THE NEWS SHARING GAP: how political context helps to explain news routines and shareworthiness on SNSs

COMPARTILHANDO MUNDOS DISTINTOS: como o contexto político ajuda a explicar as rotinas produtivas e o compartilhamento de notícias nas mídias sociais online

COMPARTIR MUNDOS DIFERENTES: cómo el contexto político ayuda a explicar rutinas productivas y compartir en sitios de redes sociales

ABSTRACT

Current literature on online news sharing presents a range of methods and results, alongside contradictory explanations for the phenomenon. To disentangle this, we compared three countries with contrasting levels of political stability (Brazil, Canada and the US). A content analysis of articles (n = 1.658) posted in 2016, on the main news pages in Facebook for each country, shows that the Canadian news pages presented far less news sharing on national politics conflicts than Brazil and the US did. We discuss how this shows the relevance of political context in explaining both news routines and shareworthiness.

Key words
Online news sharing; News routines; Facebook; Soft news; Hard news

RESUMO

A literatura sobre compartilhamento de notícias online apresenta uma gama de métodos e resultados com linhas contraditórias de explicação para o fenômeno. Para destrinchar o problema, comparamos três países com níveis contrastantes de estabilidade política (Brasil, Canadá e EUA). Uma análise de conteúdo de notícias (n = 1.658) postadas nas principais páginas noticiosas do Facebook de cada país em 2016 mostra que as páginas canadenses apresentaram muito menos compartilhamento sobre conflitos políticos nacionais do que nos EUA e Brasil. Discutimos como isso mostra a relevância do contexto político para explicar as rotinas produtivas e a compartilhabilidade.

Palavras-chave
Compartilhamento de notícias online; Rotinas produtivas; Facebook; Notícias leves; Notícias sérias

RESUMEN

La literatura actual sobre el intercambio de noticias en línea presenta una variedad de métodos y resultados junto con líneas de explicación contradictorias para el fenómeno. Para desenredar esto, comparamos tres países con niveles contrastantes de estabilidad política (Brasil, Canadá y EE. UU.). Un análisis de contenido de los artículos (n = 1.658) publicados en las principales páginas de noticias en Facebook de cada país en 2016 muestra que las páginas de noticias canadienses presentaron mucho menos intercambio de noticias sobre conflictos políticos nacionales. Lo contrario ocurrió en Brasil y Estados Unidos. Discutimos cómo esto muestra la relevancia del contexto político para explicar tanto las rutinas informativas como las compartidas.

Palabras clave
Intercambio de noticias en línea; Rutinas de noticias; Facebook; Soft news; Noticias difíciles

1 Introduction

Current literature on news sharing on social networking sites (SNSs) presents a range of studies, methods, and theories. An influential strain of this literature explains online news sharing as a function of how platform affordances make each SNS unique in its dynamics and characteristics (Bastos, 2015; Christin, 2015; Kalsnes & Larsson, 2018; Karnowski et al., 2020; Trilling et al., 2017). Alternatively or additionally, news sharing is also explained by journalism studies in terms of general human selection criteria or similar concepts from evolutionary psychology and cognitive accessibility theories (Trilling et al., 2017).

Despite the comprehensive explanatory frameworks from both of these scholarly strains, one finds conflicting results regarding what kind of content is found as shareworthy, even for a single SNS like Facebook. Thus, whereas some studies will, for example, point out that hard news and public affairs articles will be highly shared on this platform, others will suggest exactly the opposite.

To solve this kind of conundrum, we argue that, beyond ascribing fixed and permanent relations among platform characteristics, content, and news users’ motivations to understand online news sharing, it is imperative to investigate how this phenomenon is also shaped by the dynamics of the sociopolitical context. In this respect, some studies (e.g. Boczkowski & Mitchelstein, 2012; Lycarião & Leite, 2020; Ørmen, 2019) provide evidence that this is a key factor to understanding the dynamics of online news sharing. In that light, this paper investigates if (and how) political context influences online news routines and shareworthiness.

To examine this, we selected three American countries (Brazil, Canada, and the US) with contrasting levels of political activity and stability. In order to assess the extent to which the distinct political contexts of our sampled countries can explain their online news sharing differences, we conducted a content analysis of all news pieces posted on the two most-followed news pages in Facebook of each country, from 24 to 30 August 2016 (n = 1.658). We selected these countries, precisely in this period and due to their respective political situations, by checking whether they were experiencing: (a) national elections; (b) political scandals involving high authorities of the Executive branch; and, (c) political or judicial prosecutions against heads of government.

Following these criteria, we noticed that some political events in 2016 led Brazil and the US far from the calm waters of the political situation in Canada. Whereas the latter was experiencing a non-election year and facing no significant political scandal, the US was experiencing one the most polarized and hostile elections in its history, including political scandals involving both candidates (Faris et al., 2017; Patterson, 2016). Brazil, in turn, was struck by a huge corruption scandal that paved the way for the impeachment of President Dilma Rousseff. Her trial was concluded by the Senate on 31 August 2016.

Content analysis of news posts and statistical tests of news sharing outputs provided evidence for our research hypothesis that both online news routines and news sharing follow the contrasting levels of political stability and activity of the surveyed countries. In this respect, the Canadian sample (the most stable/regular country in our study) showed far less news supply and news sharing on controversial topics involving national politics and political scandals.

Overall, our main contribution to current scholarship consists of setting a unified theoretical and methodological framework that takes both domestic (political context) and formal (news page characteristics, e.g. visual cues) factors into account to understand news sharing on SNSs. Another contribution is showing how the levels of political activity and stability can predict online news routines and sharing patterns. This allows us to solve some empirical conundrums in the current literature, which tends to explain shareworthiness either as a function of platform characteristics or as a function of users’ motivations. In this paper, we argue that all these factors are so intertwined that only by taking political context and its coupling with news routines into account that it becomes possible to see the aforementioned conflicting results as pixels of a coherent picture.

2 News sharing and politics in Facebook

In the present study, we are particularly interested in the dynamics of news sharing in SNSs, which raises issues concerning other forms of news retransmission, especially that which is conducted in offline environments. More specifically, we focus our investigation on sharing news, derived from traditional news media, in SNSs. This remains an important subject for scholarly investigation, as a both necessary and influential factor for inclusive and enlightened public debate (Newman et al., 2019).

Yet, what happens on SNSs with the kind of content that is crucial for a well-informed public (i.e. hard news and opinion-oriented articles on public affairs)? Is it usually downgraded or upgraded by news users in SNSs?

The available studies present different and even conflicting answers, depending on at least four clusters of (interacting) factors: (a) the platform characteristics, (b) the sharing value of news content in SNSs, its shareworthiness; (c) the underlying motivations behind news sharing; and (d) how political context might shape and change these same motivations. Below, we will present an outline of each of these clusters to build a unified theoretical framework that will structure our empirical study.

