Schools in the media: framing national standardized testing in the Norwegian press, 2004–2018

ABSTRACT In the education sector, media outlets have been increasingly active in reporting on standardized testing. The purpose of this paper is to identify the most recurrent discursive frames used by the Norwegian regional and local press when informing their readers about national standardized testing, and to explore whether differences over time and across geographical localities exist in the pervasiveness of frames. Our analysis is guided by framing theory, and builds on a corpus of 3,046 articles that focus on national testing, published by 155 Norwegian regional and local newspapers between 2004 and 2018. The analysis identifies four different discursive frames within Norwegian press coverage, namely the frame of ‘performance’, ‘transparency and empowerment’, ‘misinterpretation and misuse’, and ‘criticism’. The four frames convey highly distinct causal and normative beliefs and realities about national standardized testing. While the dominance of the frames varies over time and across Norwegian counties, the frame of ‘performance’ is increasingly pervasive, something that potentially contributes to naturalize performative-oriented reporting and competition in education. The study highlights the importance of systematic media analyses to identify circulating principle beliefs on education, and of not limiting research to national newspapers in order to grasp geographical variation in media coverage.


Introduction
In recent decades, media outlets have been increasingly active in reporting on standardized testing. A growing body of research has examined how the media cover standardized tests, as well as its implications, focusing both on international large-scale assessments (e.g. Baroutsis & Lingard, 2017;Dixon et al., 2013;Elstad, 2012;Stack, 2006;Waldow, Takayama, & Sung, 2014) and on national standardized tests (e.g. Baroutsis, 2016;Mockler, 2013;2020;Shine, 2015;Shine & O'Donoghue, 2013;Yemini & Gordon, 2015). While these studies cover diverse educational settings, non-Anglophone contexts remain under-researched, in particular with regards to media coverage of national standardized testing. In addition, many studies have limited their analyses to a small sample of newspaper articles and/or a selection of national newspapers, while the regional and local press have been more rarely considered.
Nevertheless, in some countries, the local press is particularly important in providing local communities with information, which is viewed as more relevant and 'closer' to home (House of Lords, 2008). This is the case in Norway, which has a strong local newspaper tradition (Østbye, 2019). In 2017, regional and local newspapers in Norway made a total of 1.69 million distribution 1 (Norwegian Media Authority, 2018), i.e. reaching 42% of the population over the age of 20 (Statistics Norway, 2019). In a context such as the Norwegian one, with a very dispersed population across an extensive geographical area, an exclusive focus on national newspapers signifies the neglect of an important part of the whole media ecosystem. Moreover, the impact of regional and local press on readers may be greater because the presented content is closer to home.
The paucity of studies analyzing longitudinal samples of newspaper articles and the under-consideration of the regional/local press limit the understanding of how media coverage evolves over time and varies across geographical localities. To address these gaps, we draw on a corpus of 3,046 articles on national tests, published in Norwegian regional and local newspapers between 2004 and 2018. We aim to (a) identify the most recurrent discursive frames used by the media when informing readers about national standardized tests, as well as their main features, and (b) explore how pervasive certain frames are and describe how they evolve over time and are distributed across geographical localities. In so doing, we rely on framing theory (Entman, 1993), which forms a useful conceptual framework for understanding how one issue can be presented and framed in different ways, thereby communicating highly distinct causal and normative beliefs and realities (Chong & Druckman, 2007). It is important to gain a deeper understanding of how educational issues, such as standardized testing, are framed, since previous research has highlighted the impact that framing can have on public understanding and opinions (Chong & Druckman, 2007, as well as on public policy processes (Baroutsis & Lingard, 2017).

