 6,677) = 23.65, p = 0.000), income (X 2 (2, N = 5,880) = 16.68, p = 0.000), and education (U = 3300886, p = 0.000), but not age (U = 3471020, p = 0.065). Table 3 shows that in 2015, male audiences (52%) with high income (28%) and a bachelor's or postgraduate degree (38%) tend to use more legacy media than born-digital. Five years later, there were significant differences in all sociodemographic variables: gender (X 2 (1, N = 6,531) = 26.24, p = 0.000); income (X 2 (2, N = 5,651) = 11.90, p = 0.003), education (U = 3202889, p = 0.000); and age (U = 3082176, p = 0.000). Table 4 shows the same trend in 2019. Those who use more legacy media than born-digital media tend to be male, with higher levels of income and education than those who read more digital-born outlets. Surprisingly, respondents under 44 years read more legacy brands (45%) than born-digital media (33%), while those over 45 years use more native media (66%) than legacy media (54%).

Results
By countries, the data shows that Spain and especially France differ from the general trend, as there are no sociodemographic differences between the two groups in these countries, either in 2015 or in 2019.
With regard to the amount of interest in news and the level of news consumption (RQ2), the data shows significant differences between the two groups in 2015interest in news (U = 3051343, p = 0.000) and frequency of news use (U = 3185492, p = 0.000)-and in 2019interest in news (U = 2922834, p = 0.000) and frequency of news use (U = 3012158, p = 0.000).
In 2015 (see Table 5), those who use more legacy media were more interested in news (34% are extremely interested, versus 24% of native users) and read news more frequently (61% of legacy users access six times a day or more, versus 50% of born-digital users). Five years later (see Table 6), the data shows similar differences between the two groups: 30% of legacy users are extremely interested in news, versus 21% of native users, and the 27% of them access 6 times a day or more to news, versus the 17% of born-digital users. The differences between the two groups of users occur in all the countries analysed, except France in 2015, where there are no significant differences in both variables: interest in news (U = 120539, p = 0.257) and frequency of news use (U = 124422, p = 0.770). However, in 2019, significant differences appear between the two groups of French users.
The results on attitudes toward payment for online news content (RQ3) show also significant differences between the two groups in 2015 (X 2 (1, N = 6,557) = 20.23, p = 0.000) and 2019 (X 2 (1, N = 6,333) = 30.20, p = 0.000). In 2015 (see Table 5), those who read more legacy brands than born-digital media were more likely to pay for online news (12%) than those who use native media (8%). In 2019, the percentages were 13% and 7%, respectively (see Table 6). Significant differences occur in all countries except France (X 2 (1, N = 1,078) = 3.17, p = 0.075) and Spain (X 2 (1, N = 1,426) = 2.79, p = 0.094) in 2015, and Italy in 2019 (X 2 (1, N = 1,310) = 1.63, p = 0.201).  Finally, the analysis of the groups' media trust perception (RQ4) does not provide a clear conclusion. Although both types of readers show significant differences in 2015 (U = 3453343, p = 0.027) and 2019 (U = 3341165, p = 0.039) and those who use more legacy brands than digital-born media show higher media trust (44% of legacy respondents usually trust on news versus 40% of native users in 2015; 40% and 37% in 2019, respectively; see Tables 5 and 6), there are some differences in the five countries analysed. In 2015, the data show no significant differences between the two types of users in Spain (U = 159969, p = 0.803), Italy (U = 188917, p = 0.223), and France (U = 119280, p = 0.161). However, the similarity in the Spanish and Italian audience is around the distrust on media, while in France is around trust. As Table 5 shows, only a third of Spanish and Italian respondents usually trust the news, in contrast with 41% of the French users. That is, distrust is widespread in Spain and Italy in 2015, regardless of the type of media outlet, while in France, trust is the common ground for both groups.