(a) Platformization of journalism

To begin with, platforms of SNSs like Facebook and Twitter are not neutral and are strongly shaped by their algorithms and affordances (dos Santos et al., 2019; Lischka & Werning, 2017; Lischka, 2018). This has implications for how news editors manage news production and routines in terms of publication pace, media formats, and even what kind of content is perceived as more shareworthy on each SNS.

Some German studies indicate that, specifically on Facebook, its affordances and news feed algorithm favor the spread of soft and sensational news over that of hard news (Lischka & Werning, 2017; Lischka, 2018). The evidence offered by these studies suggests a general trend of news editors being pushed to offer a more softened supply of their news production to Facebook audiences. However, Steiner’s (2020) study of four German news media outlets challenges this by showing that there is no unified trend when one compares legacy with tabloid media. This is because the former is proving to be more reluctant to pursue news softening than the latter.

Nonetheless, there are other aspects, in addition to content, that indicate the influence of platform affordances and their algorithms on news sharing. One of these aspects is the differentiated capacity of some news media outlets to obtain better performance than others (dos Santos et al., 2019; Salgado & Bobba, 2019). Mainly, this is influenced by the number of followers, how well the new page holds a balanced pattern of publication, promotes video formats, and posts regularly. Yet, dos Santos et al. (2019) warn “each variable might be different after each platform’s algorithm tweak” (p. 416). This means that, if there is no stability of news routine effects and platform affordances, even in the same SNS (which changes its algorithm over time), this is even more acute across different platforms.

(b) What content is (not) worth sharing on Facebook?

Platform implications on how news sharing is shaped might also be noticed in the study by Trilling et al. (2017). They found in the Netherlands context, that news sharing on Twitter and Facebook “is correlated, but not identical” (p. 51). This is reflected in how their concept of “shareworthiness” turns out to be somewhat distinct on each platform. To the point that, whereas a human-interest angle in news articles has no significant effect on news sharing on Twitter, it does so on Facebook (pp. 52–53). Bastos (2015), Christin (2015), Kalsnes and Larsson (2018), Karnowski et al. (2020) found similar results by analyzing data from different countries.

On the other hand, García-Perdomo et al. (2018) found no differences between Facebook and Twitter in their cross-national analysis of news sharing in the United States, Brazil, and Argentina. Valenzuela et al. (2017) showed more varied results. In their study of six Chilean outlets on the same platforms (Twitter and Facebook), they found that “the human-interest frame is not a significant predictor of news sharing” (p. 817).

In the end, what kind of news content is more shareworthy on Facebook? Do hard news and public affairs (PA) articles have any shareworthiness, there?

As suggested above, there is little room for uncontested answers to this question because the available evidence is rather heterogeneous and even conflicting. Some studies will identify Facebook as infertile soil for hard news and PA articles (Bright, 2016; Karnowski et al., 2020; Trilling et al., 2017), others will suggest alternative accounts (Lycarião & Leite, 2020; Martin, 2019; Ørmen, 2019).

Bright’s (2016) analysis of news interactions on the SNSs of the BBC News showed that politics, government, economics, and social issues had a negative impact on news sharing on Facebook (p. 353). The author suggests the existence of a social news gap “between what people like to read and what they share” (Bright, 2016, p. 358). This news gap implies that the traditional news agenda and the social media agendas diverge in their treatment of what news media content should gain attention and importance. As a result, hard news, opinion-oriented articles on PA would not be as shareworthy as soft and sensational news.

Nonetheless, Martin (2019) came to what seems to be rather contrasting results. Ultimately, her study of 160 English language news sites and their news sharing performance on Facebook and Twitter concludes the following:

Certainly, there is, overall, a greater tendency to soft stories in everyday news sharing, but this type of sharing activity has a lesser engagement and reach than the top quartile sharing, so is not as publicly visible. Further, when people are doing ordinary low-level commendation, they are more likely to exchange PA-oriented news features and opinion content, suggesting they are regularly engaged in conversations about socially significant issues.

(Martin, 2019, p. 150).

Thus, current literature presents mixed results and accounts about how shareworthy hard news and public affairs articles are on SNSs, including Facebook. The same occurs for online news featuring conflict or controversy. Whereas some studies identified a positive effect on news sharing (García-Perdomo et al., 2018; Trilling et al., 2017), others identified a null or even a negative effect (Bright, 2016; Lycarião & Leite, 2020; Valenzuela et al., 2017).

(c) Underlying motivations

As shown above, some studies observed a negative association between news sharing and news articles featuring controversy or conflict. These results are regularly explained by SNSs users’ motivations toward impression management. This is because when one shares news in an open environment, it offers a high risk of reputational damage. After all, SNSs like Facebook allow for many acquaintances and even unknowns to become aware of one’s opinions, likes, and dislikes. Consequently, sharing news on SNSs also carries concerns about preserving one’s social status and reputation. In this regard, as conflict is present in news about crime and disasters, Bright (2016) points out that this hinders news sharing because it might imply a morbid interest in such kinds of events.

Nonetheless, we also have seen that other studies found a positive effect of conflict or controversy on news sharing, including on Facebook. For these results, there are other sorts of users’ motivations at stake. In this regard, Kim and Ihm (2020) found, through a survey with 400 Korean citizens, that they prefer to share highly controversial news on open and asymmetrical social media (OAS) like Facebook and Twitter, rather than on closed and symmetrical social media (CSS) like WhatsApp and (Facebook) Messenger. The authors explain these results by stressing news users’ interest in getting larger attention and news diffusion through the OAS.

Thus, just as with the case for hard news, articles containing conflict or controversy present mixed results not only in terms of how shareworthy they are on SNSs but also in terms of what the actual motivations are behind their news sharing. Studies on uses and gratification theory reinforce this conundrum since the corresponding results reveal a variety of motivations that trigger news sharing, but which “motivation is strongest, however, is not clear” (Valenzuela et al., 2017, p. 807).

Kim et al. (2020) provided an important clue to solving this conundrum. In their study with 406 US and 400 Korean citizens, they found evidence that, despite important cultural differences between the two countries, both samples showed that motivation for sharing a certain online news article is strongly influenced by the user’s perception of how important (newsworthy) the topic or event featured in the article is. Another factor that triggers news sharing is the user’s prediction of how similarly to them their most esteemed people will ascribe importance to that same topic or event. Thus, although determining what kind of content is regarded as important and, as a result, shareworthy might be difficult due to variation across cultures and contexts, it is possible to point out that importance attribution is key to understanding news sharing across different contexts and cultures.

