1. Introduction

In 1979, the British newspaper Birmingham Post and Mail launched Viewtel 202 –considered to be the first electronic newspaper– and thus established the starting point for all subsequent digital newspapers (Armentia, Caminos, Elexgaray, Marín & Merchán, 2000). The following computer revolution fuelled interest in new technology, particularly after the development of the World Wide Web. This project, devised by Tim Berners, would transform the global media scene and clear the way for cyberjournalism in the early nineteen-nineties, “a time when the traditional print media industry foresaw on the horizon possible difficulties in maintaining their pace of growth” (López, Silva & Toural, 2015, p. 75). Since then, news professionals have had to face more than a few challenges stemming from media convergence. This convergence is seen as a multidimensional process that, due to the implementation of new technology, affects both professional and business environments, and leads to the integration of new tools, spaces and working methods in newsrooms (Salaverría, García Avilés & Masip, 2010, p. 48).

The traditional press in Spain took their first steps towards the creation of digital versions in November 1994 with El Mundo and El Periódico de Catalunya. In 1995, newspapers such as the ABC, La Vanguardia and El Diario Vasco also went online, while El País followed a year later (López et al., 2015). Most of the first cybermedia initially opted for shovelware (Bardoel & Deuze, 2001), an editorial model that reproduced very literally on the web the content originally produced for the print edition. Even today, “many of the digital media are still directly indebted to traditional media in their genres and communicative codes”, although many others, “at least the most advanced ones, show specific forms of expression” (Salaverría, 2019).

2. State of the art

The leap to digitalisation forced newspapers to face a series of changes that revolutionised their working methods, from the selection of news events to the process of preparing and publishing content. Used to working with limited time frames so that new stories would reach readers first thing the next day, journalists had to adapt not only to the demands of a society that was changing the way it consumed information, but also to a working environment characterised by maximum immediacy, global access to sources and the new tools offered by the internet. These new tools offered the possibility of updating information, multimediality, hypertextuality and interactivity with users, all of which have great advantages, but also some trade-offs, depending to a large extent on how they are used.

The possibility of continually updating the news has changed our view of journalism, which is now inconceivable without this constant addition of the latest information. Yet it is also true that being constantly informed and updated can increase the reader’s feeling of saturation and information overload, and, furthermore, “magnifies, to an exclusively temporary degree, the latest information that appears in the most prominent place on the page” (García de Torres & Pou, 2003, p. 70). Multimediality mixes narrative formulas from the press, radio and television (Parra & Álvarez, 2004, p. 145) in the same digital medium, which makes it possible to integrate audio-visual, textual and infographic content (García de Torres & Pou, 2003) in a single story, although “the abuse of integrated formats in the same news item can have a perverse effect by pointing out too many points of interest to the reader” (p. 73). As for hyper textuality, an “open system that allows consumers to construct their own reading paths by jumping from lexia to lexia according to their interests” (Aguirre, 1997, p. 284) and decide whether to continue reading or redirect to another text (Landow, 2009) according to their needs in “a non-linear reading environment” (Porto & Flores, 2012, p. 46), this can also lead to information overload, or to losing the thread of the piece they were initially reading.

At the same time, digital newspapers face another challenge rising from technological transformation, one which has been further aggravated by the economic crisis: that of establishing a new business model. Casero-Ripollés (2014) argues that without a viable model, journalism will find it difficult to avoid abandoning some of its social and democratic functions. The solutions involve either generating income solely from advertising revenue or initiating a paid subscription model to financially sustain the media entity without depending solely on advertisers, which may thus enable a greater capacity to offer quality content. In fact, “the quality of the product is the main motivation cited by Internet users behind paying for news content [...], it is not so much about getting big exclusive scoops as it is about offering information in a differentiated, careful way, with added extras and greater visual support” (Segura-Anaya, Marta-Lazo & Nogales-Bocio, 2019). Following this line of thinking, elpais.com (El País, 2020) recently announced the launch of a subscription model “with the aim of making professional, high-quality, globally-reaching journalism sustainable for the future”.