National testing in Norway
In Norway, the municipalities are responsible for the 10 years of compulsory education, following the division of school governance responsibilities between central, county and municipal authorities. Introduced in 2004, national standardized tests constitute a central steering device to monitor whether schools and municipalities achieve central learning objectives (Skedsmo, 2011). In addition, they are meant to provide useful information to teachers, school leaders, parents, students and municipal authorities, thereby fostering school improvement efforts. Alongside the adoption of the tests, the government website the 'School Portal' was launched in 2004, and aggregated test results are published here. While the tests are one element of the country's quality assessment system (Skedsmo, 2011), they have become the prime measure to hold schools and municipalities accountable for student learning (Skedsmo & Mausethagen, 2016).
Even though national tests had long been considered controversial in Norway (Tveit, 2014), the heated debate following the below-expected results obtained by Norwegian students in PISA in 2001 contributed to a relatively broad consensus among political parties and stakeholders about the need to introduce large-scale assessments in the country. Nonetheless, the speed with which the tests were implemented, the lack of consultation of key stakeholders regarding the tests' design and administration, and the decision that test scores would be published, were heavily criticized (Camphuijsen, Møller, & Skedsmo, 2020). The first two test rounds were met by student boycotts and destructive evaluation reports (Lie, Hopfenbeck, Ibsen, & Turmo, 2005), leading the new government coalition to pause the administration of the tests in 2006, in order to improve their quality. In 2007, the tests in reading, numeracy and English were reintroduced. Since then, the annual tests have been taken at the beginning of the fifth, eighth and ninth grade. All students are required to participate, but schools can exempt students with the right to Norwegian language instruction or special needs education. From 2007 until 2014, under a red-green government coalition, 2 only municipal, county and national test results were published on the 'School Portal', while school results remained hidden. 3 Since 2014, when a Conservative government coalition returned to office, school results have again been published.

Examining media coverage on standardized testing through frame analysis
Drawing on the seminal work of the sociologist Erving Goffman (1974), framing theory was adapted and developed by scholars such as Iyengar (1990;1991) and Entman (1993) as a tool to examine how the media organize and present information on particular issues. According to Entman (1993), to frame is: Frames thus promote a particular version of reality, and do so by relying on different framing devices, such as tone, placement, repetition, keywords or culturally familiar symbols or images (Entman, 1993), as well as catch phrases (Pan & Kosicki, 1993) and metaphors (Gamson & Modigliani, 1989). As important as the selection and salience of particular aspects of a described reality, can be the exclusion of other aspects (Entman, 1993). By examining the four framing 'functions', i.e. problem definition, causal interpretation, moral evaluation, and treatment recommendation (Entman, 1993), as well as the framing devices present in a text, a deep understanding can be gained of the 'interpretative lenses' through which media texts can guide readers to notice, understand, evaluate and act upon a particular problem in a particular way (Chong & Druckman, 2007;Entman, 1993;Scheufele & Tewksbury, 2007).
In recent decades, a range of studies have relied on framing analysis to examine how education, schools and educators are represented and framed in media texts (e.g. Baroutsis, 2016;Baroutsis & Lingard, 2017;Goldstein, 2010;Ulmer, 2014). In this way, it has been revealed how media reports can create and perpetuate a discourse of professional inadequacy (Ulmer, 2014), as well as visually and textually frame and reinforce particular reforms as the sole solution to the crisis in public education (Goldstein, 2010). Other studies have identified how different frames can be present in media coverage on one educational issue (Baroutsis, 2016;Baroutsis & Lingard, 2017). These studies form important contributions to the understanding of how the media frame educational issues. Non-Anglophone contexts, however, remain under-researched.

Data and methodology
Our analysis has been conducted on a corpus of articles published in Norwegian regional and local newspapers from 2004 (the year the tests were introduced) to 2018.