Ihm and Kim (2018) also presented evidence in this regard. They conducted surveys with 400 Korean citizens and confirmed the hypothesis that news users “are more likely to share news online when they perceive their audience to have a higher receptivity to news content they are offering” (Ihm & Kim, 2018, p. 4350).

(d) Context also matters

As shown above, studies on the motivations that trigger online news sharing suggest that the relationship between news users and their audience is key to understanding this communication phenomenon. In this regard, other studies also show how this relationship can be affected by the socio-political context.

Boczkowski and Mitchelstein (2012) found that, during times of heightened political activity (national election campaigns), US news websites presented a higher level of users engagement (clicking, emailing, and commenting) on public affairs than during times of routine political activity. To explain this difference, the authors suggest that online news consumers scan their environment when they engage with news content. If the political environment changes towards a more heightened situation, this affects news users’ perceptions of their audience receptivity about hard news and public affairs articles.

Ørmen (2019) reinforces this line of reasoning with his study of online news engagement during the 2015 Danish parliamentary elections. Lycarião and Leite (2020) also point to the Brazilian political situation to explain why hard news was more shared than soft and sensational news on Facebook. According to them, the political turmoil of Dilma Rousseff’s impeachment during the time of the study (in 2016) altered the default behavior of SNS users to avoid hard news and public affairs articles.

3 For a unified framework

Overall, our literature review reinforces the assumption that online news sharing “logics are not just platform-specific, but rather platform-and-region-specific because the audience structures of SNSs vary across regions and nations” (Karnowski et al., 2020, p. 18). Moreover, current literature also provides evidence that news routines and shareworthiness vary not only across countries but also across different political contexts.

However, this variation does not derive solely from news users’ behavior but is a communication phenomenon that embraces society as a whole. This way of approaching news dynamics is not new. Eilders (2006) summarized this approach in the well-established conceptualization that “journalists and recipients process world events in very similar ways” (p. 19). This view gained a more robust empirical confirmation by Neuman et al. (2014), precisely regarding how news media and SNSs users in the US follow very similar patterns in turning their attention to politics and social issues. Thus, we argue that both media professionals and news users produce mutual expectations according to their social-political environment. When this environment turns out to be more unstable and politically heightened, they both (media professionals and news users) turn their attention to politics and the perceived sources of the aforementioned instability and activity.

Therefore, beyond ascribing fixed and permanent relations among audience profiles, content, and motivations to understand news sharing, it is crucial to investigate how mutual expectations of SNS interactions are also shaped by the dynamics of distinct levels of political activity and stability. After all, in addition to the aforementioned factors (news user profiles, motivations, etc.), there is plausible theoretical reasoning, and even some evidence to indicate that the level of political activity and the resulting political context also affect online news routines and news sharing patterns. Taking this into account, the present study was developed based on the following research question:

• Research Question (RQ): Does (and how does) political context influence online news routines and shareworthiness patterns of countries with varying levels of political stability?

• To set the conditions for our investigation, we proposed to test the research hypothesis based on our literature review and theoretical framework:

• Research Hypothesis (RH): The more unstable and politically active the country, the more its news media will post news involving the inner political controversies that trigger its instability and, similarly, users will share such news.

Below we describe how we operationalized an empirical inquiry to test this hypothesis. Afterward, we formulate specific hypotheses that translate them into expected empirical observations in terms of the variables and sampling described in our research method.

4 Method

Aiming to test our research hypothesis, we collected all news pieces posted on the two most followed news pages in Facebook of Brazil, Canada, and the US, from 24 to 30 August 2016 (n = 1.658). This was undertaken between October and November 2016 by querying the Facebook Graph Application Programming Interface. Our data collection was conducted by respecting a minimum of one month after the posting of each news article. Our data was based on Facebook because this platform remains “the most important social network for news” (Newman et al., 2019, p. 10).

We selected the news pages based on their number of followers and news quality orientation. We operationalized this by identifying at least two posts on public affairs or hard news in their last 20 publications. The news pages selected according to these criteria were: Veja and G1, for Brazil; CBC News and Global News, for Canada; and, Fox News and The New York Times, for the US.

We selected these countries precisely in this period because they presented different levels of political activity and stability in 2016, and were the most populous countries in which a predominant language could be coded by our team. For contrasting political activity levels, we examined the political situation of each country by checking whether it was experiencing: (a) national elections; (b) political scandals involving high authorities of the Executive branch; and (c) political or judicial prosecutions against the head of government.

Following these criteria, we noticed that some political events in 2016 led Brazil and the US far from the calm waters of the political situation in Canada. Whereas the latter was experiencing a non-election year and facing no significant political scandal, the US was experiencing one the most polarized and hostile elections in its history, including political scandals involving both candidates (Faris et al., 2017; Patterson, 2016). Brazil, in turn, was struck by a huge corruption scandal known as the Car Wash (Lava Jato). This operation, receiving enthusiastic coverage in the press (de Albuquerque, 2019), greatly increased the climate of mistrust in the country. This contributed to lowering the government’s approval ratings and undermining its parliamentary support in Congress (Veiga et al., 2019). This context served as a pretext for Rousseff’s dismissal for budget misconduct, which, in turn, had no objective relationship with the Car Wash operation (Nunes & Melo, 2017). The impeachment trial was concluded by the Senate on 31 August 2016.

For establishing the level of political stability of each country, we used the Political Stability and Absence of Violence Index data provided by the World Bank (TheGlobalEconomy.com, 2020). According to this, Canada was the 11th most stable country in the world in 2016, whereas the US and Brazil were 75th and 129th, respectively. We infer from these analyses the levels of political activity and instability in Brazil were more pronounced than in the US and Canada, with the latter being in the most stable and routine-alike situation.

That being said, the decision to collect data only from Facebook was based on previous research showing that different platforms present contrasting patterns of news sharing (Bastos, 2015; Christin, 2015; Kalsnes & Larsson, 2018; Trilling et al., 2017). Since Facebook was and still is the most popular platform in the world, it seemed the most representative of online news sharing patterns.

Additionally, we selected a continuous rather than a constructed week in order to control the effect of the number of news page followers’ variance on news sharing. According to dos Santos et al. (2019), journalistic pages on Facebook tend to gain new followers in the long run, which increases the news sharing of their posts. In view of that, we selected a time frame shorter than 14 days, precisely the number of days tested by dos Santos et al. (2019). Moreover, we restricted our data sample to this time frame since Facebook changed its algorithm in 2018 (Tozetto, 2018). This change significantly reduced the reach of news pages, as demonstrated by dos Santos (2019).