Journalistic quality may be affected precisely by the abundance of information available on the internet, which has favoured the emergence of a less demanding consumer, accustomed to a brief review of headlines and immediacy of information, a fact which thus marginalises in-depth analysis of the facts (Casero-Ripollés, 2014). Furthermore, the rise of social networks and other digital formats contributes to this habit of brief, superficial reading and therefore forces journalists to focus their efforts on reporting more concisely. This encourages a form of cyberjournalism that has “often given in to reporting the latest information, to the easy click and to being more concerned with traffic than with quality, and that has therefore ended up turning its back on a large part of the narrative and documentary possibilities of hypertext, taking advantage of it only as a resource to multiply visits” (Salaverría, 2019).

The quality of journalistic content in cybermedia has been the subject of academic analysis, giving rise to studies with disparate results both internationally (among others, Odriozola-Chéné, Aguirre & Bernal, 2016; Salas, Hernández & Realyvázquez, 2018) and in Spain. Ramírez de la Piscina, Gorosarri, Aiestaran, Zabalondo and Aguirre (2014) observed the evolution of the quality of news published in leading European newspapers over a decade, and concluded that all of them seemed to have forgotten journalism´s historically attributed social function as the guardian of society’s best interests. Meanwhile, Gómez-Mompart, Gutiérrez-Lozano and Palau (2015, p. 13) investigated the perceptions of quality in Spanish media held by its professionals. They highlighted shortcomings related to access to sources, data processing and verification, but also to the challenges arising from adaptation to new technology and the impositions of immediacy. Others, conversely, point to the excellent quality of texts in Spanish cybermedia, considering criteria such as content, access to information, visibility, and usability user –participation and interactivity– (Iglesias & González Díaz, 2012). More recently, Costera Meijer and Bijleveld (2016) carried out a study to investigate what readers understand by what is known as “valuable journalism”.

The terms most often used in academia to refer to the quality of information are journalistic quality, news quality and news content quality (Ramírez de la Piscina et al., 2014). The former will be used in this study. McQuail (1992, p. 17) described news quality analysis as “the independent evaluation of mass media provision according to alternative ‘public interest’ criteria through objective and systematic research methods”. He would later focus the idea of quality on the credibility and reliability of information by understanding that it should be objective in terms of accuracy, honesty, truthfulness, reliability, being factual rather than opinionated, and being sufficiently complete (McQuail, 2005).

The complexity of this concept and the differences in criteria relating to its meaning have given rise to very diverse works (among others, Merrill, 1968; Hansen, Neuzil, & Ward, 1998; Koch, 2008) and have led to the creation of different studies for its evaluation. Alessandri et al. (2001) developed a system called VAP to measure the processes of news selection and content creation. The former is studied by focussing on indicators of access to sources and information equity, and the latter by looking at guidelines of style, content, and the information´s emphasis. Vehkoo (2010) added the independence of the media, the credibility of the information, the level of interpretation and contextualisation, as well as the number of journalists responsible for producing the piece. Rodríguez-Martínez, Codina and Pedraza-Jiménez (2012) proposed a methodology to identify, through 36 indicators, “the tools, services and communication strategies used to ensure that the information reaches their users properly, as well as evaluating the quality of their media websites”. Lacy and Rosenstiel (2015) established criteria of accuracy, credibility, diversity, depth, breadth, human interest, and geographic proximity.