To obtain a complete list of all newspapers, we used Høst's media report (Høst, 2018) and data derived from the Norwegian Media Authority, the National Association of Local Newspapers, and the Norwegian Media Businesses' Association. We selected the newspapers according to their geographical distribution. Regional newspapers are those covering one or more counties, whereas local newspapers are those covering districts, cities or one or more municipalities. To examine whether the identified newspapers had reported on national testing during the 15 years being researched, we relied on the digital archive of each newspaper. This approach enabled us to obtain a more complete corpus than if we had used a media search engine such as Retriever, whose coverage of local newspapers is not comprehensive. We used the term 'national test(s)' to search all digital archives. A first scan of all retrieved articles was then performed to select only those articles with a clear main focus on national testing, i.e. (a) articles that reported news about national testing; (b) articles that expressed opinions/debates on national testing; (c) articles that reported national test scores. 4 Of all 202 regional and local newspapers, 155 (77%) had published articles focusing on national testing during the selected publication time frame. In total, 3,046 articles were retrieved after this first scan.
Informed by framing theory, we operationalized each of the four framing functions (Entman, 1993), namely (1) problem definition, (2) causal interpretation, (3) moral evaluation, (4) treatment recommendation. Based on this operationalization, we then coded all the articles using a codebook, which was the product of a dynamic, nonlinear approach. By coding for framing functions, rather than frames, one is able to identify distinct frames in a more valid and reliable manner (Matthes & Kohring, 2008). In addition to the four framing functions, the continuous iterations with the analyzed texts led us to inductively generate and add new categories to the initial ones (e.g. codes were used to identify score reporting, ranking, naming, praising, blaming, etc.). Parallel to the coding process, a longitudinal dataset was created, which included information on every article of the corpus, i.e. article identifier, title, name of the newspaper where it was published, geographical distribution and date of publication. In this database, all information about the codes used in every single article was inserted. Once we had coded all the articles, the longitudinal dataset enabled us to quantify the recurrence and coexistence of the codes. By means of different combinations of codes, we were able to identify different frames. We then relied on qualitative content analysis (Macnamara, 2005) to gain a deeper understanding of the frames identified. Finally, we analyzed the dataset to disclose trends across place and time.

Framing national standardized testing in the Norwegian press
Figure 1 provides a longitudinal overview of the analyzed corpus and shows the evolution of press coverage of national testing over time. The figure shows a peak in press coverage in 2005, one year after the tests were introduced, and a sharp decline in 2006, when the administration of the tests was paused. Ever since the reintroduction of the tests in 2007, press coverage has remained relatively stable.
Four recurrent frames could be clearly identified in our analysis. These were present in 2,475 newspaper articles (81.25% of the corpus). The remaining articles were articles where other, non-recurrent frames were present (334 in total), or articles that merely reported general information about the tests, such as practical information about the tests' administration (237 in total). In the presentation of the findings that follows, we focus on these four most recurrent frames.

Four most recurrent frames: evolution over time and geographical differences
We named the four dominant frames 'performance', 'transparency and empowerment', 'misinterpretation and misuse' and 'criticism'. Table 1 outlines each frame according to the four framing functions (Entman, 1993).
Often (in 113 cases), the same newspaper contains articles that offer different, sometimes competing, frames. That is, even newspapers that regularly report results and rankings, publish articles that harshly criticize this practice, as well as articles that present strong critiques of the validity and reliability of the test results. Nonetheless, the four frames identified were not equally present in our corpus, and their dominance varied over time and across geographical localities. Figure 2 displays the evolution of the four frames over the analyzed 15-year time span.
As depicted in the figure, the frames of 'criticism' and 'misinterpretation and misuse' were particularly pervasive during the years of the first two test rounds. Their dominance, however, declined upon the reintroduction of the tests in 2007. Ever since, the frame of 'performance' has become the most dominant frame in the articles analyzed, followed by the frame of 'transparency and empowerment'.
In addition to differences in time, the analysis identified differences in how pervasive each frame is across geographical localities. Figure 3 shows the recurrence of each of the four frames per county. 5  National test performance reflects the effort put by key actors in the acquisition of basic skills. High-performers found the right formula to acquire basic skills and succeed in the test.
National test scores are valid and relevant to measure, compare and judge the acquisition of basic skills. As test performance reflects effort, actors can be praised or blamed for their performance. Identification of recipes for success that low-performers can/should learn from high-performers. Those responsible for poor performance should urgently address this problem and take measures to assure the acquisition of basic skills.