Another important procedure worth clarifying is that we analyzed news posts directly from the Facebook news pages. This procedure gives all the posts almost the same level of editorial promotion, which contrasts with news websites. On news websites, one finds front pages and news articles with different levels of editorial emphasis in terms of placement, visual cues, and headline sizes (see e.g., Bastos, 2015; Boczkowski & Mitchelstein, 2012; Bright, 2016). This kind of emphasis exists on Facebook but to a far lesser degree. We captured much of it in our analysis with the “Post Format” variable (table 1), and our regression analysis treated it as a control variable (table 7). We also treated the news page as a control variable precisely because these news pages have an uneven number of followers (dos Santos et al, 2019).

Table 7
Multiple regression analyses for all variables and country’s samples

	Dependent variable
	Log shares
	Brazil (1)	Canada (2)	US (3)	All (4)
Control variables
Fox News				3.489***
				(0.172)
G1				1.147***
				(0.152)
Veja	-0.305*			0.875***
	(0.159)			(0.141)
Global News		-1.649***		-1.640***
		(0.116)		(0.121)
The New York
Times			-2.730***	0.925***
			(0.170)	(0.142)
Native video	-0.039	2.562***	0.456**	1.381***
	(0.407)	(0.204)	(0.209)	(0.144)
Bait video	-0.780**	-0.437	1.137**	-0.423**
	(0.314)	(0.301)	(0.556)	(0.204)
R2 Incremental	0.19%	31.89%	45.77%	48.04%
Content variables
Hard news	0.357	1.269***	0.211	0.483***
	(0.362)	(0.406)	(0.187)	(0.157)
Sensational news	0.960**	1.625***	0.111	0.876***
	(0.388)	(0.417)	(0.262)	(0.174)
Soft news	0.260	1.376***	-0.255	0.407**
	(0.384)	(0.418)	(0.227)	(0.171)
Political scandal	-0.259	-0.786**	0.833***	-0.041
	(0.222)	(0.320)	(0.286)	(0.143)
National Politics / Conflict	0.800***	-0.125	-0.209	0.152
	(0.241)	(0.186)	(0.200)	(0.122)
R2 Incremental	4.45%	2.59%	4.23%	2.78%
Constant	4.825***	2.805***	7.740***	3.634***
	(0.372)	(0.399)	(0.228)	(0.180)
Observations	488	777	393	1,658
R2	0.064	0.345	0.524	0.508
Adjusted R2	0.048	0.338	0.515	0.505
Residual Std. Error	1.671 (df = 479)	1.460 (df = 768)	1.414 (df = 384)	1.561 (df = 1645)
F Statistic	4.075*** (df = 8; 479)	50.551*** (df = 8; 768)	52.943*** (df = 8; 384)	141.734*** (df = 12; 1645)

• Note:
• ^*
  p<0.1;
• ^**
  p<0.05;
• ^***
  p<0.01

As previous studies show (dos Santos et al., 2019; García-Perdomo et al., 2018; Trilling et al., 2017), the number of shares on SNSs presents a log-normal distribution. Thus, to identify and compare news sharing patterns on Facebook between the surveyed countries, we treated this dependent variable using its log units in both the descriptive and regression analyses. By doing this, we could compare the performance of the independent variables in a normalized distribution¹ (figure 1).

Figure 1
Shares (log) distribution per country

In the content analysis of the 1.658 news posts, we followed Sampaio and Lycarião (2018) and, therefore, we performed several pre-tests. During these pre-tests, disagreements were addressed and fixed by refining the coding rules. An entire year elapsed until a reliable coding procedure among three coders was achieved. Table 1 shows the results for the final intercoder reliability test, conducted using a sample with 200 posts. Each variable was coded by the same coders and the reliability scores are presented for each variable.

Table 1
Operationalization of news posts content analysis

Content variable	Operationalization
Post format	Posts were coded in three possible formats: (1) photo/link (the post has a link, with or without a photo, to the news webpage); (2) bait video (the post warns a video is present in the news article); or, (3) native video (an embedded video). (Kα = 0.83; Pa = 98%).
News type	Four types of news posts were coded: (1) hard news (public affairs frame, useful information, or politically relevant events); (2) sensational news (dramatic events framed mainly at the individual level); (3) soft news (funny or curious events framed mainly at the individual level); (4) opinion (posts corresponding to opinion-oriented articles). (Kα = 0.69; Pa = 80%).
Political scandal	Dummy variable identifying whether or not the headline mentions events framed as political scandals. This includes judicial prosecutions of public authorities and leaks involving sensitive material that damages their public image. (Kα = 0.69; Pa = 94%).
Conflict / Controversy	Ordinal variable establishing the level of controversy/conflict depicted by the news post headline. Three levels were identified: (0) Absent; (1) Low (disagreements, controversial topics, political contends); (2) High (protests, judicial actions, violent actions). (Kα = 0.68; Pa = 75%).
Influence	Ordinal variable establishing the level of the political influence of the most powerful person or institution mentioned on the headline. Four levels were identified: (0) Absent; (1) Low (institutions or public authorities at the local level); (2) Intermediate (institutions or public authorities at the national level); (3) High (institutions or public authorities at the international level). (Kα = 0.81; Pa = 85%).
Proximity	Ordinal variable establishing the level of proximity of the central event, person, or institution depicted in the news post. Four levels were identified: (0) Absent/Not Discernible; (1) Low (event/person/institution related to a foreign country from Global South); (2) Intermediate (event/person/institution related to a foreign country member of NATO or OCDE); (3) High (event/person/institution from the same country as the news media outlet). (Kα = 0.86; Pa = 88%).
National politics with conflict / Controversy	This variable identifies news articles by assessing information of three of the aforementioned variables: proximity, influence, and conflict/controversy. This variable was automatically coded as present (1) when the news post was coded with high proximity, any positive level of influence (low, intermediate, or high), and any positive level of conflict/controversy (low or high). This variable was automatically coded as absent (0) when any of these three conditions was not fulfilled.

5 Hypotheses

Our empirical hypotheses were developed assuming that the content characteristics that appear the most frequently in news posts will be the most shared. Nevertheless, these characteristics are expected to vary according to the country’s political stability. Therefore, the most unstable countries are expected to offer more content toward public affairs and the sources of their political instability. Taking into account that Brazil was the most unstable country, and Canada the most stable, we formulate three empirical hypotheses as follows:

• H1: News production (H1a) and the sharing (H1b) of hard news will respect the order of stability among surveyed countries, Brazil > US > Canada, respectively.

• H2: News production (H2a) and the sharing (H2b) of news mentioning political scandals will follow the order of stability among surveyed countries.

• H3: News production (H3a) and the sharing (H3b) of news about national politics within a conflict frame will respect the order of stability among surveyed countries.

Since our theoretical model assumes that the content characteristics offered most frequently by news page posts will be the most shared by SNSs users, we propose a last empirical hypothesis:

• H4: Content characteristics of news posts will present negligible effects on news sharing across the surveyed countries.