All these studies are very useful for a rigorous analysis of digital media textual quality. Eight million Spaniards accessed this type of media daily between January and September 2019 (AMI&Deloitte, 2019, p. 18) and 67% of internet users in Spain believe that news providers are fulfilling their task of informing with immediacy. However, only 53% consider that they provide depth and understanding (Vara, 2019). This raises the question of whether the subject of study in this paper, feature stories –a genre that delves into context and background and whose quality has not been addressed so far, and which offers an exhaustive analysis of data, different perspectives and multiple sources– could have been particularly affected by a loss of quality with its roots in the digital transformation, despite having “undergone a clear metamorphosis” (Larrondo, 2009) more than a decade ago. In this type of digital reporting, the writer´s style is characterised by a greater freedom of language, a more variable length of article, texts that allow for different ways of telling a story and by a greater repertoire of supporting elements, both textual and visual. In terms of the ways that ideas are organised, they present an open structure because they are designed to be published in hypertext media where the journalist has the possibility of linking different documents to the same text, allowing the reader to access “a potentially infinite variety of options” (López, 2003, p. 460).

Therefore, the aim of this paper is to analyse the evolution of the journalistic quality of feature stories published in elpais.com, from the moment when the digital version of El País –the most consulted Spanish-language digital newspaper in the world– started its activity in 1996, until 2019, and to discover whether this quality has been affected by technological development. As previously explained, the media in Spain has undergone major transformations due to the onward march of new technology, both in its business models and its daily work. This leads us to the hypothesis that with the arrival of digital journalism, journalistic pieces, and in this case feature stories, have suffered a loss of quality.

3. Methodology

The universe of analysis was feature stories taken from elpais.com –the most widely read general news digital newspaper (AIMC, 2020)– published between 1996, the year in which El País created its digital version, and 2019, inclusive. A sample of 72 pieces was analysed (see appendix) drawn at random from the months of March, June and October of each year, except for 1996, from which two feature articles published in June and October were drawn, as elpais.com had not yet been created in the March of that year.

Content analysis, a research method for the objective, systematic and quantitative description of information (Berelson, 1971), in this case press information, has been used. The starting point was the quantitative methodology for the evaluation of news quality established by Ramírez de la Piscina et al. (2014), to which modifications have been applied so as to adapt it to feature stories, which is the object of our study. A total of 20 parameters referring to four main aspects of the feature stories were thus evaluated, which were scored according to the number and quality of the parameters they contain. Although this methodology is described by these authors as quantitative, the scoring process has an inevitable qualitative component, so below are the justifications of its implementation:

1. The selection process of the information (in total, 2.5 points out of 10), which includes five parameters:

• Citation of the origin of the information: number of sources cited in the feature story. The minimum score was taken as 0 sources cited, with more than three sources receiving the maximum score (0.5 points).
• Nature of the sources: variety of sources used according to their exclusivity, their temporality, their entity, and their degree of identification (Parratt, 2017, p. 51). The maximum score is achieved when there is a plurality in the nature of the sources, i.e., more than three sources of different origin, while the minimum score is awarded when there is none.
• Balance between facts and statements: considering whether there is conjecture on the part of the author. To quantify this parameter, the balance between facts –objective information without interpretations by the author– and statements –testimonies from sources–, was evaluated, and depending on the criteria of the evaluators, the feature stories have been scored.
• Degree of topicality: this refers to the connection between the subject matter and current affairs. This parameter took into consideration whether at the time the story was published, the topic was part of the media and political agenda –and if this was the case, it was awarded the maximum score. In the case that it was a timeless topic, or did not meet any current affairs criteria, it was awarded the minimum score.
• Degree of newsworthiness or journalistic interest: according to the criteria of novelty, timeliness, proximity, prominence, conflict, impact, rarity, action, human interest or usefulness (Diezhandino, 2009, pp. 8-9). If the story met five of the above-mentioned criteria, it was awarded 0.5 points, while if it did not meet any of them, it was awarded 0 points.