Frame of transparency and empowerment
There are large inequalities in student learning (between students, schools, and localities). It is crucial to know more about these inequalities to address them.
National tests and transparency in terms of learning outcomes guarantee that key actors (policymakers, stakeholders and society as a whole) become aware of these inequalities, and are empowered to address them.
National tests are valid, reliable and useful instruments to uncover students' learning difficulties and adapt education to individual and local needs. Transparency in terms of learning outcomes is both a democratic right, and a prerequisite for school improvement.
Safeguarding of the national testing system. Guarantee of transparency in terms of learning outcomes. As can be derived from the figure, geographical differences can be observed with regard to the absolute number of articles published on national testing in each county, as well as with regard to the dominance of each of the four frames. In 15 counties, the frame of 'performance' formed the most dominant frame, while in three other counties (Hedmark, Nord-Trøndelag, and Oslo), another frame was found slightly more often. Hordaland forms the sole example of a county where two frames, namely 'performance' and 'criticism', were found to be present in the same number of articles. Even though in Oslo the frame of 'criticism' was slightly more frequent, the county represents a unique case of a near balance in the dominance of all four frames.
In the next subsections, we present each discursive frame in more depth, using salient quotations to epitomize and explain the frames. 6 The frame of 'performance' The frame of 'performance' was the most frequently found frame in the Norwegian regional and local press in the time span analyzed (present in 1,190 articles). As shown in Table 1, the frame of 'performance' communicates the idea that national tests measure the acquisition of basic skills that are crucial for individual success in life and for societal progress. Consequently, low test performance should 'ring the alarm bells' (Art-0962 2017). The frame drives the message that test performance, to an important extent, reflects the effort put in by key actors to secure acquisition of basic skills (most notably students, educators and municipal authorities), and that high-performers have 'found the formula' to acquire basic skills (Art-0580 2018, november 16). This message reflects an implicit evaluation of national test scores as valid indicators to measure, compare and judge students' acquisition of basic skills. Furthermore, as performance in the national tests corresponds to the effort put in by key actors, those who obtain low scores can be criticized or even blamed. Finally, as a treatment recommendation, the frame conveys the message that those responsible for the low results urgently have to take measures to raise them and that they can learn from high-performers, who have found measures that 'work' (Art-2849 2015). Below, we provide examples of how the frame of 'performance' disseminates and reinforces the above-outlined messages.
In total, 1,117 articles (93.87% of the 1,190 presenting this frame) use test scores to make any form of comparison between the performance of different entities (e.g. school to school, school to national average, municipality to county, etc.), or over time. Even though school results were not published on the 'School Portal' between 2007 and 2014, this did not hamper the construction of school comparisons and rankings in the regional and local press during this period. 7 Test scores are occasionally presented in ways that encourage the reader to think of the national testing system as a competition of some sort. For example, in 167 articles the frame of 'performance' is constructed around the conceptual metaphor (cf. Lakoff & Johnson, 2003) 'national test performance is a competitive sport'. That is, terms and language taken from sports journalism are used to report on school test results. Examples include: 'Students break their own personal record' (Art-1042 2013) and 'Municipality X is beaten at the finish line by Municipality X and Municipality X' (Art-2310 2008). As argued by Schmitt (2005) metaphors make 'perception more automatic and ease the energy required to understand' (p. 336). However, the use of this conceptual metaphor not only enhances the power of the message and catches the reader's attention, but also conveys a range of implicit messages. For instance, sport is quintessentially meritocratic and sport competition is healthy and aimed at self-overcoming. Accordingly, the best performers can/should be praised. In this regard, in the frame of 'performance', a judgment is often assigned to the comparative position of one or more entities, blaming those that score below the desired level, while praising those that perform beyond expectations. Of the 1,172 articles within this frame that name schools, municipalities and/or counties (i.e. 98.49%), 235 (20.05%) articles include blame and 570 (48.63%) articles include praise. Examples of blaming include statements such as 'School losers in Municipality X and Municipality X' (Art-2941 2008) and 'Worst in the country at 2 + 2' (Art-0484 2011). In contrast, examples of praising include statements such as 'School A tops the list of Norway's best schools' (Art-1236 2018, october 24) and 'No one is better at counting than this group' (Art-1235(Art- 2017. This result confirms the observation by Elstad (2009) that Norwegian newspaper articles often focus on the worst and/or best-performing schools. It also seems to correspond to the tendency towards sensationalization, which previous research has pointed out as a fundamental feature of media reporting (Baroutsis, 2017). However, in contrast to what has been reported by previous research on a negative press bias, i.e. a bias towards reporting negative facts (Dixon et al., 2013;Elstad, 2009;2012), we found praising a more common practice than blaming.