6 Results

Of the 1.658 coded posts, 488 of them were posted on the Brazilian news pages, 777 on the Canadian, and 398 on the US’s counterparts. Table 2 shows the news page with the highest number of posts was Global News (525), followed by Veja (286), NYT (253), CBC News (252), G1 (202), and, finally, Fox News (140).

Table 2
News page results (control variable)

	News page	Total posts n (%)	Log shares (%)	SD (log shares)	Mean (log shares/post)
Brazil	G1	202 (41%)	43%	1.577	5.370
	Veja	286 (59%)	57%	1.797	5.0987
Canada	CBC News	252 (32%)	43%	1.671	4.273
	Global News	525 (68%)	57%	1.617	2.690
US	Fox News	140 (36%)	46.7%	1.467	8.057
	NYT	253 (64%)	53.3%	1.430	5.078

Across the three countries, the news pages having the highest rate of news posts also have the highest rate of shares. At the same time, the average values of (log) shares per post are higher for those news pages with the lowest post rates. Fox News obtained the highest average value (8.057) for (log) shares per post and Global News the lowest average value (2.69). Our multiple regression analysis (table 7, column/model 4) also depicts these news pages as having the most contrasting factors for news sharing across the three countries. Whereas Fox News significantly increases the chance of a news post being shared, Global News strongly decreases it.

Regarding post format, the most common category across all three countries was “picture/link”. Table 3 shows that this formal characteristic was also the most shared. In the Brazilian sample, it obtained the highest rate for (log) shares per post (5.26). In the Canadian sample, the highest average value was “native video” (5.69), whereas in the US sample it was “bait video” (8.85). Our multiple regression analysis (table 7) indicates that these higher mean values reflect higher chances of news sharing.

Table 3
Results for post format (control variable)

	Post format	Total posts n (%)	Log shares (%)	SD (log shares)	Mean (log shares/post)
Brazil	Native video	20 (4.1%)	4.0%	1.434	5.149
	Picture/link	437 (89.5%)	90.5%	1.732	5.264
	Bait video	31 (6.4%)	5.5%	1.471	4.495
Canada	Native video	56 (7.2%)	13%	1.543	5.692
	Picture/link	696 (89.6%)	85%	1.673	3.033
	Bait video	25 (3.2%)	2%	1.241	2.353
US	Native video	66 (16.8%)	20%	2.329	7.454
	Picture/link	320 (81.4%)	77%	1.806	5.809
	Bait video	7 (1.8%)	3%	1.835	8.853

Regarding content variables, the most common and most shared post type across the three countries was hard news (table 4). Sensational news, however, obtained the highest average values for (log) shares per post in the Brazilian and the US samples (5.58 and 6.77, respectively). In the Canadian sample, soft news obtained the highest average value (3.42).

Our first empirical hypothesis (H1) was that news production (H1a) and the sharing (H1b) of hard news would present the following hierarchical order: Brazil > the US > Canada. This hypothesis must be rejected for news production (H1a) since the three countries provided decidedly similar proportions of hard news to their audiences. Table 4 shows that the chi-square residuals reveal no significant differences for this category among Brazil, Canada, and the US. Moreover, news sharing (H1b) prevalence turned out to be BR > CA > US².¹ As a result, H1 must be fully rejected.

Table 4
Results for post type

	Post type	Total posts n (%)	Chi-Square Residuals	Log shares (%)	SD (log shares)	Mean (log shares/post)
Brazil	Hard news	228 (46.7%)	0.50	47.3%	1.649	5.271
	Opinion	27 (5.5%)	-1.90	5.1%	2.357	4.832
	Sensational news	111 (22.7%)	-0.72	24.4%	1.667	5.583
	Soft news	122 (25.0%)	1.16	23.2%	1.638	4.843
Canada	Hard news	361 (46.5%)	0.53	43.9%	1.704	3.023
	Opinion	14 (1.8%)	-6.08*	1.5%	1.377	2.713
	Sensational news	237 (30.5%)	3.46*	31.9%	1.865	3.353
	Soft news	165 (21.2%)	-0.74	22.7%	1.878	3.422
US	Hard news	160 (40.7%)	-1.31	42.2%	1.966	6.371
	Opinion	91 (23.2%)	10.67*	22.8%	2.128	6.052
	Sensational news	56 (14.2%)	-4.06*	15.7%	1.975	6.771
	Soft news	86 (21.9%)	-0.25	19.2%	1.870	5.390

• Pearson’s Chi-squared test x-squared = 187.9, p-value < 0.0001, df = 6

Our second empirical hypothesis (H2) was that news production (H2a) and the sharing (H2b) of news about political scandals would present the following hierarchical order: Brazil > the US > Canada. Table 5 shows that the Brazilian sample obtained the highest news supply rate on the political scandal (27.9%), the US sample obtained the intermediary rate (7.6%), and the Canadian sample obtained the lowest rate (2.8%). Since the chi-square residuals are significant for this category, H2a is confirmed. Examination of the (log) share rates shows that the aforementioned samples obtained 28%, 10%, and 2%, respectively, of their news sharing about political scandals. Thus, H2 is fully confirmed.

Table 5
Results for political scandal

	Political Scandal	Total Posts n (%)	Chi-Square Residuals	Log shares (%)	SD (log shares)	Mean (log shares/post)
Brazil	Absent	352 (72.1%)	-3.87*	72%	1.748	5.201
	Present	136 (27.9%)	10.84*	28%	1.622	5.236
Canada	Absent	755 (97.2%)	2.51	98%	1.788	3.242
	Present	22 (2.8%)	-7.04*	2%	1.488	1.870
US	Absent	363 (92.4%)	0.78	90%	2.004	6.007
	Present	30 (7.6%)	-2.18*	10%	1.628	7.740

• Pearson’s Chi-squared test x-squared = 193.9, p-value < 0.0001, df = 2

Our third empirical hypothesis (H3) was that both news production (H3a) and the sharing (H3b) of news about national politics within a conflict frame would present the following hierarchical order among the samples: Brazil > US > Canada. According to the data shown in table 6, the news production and sharing rates for this category followed this hypothesis. The Brazilian sample presented 30% of its news supply and 31% of its news sharing in this category. The US sample presented 24% and 27%, respectively, and the Canadian sample, 10%, and 9%, respectively. Since the chi-square residuals are significant, H3 is also fully confirmed.