2. The process of drafting the text (4.5 points in total), which includes 11 parameters. To the six criteria proposed by Ramírez de la Piscina et al. (2014), five –the last five– have been added, which particularly enrich the feature story:

• Degree of accuracy: reporting without falsifying or misrepresenting (Diezhandino, 2009, p. 11), avoiding terms with broad meanings and using the correct words to communicate specific concepts (Yanes, 2003). This parameter was assessed according to the definitions of Diezhandino (2009) and Yanes (2003).
• Degree of depth: the level at which contextualisation and background information are provided to present the facts as fully as possible (Martín Vivaldi, 1998, p. 103). This parameter was evaluated according to the concept of depth presented by Martín Vivaldi (1998).
• Use of different perspectives: the use of different points of view on a topic. In this case, the highest score was awarded when the author of the story used three or more different perspectives.
• Supporting textual elements: the use of summaries, excerpts, sidebars, and bold type. 0.5 points were awarded when the author used four or more supporting textual elements.
• Audio-visual resources (Del Olmo & Parratt, 2011): use of photographs, graphics, infographics, audio, and video. As in the previous parameter, the maximum score was reached when the author used four or more audio-visual resources to contextualise the story.
• Accuracy of language: understood from a grammatical point of view because “writing properly means never forgetting the fundamental structures of the language in which one writes” (Martín Vivaldi, 1998, p. 35). This parameter was scored with 0.5 points when the text contained no grammatical errors, and with 0 points when it contained more than four errors.
• Length: calculated by taking the highest length, the lowest length, and the overall average to obtain six sections with scores ranging from 0 to 0.5.
• Data analysis: this parameter was scored based on the author’s ability to present, explain, and interpret the data accompanying the information.
• Clear presentation of ideas: the highest score was awarded when the feature story contained a presentation of ideas correctly divided into paragraphs and through a logical structure, an essential element in the technique of interpretative reporting (Martín Vivaldi, 1998).
• Comprehensibility: this is achieved when no doubts are sown either by the overall content or by using words or images that may alter the meaning of the message.
• Additional information of practical utility: typical of service journalism, and which enriches the text with “the inclusion of a range of utilities for the user, although these are not directly linked to the strictly journalistic field” (Armentia, Caminos, Marín & Alberdi, 2005). This parameter was quantified by measuring the use of hypertexts with content of practical utility for the reader.

3. The social contribution of the feature story (1 point in total). The four criteria established by Ramírez de la Piscina et al. (2014) for news stories were reduced to two parameters –those which we consider to be of the most significance for measuring journalistic quality:

• Fostering social debate: promoting debate on social aspects of general interest.
• Combatting social marginalisation: making situations of social marginalisation or vulnerability visible.

4. A fourth characteristic, the digitisation of the text (2 points in total), has been added to the previous three. The quality of this was measured according to two parameters:

• Appropriate use of hypertext: this occurs when the hypertext allows the reader to go to external documents without losing the main thread of the initial text. The maximum score was assigned when the feature article contained more than four hypertexts and these were used appropriately, without hindering reading.
• Capacity for interaction between the user and the author of the article: this occurs when the necessary tools are provided so that the reader can interact adequately with the journalist or with other users. This aspect was mainly valued with 0.5 points when the reader had the possibility of commenting on the article and, therefore, interacting with the author and other users.

4. Analysis of the results

As indicated above, four main characteristics of each feature story were examined via their corresponding parameters to determine their qualitative evolution over time, the results of which examination are set out below.

4.1. The information selection process

The quality of the information selection process of the feature articles analysed evolved positively from 1996 to 2019, going from a score of 5.7 out of 10 in 1996 to a score of 7.2 in 2019, i.e. from a high pass to an outstanding. The score obtained in the five parameters used to measure this aspect improved in three cases –in one of them from a fail to a pass and in another from a fail to an outstanding. In one the score remained the same and in the last it fell (table 1)1.

Table 1. Evolution of the information selection process by parameters (0/10)

Source: prepared by the authors

Regarding the overall score of the information selection process, the average score over the 24 years analysed was 1.83 out of 2.5 –equivalent to 7.32 out of 10. However, the progression experienced over this time was uneven (figure 1) and marked by particularly low points in 1996 and 1999, by a positive period between 2000 and 2008, and by another negative period between 2009 and 2011, which started to improve from the latter year onwards.