The frame of 'transparency and empowerment'
The frame of 'transparency and empowerment' was the second most frequently found frame, present in 495 articles. The frame problematizes that there exist significant inequalities in student learning in Norwegian education. Before the adoption of the testing system, as well as during periods that test scores were not published at the school level, these inequalities have largely gone unnoticed, and have therefore remained unaddressed. 8 National tests provide knowledge that is presented as essential to uncover these inequalities and empower key actors (e.g. policymakers, educators and society as a whole) to address them, thereby making 'winners of those who might otherwise become losers' (Art-1110(Art- 2004.
National tests are thus evaluated as valid, reliable and useful instruments to uncover students' learning difficulties, adapt education to individual and local needs, and hold teachers and other actors accountable for low scores. What follows from this positive evaluation of testing and transparency to address social inequalities, is the recommendation to safeguard the testing system and to guarantee transparency in terms of learning outcomes. In turn, arguments against and efforts to discontinue testing and/or the publication of results are harshly criticized: It is not the tests that create the inequalitiesbut it is the tests that reveal them, thus making it possible to direct efforts to help the most struggling schools and students (…) The differences do not disappear even if you hide them away. They only become invisible and continue to destroy the future of new generations. (Art-2652(Art- 2005 One important assumption behind the requirement for transparency is that it can serve as an important pressure mechanism for schools and municipalities to reflect on their work, make changes and improve results. Hiding the results, in contrast, would imply that quality improvement measures 'lie untouched in archives, drawers and folders in public offices and with teachers' (Art-1589 2009). An important second reason behind the promotion of transparency is that it is a democratic right and it facilitates public engagement and deliberation, allowing school administrators, parents, local politicians, journalists and the general public to follow and engage with the efforts of municipalities, schools and teachers to ensure education quality and equity. In 114 of the 495 articles within this frame (i.e. 23.03%), the publication of results is promoted as a key condition to assure transparency and for the tests to be of optimal value.

The frame of 'misinterpretation and misuse'
The frame of 'misinterpretation and misuse', present in 400 articles, denounces the detrimental effects of the wrong interpretation and usage of test results, while the importance and potential usefulness of national testing itself remains unquestioned. For example, concern is expressed regarding the risk of interpreting test results as the prime indicator of education quality, while the national tests measure only a set of learning goals.
Particular blame is directed towards politicians and authorities, who are accused of making statements and drawing conclusions that are not statistically sound, thus misinterpreting and misusing the information derived from the tests, sometimes causing 'the warning lights to turn on for no reason' (Art-1358 2016). A second important factor contributing in particular to the misuse of test scores is the publication of results. This is said to generate competition and enhance a situation in which test results become a goal in itself, rather than a pedagogical tool. Moreover, this competition is said to elicit the adoption of harmful practices aimed at raising test scores at any cost, such as teaching to the test, or abuse of the exemption rule.