Table 6
Results for national politics with a Conflict / Controversy Frame

	National politics (conflict)	Total posts n (%)	Chi-Square Residuals	Log shares (%)	SD (log shares)	Mean (log shares/post)
Brazil	Absent	343 (70%)	-2.58*	69%	1.682	5.095
	Present	145 (30%)	5.31*	31%	1.758	5.485
Canada	Absent	698 (90%)	2.79*	91%	1.815	3.237
	Present	79 (10%)	-5.73*	9%	1.582	2.907
US	Absent	299 (76%)	-1.04	73%	1.967	5.894
	Present	94 (24%)	2.14*	27%	2.037	6.921

• Pearson’s Chi-squared test x-squared = 81.3, p-value < 0.0001, df = 2

The fourth and final empirical hypothesis (H4) was that content characteristics of news posts would present negligible effects on news sharing across all three surveyed countries. Aiming to establish a threshold that identifies an independent content effect on news sharing, we performed a multiple regression analysis (table 7). Examination of the results for the overall model (column 4) revealed the most powerful and significant effects were found in the control variables, specifically, Fox News (a positive effect), Global News (a negative effect), and native video (a positive effect). Moreover, the overall model shows that control variables explain 48.8% of the news sharing variation, whereas content variables explain only 2.78%.

Observation of the model based only on the Brazilian sample (column 1) shows that the most powerful and significant effects are found in the content variables, specifically in sensational news and in national politics within a conflict frame (both present a positive effect on news sharing). Nonetheless, the content variables explain only 4.45% of news sharing variation. As a result, model 1 (based on the Brazilian sample) is very weak and does not fit the news sharing variation well, compared to the other three models.

Model 2 (based on the Canadian sample) presents its most powerful and significant effects in the control variables, specifically, on native video (a positive effect) and Global News (a negative effect). Additionally, whereas the control variables explain 31.89% of news sharing variation, content variables explain less than 3%.

The third model (based on the US’s sample), also here we realize that the most powerful and significant effects are located on control variables, namely on NYT (a negative effect) and on Bait Video (positive). Just as the Canadian model, this 3rd model also has a contrasting explanatory power between control and content variables. Whereas the former can explain 45.5% of news sharing variation, the latter explains only 4.23%.

All in all, these results show that content variables explain very little of the news sharing variation and, as such, their independent effects on news sharing are negligible. Therefore, the fourth and last empirical hypothesis (H4) is also confirmed.

Table 7 - Multiple regression analyses for all variables and country’s samples

7 Discussion

This paper tested four empirical hypotheses based on the research hypothesis that news routines and, consequently, news sharing patterns on SNSs are influenced by the levels of political activity and instability. To do this, we compared samples of three countries, Brazil, Canada, and the US, with contrasting levels of these aspects. Our first hypothesis (H1) was that hard news would correspond to the level of stability among the surveyed countries. However, this was not the case. We found no significant differences among the three samples in this regard because all the samples presented high levels of hard news on news production and sharing, without significant differences.

Exploring the overall results, and taking previous studies into account, we found the H1 results suggest that hard news might be more stable and regular than expected. This could be the case, especially for legacy news media, as Steiner (2020) showed in her study of different types of news media in Germany. Legacy media are more reluctant than tabloid media to soften their news to be accommodated on Facebook. Moreover, Martin (2019), in her aggregate analysis of 160 English-language news sites, found that news pieces about public affairs are at the top of online news sharing. Thus, it is plausible that lower levels of political stability do not affect hard news supply per se, but rather the objects of its attention. This is precisely what the results for H2 and H3 suggest.

In this regard, the second empirical hypothesis (H2) was that news about political scandals would be more evident in the news production and sharing from countries with higher levels of political instability and activity. This is based on the assumption that political scandals have become one of the main sources of contemporary political instability.

In the Brazilian case, this helped to create, through the long-standing Car Wash (Lava Jato) operation, conditions for the impeachment of the democratically-elected President, Dilma Rousseff. In the US case, scandals during the 2016 national presidential campaign, involving both candidates, led to the coverage being dominated by partisan and negative news (Faris et al., 2017; Patterson, 2016).

Since the data concerning the differences among the surveyed countries confirmed the H2 hypothesis, our research provides original evidence that political scandals can affect online news routines and sharing. Moreover, as our results for H1 suggest, this effect does not necessarily change the supply and sharing of hard news. Rather, it shifts the foci of attention which become more oriented toward the conflicts and events that are causing the political instability and activity.

This view appears more robust when one compares the supply and sharing of national news within a conflict/controversy frame. In this respect, H3 was that news having this characteristic would be more common in the news production and sharing in more unstable and politically heightened environments. Confirming this hypothesis, our research provides additional evidence to previous studies that pointed to disruptive political events as a key factor to explain both news routines and patterns of news sharing on digital platforms (Boczkowski & Mitchelstein, 2012; Lycarião & Leite, 2020; Ørmen, 2019).

Moreover, our research hypothesis assumed that political context influences both news routines and news sharing on SNSs, simultaneously. This is based on the well-established tradition of journalism studies showing that both media professionals and news users scan the social environment using similar criteria for relevance, which makes their reaction to world events very similar (Eilders, 2006; Neuman et al., 2014). As a result, our fourth empirical hypothesis (H4) was that the content characteristics of news posts would present negligible effects on news sharing across the surveyed countries.

Results confirmed this hypothesis by showing that news routines and formal variables are more relevant to explain news sharing than are content variables. This is in line with previous studies that have demonstrated the strong impact that news outlets characteristics (Karnowski et al., 2020), as well as other aspects of online news routines, like the boosting of videos, and other visual cues produced on news sharing (Bastos, 2015; Boczkowski & Mitchelstein, 2012; Bright, 2016; dos Santos et al., 2019; Salgado & Bobba, 2019).

This phenomenon can be traced back to studies that investigated the content characteristics associated with recipients’ news selection and information recall from reading newspapers. In this regard, Eilders (1997) found that content variables (news factors) could not explain these aspects independently, but only in combination with editorial features like placement, size, and length of news articles (formalen Beitragsmerkmale).

This is because the news articles selected the most frequently were those with higher levels of editorial emphasis. Not incidentally, these same articles showed higher rates of news factors. Nonetheless, regression analysis showed that editorial emphasis explained between 14 and 21% of the variation in news selection. This is much higher than the content factors, which could explain only 1% of this variation (Eilders, 1997, pp. 206–207).

In view of this, it is possible to propose an explanatory framework for our research question, which is whether (and how) political context influences the online news routines and shareworthiness patterns of countries with varying levels of political stability.

In this regard, political context seems to influence the online news routines and shareworthiness of countries according to their level of political stability and activity. Important effects are expected when they present (a) contrasting levels of stability due to political scandals, and (b) distinct levels of political activity due to election campaigns. Moreover, this influence does not necessarily dramatically change the supply of hard news and public affairs articles. Rather, it shifts its foci of attention, which turns more toward the sources of the instability (i.e., political scandals) and of the political activity (i.e., election campaigns). Thus, when national realities present contrasting levels of political stability and activity, we argue that they present an online news sharing gap, precisely on these foci of attention.