Figure 1. Evolution of the quality of the information selection process

Source: prepared by the authors

It is striking that the lowest peaks detected –1996 and 1999– occurred at times when El País readership had declined and, in the case of 1999, coincided with the economic crisis and with the date when this newspaper reached its lowest audience penetration since 1987 (AIMC, 2000). In contrast, the positive period that began at the turn of the 21st century occurred at a time when elpais.com still used shovelware (Bardoel & Deuze, 2001) and therefore published the same stories as the paper version (it stopped doing so in 2009). Furthermore, this uptick coincides with the boom years for El País, whose audience penetration in print newspapers increased from 1.7% to 5.9% at this time (AIMC, 2008a). Another drop in quality, between 2009 and 2011, may have been influenced not only by the economic crisis, but also by the fact that this was the first stage of a differentiation between elpais.com and the print edition.

4.2. The process of drafting the text

As with the previous characteristic, the quality of the feature stories with respect to the content production process developed positively, going from an overall score of 5.76 in 1996 to a score of 7 in 2019, i.e. from a high pass to a “B”. As can be seen in table 2, the score rose in six of the 11 parameters used to measure this characteristic, remained the same in one and dropped slightly in four. Also illustrated in table 2, the score rose in six of the 11 parameters, remained the same in one and fell slightly in four.

Table 2. Evolution of the process of drafting the text by parameters (0/10)

Source: prepared by the authors

The analysis reveals interesting data regarding some parameters. This is the case with the use of audio-visual resources, which was nil –this being the lowest score of all 11– when elpais.com began its activity. It improved, however, as the possibilities offered by technology developed, reaching its highest peak at the end of 2013. It was not until 2000 –from which date it would always be so– that feature stories began to be published with photographs, either due to a lack of resources or to the high prices of digital cameras at the time; and it was not until 2009 that various visual elements began to be incorporated simultaneously. Infographics were scarce, appearing only –and then only rarely– from 2012 onwards, while it was only in 2018 that a video was included for the first time. Conversely, the use of supporting textual elements has had a more uneven development, improving in 2016 and 2019, but remaining notably low.

On the other hand, it is striking that some highs and lows in the quality of certain parameters coincide with circumstances related to the economic situation of the newspaper. For example, one of the three exceptions to the general positive trend in the degree of depth occurred in 2011, a year in which the Prisa Group made a net loss of 451 million euros (Grupo Prisa, 2012, p. 2). However, the use of different perspectives was outstanding between March 2000 and June 2003, at a time when revenues increased by 5% (Grupo Prisa, 2003, p. 2). For its part, the degree of comprehensibility attained, while remaining stable, experienced its highest score between 2005 and 2009, which is exactly the period in which El País generated positive data in audience penetration both in its print and digital versions (AIMC, 2019).

The degree of accuracy is outstanding between 2000 and 2008, which were years in which the reports in the digital version were the same as those appearing in print, though from 2009 onwards, the degree of accuracy is lower. This was also a period of economic prosperity for the daily El País print edition, which grew in audience penetration at this time, increasing by 40% between 2000 and 2007 (AIMC, 2008a), while elpais.com was placed as the second biggest website in terms of media audience penetration in Spain (AIMC, 2008b). Similarly in extension, the highest possible scores were attained in 2013, when elpais.com consolidated its position as the leading digital newspaper for general news and as the second most significant website in terms of daily readers (AIMC, 2013); and also in 2016, when it managed to overtake marca.com and become the leader in terms of daily users of digital media (AIMC, 2016).

A final noteworthy aspect is the fact that the scores on some parameters continue to fail, such as the ones for data analysis, or for the incorporation of additional information of practical use to the reader –which actually decreases. Other parameters, while still receiving good ratings, have dropped slightly, such as the degree of depth, the clear presentation of ideas and the degree of comprehensibility.