The media themselves are blamed for contributing to this competition by awarding test winners and losers, oversimplifying educational debates, stigmatizing schools using 'brutal' and 'distasteful' headings (e.g. Art-23402016) and caring more about 'newspapers' sales figures' than the impact of their reporting (Art-2581(Art- 2007. Moreover, journalists are accused of misleading usage of statistics, which leads to distorted descriptions of the real situation in the educational system. As argued in one article, 'Sometimes a brilliant idea is ruined by poor execution. This may have happened to the national tests in schoolsnot least by the way that the media have misused this information' (Art-2433(Art- 2005.
As such, this frame conveys the message that the tests themselves are useful tools, while serious problems emerge when test results are misinterpreted, not understood or badly used. The frame thus suggests that the publication of results must be abandoned and that correct interpretation and use of national tests results should be promoted: If you stop using national tests to rank schools, you will automatically end different practices around exemption. We believe that national tests can have a value in relation to the individual student's learning. Then it is important that students are not taken out for better results. (Art-2034(Art- 2014 The frame of 'criticism' The frame of 'criticism', found in 390 articles, conveys a skeptical message towards the usefulness of the national tests, and is highly critical of the detrimental consequences of the test system. Demanded and developed by politicians, rather than school professionals, the national tests are described as adding little to an already 'good selection of mapping tests and other tests that give better insight' (Art-0851 2017). Even more so, the tests are framed as providing 'a very thin foundation for commenting on learning outcomes' (Art-0517 2010). In addition to not living up to the promise of serving as a useful pedagogical and organizational tool, the test system is described as disruptive and as having a negative impact on education quality and equity.
That is, the existence of the national testing system causes a number of negative sideeffects. For example, it is argued that the excessive focus on the few measurable educational goals runs the risk of constraining autonomy, diversity and innovation. Similarly, the test system is said to damage students, by reducing motivation, and enhancing conformity of thought, pressure and stigmatization. At the same time, the pressure felt by teachers and principals to obtain high scores is said to lead to the adoption of a number of undesirable and harmful practices. Curriculum narrowing, teaching to the test, excessive test preparation, educational triage, and exclusion of low-performing students are particularly denounced: It has already been revealed that schools drill students in tests similar to the national tests to secure the best possible results. (Art-0672 2005) A schoolboy was ill on the day of national tests. The principal called home, asked how sick the boy was, brought the boy in his own car to school and then drove him home again after the end of the test. Hi and ho and up the average! (Art-2097(Art- 2015 As such, this frame conveys the message that the national tests are not useful, not valid and not reliable. At the same time, the test system hampers good education by causing the adoption of harmful practices and introducing elements into the system with detrimental consequences for education quality and equity. It is therefore recommended that the national testing system is abandoned and that assessment measures are promoted that are considered to be more suitable by school professionals.
The analysis revealed differences in the pervasiveness of each frame over time. During the first two test rounds, which were met by severe societal discontent and destructive evaluation reports (Lie et al., 2005), the three frames that express explicit opinions on the test and its use were most pervasive. Upon reintroduction of the tests in 2007, the dominance of these frames declined, which might reflect the more favorable public opinion towards the quality of the tests, and their conditions of administration, thereby giving rise to less debate in the media. At the same time, the frame of 'performance' grew in importance from 2007 onwards and ever since has been the most dominant frame found in Norwegian regional and local press coverage. In contrast to the other three frames, within the frame of 'performance', the existence and validity of the national tests is neither questioned nor debated. Rather, performative-oriented ranking and reporting are common practice, with the potential effect of naturalizing and enhancing competition in education.
While the reasons behind the increasing pervasiveness of the frame of 'performance' are not clear, possible explanations can be advanced that could feed future research. The predominance of this frame may reflect an increasingly widespread acceptance of national tests as valid and useful measures of crucial learning outcomes. The dominance of the frame of performance could also reflect the seductions of quantification (Espeland & Stevens, 2008). That is, indicators such as test scores appear to provide objective, scientific, concrete and politically neutral knowledge of a complex world, thereby appealing 'to the desire for simple, accessible knowledge and to a basic human tendency to see the world in terms of hierarchies of reputation and status' (Merry, 2016, p. 1). This would correspond to a fetish for numbers, measurements and comparisons, as has already been theorized by several educational researchers (e.g. Ball, 2015;Ozga, 2008). Simultaneously, comparisons between entities based on averages, as well as over time, could correspond to the easiest, cheapest or most sensational way for non-technical stakeholders such as journalists, who may not have the expertise to fully understand the complexity of the results (Wu, 2010), to make sense of and report on national tests.