8 Conclusion

Current literature on online news sharing presents a range of approaches, methods, and results. Our literature review pointed out that some studies seem to contradict each other in their conclusions about online news sharing logic. To disentangle this, we applied a unified theoretical and methodological framework that allowed us to express the importance of socio-political context within a broader understanding of the phenomenon.

We have been able to provide evidence that social-political context can alter public attention in a way that affects both online news routines and audiences simultaneously. We argue that analyzing only content characteristics for explaining online news sharing should be regarded as a serious blind spot in an important part of the current scholarship on the subject. We have seen that, even when content variables are statistically significant, the scale of what they explain is negligible in comparison with news page characteristics (control variables).

Thus, in addition to its unified research framework, this paper also contributes to current scholarship by showing how political context can be regarded as an important factor for shareworthiness. After all, if news editors believe that they need to soften or sensationalize their news supply to conform to Facebook’s algorithm, they run the risk of losing much of their potential reach, especially in contexts of political instability and high activity.

Nonetheless, broader, updated, and more diversified empirical analyses of countries having varied levels of political competitiveness and stability would allow testing of how far this understanding could travel. In this regard, an important procedure would be to replicate the Boczkowski e Mitchelstein (2012) research design of comparing a society at different moments, contrasting “heightened” political moments with “regular” ones.

NOTES

• 1
  The codebook, coded data, and the R Markdown script, along with the analyses are available at the following links: <https://www.doi.org/10.6084/m9.figshare.16702417> (Codebook – pdf) | <https://doi.org/10.6084/m9.figshare.16702438.v1> (Coded data – csv) | <https://doi.org/10.6084/m9.figshare.16702456.v1> (Script – Rmd)
• 2
  To the best of our knowledge, the proper way to test the “b” component of the first three hypotheses is to use percentages rather than mean values. First, mean values cannot be used directly to compare the news sharing of the three countries since they have very different values from each other. Furthermore, our fourth hypothesis tests the understanding that news sharing behaves, as a rule, as a function of the content offered by the news supply. This implies that news sharing would have negligible effects as a phenomenon, independent of this supply. As this hypothesis is confirmed, it makes perfect statistical and theoretical sense, in our view, to test the first three hypotheses as a function of percentages rather than means.

• ENGLISH REVISION: PETER LASPINA
• One review used in the evaluation of this article can be accessed at: https://osf.io/cavx2/ | Following BJR’s open science policy, the reviewers authorized this publication and the disclosure of his/her names.