As for the overall score for the quality of the content production process, the average for the entire period analysed is 3.3 points out of 5 (equivalent to 6.6 out of 10), showing a gently upward trend but with slight ups and downs. Between 1996 and 2000, for example, the reports barely scraped over the pass mark. From the latter year onwards, they improved and there are only two rather low points thereafter, in 2012 and 2014. The highest peak was reached in 2016, coinciding with elpais.com´s leadership in the numbers of daily users of digital media (AIMC, 2016), (Figure 2).

Figure 2. Evolution of the process of drafting the text

Source: prepared by the authors

4.3. Social contribution

The quality of the third characteristic of the feature stories, which is their social contribution, has evolved slightly from a score of 6.4 in 1996 to a 7 in 2019, i.e., from a pass to a good. As shown in Table 3, the score obtained for one of the two parameters used to assess this characteristic dropped from an excellent to a good, but that of the other improved considerably, from a pass to a good, and thus the overall progression is positive.

Table 3. Evolution of social contribution by parameters (0/10)

Source: prepared by the authors

Regarding the overall score for the whole period, the average score for this aspect is 0.5 out of 1 –or 5 out of 10–, with a very irregular progression (Figure 3). There was only one relatively stable period in the parameter of fostering social debate, between the years of 2012 and 2017. This could be related to the coverage that elpais.com gave during those years to issues of great social interest such as the end of ETA at the end of 2011, the Syrian conflict of the same year, immigration –which reached its peak in Spain from 2010– refugees, the rise of social movements such as the Arab Spring, the fourth wave of feminism, European nationalism, the LGTB+ movement, or the economic and financial post-crisis, among others.

Figure 3. Evolution of social contribution

Source: prepared by the authors

4.4. Digitisation

The overall quality of the last aspect, which is the digitisation of feature stories, shows a considerable improvement, rising from a very low 0.7 in 1996 to 6.3 in 2019 –from a fail to a pass. Between 1996 and 2002 the use of hypertext was practically non-existent, probably because elpais.com was still in the early stages of the transition from traditional to digital journalism. The journalistic pieces it published at this time were the same as those of the printed version –where, obviously, there is no hypertextuality, although its presence was starting to be felt. Similarly, the capacity for user interaction between both journalists and the other users was non-existent between 1996 and 2011, though from 2012 onwards it has increased and remains stable (Table 4).

Table 4. Evolution of the quality of digitalisation by parameters (0/10)

Source: prepared by the authors

The progression of the quality of digitisation between 1996 and 2019, with an average score of just 0.15 points out of 0.5 –equivalent to a 3 out of 10– shows an uneven but rising trend. However, it does not reach a pass mark until 2012 and, although it obtained its highest score in 2016, it has fallen again more recently (Figure 4).

Figure 4. Evolution of digitalisation

Source: prepared by the authors

4.5. Overall evolution of journalistic quality

The sum of the scores achieved by the four characteristics which were analysed –based on some 20 parameters– allows us to describe the progression in quality of the elpais.com feature stories under study over the last 24 years. As can be seen in table 5, the overall quality has improved from a score of 5.2 in 1996 to 6.9 in 2019. 82% of the pieces analysed pass the quality test with an average score of 6.37, i.e., from a pass to almost a “B”. The differences are minimal between the first three characteristics, which in 2019 all obtained a “B” level score. However, the same is not true for the fourth, digitalisation, which, despite having undergone a much greater change over the same period, still fails to achieve a “B” equivalent.

Table 5. Evolution of overall quality by aspects (0/10)

Source: prepared by the authors

The line in figure 5 shows a general upward trend. The worst score occurs in 1999, when El País achieved its lowest percentages in audience penetration of the previous nine years (AIMC, 2000), while the other two low points, in 2009 and 2011, occurred in the midst of the economic crisis. In contrast, the highest scores came in 2013, coinciding with the consolidation of elpais.com as the leading digital general newspaper and the second biggest website in terms of daily readership (AIMC, 2013); and in 2016, after overcoming the economic crisis and becoming the platform with the greatest number of daily users of digital media in Spain (AIMC, 2016). However, it cannot be ruled out that the personal criteria of the feature stories’ authors themselves may have influenced the quality of the pieces.