Further research could also look at the impact of different forms of media coverage. To date, most studies have focused on the stress responses of educators in blamed schools (e.g. Elstad, 2009), while less is known about how other media practices, such as praising, rankings and comparisons, which we found to be more common, affect different schools, including those ignored by the media. Such research might be particularly relevant in contexts where schools face increasing competitive pressures and test scores affect the school's reputation. For example, whereas praising may seem like a neutral or harmless media practice, in small municipalities where there are only a few schools, praising and exalting the performance of one or few schools means de facto neglecting and subtly, implicitly blaming the few other local schools. This could therefore correspond to a silent form of 'public shaming and blaming' (Takayama, 2008, p. 388). For neglected schools, the consequences of praising practices might be as harmful as those reported by Elstad (2009) for schools with negative media coverage.
In addition to differences over time, our analysis shows the existence of geographical differences in the pervasiveness of each of the four frames. Whereas in some counties, such as Oslo and Hordaland, where Norway's biggest cities are located, all four frames are almost equally present, in other counties, such as the more rural Finnmark and Nordland, the frame of 'performance' strongly dominates. Although descriptive, these findings point to the need for more scholarly attention for geographical differences in media coverage. More specifically, they confirm the importance of not limiting research to national media coverage, which could erroneously lead to the impression of uniformity throughout the country. This might be particularly important considering that previous research has highlighted that media coverage can have an important impact on school actors' experiences and actions (e.g. Elstad, 2009;Stack & Boler, 2007). As such, different media coverage within the same country may have far-reaching effects on school practices and responses to test-based accountability reforms that would otherwise remain difficult to explain. These findings also open up a new line of inquiry, which might want to look at the reasons behind the geographical differences found. A number of factors, such as for example, different local media landscapes, local policy contexts, demographic factors, cultural specificities or different levels of performance in the national tests could be related to these differences and determine them. Geographical differences in media coverage thus deserve a critical spotlight. Notes 1. Of these distributions, 1.4 million included paid-for subscriptions, while nearly 300,000 distributions comprised free newspapers. For seven free newspapers, no data can be found on the number of distributions (Norwegian Media Authority, 2018). 2. The coalition was made up of the Labor Party, Socialist Left Party and the Centre Party. The appellative of 'red-green' refers to the party colors. 3. Even though school results remained hidden on the 'School Portal', anyone interested could request access to these data, following the Freedom of Information Act. 4. We included all types of contributions, so the corpus includes articles written by journalists, editorials and opinion and commentary articles. 5. As of January 2020, the former 19 counties were merged into 11 counties. However, as the former county distribution was in place during the analyzed time span, and the official data on the geographical distribution of newspapers relies on this county distribution, we present the data for the 19 former counties. 6. Considering the size of the corpus, it was not possible to represent all of the analyzed articles.
We therefore selected quotations that represent typical examples of commonly expressed statements. All quotations have been translated from Norwegian to English by the authors. We refer to each newspaper article by its article identifier. 7. Our analysis shows that, even though school results were not published on the 'School Portal', during this period, 619 articles of our corpus were presenting school comparisons and/or rankings. It was found that in 18 cases, school rankings were based on results presented in national newspapers, in 9 cases school results were provided by county/municipal authorities, and in 3 cases local newspapers requested school data following the Freedom of Information Act. In all other cases, the source of the school data was not provided. 8. In making this claim, no mention is made of the fact that, as highlighted in footnote 6, some media and municipalities reported school results during the years that they were not published on the 'School Portal'.