REFERENCES

• Bastos, M. T. (2015). Shares, pins, and tweets: News Readership from Daily Papers to Social Media. Journalism Studies, 16(3), 305–325. DOI: 10.1080/1461670X.2014.891857
  » https://doi.org/10.1080/1461670X.2014.891857
• Boczkowski, P. J., & Mitchelstein, E. (2012). How Users Take Advantage of Different Forms of Interactivity on Online News Sites: Clicking, e-Mailing, and Commenting. Human Communication Research, 38(1), 1–22. DOI: 10.1111/j.1468-2958.2011.01418.x
  » https://doi.org/10.1111/j.1468-2958.2011.01418.x
• Bright, J. (2016). The Social News Gap: How News Reading and News Sharing Diverge. Journal of Communication, 66(3), 343–365. DOI: 10.1111/jcom.12232
  » https://doi.org/10.1111/jcom.12232
• Christin, A. (2015). “Sex, Scandals, and Celebrities”? Exploring the Determinants of Popularity in Online News. Sur le journalisme, 4(2), 28–47. Retrieved from www.surlejournalisme.kinghost.net/rev/index.php/slj/article/view/215
  » www.surlejournalisme.kinghost.net/rev/index.php/slj/article/view/215
• de Albuquerque, A. (2019). Protecting Democracy or Conspiring against it? Media and Politics in Latin America: A Glimpse from Brazil. Journalism, 20(7), 906–923. DOI: 10.1177/1464884917738376
  » https://doi.org/10.1177/1464884917738376
• dos Santos, M. A., Jr. (2019). Desarranjo da visibilidade, desordem informacional e polarização no Brasil entre 2013 e 2018 [doctoral dissertation, Universidade Federal Fluminense]. Repositório institucional UFF.
• dos Santos, M. A., Jr., Lycarião, D., & Aquino, J. A. (2019). The Virtuous Cycle of News Sharing on Facebook: Effects of Platform Affordances and Journalistic Routines on News Sharing. New Media & Society, 21(2), 398-418. DOI:10.1177/1461444818797610
  » https://doi.org/10.1177/1461444818797610
• Eilders, C. (1997). Nachrichtenfaktoren und Rezeption: eine empirische Analyse zur Auswahl und Verarbeitung politischer Information Springer-Verlag. Retrieved from www.springer.com/de/book/9783531129136
  » www.springer.com/de/book/9783531129136
• Eilders, C. (2006). News Factors and News Decisions. Theoretical and Methodological Advances in Germany. Communications, 31(1), 5–24. DOI: 10.1515/COMMUN.2006.002
  » https://doi.org/10.1515/COMMUN.2006.002
• Faris, R., Roberts, H., Etling, B., Bourassa, N., Zuckerman, E., & Benkler, Y. (2017). Partisanship, Propaganda, and Disinformation: Online Media and the 2016 US Presidential Election Berkman Klein Center for Internet & Society Research Paper. Retrieved from http://nrs.harvard.edu/urn-3:HUL.InstRepos:33759251
  » http://nrs.harvard.edu/urn-3:HUL.InstRepos:33759251
• García-Perdomo, V., Salaverría, R., Kilgo, D. K., & Harlow, S. (2018). To Share or not to Share: The Influence of News Values and Topics on Popular Social Media Content in the United States, Brazil, and Argentina. Journalism Studies, 19(8), 1180–1201. DOI: 10.1080/1461670X.2016.1265896
  » https://doi.org/10.1080/1461670X.2016.1265896
• Ihm, J., & Kim, E. M. (2018). The Hidden Side of News Diffusion: Understanding Online News Sharing as an Interpersonal Behavior. New Media & Society, 20(11), 4346–4365. DOI: 10.1177/1461444818772847
  » https://doi.org/10.1177/1461444818772847
• Kalsnes, B., & Larsson, A. O. (2018). Understanding News Sharing Across Social Media: Detailing Distribution on Facebook and Twitter. Journalism Studies, 19(11), 1669–1688. DOI: 10.1080/1461670X.2017.1297686
  » https://doi.org/10.1080/1461670X.2017.1297686
• Karnowski, V., Leiner, D. J., Kümpel, A. S., & Leonhard, L. (2020). Worth to Share? How Content Characteristics and Article Competitiveness Influence News Sharing on Social Network Sites. Journalism & Mass Communication Quarterly, 98(1), 59–82. DOI: 10.1177/1077699020940340
  » https://doi.org/10.1177/1077699020940340
• Kim, E. M., & Ihm, J. (2020). More than Virality: Online Sharing of Controversial News with Activated Audience. Journalism & Mass Communication Quarterly, 97(1), 118–140. DOI: 10.1177/1077699019836950
  » https://doi.org/10.1177/1077699019836950
• Kim, J., Namkoong, K., & Chen, J. (2020). Predictors of Online News-Sharing Intention in the US and South Korea: An Application of the Theory of Reasoned Action. Communication Studies, 71(2), 315–331. DOI: 10.1080/10510974.2020.1726427
  » https://doi.org/10.1080/10510974.2020.1726427
• Lischka, J. A. (2018). Logics in Social Media News Making: How Social Media Editors Marry the Facebook Logic with Journalistic Standards. Journalism, 22(2), 1–18. DOI: 10.1177/1464884918788472
  » https://doi.org/10.1177/1464884918788472
• Lischka, J. A., & Werning, M. (2017). Wie Facebook den Regionaljournalismus verändert: Publikumsund Algorithmusorientierung bei der Facebook-Themenselektion von Regionalzeitungen. Kommunikation @ gesellschaft, 18(2), 1–25. DOI: 10.15460/kommges.2017.18.2.583
  » https://doi.org/10.15460/kommges.2017.18.2.583
• Lycarião, D., & Leite, A. B. (2020). Política no Facebook: Emergência de novos padrões de compartilhamento de notícias em tempos de crise. E-Compós, v.23, 1–26. DOI: 10.30962/ec.1821
  » https://doi.org/10.30962/ec.1821
• Martin, F. (2019). What We Share: Genre and Topicality on Facebook and Twitter. In F. Martin & T. Dwyer (Eds.), Sharing News Online: Commendary Cultures and Social Media News Ecologies (pp. 129–156). Springer-Verlag. DOI: 10.1007/978-3-030-17906-9
  » https://doi.org/10.1007/978-3-030-17906-9
• Neuman, W. R., Guggenheim, L., Jang, S. M., & Be, S. Y. (2014). The Dynamics of public attention: Agenda-setting theory meets big data. Journal of Communication, 64(2), 1936–2214. DOI: 10.1111/jcom.12088
  » https://doi.org/10.1111/jcom.12088
• Newman, N., Fletcher, R., Kalogeropoulos, A., & Nielsen, R. (2019). Reuters Institute Digital News Report 2019 Reuters Institute for the Study of Journalism. Retrieved from https://reutersinstitute.politics.ox.ac.uk/sites/default/files/2019-06/DNR_2019_FINAL_0.pdf
  » https://reutersinstitute.politics.ox.ac.uk/sites/default/files/2019-06/DNR_2019_FINAL_0.pdf
• Nunes, F., & Melo, C. R. (2017). Impeachment, Political Crisis and Democracy in Brazil. Revista de Ciencia Política, 37(2), 281–304. DOI: 10.4067/s0718-090x2017000200281
  » https://doi.org/10.4067/s0718-090x2017000200281
• Ørmen, J. (2019). From Consumer Demand to User Engagement: Comparing the Popularity and Virality of Election Coverage on the Internet. The International Journal of Press/Politics, 24(1), 49–68. DOI: 10.1177/1940161218809160
  » https://doi.org/10.1177/1940161218809160
• Patterson, T. T. E. (2016). News Coverage of the 2016 General Election: How the Press Failed the Voters Harvard Kennedy School. DOI: 10.2139/ssrn.2884837
  » https://doi.org/10.2139/ssrn.2884837
• Salgado, S., & Bobba, G. (2019). News on Events and Social Media: A Comparative Analysis of Facebook Users’ Reactions. Journalism Studies, 20(15), 2258–2276. DOI: 10.1080/1461670X.2019.1586566
  » https://doi.org/10.1080/1461670X.2019.1586566
• Sampaio, R., & Lycarião, D. (2018). Eu quero acreditar! Da importância, formas de uso e limites dos testes de confiabilidade na Análise de Conteúdo. Revista de Sociologia e Política, 26(66), 31–47. DOI: 10.1590/1678-987318266602
  » https://doi.org/10.1590/1678-987318266602
• Steiner, M. (2020). Soft Presentation of Hard News? A Content Analysis of Political Facebook Posts. Media and Communication, 8(3), 244–257. DOI: 10.17645/mac.v8i3.3152
  » https://doi.org/10.17645/mac.v8i3.3152
• TheGlobalEconomy.com. (2020). Political stability – Country rankings Retrieved from www.theglobaleconomy.com/rankings/wb_political_stability/
  » www.theglobaleconomy.com/rankings/wb_political_stability/
• Tozetto, C. (2018, January 13). Facebook muda algoritmo e reduz alcance de notícias Estadão. Retrieved from https://link.estadao.com.br/noticias/empresas,facebook-muda-algoritmo-e-reduz-alcance-de-noticias,70002149049
  » https://link.estadao.com.br/noticias/empresas,facebook-muda-algoritmo-e-reduz-alcance-de-noticias,70002149049
• Trilling, D., Tolochko, P., & Burscher, B. (2017). From Newsworthiness to Shareworthiness: How to Predict News Sharing Based on Article Characteristics. Journalism & Mass Communication Quarterly, 94(1), 38–60. DOI: 10.1177/1077699016654682
  » https://doi.org/10.1177/1077699016654682
• Valenzuela, S., Piña, M., & Ramírez, J. (2017). Behavioral Effects of Framing on Social Media Users: How Conflict, Economic, Human Interest, and Morality Frames Drive News Sharing. Journal of Communication, 67(5), 803–826. DOI: 10.1111/jcom.12325
  » https://doi.org/10.1111/jcom.12325
• Veiga, L. F., Dutt-Ross, S., & Martins, F. B. (2019). Os efeitos da economia e da Operação Lava-Jato na popularidade da Presidente Dilma Rousseff no período pré-impedimento. Revista de Sociologia e Política, 27(72), 1–21. DOI: 10.1590/1678-987319277202
  » https://doi.org/10.1590/1678-987319277202

Desk Review Editor: Lia Seixas

Publication Dates

• Publication in this collection
  15 Apr 2022
• Date of issue
  Sep-Dec 2021

History

• Received
  16 May 2021
• Reviewed
  21 June 2021
• Reviewed
  06 Aug 2021
• Accepted
  20 Sept 2021