Five years later (see Table 6), there are no differences between the two types of readers in France (U = 125871, p = 0.966), Germany (U = 140053, p = 0.063), and Italy (U = 137063, p = 0.181). In this case, French users coincide in their media distrust, regardless of the kind of outlet (only a quarter of French respondents usually trust media), while in Germany and Italy it is the opposite.
Additionally, we found no evidence of the third person effect (Davison, 1983;Perloff, 2009) in the research. The differences between groups in all countries (except Spain) are identical in 2015 (U = 3463762, p = 0.037) and 2019 (U = 3315316, p = 0.011) when users were asked about the trust they have in the news that they usually read.

Conclusion
The data confirms that, although the online news market continues to be dominated by legacy brands (Bruno & Nielsen, 2012) and the most popular sites are those of major news organisations (McDowell, 2011), digitalborn media have become part of the media ecosystem in all the countries analysed. This article confirms the divergence between legacy media users and native users. By and large, legacy users tend to be male and with higher income and education levels than native users (RQ1). However, cross-national data are not homogeneous in all countries. It is interesting to note that the two groups of users differ more when journalistic variables are analysed (RQ2 and RQ3). In all countries (except France in both years), legacy brands audiences show greater interest, more news consumption, and more willingness to pay for digital news than the digital-born users.
In relation to media trust, data about the existence (or not) of differences between the two groups are less clear (RQ4). Global data shows significant differences in 2015 and 2019, and those who use more legacy brand than digital-born outlets are more likely to trust the news. In that sense, media scepticism is more widespread among native users.
Finally, the longitudinal analysis shows that in 2019, the percentage of very interested and heavy users decreased in both groups from 2015, although more intensely among native users. Additionally, trust in news also declined in 2019 in both types of users, especially among digital-born outlet readers (from 40.9% to 37.6%).
This data could demonstrate the strength of legacy brands, and its ability to get and keep loyal customers. As a whole, all the indicators analysed (interest, reading frequency, payment for news, and trust) are more positive to legacy media. And when they decreased in 2019, they did with less intensity than the native ones, widening the gap between the two groups. From a managerial viewpoint, legacy brands have three strong competitive advantages over digital-born media. Firstly, they enjoy a better differentiation in relation to competitors, not only in sociodemographic terms. Secondly, legacy media users show a higher engagement and tendency to loyalty to their brands. Thirdly, this engagement drives to a higher desire to continue buying the same brand, at least in a higher percentage than native users (Chan-Olmsted, 2011). Any global news branding strategy should be based on these ideas, as brand recognition constitute the baseline from which to get and keep loyal customers. By far, this conclusion does not underestimate the role that digital-born outlets could play in the media landscape. Even if legacy media organisations succeed in moving readers towards paywalls, free alternatives are likely to remain in news markets (Sjøvaag, 2016).
Although a complete review of the factors contributing to media trust is beyond the scope of this study, a deeper look at this variable is necessary. Taken as a whole, the data shows the existence of significant differences among legacy users and native users in both years, as the former are more likely to rely more on news than the latter. This is not surprising, as native media were born in many cases as an alternative to legacy options, perceived as unreliable, mostly by younger people, as confirmed in previous studies (Tsfati, 2010;Tsfati & Cappella, 2003. Brand reputation could be a crucial factor in the trust transfer process from traditional media outlets to digital ones. Those who read more legacy media trust more in news, no matter the channel (traditional or digital) used. This is an additional competitive advantage for legacy brands, and any media brand extension strategy should be based on this idea. However, the analysis by countries suggests a need for deeper research on other factors (like national media systems or institutional media trust) that could explain better why France and Italy there are no significant differences between the two groups in 2015 and 2019; why native users in Germany show greater confidence in the news than legacy users in both years; or why, when institutional trust in media is widespread (such as in France in 2015 or Germany in 2019), there are no differences between legacy and native users. The relevance of these studies could improve the strategies for the media to follow in order to avoid the dilution of the differential value of legacy brands.