Figure 5. Overall evolution

Source: prepared by the authors

5. Conclusions

The aim of this study was to analyse the progression of the journalistic quality of feature stories published on elpais.com, from the time of the creation of the digital version of El País in 1996 up until 2019, and to investigate whether this quality was affected by technological development. To this end, a sample of 72 interpretative feature stories was subjected to a quality test, from which a total of 20 parameters, relating to four main characteristics, were evaluated: the process of selecting the information, the process of preparing the content, the social contribution of the story and the digitisation of the text.

The quality of the information selection process has undergone an uneven but positive progression, from a score of 5.7 out of 10 in 1996, to 7.2 in 2019. The biggest change observed was in the parameters of citing the origin of the information and in the variety of the nature of the sources, both of which have improved significantly.

In terms of the content production process, the quality of the feature stories has also advanced positively, rising from 5.76 in 1996 to 7 in 2019. However, some parameters such as the analysis of data or the incorporation of additional information of practical use for the reader – which in fact deteriorates– continue to fail, while the use of supporting textual elements also demonstrates some shortcomings. Scores for other parameters, such as the degree of depth, the clear presentation of ideas or the degree of comprehensibility, have, although continuing to be good, slightly decreased. On the other hand, the inclusion of audio-visual resources which began at zero when elpais.com went online, has improved hand in hand with the possibilities offered by new technological developments.

The quality of the third characteristic, social contribution, improved slightly from 6.4 in 1996 to 7 in 2019. In its generally upwards but uneven path, a relatively stable period stands out in the promotion of social debate parameter. This period coincides with the coverage given by elpais.com during these years to issues of great social interest, such as the end of the terrorist group ETA at the end of 2011, the Syrian conflict of the same year, immigration –which reached its peak in Spain in 2010– refugees, the rise of social movements and the post-crisis economic and financial situation, among others.

The final characteristic, the digitisation of information, has undergone significant improvement, rising from a meagre 0.7 in 1996 to 6.3 in 2019, with an uneven but upward-trending progression. The parameters for the use of hypertext and capacity for interaction were practically non-existent when elpais.com began its transition from traditional to digital journalism, when the journalistic pieces it published were identical to those in the print version, but their significance has gradually increased –although more recently a downturn has been noted.

The sum of the scores obtained over the four characteristics previously discussed allows us to describe their overall progression as ascending, since the feature stories analysed have risen from a general score of 5.2 in 1996 to one of 6.9 in 2019. This, from the outset, would seem to disprove our initial hypothesis that since the advent of digital journalism, journalistic pieces, and in this case, feature articles, have suffered a loss of quality. However, although the scores for the first three characteristics are similar, with all of them achieving a “B”, the scores obtained in some of their parameters, such as the use of supporting textual elements –which only scrapes a pass; the analysis of data –which remains as a fail; or the presence of additional information of practical use –which fails with an even worse mark than in 1996, show that there are still shortcomings. Similarly, digitisation, in contrast to the three other analysed characteristics and despite experiencing great change since the incorporation of digital tools into journalistic texts, could also be improved, as it still fails to achieve a “B”. Therefore, the hypothesis is only partially confirmed.

However, it should not be forgotten that some of the low points detected coincide with moments of economic difficulty for the newspaper. This leads us to speculate that the loss of journalistic quality may not be related solely to the adaptation to new technology and the impositions of immediacy derived from the leap to digitalisation, but also to economic factors, and possibly to the personal criteria of the authors of the feature articles themselves. It would be interesting not only for future research to enlarge the sample size of feature articles analysed, but also to investigate a variety of non-Spanish digital newspapers in order to see if there are significant differences between them and the national sample, or whether, conversely, overall conclusions can be drawn for the Spanish press as a whole.